J Clin Periodontol 2003; 30: 671681 Copyright r Blackwell Munksgaard 2003

Printed in Denmark. All rights reserved

Review Paper
Influence of sex hormones on the Paulo Mascarenhas, Ricardo Gapski,
Khalaf Al-Shammari and
Hom-Lay Wang

periodontium Department of Periodontics/Prevention/

Geriatrics, School of Dentistry, University of
Michigan, Ann Arbor, MI, USA

Mascarenhas P, Gapski R, Al-Shammari K, Wang H-L: Influence of sex hormones on

the periodontium. J Clin Periodontol 2003; 30: 671681. rBlackwell Munksgaard,
2003.

Abstract
Objectives: Sex hormones have long been considered to play an influential role on
periodontal tissues, bone turnover rate, wound healing and periodontal disease
progression. The objectives of this review article are to (1) address the link between
sex hormones and the periodontium, (2) analyse how these hormones influence the
periodontium at different life times and (3) discuss the effects of hormone
supplements/replacement on the periodontium.
Materials and Methods: Two autonomous searches were performed in
English language utilizing Medline, Premedline and Pubmed as the online
databases. Publications up to 2002 were selected and further reviewed. In addition,
a manual search was also performed including specific related journals
and books.
Results: It is certain that sexual hormones play a key role in periodontal disease
progression and wound healing. More specifically, these effects seem to differentiate
by gender as well as lifetime period. In addition, the influence of sex hormones can be
minimized with good plaque control and with hormone replacement. Key words: hormones; estrogen; periodontium;
Conclusion: Despite profound research linking periodontal condition with sex pregnancy; periodontal diseases
hormones kinetics, more definitive molecular mechanisms and therapy still remain to
be determined. Accepted for publication 3 September 2002

Bacterial plaque has been established as
the primary etiologic factor for the
initiation of periodontal disease (Loe
et al. 1965). However, it has also been
shown that without a susceptible host
the periodontal pathogens are necessary
but not sufficient for disease to occur.
Hence, the systemic factors/conditions
of the host must be understood since
they may affect disease prevalence,
progression, and severity (Lang et al.
1983). Among these, sexual hormones
have been suggested as important mod-
ifying factors that may influence the
pathogenesis of periodontal diseases
(Mariotti 1994, Parkar et al. 1998b,
Hofbauer & Heufelder 2001).
Hormones are specific regulatory mo-
lecules that modulate reproduction,
growth and development, maintenance
of the internal environment, as well as
energy production, utilization, and sto-
rage (Mariotti 1994). Hormonal effects
reflect physiological/pathological changes
in almost all types of tissues of the body.
Targets for a number of hormones such
as androgens, estrogen, and progesterone
have also been localized in periodontal
tissues (Gornstein et al. 1999). Conse-
quently, systemic endocrine imbalances
may have an important impact in period-
ontal pathogenesis.
Researchers have shown that changes
in periodontal conditions might be asso-
ciated with variations in sex hormone
levels. This association is evident in the
recent periodontal disease classification
(Armitage 1999), which includes the
following hormone-related disease cate-
gories: puberty-associated gingivitis,
menstrual cycle-associated gingivitis and
pregnancy-associated gingivitis. There-
fore, the goal of this review paper is to
discuss how sex hormones may influence
the periodontium and periodontal wound/
bone healing.
Steroid Sex Hormones
Because of their complexity and diver-
sity, hormones are difficult to arrange
into discrete groups, although they can
be divided into four subgroups based
upon their chemical structure steroids,
glycoproteins, polypeptides, and amines.
Steroid sex hormones are derived
from cholesterol and as a common
structure they have three rings of six
carbon atoms. They are believed to play
an important role in the maintenance of
the skeletal integrity, including the
alveolar bone. The steroid sex hormones
such as estrogen and estradiol have been
672 Mascarenhas et al.
known for their effect on bone mineral
metabolism. Other bone turnover-re-
lated hormones include progesterone,
testosterone and dihydrotestosterone,
androstenedione, dihydroepiandrostene-
dione, and sex hormone-binding globu-
lin (Katz & Epstein 1993). Among
these, estrogens, progesterone, and tes-
tosterone have been most linked with
periodontal pathogenesis and therefore
will be described in detail in the paper.
Estrogen and progesterone
Estrogen and progesterone are responsi-
ble for physiological changes in women
at specific phases of their life, starting in
puberty. Estrogen induces several of the
pubertal developmental changes in fe-
males, and progesterone acts synergisti-
cally with estrogen to control the
menstrual cycle and to inhibit follitroppin
secretion by the anterior pituitary gland
(Amar & Chung 1994). Both hormones
are also known to promote protein
anabolism and growth (Soory 2000a).
Both hormones have significant bio-
logical actions that can affect different
organ systems including the oral cavity
(Lundgren 1973a, b, Lundgren et al.
1973, Lopatin et al. 1980, Pack &
Thomson 1980, Sooriyamoorthy &
Gower 1989a, Ojanotko-Harri et al.
1991, Mariotti 1994). Specifically, es-
trogens can influence the cytodifferen-
tiation of stratified squamous epithelium
as well as the synthesis and mainte-
nance of fibrous collagen (Amar &
Chung 1994). Estrogen receptors found
in osteoblast-like cells provide a me-
chanism for the direct action on bone
(Klinger & Sommer 1978, Aufdemorte
& Sheridan 1981, Eriksen et al. 1988,
Komm et al. 1988). These receptors
were also located in periosteal fibro-
blasts, scattered fibroblasts of the lami-
na propria (Aufdemorte & Sheridan
1981), and periodontal ligament (PDL)
fibroblasts (Nanba et al. 1989a), proving
the direct action of sex hormones on
different periodontal tissues.
Clinically, estrogen-sufficient pa-
tients have been reported to have more
periodontal plaque without increased
gingival inflammation when compared
to patients with deficient levels of
estrogens (Reinhardt et al. 1999). This
suggests that inflammatory mediators
may be affected by estrogen hormone
level, which may be attributed to the
production of prostaglandins (PGs) by
the involvement of estradiol and pro-
gesterone. It is, therefore, speculated
that normal circulating estrogen levels estingly, the number of receptors in
might be essential for periodontal pro- fibroblasts tends to increase in inflamed
tection. In fact, the amount of circulat- or overgrown gingiva (Ojanotko et al.
ing estradiol seems to be inversely 1980). It is believed that an increasing
correlated with the prevalence of peri- matrix synthesis occurs on periodontal
odontal disease (Plancak et al. 1998). cells under testosterone influence (Oja-
Other effects of the estrogens on the notko et al. 1980, Kasperk et al. 1989,
periodontium are listed in Table 1. Sooriyamoorthy & Gower 1989a).
Progesterone is another sex hormone Testosterone has also been associated
that has also been demonstrated to have with bone metabolism, playing a role in
direct effects on the periodontium the maintenance of bone mass (Morley
(Table 2). Experimental, epidemiologic, 2000). A study performed on a group of
and clinical data have demonstrated that men who were castrated for sexual
progesterone is active in bone metabo- offences showed that bone density
lism and may play an important role in suffered a rapid decrease that was
the coupling of bone resorption and sustained for a number of years after
bone formation (Dequeker et al. 1977, castration (Stepan et al. 1989). Kasperk
Dequeker & De Muylder 1982, Lobo et et al. (1997) observed that both gonadal
al. 1984, Gallagher et al. 1991). Studies androgen dihydrotesterone (DHT) and
have shown that progesterone may exert adrenal androgen dehydroepiandroster-
its action directly on bone by engaging one (DHEA) have a positive impact on
osteoblast receptors or indirectly by bone metabolism, by stimulating bone
competing for a glucocorticoid receptor cell proliferation and differentiation. A
(Feldman et al. 1975, Chen et al. 1977). very effective way to analyze the effect
of androgens on bone metabolism is the
evaluation of the presence of biochem-
Androgens (testosterone)
ical markers of bone remodeling on
Androgens are hormones responsible bone tissue under the influence of those
for masculinization. Testosterone can hormones. One of the bone remodeling
be produced in the adrenal cortex, markers that has been used for this
although the one from the testes is the objective is osteoprotegerin (OPG),
most active form (Ganong 1997). Its which is a secreted decoy receptor that
secretion is regulated by ACTH and by inhibits osteoclast formation and activa-
pituitary adrenal androgen-stimulating tion by neutralizing its cognate ligand
hormone. The adrenal androgen andros- (Kong et al. 1999, Teitelbaum 2000).
tenedione is converted to testosterone This OPG action has been associated
and to estrogens in the circulation and re- with a reduction in the loss of bone
presents an important source of estrogens mineral density that is observed during
in men and in postmenopausal women. periodontal disease progression (Kong
Specific receptors for this hormone et al. 1999). Szulc et al. (2001) found
have been isolated in the periodontal that serum concentrations of OPG
tissues (Wilson & Gloyna 1970). Inter- increased significantly with age, and
Table 1. Effects of estrogen in the Periodontium
 increased amount of plaque with no increase of gingival inflammation (Reinhardt et al. 1999)
 inhibit proinflammatory cytokines release by human marrow cells (Gordon et al. 2001)
 reduce T-cell-mediated inflammation (Josefsson et al. 1992)
 suppress leukocyte production from the bone marrow (Josefsson et al. 1992, Cheleuitte et al.
1998)
 inhibit PMN chemotaxis (Ito et al. 1995)
 stimulate PMN phagocytosis (Hofmann et al. 1986)
Table 2. Effects of progesterone in the periodontium
 increase production of prostaglandins
(self-limiting process) (ElAttar 1976b, Smith et al. 1986)
 increase polymorphonuclear leukocytes and PGE2 in the GCF
 reduce glucocorticoid anti-inflammatory effect (Feldman et al. 1975, Chen et al. 1977)
 altered collagen and noncollageneous protein synthesis (Willershausen et al. 1991)
 alter PDL fibroblast metabolism (Nanba et al. 1989b, Sooriyamoorthy & Gower 1989b,
Tilakaratne & Soory 1999a, b)
 increase vascular permeability (Abraham-Inpijn et al. 1996)

were positively correlated with free
testosterone index and free estradiol
index. They concluded that age as well
as androgen and estrogen status are
significant positive determinants,
whereas parathyroid hormone (PTH) is
a negative determinant of OPG serum
levels in men (Szulc et al. 2001). These
data suggest that OPG may be an
important paracrine mediator of bone
metabolism in elderly men and high-
light the role of androgens in the
homeostasis of the male skeleton.
Studies have also examined how the
function of these hormones is controlled
or regulated in the periodontium, look-
ing specifically at the influence of
different growth factors on the stimula-
tion of DHT synthesis. Kasasa & Soory
(1995) found significant stimulation of
DHT synthesis by insulin-like growth
factor (IGF) in gingiva and cultured
fibroblasts. This finding suggests a
possible mechanism of mediating in-
flammatory repair via the androgen
metabolic pathway. The same authors
later investigated the effects of inter-
leukin-1 (IL-1) on the metabolism of
androgens from chronically inflamed
human gingival tissue (HGT) and
PDL. In response to IL-1, HGT demon-
strated a two-fold increase in DHT
synthesis and a 3.5-fold increase in 4-
androstenedione formation over control
gingival tissue; the PDL showed a 9-
fold increase in DHT synthesis in
response to IL-1 and a 6-fold increase
in 4-androstenedione formation over
control ligament tissue. The observation
of increased androgen metabolic capa-
city of PDL over HGT in response to
IL-1 insult might be relevant to repair
processes during inflammatory period-
ontal disease (Kasasa & Soory 1996).
Androgens also have a reciprocal
effect on other important mediators of
inflammation, more specifically on IL-6.
This cytokine plays a major role in
tissue damage during periodontal dis-
eases, and is secreted by many cell
types, including oral fibroblasts. In
1998, Parkar et al. investigated whether
DHT affects the expression and regula-
tion of IL-6 in gingival fibroblasts.
Using ELISA, they observed that in-
creasing DHT concentrations progres-
sively reduced IL-6 production by
gingival cells from both normal indivi-
duals and patients with gingival inflam-
mation and gingival hyperplasia (Parkar
et al. 1998a). Similar results were also
reported by Gornstein et al. (1999).
They found androgen receptors to be
Sex hormones and periodontium
Table 3. Effects of androgens in the periodontium
 stimulate matrix synthesis by osteoblasts and periodontal ligament fibroblasts (Ojanotko et al.
1980, Kasperk et al. 1989, Sooriyamoorthy & Gower 1989c)
 stimulate osteoblast proliferation and differentiation (Kasperk et al. 1997, Morley 2000)
 reduce IL-6 production during inflammation (Parkar et al. 1998b, Gornstein et al. 1999)
 inhibit prostaglandin secretion (ElAttar et al. 1982)
 Enhance OPG concentration (Szulc et al. 2001)
present in both human gingival and
periodontal ligament fibroblasts, and
reduced the production of IL-6 in the
presence of androgens. It was suggested
that elevated levels of androgens, more
specifically testosterone and dihydrotes-
tosterone, could affect the stromal cell
response to an inflammatory challenge
through downregulation of IL-6 produc-
tion.
An in vitro study analyzed the
relationship between various concentra-
tions of male testosterone and the
formation of radioactive PGs from
arachidonic acid by gingival homoge-
nate (ElAttar et al. 1982). They reported
that testosterone had inhibitory effects
in the cyclooxygenase pathway of
arachidonic acid metabolism in the
gingiva, and speculated that this sex
hormone may have anti-inflammatory
effects in the periodontium. These
steroids can exert an anabolic effect on
the tissues even when their anabolic
capacity is decreased, as is the case
during inflammation. These findings
support the concept that androgens
may have a limiting effect on period-
ontal inflammation during periodontal
disease progression. From the research
reported above, it can be concluded that
the production of androgens is stimu-
lated by the presence of proinflamma-
tory cytokines commonly found on
periodontally diseased tissues and is
downregulated by proinflammatory cy-
tokines concentration as well as the
concentration of prostaglandins.
Table 3 lists the effects of androgens
on the periodontium. Overall, androgens
may protect the periodontium via a
positive anabolic effect on periodontal
cells, a negative effect on the production
and presence of mediators of inflamma-
tion, and an inhibitory effect on osteo-
clastic function.
Factors Influencing Sex Hormone
Effects on the Periodontium
Several factors may also influence how
sex hormones affect the periodontium.
673
These include gender, age, and hormone
supplements. These interactions will be
thoroughly discussed in the following
sections.
Gender
It is well understood that gender plays
an important role in changes of the bone
density throughout the entire skeleton. It
is also known that women are much
more affected than men (e.g. osteoporo-
sis). Lau et al. (2001) reported that 80%
of the osteoporotic patients are female,
correlating with the higher frequency of
hip fractures in females, who are also
more likely to experience hormonal
imbalance throughout their lives than
males. In addition, when the influence
of gender on periodontal disease was
studied, females were considered for
several years to be more affected than
males (Marques et al. 2000, Novaes Jr
& Novaes 2001), although contradicting
data have been reported (Novaes Jr et al.
1996, Ship & Beck 1996, Albandar et
al. 1999, Novaes Jr & Novaes 1999,
Pihlstrom 2001, Yuan et al. 2001a, b).
This disparity seems to be simply
correlated with the fact that females
are more likely to seek dental care than
males (Hart et al. 1992, Novaes Jr &
Novaes 2001). One observation that
supports this notion is the presence of
the same quantity and quality of sub-
gingival bacteria, which is the most
important risk factor for periodontal
disease, in both men and women
(Schenkein et al. 1993). Overall, the
similarities/differences of periodontal
pathogenesis among different sexes still
requires much clarification. The role of
the sex hormones in wound healing
raises another question of how the
endocrine system influences this dis-
ease, since sexual hormones differ with
gender.
Regarding periodontal anatomic dif-
ferences between genders, when the
shape and height of the residual alveolar
ridge were compared, it was found that
the residual ridge in women is lower
than that in men (Hirai et al. 1993). This
674 Mascarenhas et al.

might be associated with the decreased
amount of circulating estrogen found in
women during menopause, since this
condition is associated with a higher
frequency of alveolar bone height loss,
as well as decreased crestal and sub-
crestal bone density (Payne et al. 1999).
Those effects seem to be correlated with
hormone concentrations and not to the
increased susceptibility of the period-
ontal tissues, since it has been shown
that males and females tend to have the
same amount of receptors for sex
hormones in periodontal tissues (South-
ren et al. 1978).
Table 4. Clinical and microbiologic changes in the periodontium
During puberty
 enhanced blood circulation in the end terminal capillary loops and associated increased
prevalence of gingivitis/bleeding tendency (Muhlemann 1948, Massler et al. 1950,
Curilovic et al. 1958, Sutcliffe 1972, Daniell 1983)
 higher bacterial counts (especially Prevotella intermedia (Pi) and Capnocytophaga species)
(Kornman & Loesche 1982, Mombelli et al. 1990, Mariotti 1994)
During pregnancy
 increased tendency for gingivitis and larger gingival probing depths (Loe & Silness 1963,
Silness & Loe 1964, Miyazaki et al. 1991, Robinson & Amar 1992, Machuca et al. 1999,
Soory 2000a) and periodontitis (Robinson & Amar 1992)
 increased susceptibility to infection (Cohen et al. 1969, Brabin 1985)
 decreased neutrophil chemotaxis and depressed antibody production (Sooriyamoorthy &
Gower 1989b, Raber-Durlacher et al. 1991, Raber-Durlacher et al. 1993)
 increased numbers of periodontopathogens (especially Porphyromonas gingivalis and Pi)
(Kornman & Loesche 1980, Tsai & Chen 1995)
 increased synthesis of PGE2 (ElAttar 1976b)

Age
The biological changes on the period-
ontal tissues during different time points
such as puberty, the menstrual cycle,
pregnancy, menopause, and oral contra-
ceptive use have heightened interest in
the relationship between steroid sex
hormones and the health of the period-
ontium. Females seem to be more prone
to hormone imbalance than males and
therefore have been more extensively
studied. It is important, however, to note
that males also suffer from these varia-
tions (Morley 2000).
Puberty
Puberty is a complex process of sexual
maturation resulting in an individual
capable of reproduction (Ford & DOc-
chio 1989, Halpern et al. 1998). It is
also responsible for changes in physical
appearance and behavior (Buchanan et
al. 1992, Angold & Worthman 1993,
Angold et al. 1999) that are related with
increased levels of the steroid sex
hormones, testosterone in males and
estradiol in females.
During puberty, the production of sex
hormones increases to a level that
remains constant for the entire normal
reproductive period. A peak prevalence
of gingivitis was determined at 12 years,
10 months in females and 13 years, 7
months in males, which is consistent
with the onset of puberty (Sutcliffe
1972). Changes in hormone levels have
been related with an increased preva-
lence of gingivitis followed by remis-
sion (Massler et al. 1950, Curilovic et
al. 1958, Sutcliffe 1972, Daniell 1983),
a situation that is not necessarily
associated with an increase in the
amount of dental plaque (Sutcliffe
1972).
Table 5. Effect of osteoporosis upon periodontium
 Poor wound healing: less attachment formation (von Wowern et al. 1994)
 Reduced bone mineral content in the jaws (von Wowern et al. 1994, Payne et al. 1999)
 Increase of periodontosis and tooth loss (Mittermayer et al. 1998)
The subgingival microflora is also
altered during this period since the
bacterial counts increase in number,
and there is a prevalence of certain
bacterial species such as Prevotella
intermedia (Pi) and Capnocytophaga
species (Yanover & Ellen 1986, Gus-
berti et al. 1990). Pi has been shown to
possess the ability to substitute estrogen
and progesterone for menadione (vita-
min K) as an essential growth factor
(Kornman & Loesche 1982). This may
explain the association between in-
creased estrogen concentrations and
the elevated counts of Pi. On the other
hand, Capnocytophaga species, which
often increase during puberty, have
been associated with the increased
bleeding tendency observed during this
period of time (Gusberti et al. 1990).
Mombelli et al. (1990) evaluated long-
itudinal changes, during a period of 4
years, in the composition of the sub-
gingival microbiota of children between
11 and 14 years of age. They observed
that children who developed marked
and sustained puberty gingivitis had
higher frequencies of spirochetes, Cap-
nocytophaga sp., Actinomyces viscosus,
and Eikenella corrodens than the chil-
dren without puberty gingivitis. They
also noted that a resurgence of Pi was
correlated with a high bleeding ten-
dency in these patients.
The data strongly indicate that with
the influence of sex hormones, children
in puberty experience an exaggerated
gingival inflammatory response to pla-
que. Table 4 lists the clinical and
microbiological changes observed in
the periodontium during puberty.
Menstrual cycle
The menstrual cycle is controlled by the
secretion of sex hormones over a 2530-
day period and is responsible for
continued ovulation until menopause
(Ganong 1997, McCartney et al.
2002). In humans, the menstrual cycle
can be divided into two phases: a
follicular or proliferative phase, and a
luteal or secretory phase. During the
first phase, there is an increase in
estrogen levels. At the same time, the
luteinizing hormone stimulates proges-
terone secretion and ovulation. After
ovulation, the luteal phase is character-
ized by an increase in progesterone and
estrogen secretion. At the end of this
phase, and if fertilization has not
occurred, the plasma levels of proges-
terone and estradiol decline because of
the demise of the corpus luteum (Laufer
et al. 1982).
Generally, the periodontium does not
exhibit evident changes during the
menstrual cycle. Nonetheless, two dif-
ferent clinical findings have been ob-
served in the oral cavity: gingival
bleeding and increased production of
gingival exudate (Kribbs & Chesnut
1984, Kribbs et al. 1989, Kribbs 1990,
1992, Grodstein et al. 1996a, b). In
addition, ulcerations of the oral mucosa
and vesicular lesions have also been

noted (Segal et al. 1974) in the luteal
phase of the menstrual cycle, although
the incidence is low (Segal et al. 1974,
Ferguson et al. 1978, 1984). However,
the specific mechanism of how steroid
sex hormones influence vesicle/ulcera-
tion formation remains to be determined
(Mariotti 1994).
Pregnancy
Some of the most remarkable endocrine
alterations accompany pregnancy. Dur-
ing this period, both progesterone and
estrogen are elevated due to continuous
production of these hormones by the
corpus luteum. By the end of the third
trimester, progesterone and estrogen
reach peak plasma levels of 100 and
6 ng/ml, respectively, which represent
10 and 30 times the levels observed
during the menstrual cycle (Zachariasen
1989, Amar & Chung 1994, Mariotti
1994).
Susceptibility to infections (e.g. per-
iodontal infection) increases during
early gestation due to alterations in the
immune system (Cohen et al. 1969,
Brabin 1985) and can be explained by
the hormonal changes observed during
pregnancy (Hansen 1998, Smith 1999),
suppression on T-cell activity (Priddy
1997, Taylor et al. 2002), decreased
neutrophil chemotaxis and phagocyto-
sis, altered lymphocyte response and
depressed antibody production (Soor-
iyamoorthy & Gower 1989a, Raber-
Durlacher et al. 1991, 1993, Zacharia-
sen 1993), chronic maternal stress
(Culhane et al. 2001), and even nutri-
tional deficiency associated with in-
creased nutritional demand by both the
mother and the fetus (Wellinghausen
2001). Consequently, increased suscept-
ibility for certain infections such as
Helicobacter pilori (Lanciers et al.
1999), Coxiella burnetii, Listeria mono-
cytogenes, Toxoplasma gondii (Smith
1999), and virus infections (e.g. hepati-
tis E virus rubella, herpes, and human
papilloma virus) has also been observed
and reported (Priddy 1997).
These immunologic changes might
also be responsible for periodontal
pathologic conditions observed during
pregnancy such as pregnancy gingivitis
(Loe & Silness 1963, Silness & Loe
1964, Miyazaki et al. 1991, diLauro &
Tarturo 1971, Machuca et al. 1999,
Soory 2000a), pregnancy granuloma,
periodontitis, and dental caries (diLauro
& Tarturo 1971). The increased synth-
esis of PGE2 observed when estradiol
Sex hormones and periodontium
and progesterone are present in higher
concentrations, such as occurs during
pregnancy (ElAttar 1976a), may also
contribute to these pathologic changes
(Offenbacher et al. 1984, 1986). On the
other hand, periodontal pathogens such
as Pi and Porphyromonas gingivalis
(Pg) can also use female sex hormones
such as progesterone or estradiol as a
source of nutrients. These bacteria are
generally increased in the gingival
crevicular fluid of pregnant women, a
situation that is positively correlated
with the severity of pregnancy gingivitis
(Kornman & Loesche 1980, Tsai &
Chen 1995). These microbiological
shifts usually do not last postpartum
(Raber-Durlacher et al. 1994). None-
theless, (Jonsson et al. 1988) found that
Pi remains consistent even after the end
of the pregnancy.
Aside from the transient increases in
bleeding (Arafat 1974, Miyazaki et al.
1991), gingivitis, larger gingival prob-
ing depths (Hugoson 1971, Miyazaki et
al. 1991), increased gingival crevicular
fluid flow (Hugoson 1971), and the
subgingival microbial shift, pregnant
women in good heath are unlikely to
experience any significant gingival re-
sponses that would have serious clinical
implications (Amar & Chung 1994).
Menopause and postmenopause
In the premenopausal women, the
principal circulating estrogen is 17b-
estradiol (Katz & Epstein 1993). As
women approach menopause, the levels
of estrogen begin to drop mainly during
the late follicular and luteal phase of the
menstrual cycle (Sherman & Korenman
1975). As a result of this physiologic
situation, irregular cycles start to occur.
Frequently, the time frame between
regular cycles and the cessation of
menstrual periods, called perimenopau-
sal transition, is 27 years (Treloar
et al. 1970). During this period, the
concentration of circulating estrogen
decreases while follicle-stimulating hor-
mone (FSH) and luteinizing hormone
(LH) concentrations increase (Monroe
& Menon 1977). Consequently, the
effects of estrogen listed in Table 1
are reduced, therefore compromising
the anti-inflammatory effect of this
hormone on the periodontium.
Progesterone is another sex hormone
that may play an important role in bone
metabolism during pre- and postmeno-
pause (Katz & Epstein 1993). It is
believed that ovarian deficiency and
675
associated alterations, but not aging,
are the predominant causes of bone loss
during the first two decades after
menopause (Richelson et al. 1984).
Researches have shown that progester-
one may compete with glucocorticoids
for an osteoblast receptor and inhibit the
glucocorticoid-induced osteoporosis.
Therefore, postmenopausal bone density
reduction may be the result of a
combination of the inhibition of osteo-
clast downregulation by reduced estro-
gen and the increased cortisol inhibition
of osteoblasts via the reduction of
competition with progesterone (Katz &
Epstein 1993).
Menopause also affects the concen-
tration of circulating androgens. Before
menopause, 50% of the circulating
androstenedione is derived from the
ovaries and 50% from the adrenals.
With the ovarian failure that typically
occurs with menopause, the concentra-
tion of circulating androgens is reduced
by about 50% (Judd 1976). It has been
suggested that the peripheral conversion
of androgens to estrogens may be the
main factor for protecting bone (Katz &
Epstein 1993), since, as previously
mentioned, estrogens have an inhibitory
effect on osteoclastic action. Testoster-
one was also found to be positively
correlated with bone density, a finding
that is supported by the evidence of low
concentrations in osteoporotic patients
(Davidson et al. 1982, Steinberg et al.
1989).
A number of studies have linked
menopause with some periodontal con-
ditions, although the different methods
applied to assess osteoporosis, alveolar
bone loss, and periodontitis make that
literature difficult to analyze. As sug-
gested, osteoporosis is responsible for
less crestal alveolar per unit volume, a
condition that may promote quicker
bone loss when encountered with infec-
tions such as periodontal infections
(Wactawski-Wende et al. 1996).
It has been reported that the inci-
dence of periodontitis correlates with
signs of generalized osteoporosis, and
lower bone density of the mandible with
an increased incidence of periodontal
disease (Groen et al. 1968, Klemetti et
al. 1994, Krall et al. 1994, Krall 2001,
Wactawski-Wende 2001), although
other studies do not show consistent
relationships between those aspects
(Kribbs 1990, Elders et al. 1992).
Reinhardt et al. (1999) reported that
patients with estrogen deficiency
showed more bleeding on probing
676 Mascarenhas et al.
(BOP) and a higher frequency of
X2 mm of clinical attachment loss than
patients without estrogen deficiency.
They then speculated that estrogen
supplementation is important in redu-
cing gingival inflammation and limiting
the frequency of clinical attachment loss
in osteopenic/osteoporotic women in
early menopause. However, when Win-
grove et al. (1979) compared gingival
biopsies taken from postmenopausal
females and from younger women with
regular menses, no significant differ-
ences in the gingival tissues associated
with changes in the levels of sex
hormones were found.
The same correlation between estro-
gen deficiency, osteopenia/osteoporosis,
and the prevalence of tooth loss was
also reported (Daniell 1983, Kribbs et
al. 1989, Elders et al. 1992, von
Wowern et al. 1994, Paganini-Hill
1995a, Grodstein et al. 1996a,, Mitter-
mayer et al. 1998, Reinhardt et al.
1999). An important idea that should
be kept in mind is that in many of these
cases, tooth loss might be influenced by
factors other than periodontal disease.
The severity of osteoporosis was
found to be statistically significantly
associated with the height of the re-
sidual alveolar ridge (Hirai et al. 1993).
Payne et al. (1999) in a 2-year long-
itudinal study examined alveolar bone
height and density changes in osteo-
porotic/osteopenic women compared
with women with normal lumbar spine
bone mineral density (BMD). In total,
38 postmenopausal women with a past
history of periodontal disease on a 3- or
4-month periodontal maintenance inter-
val were monitored; 21 had normal
bone mineral density and 17 had
diagnosed osteoporosis. Results from
this study indicated that osteoporotic/
osteopenic women exhibited a higher
frequency of alveolar bone height loss,
as well as crestal and subcrestal density
loss when compared to women with
normal bone density. This corresponds
to the decreased amount of circulating
estrogen present in the osteoporotic/
osteopenic women.
Even though osteoporosis has been
considered a risk for periodontal disease
(Page & Beck 1997, Krejci & Bissada
2000, Pihlstrom 2001), many authors
still question if the associations reported
above are just concomitant findings or if
osteoporotic patients have an increased
susceptibility to periodontal disease,
which would explain the increased
incidence of edentulous patients and
more severe forms of periodontal dis-
ease. In 1997, Streckfus et al. studied
the relationship among alveolar bone
loss, alveolar bone density, second
metacarpal density, salivary and gingi-
val crevicular fluid IL-6, and IL-8
concentrations in premenopausal and
postmenopausal healthy women receiv-
ing estrogen therapy. Other than obser-
ving that postmenopausal women on
estrogen therapy had more alveolar
bone loss, more missing teeth, and
reduced alveolar and second metacarpal
bone density than premenopausal wo-
men, the authors also noticed that
postmenopausal women on estrogen
therapy had higher salivary IL-6 con-
centrations than premenopausal women.
From this study it was concluded that
levels of bone resorptive cytokines in
saliva may be secondary to changes in
menopausal status, and that these
changes may predispose loss of alveolar
bone with resultant loss of teeth (Streck-
fus et al. 1997).
Although research shows a linkage
between menopause and increased signs
of periodontal disease and a reduction in
alveolar bone height, it is important to
note that both menopausal and postme-
nopausal women who are in good
gingival health should not be considered
to be at an increased risk of periodontal
disease, although these physiologic con-
ditions may affect the severity of the
present disease (Amar & Chung 1994).
Hormone replacement
As addressed above, females experience
hormonal changes under both physiolo-
gical (e.g. menstrual cycle, pregnancy)
and nonphysiological conditions (e.g.
hormone therapy, use of oral contra-
ceptives). The effects of these treat-
ments on the periodontium have been a
center of focus for understanding the
interaction between sex hormones and
periodontium health (Soory 2000b).
Contraceptives
Hormonal contraceptives induce a hor-
monal condition that stimulates a state
of pregnancy to prevent ovulation by
the use of gestational hormones. Oral
contraceptive agents are one of the most
commonly used classes of drugs. The
most commonly used contraceptives
nowadays consist of low doses of
estrogens (30 mg/day) and/or Progestins
(1.5 mg/day) (Chihal et al. 1975, Brown
et al. 2001, 2002).
The influence of contraceptives on
the periodontium is not limited to
increases in inflammation and in the
amount of gingival exudates (Lindhe &
Bjorn 1967, Lynn 1967, Groen et al.
1968, el-Ashiry et al. 1971, Knight &
Wade 1974, Pankhurst et al. 1981).
These drugs have also been associated
with an increase in the prevalence of dry
socket after dental extraction (Catellani
1979), and accelerated progression of
periodontal disease (higher gingival
index scores and more loss of attach-
ment) when they are used long-term. In
contrast, some authors have found no
significant influences on the periodontal
clinical parameters when comparing
oral contraceptives to non-medicated
control groups (Moshchil et al. 1991).
Hormone replacement therapy in
postmenopausal women
Osteoporosis, which is defined as a
systemic condition characterized by a
decrease in the bone mineral density of
at least 2.5 times the normal values in a
healthy young female, is a major health
problem in postmenopausal women. In
Western societies, more than one-third
of the female population above the age
of 65 years suffers from signs and
symptoms of osteoporosis, a disorder
characterized by low bone mass. Estro-
gen deficiency is the dominant patho-
genic factor for osteoporosis in women
(Reinhardt et al. 1999). Although hor-
monal replacement in an adequate
dosage can slow or prevent bone loss
(Ettinger 1993, Allen et al. 2000), only a
small percentage of postmenopausal
women receive such therapy, and many
who do fail to comply with the pre-
scribed regimen because of the fear of
cancer, irregular bleeding, and other
minor side effects (Bjorn & Backsrom
1999, Bai et al. 2000, Kenemans et al.
2001, Schneider 2001).
Progesterone alone is not effective in
preventing postmenopausal bone and
tooth loss (Ettinger 1988, Jeffcoat
1998), but when combined with estro-
gen it is believed to uncouple formation
and resorption to diminish bone resorp-
tion induced by estrogen (Christiansen
et al. 1985).
Paganini-Hill (1995a) analyzed the
effects of estrogen replacement therapy
(ERT) on the prevention of tooth loss
and the need for dentures in older
women. Results from this study indi-
cated that the proportion of edentulous

women decreased with increasing dura-
tion of ERT. The author therefore
concluded that ERT may be beneficial
in preventing tooth loss and the need for
dentures in older women
Reinhardt et al. (1999) analyzed pro-
spectively the influence of serum estra-
diol (E2) levels and osteopenia/osteo-
porosis on common clinical measure-
ments of periodontal disease over a 2-
year period. E2 status, which was condi-
tioned by the presence of ERT, did not
influence the percentage of sites losing
clinical attachment level for either per-
iodontitis or nonperiodontitis groups, but
when nonsmoking osteopenic/osteoporo-
tic periodontitis patients were evaluated,
E2-deficient subjects had more BOP and
a trend toward a higher frequency of
X2.0 mm clinical attachment loss than
E2-sufficient subjects. These data suggest
that E2 supplementation (serum E2440 pg/
ml) is associated with reduced gingival
inflammation and a reduced frequency
of clinical attachment loss in osteope-
nic/osteoporotic women in early meno-
pause.
Other than just having a positive
effect on the alveolar and skeletal bone
density, hormonal replacement therapy
may have other positive effects such as
reduction of the risk of fatal and
nonfatal myocardial infarction, is-
chemic heart disease, and stroke by
2040%, and even a reduction in the
prevalence of Alzheimers disease (Pa-
ganini-Hill 1995b, c).
The available data indicate that hor-
mone replacement therapy should be
suggested for women during menopause
since several pathologic conditions com-
mon during this period of time can be
avoided or at least reduced in severity.
Influence of Sex Hormones on
Periodontal/Implant Wound Healing
The effects of sex hormones in the
wound healing of different organs have
been studied and, in some cases, proved.
Even though the direct effect of these
hormones in periodontal wound healing
is still far from being completely studied
and clarified, other systems like the brain
(Chowen et al. 2000) and the ligaments
(Frank et al. 1999) are already known to
be targets for sex hormones during the
wound healing processes. At a molecular
level, research has also shown that sex
hormones have a regulatory effect on
growth factors involved in the wound
healing such as the keratinocyte growth
factor (Rubin et al. 1995), which has
Sex hormones and periodontium
been known to have wound healing
regulatory effect including stimulation
of proliferation, migration, and morpho-
genesis of pluripotential cells.
However, the influence of sex hor-
mones on periodontal wound healing is
still largely unknown. As discussed
above, people with hormone deficiencies
may show more periodontal disease/
destruction. However, the mechanisms
of how this occurs remain to be
determined. Overall, lack of sex hor-
mones often causes the reduction of bone
density (Reinhardt et al. 1999). There-
fore, some authors consider the osteo-
porotic status a risk factor for implant
success (Roberts et al. 1992) and others
disagree. For example, Cuenin et al.
studied the association between sex
hormone levels and dental implant
success in three patient groups: (1) male
controls, (2) females with high estrogen,
and (3) females with low estrogen. They
found that the balance of alveolar
osseous wound healing was not influ-
enced by temporal fluctuations of the
ovulatory cycle (Cuenin et al. 1997).
More research is definitely needed to
improve the understanding of how sex
hormones influence the overall period-
ontal and implant wound healing.
Conclusion
naux de substitution). Aussi, cette revue critique
Sexual hormones play an important role
de la litterature se propose (1) de faire le point
in influencing periodontal disease pro-
gression and wound healing. These
effects are different depending on the
gender as well as the lifetime period
analyzed. It is also clear that not all
patients and their periodontium respond
in the same way to similar amounts of
circulating sexual hormones.
In addition, the influence of sex
hormones can be minimized with good
plaque control as well as with hormone
replacement therapies; however, the
true mechanism of how these interac-
tions actually occur remains to be
determined.
Disclaimers
The authors do not have any financial
interests, either directly or indirectly, in
the products listed in the study.
Acknowledgements
This work was partially supported by
the University of Michigan, Periodontal
Graduate Student Research Fund.
677
Zusammenfassung
Einfluss von Sexualhormonen auf das Parodon-
tium
Mannliche und weibliche Sexualhormone wur-
den schon lange einen wichtigen Einfluss auf
das parodontale Gewebe, die Knochenumsatz-
rate, die Wundheilung und die parodontale
Erkrankungsprogression ausubend betrachtet.
Der Einfluss dieser Hormone auf das Parodon-
tium unterscheidet sich zu verschiedenen phy-
siologischen Phasen (z.B. Pubertat,
Schwangerschaft, post Menopause) und mit
der Einnahme von Pharmaka (z.B. Antikonzep-
tiva, Hormonsubstitution). Deshalb ist der
Zweck dieses Reviewartikels (1) die Beziehung
zwischen Sexualhormonen und dem Parodon-
tium zu beschreiben, (2) die Analyse des
Einflusses dieser Hormone auf das Parodontium
zu unterschiedlichen Lebenszeiten und (3) die
Effekte von Hormonunterstutzung/substitution
auf das Parodontium zu diskutieren.
Resume
Influence des hormones sexuelles sur le par-
odonte
On a longtemps considere que les hormones
sexuelles, aussi bien masculines que feminines,
jouaient un role important sur les tissus
parodontaux, le taux de remaniement osseux,
la cicatrisation et la progression de la maladie
parodontale. Linfluence de ces hormones sur le
parodonte est differente en fonction des divers
conditions physiologiques (par exemple, la
puberte, la grossesse, et apres la menopause)
et les prises de medicaments (par exemple, la
pillule contraceptive et les traitements hormo-
sur les liens entre les hormones sexuelles et le
parodonte (2) d analyser la facon dont ces
hormones influencent le parodonte lors des
differentes etapes de la vie , et (3) discuter les
effets des hormones de substitution sur le
parodonte.
References
Abraham-Inpijn, L., Polsacheva, O. V. &
Raber-Durlacher, J. E. (1996) The signifi-
cance of endocrine factors and microorgan-
isms in the development of gingivitis in
pregnant women. Stomatologiia (Mosk) 75,
1518.
Albandar, J. M., Brunelle, J. A. & Kingman, A.
(1999) Destructive periodontal disease in
adults 30 years of age and older in the
United States, 19881994. Journal of Period-
ontology 70, 1329.
Allen, I. E., Monroe, M., Connelly, J., Cintron,
R. & Ross, S. D. (2000) Effect of post-
menopausal hormone replacement therapy on
dental outcomes: systematic review of the
literature and pharmacoeconomic analysis.
Management Care Interface 13, 9399.
Amar, S. & Chung, K. M. (1994) Influence of
hormonal variation on the periodontium in
women. Periodontology 2000 6, 7987.
Periodontology 2000, Vol. 61, 2013, 103124
Printed in Singapore. All rights reserved
Sex and the subgingival
microbiome: Do female sex
steroids affect periodontal
bacteria?
P U R N I M A S. K U M A R
The human female experiences periodic increases in
sex steroid levels during her reproductive life, which
begins at menarche and ends with menopause. Both
endogenous hormones and exogenously adminis-
tered steroids may contribute to these fluctuations.
Gingival inflammation has been extensively reported
in association with elevated female hormonal surges,
and the biological mechanisms underlying this florid
inflammatory condition have been examined for over
a century. Gingival enlargement during pregnancy
was first described in 1844 by Hulihen (39). He called
this condition an aneurysm of the gingiva. In the
same year, Westcott (116) described pregnancy gin-
givitis as a condition of uterine irritation. This con-
dition was attributed to several causes, including
metastasis of the menstrual secretion from the
uterus to the mouth (39). It was not until several
decades later that bacteria in dental plaque were
recognized as the primary etiological agents of peri-
odontal diseases, leading to research on correlations
between elevated female sex steroid levels and the
composition of the subgingival microbiome.
The effect of female sex steroids on the composi-
tion of the subgingival biofilm has been examined
using in vivo investigations during periods of change
in hormonal levels and in vitro investigations using
bacterial cultures. Pregnancy, puberty, menstruation
and menopause are four physiological conditions
that offer a means to examine the changes that occur
in the oral microbial community during fluctuations
in endogenous hormonal levels. Oral contraceptive
use and hormone replacement therapy have been
used to examine the effect of exogenous hormonal
fluctuations on oral microbial profiles. In addition,
several studies have examined the effects of these
 2013 John Wiley & Sons A/S
PERIODONTOLOGY 2000
hormones on bacterial growth in vitro and in animal
models. This review examines the evidence from
these studies to determine the contributions of
female sex steroids to altering the composition of the
subgingival microbiome.
Factors affecting studies on the
composition of the subgingival
microbiome
Methods of bacterial identification and
characterization
Several different methodologies have been employed
to examine the association between sex hormones
and subgingival bacteria. The methodologies are
summarized in Fig. 1. The earliest investigations were
based on cultivation and microscopic examination.
Cultivation is an open-ended approach, which does
not require a priori knowledge of the bacterial com-
munity under investigation. However, this approach
is very selective, as it precludes identification of
fastidious organisms and species whose growth
requirements are not known (109). Socransky et al.
(104, 105) showed that anaerobic cultivation
techniques can recover only 1020% of the total
microbial counts from gingival debris and dental
plaque. Thus what grows on a culture plate is not
necessarily representative of the bacterial community
under investigation, and this presents a serious
drawback in any quantitative investigation of micro-
bial communities.
Neither cultivation nor microscopic examination
provide accurate bacterial identification, as they use
103
Kumar

Statistical
Sampling method Hormonal measurements Clinical measurements Bacterial identification
method

Study design

Tanner stages

Gingival Index
Progesterone

Probe Depth
Testosterone

Plaque Index

hybridization
Skeletal age
Paper point

Microscopy

Parametric

parametric
Cultivation

DNA-DNA
State of hormonal

Estrogen

Bleeding
Gingival
Reference

Curette
elevation

Saliva

Other

Non-
PCR
Kelstrup (50) CS

Yanover & Ellen (122) LS

Delaney et al. (19) CS

Wojcicki et al. (118) CS

Morishita et al. (80) CS

Van Oosten et al. (114) CS

rty
be
Pu

Gusberti et al. (33) LS

Mombelli et al. (77) LS

Moore et al. (79) LS

Nakagawa et al. (83) LS

Mombelli et al. (78) LS

Tsuruda et al. (112) CS

Kornman & Loesche (57) LS

Jensen et al. (41) CS

Jonsson et al. (44) CS

Raber-Durlacher et al. (94) LS

y
nc
na
eg

Buduneli et al. (9) CS

Pr

Yokoyama et al. (124) CS

Gursoy et al. (31) LS

Adriaens et al. (1) LS

e
iv

Jensen et al. (41) CS

pt
ce
tra

LS
on

Klinger et al. (54)

C

Prout & Hopps (93) LS

n

LS
io

Calil et al. (10)

at
ru
st
en

Fischer et al. (26) LS

M

Fig. 1. Summary of clinical, microbiological and statistical methods. CS, Cross-sectional; LS, Longitudinal; PCR,
polymerase chain reaction.

phenotypic characteristics, e.g. cell or colony mor- organisms cultured from a single sample of subgingi-
phology, biochemical responses to stains, substrates val plaque are mis-identified using phenotypic
etc., for bacterial identification (40). Several bacteria characterization. As an example, we discuss how the
demonstrate similar morphological characteristics. taxonomy of black-pigmented Bacteroides has evolved
For example, species belonging to the genera Pasteu- as a result of advances in bacterial characterization
rella and Actinobacillus do not have a single pheno- methods. Thus, the limited discriminatory powers of
typic characteristic that reliably distinguishes between phenotypic and biochemical characterization should
them. Similarly, Slots (102) found that it was not pos- be taken into consideration when interpreting the
sible to characterize all Gram-negative rods into indi- findings of cultivation-based investigations.
vidual species based on culture characteristics, and More recent studies have used DNA-based
Kroes et al. (58) demonstrated that at least 20% of approaches that circumvent the need for cultivation.

104

Targeted molecular approaches, for example poly-
merase chain reaction (PCR), real-time PCR, in situ
hybridization and DNADNA hybridization (check-
erboard), overcome the limitations of cultivation-
based approaches, as they are capable of identifying
and enumerating species that are as yet uncultivated.
However, probes or primers can be designed only
when the organism is known and its sequence
information available. Hence, directed DNA ap-
proaches are limited to examining the presence or
levels of previously known species. It is also chal-
lenging to determine the proportion of each species
in the community using these methodologies, as
gene copy numbers (for example, the 16S rRNA gene,
the GroEL gene, HSP genes etc.) differ between
species (12, 27). Further, the biofilm is composed
of both live and dead bacteria, and DNA-based
methods do not differentiate between viable and
dead organisms.
The methodology used for bacterial identification
and enumeration plays a critical role in the findings
of any investigation. Hence, the merits and limita-
tions of each approach must be given serious con-
sideration when examining the literature on female
sex steroids and the composition of the subgingival
microbiome.
Statistical methods used in bacterial
comparisons
A careful examination of the research on sex hor-
mones and subgingival bacteria reveals that both
parametric and non-parametric tests have been
used for comparing bacterial communities (Fig. 1).
Parametric analyses test for differences in means or
variances (t test or analysis of variance) based on
the assumption that the data are normally distrib-
uted. Parametric tests also assume homogeneity of
variances and independence between mean and
variance. Bacterial populations usually exhibit non-
normal distributions, which can lead to erroneous
results when the null hypothesis is tested (43, 98,
111). Power transformations or non-parametric
tests overcome this limitation, either by stabilizing
the variance (power transformations) or by assum-
ing non-normality (non-parametric analysis). Re-
sults using these methodologies are usually more
conservative than parametric tests on non-trans-
formed data (6, 42, 96); however, they are more
accurate (43). Thus, the implications of using
parametric tests over non-parametric tests must be
taken into account when investigating bacterial
communities.
Sex steroids and subgingival bacteria
Sampling methods
The studies described here have employed several
methodologies for collecting subgingival bacterial
samples: endodontic broaches, paper points, scalers
or curettes (Fig. 1). The effect of sampling method-
ology on the quantity and quality of plaque has been
examined in several studies. One study used culti-
vation to enumerate bacterial species, and showed
that samples collected on paper points demonstrated
significantly higher colony counts and spirochete
numbers than those collected using curettes (96).
When a molecular methodology was used to enu-
merate both total bacteria and the levels of selected
species, curette samples were found to yield higher
numbers of total bacteria; however, the samples were
not qualitatively different from the paper point
samples (42). Therefore, the evidence suggests that
the methodology used for sample collection can be a
factor in influencing the results of bacterial com-
munity studies.
Thus, in addition to inter-individual variations in the
composition of the subgingival microbial community,
several methodological factors can significantly affect
the outcomes of microbiological investigations. The
influence of bacterial characterization, sampling
technique, and analysis and interpretation of the data
have been briefly discussed here. It is important to
understand the impact of such methodological dif-
ferences when evaluating the results of various studies.
Gender and the subgingival
microbiome Do women have a
different bacterial profile to men?
The incidence and severity of periodontal disease
appear to have a gender predilection, being greater in
males than in females (2, 89). In contrast, females
have a greater susceptibility to caries than males (24).
The primary etiological agents of both diseases are
bacteria, and a logical assumption would be that oral
bacterial profiles differ between genders. A second
logical assumption, based on the prevailing paradigm
that sex steroids affect subgingival bacteria, would be
that the cumulative effects of puberty, menstruation,
oral contraceptives and pregnancy cause long-term
changes in the female subgingival microbiome. Of
the few studies in the literature that examine the
association between gender and the composition of
the oral microbiome in health and disease (14, 70,
101, 103, 113), only one has suggested an association
between gender and carriage of a specific organism.
105
Kumar
Umeda et al. (113) found that males, not females,
have greater odds of carrying Prevotella intermedia in
saliva, subgingival and supragingival plaque. Thus,
there is very little evidence in the literature to indi-
cate a gender predilection for oral bacterial coloni-
zation, suggesting that the elevated sex steroid levels
in females do not produce long-term changes in the
oral microbiome.
Sex and the human microbiome
Do sex steroids affect all mucosal
ecosystems?
It is known that the gingiva contains receptors for
ovarian hormones, making it a target for the actions
of estrogen and progesterone. The gingival inflam-
mation observed during pregnancy and puberty, for
example, is mediated by the actions of these hor-
mones on the gingival receptors. The currently
accepted model of periodontal disease during hor-
monal surges is that inflammation occurs due to the
effect of the hormones on both the gingival tissues as
well as the subgingival ecosystem. As these hormonal
variations are systemic, it may be useful to investigate
whether host-associated bacterial ecosystems else-
where in the body respond in a similar manner to
those in the subgingival habitat.
Mucosal receptors for ovarian hormones are found
in the nasopharynx, gastrointestinal tract and female
uro-genital system. In addition to florid inflamma-
tion of the gingiva, pregnant women are also reported
to suffer from nasal congestion, inflammation of the
nasal mucosa and coryza (20). However, there is no
reported association between bacterial community
shifts and pregnancy rhinitis; instead, it has been
shown that estrogen-induced mucosal hyper-secre-
tion and tissue enlargement lead to pregnancy rhi-
nitis (92). Similarly, although it is known that several
broncho-pulmonary infections exhibit a gender bias
(23), large-scale studies have shown that this sus-
ceptibility is due to immuno-modulatory effects of
estrogen (53).
Several lines of evidence indicate that the levels of
sex steroids influence bacterial colonization in the
vagina; however, the exact nature of this influence is
not clear. The high levels of estrogen seen during
menarche and pregnancy promote normal bacterial
colonization by increasing the amount of glycogen
produced by the epithelial cells, providing a
nutritional source for lactobacilli (5). High levels of
lactobacilli lead to a decrease in pH, which regulates
106
the composition of the vaginal, cervical and uterine
microbial communities (4). A decrease in the levels of
estrogen, for example during menopause, is associ-
ated with a reduction in levels of commensal species
and increased susceptibility to bacterial vaginosis (91,
117). The incidence and prevalence of bacterial vag-
inosis decreases in post-menopausal women after
initiation of hormone replacement therapy (28, 36).
Evidence suggests that cells of the cervical mucosa
also demonstrate a hormone-dependent antibacterial
activity, as human uterine epithelial cells from pre-
menopausal women inhibit bacterial growth, while
cells from post-menopausal women do not (22).
Progesterone has been shown to have a dose-
dependent bacteriostatic or bactericidal effect against
Neisseria and Staphylococcus spp. in vitro (125).
However, greater susceptibility to gonococcal,
candidal and chlamydial pelvic infections has been
reported in pregnant women and those on oral con-
traceptives (25, 106). These infections also have a
higher incidence during the earlier part of the men-
strual cycle (49). Taken together, it appears that local
environmental factors, rather than systemic hor-
monal surges, play a major role in determining the
composition of the uro-genital microbiome.
Thus, the available data indicate that, although
female sex steroids alter mucosal surfaces, an effect
on mucosa-associated biofilms is not as readily
apparent. Thus, there are divergent opinions on the
effect of ovarian hormones on complex microbial
ecosystems in the human body.
Endogenous sex steroids and the
subgingival biofilm
The human female has distinctive periods of surges
in endogenous sex hormones, beginning at menarche
and ending with menopause. Therefore, the majority
of data are derived from female populations of
reproductive age.
Puberty
Endocrinal changes in puberty
Reproductive endocrine development is regulated by
the anteroventricular periventricular (AVPV) nucleus
and the arcuate nucleus in the hypothalamus, which is
suppressed during childhood (Fig. 2). De-inhibition of
the arcuate nucleus leads to menarche or onset of
puberty in females. This de-inhibition occurs in
response to stimulation of the Kiss neurons in the

Metabolic and
Arctuate
nutritional cues nucleus
(leptin pathway)
Infant and childhood Anteroventral
periventricular
nucleus
Fig. 2. Hormonal and neuronal control of puberty.
anteroventral periventricular and arcuate nuclei by
metabolic cues that operate through the leptin path-
way. This de-inhibition is characterized by pulsed
release of gonadotrophin-releasing hormone (GnRH),
which stimulates the release of luteinizing hormone
(LH) and follicle-stimulating hormone (FSH) from the
anterior pituitary, which in turn release testosterone
from the thecal cells of the ovary. In females, most of
this testosterone is converted to estradiol by the
granulosa cells of the ovary. The growth spurts and
sexual changes of puberty are primarily due to the
action of estradiol on target tissues. Puberty in males
begins similarly, with pulsed release of GnRH. The
Leydig cells of the testicles secrete testosterone, which
is responsible for the changes associated with male
puberty (virilization). However, a significant portion of
the testosterone is also converted to estradiol in males,
which mediates growth spurts, bone maturation and
epiphyseal closure, similar to females. Thus, during
puberty, both males and females have high circulating
levels of estradiol.
Methods of assessing puberty
Investigators have used serum hormonal levels,
chronological age, bone age and or Tanner staging
to delineate subjects into various groups (Fig. 1). The
sexual maturity rating or Tanner staging is based on
development of secondary sexual characteristics
during puberty (73). It measures the change in size of
external genitalia, development of pubic hair and
breast development in girls. There are five stages of
puberty in this classification system, ranging from
stage 1 (pre-pubertal) to stage 5 (mature adult).
Hormonal levels vary significantly between levels
(95), so it is important to assess the Tanner stage of
puberty in order to understand the correlation
between hormones and gingival changes. Ethnic dif-
ferences in body structure, obesity and body mass
index can affect Tanner staging (47, 48). For example,
Sex steroids and subgingival bacteria
obesity can be mistaken for breast development in
females. Several investigations have used Tanner
staging to identify onset of puberty (19, 33, 77, 83, 118,
Kiss neurons GnRH neurons
121); however, only Delaney et al. (19) identified dif-
ferent stages of maturation within the pubertal group.
GnRH
Bone age measurements based on wrist radio-
graphs have also been used to assess puberty.
Anterior
pituitary
Metacarpal width and density differ significantly
LH/FSH
between Tanner stages 2 and 3 and stages 3 and 4
Testes Ovary (62). Therefore, a combination of the two methods
T/E2 E2
would provide better discrimination of pubertal
Puberty
stages than either one individually.
Pubertal changes begin approximately at 10 years in
girls and 11 years in boys, and last for 4 years (18, 21,
37). Several factors, notably nutrition and body mass
index, have a direct bearing on the onset of puberty
(48). Therefore, chronological age may be a poor
indicator of the onset or duration of puberty.
The methodologies used to assess puberty must be
considered in examining the results of these studies,
since Tanner staging, skeletal maturity and chrono-
logical age have been used as surrogate measures of
hormonal increases in puberty.
Subgingival bacteria in puberty
Research on the effect of pubertal hormones on
microbial changes occurring in the gingival sulcus
can be divided into two broad categories: (i) investi-
gations examining compositional changes occurring
in the subgingival microbiome before, during and
after puberty, and (ii) investigations examining
the prevalence and abundance of black-pigmented
Bacteroides before, during and after puberty.
The clinical and microbiological findings of each
study and the conclusions drawn are summarized in
Table 1.
Changes in the microbiome
The predominant subgingival organisms in pre-
pubertal children are reported to be Lactobacillus,
Streptococcus, Capnocytophaga, Eikenella, Prevotella
and Actinomyces spp., and Peptostreptococcus (17, 19,
45, 46). During the pre-pubertal period, these children
transition from primary to mixed dentition, but with
very small changes in their subgingival microbial
community (45, 46). After the onset of puberty, there
appears to be a significant shift in the composition of
this community, with an increase in the levels of cocci
(Streptococcus and Veillonella), rods (Actinomyces,
Prevotella, Fusobacterium), spirochetes and gliders
(Capnocytophaga, Eikenella and Wolinella) (33, 79,
122). Moore et al. (79) reported increases in the levels
of Lactobacillus rimae (now Atopobium rimae),
107
Kumar

Table 1. Summary of evidence: studies on pubertal children

References Study Number of subjects, Clinical Bacterial profile Authors conclusions and
duration gender (age range at outcomes reviewers comments
baseline, years)

Kelstrup Cross- Five primary Gingivitis absent B. melaninogenicus Positive association

(50) sectional dentition (45) only in primary seen more often in between
14 mixed dentition dentition adolescent and adult B. melaninogenicus and
(68) dentition. In adults, age.
16 adolescent detection frequency Adults and adolescents
permanent dentition greater in deeper exhibit similar frequency of
(1315) sulci. No correlation detection.
Nine adult dentition with GI or PI No association between
(2750) B. melaninogenicus and
gingival inflammation
Bacterial presence or
absence rather than levels
were measured
Sexual maturity of subjects is
not known.
Delaney Cross- 12 pre-pubertal BOP, PI and PD Actinomyces, Eikenella, Positive association
et al. (19) sectional females similar across Capnocytophaga, between B. intermedius
Four circum-pubertal groups. All Wolinella and and pre-pubertal gingivitis.
females subjects had Bacteroides No association between
Six post-pubertal gingival predominate at all B. intermedius and pubertal
females (716) inflammation ages. B. intermedius gingivitis.
in sampled sites levels were higher in No evidence of puberty
pre-pubertal females gingivitis.
within 1 year of Small sample size.
menarche Correlations made with
sexual maturity, rather than
chronological age.
Phenotypic and biochemical
characterization used for
bacterial identification

Yanover Longitudinal 18 females No change in GI No change in No correlation between

& Ellen with normal puberty or PI following B. intermedius with puberty and the presence
(122) (9.515.5) Eight menarche age or puberty. or levels of BPB, except in
females with increase BPB in sites precocious puberty.
precocious puberty with increased GI Small sample size with
(7-11) Correlation between two-year follow-up.
Two-year follow-up increase in E2- Phenotypic and biochemical
with 4-month recalls estradiaol and the characterization used for
presence of BPB (but bacterial identification.
not BPB levels) in BPB not speciated.
precocious puberty
Wojcicki Cross- 12 pre-pubertal Higher BOP and Level of B. intermedius Positive association between
et al. sectional children (78) PD in all higher in circum- puberty and BPB.
(118) 16 circum-pubertal circum-pubertal pubertal children. Small sample size,
children (1213.5) and several Levels of spirochetes, cross-sectional study,
14 post-pubertal post-pubertal B. melaninogenicus several media plates in the
children (1619) subjects and B. denticola pre-pubertal and
loeschii greater in post-pubertal groups were
post-pubertal chil- lost due to contamination;
dren bacterial counts therefore
not equally distributed
between groups.

108
 Sex steroids and subgingival bacteria

Table 1. (Continued)

References Study Number of subjects, Clinical Bacterial profile Authors conclusions and
duration gender (age range at outcomes reviewers comments
baseline, years)

Morishita Cross- 132 subjects: 63 Lower PD in Lower total bacteria, Negative association
et al. (80) sectional males, 69 females males and cocci, rods and between progesterone
(1215) females with spirochetes in males and bacterial counts as
low and females with well as probe depths.
progesterone higher progesterone Sexually mature as well as
Greater BOP in immature subjects were
males with included in study; with
higher estradiol no segregation.
Bacterial characterization
based on microscopic
morphology.

van Oosten Cross- 22 males, 20 females Gingivitis BPB and spirochetes Positive association
et al. (114) sectional (1111.8) present in detected in 75% of between BPB, spir
Four-year follow-up approximately children ochetes and gingivitis in
with 45-month 50% of pre- B. intermedius was the pre-pubertal children.
recalls pubertal most common BPB
Focus on pre-pub children Higher levels of BPB
ertal conditions and spirochetes were
found in sites with
bleeding; none were
detected in non-
bleeding sites
Gusberti Longitudinal 22 males, 20 females BOP increases Increase in detection Temporary shifts in
et al. (33) (1111.8) during onset of frequency of microbial community
Four-year follow-up puberty B. intermedius, seen during onset of
with 45-month B. melaninogenicus puberty, along with
recalls and A. odontolyticus temporary increase in
Focus on longitudinal at onset of puberty in gingival bleeding.
bacterial changes boys Positive correlation
during puberty A. odontolyticus, and between A. odontolyticus
Capnocytophaga sp. and sexual maturation in
increased with age in both boys and girls.
boys and girls No correlation between
BPB and puberty.
Bacterial presence or
absence rather than
levels were measured.
Longest follow-up of
pubertal children
available in the literature.

Mombelli Longitudinal 22 males, 20 females BOP increases Increase in Eikenella No correlation between
et al. (77) (1111.8) during onset of and spirochetes and Bacteroides and pubertal
Four-year follow-up puberty decrease in changes.
with 45-month A. viscosus in puberty BPB detected after
recalls gingivitis gingivitis was
Focus on puberty Increase in BPB after established.
gingivitis bleeding established Bacterial presence or
absence rather than
levels were measured.
Longest follow-up of
pubertal children
available in the literature.

109
Kumar

Table 1. (Continued)
References Study Number of subjects, Clinical Bacterial profile Authors conclusions and
duration gender (age range at outcomes reviewers comments
baseline, years)
Moore Longitudinal 20 male monozygotic GI and bleeding V. atypica, P. denticola Positive correlation
et al. (79) twins (11) index increase and between P. denticola,
20 male dizygotic with time P. melaninogenica P. melaninogenica and
twins (11) increased in puberty. V. atypica and puberty.
10 unrelated children Prevotella D1C20 Strong evidence that
(46) higher in pre-pubertal subgingival microbial
11 unrelated adults children. profile at age 14 is
(2434) Only V. atypica different from that of
Three-year follow-up increased in both twin younger children as well
of twins with groups. as adults. However, the
18-month recall Pre-pubertal and strongest association was
pubertal children with between puberty and
gingivitis had no V. atypica, not
difference in flora P. denticola or
Sample size was too P. melaninogenica, each
small to investigate of which was increased
correlation between in only one of the twin
testosterone levels groups.
and bacterial levels No association was
demonstrated between
hormonal levels and
bacterial profiles.
Nakagawa Longitudinal 24 subjects: 12 with GI increased in P. intermedia in Positive correlation
et al. (83) gingivitis, 12 healthy all females creased in males and between pubertal
(812) circum- females with gingivitis and levels of
Unspecified duration, pubertally, but prepubertal gingivitis. P. intermedia.
4-month recalls only in males P. intermedia Small sample size, single
with increased during examiner.
pre-pubertal puberty in Strongest correlation was
gingivitis gingivitis-free between pre-existing
females. Antibodies to gingivitis in both males
P. intermedia and females and pubertal
increased in all levels of P. intermedia
females and males
with pre-existing
gingivitis

Tsuruda Cross- 42 pubertal children PD and BOP Motile rods, No differences between
et al. (112) sectional (1215) greater in both spirochetes and puberty and adulthood
18 adults (2125) adults and P. intermedia greater in terms of levels of
children with in gingivitis for both P. intermedia.
gingivitis adults and children. A cluster of P. intermedia,
No difference between E. corrodens and motile
puberty and adults rods may indicate a high
risk for inflammation.
Cluster analysis was
performed only for
children, not young
adults, and comparisons
are not possible. No
clinical differences, only
microbial differences,
between clusters 1 and 2.
Pubertal status was not
verified in the children.

110

Table 1. (Continued)
References Study Number of subjects, Clinical
duration gender (age range at outcomes
baseline, years)
Mombelli Longitudinal 22 males, 20 females
et al. (78) (1111.8)
Six-year post-
pubertal follow up
Focus on
post-pubertal
changes
PD, probing depth; GI, gingival index; PI, plaque index; BOP, bleeding on probing; BPB, black-pigmented Bacteroides.
Prevotella denticola, Actinomyces naeslundii and Ve-
illonella atypica in pubertal boys, and Wojcicki et al.
(118) found higher levels of Selenomonas spp. and
Bacteroides intermedius in pubertal children. Thus,
during puberty, there appears to be a large-scale shift
in the composition of the subgingival microbiome.
This shift mostly occurs as a result of the change in
relative abundance of several species, and acquisition
of new species plays a smaller role.
When the subgingival microbial profile of pubertal
children was examined with respect to periodontal
health status in cross-sectional and longitudinal
studies, subjects with gingivitis appeared to have a
significantly different flora than periodontally healthy
individuals (77, 112). Gingivitis was associated with
elevated levels and or prevalence of Eikenella,
B. intermedius, Fusobacterium and spirochetes, while
periodontally healthy children had higher levels of
Actinomyces viscosus. According to the findings of
Tsuruda et al. (112), the flora of healthy pubertal
children and healthy young adults were highly simi-
lar, and the only differences detected were between
healthy subjects and those with gingivitis in both age
groups. Thus, the presence or absence of disease
appears to be a determinant of the microbial changes
seen in this group.
Taken together, it appears that puberty is a period
when large-scale changes occur in the subgingival
microbiome. These changes result in a microbial
profile that is similar to that of young adults. The
presence of disease plays a significant role in altering
the microbial community, and the composition
of disease-associated communities does not differ
between pubertal children and young adults.
Sex steroids and subgingival bacteria
Bacterial profile Authors conclusions and
reviewers comments
Increase in BOP Spirochetes and Positive correlation
and attachment A. actinomyecetem- between sites with
loss following comitans present spirochetes and
pubertal post-pubertally in P. intermedia before and
gingivitis subjects with pubertal after puberty.
gingivitis. Sites of No correlation between
puberty gingivitis pubertal gingivitis and
contained spirochetes P. intermedia.
during puberty and Bacterial presence or
for six years post- absence rather than
pubertally levels were measured.
Longest follow-up of
post-pubertal children
available in the literature
Changes in black-pigmented Bacteroides
The association between black-pigmented Bactero-
ides and puberty has been investigated both cross-
sectionally and longitudinally, with equivocal results.
These bacteria have been detected in the subgingival
plaque of 27100% of pre-pubertal children (17, 19,
50, 83, 114, 118). The levels of these organisms were
reported to correlate with the presence of gingivitis in
this cohort (83, 114). Thus, it appears that black-
pigmented Bacteroides colonize the subgingival
microbial community during early childhood, and
are found in greater abundance in association with
disease in pre-pubertal children.
The largest evaluation of pre-, circum- and post-
pubertal children was carried out by Mombelli et al,
who longitudinally monitored clinical and micro-
biological changes in 42 children (22 boys and 20
girls) over 10 years [46, 47, 62, 63]. Both Tanner
staging and skeletal growth were used to assess
puberty. Black-pigmented Bacteroides were fre-
quently found in pre-pubertal children, and their
levels correlated with the degree of gingivitis (114).
Puberty was associated with an increase in gingival
bleeding in both boys and girls. Immediately after the
onset of puberty, P. intermedia and Prevotella mel-
aninogenica were detected more frequently in boys
(P = 0.05), while Actinomyces odontolyticus was fre-
quently detected in girls (P < 0.01) (33). However,
these initial increases were not sustained throughout
puberty, during which the levels of sex hormones are
expected to increase (95). Further, the proportions of
these species did not increase during any time
period. Among both boys and girls who developed
111
Kumar
gingivitis, an increase in both the frequency and
proportions of spirochetes and Eikenella corrodens
was observed (77). Species belonging to the genus
Capnocytophaga were detected more frequently prior
to the onset of bleeding, while black-pigmented
Bacteroides were detected more frequently after gin-
gival bleeding was established. This increase in de-
tection frequency suggests that either acquisition of
black-pigmented Bacteroides or increase in levels
above detection threshold is favored by the presence
of gingival inflammation.
Yanover & Ellen (122) longitudinally evaluated 18
females progressing normally through puberty over
2 years, and found no correlation between the levels
of black-pigmented Bacteroides and either physical
maturation, as measured by Tanner staging, or plas-
ma estradiol concentrations. However, an increase in
gingival index was associated with increased detec-
tion frequency of black-pigmented Bacteroides, sug-
gesting once again that gingival inflammation is a
greater determinant of colonization by black-pig-
mented Bacteroides than puberty.
In contrast, a longitudinal evaluation of 24 children
by Nakagawa et al. (83) revealed a significant associ-
ation between serum estrogen and progesterone levels
and proportions of P. intermedia Prevotella nigres-
cens in both the presence and absence of gingivitis.
However, the levels of testosterone only correlated
with these species in the presence of gingivitis.
As with the longitudinal studies, the results of the
cross-sectional investigations are ambivalent with
regard to a correlation between black-pigmented
Bacteroides and the levels of sex steroids. Tsuruda
et al. (112) reported higher proportions of P. inter-
media in pubertal children with gingivitis than in
healthy children; this increase was also seen in young
adults with gingivitis compared to an age-matched
healthy group. Wojcicki et al. (118) also found higher
proportions of black-pigmented Bacteroides in
pubertal children as well as young adults. Both
Moore et al. (79) and Kelstrup (50) found a higher
112
prevalence of black-pigmented Bacteroides in
14-year-old children compared to pre-pubertal chil-
dren; however, Delaney et al. (19) found that pre-
menarchal girls demonstrated higher levels of these
bacteria than pubertal girls. van Oosten et al. (114)
reported a greater prevalence of these organisms in
pre-pubertal girls than in pre-pubertal boys.
In summary, puberty is a time of change in the
subgingival ecosystem. Changes in relative abundance
contribute to this compositional shift, and several
species belonging to the genera Veillonella, Strepto-
coccus, Capnocytophaga, Fusobacterium, Prevotella
and Actinomyces, increase in abundance. Available
evidence indicates that in pre-, circum- and post-
pubertal children, as well as young adults, increases in
the detection frequency and proportion of black-pig-
mented Bacteroides are preceded by, and correlate
with, an increase in gingival inflammation. In pubertal
children, correlations between an increase in sex ste-
roid levels and the proportions of these organisms
have not been demonstrated except in one study. Ta-
ken together, these data suggest that the presence of
gingival inflammation, rather than the increase in
levels of sex steroids, correlates strongly with the
prevalence and levels of black-pigmented Bacteroides.
Menstrual cycle
Endocrinal changes during the menstrual cycle
Females of reproductive age exhibit cyclic changes in
the levels of circulating sex steroids over 2530-day
periods, referred to as the menstrual cycle. The
menstrual cycle in humans is divided into the follic-
ular or proliferative phase, the luteal or secretory
phase, and menstruation (Fig. 3). The transition from
follicular to luteal phase is defined to the onset of
ovulation. Estrogen levels begin to increase in the
follicular phase, initiating a surge in the level of
luteinizing hormone and culminating in ovulation.
After ovulation, the levels of progesterone and
estrogen increase during the luteal phase; however, if
Fig. 3. Hormonal and neuronal
control of menstruation and preg-
nancy.

the ovum is not fertilized, the corpus luteum invo-
lutes, causing a sharp decrease in the levels of
estrogen and progesterone, leading to menstruation.
Gingival changes during the menstrual cycle
Increases in gingival inflammation, bleeding, crevic-
ular fluid flow and the presence of oral ulcers have
been reported during the menstrual cycle. Muhle-
mann (81) was the first to report on gingival changes
during the menstrual cycle, and coined the term
gingivitis intermenstrualis to describe a condition of
bright red, hemorrhagic lesions in the interdental
papilla. The literature on the effect of hormonal
fluctuations during the menstrual cycle on the gingiva
is sparse, and although some studies have indicated
an association between ovulation and the amount of
gingival crevicular fluid (38, 64, 65), the majority have
demonstrated this only in female subjects with pre-
existing gingivitis (38, 64). The incidence of gingival
inflammation during the menstrual cycle is also the
subject of controversy, with Holm-Pederson & Loe
(38) reporting an increase in pre-existing gingivitis in
the pre-menstrual interval, and Machtei et al. (71)
finding no association. Thus, the evidence linking the
menstrual cycle to gingival changes is extremely
tenuous, with no clear consensus.
Bacterial changes during the menstrual cycle
Changes in levels of subgingival and salivary bacteria
during the menstrual cycle have been examined in a
few studies (10, 26, 93). Prout & Hopps (93) found a
possible association between ovulation and the levels
of anaerobic bacteria in saliva; however, the findings
of this study are based on a sample size of only six
subjects, several of whom were not available for
sampling during ovulation and menstruation, caus-
ing a significant reduction in the number of samples
available for bacterial analysis during these periods.
This meant that statistical comparisons could not be
performed, calling into question the conclusions of
the authors. In contrast, Calil et al. (10) found no
association between the menstrual cycle and the
salivary levels of anaerobic bacteria in a group of 17
women. The only study that has examined the asso-
ciation between menstruation and subgingival bac-
terial levels did not detect any correlation between
the periodicity of hormonal changes and fluctuations
in the microbial profile (26).
In summary, the scant evidence in the literature
does not support the hypothesis that hormonal fluc-
tuations during the menstrual cycle affect either the
gingiva or the subgingival biofilm. However, the suit-
Sex steroids and subgingival bacteria
ability of the menstrual cycle as a model to examine
the effects of hormonal changes on subgingival bacteria
must be carefully and critically examined, as these
hormonal fluctuations are rapid, occur over a relatively
short space of time (sometimes only a few hours), and
are repeated in periodic cycles. It should be questioned
whether these repeated cyclical fluctuations have a
long-term effect on a relatively stable, well-established
community, and whether studies can be designed to
effectively examine any possible effect on the bacterial
community over these small intervals of time.
Pregnancy
Endocrinal changes in pregnancy
Hormonal upsurges in pregnancy begin during the
luteal phase and steadily increase for 40 weeks until
parturition (86). Progesterone levels increase from 26
91 gmol/L at 5 weeks gestation to 3141087 gmol/L
during the 40th week. In contrast, the level of 17-hy-
droxyprogesterone remains steady until 9 weeks and
then increases almost threefold by the end of the third
trimester. Estradiol levels also remain stable during
the first trimester (11.13 gmol/L at 9 weeks), and
show a fivefold increase in the third trimester. Thus,
towards the end of pregnancy, estrogen and proges-
terone levels increase to approximately 10 and 30
times those of a normal menstrual cycle.
Gingival changes in pregnancy
It has been known for a number of years that hor-
monal surges of pregnancy are associated with
increased gingival inflammation (39, 116). Gingival
changes in pregnant women have been followed over
the three trimesters of pregnancy (first trimester,
weeks 112; second trimester, weeks 1326; third
trimester, week 27 to parturition) as well as post-
partum. The prevalence and severity of gingivitis
have been reported to be greater in pregnant women
than in non-pregnant controls in cross-sectional and
longitudinal studies (7, 32, 67, 99). Pregnant women
also have greater amounts of gingival inflammation,
deeper probe depths, higher gingival crevicular fluid
levels and more bleeding during all three trimesters
than after parturition. Localized gingival enlarge-
ments have also been reported during pregnancy (7).
These increases in gingival inflammation correlate
with increases in the levels of sex steroid hormones.
Subgingival bacteria in pregnancy
The association between pregnancy gingivitis and
subgingival bacteria has been examined using
113
Kumar

Table 2. Summary of evidence: studies on pregnant women

References Study Number of Clinical Bacterial profile Authors conclusions and

duration subjects, gender outcomes reviewers comments
(age range at
baseline in years)

Kornman & Longitudinal 20 pregnant, below GI and number B. melaninogenicus ss. Positive correlation
Loesche (57) 13 weeks gestation of bleeding intermedius increased between hormonal levels
(2033) sites increased between 17 and and B. intermedius
11 non-pregnant from 13 to 24 weeks and during weeks 1724 of
(2236) 24 weeks and decreased from pregnancy
decreased 24 weeks to post- Positive correlation
thereafter partum between hormonal
Post-partum levels uptake by plaque
were similar to 1st bacteria and gestational
trimester levels age in 2nd trimester
Small sample size;
correlation between
hormones and BPB
evident only between
weeks 17 and 24; BPB
levels decreased with
increasing hormonal
levels after 24 weeks.
Jensen Cross- 54 pregnant Pregnant group BPB levels were greater Positive association
et al. (41) sectional 27 non-pregnant had greater GI in pregnant and oral between pregnancy,
23 non-pregnant on than non- contraceptive groups contraceptive use and
oral contraceptives pregnant and than controls BPB levels.
(1840) contraceptive No association between
groups clinical inflammation
Pregnant group and BPB levels.
had greater Hormone levels were not
gingival measured; no
Crevicular Fluid information on stage of
than non- pregnancy
pregnant group

Jonsson Cross- 9 males (2064) No difference in B. intermedius levels No correlation between

et al. (44) sectional 14 non-pregnant PD, BOP and PI similar in all females clinical parameters and
females (2238) 7 between female B. intermedius levels pregnancy.
in the 1st trimester groups correlated with No correlation between
and 7 in the 2nd Males had progesterone and BPB and pregnancy.
trimester (1940) greater PD and cortisol in males and No correlation between
PI than females with cortisol in BPB and estradiol in
BOP and PD did pregnant females females.
not increase Small sample size;
during cross-sectional study.
pregnancy
Raber- Longitudinal 9 females in 2nd More swelling, Subgingival levels of No differences between
Durlacher trimester (2034) redness and P. intermedia pregnancy and post-
et al. (94) 8 post-partum bleeding during increased over partum in terms of levels
experimental 14 days of of P. intermedia during
gingivitis in experimental development of gingivitis.
pregnancy than gingivitis in Positive correlation
post-partum pregnancy, but not between inflammation
post-partum and P. intermedia during
No differences in levels pregnancy, but not
of P. intermedia post-partum
between pregnancy Longitudinal study
and post-partum at examining same subjects
any time point over during and after
14 days pregnancy
Single examiner

114
 Sex steroids and subgingival bacteria

Table 2. (Continued)

References Study Number of Clinical Bacterial profile Authors conclusions and

duration subjects, gender outcomes reviewers comments
(age range at
baseline in years)

Adriaens Longitudinal 20 females at No change in Levels of N. mucosa No correlation between

et al. (77) 12 weeks gestation BOP or PD dur- increased from week gestational age and
(2642) ing pregnancy or 12 to post-partum, composition of
post-partum and levels of 17 subgingival biofilm
species decreased
during the same
period
No changes were
observed during the
three trimesters of
pregnancy

Yokoyama Cross- 22 pregnant (2541) Pregnant group No difference between Positive correlation
et al. (124) sectional 15 non-pregnant had a greater groups in terms of between pregnancy and
(2440) number of sites levels of salivary C. rectus levels.
with PD > C. rectus Salivary levels of bacteria
4 mm and BOP C. rectus levels measured.
correlated with E2- No information on
estradiaol levels in all gestational age of
subjects subjects.
C. rectus levels Stronger correlations in all
correlated with PD > subjects than in pregnant
4 mm in all subjects groups alone. Stronger
P. gingivalis and correlation between
F. nucleatum levels P. gingivalis and
correlated with E2 F. nucleatum and E2-
levels in pregnant estradiaol levels than
group between C. rectus and
E2-estradiaol levels in
pregnant group.
Gursoy Longitudinal 30 pregnant (2632) Pregnant group Subgingival No correlation between
et al. (31) 24 non-pregnant had greater BOP P. intermedia levels subgingival and salivary
(2733) BOP increased increased between levels of P. intermedia
from the 1st to 3rd trimester and and gestational age.
2nd trimester post-partum Salivary The authors indicate a
and decreased P. intermedia levels trend towards increasing
thereafter decreased between levels of P. intermedia in
post-partum and end the 1st and 2nd
of lactation trimesters, but this is not
statistically significant.
PD, probing depth; GI, gingival index; PI, plaque index; BOP, bleeding on probing; BPB, black-pigmented Bacteroides.

cultivation, microscopy, DNADNA checkerboard
hybridization and quantitative real-time PCR (1, 9,
41, 44, 54, 57, 79, 82). The findings of these studies are
summarized in Table 2. Cultivation-based investiga-
tions have shown an increase in the prevalence and
levels of black-pigmented Bacteroides, particularly
Bacteroides melaninogenicus ss. intermedius, during
the second trimester of pregnancy (41, 57, 82, 94).
These increases were reported to occur indepen-
dently of increases in plaque levels; however, they
were accompanied by increases in Gingival Index and
subgingival levels of anaerobic bacteria. The cross-
sectional study by Jensen et al. (41) reported a 50-fold
greater level of B. melaninogenicus during pregnancy,
and a longitudinal investigation by Kornman & Loe-
sche (57) reported a fivefold increase from the first to
the second trimester of pregnancy. It is interesting to
note that the levels of Bacteroides decreased in the

115
Kumar
third trimester of pregnancy, despite increased serum
estradiol levels (57). Closer examination of the results
of the longitudinal investigation revealed that the
greatest level of gingival inflammation was observed
during the second trimester, and decreased sharply
immediately thereafter. Further, the number of
bleeding sites increased between 13 and 28 weeks,
with a peak at 2124 weeks and a decrease during the
third trimester. Thus, neither bacterial levels nor gin-
gival inflammation correlated with serum estradiol
levels, a very strong temporal relationship was ob-
served between gingival bleeding and the level of
black-pigmented Bacteroides. The cross-sectional
investigation did not specify the gestational age of the
pregnant women, making it difficult to confirm or
disprove the findings of the longitudinal investigation.
The association between levels of Bacteroides and
gingival inflammation was also investigated by
experimental induction of gingivitis during the
second trimester of pregnancy and post-partum in
the same women (94). Sites with inflammation were
colonized by greater numbers of P. intermedia during
pregnancy compared to post-partum. However, the
level of gingival inflammation was also significantly
higher during pregnancy than post-partum. Hence, it
is possible that the gingival inflammation had a
greater contribution in altering the composition of
the subgingival microbial community than female
sex steroids.
When men, pregnant and non-pregnant women
with similar levels of gingival inflammation (based on
probing depth and bleeding score) were compared,
there were no differences in the proportions of
Bacteroides between the three groups, nor was a
correlation detected between the levels of the
organisms and progression of pregnancy (44).
A recent investigation using DNADNA hybridiza-
tion for 37 species found no significant change in the
levels or prevalence of any species during the course
of pregnancy (1). The level of Neisseria mucosa was
found to increase from early pregnancy to post-
partum, and the levels of 17 species, including
P. melaninogenica and P. intermedia, decreased from
week 12 to post-partum.
Another cross-sectional study that used molecular
methods to examine salivary levels of five periodontal
pathogens (Porphyromonas gingivalis, Aggregati-
bacter actinomycetemcomitans, Fusobacterium nucle-
atum, P. intermedia and Campylobacter rectus) found
significant correlations between salivary estradiol
levels and C. rectus in pregnant women (124).
At the time of writing this article, the most recent
evidence comes from two longitudinal evaluations of
116
pregnant women and non-pregnant controls (11, 31).
Gursoy et al. (32) followed pregnant women through
to the end of lactation, and P. intermedia and P. ni-
grescens were characterized using both cultivation
and PCR methods. The only statistically significant
finding was an increase in the subgingival levels of
P. intermedia after parturition, and a decrease in the
salivary levels of the same organism after lactation.
There were no significant increases in the levels of
these organisms during the various trimesters of
pregnancy, even though bleeding on probing in-
creased during the first and second trimesters and
was significantly greater than in non-pregnant con-
trols (32). Although the authors concluded that
P. intermedia levels increased during the first and
second trimesters of pregnancy, these increases were
not statistically significant. Carrillo-de-Albornoz et al.
(11) found that the levels of C. rectus and P. inter-
media increased significantly post-partum compared
to the three trimesters of pregnancy.
In summary, although early evidence from
cultivation-based approaches indicated that hor-
monal surges during pregnancy may play a role in
increasing the subgingival levels of black-pigmented
Bacteroides, more recent investigations using molec-
ular methods did not corroborate these findings. As
in puberty, the levels of black-pigmented Bacteroides
demonstrate a stronger correlation with gingival
inflammation during pregnancy than with the levels
of circulating hormones. Thus, at present, there is
very little evidence linking pregnancy and preferen-
tial colonization by certain bacterial species.
Biological effects of sex steroids on
bacteria
Several mechanisms have been proposed to explain
the action of sex steroids on oral bacteria. It has
been suggested that estradiol and progesterone
substitute for vitamin K compounds as electron
carriers in bacterial anaerobic respiration (56).
Anaerobic respiration in bacteria is mediated by
transfer of electrons from an electron-rich donor,
such as lactate or formate, to a quinone molecule.
This quinone molecule subsequently transfers these
electrons to the terminal electron acceptor fumarate.
By substituting for naphthoquinone (vitamin K), fe-
male sex steroids may play a critical role in bacterial
respiration.
It has also been hypothesized that sex steroids
increase the levels of formate-producing bacteria
(e.g. P. intermedia), thereby indirectly increasing the

growth of C. rectus, a formate metabolizer (123, 124).
However, this has not been proven to date.
Several bacterial species, for example Trepo-
nema denticola, P. gingivalis, A. actinomycetemcom-
itans and P. intermedia, are capable of stimulating
5a-reductase activity in human gingival fibroblasts,
leading to the formation of dihydroxytestosterone
from testosterone (107, 108). It is thought that tes-
tosterone metabolites contribute to bacterial growth
and virulence; however, there is no evidence in the
literature to support this hypothesis.
Adverse pregnancy outcomes and
bacteria
Almost 70% of all perinatal deaths are attributable
to pre-term births. The infection inflammation
hypothesis of pre-term birth suggests that subclinical
infection triggers an inflammatory cascade that
causes cervical ripening, myometrial contractions,
membrane rupture and pre-term birth. Symptomatic
maternal infections of the respiratory and urinary
tracts, as well as subclinical infections of the genital
tract, for example bacterial vaginosis, have been
associated with pre-term birth, cerebral palsy, respi-
ratory depression and other neonatal sequelae (74).
Gibbs (29) has extensively reviewed the evidence
linking adverse pregnancy outcomes and maternal
infection. Briefly, the presence of certain organisms
in the lower genital tract, positive cultures of amni-
otic fluid and various markers of infection are asso-
ciated with pre-term births. Pre-term birth has been
observed in animal models after maternal infection.
Several clinical trials have also shown reduced rates
of pre-term births or a deferral of premature labor
after antibiotic therapy.
Pregnancy outcomes and subgingival
bacteria
The infection hypothesis of pre-term birth, together
with evidence implicating female sex steroids in
altering subgingival microflora, has led to several
lines of research examining the association between
disease-associated periodontal bacteria and neonatal
health; however, the results have varied widely (15,
76, 84, 85, 88, 100). Several cross-sectional and lon-
gitudinal studies have suggested that the presence of
certain oral bacteria increases the risk of adverse
pregnancy outcomes such as pre-term birth or
low-birth-weight infants. Pre-term low birth weight
has been found to be associated with high maternal
Sex steroids and subgingival bacteria
levels of Tannerella forsythia, C. rectus, P. interme-
dia, P. nigrescens and P. gingivalis in subgingival
plaque (76, 88, 100). P. gingivalis was detected in
both amniotic fluid and subgingival plaque of 31%
of women with threatened pre-term labor (61).
Higher maternal salivary levels of A. naeslundii and
low levels of Lactobacillus casei have also been found
to be associated with pre-term births (15). Several
mechanisms (described below) have been proposed
and tested to explain the correlation between
adverse pregnancy outcomes and periodontal infec-
tion.
Initiation of intrauterine inflammation by
circulating pro-inflammatory mediators
It has been shown that periodontal bacteria elicit a
florid systemic host response, resulting in high levels
of prostaglandin E2 and cytokines in the circulation
and the placenta (66, 72). Periodontal therapy has
been shown to decrease both the subgingival bacte-
rial load and the serum levels of interleukin-6 (88).
This was accompanied by a 3.8-fold decrease in the
rate of pre-term births. Together, these results sug-
gest that periodontal bacteria may contribute to pre-
term birth by indirectly inducing inflammation in the
uterus.
Fetal immune response to maternal pathogens
Certain oral bacteria, for example F. nucleatum, some
streptococci and C. rectus, are able to translocate
across the placental barrier (3, 34, 72). Pre-term infants
showed an increase in the levels of circulating anti-
bodies to C. rectus compared to full-term newborns
(72), suggesting another possible mechanism by which
these organisms contribute to pre-term birth.
However, other studies, including a recent large-
scale investigation on 823 pregnant women, found no
association between the presence or levels of sub-
gingival periodontal pathogens and adverse preg-
nancy outcomes (75, 84, 85). These studies also found
little or no improvement in pregnancy outcomes
after periodontal therapy. A recent systematic review
concluded that periodontal disease may not play a
causal role in the pathogenesis of pre-eclampsia (60).
It has been suggested that poor pregnancy outcomes
in women with periodontal disease may be the
result of an epiphenomenon of an exaggerated
inflammatory response to pregnancy.
In summary, a large body of evidence indicates an
association between periodontal pathogens, poor
periodontal health and adverse pregnancy outcomes,
specifically low-birth-weight infants and pre-term
birth. However, there is very little evidence that
117
Kumar
periodontal disease plays a causal role in these dis-
eases; rather, emerging evidence suggests that both
these conditions may be manifestations of a hyper-
inflammatory response common to both.
Exogenous hormone
administration
Exogenous estrogens or progestins are used pre-
dominantly as oral contraceptives in pre-menopausal
women and as hormonal replacements post-meno-
pausally. Although exogenous hormonal ingestion
may begin after menarche and continue after men-
opause, the formulations and doses used in these
two treatments are significantly different. Another
significant difference is that contraceptive regimens
require periodic interruption of the exogenous hor-
mones for a week, altering the levels of endogenous
steroids during that period. The following sections
discuss the current state of knowledge on the effect of
exogenous hormone administration on the peri-
odontium and subgingival bacteria.
Contraceptive therapy
Oral contraceptives containing 17a-ethinyl estradiol
and progestins such as norethindrone, levonorgestrel
or norgestimate were approved for use in the
USA almost 50 years ago. The levels of estradiol
have been progressively reduced over the years,
and current prescriptions contain no more than
35 lg dose. In addition, other oral contraceptive
regimens that result in four menstrual cycles per
year are available. These formulations use 84 days of
continuous exogenous hormone therapy followed by
a week-long pill-free period. Non-oral contraceptive
methods include hormone formulations incorpo-
rated into vaginal rings, transdermal patches,
injectable depot preparations and progesterone im-
plants. As 8086% of sexually active women use oral
contraceptives during their reproductive years (16),
it is important to understand the effect of this
cyclical fluctuation in endogenous and exogenous
hormonal levels on various organ systems, including
the periodontium.
Subgingival bacteria in women using oral
contraceptives
Two studies, one cross-sectional and the other
longitudinal, have examined shifts in subgingival
bacteria associated with oral hormonal contraceptive
use (41, 54). The cross-sectional study did not define
118
the type of oral contraceptive used by the women,
and the longitudinal study examined two groups: one
using ethinyl estradiol and desogestrel and the other
using ethinyl estradiol and dienogest. Both types of
contraceptive contain third-generation progestins.
The cross-sectional study found a 16-fold increase in
black-pigmented Bacteroides in the oral contracep-
tive users compared to non-contraceptive users, and
the longitudinal study found an increase in the levels
of P. intermedia after 20 days of oral contraceptive
use. There was also a difference in the levels of this
organism between the two groups of oral contra-
ceptive users at 20 days.
The evidence from these two studies indicates that
oral contraceptives influence the composition of
the subgingival microbial community. However, as
mentioned above, hormonal formulations and dos-
ages, as well as modes of administration, have
changed significantly over recent years, especially
within the last decade, and further studies are re-
quired to examine whether newer-generation oral
contraceptives influence the composition of the
subgingival microbiome.
Hormone replacement therapy
Hormone Replacement Therapy (HRT) or Hormone
Treatment (HT) is given to surgically menopausal
(women who have undergone hysterectomy, ovar-
ectomy or pan-hysterectormy), peri-menopausal and
post-menopausal women to reverse the effects of
lower levels of circulating estrogen. While estrogen
alone is capable of reversing post-menopausal
symptoms, these women are at high risk of devel-
oping uterine carcinoma. Progesterone is therefore
given together with estrogen in a combina-
tion therapy to protect the uterus from neoplastic
changes.
Conjugated equine estrogens are commonly used
in combination with progestins in hormone
replacement therapy preparations. Recently, bio-
identical human estrogens have been introduced in
an effort to decrease the cardiovascular side-effects
seen with the equine estrogens. The traditional form
of hormone replacement therapy is sequential com-
bined hormone replacement therapy, which delivers
daily doses of estrogen alone for 2 weeks followed by
a combination of estrogen and progesterone for the
next 2 weeks. Newer regimens have been introduced
that deliver constant daily doses of both hormones
(i.e., continuous combined hormone replacement
therapy). Hormones can be delivered by several
methods; while oral pills are the most common form,

transdermal patches, vaginal rings and injections are
being more commonly administered.
Hormone replacement therapy and gingival health
There is a lack of studies examining gingival chan-
ges after menopause and hormone replacement
therapy. Hormone replacement therapy has been
reported to improve probing depths in post-meno-
pausal women (69). There are no studies examining
the effect of menopause or hormone replacement
therapy on the prevalence and levels of subgingival
bacteria.
Black-pigmented Bacteroides and
periodontal disease
The term black-pigmented Bacteroides refers to
a large group of gram-negative, anaerobic rods
that produce black or brown-pigmented colonies
on blood agar plates. Three large groups (Bactero-
ides asaccharolyticus, Bacteroides intermedius and
Bacteroides melaninogenicus) were initially identi-
fied based on phenotypic characteristics. This tax-
onomy became obsolete after advances in bacterial
characterization led to identification of at least 10
genetically and phenotypically distinct species.
Several of these species have now been identified
as belonging to other genera, for example Por-
phyromonas, Bacteroides and Prevotella. Recent
investigations using DNA sequencing have revealed
the existence of a large number of as yet unchar-
acterized Prevotella and Porphyromonas species
(97). It is possible that several of these species will
Bacteroides melaninogenicus
Based on phenotypic characteristics
B. asaccharolyticus B. Intermedius
Prevotella asaccharolytica Prevotella Intermedia
DNADNA homology DNADNA homology
B. loescheii
Porphyromonas Porphyromonas Prevotella Prevotella
gingivalis endodontalis Intermedia corporis
Prevotella
nigrescens
Sex steroids and subgingival bacteria
also fall into the category of black-pigmented
Bacteroides. This taxonomic evolution, shown in
Fig. 4, presents a challenge when comparing studies
that use different nomenclatures.
Several studies prior to 1990 used production of
black-pigmented colonies on blood agar plates to
screen for these bacteria. However, certain strains of
P. melaninogenica and P. denticola are slow pigment
producers, with some not producing pigments even
after 3 weeks of incubation (119). Phenotypically, a
non-pigmented colony of P. melaninogenica or
P. denticola is indistinguishable from Prevotella oris
or other oral Prevotella species (90). Therefore, pig-
mentation on blood agar is not considered a reliable
characteristic for identification of black-pigmented
Bacteroides.
Virulence of black-pigmented
Bacteroides
Although it has been known for several years that
Bacteroides spp. possess several virulence factors,
the characteristics of individual species have not
been well elucidated due to limitations associated
with bacterial identification and nomenclature. One
member of the black-pigmented Bacteroides group,
P. gingivalis has been extensively studied for several
decades; however, recent investigations have revealed
several mechanisms associated with members of the
genus Prevotella. These gram-negative bacteria affect
the orientation of Th1 Th2 lymphocytes through
stimulation of Toll-Like Receptor 2 (TLR2) (51). They
also produce beta-lactamase, a property that confers
antibiotic resistance to drugs carrying the beta-lactam
B. melaninogenicus
Prevotella melaninogenica
DNADNA homology
B. denticola
Prevotella Prevotella
loescheii melaninogenica
Prevotella Fig. 4. Taxonomy of black-pig-
denticola mented Bacteroides.
119
Kumar
ring, for example penicillin. This not only allows them
to survive penicillin therapy, but also shields penicil-
lin-susceptible bacteria in the biofilm from the drug
(8, 35). They contain cysteine and serine proteases
that enable heme and protein degradation (30, 121).
This proteolytic activity may be responsible for
causing spreading odontogenic infections. Prevotella
species also produce a potent lipopolysaccharide that
has been shown to stimulate osteoclastogenesis and
release of nitric oxide from macrophages in vitro (13,
52). Another recently identified virulence factor is an
exopolysaccharide that appears to confer resistance
to phagocytosis by polymorphonuclear leukocytes
(120). All of these virulence factors may contribute to
their role in the pathogenesis of periodontal diseases.
Prevalence of Prevotellae in health and
disease
Prevotella species are constituents of host-associated
ecosystems in the upper respiratory tract, female
genital tract and oral cavity. These organisms con-
tribute to polymicrobial infections such as bacterial
vaginosis, chronic sinusitis and periodontitis (8, 87,
126). Recent investigations have revealed several
novel, previously uncultivated, phylotypes within
these communities (87). Within the oral envi-
ronment, Prevotella species (P. intermedia and
P. nigrescens) have been found in both periodontal
healthy sulci and diseased pockets, although the de-
tection frequencies and proportions in health are
significantly lower than in disease (63, 126). Prevo-
tella species have been identified in acute necrotizing
gingivitis (68), and have been detected in saliva and
supragingival plaque of periodontally healthy chil-
dren as young as 6 years of age (59, 110). There is
evidence for intra-familial and maternal transmission
of these organisms (55, 115). Several researchers have
found a correlation between deep probe depths and
the levels of these species in subjects with gingivitis
as well as periodontitis (50, 126).
In summary, Black-pigmented Bacteroides is a
classification term used to signify a genetically het-
erogeneous group of organisms. This classification
was based on phenotypic characterization and is now
obsolete. These bacteria are found in the subgingival
habitat in both health and disease, and in signifi-
cantly greater abundance in association with gingi-
vitis and periodontitis. These organisms express
several virulence factors that may contribute to their
role in the pathogenesis of both gingvitis and
periodontitis.
120
Summary and conclusions
Puberty, the menstrual cycle, pregnancy and the peri-
menopause are periods of significant hormonal
surges, while a dramatic decrease occurs after men-
opause. An increase in gingival inflammation has
been associated with these periods of increased
hormonal levels, and it was thought that sex steroids
play an etiological role by altering the gingival
microanatomy as well as the subgingival bacterial
community. Early investigations suggested that in-
creases in female sex steroids led to preferential
colonization by black-pigmented Bacteroides, espe-
cially P. intermedia. Evidence to support this comes
from the following data:
1. Black-pigmented Bacteroides were found at higher
levels in pregnant than non-pregnant women.
2. Black-pigmented Bacteroides were found at higher
levels in women using oral contraceptives.
3. Estrogen and progesterone substitute for Vitamin
K, an important nutrient for black-pigmented
Bacteroides.
However, several confounding variables affect
these studies, the most important of which is that the
subgingival microbiome is affected by several
environmental factors, including inflammation. As
ovarian hormones exert a predominantly pro-
inflammatory effect on the gingiva, it is difficult to
separate the direct effects of female sex steroids from
the effects of gingival inflammation on the gingival
community. Other factors that need to be considered
are that very few studies demonstrate a positive
correlation between hormonal and bacterial levels,
and positive correlations between gingival inflam-
mation and bacterial profiles are consistent in all
studies. Prevotella species are found in periodontal
health and disease, and show a higher prevalence at
sites with deep probing depths and inflammation.
Prevotella species have also been found in all forms
of gingivitis, including necrotizing ulcerative gingivi-
tis and HIV-related gingivitis. Further, studies using
cultivation-based approaches for bacterial identifi-
cation diverge significantly in terms of bacterial
nomenclature, and are also limited since these
methods do not allow for highly discriminatory bac-
terial identification, and make it difficult to draw
definitive conclusions.
In conclusion, at present, there is no definitive
evidence linking elevated states of ovarian hormones
to preferential enrichment of the subgingival mi-
crobiome for selected species.
Periodontology 2000, Vol. 61, 2013, 125159
Printed in Singapore. All rights reserved
Oral contraceptives and the
periodontium
P H I L I P M. P R E S H A W
Oral contraceptives were introduced in the 1960s.
The first oral contraceptives contained high doses of
estrogens (150 lg) and progestins (9.85 mg), and
were associated with a high risk of cardiovascular
events (122). From the outset, epidemiological stud-
ies identified an unacceptably high risk of adverse
events, and this drove the development of oral con-
traceptives that contained progressively lower doses
of estrogen and progesterone. For example, a review
of almost 2000 safety reports submitted to the UKs
Committee on Safety of Medicines from 1964 to 1977
revealed that use of oral contraceptives containing
30 lg estrogen was associated with significantly fewer
instances of death or ischemic heart disease than use
of those containing 50 lg. Similarly, there was a sig-
nificant positive association between an increase in
the dose of the progestin norethisterone acetate from
1.0 to 4.0 mg and deaths from stroke and ischemic
heart disease (99).
Observations such as these led to the development
of modern, low-dose formulations of oral contracep-
tives, which typically contain estrogen doses of 20
35 lg day and progestin doses of 0.51.0 mg day,
and are associated with a much lower risk of arterial
and venous events than the original formulations.
Indeed, it can now be concluded that for healthy, non-
smoking women there is no increased risk of myo-
cardial infarction or hemorrhagic or ischemic stroke
associated with oral contraceptive use, though there is
an increased risk of venous thromboembolism
(though this risk is approximately half that associated
with pregnancy (103). However, the product labeling
of currently available oral contraceptives still reflects,
to some degree, the findings of the early epidemio-
logical studies that investigated the unwanted effects
of high-estrogen oral contraceptives, even though
these are no longer marketed. In one survey in the
early 1990s, approximately half of the women in-
volved believed that use of oral contraceptives carried
significant health risks (111). Furthermore, 8090% of
 2013 John Wiley & Sons A/S
PERIODONTOLOGY 2000
the women surveyed were unaware of the health
benefits associated with oral contraceptive use, such
as decreased risk of ovarian and endometrial cancers,
even though these non-contraceptive benefits, and
others, are listed in the product labeling.
Clinical reports of an increased prevalence of gin-
gival (and possibly periodontal) disease associated
with increasing levels of plasma sex steroid hormones
have appeared on an intermittent basis in the dental
and medical literature throughout the last century
(95). For example, clinical studies reported the inci-
dence and severity of gingival disease to be positively
correlated with increased sex steroid hormone con-
centrations during puberty (132), the menstrual cycle
(73) and pregnancy (51, 52). Reports from the 1960s
and 1970s also linked the use of oral contraceptives
with gingival diseases. The key point to remember is
that these early studies linking gingival inflammation
to oral contraceptives involved investigations of the
high-dose oral contraceptives then available. As a
result, many practitioners and patients have labored
under the belief that oral contraceptives constitute a
risk factor for gingivitis, despite the fact that this
conclusion is based on clinical data that are now 30
40 years old. As a result, periodontal management of
women using oral contraceptives has sometimes
been unnecessarily time-consuming, over-zealous
and probably inappropriate (95).
The purpose of this review is to consider the lit-
erature pertaining to the use of oral contraceptives
and the impact, if any, of these commonly used drugs
on the gingival and periodontal tissues. The historical
perspective is important, as moves to decrease hor-
mone concentrations in oral contraceptives to reduce
the risk of significant medical adverse events also had
a positive impact on reducing the risk for gingival
disease. The use of oral contraceptives worldwide
continues to increase, due to the combined effects of
population growth and increased availability and
choice of contraceptives. It is therefore important
125
Preshaw
that all members of the dental team are aware of the
development and pharmacology of these agents, and
that they possess up-to-date knowledge regarding
any effect on the periodontium in order to be able to
adequately advise patients who are using oral con-
traceptives and address any concerns that they may
have.
A brief history
As many women are always doing, doctoring
themselves Hippocrates (circa 400 BC), Illnesses of
Women
In the world of the ancient Greeks, families were
relatively small, life expectancy was short (for those
who survived childhood, life expectancy was
approximately 45 years for men and 36 years for
women), and recurrences of famines and war had
devastating effects on population growth (97). If
children were conceived, diseases during pregnancy
and maternal malnutrition resulted in high levels of
fetal death. Interest in influencing procreation often
focused on trying to predetermine the sex of children
in a somewhat cryptic fashion (potential fathers must
live according to a regimen inclining more to water
if they wanted to have a girl, and inclining more to
fire if they wanted a boy). The most effective form of
contraception was coitus interruptus, although a
number of herbal potions were believed to have
contraceptive effects. De Materia Medica, written by
the Greek physician Dioscorides, listed ingredients
for drinks that would induce sterility, such as leaves
from hawthorn trees, ivy, willow and poplar (97).
These were supplemented by topically applied agents
or barrier methods of contraception. The risks of
abortion were significant, and therefore unwanted
pregnancy was to be avoided at all costs.
The ancient Romans took a similar interest in
methods to control procreation. At times, rewards
were given to those families who had greater num-
bers of children (such as the allocation of land to
fathers of three or more children instigated by Julius
Caesar in 59 BC) as a means of expanding the Roman
influence through population growth. However,
among the elite of Roman families (and these are the
only ones for whom records usually remain, for
example in the form of funeral inscriptions), there
were typically only one or two children per family.
This undoubtedly partly resulted from high infant
mortality rates, but also probably from attempts to
control pregnancy, particularly by females, given the
126
risks of childbirth, and also as a means of providing a
more secure inheritance for the existing children. A
variety of pharmaceutical approaches to contra-
ception were employed, including use of willow
seeds, bryony mixed with ox urine, wine, wallflower
seed, myrtle, myrrh and white pepper, in addition to
barrier methods of birth control.
During the Middle Ages, much of the established
medical knowledge of the Ancients was lost. How-
ever, the idea of controlling pregnancy was not
completely abandoned during this period. Family
size was small, partly due to short life expectancy, but
also because society was primarily rural and small
farms could not support large families. Extended
periods of breastfeeding of existing babies were
sometimes used as a means of preventing further
pregnancies, but there are also references to other
birth control techniques. In the Parsons Tale,
Chaucer refers to women drynkynge venenouse
herbes thurgh which she may nat conceyve, in
addition to descriptions of barrier methods and
abortions.
The birth control movement in more recent times
is typically considered to have begun in France in the
latter parts of the 18th century, when practices to
curb population growth were proposed and de-
fended. Advocates for birth control later became
known as neo-Malthusians after Thomas Malthus
who, although opposed to contraception, wrote an
Essay on Population in 1798 that considered the so-
cial impact of fertility, and particularly challenged the
view widely held by governments that greater num-
bers in the population led to greater prosperity. He
believed that, beyond a certain point, this argument
no longer held true, as population growth always
places pressure on resources. Through the 19th and
early 20th centuries, coitus interruptus and ineffec-
tive barrier methods continued to be the most
important forms of birth control.
In the early part of the 20th century, a New York
housewife, Margaret Sanger (18791966), was in-
spired by her experiences as a nurse working with
poor women who had unwanted pregnancies to
promote methods of controlling family size. She
introduced the term birth control as a more positive
description of methods of controlling fertility than
the previously used neo-Malthusianism, thereby
distancing what we now refer to as family planning
from some of its earlier economic and political con-
notations. However, a monograph that she wrote in
1914 entitled Family Limitation led to prosecution by
the US government for violation of postal obscenity
laws, on the basis that the material she wrote con-

tained obscene, lewd or lascivious material sent
through the US postal system. She left the country for
Europe, but returned in 1916, whereupon, as a result
of the positive public support that the charges against
her had generated, all charges were dropped. Sanger
then established a birth control clinic in Brooklyn,
which was promptly shut down by police and she was
arrested and jailed. However, she continued to cam-
paign for legislative reforms, and in 1923, changes to
the law made it legal for physicians to dispense
contraceptive information (97).
In the middle part of the 20th century, Sanger
continued to pursue the concept of a reliable,
simple, safe and effective means of contraception.
Researchers had found that ovulation in animals
could be prevented by injections of estrogen (97), and
one researcher, Gregory Pincus, was introduced to
Sanger who secured financial backing for his re-
search. In 1951, Pincus found that progesterone
inhibited ovulation, and clinical trials in humans
were begun in Puerto Rico in 1956 in collaboration
with John Rock, a gynaecologist from Harvard. He
had initially opposed contraception, but was alarmed
by the ever-increasing population, and declared that
the pill was a natural contraceptive that could be
used by Catholics (which he was himself). This
opinion was not shared by the Vatican. However, in
1960, the US Food and Drug Administration ap-
proved the first oral contraceptive, which was pro-
duced by Searle, and other companies quickly began
to develop their own oral contraceptive formulations.
The oral contraceptive pill soon began to be ac-
cepted more widely by the medical profession, as it
struck a chord with physicians concepts of phar-
maceuticals being used in the name of scientific
progress. After all, physicians prescribe medications,
and it was also much easier at that time for doctors to
prescribe a pill than suffer the embarrassment of
demonstrating the use of fiddly barrier methods of
contraception. The oral contraceptive pill was viewed
as something of a panacea, despite the emergence of
a variety of side-effects with the earliest formulations
of these drugs, and their association with increased
incidences of cancer and blood clotting. By 1970,
nearly six million married women of reproductive age
in the USA were using oral contraceptives (more than
one in five) (149). Oral contraceptives were by far the
most popular method of contraception, accounting
for more than one-third of all contraceptive practices,
and were disproportionately more popular among
younger women. By the 1980s, approximately 90% of
married couples in most developed countries were
using contraceptives, and approximately 20% were
Oral contraceptives and the periodontium
using oral contraceptives. In 2002, there were
approximately 120 million users of oral contracep-
tives worldwide. This number is likely to increase
significantly in the future as a result of population
growth and increased availability of contraceptives,
with increases in the use of oral contraceptives being
particularly great in developing countries (8). In 2009,
according to the United Nations Population Division
of the Department of Economic and Social Affairs,
the contraceptive pill was the third most commonly
used form of contraception (after female sterilization
and intrauterine devices), used by 8.8% of women
worldwide between the ages of 15 and 49 who
are married or in union (http://www.un.org/esa/
population/unpop.htm ). In more developed regions,
the pill was found to be used by 18.1% of women
(and was the most commonly used form of contra-
ceptive), compared to 7.2% of women in less devel-
oped regions.
Three generations, a pill scare, and
legal wranglings
The earliest concepts of the contraceptive pill were
based on the idea of a synthetic progestin that would
inhibit ovulation. Progestin-only methods suffer from
certain disadvantages, in that, although they inhibit
ovulation, they are associated with irregular bleeding.
Formulations that also include estrogen maintain the
endometrium and prevent unscheduled bleeding.
These formulations are referred to as combined oral
contraceptives. Modern combined oral contracep-
tives contain two main ingredients: a low dose of
estrogen and a progestin. In broad terms, the oral
contraceptives have been classified according to
generation (first-, second-, third-, and, most re-
cently, fourth-generation). This terminology usually
refers to the timing of introduction of the product,
although confusingly it sometimes refers to the tim-
ing of introduction of the progestin, and sometimes
to the structure of the carbon ring from which the
progestin is derived (estrane or gonane), occasionally
leading to the same oral contraceptive formulation
being classified differently in different studies. In this
review, the description will refer to the timing of
introduction of the particular oral contraceptive
product. First-generation products include those
approved for marketing in the US before 1973, sec-
ond-generation products include those approved for
marketing in the US between 1973 and 1989, third-
generation products include those approved for
marketing in the US or Europe between 1990 and
127
Preshaw
2000, and fourth-generation products are those ap-
proved for marketing in the US after 2000 (113).
The first generation of combined oral contracep-
tives were those initially developed in the 1960s.
First-generation progestins (norethindrone type) in-
cluded (among others) norethindrone (known as
norethisterone in Europe), ethynodiol diacetate and
lynestrenol. First-generation oral contraceptives
used a very high concentration of estrogen as the
primary means of contraception, but were associ-
ated with unacceptable adverse effects, particularly
an increased risk of deep vein thrombosis (see be-
low) and hypertension (99, 100). For example, eth-
inylestradiol (the estrogen component of most
combined oral contraceptives) increases the syn-
thesis of hepatic globulins, including angiotensino-
gen, which, when converted to angiotensin, in-
creases blood pressure (102). Furthermore, positive
correlations were identified between the dose of
estrogen and the risk of pulmonary embolism, deep
vein thrombosis, cerebral thrombosis and coronary
thrombosis (54).
Concerns about the unwanted effects of the early
combined oral contraceptives led to development
of the so-called second generation of oral contracep-
tives, which contained lower hormone concentrations
in an attempt to reduce the risk of adverse effects.
These formulations (which were introduced in the
mid-1970s) typically contained greatly reduced con-
centrations of estrogens and progestins, with ethiny-
lestradiol doses of 2035 lg, for example. At first, they
were not widely used, as clinicians worried that there
would be more breakthrough bleeding as a result of
the lower estrogen dose. However, it soon became
clear that this was not the case for doses of 3035 lg
ethinylestradiol, although there is some evidence of
more breakthrough bleeding with oral contraceptives
containing 20 lg ethinylestradiol (1, 3). The doses of
progestins also fell from typical doses of up to 10 mg
in the original first-generation combined oral con-
traceptives to <1 mg by the late 1980s as it became
clear that the same contraceptive efficacy was
achieved, but with fewer unwanted progestogenic
effects such as acne and weight gain (102, 103). Sec-
ond-generation progestins (norgestrel type) include
levonorgestrel and norgestrel. Levonorgestrel is
probably the most widely used progestin in modern
combined oral contraceptives, and is typically com-
bined with 30 lg ethinylestradiol. Today, combined
oral contraceptive formulations containing higher
doses of estrogens (e.g. >50 lg) are no longer indi-
cated because of their associated increased cardio-
vascular risks.
128
In the 1990s, a third generation of oral contracep-
tives was developed, which contained new progestins,
including norgestimate, gestodene and desogestrel.
These have greater affinity for progesterone receptors
than for androgen receptors compared with older
progestins, permitting a reduction in androgenic side-
effects (17). They were claimed to be as efficacious as
second-generation oral contraceptives in preventing
pregnancy, but contained lower doses of progestins.
Further, they were reported to result in fewer side-
effects such as headaches, weight gain and fluid
retention, and were developed with the aim of pro-
ducing less androgenic adverse effects such as the
hirsutism and acne that were typically associated with
first- and second-generation oral contraceptives. It
should be noted that although norgestimate was
developed after most second-generation oral contra-
ceptives and is therefore sometimes referred to as a
third-generation progestin, most researchers consider
it to be more akin to a second-generation progestin
because it is partially metabolized to the second-
generation progestin levonorgestrel and also because
the problems linked to third-generation oral contra-
ceptives (see below) are limited to desogestrel and
gestodene. For the purposes of this review, the term
third generation will now be used to refer only to oral
contraceptives containing desogestrel or gestodene.
Problems started to emerge with respect to third-
generation oral contraceptives in the mid-1990s
when a series of studies found that they approxi-
mately doubled the risk of venous thrombosis com-
pared to second-generation contraceptives (7, 56,
129, 152, 153). Four of these studies were published
together in one issue of The Lancet in 1995 (7, 56, 152,
153). The conclusions from these studies were that
users of low-estrogen dose combined oral contra-
ceptives containing desogestrel or gestodene were at
a higher risk of venous thromboembolism than users
of combined oral contraceptives containing levo-
norgestrel. In one of these studies, performed as part
of the World Health Organization Collaborative Study
of Cardiovascular Disease and Steroid Hormone
Contraception, it was found that, whereas users of
low-estrogen combined oral contraceptives contain-
ing levonorgestrel had 3.5 times the risk of venous
thromboembolism compared to non-users, users of
oral contraceptives containing desogestrel or ges-
todene had 9.1 times the risk of venous thrombo-
embolism compared with non-users (153). However,
it was not possible to confirm with certainty whether
these unexpected findings were due to chance, bias
or confounding. Another study estimated that the
excess risk for non-fatal venous thromboembolism

associated with combined oral contraceptives con-
taining desogestrel or gestodene compared to levo-
norgestrel was 16 per 100,000 woman-years (56). It
was also reported that the absolute risk of venous
thromboembolism associated with these oral con-
traceptives was particularly high among carriers of
the factor V Leiden mutation (which is the com-
monest cause of inherited thrombophilia and results
in an increased risk of venous thrombosis), and also
among women with a family history of thrombosis
(7).
Publication of these studies created a furore. There
was extensive criticism of these non-randomized
epidemiological studies from some quarters, as might
be expected given the multi-billion dollar nature of
the worldwide oral contraceptives market and the
considerable investments made by pharmaceutical
companies to develop the third-generation drugs.
However, as discussed in a commentary in the same
issue of The Lancet that carried four of the reports,
such large-scale epidemiological studies are the pri-
mary source of evidence in this situation because the
differences between oral contraceptives in terms of
their influence on coagulation variables are typically
small (and may or may not be clinically relevant), and
it is not possible to conduct randomized controlled
trials that are large enough to detect differences be-
tween drugs in terms of adverse events that occur in
only a tiny proportion of the oral contraceptive users
(148). The studies published in The Lancet were large,
and each independently arrived at the same conclu-
sion: that there was an approximately twofold in-
creased risk for venous thromboembolism in users of
third-generation oral contraceptives containing des-
ogestrel or gestodene compared to other oral con-
traceptives. The increased risk was similar for both
desogestrel and gestodene, and was significant after
adjustment for known risk factors for venous
thromboembolism such as age, weight, smoking,
parity (the number of times a woman has given birth)
and varicose veins (148).
It should be noted that the issue of The Lancet that
carried these articles was published on 16 December
1995. However, approximately 2 months earlier, on
18 October 1995, the UK Committee on Safety of
Medicines had already announced that the third-
generation oral contraceptives containing desogestrel
or gestodene carried an increased risk of venous
thromboembolism and should no longer be routinely
prescribed. Doctors in the UK were told that women
using such products should be advised of the risk and
given the opportunity to switch to another brand.
This decision was based on 3 of the studies that at the
Oral contraceptives and the periodontium
time were unpublished and unavailable, but that
would be published in the Lancet two months later.
In addition, confusion arose because many sub-
sequent commentaries on this issue did not specify
that the original statement was limited to combined
oral contraceptives containing desogestrel or ges-
todene and did not refer to all oral contraceptives,
despite the fact that the UK Committee on Safety of
Medicines had stated that no change in prescribing
practice is required for any other combined oral
contraceptive. The Committee on Safety of Medi-
cines statement indicated that oral contraceptives
containing desogestrel or gestodene should no longer
be routinely prescribed, and should be limited to
women who were intolerant of other combined oral
contraceptive preparations and who were prepared
to accept an increased risk of thromboembolism.
A major criticism of this whole episode is that
information was available to the general public via
the media more rapidly than scientific information
was available to physicians. The British public re-
acted quickly to the headlines, and many thousands
of women abruptly discontinued the use of oral
contraceptives (37, 145). From mid-October 1995
until June 1996, there were approximately 30,000
extra conceptions in the UK, a figure estimated from
the likely number of conceptions that would have
been expected to occur over the same time period
based on existing trends had the announcement from
the Committee on Safety of Medicines not been
made. It was also reported that 10,000 of these extra
conceptions were terminated (145, 151). Anxious
patients contacted helplines and their doctors clinics
in huge numbers. An editorial in The Lancet lamented
the course of events, and made a plea that clinical
alert systems must be improved to prevent alarming
patients, and to ensure that the medical community
and the public receive appropriate and full informa-
tion at the same time (27).
The controversy didnt fade away. Of course, as
stated in an editorial in the British Medical Journal, a
statistical association in an epidemiological study
does not prove causality (91). The epidemiological
studies in question attempted to control for con-
founding factors such as age, smoking, alcohol in-
take, body mass index and duration of use of oral
contraceptives. However, controlling for bias was
more difficult. For example, newer (i.e. third-gener-
ation) oral contraceptives may have been prescribed
more frequently to new users and to patients who
were considered to be at higher risk because their
doctors might have perceived the new pills to be
safer. It was also postulated that the increased risk
129
Preshaw
associated with these drugs could have resulted from
their use by younger women, who, as a group, may be
expected to include a greater proportion of individ-
uals who were undetected carriers of the factor V
Leiden mutation, which is known to increase the risk
of thromboembolism (7). Older women who were
carriers of the Leiden mutation would probably have
already been identified, and would be extremely un-
likely to have been prescribed any oral contraceptives
(147). The third-generation agents, being newer, were
more commonly prescribed to new (i.e. younger)
users, and the increased association with venous
thromboembolism may have arisen not because of
any direct effect of the drugs themselves, but because
of differential prescribing to a (younger) group in
which carriers of the factor V Leiden mutation were
as yet undetected.
In the UK, questions were subsequently raised as to
whether the advice from the Committee on Safety of
Medicines was arrived at lawfully, because of ques-
tions about process (i.e. using the data before it was
published) and whether the advice was dispropor-
tionate and beyond the authority of the committee
(146). Furthermore, one of the authors of the reports
complained that the committee had mis-used his
data and had acted too hastily (128). He reported
that, prior to publication, he and his colleagues had
shared the data in confidence with the committee,
which then, without further consultation with the
researchers, made the decision to send out the Dear
Doctor letter to 190,000 doctors and pharmacists in
the UK advising them that third-generation agents
containing gestodene or desogestrel should no longer
be prescribed as a matter of routine. The author re-
ported that he was unconvinced whether the findings
in his research resulted from bias or were causally
significant, and that use of the term twice the risk of
venous thromboembolism in the committees letter
to doctors was misleading and alarmist (128).
Nonetheless, other countries followed suit. In
Germany, the Federal Institute for Medical Drugs
and Products imposed limits on the use of pills
containing desogestrel or gestodene, and stated that
they should no longer be prescribed for women un-
der the age of 30 who were using contraceptives for
the first time (12). The Norwegian Medicines Control
Agency also decided to advise women against taking
these agents, and recommended that women should
discuss the issues with their doctor. However, no
advice was provided to the medical community on
what they should tell the patients. In Canada, a more
cautious approach was followed, and the Federal
Drugs Directorate promised to review any studies
130
once they were published and then make appropri-
ate recommendations. In the USA, the controversy
failed to attract much media attention and the Food
and Drug Administration concluded, after reviewing
the three unpublished studies, that the risk is not
great enough to justify switching to other products
(12).
The paradox is that the decision by the Committee
on Safety of Medicines in the UK was based solely on
the risk for venous thrombosis, whereas the rationale
for marketing desogestrel and gestodene was that
they were expected to be less likely to increase
the risk of arterial disease, particularly myocardial
infarction and stroke, which, although rare in young
women, are more likely to be fatal if they do occur
than venous thromboembolism is (91). In an editorial
which appeared in the British Medical Journal after
the Dear Doctor letter from the Committee on Safety
of Medicines but before the publications concerned
appeared in The Lancet, doctors were reassured
about the safety of these agents but were advised that
women should be informed that, if they use a low-
dose estrogen plus desogestrel or gestodene contra-
ceptive, they may increase their risk of venous
thromboembolism by up to 15 per 100,000, but that
they may also be less likely to suffer a myocardial
infarction or stroke than if they were taking an oral
contraceptive brand containing levonorgestrel or
norethisterone (40).
Subsequent reviews further dissected the possible
reasons for the increased risk of venous thrombo-
embolism in the users of third-generation oral con-
traceptives. It is generally accepted now that bias was
not introduced by selective prescribing, for example
by inadvertently selectively prescribing these drugs to
high-risk women, and it is extremely unlikely that
factors strong enough to affect prescribing choices
would not have been recognized and adjusted for
(145). Furthermore, there is no evidence that there
was preferential diagnosis of venous thromboembo-
lism in users of third-generation pills, or more
intensive investigation of women using these drugs,
which in any case would be at odds with the initial
marketing of these products, which suggested they
were associated with fewer adverse events than sec-
ond-generation oral contraceptives. There was also
no evidence of any tendency for high-risk women to
shift from older products to newer ones that could
explain the increased risk, or that the increased risk
with the third-generation products was associated
with a shorter average duration of use.
To summarize, the scientific evidence to date
indicates that there is a small but increased risk for

venous thromboembolism in women who used
combined oral contraceptives containing low-dose
estrogen and desogestrel or gestodene compared to
users of second-generation contraceptives. Sugges-
tions of bias as a result of confounding and pre-
scribing variations have been made, but such bias
would probably be inadequate to sufficiently account
for the observed increased risk, and these suggestions
do not refute the contention that the epidemiological
studies uncovered a direct causal effect (145).
Four years after the original publications appeared
in The Lancet, the UKs Committee on Safety of
Medicines and the Medicines Control Agency chan-
ged their stance and issued new advice in which
doctors were advised they could again prescribe
third-generation pills as a first-line contraceptive
method (18). The new guidance from the Committee
on Safety of Medicines was that provided women are
fully informed of the very small increased risks and
do not have any other medical contraindications
(such as a history of venous thromboembolism), it
should be a matter of clinical judgement and per-
sonal choice which types of oral contraceptives [are]
prescribed. For some individuals, the choice of oral
contraceptive used may be made for reasons not re-
lated to the contraceptive benefits of the drug. For
example, the potential benefits of the third-genera-
tion pills, such as reduced acne, may outweigh the
small additional risk of thrombosis. However, on a
population basis, it has been stated that, although
third-generation contraceptives cannot be regarded
as unsafe, it should be remembered that, overall,
second-generation contraceptives are (slightly) safer
(108). The decision about which contraceptive to use
must be based on personal choice and communica-
tion between the doctor and patient. It is generally
now accepted that a second-generation product (e.g.
containing levonorgestrel or norethindrone) should
remain the first choice for first-time users, because
first-time users will include an unknown subgroup
who are more predisposed to venous thrombo-
embolism (41).
The Cochrane Collaboration reviewed the literature
regarding combined oral contraceptives with respect
to effectiveness, safety, cycle control, side-effects and
use continuation rates in 2004 (92). They stated that,
since the introduction of oral contraceptives, the
estrogen dose has gradually been reduced to 30 lg or
less to minimize unwanted effects, but it is not pos-
sible to decrease the progestin dose further as this
could allow an oocyte to be released from the ovary,
possibly resulting in pregnancy. The effectiveness of
oral contraceptives in preventing pregnancy is not
Oral contraceptives and the periodontium
only dependent on the dose of hormones but is also
affected by the acceptability to the users, and
breakthrough bleeding, spotting and cycle control all
affect acceptability. Twenty-two trials were consid-
ered by the Cochrane Collaboration, of which 18 were
sponsored by the pharmaceutical industry and only
five reported attempts at blinding. The review found
that discontinuation of use was significantly lower
with second- vs. first-generation products (relative
risk [RR] of discontinuation of use 0.79; 95% confi-
dence intervals [CI] 0.690.91) indicating better
acceptability with the second-generation agents. Cy-
cle control was also better with second-generation
than first-generation drugs (RR 0.61; 95% CI 0.52
0.91). Contraceptive effectiveness, spotting, break-
through bleeding and absence of withdrawal bleeding
were similar for third-generation gestodene com-
pared to second-generation levonorgestrel, but there
was less inter-menstrual bleeding in the gestodene
group (RR 0.71; 95% CI 0.550.91). The acceptability
indices all confirmed that third- and second-
generation products were preferred compared to
first-generation contraceptives. Most pregnancies
occurred when pills that contained 20 lg ethinylest-
radiol were used. Of the third-generation products,
those containing gestodene resulted in better cycle
control than those containing desogestrel, although
the continuation rate was higher in the desogestrel-
containing pills.
Two independently conducted meta-analyses
published in 2001 that considered the risk of venous
thrombosis when using third-generation oral con-
traceptives both reported a 1.7-fold increased risk
compared to second-generation compounds (49, 63).
These studies concluded that third-generation oral
contraceptives are associated with an increased risk
of venous thrombosis compared with second-gener-
ation oral contraceptives, and the increase could not
be explained by bias or confounders. The increased
risk translated to an incremental risk of venous
thromboembolism of approximately 10 or 11 per
100,000 women per year (49), and the risk appears to
be highest in first-time users.
The last word has yet to be written on this subject.
In February 2007, the advocacy group Public Citizen
filed a petition with the US Food and Drug Admin-
istration, demanding that sale of third-generation
oral contraceptives containing desogestrel be banned
in the USA (115). Their claim is that, by switching
from third- to second-generation contraceptives,
approximately four deaths per 1,000,000 women-years
could be prevented, and that the third-generation
drugs do not provide any proven superior clinical
131
Preshaw
benefit compared to other classes of oral contracep-
tives with the same efficacy profile. At the time of
writing this review (August 2010), the Food and Drug
Administration has yet to respond, other than stating
that they will carefully review the petition.
Benefits and unwanted effects of
oral contraceptives
Currently available oral contraceptives are extremely
effective when used correctly, with an estimated
0.3% of women experiencing an unintended preg-
nancy during the first year of use if used perfectly
(although the figure rises to approximately 8% when
considering typical use) (41). Combined oral contra-
ceptives primarily prevent ovulation, making the
method a highly effective form of birth control.
Typical hormone doses found in currently available
combined oral contraceptives are presented in
Table 1. Details of the composition of oral contra-
ceptives used worldwide, listed by country and drug
name, are available in the directory of hormonal oral
contraceptives at www.ippf.org.
The contraceptive benefits of combined oral con-
traceptives are readily apparent, namely their effec-
tiveness, convenience, reversibility and simplicity of
use, including the fact that they do not need to be
used directly at the time of sexual intercourse. Some
of the non-contraceptive benefits of oral contracep-
tives are listed in Table 2, and these may sometimes
be the primary indication for using these drugs. For
example, the pill can reduce irregular bleeding and
painful periods, and is often prescribed for control of
dysmenorrhea. Indeed, combined oral contraceptives
are now one of the first-line medical treatments for
Table 1. Formulations of a selection of marketed combined oral contraceptives, adapted from (41)
Pill type Marketed name
Ethinylestradiol norethisterone Loestrin 20
Loestrin 30
Norimin
Ethinylestradiol levonorgestrel Microgynon 30
Ethinylestradiol desogestrel Mercilon
Ethinylestradiol norgestimate Cilest
Ethinylestradiol drospirenone Yasmin
Mestranol norethisterone Norinyl-1
*Converted to norethisterone as the active metabolite.
Norgestimate is partially metabolized to levonorgestrel.
132
women who suffer from heavy menstrual blood loss.
There are also reports of reduced headaches, tired-
ness, bloating and menstrual pain among users, and
most combined oral contraceptives improve acne by
raising the levels of sex hormone-binding globulins,
thereby mopping up free unbound testosterone (93).
Table 2. Non-contraceptive benefits of combined oral
contraceptives, adapted from Guillebard (41)
Improvement in most menstrual cycle disorders: de-
creased bleeding, reduced dysmenorrhea, more regular
bleeding (the timing of which can be controlled), no
ovulation pain
Reduced risk of ovarian and endometrial cancer, and
possibly also colorectal cancer
Fewer functional ovarian cysts (because abnormal ovu-
lation is prevented)
Fewer extra-uterine pregnancies (because normal ovu-
lation is inhibited)
Reduction in pelvic inflammatory disease
Reduction in benign breast disease
Probable reduction in thyroid disease (whether over- or
under-active)
Probable reduction in the risk of rheumatoid arthritis
Fewer sebaceous disorders, especially acne
Possible reduction in risk of endometriosis
Possibly fewer duodenal ulcers (not well established)
Reduction in Trichomonas vaginalis infections
Possible lower incidence of toxic shock syndrome
No toxicity in overdose
The list is based on data from the prospective Royal College of General
Practitioners, Oxford Family Planning Association and US Nurse Studies,
supplemented by casecontrol studies and some randomized controlled trials
performed by the World Health Organization and other bodies.
Estrogen content (lg) Progestin content (lg)
20 1,000 (norethisterone acetate*)
30 1,500 (norethisterone acetate*)
35 1,000 (norethisterone)
30 150
20 150
35 250
30 3,000
50 1,000

A major non-contraceptive benefit is that carcino-
mas of the ovary and endometrium are less frequent
in users of oral contraceptives, with the incidence of
both being approximately halved in all oral contra-
ceptive users and reduced to one-third in long-term
users (41). The most likely explanation for this ben-
eficial effect is the suppression of ovulation and
normal mitotic activity in the endometrium that re-
sults from using oral contraceptives. In other words,
suppression of ovulation in users reduces the cellular
damage in the ovarian epithelium that would other-
wise occur as a result of repeated ovulation. The risks
of ovarian cancer decrease with increased numbers
of live births, increased numbers of incomplete
pregnancies, and length of time using oral contra-
ceptives, highlighting the importance of periods
without regular ovulation in reducing the risk for this
disease (13). It has also been suggested that reduced
menstrual blood loss in those taking oral contracep-
tives may prevent retrograde transportation of car-
cinogens via the Fallopian tubes, and also that the pill
may have a direct effect on the ovarian epithelium,
influencing apoptosis of cells in precancerous lesions
(46).
The combined oral contraceptive may now be
indicated primarily for its protective effect against
ovarian cancer in women known to be at increased
risk. It has been estimated that use of oral contra-
ceptives worldwide has already prevented approxi-
mately 200,000 cases of ovarian cancer and 100,000
deaths from the disease, with predictions of preven-
tion of 30,000 cases of ovarian cancer per year in the
future as a result of the increasing age of past users
and increasing numbers of new users (16).
A small protective effect of the combined oral
contraceptive against colorectal cancer has also been
suggested, with an 18% reduction in risk among
individuals who had ever used oral contraceptives
(32). Possible mechanisms are unclear, but may be
related to altered bile synthesis and secretion, leading
to a reduced concentration of bile acids in the colon
(98). Also, estrogens have been shown to inhibit the
growth of colon cancer cells in vitro (87) and estrogen
receptors have been identified in normal and neo-
plastic colon epithelial cells (136). However, this
protective effect of oral contraceptives against colo-
rectal cancer has yet to be fully proven (41), and the
authors of the meta-analysis that identified this effect
stated that some aspects remain undefined and the
issue of causality for the observed association is still
open to discussion (32).
There are also some broader sociological and
environmental benefits of oral contraceptives in
Oral contraceptives and the periodontium
addition to those to the individual user. Clearly, the
principal benefit of oral contraceptives is that of
simple and effective contraception, which, at the
population level, has benefits in controlling popula-
tion growth and permitting women to make choices
about education and careers as well as family plan-
ning. As stated by James Grant (Executive Director of
UNICEF 19801995) in 1992, family planning could
bring more benefits to more people at less cost than
any other single technology now available to the
human race (41). However, the introduction of con-
traceptives alone does not seem to directly influence
birth rates in given populations. For example, falling
birth rates in developing countries in the 1970s were
not simply due to the adoption of modern birth
control practices, which had, after all, been available
from the 1950s and 1960s. Instead, it was the use of
contraception combined with a shift to later ages of
marriage, and cultural changes in which families with
lots of children were no longer required for reasons of
wealth and social status, that made the difference.
Thus, contraceptives can be most confidently pre-
dicted to have sociological benefits in populations in
which birth rates are already falling (97).
In addition to the obvious benefits of these drugs,
there are many adverse and unwanted effects that
have been associated with using oral contraceptives,
particularly links with various cancers and increased
blood clotting. Given that oral contraceptives are ta-
ken by millions of young women who are not ill, the
safety expectations are high, and risks that may be
considered small if drugs were used for management
of a medical condition tend to be greatly magnified. It
is also important to remember that oral contracep-
tives do not protect against sexually transmitted
diseases. Some of the unwanted effects of oral con-
traceptives are discussed below.
Breast cancer
The Collaborative Group on Hormonal Factors in
Breast Cancer published their findings in 1996, in
which they re-analysed data relating to 53,297 wo-
men with breast cancer and 100,239 women without
breast cancer from 54 studies in 25 countries (15).
These 54 studies represented approximately 90% of
the worldwide epidemiological evidence on breast
cancer risk and the use of hormonal contraceptives
available at that time. They reported that for women
currently taking combined oral contraceptives and
within 10 years after discontinuing use, there is a
small increase in the relative risk of diagnosis of
breast cancer. In current users, the relative risk was
133
Preshaw
1.24 (95% CI 1.151.33), at 14 years after discon-
tinuing use it was 1.16 (95% CI 1.081.23), and at 5
9 years after discontinuing use it was 1.07 (95% CI
1.021.13). There was no increased risk of having
breast cancer diagnosed 10 years or more after dis-
continuing use (relative risk 1.01, 95% CI 0.961.05).
Furthermore, the cancers diagnosed in women who
had used combined oral contraceptives were less
advanced clinically than those who had never used
these contraceptives. This study indicated that cur-
rent users have a 24% increased risk of developing
breast cancer while taking the pill, and this decreases
to zero over the 10 years after discontinuing use.
There were no relationships between risks for breast
cancer and duration of use, dose or the types of
hormone present in the combined oral contra-
ceptives.
However, two studies published after the above re-
analysis reported that there is no increased risk of
breast cancer associated with low-dose combined
oral contraceptives. One of these studies was a pop-
ulation-based casecontrol study of over 9,000 wo-
men, and concluded that, among women 3564 years
of age, current or former oral contraceptive use was
not associated with a significantly increased risk of
breast cancer (94). The other report was based on
data from the Nurses Health Study cohort in the
USA, and again found that long-term previous oral
contraceptive use, either overall or prior to a first full-
term pregnancy, does not result in any appreciable
increase in breast cancer risk in women over 40 years
old (45). Taken collectively, these more recent studies
suggest that the risk of breast cancer in women taking
modern formulations of oral contraceptives is very
slight, if it exists at all.
Other neoplasms
Cervical cancer has been associated with the use of
oral contraceptives. Human papilloma virus types 16
and 18 are implicated as the principal carcinogen with
a sexual transmission route. Some studies that have
investigated links between oral contraceptives and
cervical cancer attempted to identify and control for
the presence of human papilloma virus, leading to the
conclusion that combined oral contraceptives act as a
co-factor, speeding transition through the stages of
cervical intra-epithelial neoplasia (41). The odds ratio
for cervical cancer in oral contraceptive users in-
creases with increasing duration of use, and may
persist in ex-users.
Combined oral contraceptives also increase the
relative risk of liver tumors, including benign ade-
134
noma or hamartoma. The background incidence of
these lesions is very low (13 per million women per
year), and any additional risk due to the pill is very
slight (41). Most case reports have been in long-term
users of relatively high-dose contraceptives.
Thromboembolism
The risks of venous thromboembolism have been
discussed extensively above. To recap, the spontane-
ous incidence of venous thromboembolism in heal-
thy, non-pregnant women who are not taking any oral
contraceptives is approximately five cases per 100,000
women per year. The incidence when using second-
generation oral contraceptives is approximately 15
per 100,000 women per year, and the incidence with
third-generation contraceptives (i.e. containing des-
ogestrel or gestodene) is approximately 25 cases per
100,000 women per year (18, 41). The risk of deep vein
thrombosis in pregnant women is approximately 60
cases per 100,000 women per year (43).
The increased incidence of venous thromboem-
bolism with the third-generation contraceptives has
not been explained by bias or confounding, and the
level of risk increases with age and the presence of
other known risk factors such as obesity. Mortality
from venous thromboembolism has been estimated
to be approximately 12%, and therefore a risk of
thromboembolism of 25 cases per 100,000 for the
third-generation contraceptives translates to mortal-
ity rates of 2.55 individuals per million users (40).
This figure is low compared to mortality in pregnancy
and other risks of daily life. For example, the time
required for a 1 in a million risk of dying has been
estimated to be 1 min when riding a motor bike, 1 h
when driving a car, or 1 month when taking the oral
contraceptive pill (if a non-smoker) (41). The annual
risks of death per 1,000,000 women in various cir-
cumstances are indicated in Table 3.
The biological rationale for the association
between oral contraceptives and increased risk of
venous thromboembolism is beginning to be more
completely understood. Use of oral contraceptives
causes disturbances in the complex balance between
pro-coagulant, anti-coagulant and fibrinolytic path-
ways. In a research study to investigate this issue, 28
women who were not using oral contraceptives were
given either a second-generation oral contraceptive
(30 lg ethinylestradiol + 150 lg levonorgestrel) or a
third-generation oral contraceptive (30 lg ethiny-
lestradiol + 150 lg desogestrel) in a randomized
cross-over design with a 2-month wash-out period
between the drugs (135). Parameters investigated

Table 3. Annual risks of death per 1,000,000 women,
adapted from Guillebard (41)
Numbers
of deaths
Childbirth in the UK (all direct causes of 60
death)
Venous thromboembolism associated 20
with childbirth in the UK
Venous thromboembolism from using a 5 extremely rare, unless users also smoke or have dia-
third-generation pill
Venous thromboembolism from using a 3 creased risk of acute myocardial infarction in users
second-generation pill
Venous thromboembolism in a 12
non-pregnant, non-pill user
Risk of death from all causes 10
(not just thromboembolism) if using
the pill (in a non-smoking woman)
Home accidents 30
Road accidents 80
Scuba diving 220
Hang-gliding 1,500
Cigarette smoking (in the next year, if 1,670
aged 35)
Pregnancy childbirth in rural Africa >10,000
included levels of anti-thrombin, a2 macroglobulin,
protein C (a vitamin K-dependent physiological
anti-coagulant), protein C inhibitor, and protein S
(which functions as a co-factor for protein C). The
researchers found that, when using the third-gener-
ation pill, the levels of a2 macroglobulin, protein C
and protein C inhibitor all increased significantly,
and the levels of anti-thrombin and protein S de-
creased significantly. When using the second-gener-
ation pills, similar trends were identified, but the
increases in anti-thrombin and protein C inhibitor
did not reach significance, and protein S levels
increased slightly but not significantly. The signifi-
cant decrease in total and free protein S levels with
third-generation pills indicates that the activity of
anti-coagulant pathways in users of desogestrel-
containing oral contraceptives is more extensively
impaired than in users of second-generation levo-
norgestrel-containing oral contraceptives (135). Put
more simply, levonorgestrel reduced the pro-coagu-
lant effects of ethinylestradiol and decreased the
reductions in protein S levels that were seen in the
users of third-generation products, and therefore
reduced the risk of venous thromboembolism com-
pared to a dose of ethinylestradiol alone. Desogestrel
Oral contraceptives and the periodontium
did not have as much of an opposing action against
the pro-coagulant effects of estrogen, which may
explain the increased risk of venous thromboembo-
lism in the users of these drugs.
Arterial disease
Arterial disease among users of combined oral con-
traceptives with an ethinylestradiol dose <50 lg is
betes or hypertension. For example, there is no in-
who do not smoke compared to non-users of oral
contraceptives (14), but the risk in smokers who use
the oral contraceptive is ten times greater, or more
(41). There is no additional risk of either ischemic or
hemorrhagic stroke for users of low-dose oral con-
traceptives, including those who are 35 years old,
smokers, or who are obese compared to non-users.
However, there is an approximately twofold increase
in the risk of ischemic and haemorrhagic stroke in
users of oral contraceptives who also report a history
of migraine (120, 137).
These risks associated with the use of oral contra-
ceptives require that, before prescribing these drugs,
a careful personal and family history with particular
reference to cardiovascular risk markers should be
taken, and blood pressure should be measured
accurately (47). For the majority of women, no fur-
ther investigations are normally required. Absolute
contraindications to using combined oral contra-
ceptives include any history of thrombosis (venous or
arterial), ischemic heart disease, angina, known pro-
thrombotic states, and estrogen-dependent neo-
plasms (such as breast cancer) (41).
Mechanisms of action of oral
contraceptives
Modern combination oral contraceptives typically
contain a low dose of estrogen (35 lg) and a pro-
gestin, the concentration of which is typically in the
range 0.11.5 mg. The dose of these hormones may
be fixed during the period of taking the medication
(monophasic), or may vary (multiphasic). The mul-
tiphasic preparations increase the dose of the pro-
gestin or estrogen in two or three steps to mimic
more closely the natural menstrual cycle, and also to
minimize unwanted effects such as breakthrough
bleeding. Both monophasic and multiphasic types of
oral contraceptives are taken for 21 days, followed by
7 days free of medication. In addition to the com-
135
Preshaw
bined oral contraceptives, there is also a progestin-
only oral contraceptive (sometimes referred to as the
mini-pill), which contains a small dose of progestin,
norethindrone or norgestrel. This pill must be taken
every day, at the same time of day, with no medica-
tion-free period.
The release of luteinizing hormone by the pituitary
gland is controlled by pulses of gonadotropin-
releasing hormone from the hypothalamus, and these
pulses are subject to estrogen feedback from the go-
nads. The menstrual cycle in humans is divided into a
follicular phase and a luteal phase. During the fol-
licular phase (which occurs prior to the peak in
luteinizing hormone), serum estradiol levels increase
above a threshold of 150200 pg ml for approxi-
mately 36 h. This has a brief positive feedback effect
on the pituitary, triggering the pre-ovulatory surge of
luteinizing hormone and follicle-stimulating hor-
mone. The surges of these gonadotropins stimulate
progesterone secretion and follicle rupture and ovu-
lation within the next 12 days. The luteal phase
commences after the luteinising hormone peak and
ovulation, and the ruptured follicle develops into the
corpus luteum, which, under the influence of
luteinizing hormone, produces large amounts of
progesterone and lower amounts of estrogen. To-
wards the end of the luteal phase, if fertilization has
not occurred, the plasma levels of progesterone and
estradiol decrease as the corpus luteum ceases to
function, and menstruation occurs as the endome-
trium is shed.
During the follicular (or proliferative) phase,
estrogen begins the process of rebuilding the endo-
metrium by stimulating proliferation and differenti-
ation. In the luteal (or secretory) phase, elevated
progesterone levels limit the proliferative effects of
estrogen on the endometrium by stimulating differ-
entiation of the cellular lining in preparation for
implantation, including stimulation of epithelial
secretions that are important for implantation of the
blastocyst. The window of opportunity for implan-
tation is relatively narrow, spanning days 1924 of the
endometrial cycle.
Combined oral contraceptives work primarily by
inhibiting ovulation, by action on the hypothalamic
pituitaryovarian axis to reduce the levels of lutein-
izing hormone and follicle-stimulating hormone
(65). Both the progestin and estrogen in oral con-
traceptives suppress luteinizing hormone secretion,
thereby preventing ovulation. A major action of
progestins is to decrease the frequency of gonado-
tropin-releasing hormone pulses from the hypo-
thalamus. Further, progestins in oral contraceptives
136
function partly by resulting in production of an
endometrial surface that is not receptive to
implantation. Progestins thicken cervical mucus so
that sperm cannot penetrate the uterus, and produce
an endometrium that is unreceptive to ovum
implantation even if an egg is fertilized. Estrogens
and progestins also have important effects on the
Fallopian tubes, such that estrogens increase and
progestins decrease tubal contractility, affecting the
transit time of the ovum to the uterus. These chan-
ges to the secretory and peristaltic activity within the
Fallopian tubes also contribute to contraceptive
efficacy (122). Follicle-stimulating hormone is also
suppressed by estrogens and progestins, thereby
preventing the development and emergence of a
dominant follicle. By preventing ovulation and sta-
bilizing the endometrium, irregular shedding is
prevented and breakthrough bleeding is reduced in
users of oral contraceptives.
The estrogen component of the combined oral
contraceptives also potentiates the contraceptive
efficacy of the progestin component (probably by
increasing the level of intracellular progestin recep-
tors), and therefore only a minimal dose of estrogen
is required to maintain the effectiveness of these
drugs (122). During the usual pill-free 7 days, the
endometrium sheds and withdrawal bleeding occurs.
Contraceptive protection is maintained during the
pill-free interval as long as the pills are taken cor-
rectly and consistently before and after this interval
(6567).
The progestin-only oral contraceptive relies on the
effects of progestin on cervical mucus, the endome-
trium and possibly also the Fallopian tube for its
efficacy. This pill must be taken at the same time
every day to maintain the correct progestin level,
which is approximately 75% lower than the level
resulting from combined oral contraceptives. The
impermeability of the cervix resulting from use of the
progestin-only oral contraceptive diminishes within
2224 h of ingesting the pill. Use of this pill may also
result in more irregular menstrual bleeding, and has
also been associated with acne. The acne results from
the androgenic activity of the progestin (despite the
very low concentration), which decreases the level of
circulating sex hormone-binding globulin. As a result,
there is increased availability of progestin and tes-
tosterone in the circulation, leading to increased acne
(122). Combined oral contraceptives are not associ-
ated with acne (and may even reduce it) because
estrogen increases the level of sex hormone-binding
globulin, and therefore counteracts the effects of
progestin. The progestin-only mini-pill does not
 Oral contraceptives and the periodontium

significantly alter carbohydrate levels, lipid metabo-
lism or blood coagulation.
Physiology and pharmacological
actions
Estrogens and progestins have numerous physiolog-
ical actions. In women, these include developmental
effects (pubertal development, skeletal growth and
secondary sexual development), control of ovulation,
the cyclical preparation of the reproductive tract for
fertilization and implantation of a fertilized egg,
neuroendocrine effects, and effects on the metabo-
lism of lipids, carbohydrates, proteins and minerals.
Estrogens also have effects in males, with deficiency
being associated with diminished pubertal develop-
ment and delayed skeletal maturation (88). The
therapeutic use of these hormones in oral contra-
ceptives reflects the physiological actions of these
compounds. Antagonists of these hormones are also
used therapeutically. For example, estrogen receptor
antagonists are used in the treatment of some forms
of breast cancer and infertility. Anti-progestins are
primarily used for medical abortion (88). The wide-
ranging and profound effects of sex steroid hormones
are remarkable given that they exist at extremely low
concentrations (in the femtomolar to nanomolar
range) in the circulation, and that their distinct
biological effects depend on very slight structural
differences between relatively small molecules
(molecular mass approximately 300 Da) (95).
is catalysed by a cytochrome P450 monooxygenase
enzyme complex (aromatase or CYP19) using the
reduced form of nicotinamide adenine dinucleotide
phosphate and molecular oxygen as substrates. The
ovaries are the main source of circulating estrogen in
premenopausal women, and estradiol is the primary
secretory product. In men, estrogens are produced by
the testes and also by aromatization of circulating
steroids such as androstenedione. All three naturally
occurring estrogens are excreted in the urine. The
structures of estradiol, ethinylestradiol and mestranol
are shown in Fig. 1.
The phenolic A ring of these molecules is respon-
sible for high-affinity binding to estrogen receptors.
There are two identified estrogen receptors: estrogen
receptor a (discovered in the 1960s) and estrogen
receptor b (identified in 1996). These are both estro-
gen-dependent nuclear transcription factors that
have wide tissue distributions and exert regulatory
effects on a large number of target genes (48).
Estrogen receptor a is expressed most abundantly in
the female reproductive tract, particularly the ovaries,
uterus and vagina, as well as in the breast, hypo-
thalamus, endothelial cells and smooth muscle cells.
Estrogen receptor b is expressed mainly in the pros-
tate and ovaries, with lower expression in bone,
lungs, brain and the vasculature. Many cells express
Estradiol OH
CH3 H
C18H24O2 17
13 D
1 C
A B
3

Estrogens HO
6

Many steroidal and non-steroidal compounds pos- OH

Ethinylestradiol CH3 C CH
sess estrogenic activity. Of the three main naturally 17
occurring estrogens in humans (estrone, estradiol C20H24O2 13 D
1 C
and estriol), estradiol-17b is the most potent, and is
the major secretory product of the ovary. Structural A B
3
alteration of natural estrogens makes them orally HO
6
active (i.e. pharmacologically active if taken orally).
Ethinylestradiol is one of the most potent oral Mestranol OH
estrogens, and has an ethinyl group substitution at CH3 C CH
C21H26O2 17
C17. Addition of the ethinyl group inhibits hepatic 13 D
1 C
first-pass metabolism, which is why ethinylestradiol
has far greater oral potency than estradiol. The most A B
3
commonly used estrogens in oral contraceptives are
CH3O 6
ethinylestradiol and mestranol, which have slight
Fig. 1. Structural formulae of selected estrogens. Estra-
structural variations compared to estradiol in order to
diol is the major naturally occurring estrogen. Ethiny-
increase their oral potency. lestradiol is the most commonly used estrogen in oral
Estrogens arise from androstenedione or testos- contraceptives, and mestranol is used in some oral con-
terone by aromatization of the A ring, a reaction that traceptives.

137
Preshaw
both receptors, which form homo- or hetero-dimers.
When the receptors are expressed together, estrogen
receptor b tends to inhibit estrogen receptor a-med-
iated transcriptional activity.
Both receptors are ligand-activated transcription
factors that increase or decrease the transcription of
target genes. The hormone enters the cell by passive
diffusion across the plasma membrane, and binds to
the estrogen receptor in the nucleus. The receptor is
present in the nucleus as an inactive monomer
bound to heat shock proteins, but, when estrogen
binds, the heat shock proteins disassociate and the
receptor dimerizes, increasing the affinity and rate of
binding of the receptor to DNA. The receptor dimer
binds to DNA resulting in a cascade of co-activator
proteins to the promoter regions of target genes
resulting in initiation of transcriptional events. A
better understanding of the interactions between
ligand (i.e. estrogen) and target receptors has com-
pletely changed perceptions of estrogen pharmacol-
ogy. Previously, it was considered that estrogen
molecules bound to a single estrogen receptor, and
that this complex subsequently affected transcription
by the same molecular mechanisms in all tissues.
However, the altered conformation of the two estro-
gen receptors upon binding to ligand results in
varying interactions with co-activators and co-re-
pressors in cell- and promoter-specific ways, thus
greatly amplifying the spectrum of effects of receptor
binding. Thus ligands may have anti-estrogenic ef-
fects in some tissues, partially estrogenic and par-
tially anti-estrogenic effects in other tissues, and
purely estrogenic effects in others (88).
Progestins
The progestins include the naturally occurring hor-
mone progesterone, the 17a-acetoxyprogesterone
derivatives in the pregnane series, the 19-nortestos-
terone derivatives (estranes) including norethin-
drone, and also norgestrel and related compounds in
the gonane series (including norgestimate, desoge-
strel and gestodene). The 19-nortestosterones were
developed for use as progestins in oral contracep-
tives, and although they mainly have progestational
activities, they also demonstrate some androgenic
properties. The gonanes are 19-nor compounds that
contain an ethyl rather than a methyl substitution at
the C13 position, and they have reduced androgenic
activities compared to the 19-nortestosterones. These
two classes of 19-nortestosterones derivatives (i.e. the
estranes and the gonanes) are the progestational
components of all oral contraceptives (and also some
138
long-acting injectable contraceptives) (88). The syn-
thesis of the 19-nor compounds in the early 1950s
(specifically norethindrone and the isomer nore-
thynodrel) led to development of effective oral con-
traceptives. Collectively, compounds with biological
activities similar to those of progesterone are referred
to as progestins, but they are also described as pro-
gestational agents, progestogens, progestagens, gest-
ogens or gestagens. Progesterone, the naturally
occurring progestin, is secreted by the ovary (mainly
the corpus luteum during the second half of the
menstrual cycle), and its synthesis and secretion are
stimulated by luteinizing hormone produced in the
pituitary gland.
Like estrogens, progestins also exert their effects
via a receptor, the progesterone receptor. Unlike the
estrogen receptor, whose ligand must possess a
phenolic A ring for high-affinity binding, the pro-
gesterone receptor favors a D4-3-one A ring structure
in an inverted 1b,2a conformation for high-affinity
binding (25). The non-phenolic A ring structure of
the progestins is also bound by other steroid hor-
mone receptors, but optimal binding is achieved with
the progesterone receptor. However, some synthetic
progestins, particularly the 19-nor compounds, do
display some binding activity to glucocorticoid,
androgen and mineralocorticoid receptors, contrib-
uting to the non-progestational activities of these
progestins. The specific properties of progestin mol-
ecules depend greatly on their substitutions, such as
the C17 substituent in the D ring, the methyl group at
C19 and the ethyl group at C13. These slight changes
in molecular structure account for highly significant
differences in biological activity. For example, the 19-
nortestosterone derivatives (the estranes) are deriva-
tives of testosterone but lack the methyl group at C19
and display progestational rather than androgenic
properties. The substitution of an ethyl group at C17
results in formation of 19-nortestosterone analogs
such as norethindrone, norethindrone acetate, nore-
thynodrel and ethynodiol diacetate. These are all
orally active molecules (i.e. active when taken orally),
and the ethyl substitution at C17 reduces hepatic
metabolism. The latter three compounds are rapidly
converted to norethindrone, are less specific than
progesterone itself, and possess varying degrees of
androgenic activities.
Replacement of the methyl group at C13 of nor-
ethindrone with an ethyl substituent results in the
gonane norgestrel. This is a more potent progestin
than the parent compound norethindrone, but has
less androgenic activity. Norgestrel is a racemic
mixture (i.e. a mixture that contains equal quantities

of left- and right-handed enantiomers of a chiral
molecule) of an inactive dextrorotatory isomer and
the active levorotatory isomer, levonorgestrel. Thus, a
preparation that contains only levonorgestrel has the
same pharmacological activity as a preparation con-
taining twice as much norgestrel. The gonanes,
including norgestimate, desogestrel and gestodene,
have been reported to have very little if any andro-
genic activity at therapeutic doses. Newer steroidal
preparations include the spironolactone derivative
drospirenone, which is used in a combination oral
contraceptive (Table 1). Non-steroidal progestins are
under development, as these are likely to have less
affinity for other steroid receptors (88). The structures
of some of the principal progestins relevant to oral
contraceptives are shown in Fig. 2.
Progesterone has many physiological effects,
including decreasing the frequency of pulses of gon-
adotropin-releasing hormone from the hypothalamus,
which is a major mechanism of action of progestin-
containing contraceptives. Progesterone decreases the
endometrial proliferation stimulated by estrogen, and
results in development of a secretory endometrium in
21 CH
Progesterone
C21H30O2 20 C
17
11
19
Norethindrone OH
C20H26O2
O
H 3C
Norgestrel
C21H28O2 H2C OH
O
Fig. 2. Structural formulae of selected progestins. Pro-
gesterone is the major naturally occurring progestin.
Norethindrone (referred to as norethisterone in Europe) is
a derivative of 19-nortestosterone, and is a first-genera-
tion progestin. Norgestrel is a racemic mixture of an
inactive isomer and the active isomer levonorgestrel
(which has the same chemical structure as norgestrel).
Norgestrel and levonorgestrel are found in second-gen-
Oral contraceptives and the periodontium
preparation for implantation of the blastocyst. Pro-
gesterone also modifies the secretions of the endo-
cervical glands so that the watery secretions of the
estrogen-stimulated structures are changed to a much
more viscous composition which decreases the pen-
etration of the cervix by sperm. Progesterone is also
important in maintaining pregnancy by suppressing
menstruation and uterine contractility, among other
effects.
The progesterone receptor exists in two isoforms
(progesterone receptors A and B), which are encoded
by a single gene. The relative proportions of the two
isoforms of the receptor vary in reproductive tissues
according to the tissue type, developmental status
and hormone levels. The ligand-binding domains of
the two isoforms are identical, so there is no differ-
ence in ligand binding based on the isoform that is
present. In its unbound state, the receptor is present
in the nucleus as an inactive monomer bound to heat
shock proteins (HSP90, HSP70 and p59). When
progesterone binds to the receptor, the heat shock
proteins dissociate and the receptor is phosphory-
lated and subsequently forms homo- and hetero-di-
3
Norgestimate
O C23H31NO3 H 3C
OCOCH3
H2C
C CH
HON
C CH
Desogestrel
H 3C
C22H30O
H2C OH
H2C C CH
C CH
eration oral contraceptives. Norgestimate was developed
after the second-generation oral contraceptives and was
sometimes referred to as a third-generation progestin, but
is now considered to be akin to a second-generation
progestin because it is partially metabolized to the sec-
ond-generation progestin levonorgestrel. Desogestrel is
found in third-generation oral contraceptives.
139
Preshaw
mers that bind with high selectivity to progesterone
response elements located on target genes (36).
Transcriptional activity by the progesterone receptors
then occurs via recruitment of co-activators, and the
receptor co-activator complex undergoes further
interactions with additional binding proteins that
possess histone acetylase activity. Histone acetylation
results in remodeling of chromatin, increasing the
accessibility of the target promoter sequence to
transcriptional proteins such as RNA polymerase II
(88).
Although the ligand binding domains of proges-
terone receptors A and B are identical, the biological
activities of the two isoforms are very distinct and
depend on the target gene. For example, the
anti-proliferative effect of progesterone on the
estrogen-stimulated endometrium is lost in proges-
terone receptor A knockout mice (88). Typically,
progesterone receptor B mediates the stimulatory
activities of progesterone whereas progesterone
receptor A inhibits this action of progesterone
receptor B (142) and also inhibits transcriptional
activity resulting from binding of other steroid
receptors. Co-activators and co-repressors interact
differentially with progesterone receptors A and B,
partly accounting for the different activity of the two
isoforms (35).
Pharmacokinetics
Estrogens
Estrogens are essentially lipophilic molecules, and
thus absorption is generally good. Estrogens are
therefore available for oral, parenteral, transdermal
or topical administration. In the context of oral
contraceptives, ethinylestradiol is the most com-
monly used estrogen as the ethinyl group at the C17
position inhibits first-pass hepatic metabolism,
thereby significantly increasing potency. Absorption
occurs across the intestinal epithelium. Ethinylest-
radiol is extensively bound to serum albumin in cir-
culation, but not to sex hormone-binding globulin (in
contrast to naturally occurring estrogens, which are
mainly bound to sex hormone-binding globulin, and
to a much lesser extent to serum albumin). The active
(i.e. unbound) molecules are widely and extensively
distributed on account of their size and lipophilic
nature.
Metabolism of estrogens varies according to the
stage of the menstrual cycle, menopausal status and
genetic polymorphisms. Estradiol has a plasma half-
140
life measured in minutes, and undergoes rapid hepatic
biotransformation in which it is converted by 17b-
hydroxysteroid dehydrogenase to estrone, which is
then converted by 16a-hydroxylation and 17-keto
reduction to estriol, which is the main urinary
metabolite. Ethinylestradiol is metabolized much
more slowly by the liver than the naturally occurring
estradiol. The elimination phase half-life varies from
13 to 27 h. In contrast to estradiol, the main route of
biotransformation of ethinylestradiol is via 2-hydrox-
ylation and subsequent formation of 2- and 3-methyl
ethers. Mestranol, which is another semi-synthetic
estrogen and a component of some oral contracep-
tives, is the 3-methyl ether of ethinylestradiol. It is
metabolized rapidly in the liver by demethylation to
form ethinylestradiol, which is its active form (34).
Progestins
Similar to estrogens, the progestins are generally
rapidly absorbed after oral administration. The nat-
urally occurring hormone, progesterone, undergoes
rapid first-pass metabolism, which leads to low bio-
availability and the requirement for high doses (34).
However, hepatic metabolism of the 19-nor steroids
used in combined oral contraceptives is significantly
slowed because of the ethinyl substituent at C17,
thereby increasing their oral activity. Progesterone is
bound by albumin and corticosteroid-binding glob-
ulin in the plasma, but is not significantly bound to
sex hormone-binding globulin. In contrast, the 19-
nor compounds (norethindrone, norgestrel and
desogestrel) bind to sex hormone-binding globulin
and also albumin. The synthetic compounds are
extensively (>90%) bound to plasma proteins.
The half-life of progesterone is approximately
5 min, with the hormone being rapidly metabolized
in the liver to hydroxylated metabolites and their
sulfate and glucuronide conjugates, which are then
excreted in the urine. The half-life of the synthetic
progestins is much greater, approximately 7 h for
norethindrone, 16 h for norgestrel and 12 h for ges-
todene. It is believed that metabolism of the synthetic
progestins mainly occurs in the liver, and elimination
of the metabolites is via the urine, although the pre-
cise mechanisms are not well defined (88).
Effects of oral contraceptives on
the periodontium
It has long been recognized that increases in circu-
lating levels of sex steroid hormones can have an

impact on the gingival and periodontal tissues, and
these effects can be most obvious during pregnancy.
Pregnant women near or at term produce large
quantities of sex steroid hormones on a daily basis
(20 mg estradiol, 80 mg estriol and 300 mg proges-
terone) with sometimes dramatic effects on the
periodontium (95), and the prevalence and severity of
gingivitis have been reported to be elevated during
pregnancy (52, 85, 86, 138). Indeed, some of the
classic periodontal literature pertains to investiga-
tions of the gingival tissues during pregnancy, and
studies by Loe and Silness in the early 1960s led to
the development of indices for quantifying plaque
and gingival inflammation that are still widely used
today (86, 123). In their seminal study in 1963, Loe
and Silness assessed the gingival condition of 121
pregnant and 60 post-partum women, and found that
all participants (whether pregnant or post-partum)
showed signs of gingival inflammation (86). However,
the severity of inflammation was significantly greater
in the pregnant women, whether assessed using the
Loe and Silness gingival index or the Russell peri-
odontal index (118). Mean probing depths were also
slightly greater during pregnancy compared to post-
partum (86), a finding that has been reported in other
studies (104), and that is considered to result from
gingival enlargement and inflammation as opposed
to attachment loss. The impact of pregnancy on the
periodontium is considered in more detail elsewhere
in this volume.
Observations that pregnancy was associated with
exacerbated gingival inflammation led to a number of
studies that investigated whether similar findings
resulted from use of oral contraceptives. Such studies
were especially prevalent in the late 1960s and early
1970s, probably as a result of the high hormone
concentrations used in first-generation oral contra-
ceptives. The first few volumes of the Journal of
Periodontal Research carried a large number of such
studies. Studies that investigated the impact of sex
steroids on the periodontium are supported by
observations that the gingival and periodontal tissues
are target tissues for these hormones, and localiza-
tion of androgens, estrogens and progestins has been
reported in the periodontium in a variety of species.
For example, autoradiographic studies have con-
firmed the localization of estrogen receptors in
human gingival epithelium cells, fibroblasts and
endothelial cells (144). More recently, estrogen
receptors a and b were both shown to be expressed in
the periodontal ligament cells of rats (11). In humans,
it has been reported that periodontal ligament
cells express estrogen receptor b but not estrogen
Oral contraceptives and the periodontium
receptor a, implying that estrogenic effects in the
periodontal ligament are mediated via estrogen
receptor b (58). Furthermore, it was reported in 1971
that human gingival tissues can metabolise proges-
terone (30), and in 1974 that progesterone is localised
to the oral mucosa of rabbits (105). Specific proges-
terone receptors were subsequently detected in rab-
bit gingiva (143). However, there is no evidence for
the presence of progesterone receptors in the human
periodontal ligament, and progesterone appears to
have no direct effect on periodontal ligament cells
(57, 59). Therefore, it is clear that the periodontium is
a target tissue for sex steroid hormones, although not
to the same extent as other tissues. The primary
target tissues of sex steroid hormones include the
breast, female reproductive tract (uterus, ovary),
bone, the vascular system and central nervous sys-
tem, and, to a lesser extent, the genitourinary tract,
male reproductive tract, gastrointestinal tract, im-
mune system, skin, kidney and lung (150).
The biological significance of the human peri-
odontium as a target tissue remains unclear. It is
difficult to imagine an evolutionary advantage con-
ferred as a result of possessing gingival tissues whose
inflammatory response to plaque is affected by the
circulating levels of sex steroid hormones. It is per-
haps more likely that the periodontal tissues are a
target tissue by virtue of the fact that sex steroid
hormone receptors are found throughout the body
because of the fundamental role played by these
hormones in human homeostasis. In the context of
these observations, the literature regarding the im-
pact of oral contraceptives on the periodontium is
discussed below.
Studies in experimental animals
A number of studies have been published in which
exogenous sex steroid hormones were injected or
applied topically in various animal model systems,
often at doses that far exceed those encountered in
normal human physiology and using experimental
designs that appear to have little relevance with re-
gard to the impact of oral contraceptives on the
periodontium. Many of these studies attempted to
investigate the impact of sex steroid hormones on
inflammation of the periodontium rather than spe-
cifically studying the impact of oral contraceptives on
gingival or periodontal tissues.
Early studies used the so-called granuloma pouch
technique, in which 25 ml of air is injected into the
subcutaneous layer in the back of a rat, creating a
space in which a granuloma develops when an irri-
141
Preshaw
tant such as croton oil is subsequently injected
(121). Using this model, in 1966, if 1 ml of a 1%
solution of dydrogesterone, a synthetic progestin
similar to progesterone, was injected at the same
time as the croton oil, the granulomatous membrane
that formed was much thinner than that induced by
the croton oil alone, and contained far fewer leu-
kocytes, indicating that the progestin may depress
the inflammatory reaction of the connective tissue
(83). The authors acknowledged that the findings of
their research may not directly relate to any antici-
pated effects of progestins on gingival inflammation
because the granuloma pouch is characterized by
acute inflammatory responses as opposed to the
chronic inflammation seen in gingivitis or peri-
odontitis. In a follow-up study (1967), the same
experimental model was used, this time with addi-
tional injections of the same progestin (1 mg ml),
estrogen (1 mg ml) or chorionic gonadotropin
(1,500 IU ml) (84). The pouch wall thickness was
significantly reduced when the connective tissue was
perfused with any of the hormones, suggesting an
anti-inflammatory effect of the hormones. The au-
thors recognized that the concentrations of the
hormones administered far exceeded the concen-
trations occurring naturally in humans.
A subsequent study (1967) investigated the effect of
sex steroid hormones (estrogens, progestins and
gonadotropin) on the microvascular system by
means of vital microscopy in the hamster cheek
pouch (in which the cheek pouch is everted over a
glass plate and viewed under the microscope) fol-
lowing injection of 0.1 ml preparations of hormones
into the tissues of the cheek pouch, or topical
administration onto the surface of a wound created
in the cheek pouch (78). The sex steroid hormones
had rapid effects on the microvasculature in the
cheek pouch, described as vascular impairment and
characterized by reductions in flow rate, adherence of
granulocytes and platelets to vessel walls, formation
of microthrombi, and disruption (degranulation) of
mast cells. The magnitude of the observed changes
was greatest with the highest concentrations of the
applied hormones (particularly 3,000 IU ml chori-
onic gonadotropin and 25 mg ml progesterone).
The authors concluded that, although high doses
were used and the hormones were applied directly
onto or into the tissue, the sex hormones had a def-
inite effect on the microvasculature, and hypothe-
sized that gingivitis in pregnancy may result from a
reduced threshold for tissue injury in response to
plaque induced by elevated concentrations of sex
steroid hormones (78). It should be noted that, in the
142
context of modern oral contraceptives, which are
systemically administered and typically contain daily
doses of 35 lg estrogen and 0.11.5 mg progestin,
the concentrations applied topically in this early
animal experiment (520 mg ml estrogens and 525
mg ml progestins) resulted in hugely elevated local
concentrations that would never be achieved in hu-
mans following ingestion of oral contraceptive pills.
The same experimental model, but using 0.05 ml
volumes of 5 mg ml formulations of estrogens or
progestins delivered locally by injection into the
hamster cheek tissues, was used (in 1967) to study the
effect of sex steroid hormones on vascular perme-
ability (79). Vascular permeability was assessed by
leakage of a protein-bound dye (Evanss blue) that had
previously been administered by intravenous injec-
tion. In contrast to chorionic gonadotropin, estrogens
and progesterone were found to result in endothelial
damage and increased vascular permeability, and the
effect of progesterone was greater than that of estro-
gen. The same model (again published in 1967) was
also used to demonstrate that estrogen and proges-
terone stimulated vascular proliferation at the site of
injection (82). Taken collectively, these studies con-
firm that estrogens and progesterone, when delivered
locally and at high concentrations, result in impaired
vascular function and increased vascular permeability
and stimulate vascular proliferation.
To further elucidate the pathophysiological
effects of these hormones acting via the circulation
(as opposed to the effects resulting from topical
administration), a further series of experiments was
performed using experimental animal model sys-
tems. Following removal of the ovaries of female
hamsters and intramuscular administration of estro-
gen or progesterone, it was found in a publication in
1968, using the cheek pouch model, that estrogen
(0.2 ml at 5 mg ml) resulted in only minor vascular
changes, whereas progesterone (0.2 ml at 5 mg ml)
resulted in enhanced inflammation in the cheek
pouch, characterized by swelling of the endothelium
lining the veins, and an increased number of adher-
ent leukocytes and microthrombi (81).
The microvascular changes following increased
production of female sex hormones in the rabbit were
also studied (1968) using an ear chamber model,
which allows monitoring of revascularization fol-
lowing removal of a small core of tissue from the ear.
An intravenous dose (500 IU) of chorionic gonado-
tropin was administered, resulting in a physiological
state called pseudo-pregnancy that is characterized
by a high production of progesterone (reaching a
peak 1012 days after injection), whereas production

of estrogens remains low. A saline injection was given
in the control group. Following injection, the ear
chambers were examined daily for 4 weeks. In the ear
chamber of the pseudo-pregnant rabbits, vascular
dilatation was evident almost immediately after
hormone injection, the micro-vessels exhibited in-
creased blood flow, and new venules appeared,
demonstrating an increased vascularity of the gran-
ulation tissues that was not observed in the saline-
treated animals (80).
In a study of gingivitis-free experimental dogs
published in 1968, the effect of sex steroid hormones
administered intramuscularly on gingival exudate
was assessed. Estrogen and progesterone were
administered together in doses ranging from 0.5
1.0 mg estrogen and 2550 mg progesterone.
Administration of these hormones resulted in an in-
crease in crevicular fluid flow in these dogs, with a
return to baseline at the end of the period of hor-
mone administration (75). Similar significant in-
creases in gingival exudate were seen when high
doses of estrogen (1 mg every other day) and pro-
gesterone (25 mg every day) were administered to
dogs with chronic gingivitis, and the authors con-
cluded that this effect resulted from an effect of the
sex hormones on the permeability of the dentogin-
gival vessels (74).
In a study published in 1972, the ovaries of female
squirrel monkeys were removed, and either 10 mg
progesterone and 0.2 mg estrogen were administered
twice daily for 50 days, or the animals received pla-
cebo injections (23). Experimental gingivitis was in-
duced using silk ligatures and or by placement of
cotton thread soaked in ovalbumin into the gingival
sulcus. Gingival inflammation was assessed by his-
tological examination, and the most severe gingival
inflammation was seen in the animals who were
receiving placebo injections and had very low circu-
lating levels of sex steroid hormones and cortisol
(following removal of their ovaries). It was postulated
that this inflammation may result from the absence
of the anti-inflammatory effects of cortisol (23). The
authors recognized that their results were at variance
with those of studies that had identified increased
severity of gingivitis during pregnancy (52, 86), and
attributed this to methodological differences between
the studies, and also the fact that their study design
made it impossible to discriminate between the ef-
fects of cortisol, estrogen and or progesterone (23).
In a study that investigated the microvasculature of
the rabbit gingiva following intramuscular injections
of high doses of progesterone (0.5 mg kg body
weight), it was reported (in 1974) that progesterone
Oral contraceptives and the periodontium
resulted in increased permeability of the gingival
vessels as a result of formation of gaps between
endothelial cells that closed after approximately
90 min (106).
Taken collectively, it is difficult to draw firm con-
clusions from these studies of experimental animals,
which appear to present conflicting results. For
example, some studies identified an anti-inflamma-
tory effect of estrogen and progesterone (83, 84),
others identified an effect of sex steroid hormones on
the vasculature (increased vascular permeability and
vascular proliferation) (80, 81), and others reported
increased gingival exudation (74, 75). The results of
these various studies need to be considered in light of
the fact that doses of estrogens up to 400 106 times
the plasma concentration found in non-pregnant
human females and doses of progesterone up to
1 106 times the plasma concentration found in
non-pregnant human females were administered
(95). Furthermore, these studies did not directly at-
tempt to ascertain whether oral contraceptives
themselves have an effect on gingival inflammation,
and indeed suffer from disadvantages when consid-
ered for this purpose, including the use of extremely
high doses of sex steroid hormones and overly com-
plex experimental designs.
One of the first studies (published in 1972) to di-
rectly address the question of whether oral contra-
ceptives affect the periodontium involved female rats
receiving large subcutaneous injections of steroids in
the form of a commonly used oral contraceptive pill
(117). The rats received daily injections of either 104
or 416 lg 100 g body weight of a combined oral
contraceptive pill containing mestranol and nore-
thynodrel (except at weekends). The low dose was
just above the dose used to prevent fertility in rats,
and the high dose was higher than this. A control
group received injections that contained no hor-
mones. Alveolar bone loss was identified over the
91 days of the study in all three groups, but there
were no overt signs of gingivitis or inflammation in
the hormone-treated groups. There was a trend to-
wards the greatest alveolar bone loss over the 91 days
in the control group, with moderate bone loss in the
low dosage group and the least bone loss in the high
dosage hormone group, with significantly more bone
loss in the controls compared to the high dose
hormone group. This difference was no longer sig-
nificant when the bone loss was expressed as a per-
centage of the total mandibular size. The authors
concluded that injection of sex steroid hormones did
not result in any adverse effects on the gingiva or the
alveolar bone (117), and speculated that their find-
143
Preshaw
ings were at variance with previously published
studies investigating the effect of sex steroid hor-
mones on the periodontium as they used much lower
hormone doses than previous studies. Furthermore,
in another study in which the effect of sex hormones
was assessed in oophorectomized female hamsters
vs. control hamsters (who underwent sham opera-
tions), who were then maintained on periodontitis-
inducing diets and received intramuscular injections
of estrogen and or progesterone for 12 weeks, it was
found that administration of the sex hormones had
no significant effect on the degree of alveolar bone
resorption (89).
Since these studies in the late 1960s and early
1970s, there have been very few, if any, research
articles involving use of experimental animals to
investigate the effect of oral contraceptives on the
periodontium, which is entirely appropriate given the
strong evidence in humans (see below) that modern,
low-dose oral contraceptive formulations do not in-
crease the risk for gingival or periodontal disease.
Case reports
A number of case reports of apparent associations
between oral contraceptives and gingival conditions
were published in the 1960s and 1970s. In 1967, the
first description of what was described as oral con-
traceptive hypertrophic gingivitis was published
(127). A 21-year-old female patient presented with
painful gingival hypertrophy and bleeding that was
particularly affecting the interdental papillae of the
anterior teeth. There was also a history of possible
exposure to Vincents infection (which would now
probably be referred to as necrotizing ulcerative
gingivitis). Initial treatment involved use of an oxy-
gen-liberating mouthrinse and scaling, but the con-
dition did not improve. This prompted the clinician
to obtain a more complete history, including whether
the patient was taking an oral contraceptive. It is
perhaps a sign of the times that this question was
only asked after the lack of response to the initial
treatment, and even more notably, that smoking
status was not reported at all. It was established that
the patient was using an oral contraceptive contain-
ing 2 mg norethindrone and 0.1 mg mestranol. This
led to the diagnosis of contraceptive hormone gin-
givitis (although a differential diagnosis of Vincents
infection was also considered). The clinical condition
improved after further oral hygiene instruction and a
change in contraceptive to one containing a lower
dose. The author concluded that a full investigation
of drugs taken by women, including oral contracep-
144
tives, is considered advisable in dental practice. In
the same year, another case report was published of a
19-year-old female who self-medicated daily with
30 mg of an anti-pregnancy pill containing noreth-
indrone and mestranol, and who developed marked
gingival inflammation with very inflamed hypertro-
phic gingiva (90). Following correct prescription of
a 5 mg oral contraceptive by a gynaecologist, the
gingival tissues returned to their normal healthy
state.
In 1969, a lesion that was similar in appearance to a
pregnancy tumor in a 24-year-old women using an
oral contraceptive was reported (61). The individual
concerned had been using a combined oral contra-
ceptive containing the progestin ethynodiol diacetate
(1 mg) and 100 lg mestranol for the last 18 months.
In addition to the lesion itself, there was generalized
gingivitis affecting the anterior regions of the maxilla
and mandible. Oral hygiene instruction was provided
and the oral contraceptive was discontinued. Within
2 months, the lesion had reduced to less than half the
original size, was notably less inflamed, and was
subsequently excised.
In 1974, another case report described another
pregnancy tumor associated with use of an oral
contraceptive (110). A 21-year-old female presented
with a 4-week history of gingival swelling at the
interdental papilla between the central incisors, to-
gether with generalized gingivitis and widespread
plaque deposits. The labial frenum extended close to
the free gingival margin of the papilla concerned,
resulting in difficulties with oral hygiene in that re-
gion. The patient had been taking an oral contra-
ceptive containing 3 mg norethisterone acetate and
50 lg ethinylestradiol for the preceding 2 years. Oral
hygiene instruction was provided, the lesion was ex-
cised, and a partial frenectomy was performed to
improve access for cleaning. Histological examina-
tion confirmed the diagnosis of a pyogenic granu-
loma. Gingival health improved throughout the
mouth, and these improvements were maintained for
the next 21 months. The patient continued to take
the oral contraceptive throughout, and the author
concluded that the hormones played a secondary role
in the etiology of the lesion by producing an exag-
gerated response to the local accumulation of plaque
rather than directly causing the lesion (110). Of
course, with the benefit of hindsight, it is perfectly
possible that the oral contraceptives played no role at
all in the development of this lesion, which most
likely arose as a result of very poor oral hygiene
associated with a prominent labial frenum, as evi-
denced by the photographs presented in the report.

In addition to case reports describing effects of oral
contraceptives on the gingiva, there have also been
reports suggesting links between oral contraceptive
use and dry socket formation following extraction of
third molars (although it is notable that smoking,
which is a risk factor for dry sockets, was not as-
sessed) (133). In addition, it was reported that users
of oral contraceptives were more likely to have evi-
dence of radiopacities (of unknown nature) in the
mandible (as were females in general, compared to
males), an effect that was suggested to be linked to
the effect of estrogens on bone formation (20).
Associations between use of oral contraceptives and
melanotic pigmentation of the gingival tissues have
also been reported (50).
Studies in humans: 1960s and early
1970s
Recognizing that early studies of the effect of oral
contraceptives on gingival inflammation focused on
experimental animals or were reports of individual
cases, a number of researchers began to undertake
clinical investigations in humans. One of the earliest
studies (1967) focused on the quantity of gingival
crevicular fluid before and during 12 months of reg-
ular use of contraceptive preparations in 115 women
aged 1835 years (77). Crevicular fluid was collected
on 10 1.5 mm strips of filter paper that were in-
serted into the sulcus until resistance was felt, held in
place for 3 min, and then ninhydrine-stained to
determine the distance along the strip that the exu-
date travelled. Five samples were collected per pa-
tient from incisors and canines on the basis that it
had previously been reported that the greatest in-
creases in gingivitis during pregnancy occurred at the
interproximal sites of anterior teeth (86). Samples
were collected before oral contraceptive therapy,
then after 2, 6 and 12 months of use, and were col-
lected by an individual who was not aware of the
purpose of the study. Two first-generation combined
oral contraceptives were used: one contained 5 mg
megestrol acetate (a synthetic progesterone) and
0.1 mg mestranol, but the constituents of the other
oral contraceptive preparation were not reported.
The authors found that regular use of the oral con-
traceptive for 12 months increased the amount of
gingival exudate from the anterior teeth, and the in-
crease was statistically significant for both types of
oral contraceptives investigated. For the preparation
containing megestrol acetate and mestranol, the in-
crease in gingival exudate occurred during the first
6 months of contraceptive use. For the other con-
Oral contraceptives and the periodontium
traceptive, no increase was found in the first
6 months, but there was an increase in exudation
between 6 and 12 months. No explanation was of-
fered for this differing response between the two oral
contraceptive preparations. There were no significant
changes in plaque scores during the study, which was
interpreted as indicating that the increases in gingival
exudation could not be attributed to poorer oral hy-
giene but had occurred as a consequence of oral
contraceptive therapy (77).
These authors then went on (in 1969) to investigate
the effect of a raised level of progestins on gingival
exudation in young females by using exudation dur-
ing the normal menstrual cycle as a reference (76).
Crevicular fluid samples were collected from 14 wo-
men aged 1924 years who demonstrated only slight
signs of gingival inflammation clinically, and quan-
tified as described above. Samples were collected
every other day for 2 months (two complete cycles) to
establish baseline variations in gingival exudate, and
then samples were collected for a further 2 months
during which the subjects took 0.5 mg chlormadi-
none acetate (a synthetic preparation with a marked
progestogenic effect) daily. The mean individual
exudate value during the normal menstrual cycle was
0.82 mm, with a statistically significant increase
to 1.01 mm during the period of use of the oral
contraceptive. The authors concluded that use of
contraceptives containing progestins may aggravate
pre-existing chronic inflammation of the periodontal
tissues (76).
A group of researchers in India reported (in 1971)
the findings of a study of 100 users of oral contra-
ceptives and 100 non-users of a similar age and from
the same social background (21). Periodontal status
was assessed using the Russell periodontal index
(118), and oral hygiene was evaluated using the
Greene oral hygiene index (38). The Greene oral hy-
giene index is based on the combined scores for an
oral debris index (extent of coronal coverage by
mucin, bacteria and food) and an oral calculus index
(extent of coronal coverage by supragingival calcu-
lus), both ranging from 0 to 3, where 0 indicates no
coverage, 1 indicates up to one-third coverage, 2
indicates up to two-thirds coverage, and 3 indicates
more than two-thirds coverage). The oral hygiene
index is calculated by addition of the oral debris in-
dex and the oral calculus index, and is assessed at 12
periodontal sites. The Russell periodontal index as-
signs a score of 0 (no inflammation or periodontal
breakdown), 1 (mild gingivitis not entirely circum-
scribing the tooth), 2 (gingivitis that completely
circumscribes the tooth), 6 (gingivitis with suspected
145
Preshaw
pocket formation as a result of attachment loss) or 8
(advanced periodontal breakdown with mobile teeth,
loss of masticatory function, or drifting). Oral con-
traceptive users had significantly higher mean oral
debris index scores than non-users (1.39 0.42 vs.
1.06 0.20) and significantly higher mean oral cal-
culus index scores (1.49 0.43 vs. 1.13 0.31), and
thus significantly higher mean oral hygiene index
scores (2.88 1.65 vs. 2.20 1.02) (21). The mean
Russell periodontal index was also significantly
higher in users of oral contraceptives compared to
non-users (1.12 0.25 vs. 1.03 0.54). Periodontal
disease was seen in practically all patients in both
groups. The authors concluded that oral hygiene was
significantly poorer in those patients taking oral
contraceptives, and that this may have been the
cause of the observed poorer periodontal health in
this group of patients (as opposed to any direct effect
of the oral contraceptives themselves) (21). Closer
examination of the report reveals that the examina-
tion was performed in a portable dental chair at a
family planning clinic in good daylight, that only six
teeth were assessed in each patient (although it is not
stated which teeth), and that there appeared to have
been no attempt to blind the examiner to oral con-
traceptive status. It was also reported that 60% of
users of oral contraceptives demonstrated what was
described as early destructive periodontal disease
(corresponding to a mean Russell periodontal index
score of 0.71.9), compared to 37% of non-users.
Russell himself suggested that a score of 0.71.9 re-
flects beginning periodontal disease or significant Periodontal Index were significantly higher in oral
gingivitis ranging upward toward incipient destruc-
tive disease whereas patients with frankly estab-
lished destructive disease demonstrate mean peri-
odontal index scores of 1.55.0 (119). Thus, the
interpretation in the Indian study that a mean Russell
periodontal index of 0.71.9 demonstrated early
destructive periodontal disease is not quite in keep-
ing with what was proposed by Russell. As a result,
and given the methodological weaknesses noted
above, the findings of this study are difficult to
interpret. The authors themselves presented no
explanation of why oral hygiene was worse in the
users of oral contraceptives compared to non-users.
In a follow-up report, the same authors described
the frequency of occurrence and severity of
periodontal disease in 100 patients taking oral con-
traceptives compared to 100 individuals not taking
oral contraceptives (22). It is not stated whether
these were the same patients as in the first study,
but it is likely that both studies involved the same
patients, as both studies reported data on exactly
146
200 patients, the reported age range was identical in
both studies (1840 years), and the first study was
published in March 1971 while the second study
was published in August 1971. Most importantly, the
mean Russell periodontal index scores reported for
the users and non-users of oral contraceptives were
the same in both studies. In the second report, the
authors evaluated periodontal status at four sites per
tooth for six teeth per patient, this time identified as
the Ramfjord teeth (116), using the Loe and Silness
gingival index (86) and the Russell periodontal index
(118). Mean gingival index scores were significantly
higher in the group taking oral contraceptives
compared to those who were not (1.87 0.51 vs.
0.72 0.34), as were mean Russell periodontal index
scores (1.12 0.25 vs. 1.02 0.54; same data as
previous study). Gingival index scores were higher
for anterior teeth than posterior teeth, and were
higher at interproximal sites compared to smooth
surface sites in both groups. Mean probing depths
were 1.99 mm in the oral contraceptive users and
1.90 mm in non-users (no standard deviations were
presented), with no significant difference between
groups. In the oral contraceptive group, 19% had
gingivitis (meaning they had no Russell periodontal
index scores >2.0), and 81% showed destructive
periodontal disease (meaning they had at least one
site with a Russell periodontal index score >2.0). In
the non-users, the corresponding figures were 36%
(gingivitis) and 64% (destructive periodontitis). The
authors concluded that Gingival Index and Russell
contraceptive users compared to non-users, that
destructive periodontal disease was significantly
higher in patients taking oral contraceptives, yet
(confusingly), that oral contraceptives had no effects
on pocket depths.
Clearly, these studies suffered from significant
methodological weaknesses, such as lack of blinding,
vague and confusing interpretation of periodontal
indices (some of which are themselves recognized
as suffering from methodological problems), and
incorrect presentation of data. Furthermore, the
second study by the Indian group was undertaken
(according to the wording in the first report), to find
out the effects of oral contraceptives after the soft
debris and calculus deposits are removed, but the
second study made no mention of any treatment
being provided, and the assertion in the second study
that their findings confirm the common belief that
the gingival tissues are influenced by oral contra-
ceptives is based purely on the Loe and Silness gin-
gival index and Russell periodontal index data and

does not take into account the level of oral hygiene
which, indicated in the first report, was significantly
worse in the users of oral contraceptives for reasons
unknown.
In the same year (1971), an Egyptian group re-
ported on the gingival condition around the anterior
teeth in 125 females aged 1839 who were regularly
taking an oral contraceptive containing 4 mg me-
gestrol acetate and 0.05 mg ethinylestradiol, and in
50 women of comparable age and socioeconomic
status who were not taking oral contraceptives (29).
Gingival inflammation was assessed using a scoring
system that appears to be loosely based on the Loe
and Silness gingival index (86), and calculus was
scored according to the Greene simplified oral hy-
giene index (39), which examines just six surfaces
(on four posterior and two anterior teeth), but
otherwise uses the same scoring criteria as the
Greene oral hygiene index (38). There was no indi-
cation of any attempt to blind the (single) examiner
regarding oral contraceptive status. The users of the
oral contraceptives consistently had significantly
more gingival inflammation (mean scores from 1.5 to
1.9 depending on duration of use, no standard
deviations presented) compared to non-users (mean
score 0.6), irrespective of the duration of use of oral
contraceptives. There were no differences in oral
calculus index scores between the users and non-
users of the contraceptives. The authors reported
a progressive increase in gingival inflammation
between 3 and 9 months of contraceptive use
(increase in the Loe and Silness gingival index from
1.5 to 1.9), with a progressive decrease between 9
and 24 months to the levels seen at 3 months. Mul-
tiple statistical testing revealed that the differences in
gingivitis scores observed at various time points
within the oral contraceptive group sometimes
achieved statistical significance and sometimes did
not. The authors interpretation of the increase in
gingival inflammation during the first 9 months fol-
lowed by a decrease afterwards is somewhat specu-
lative, and appears to have been made to fit in with
the findings of other researchers (described above),
who reported that increases in gingival exudate pri-
marily occurred within the first 6 months of use of
one oral contraceptive preparation investigated, but
not another (77).
The same data regarding gingival inflammation
and calculus scores in the 125 oral contraceptive
users and 50 control patients were also included in
another report (1970) by the same group in which
they additionally reported the gingival status of 120
pregnant Egyptian women aged 1839 years old (28).
Oral contraceptives and the periodontium
The authors concluded that women taking the oral
contraceptive pill for one or 2 years had gingival
conditions similar to those observed in women dur-
ing the first or second trimester of pregnancy,
whereas those taking the pill for 6 or 9 months had
gingival conditions similar to those seen in the third
trimester of pregnancy. Again, there are no reports of
blinding of the examiner, and the increases and de-
creases in gingival inflammation scores observed in
the various trimesters of pregnancy and after various
durations of oral contraceptive use did not differ
significantly, meaning that it is difficult to draw
conclusions about the similarity or lack thereof of
gingival conditions between groups.
Mindful of the limitations of these early clinical
studies, in 1974 a British research group reported on
an evaluation of periodontal status in 89 women aged
1723 years who were taking an oral contraceptive,
compared to 72 women in the same age range who
had never used oral contraceptives (70). The exam-
iner was blind to the contraceptive status, duration of
contraceptive use, and age of participants. The Loe
and Silness gingival index (86), plaque index (123)
and attachment level were recorded at two sites per
Ramfjord tooth. The progesterone content of the
various oral contraceptives that were used varied
from 1 to 4 mg, and they contained either 0.05 mg
ethinylestradiol or 0.05 mg mestranol. There were
no significant differences between the two groups
in terms of mean plaque scores (0.81 0.08 in the
control group and 0.76 0.07 in the contraceptive
group) or mean Loe and Silness gingival index
scores (0.70 0.05 in the control group and
0.75 0.05 in the contraceptive group). Similarly,
there were no significant differences in mean
attachment loss measurements between the control
group (0.62 0.08 mm) and the contraceptive group
(0.72 0.07 mm).
Faced with the lack of any significant differences
between the study groups in all the parameters that
were tested, the authors proceeded to subdivide the
dataset into four subgroups (it is not indicated whe-
ther or not this subgroup analysis was planned
a priori, though it appears that it was not). The
contraceptive group was divided into those who had
been taking the oral contraceptives for 1.5 years
(and had a mean age of 19.3 years), and those who
had been taking the oral contraceptives for
>1.5 years (and had a mean age of 22.3 years). In
order to divide the control group into two subgroups
with similar mean ages to the two subgroups created
by splitting the contraceptive group, the authors di-
vided the control group into individuals <20 years old
147
Preshaw
(mean age 19.1) and individuals 20 years old (mean
age 22.5). The division of the two main study groups
into these four subgroups appears to have been
entirely arbitrary, but did have the result of revealing
a statistically significant greater mean loss of attach-
ment in those women who had been taking oral
contraceptives for >1.5 years (0.80 0.06 mm) com-
pared to those taking contraceptives for 1.5 years
(0.64 0.07 mm) and the control subgroup of a sim-
ilar age (0.60 0.08 mm). There were no significant
differences in gingival index or plaque index scores
between the four subgroups, according to the results
section of the paper, though the abstract does
misleadingly report lower plaque index and higher
gingival index scores in those who had been using
oral contraceptives for > 1.5 years compared to those
using oral contraceptives for 1.5 years while
simultaneously failing to mention that these differ-
ences did not achieve statistical significance (70).
Studies in humans: late 1970s to early
1990s
During the mid-1970s, the second-generation oral
contraceptives were introduced, containing much
lower hormone doses than first-generation contra-
ceptives. In 1978, research findings were published
relating to gingival inflammation in 93 women aged
1835 years who were taking oral contraceptives,
compared to 75 women who were not taking oral
contraceptives (60). The study was based in the USA.
Many of the women involved in the study were
continuing to take what would now be described as
first-generation oral contraceptives. There was no
attempt to match individuals between the groups. A
modified gingival inflammatory index based on the
Loe and Silness gingival index (86) and an oral debris
index based on coronal coverage of the tooth by
plaque was recorded at one site at each of four teeth.
One examiner with no knowledge of the oral con-
traceptive status of the women performed the
examinations. Overall, the mean gingival inflamma-
tory index was significantly higher in the women
taking oral contraceptives (1.49 0.04) compared to
those who were not (1.20 0.05), whereas oral debris
index scores were significantly lower in those taking
contraceptives (0.81 0.10) compared to those who
were not (0.95 0.04). Various oral contraceptives
were being used by the women, with estrogen con-
tents varying from 50 to 100 lg and progestin con-
tents varying from 0.5 to 25 mg.
A subgroup analysis was then performed, com-
paring the oral findings between the groups taking
148
the various oral contraceptive brands. Three brands
were excluded from this subgroup analysis, including
one first-generation contraceptive with a progester-
one content of 25 mg per day on the grounds that the
drug had since been removed from the market by the
US Food and Drug Administration. The comparison
of overall mean gingival inflammatory index scores
between the contraceptive user and control groups
was not repeated following exclusion of these three
brands. The inter-brand comparisons revealed that
users of certain brands had higher gingivitis and or
lower plaque scores than users of other brands, but,
as reported by the author, these findings were not
conclusive because of the small numbers in each
group and the multiple statistical testing that was
applied. There was no relationship between total
duration of use of oral contraceptives and gingivitis
or plaque scores. The author concluded that current
users of oral contraceptives had higher mean gingi-
vitis scores and lower mean plaque scores than non-
users, and that, although users of certain brands had
higher gingivitis scores than others, no relationship
could be identified between gingivitis scores and the
various progesterone or estrogen contents of the
various brands (60). With regards to the lower plaque
scores observed in the oral contraceptive users, the
author stated that members of the experimental
sample who were taking oral contraceptives [provided]
histories of more social activity and demonstrated
better personal hygiene habits than the control group,
although no information is provided to support this
claim and it is not clear how this assumption was
arrived at or exactly what it means.
In 1981, a study was undertaken in Minnesota,
USA, to assess the subgingival microflora of 54
pregnant women, 27 non-pregnant women and 23
non-pregnant women taking oral contraceptives, all
aged between 18 and 40 (55). It was not stated which
oral contraceptives were used or what the hormonal
dose was. The authors collected plaque samples and
crevicular fluid samples from the mesiobuccal sites of
maxillary first molars, and measured the Loe and
Silness gingival index (86) and probing depths at
Ramfjord teeth. Gingival crevicular fluid volumes
were significantly greater in the pregnant individuals
compared to non-pregnant individuals who were not
using oral contraceptives, but there were no other
significant differences between the groups. The Loe
and Silness gingival index scores were significantly
higher in the pregnant patients compared to the non-
pregnant groups (users or non-users of contracep-
tives), and there were no significant differences in
gingival index scores between the two non-pregnant

groups. Finally, there were no significant differences
in probing depths between any of the groups. Thus,
no statistically significant clinical differences were
found between the group taking oral contraceptives
and the non-pregnant group who were not taking
oral contraceptives, although the authors did observe
a 16-fold increase in Bacteroides species in the group
taking oral contraceptives (55).
In the same year, a research group in London
performed a study to investigate the relationship
between the response of the periodontal tissues to
plaque in individuals who had been taking oral
contraceptives for various periods of time compared
to control subjects who were not using oral contra-
ceptives (109). Participants were aged 1840 years,
and were divided into three groups depending on
duration of contraceptive use: <5 years (n = 63),
>5 years (n = 49), or not using oral contraceptives
(either never had used oral contraceptives, or had
terminated a course of therapy of less than one
years duration at least 12 months previously)
(n = 39). The contraceptives used by the women had
an estrogen content of 2050 lg and a progestin
content of 0.154.0 mg, and the authors stated that
these included high-dose oral contraceptives which
were at one time in common use, but are now no
longer favored. Thus, individuals who had been on
long-term contraceptive therapy (i.e. those in the
group using contraceptives for >5 years) were more
likely to have received higher-dose formulations
during the earlier stages of use. Assessments were
recorded at four sites per Ramfjord tooth, and in-
cluded the plaque index (123), a gingival index based
on the Loe and Silness gingival index (86), probing
depth and loss of attachment. There is no mention of
blinding of the examiner(s) with respect to oral
contraceptive status, duration of use or participants
ages. Rather than presenting overall mean data for
all the subjects in the three groups, the authors
analysed the data for patients aged 1823 years
separately from those for patients aged 2440 years.
The justification for this was that the authors wished
to be certain that any effects of contraceptive use
identified were not simply a result of age. Thus, in
the 1823 year olds, the mean gingival index scores
were 0.58 0.33 for non-users (n = 28), 0.73 0.32
for users <5 years (n = 39) and 0.78 0.36 for users
>5 years (n = 4). In the 2440 year olds, the corre-
sponding scores were 0.85 0.45 for non-users
(n = 11), 0.81 0.48 for users <5 years (n = 24) and
0.89 0.36 for users >5 years (n = 45). It was re-
ported that there was an increase in gingival index
with duration of therapy, particularly in the younger
Oral contraceptives and the periodontium
group, and analysis of variance of the combined data
for both age groups indicated that this was statisti-
cally significant. The authors also reported that gin-
gival inflammation scores were significantly greater
in the older group compared to the younger group
within each oral contraceptive status category.
With regard to attachment loss, attachment
level measurements in the younger group were
0.36 0.22 mm (non-users), 0.35 0.23 mm (users
<5 years users) and 0.48 0.16 mm (users >5 years),
and the corresponding figures in the older group were
0.64 0.46 mm (non-users), 0.66 0.42 mm (users
<5 years) and 0.66 0.40 mm (users >5 years). There
were no significant differences between the oral
contraceptive groups with regard to loss of attach-
ment (i.e. no effect of contraceptive therapy on
attachment levels), but there was a significantly
greater loss of attachment in the older group com-
pared to the younger group.
In order to account for the influence of plaque on
the gingival index and attachment levels, the authors
combined the age groups and then separated the
combined group into three subgroups based on pla-
que scores (low plaque index <0.6, moderate plaque
index 0.60.12, high plaque index >1.2). The distri-
bution of younger and older participants within these
plaque score groups was not reported. The authors
reported that, in general, within the plaque score
categories, gingival index significantly increased with
increasing duration of use of oral contraceptives, but
no post hoc tests were performed to identify exactly
where the significant increases could be found. In the
low and moderate plaque index groups, gingival index
scores increased progressively from non-users to
users for <5 years to users for >5 years (although it is
not stated whether the increases were statistically
significant). In the high plaque index group, gingival
index scores were greater in users for <5 years than
users for >5 years. Gingival index scores significantly
increased from the low plaque index group to the
moderate plaque index group to the high plaque index
group, as expected. In a similar analysis of attachment
levels based on the plaque score subgroups, there was
no significant increase in attachment loss associated
with duration of use of oral contraceptives. However,
as reported above, there was a significant increase in
loss of attachment with age.
The authors concluded that duration of oral con-
traceptive use has a significant impact on gingival
inflammation, but that there were no differences in
attachment levels between groups that could be
attributed to contraceptive use or duration of therapy
(109). However, the limitations of this paper include
149
Preshaw
the methods of analysis, the lack of blinding of
examiner(s), and the fact that the older age group was
likely to contain a higher proportion of women who
had been taking high-dose oral contraceptives (al-
though this was not quantified).
A study of dental status in relation to oral contra-
ceptive use undertaken in Gothenburg, Sweden,
published in 1989, found that, for 356 38-year-old
women, there were no differences in the number of
remaining teeth between current or previous users of
oral contraceptives compared to non-users (44).
Furthermore, there were no significant differences in
interproximal alveolar bone levels determined from
panoramic radiographs between users and non-users
of oral contraceptives. The precise contraceptives
used and their hormonal content were not reported.
The authors concluded that there is no correlation
between the use of oral contraceptives and dental
state.
Studies in humans: mid-1990s and later
By the 1990s, it had become clear that many of the
unwanted effects elicited by oral contraceptives were
dose-dependent, and this realisation led to intro-
duction of the current low-dose oral contraceptive
formulations. Despite numerous studies evaluating
the improved safety of these oral contraceptives in
other body systems (5, 19, 122), very few studies have
investigated the effect of modern low-dose oral con-
traceptives on gingival and periodontal health.
In 1998, a German research group reported find-
ings from a study in which they investigated
the effects on the subgingival microflora of a
third-generation oral contraceptive containing 20 lg
ethinylestradiol and 0.15 mg desogestrel, and a
contraceptive formulation containing 30 lg ethiny-
lestradiol and 2 mg diegonest (a synthetic proges-
terone that is not widely used outside Germany)
(69). Plaque samples were collected in 29 women
aged 2032 years before starting to use the oral
contraceptives, and at 10 and 20 days after starting
the medication. There was no effect of the contra-
ceptives on bleeding on probing, probing depths or
plaque scores recorded at the upper central incisors
during the 3 weeks of the study. In both contra-
ceptive groups, there were significant increases in
the levels of total anaerobic colony-forming units
10 days after starting to use the drugs, but signifi-
cant differences compared to pre-treatment levels
were no longer evident at 20 days. The authors noted
a large increase in the counts of Prevotella inter-
media 20 days after starting to use the oral contra-
150
ceptive preparation containing desogestrel whereas at
10 days, there had been no difference from pre-
treatment levels. No such increase was noted for
the other oral contraceptive preparation studied.
In 2000, a study that investigated the effects of
hormonal contraceptives on the periodontium in a
population of rural Sri Lankan women (aged 17
36 years) was reported (139). The study comprised 32
women who had been using hormonal contracep-
tives for <2 years, 17 who had used them for 2
4 years, and a matched control group of 39 non-
users. This study differed from those described above
in that the contraceptive users were not using only
oral contraceptives. Thus, of the 32 women who had
used contraceptives for <2 years, an unspecified
number were using oral contraceptives and an
unspecified number had received progesterone
injections. All of the 17 women who had used con-
traceptives for 24 years had received progesterone
injections. Matching between test and control groups
was based on age, socio-economic factors, smoking
status, betel chewing, ethnicity, education level,
occupation and childbirth history (one child each).
The oral contraceptive contained 30 lg ethiny-
lestradiol and 0.15 mg levonorgestrel daily. The
progesterone injection contained 150 mg medroxy-
progesterone acetate and was administered every
12 weeks.
A baseline clinical examination was performed,
followed by three further examinations during a
period of < 2 years and 24 years after baseline. The
cross-sectional analyses between groups reported
were based on the clinical findings at the fourth
examination (undertaken within the period 24
years after baseline), and thereby defining the three
study groups based on duration of use of oral con-
traceptives. The plaque index (123) and Loe and
Silness gingival index (86) were measured on all
surfaces of all teeth, and loss of attachment was
measured at the mesial and buccal surfaces of the
teeth. Attachment loss was recorded using a pres-
sure-controlled probe set at 20 g, and intra-exam-
iner reproducibility (with calculation of kappa
scores) was determined by repeated assessments of
gingival index and attachment loss in 10 randomly
selected females who were not participating in the
main study. There is no mention as to whether the
examiner was blind to the oral contraceptive status
of the participants.
There were no significant differences in plaque
scores between the three groups, but there were
significant differences in gingival index scores be-
tween non-users (0.94 0.3) and both users for

<2 years (1.18 0.3) and users for 24 years (1.24 in each quadrant by a single trained and calibrated
0.3). Attachment loss was significantly greater in
users for 24 years (0.56 0.3 mm) compared to
non-users (0.23 0.2 mm), which in turn did not
differ significantly from that in users for <2 years
(0.24 0.2 mm).
The authors concluded that hormonal contracep-
tive use was associated with significantly more
gingival inflammation compared to non-users,
particularly if the duration of use was longer, and also
that the usage of hormonal contraceptives for
24 years resulted in a significant increase in peri-
odontal breakdown (139). This statement, which
suggests a causal link between contraceptive use and
attachment loss, does not appear to be justified by
the data presented in this association study, which
contained no description of the baseline data
recorded 24 years previously or progression of dis-
ease (if any) since then. Further concerns relate to the
lack of blinding of the examiner, the lack of data
regarding age ranges within the study groups, and
also issues relating to normality of data. A concern is
that a very small number of patients with periodontal
disease could have skewed the data, and after all, the
differences in attachment levels were extremely slight
between the study groups. Finally, this study com-
bined data from oral contraceptive users and injected
progestin users, which this is particularly important
as the daily dose of the injected progestin was
approximately 10 times higher than the dose of pro-
gestin in the oral contraceptives.
The impact of modern, low-dose oral contracep-
tives on the development of gingivitis was studied
prospectively in a 3-week experimental gingivitis
study in a population of 30 non-smoking, periodon-
tally healthy women aged 2045 years in the USA
(114). Fourteen of the women were taking oral con-
traceptives of various types, with estrogen contents
2040 lg and progestin contents 501,000 lg, and 16
women were not using oral contraceptives. The mean
duration of use in those taking contraceptives was
22 16 months (range 848 months). All subjects
were scheduled to begin the experimental gingivitis
protocol (day 0) immediately following menses.
Professional prophylaxis was provided 7 days before
day 0, together with oral hygiene instruction. A
split-mouth design was used in which oral hygiene
practices were suspended for the 3 weeks of the
experimental gingivitis phase in one randomly se-
lected quadrant in the maxilla (the test quadrant), but
continued as normal in the contralateral (control)
quadrant. The plaque index (123) and Loe and Silness
gingival index (86) were recorded at six sites per tooth
Oral contraceptives and the periodontium
examiner who was blind to the oral contraceptive
status of the participants. Gingival crevicular fluid
(30-second samples) was collected from four sites per
quadrant (the mesiobuccal and mesiopalatal sites of
the most anterior permanent premolar and molar
teeth present at which there were no interproximal
restorations and there was a normal interproximal
contact point), and the volume was determined using
a calibrated Periotron 8000.
There were no significant differences in mean ages
between the two groups (users 27.3 4.0 years; non-
users 29.3 6.8 years). Also, there were no significant
differences in gingivitis or plaque scores between the
groups at day 0, despite the fact that the oral con-
traceptive users had been taking the medication for
an average of 22 months (although it should be
remembered that all subjects received a dental pro-
phylaxis 1 week before day 0). Statistically significant
increases in plaque and gingivitis scores were ob-
served in the test quadrants (no oral hygiene) but not
in the control quadrants (normal oral hygiene) over
the 21 days of the study, as expected. However, there
were no significant differences in plaque or gingivitis
scores for test or control quadrants between the
groups at any time point, and the users of oral con-
traceptives did not show any significant differences
in plaque and gingivitis scores compared to the group
not taking oral contraceptives. There were also sig-
nificant increases in crevicular fluid volumes in the
test quadrants over the 3 weeks of the study (as ex-
pected following cessation of oral hygiene practices),
with no changes in crevicular fluid volume in the
control quadrants (normal oral hygiene). Again, oral
contraceptive status did not significantly affect cre-
vicular fluid volumes in test or control quadrants at
any point in the study.
The authors commented that these findings
represented a paradigm shift in perceptions of the
relationship between oral contraceptive usage and
gingival disease. The finding that oral contraceptives
did not influence gingival inflammation was sup-
ported by the lack of any effect of oral contraceptives
on crevicular fluid volumes. The authors recognized
that their findings of no effect of oral contraceptives
on gingival inflammation were at variance with the
results of previous studies (21, 22, 28, 77, 109),
although, notwithstanding the methodological con-
cerns with some of the early studies, it should be
remembered that the concentrations of estrogen and
progestins in those early studies were 220 times
greater than those found in current oral contracep-
tive formulations. The authors concluded that mod-
151
Preshaw
ern, low-dose oral contraceptives do not increase the
risk for gingivitis in premenopausal women with no
attachment loss (114).
Perhaps the most compelling and powerful data
regarding any impact of modern oral contraceptives
on the periodontium were published in 1995 (134).
This large-scale, cross-sectional association study
used data from the first (19711974) and third (1988
1994) National Health and Nutrition Examination
Surveys (NHANES) in the USA. The critical develop-
ment that occurred between NHANES I and NHA-
NES III was the switch from high-dose first-genera-
tion oral contraceptive formulations to the modern
low-dose second- and third-generation oral contra-
ceptives that are still used today. The purpose of the
study was to evaluate the historically suggested
association between use of high-dose oral contra-
ceptives and periodontal diseases, and to determine
whether such an association still existed in users
of modern low-dose oral contraceptives aged 17
50 years in the US population (134).
Data relating to 4930 (NHANES I) and 5001
(NHANES III) non-pregnant, premenopausal women
were abstracted. Data for women using non-oral
forms of contraception were excluded. Furthermore,
four women from the NHANES III dataset were ex-
cluded as they were using a brand of oral contra-
ceptives containing >50 lg estrogen per day. A
weakness of the NHANES I data is that periodontal
status was assessed using the Russell periodontal
index (118). The presence of calculus was measured
at six teeth using the simplified oral hygiene index
(39). In NHANES III, periodontal assessments were
performed in randomly assigned half-mouths (one
upper and one diagonally opposite lower quadrant),
and included clinical attachment loss, probing depth,
gingival bleeding and the amount of calculus at the
buccal and mesiobuccal aspect of each tooth.
In NHANES I, gingivitis was defined as at least one
tooth with a Russell periodontal index score 1, and
in NHANES III, gingivitis was defined as bleeding on
probing at 1 site. Thus, both surveys had very low
thresholds for defining the presence of gingivitis. In
NHANES I, periodontitis was defined as at least three
teeth with a Russell periodontal index score of 6. In
NHANES III, periodontitis was defined as at least two
sites with 4 mm attachment loss and probing depth
4 mm. In both surveys, data were abstracted to
allow comparisons between current users of oral
contraceptives, users within the last 6 months, and
non-users. Social and demographic data were re-
ported, including smoking status. The dependent
variables of gingival inflammation and periodontitis
152
were dichotomized as present or absent. Multivari-
able logistic regression analyses were used to assess
the independent effect of oral contraceptive use on
periodontal status while controlling for all other
covariates in the study.
Complete data on oral contraceptive use were
available for 4,665 women in NHANES I and 4,510
women in NHANES III. In NHANES I, 22.1% of wo-
men aged 1750 were current oral contraceptive
users, compared with 20.0% in NHANES III. Oral
contraceptive users were more likely to be <33 years
old, white and non-smokers, with higher levels of
education and income. The mean age of those who
underwent a periodontal examination was 30.0
0.22 years in NHANES I and 32.0 0.12 in NHANE-
S III, and the age of the current oral contraceptive
users within each survey was significantly lower
than that of non-current users (NHANES I: 27.1 vs.
31.1 years, respectively; NHANES III: 25.8 vs. 34.1
years, respectively). No significant differences were
observed in terms of the prevalence of gingivitis or
calculus between current or non-current users in ei-
ther survey. Furthermore, current oral contraceptive
users had a significantly lower prevalence of peri-
odontitis than non-current users in both surveys
(NHANES I: 8.9% vs. 13.3%, respectively; NHANE-
S III: 4.4% vs. 12.5%, respectively).
After stratifying for individual risk indicators such
as age and smoking, there were still no significant
differences in the prevalence of gingivitis between
current oral contraceptive users and non-users. With
regard to periodontal disease, even after stratifying
for risk indicators, current oral contraceptive users
had a consistently lower prevalence of periodontitis
compared to non-current users. The prevalence of
periodontitis increased with age and smoking status.
Interactions between smoking status and oral con-
traceptive use were not statistically significant in ei-
ther NHANES survey, although NHANES I had low
power to detect such interactions because of the
small number of subjects for whom smoking data
were collected.
Multiple regression modeling after adjustment for
confounding variables revealed that, in NHANES I,
oral contraceptive use was associated with a de-
creased risk of gingivitis [adjusted odds ratio
(OR) = 0.65; 95% CI = 0.421.01], although this
association did not reach statistical significance.
There was a significant protective association be-
tween current oral contraceptive use and moderate
periodontitis (adjusted OR = 0.36; 95% CI = 0.13
0.96). In NHANES III, oral contraceptive use was
associated with a slight, but non-significant, de-

creased prevalence of gingivitis (adjusted OR = 0.80;
95% CI = 0.611.02), and a slight, but non-signifi-
cant, decreased prevalence of periodontitis (adjusted
OR = 0.73; 95% CI = 0.501.07). The authors con-
cluded that modern, low-dose oral contraceptives do
not increase the prevalence of gingivitis or peri-
odontitis, and even that there was evidence of a
possible protective effect of contraceptive usage on
gingivitis and periodontitis. The most likely expla-
nation for this is that oral contraceptive use may
represent a proxy for lifestyle behaviors that are
associated with better health care behaviors, higher
levels of income and better periodontal health. The
authors acknowledged that the putative protective
effect of oral contraceptive use on gingivitis and
periodontitis that they observed may be spurious and
a consequence of healthier lifestyles resulting from
healthy women bias. The authors stated that their
analysis did not substantiate the previous theory that
high-dose oral contraceptive use is associated with
gingivitis and periodontitis, and also that there is no
detrimental association between modern low-dose
oral contraceptive use and periodontal disease for the
majority of women who use them. Furthermore, the
authors postulated that the apparent slight protective
effect of oral contraceptives requires re-examination
of the perceived relationship between oral contra-
ceptive use and periodontal diseases.
In 2007, a study was published that bucked the
trend of recent publications reporting a lack of any
impact of oral contraceptives on the periodontium.
The study was performed in Northern Ireland, and
the stated aim was to investigate the relationship
between oral contraceptive use and the presentation
of aggressive periodontitis in young females (107).
However, the clinical protocol was not devised with
this specific question in mind. Instead, a referral
process had been set up in which general dental
practitioners were encouraged to identify aggressive
periodontitis in patients 1835 years of age, and to
then refer them to a secondary care centre for con-
firmation of the diagnosis. The patients underwent a
periodontal examination including the Loe and Sil-
ness gingival index (86), the plaque index (123),
probing depths and attachment levels at six sites per
tooth (full-mouth examinations). The examiner was
blind to the oral contraceptive status of the patients,
and had previously undergone calibration and train-
ing. Mean gingival index and plaque index scores
were not reported; instead, the authors reported the
means of the percentage of sites for each patient that
had a plaque index score or a gingival index score 1.
A diagnosis of aggressive periodontitis was made
Oral contraceptives and the periodontium
based on the percentage of periodontal sites ex-
hibiting attachment loss 3 mm, though it is not
stated how many such sites were required for the
diagnosis of aggressive periodontitis to be assigned.
The aggressive periodontitis was considered localized
if <50% of sites demonstrated 3 mm attachment
loss (and when the sites primarily involved molars
and incisors), and generalized if >50% of sites
demonstrated 3 mm attachment loss. Individuals
exhibiting incidental sites with attachment loss that
were not typical of aggressive disease were assigned a
diagnosis of chronic periodontitis.
One of the principal findings was that 60% of pa-
tients with generalized aggressive periodontitis were
current users of oral contraceptives, compared to 30%
in the localized aggressive periodontitis group and
30% in the chronic periodontitis group. Overall, 50
females were identified by their general dentists and
referred for periodontal examination, of whom 21
women (42%) were taking oral contraceptives at the
time of examination, 21 (42%) were previous users,
and eight (16%) had never used oral contraceptives. In
the statistical analyses, data from the 21 previous
users were combined with those for the eight who had
never used oral contraceptives to form a non-user
group of 29 patients. A very large proportion of the
group were smokers: 67% of contraceptive users and
59% of non-users. The mean duration of oral contra-
ceptive use in the current users was 8.3 4.9 years,
and varied from a few months to 16 years. The mean
duration of oral contraceptive use in the previous users
was 5.1 4.6 years. Details of the types of contraceptive
used were not provided by all patients, but 14 were
taking oral contraceptives containing 30 lg ethinylest-
radiol and four were taking a formulation containing
35 lg ethinylestradiol.
There were no significant differences in the mean
percentage of sites with a plaque or gingival index
score 1 between the oral contraceptive user and non-
user groups. Oral contraceptive users had significantly
more sites with bleeding on probing (44.0 21.3%)
than non-users (31.1 15.6%). Mean probing depths
were significantly greater in oral contraceptive users
(3.3 1.0 mm) than non-users (2.7 0.5 mm), and
users had a significantly greater proportion of peri-
odontal sites with probing depths >4 mm (27 20%)
than the non-user group (14 10%). Similarly, mean
attachment levels were significantly greater in con-
traceptive users (2.6 1.5 mm) than non-users
(1.7 1.0 mm). Regression analyses revealed a rela-
tionship between oral contraceptive use and the
severity of periodontal pocketing which seemed to be
independent of smoking status and age. The authors
153
Preshaw
reported that the independent variable smoking ex-
plained 33% of the variance in mean clinical attach-
ment loss, whereas current medication with oral
contraceptives accounted for 31% of the variance
(both were statistically significant, but age, which
accounted for 24% of the variance, did not achieve
statistical significance). The authors concluded that
patients taking oral contraceptives had poorer peri-
odontal health than non-users, and describe a rela-
tionship between pill use and the extent and severity
of periodontal pocketing that is unrelated to smoking
status and age. Those who had never used oral con-
traceptives were reported to have significantly better
periodontal health than previous users (data not
shown), but were also significantly younger (24.7
4.5 years) than the previous and current users com-
bined (30.6 3.9 years).
There are some methodological questions regard-
ing this paper. Firstly, the population was a selected
population based on possible diagnosis of aggressive
periodontitis by the referring dentist, and the referral
process was not designed to study the impact of oral
contraceptives on aggressive periodontitis. Secondly,
certain data regarding the smoking history are lack-
ing, such as pack-years of smoking, and whether
individuals were ex-smokers. A large proportion of
subjects were current smokers (67% of contraceptive
users and 59% of non-users). This is considerably
greater than the proportion of smokers in the general
Northern Ireland population, in which 31.4% of fe-
males aged 1624 years and 32.7% of those aged 25
34 years are current smokers (31). Furthermore, in
Northern Ireland, 60.5% of females aged 1624 years
and 57.5% of those aged 2534 years are current or
former smokers (31). If the same logic is applied to
the study in question, in which somewhere between
59% and 67% of the study population were current
smokers, then it is highly likely that almost all of the
subjects would have had some history of smoking
(i.e. would either be current or former smokers), and
that there would be virtually no never smokers. If this
is the case, given the known significance of smoking
as a major risk factor for periodontitis, then the
smoking history (whether participants were current
or former smokers) could outweigh oral contracep-
tive status in terms of risk for advanced and aggres-
sive forms of periodontal disease.
Summary
The issue of whether oral contraceptives have any
impact on gingival inflammation, plaque scores,
154
probing depths or attachment levels has presented an
interesting problem for dentists. Early studies per-
formed shortly after the introduction of oral contra-
ceptives, when hormone doses were very high, found
evidence of increased gingival inflammation and
possibly also increased probing depths, often despite
better oral hygiene in the users of the oral contra-
ceptives. However, many of these studies suffered
from significant methodological weaknesses that
would probably not pass the peer review and refer-
eeing processes used by scientific journals today.
Common errors included lack of blinding of exam-
iners (with inherent problems of bias), poor study
designs that were often overly complex, reporting of
identical data in more than one original research
publication, and inappropriate data analyses includ-
ing multiple statistical testing and unfounded sub-
group analyses that sometimes give the impression of
having been performed simply to identify at least
some significant differences. At the same time, a large
number of review articles were published that cited
the same research papers over and over again, often
(but with one or two notable exceptions) without any
form of critical appraisal despite the inadequacies of
many of those original research papers. Indeed, it
appears that more review articles (2, 9, 10, 24, 26, 33,
42, 64, 68, 71, 72, 95, 96, 101, 112, 124126, 130, 131,
154156) have been published on this topic than
original research articles, and this list of reviews does
not include non-English review papers, of which
there are also a very large number. The multiplicity of
review papers undoubtedly contributed to the per-
vading sense among the dental profession that oral
contraceptives increased the risk of gingivitis and
periodontal disease, and perhaps the thinking of
many colleagues followed that described by a corre-
spondent to the British Medical Journal, who, in 1967
reported that the hypertrophic gingivitis of preg-
nancy is also seen in the patient on the pill and
that he now finds [himself] looking for gingivitis in
[his] young female patients (6).
Perhaps this was somewhat inevitable, given that
early conclusions on the effects of oral contraceptives
on the periodontium were based on findings that
gingival inflammation occurs during pregnancy due
to increases in circulating sex steroid hormone con-
centrations. There is no doubt that high doses of sex
steroid hormones have effects on the vasculature. It is
also possible that the early oral contraceptive for-
mulations were associated with exacerbated gingival
inflammation. It is intriguing that many of these early
studies also reported lower plaque scores in the users
of oral contraceptives, a finding that was attributed to

enhanced social activity by the oral contraceptive
users, with the papers typically not elaborating fur-
ther on what this meant or how it was assessed. It is
also worth remembering that oral hygiene practices
have doubtless improved over the 4050 years since
oral contraceptives were introduced. For example,
data from the UK National Adult Dental Health Sur-
veys reveal that the percentage of UK adults who
brush their teeth twice per day or more has risen
progressively from 51% in 1968 to 64% in 1978, 67%
in 1988 and 74% in 1998 (62, 140, 141). Similar pat-
terns have also been seen in other countries (53), and
it is also very probable that gingival inflammation
(irrespective of oral contraceptive use) has declined
over the same time period.
Today, oral contraceptives provide a safe, conve-
nient and effective form of birth control, but the
dental profession still appears to consider that the
use of modern, low-dose oral contraceptives in-
creases the risk for (at least) gingivitis and (possibly)
periodontitis. Indeed, the most recent (1999) World
Workshop on the Classification of Periodontal Dis-
eases included a classification criterion for oral
contraceptive-induced gingivitis under the sub-
heading gingival diseases modified by medications
(4). Most of the data that were considered in support
of this classification criterion were derived from small
clinical studies that were mostly published a quarter
of a century ago, many of which were deeply flawed.
Modern low-dose formulations have apparently not
been subject to the same level of scrutiny regarding
their possible impact on the periodontium compared
to their high-dose predecessors in terms of the
number of reports, but fortunately, in most cases, the
quality of recent studies is undoubtedly better than
that of those performed in the 1960s and 1970s.
Drawing firm conclusions can be difficult if the
quality of the evidence is questionable. It is clear that
a thorough medical history should always include
questions about the use of oral contraceptives. Pa-
tients, whether using oral contraceptives or not,
should always undergo a thorough clinical examina-
tion in order to diagnose gingivitis or periodontitis,
and correct preventive and treatment strategies
should then be initiated. Risk factor assessment
should be undertaken for all patients, including
evaluation of both local and systemic risk factors.
Current oral contraceptive formulations contain
greatly reduced estrogen and progestin doses com-
pared to the first-generation products. It is no longer
appropriate to apply the findings of the earliest
studies that investigated the impact of oral contra-
ceptives on the periodontium to the current situa-
Oral contraceptives and the periodontium
tion, as those studies investigated a type of pharma-
ceutical that is no longer used. In any case, the early
studies contained many methodological weaknesses,
although they do have historical interest.
The strongest evidence as to whether modern for-
mulations still constitute a risk factor for gingivitis or
periodontitis comes from studies performed in the
1990s and later. It is impossible to undertake a meta-
analysis using the outcomes from these studies as the
study designs were all completely different. With one
notable exception (the study from Northern Ireland,
which is discussed above), the most recently per-
formed studies have indicated that oral contraceptives
do not place individuals at any greater risk for devel-
opment of gingivitis or periodontitis. As clinicians and
scientists, we are required to evaluate the evidence
presented to us, to interpret that information, and to
make decisions to promote the best interests of our
patients. The concept of equipoise is useful in this
context, and we are certainly not at equipoise when
considering whether modern oral contraceptives have
an impact on the periodontium. Instead, the balance
of evidence supports the conclusion that there is no
impact of modern oral contraceptives on the peri-
odontal and gingival tissues, and, having considered
the evidence in detail, it is clear that oral contracep-
tives can no longer be considered to constitute a risk
factor for gingivitis or periodontitis.
Acknowledgments
The author wishes to acknowledge the help of Han-
nah Fraser and Kerry Stone during preparation of this
review.
References
1. Akerlund M, Rde A, Westergaard J. Comparative profiles
of reliability, cycle control and side effects of two oral
contraceptive formulations containing 150 micrograms
desogestrel and either 30 micrograms or 20 micrograms
ethinyl oestradiol. Br J Obstet Gynaecol 1993: 100: 832838.
2. Amar S, Chung KM. Influence of hormonal variation
on the periodontium in women. Periodontol 2000 1994: 6:
7987.
3. Appel TB, Arman KA, Birdsall C, Christenson DA, Clark
DO, DeKoning JR, Dreisbach LA, Frost JF, Gustin AE,
Halikis M, Hart TJ, Hlayackek JH, Hoffpauir NS, Hubbell
W, Kangayappan S, Kuk DS, Lackey OM, Lesko CA, Levi-
sion J, Levitan DB, Lucca PA, Maxymiv GW, McWhorter JR,
Melton RW, Minard WD, Polich JJ, Ries T, Stewart SK,
Thornton TL, White WR. A comparison of a new graduated
estrogen formulation with three constant-dosed oral
contraceptives. Contraception 1987: 35: 523532.
155