International Journal of Gynecology and Obstetrics 83 (2003) 141150

Article

Acetic-acid guided visual inspection vs. cytology-based screening

for cervical cancer in the Philippines
C.A. Ngelangel*, G.M. Limson, C.P. Cordero, A.D. Abelardo, J.M. Avila, M.R. Festin,
UP-DOH CCSHOSG
University of the Philippines, Department of Health, Cervical Cancer Screening Health Operations Study Group,
National Institutes of Health, Pedro Gil, Ermita, Manila, Philippines
Received 23 December 2002; received in revised form 15 May 2003; accepted 20 May 2003

Abstract
Objectives: To compare the validity and acceptability of acetic-acid visualization (VIA), magnified acetic-acid
visualization (VIAM), spatulaqcotton swabPapanicolaou (Pap) smear (SS), and cervical brushPap smear (CB)
in the detection of precursoryearly cervical cancer lesions. Methods: A total of 12 992 women aged between 25 and
65 years from 14 Philippine centers were randomly allocated to the four tests. The gold standard was colposcopy
with biopsy for positiveysuspicious cases. Results: Sensitivity rates w95% confidence intervals (CIs)x were 37 (CI,
26.848.5), 34.1 (CI, 24.844.8), 14.3 (CI, 6.427.8), and 19.1 (CI, 9.234.6) for VIA, VIAM, SS, and CB,
respectively. Specificity rates were 90.7 (CI, 89.691.7), 90.7 (CI, 8991.1), 97.5 (CI, 96.898), and 97.9 (CI,
97.398.4), respectively. Kappa for the Pap smear (PS) within centers ranged from y0.154 to 0.783, and between
centers from y0.028 to 0.364. Screeners preferred CB; screened-women preferred VIA. Conclusions: The acetic-acid
visualization and VIAM methods are recommended for initial cervical cancer screening in the Philippines.
2003 International Federation of Gynecology and Obstetrics. Published by Elsevier Ireland Ltd. All rights reserved.
Keywords: Philippines; Cervical cancer screening; Visual inspection of the cervix

1. Introduction
The cervical cancer screening model involving
a Papanicolaou (Pap) test at intervals of 3 years
or more, followed, if necessary, by intensive treatment of dysplasia, is inappropriate in most developing countries w1x. In these countries, resources
are lacking to implement such a model. Cytologic
screening requires infrastructure, technical expertise, and regular tracking of women who test
*Corresponding author. Tel.yfax: q63-2-525-6523..

positive w2x. In India (1992), even with a 12-fold

increase in staff trained in cytologic evaluation,
only one-fourth of the female population aged 35
years or more could possibly be screened by the
next century w3x. Socio-cultural problems exist
(many women have little knowledge of cervical
cancer symptoms and potential cure, and of the
necessity of follow-up visits; moreover, they have
a poor attitude towards cancer).
Continuing education and alternative screening
methods such as cervix visualization may prove
more efficient in reducing cervical cancer in devel-

0020-7292/03/$30.00 2003 International Federation of Gynecology and Obstetrics. Published by Elsevier Ireland Ltd. All rights
reserved.
doi:10.1016/S0020-7292(03)00265-0

142

C.A. Ngelangel et al. / International Journal of Gynecology and Obstetrics 83 (2003) 141150

oping countries. Visual inspection sensitivity was

62.6% in India w3x and 76.7% in Zimbabwe w4x.
As results are known at screening, immediate
treatment can be administered as needed, which
minimizes the number of women lost to followup.
Visual cervical inspection may be a suboptimal
strategy compared with cytologic evaluation if
both programs are done in well-organized facilities
with expert support. However, visual inspection
may be more useful where cancer is common and
cytologic screening is not available, adequate, or
organized.
In the current state of expertise and infrastructure for cervical cancer screening in the Philippines, this study: (1) evaluated the sensitivity rates
(SNR) and the specificity rates (SPR) of VIA,
VIAM, SS, and CB using colposcopy (with biopsy
for positiveysuspicious cases) as the gold standard;
(2) compared the endocervical cell and optimal
smear slide yields of SS and CB; (3) assessed the
reliability of pathologists performing PS cytologic
evaluations; and (4) assessed the acceptability of
the screening methods by the women who underwent the screening and the health workers who
performed it.
2. Methods
This standard criterion study evaluated the current state of cervical cancer screening in the
Philippines and randomly allocated eligible women
to the four screening tests.
Women aged between 25 and 65 years were
recruited from nine government regional hospitals
and five municipal community health units (14
centers) in the Philippines (LuzonVisayasMindanao). Women who were virgins, pregnant, without a cervix, or with symptomatic cervical cancer
were excluded. Women who were at least 6-weeks
postpartum, andyor 3 days from their last menstrual day, were included. An active information and
recruitment campaign was conducted in participating municipalities and hospitals.
Each center had a separate randomization
scheme provided by the RALLOC Software Program, PT Flanagan, 1986. Women were screened
and referred immediately for colposcopic evalua-

tion after VIAyVIAM or 1 week after SSyCB.

However, several women assigned to SSyCB did
not comply with the colposcopy referral. Hence,
the process was changed to immediate colposcopy
referral after SSyCB. If colposcopic evaluation
was positive or suspicious, a biopsy was performed. If the histologic evaluation was negative,
the patient was given reassurance and advised
accordingly; if it was positive, further management
followed.
The PS slides were read first by the cytologic
screener, then by the pathologist.
2.1. Procedure standardization
The University of the Philippines-College of
Medicine (UP-CM) Gynecology and Pathology
Departments conducted the training necessary for
the study. Participating personnel from municipalities and hospitals were trained to understand
standardizing definitions, and to perform procedures and evaluate resultsi.e. they honed their
screening skills and their skills at performing
biopsies; colposcopic, cytologic, and histopathologic evaluations; and treatment. The screening
personnel (nurses and physicians), gynecologist
colposcopists, cyto-screeners, and pathologists
were trained in five separate, 1-month workshops
(-15 persons per workshop). These workshops
consisted in a 1-week refresher course and a 3week study-site practicum. The practicum focused
on screening methods, colposcopyybiopsy, cytology, and cytologyypathology practices, respectively.
Topics included cervix embryology; the origin of
the transformation zone; the methods and pitfalls
of screening; an interactive lecture and wet-clinic
exercises on VIA, VIAM, SS, CB, colposcopy,
cytology, biopsy, and pathology practices; genital
infectionsyHPVySILycancer and their treatment;
setting up a screening clinic; quality assurance;
gender sensitivity; recording; reporting; and referral flow.
Each center had one cyto-screener and one or
two physician screeners (hospitalycommunity) or
nurse screeners (community). Each referral hospital had one colposcopist and one pathologist.
The study staff was the actual screening personnel
of these centers.

C.A. Ngelangel et al. / International Journal of Gynecology and Obstetrics 83 (2003) 141150

One research assistant per center filled out the

pre-test data collection forms and aided in the
laboratory delivery of biopsy specimens to be
evaluated.
All positive slides and a random sample of
negative slides per center were given to one
reference pathologist at the UP-CM Pathology
laboratory for reliability. A consensus reading of
two pathologists was needed if questions occurred.
For intra-observer agreement, the pathologists
did a second reading of 10 randomly selected
slides from their respective hospitals. For interobserver agreement, three to five pathologists read
10 slides from each hospital but not from their
own hospital. There were 4060 readings per
hospital.
2.2. Screening methods
Before the screening procedure, a non-lubricated
speculum gently exposed the cervix. The speculum
had an attached light source (battery-operated speculight; Reister, Germany). Naked-eye visualization was done prior to the screening test.
For VIA, the cervix was washed for approximately 2 min with a large cotton swab that had
been immersed in a 4% acetic acid solution. Cervix
visualization followed within 1 min of swab
removal. For VIAM, the VIA procedure was followed under speculoscope (6=16 magnification,
Trylon, USA). For SS, the ayre spatula was inserted into the cervical external os and rotated 180
degrees circumferentially, gathering cells that were
immediately smeared on a glass slide. Then, a
cotton swab moistened with a saline solution was
inserted into the endocervix and rotated 180
degrees circumferentially, gathering cells that were
smeared on another slide. For CB, a cervical brush
(Rovers BV, the Netherlands) was inserted into
the endocervix and rotated 180 degrees, with
bristles bent fully against the ectocervix and collecting cells that were then smeared on a slide.
The slides were air-dried, immediately fixed,
coded, and Pap-stained after 50% glycerinwater
rehydration.
2.3. Positive results
Positive SSyCB results were ASCUS, AGUS,
or SIL, using the Bethesda reporting system. Pos-

143

itive VIA, VIAM, or colposcopic results were a

grossly elevated mass, aceto-whitening, leukoplakia, punctuation, vesicular lesion, or mosaicism.
Positive histopathologic results were CIN II-III,
CIS, adenocarcinoma, or squamous cell carcinoma,
using standard CIN I-IIIyICD-10 nomenclature.
2.4. Blinding
The cyto-screeners and pathologists were blinded to PS specimen collection method or to visual
inspection method when identifying biopsy samples. The colposcopists were blinded to the screening results.
2.5. Ethical considerations
The Pap test is currently done in the Philippines,
and the acetic-acid wash is a standard adjunct to
colposcopy. Improving the accuracy of cervical
cancer screening would benefit women and society.
Informed consent, privacy, patient autonomy, and
confidentiality were observed.
The ethics committee did not allow biopsy of a
healthy cervix by colposcopy. Treatment followed
if CIN III, CIS, or cancer lesion was detected. The
subject could withdraw from the study at any time
without being penalized or treatment prejudice.
2.6. Data analysis
Analysis used the EPI-INFO 6.04b, SPSS, and
Stata 5 software. Statistical tests were done at as
0.05 (two-tailed) and 95% confidence intervals
(CIs).
For screening test validity: for the four tests,
SNR and SPR were estimated with 95% CIs, and
adjusted by center using the Cochran (1977) method for estimating proportion for stratified random
sampling. For cytologic evaluation reliability: (1)
the kappa agreement statistics was estimated for
the cytologic readings of the 14 centers cytoscreeners vs. the readings of the pathologists; (2)
UP-CM vs. pathologists from the other centers;
and (3) intra- and inter-hospital pathologists.
The study was conducted in 14 different centers
of three distant regions of the Philippines, and this
limited the transfer of subjects and personnel from

144

C.A. Ngelangel et al. / International Journal of Gynecology and Obstetrics 83 (2003) 141150

one center to another. Thus, the reliability of the

examination by all screening and colposcopy personnel of a random 30 subjects per center was not
assessed.
PS Quality: Using a chi-square test, the SS
group was compared with the CB group according
to their optimal smear and endocervical cell yield
and the co-presence of canceryprecursor cells and
endocervical cells.
Screen Acceptability: The acceptability of the
screening procedure by the screener and the
screened women was assessed using the Kruskal
Wallis test. Follow-up compliance was described.
2.7. Sample size
Given SNR estimates (p) of 50% for VIA w3,4x,
50% for VIAM w5x, 82% for SS w6x, and 99% for
CB w6x, 97, 97, 57 and four women, respectively,
were needed for each test.
Dividing each sample size by 2% (the precursoryearly cancer prevalence assumption) allowed
the final required sample size of 4850, 4850, 2850,
and 200 women to be derived, for each corresponding test. These sample sizes allowed for SNR
estimation of each test at the 95% CI level with a
precision of "10%.
3. Results
Of the 14 440 women who were screened, 90%
(12 992) underwent colposcopy and data analysis.
The overall dropout rate (randomized women with
a negative screening result but no colposcopic
evaluation) was 10% (1448), mostly from the SS
(32%) and CB (35%) groups. The main problems
were due to logistics (no transportation, inaccessibility of the screeningyreferral center, and
appointment difficulties), patients misconceptions, and low health priority. There were 3316,
3447, 3195, and 3034 women who underwent
VIA, VIAM, SS, and CB, respectively; 3044
women underwent cervical biopsy.
The screening groups were comparable (P)
0.05) with respect to demographic characteristics
and reproductive history (Table 1).

The VIAM and VIA groups were similar with

respect to the number of positive lesions detected,
which was significantly higher than that of SS or
CB by estimation analysis (Table 2).
An aceto-white epithelium was a frequent positive finding of visual inspection (41%, 49%, and
45% for the VIA, VIAM, and colposcopy groups,
respectively). In the SS group, positive lesions
were 76% ASCUS, 16% LSILyHPV changes, 5%
AGUS, and 2% HSIL, compared with 66%, 15%,
11%, and 7% in the CB group, respectively. In the
SS and CB groups, HSILyAGUS findings were
similar in the menopausal and premenopausal
groups of women (P)0.05).
Biopsy and histologic evaluation (ns3032)
detected 0.6% of CIS to Stage Ia-IIb cancers (55%,
20%, 10%, and 15% for the VIA, VIAM, SS, and
CB groups, respectively); 1.8% of CIN IIIII
(29%, 36%, 20%, and 14%, respectively); and
5.8% of CIN I lesions (28%, 36%, 20%, and 16%,
respectively).
Treatment (observation, antibiotics, cryosurgery,
LEEP, conization, polypectomy, dilatationycurettage, hysterectomy, or radiotherapy) was given to
1324 colposcopy- and biopsy-positive women, of
whom 303 (22.9%) were non-compliant.
3.1. Screen method sensitivity and specificity
With colposcopy (followed by biopsy and histologic evaluation) as the gold standard, all screening tests had low SNR and high SPR. By
estimation analysis, VIA and VIAM had similar
SNRs, and they were significantly higher than
those of SS or CB (Table 3).
There were similar, low SNRs among menopausal (ns1671) and premenopausal women (ns
11 268) with VIA (34.9; 95% CI, 21.550.9 vs.
52.8; 95% CI, 46.659), VIAM (50; 95% CI,
33.366.7 vs. 48.7; 95% CI, 43.354.1), SS (21.1;
95% CI, 10.237.7 vs. 21.4; 95% CI, 16.627.1),
and CB (20.7; 95% CI, 8.840.2 vs. 16.7, 12.2
22.5), respectively.
Compared with community centers, the hospital
centers had: (1) higher SPRs for all four tests; (2)
similar, low SNRs for SS and CB; and (3) similar,
higher SNRs for VIA and VIAM vs. SSyCB
(Table 4a,b).

C.A. Ngelangel et al. / International Journal of Gynecology and Obstetrics 83 (2003) 141150

145

Table 1
Profile of women by screen test
Characteristics

VIA
(ns3336)

VIAM
(ns3467)

SS
(ns3238)

CB
(ns3064)

P-value*

Agemean (S.D.)

39.33
(8.36)

39.40
(8.34)

39.06
(8.56)

39.14
(8.07)

0.33

Marital status
Marriedycohabiting
Separatedydivorced
Widow
Single

3083 (92%)
126
116
11

3225 (93%)
142
91
10

3041 (94%)
103
89
6

2835 (92%)
122
104
7

0.45

Area lived longest

City
Province
Other country

2906 (87%)
427 (13%)
3

3021 (87%)
441 (13%)
5

2828 (87%)
402 (12%)
8

2693 (88%)
362 (12%)
9

0.48

Educational attainment
No schooling
Primary
Secondary
College
Vocational
Post-graduate
Adult school

8
617 (18%)
1183 (35%)
1298 (38%)
61
150
19

6
640 (18%)
1208 (34%)
1368 (39%)
71
149
25

11
577 (17%)
1155 (35%)
1267 (39%)
69
131
28

6
559 (18%)
1079 (35%)
1202 (39%)
58
141
22

0.99

Job longest held

Agricultureyfarmingyfishing
Textileymanufacturingyindustry
Administrativeyoffice
Servicesysales
Domestic
Barymassageydisco
Overseas Contract Worker
Not applicable

78 (2%)
266 (8%)
657 (19%)
778 (23%)
177 (5%)
7
12
1347

83 (2%)
256 (7%)
777 (22%)
763 (22%)
169 (4%)
32
11
1376

82 (2%)
254 (8%)
658 (20%)
728 (22%)
173 (5%)
19
8
1316

77 (2%)
208 (7%)
606 (19%)
738 (24%)
160 (5%)
28
10
1217

0.04

3031 (90%)
180 (5%)
125

3133 (90%)
209 (6%)
125

2930 (90%)
191 (5%)
117

2779 (90%)
169 (5%)
119

0.67

22.7 (4.7);
1349

22.6 (4.6);
1254

22.7 (4.7);
1152

22.5 (4.6);
1048

0.40

13.5 (1.9);
947
2912 (87%)
3181 (95%)

13.6 (1.8);
930
3039 (87%)
3330 (96%)

13.5 (1.7);
920
2800 (86%)
3063 (94%)

13.5 (1.7);
934
2634 (85%)
2913 (95%)

0.49

3.4 (1.9);
114

3.4 (1.8);
113

3.4 (1.8);
115

3.4 (1.9);
113

0.68

1.3 (0.6);
14

1.2 (0.5);
14

1.2 (0.4);
13

1.3 (0.6);
13

0.33

1.4 (0.7);
15

1.4 (0.8);
18

1.4 (0.7);
17

1.4 (0.8);
17

0.63

Smoking history
Non-smoker
Current smoker
Ex-smoker
Age at first sexual intercourse
Mean (S.D.); range
Age at first menstruation
Mean (S.D.); range
Still menstruating
Ever been pregnant
Number of live births
Mean (S.D.); range
Number of still births
Mean (S.D.); range
Number of miscarriages
Mean (S.D.); range

0.17
0.18

146

C.A. Ngelangel et al. / International Journal of Gynecology and Obstetrics 83 (2003) 141150

Table 1 (Continued)
Characteristics

VIA
(ns3336)

VIAM
(ns3467)

SS
(ns3238)

CB
(ns3064)

P-value*

Number of abortions
Mean (S.D.); range

1.4 (0.7);
15

1.3 (0.6);
13

1.6 (0.9);
15

1.3 (0.6);
13

0.07

History of genital infection

183 (5%)

207 (5%)

183 (5%)

208 (6%)

0.13

47.3 (4.5);
2955

46.8 (5.2);
3356

47.4 (4.6);
2656

47.1 (4.8);
2557

0.32

Age at menopause
Mean (S.D.); range

*ANOVA-means; Chi-squareyFishers exact testproportions.

Table 2
Frequency of test positives

12.568.2) for VIA, VIAM, SS, and CB,

respectively.

Test

n (q)

Percentage (95% CI)

VIA
VIAM
SS
CB
Colposcopy
Biopsy

3316
3447
3195
3034
12 992
3032

331
364
87
71
1226
168

9.98 (911.1)
10.7 (9.711.8)
2.07 (2.23.4)
2.3 (1.83)
8.6 (8.29.1)
5.2 (4.56.1)

If we take only the population who had biopsy

and histologic evaluation (ns3032), the SNRs are
63 (95% CI, 42.680), 47.8 (95% CI, 27.568.8),
23.1 (95% CI, 6.353.8), and 36.4 (95% CI,

3.2. PS slide quality

There were 80.3% and 83.2% satisfactory slides
in the CB and SS groups, respectively. Among
SSyCB cases that were biopsy-positive, the smears
were satisfactory in 25.2% of SSyCB () and
88.2% of SSyCB (q) slides.
There were 91.9% of CB- and 95.2% of SSslides with endocervical cells. There were 91%
and 94% of slides with endocervical cells from
the premenopausal and menopausal subjects,

Table 3
SNRySPRa of screen test, all centers
Test property

VIA (ns3316)

VIAM (ns3447)

SS (ns3195)

CB (ns3034)

SNR (95% CI)

37%
(26.848.5%)
90.7%
(89.691.7%)

34.1%
(24.844.8%)
90.1%
(89.091.1%)

14.3%
(6.427.8%)
97.5%
(96.898%)

19.1%
(9.234.6%)
97.9%
(97.398.4%)

SPR (95% CI)

a

Colposcopyqbiopsy as gold standard wonly colposcopy (q) underwent biopsyx; population included all subjects who underwent
colposcopy; colposcopy (y) considered as negative.
Table 4
SNRySPR of screen test: community and hospital centers
Test property

VIA (ns539)

VIAM (ns568)

SS (ns586)

CB (ns552)

(a) Community Centers

SNR (95% CI)
SPR (95% CI)

42.3% (29.156.7%)
74.1% (7077.9%)

45.3% (3457.2%)
76.1% (7279.7%)

15.2% (7.127.4%)
96.8% (94.898.1%)

10.6% (4.821.2%)
96.3% (94.197.7%)

VIA (ns2777)

VIAM (ns2879)

SS (ns2609)

CB (ns2482)

52% (45.758.2%)
98% (97.398.5%)

50% (44.353.7%)
97.7% (9798.2%)

22.7% (17.628.8%)
99.6% (99.399.8%)

19.5% (14.226%)
99.6% (99.299.8%)

(b) Hospital Centers

SNR (95% CI)
SPR (95% CI)

C.A. Ngelangel et al. / International Journal of Gynecology and Obstetrics 83 (2003) 141150

147

Table 5
Acceptability scores of screen methods
VIA (ns3336)

VIAM (ns3467)

SS (ns3238)

CB (ns3064)

Screener
Mean (S.D.)
Mean rank

1.60 (0.76)
6331

1.88 (0.95)
7372

1.59 (0.74)
6315

1.54 (0.72)
6118

Screened
Mean (S.D.)
Mean rank

1.82 (0.79)
6323

1.95 (0.87)
6772

1.88 (0.77)
6592

1.86 (0.77)
6503

P-valuea
0.000

0.000

KruskallWallis test.

respectively, who underwent SS. In the CB group,

the rates were 96% and 95%, respectively. Only
the slides with endocervical cells revealed CIN
lesions: 0.8% in the SS group and 0.6% in the CB
group.
3.3. PS interpretation reliability
The cyto-screeners and pathologists had good
agreement (Kappa, 0.86; 95%CI, 0.80.9) in reading the PS slides.
Reference pathologists and pathologists from
other study centers had poor agreement in their
cytologic evaluations (172 slides; Kappas
y0.002) and biopsy evaluations (320; 0.202).
Intra-observer agreements on PS reading for the
nine hospitals (ns98 random slides, 1018 slides
per hospital) were generally poor (Kappa ranges
y0.1540.783). Only two centers had Kappa)
0.50. Overall intra-hospital reliability was low
(Kappas0.30; 95% CI, 0.110.49). Inter-hospital
Kappa estimates ranged from y0.028 to 0.364.
3.4. Screen method acceptability
The screeners assessed the ease of the screening
procedure from 1 (very easy) to 5 (difficult). The
screened women assessed the screening procedure
comfort from 1 (very comfortable) to 5 (much
discomfort). The screeners and screened women
preferred CB and VIA, respectively (Table 5).
The screeners performed the CB procedure
(mean"S.D.) 3.82"2.52 minutes faster (Ps0.00,
ANOVA)
than
SS
(4.08"2.75),
VIA
(5.09"2.73), or VIAM (6.01"3.15).

3.5. Other observations

Of 12 992 subjects, 8% had painydiscomfort
and 0.5% slight bleeding with speculum insertion.
Of 6763 subjects, 0.03% had slight mucosal burn
and 7% discomfortypain with the VIAyVIAM
procedure. Of 3195 subjects, 0.6% had slight
bleeding with the SS procedure. Of 3034 subjects,
1% had slight to moderate bleeding with the CB
procedure.
4. Discussion
Several studies have explored direct cervical
visualization as an alternative to cytology-based
cervix cancer screening, mostly in low-resource
settings, where cervical cancer is of great concern.
Visual inspection of the cervix can have different
SNRs depending on its gold standard; its SNR can
be comparable with that of the Pap test. Cronje et
al., in South Africa w7x estimated the SNR of
speculoscopy (53.6) similar to that of VIAM
(51.2), using cervicography as the gold standard.
Denny et al., also in South Africa w2x, found
comparable SNRs for VIA (75; 95% CI, 6189),
PS (82; 95% CI, 7194), HPV DNA (72; 95%
CI, 5986), and cervicography (71; 95% CI, 58
85), with histology as the gold standard. Loiudice
et al., in Italy w8x, found a lower SNR for VIAM
plus PS (40; 95% CI, 3446) compared with PS
alone (82; 95% CI, 7787), using colposcopy
(with biopsy if positive) as the gold standard.
Several studies used colposcopy (with or without biopsy) as the gold standard and paramedical
personnel as screeners. Chirenje et al. w9x evaluated
the SNR of VIA (68.4; 95% CI, 53.683.2) in

148

C.A. Ngelangel et al. / International Journal of Gynecology and Obstetrics 83 (2003) 141150

primary health clinics in Zimbabwe. Londhe et al.

w10x, in India, found a higher SNR for VIA (72.4)
than for PS (13.2). University of Zimbabwey
JHPIEGO w4x also found a higher SNR for VIA
performed by nurse-midwifes (76.7; 70.382.3)
than for PS (44.3; 37.351.4).
VIA and PS performed by nurses were found
positive in 3% and 11% of cases, respectively, in
a study by Megevand et al. in South Africa w11x.
In contrast, Sankaranarayanan et al. w12x, in India,
found 9.9% and 10.2% of positive cases in VIA
and PS groups examined by female technicians,
respectively.
The above studies concluded that cervix lesion
detection by VIAyVIAM should be considered in
countries where the equipment for cytology screening could not be completely established and
sustained.
Nanda et al. w13x reviewed the accuracy of
conventional and new cytology-based screening
tests. They concluded that the best estimates were
only moderately accurate and did not achieve high
sensitivity, and that most studies were severely
biased. Cytology has a 3089% SNR in detecting
precancerous lesions. Most SNR were between
60% and 85%, predominantly 8085%, with a 5
20% false negative rate (FNR). Pap test assessment depended on sampling and evaluation.
Inadequate specimen sampling and failure to identify or interpret abnormal cells accurately contributed 50% each to Pap test FNR. Confortini et al.
in Italy w14x, analyzed FNRs and underreported
smears, indicating errors from sampling (74%)
and interpretation (26%).
This study evaluated VIAyVIAM for cervix
cancer screening against SSyCB, under the actual
state of infrastructure and expertise of screening
centers in the Philippines. With colposcopy plus
biopsy as gold standard, VIA, VIAM, SS, and CB
had a 37%, 34%, 14%, and 19% SNR, respectively,
lower than those of the similar studies just
described w4,810xexcept for the SSyCB SNR,
which was similar to that of Londhe et al. w10x in
India. These SNRs improved, if only the population who underwent biopsy was considered. Biopsy
and histologic evaluation detected more CIN IIIIIyCISycancer lesions from subjects screened
with VIA (36%) and VIAM (31%), than from

subjects screened with SS (18%) and CB (15%).

These four tests had high SPR ()90%), particularly the PS methods (97%), reflecting better
diagnostic than screening utility. The hospital centers were comparable with the community centers
as screening areas (similar SNRs) but were better
diagnostic centers (higher SPRs). These centers
would prove better in detecting precursorycervical
cancer cases in the clinics, particularly those with
symptoms, than in the asymptomatic community
at large.
In this study, the cervix specimen sampling was
relatively adequate. However, there was poor cytologyybiopsy interpretation reliability, perhaps largely explaining the low SNR of the SS and CB
groups.
Moreover, colposcopy was used as a complementary gold standard, which in this study had a
low 64.7% SPR (95% CI, 62.966.4) and a 68.9%
SNR (95% CI, 5779; biopsy as gold standard),
compared with the 8596% SNRs of other studies
w1518x. This could be due to inadequate training
and lack of experience, the most frequent pitfalls
of colposcopy w19x. Both low colposcopy SNR
and SPR would bias towards low SNR for all the
screen tests.
5. Conclusion
This study evaluated four cervical cancer screening methods in the context of the current local
expertise and infrastructure of the Philippines, and
found them in a poor state. This could explain
why cervical cancer was the second most important
cancer among women of the Philippines from 1980
to 1995 (20.526.4 per 100 000 population) w20x,
with a 5-year relative age-standardized survival
rate of only 25.2% w21x, reflecting mostly
advanced cancer stages at first diagnosis.
Such a cervical cancer burden in the Philippines
should not be ignored. Control of cervical cancer
must be improved. If a nationwide screening program is to be established, it must be well organized. Although currently the SNR of VIAyVIAM
is low, cytology would prove a less effective
screening method in the Philippines. VIAyVIAM
is recommended as the initial screening method of
choice for such low-resource settings. In another

C.A. Ngelangel et al. / International Journal of Gynecology and Obstetrics 83 (2003) 141150

article using data from this study w22x, VIAyVIAM

presented the highest net benefit for the Philippine
government (US $21M per 100 000 women
screened each), compared with SS and CB programs (US $7M and US $5M, respectively). Traditional cytology and or colposcopybiopsy must
follow positive screening for definitive histopathologic diagnosis.
Consequently, these findings imply changes in
public health policy, mainly in the improvement
of screening, colposcopy, and pathology units, as
well as in the improvement of continuing education
for nurseyphysician screening personnel, of strategies for the compliance of target women, and of
health insurance coverage for preventive services.
The Department of Health must take the lead, with
collaboration from various government and nongovernment agencies. Regular public information
and education campaigns must be a mainstay. The
long-term evaluation of the relative protection
given by VIAyVIAM should be incorporated in
the program.
If the establishment of a nationwide screening
program seems frustrating to pursue, other alternative cervical cancer control initiatives (e.g. HPV
vaccine) should be considered.

w8 x

w9 x

w10x

w11x

w12x

w13x

w14x

References
w1x Symonds RP. Screening for cervical cancer: different
problems in developing and developed world. Eur J
Cancer Care 1997;6:275 279.
w2x Denny L, Kuhn L, Risi L, Richart RM, Pollack A,
Lorincz A, et al. Two-stage cervical cancer screening:
an alternative for resource-poor settings. Am J Obstet
Gynecol 2000;183:383 388.
w3x Singh V, Sehgal A, Luthra UK. Screening for cervical
cancer by direct inspection. Br Med J 1992;304:534
535.
w4x University of ZimbabweyJHPIEGO Cervical Cancer
Project. Visual inspection with acetic acid for cervicalcancer screening: test qualities in a primary-care setting.
Lancet 1999;353:869 873.
w5x Sherris J, Wells ES, Tsu VD, Bishop A. Cervical cancer
in developing countries: a situation analysis. Seattle,
WA: Program for Appropriate Technology in Health,
1993.
w6x US Preventive Service Task Force. 2nd ed. Guide to
clinical preventive services. Baltimore, MD: Williams
& Wilkins, 1996. p. 106.
w7x Cronjie HS, van Rensburg E, Cooreman BF, Niemand
I, Beyer E. Speculoscopy versus acetic acid test for

w15x

w16x

w17x

w18x

w19x

w20x

149

cervical neoplasia. Int J Gynecol Obstet 2000;69:249

253.
Loiudice L, Abbiati R, Boselli F, Cecchini G, Costa S,
Grossi E, et al. Improvement of Pap smear sensitivity
using visual adjunctive procedure: a co-operative Italian
study on speculoscopy (GISPE). Eur J Cancer Prevention 1998;7:295 304.
Chirenje ZM, Chipato T, Kasule J, Rusakaniko S,
Gaffikin L, Blumenthal P, et al. Visual inspection of
cervix as a primary means of cervical cancer screening:
results of a pilot study. Central African J Med
1999;45(2):30 33.
Londhe M, George SS, Seshadri L. Detection of CIN
by naked eye visualization after application of acetic
acid. Indian J Cancer 1997;34(2):88 91.
Megevand E, Denny L, Dehaeck K, Soeters R, Bloch
B. Acetic acid visualization of cervix: an alternative to
cytologic screening. Obstet Gynecol 1996;88:383 386.
Sankaranarayanan R, Wesley R, Somanathan T, Dhakad
N, Shyamalakumary B, Sreedevi AN, et al. Visual
inspection of uterine cervix after application of acetic
acid in detection of cervical carcinoma and its precursors. Cancer 1998;83(10):2150 2156.
Nanda K, McCrory DC, Myers ER, Bastian LA, Hasselblad V, Hickey JD, Matchar DB. Accuracy of Papanicolaou test in screening for and follow-up of cervical
cytologic abnormalities: a systematic review. Ann Intern
Med 2000;132(10):810 819.
Confortini M, Ciatto S, Bonardi L, Bulgaresi P, Cariaggi
MP, Carozzi F, et al. Analysis of false-negative and
underreported smears in Florence district screening program for cervical carcinoma. Tumori 1997;83(6):880
883.
Mitchell MF, Schottenfeld D, Tortolero-Luna G, Cantor
SB, Richards-Kortum R. Colposcopy for diagnosis of
squamous intraepithelial lesions: a meta-analysis. Obster
Gynecol 1998;91(4):627 631.
Cecchini S, Bonardi R, Iossa A, Zappa M, Ciatto S.
Colposcopy as a primary screening test for cervical
cancer. Tumori 1997;83(5):810 813.
Pete I, Toth V, Bosze P. The value of colposcopy in
screening cervical carcinoma. Eur J Gynecol Oncol
1998;19(2):120 122.
Carta G, Stefano L, Catellani PA, Toro G, Moscarini
M. Colposcopy, cytology, histology in diagnosis of
squamous intraepithelial lesions of cervix. Clin Exp
Obstet Gynecol 1999;26(2):60 66.
van Niekerk WA, Dunton CJ, Richart RM, Hilgarth M,
Kato H, Kaufman RH, et al. Colposcopy, cervicography,
speculoscopy and endoscopy. International Academy of
Cytology Task Force summary. Diagnostic Cytology
Towards 21st Century: An International Expert Conference & Tutorial. Acta Cytol 1998;42(2):33 49.
Laudico AV, Esteban DB, Reyes L. Liquido. Philippines
Cancer Facts & Estimates. Manila: Philippine Cancer
Society Inc, 1998.

150

C.A. Ngelangel et al. / International Journal of Gynecology and Obstetrics 83 (2003) 141150

w21x Esteban DB, Ngelangel CA, Lacaya L, Robles E,

Munson ML. Cancer survival in the Philippines
Department of Health Rizal Cancer Registry, Ch. 12.
In: Sankaranarayan R, Block RJ, Parkin DM, editors.
Survival from Cancer in Developing Countries. IARC
Publication, 1998.

w22x UP-DOH Cervical Cancer Screening Group. Delineation

of an appropriate & reliable cervical cancer screening
program for Filipino women. Manila, Philippines:
Department of Health, 2001.