JOURNAL OF CLINICAL MICROBIOLOGY, July 2006, p. 24682474 Vol. 44, No.

7
0095-1137/06/$08.000 doi:10.1128/JCM.01882-05
Copyright 2006, American Society for Microbiology. All Rights Reserved.

Comparative Study of the Epidemiology of Rotavirus in Children from

a Community-Based Birth Cohort and a Hospital in South India
Indrani Banerjee,1 Sasirekha Ramani,1 Beryl Primrose,2 Prabhakar Moses,3 Miren Iturriza-Gomara,4
James J. Gray,4 Shabbar Jaffar,5 Bindhu Monica,1 Jaya Prakash Muliyil,2 David W. Brown,4
Mary K. Estes,6 and Gagandeep Kang1*
Department of Gastrointestinal Sciences,1 Community Health,2 and Child Health Unit III,3 Christian Medical College, Vellore,
India; Virus Reference Department, Centre for Infection, Health Protection Agency, London, United Kingdom4;
London School of Hygiene and Tropical Medicine, London, United Kingdom5; and Department of
Molecular Virology and Microbiology, Baylor College of Medicine, Houston, Texas6
Received 9 September 2005/Returned for modification 14 November 2005/Accepted 6 May 2006

Rotavirus gastroenteritis is the major cause of severe dehydrating diarrhea in children worldwide. This study
compares rotavirus diarrhea in 351 children in a community-based cohort and 343 children admitted to a
hospital during the same period. Clinical information and fecal specimens were obtained during diarrheal
episodes. Fecal samples were screened for VP6 antigen, and the positive samples were G and P typed by reverse
transcription-PCR. Rotavirus was detected in 82/1,152 (7.1%) episodes of diarrhea in the community and
94/343 (27.4%) cases in the hospital. The median age of affected children (7.5 versus 10.5 months) and the mean
severity of symptoms (Vesikari score, 7.6 3.4 versus 11 2.5) were lower in the community. A larger
proportion of children in the community were breast-fed than were children admitted to the hospital (73%
versus 34.8%). In the community, the genotypes identified in symptomatic patients, in order of frequency, were
G1 (36.5%), G10 (17.1%), G2 (15.9%), and G9 (7.3%) and mixed infections (7.3%). The most common G-P
combinations were G1P[8], G2P[4], G1P[4], and G10P[11]. The distribution of G types from hospitalized
children was G1 (46.8%), G9 (19.1%), G2 (8.5%), G10 (1.1%), and 4.3% mixed infections. The most common
G-P combinations were G1P[8] and G9P[8]. This study documents significant genetic heterogeneity of rota-
viruses in the community and the hospital. G10P[11] strains resembling a vaccine candidate strain caused
disease in the community, indicating the need for careful epidemiological studies as well as safety studies for
the vaccine candidates.

Rotaviruses are the major etiological agents causing diar-
rhea in infants and young children. A recent review estimates
that more than 600,000 deaths occur annually in children un-
der the age of 5 years due to rotavirus infection, with a majority
of these deaths occurring in developing countries (36).
Rotaviruses belong to the family Reoviridae and possess a
genome of 11 segments of double-stranded RNA that are
located inside triple-layered viral particles. Rotaviruses are
classified into seven different serogroups (A to G), based on
the antigenic specificities of the capsid proteins. For group A
viruses, typing schemes were introduced based on antigenic
epitopes on the proteins that form the inner capsid (VP6;
subgroups I and II) and on proteins of the outer capsid, the
glycoprotein VP7 (G serotypes) and the spike protein VP4 (P
serotypes). In recent years, reverse transcription-PCR was
used in molecular epidemiological studies. All known G sero-
types have been correlated with genotypes; however, there are
more P genotypes than serotypes identified, leading to a sero-
type/genotype dual nomenclature for P types. At least 15 dif-
ferent G types and 26 different P types have been found in
human and animal infections (11, 24, 29, 46). The incidence
and distribution of group A rotavirus serotypes and genotypes
vary between geographical areas during a rotavirus season and
* Corresponding author. Mailing address: Department of GI Sci-
ences, Christian Medical College, Vellore 632004, India. Phone: 0091-
416-2282052. Fax: 0091-416-2232035. E-mail: gkang@cmcvellore.ac.in.
from one season to the next. Globally, different surveys indi-
cate that G1P[8], G2P[4], G3P[8], G4P[8], and G9P[8] are the
most common G and P types (1, 2, 14, 18). However, in several
regions, viruses of other G-P combinations have been found:
G5P[8] in Brazil, G8P[6] in Malawi, G9P[6] in India and Ban-
gladesh, and G10P[11] in India (6, 20, 28, 39, 44).
The first vaccine to be licensed, RotaShield, included reas-
sortants of the most frequently encountered viral strains G1 to
G4 (38). The vaccine was withdrawn due to a temporal asso-
ciation with intussusception (33). Promising results of large-
scale clinical trials on two new vaccines, Rotateq and Rotarix,
have recently been published (40, 47). The anticipated avail-
ability of an effective vaccine highlights the need to better
define the epidemiology and disease burden associated with
rotavirus. It is important to study local distribution of rotavirus
strains and conduct active surveillance programs for emerging
reassortant strains prior to and after the introduction of vac-
cines.
Most of the body of knowledge on rotavirus infections in
children is based on hospital-centered studies. There is an
inherent referral bias if these studies alone are considered. A
clear understanding of the spectrum of disease can result only
from a combination of community- and hospital-based studies
which investigate both mild and severe disease. A study based
in the community provides geographically representative infor-
mation on the disease burden, strain prevalence, and incidence
rates of rotavirus infection in the community. In contrast, the
hospital-based surveillance system provides detailed informa-

2468
VOL. 44, 2006 COMMUNITY- AND HOSPITAL-BASED ROTAVIRUS STUDY IN INDIA 2469

tion on severity and strain prevalence in children presenting to TABLE 1. Characteristics of children with rotavirus diarrhea
a hospital with diarrhea. In this report, we highlight the differ- in the community and the hospital
ences between rotavirus diarrhea in a community birth cohort Value for cohort
and in a hospital in southern India over a 2-year period. Variable P value
Community Hospital

No. of children with 351 347

MATERIALS AND METHODS diarrhea
Patients from the community. A cohort of 452 children was recruited from an No. of episodes of 1,152 343a
urban slum area in Vellore, South India, over a period of 18 months extending diarrhea
from March 2002 to August 2003. The population of this area was 33,390, of No. (%) of children 82 (7.1) 94 (27.4) 0.001
which 2,895 (8.7%) were under the age of 5 years. Sixty-seven children (14.6%) with rotavirus
dropped out of the study. Thirty-eight children moved out of the study area, 25 diarrhea
refused to continue in the cohort, and 4 children died after recruitment. The Median age (mo) of 7.5 (4.710.2) 10.0 (7.512.5) 0.007
remaining 385 children were monitored between May 2002 and May 2004. Fecal detection (IQR)
samples were collected from these children in the neonatal period, and a baseline Male/female 44/38 60/34 0.15
assessment of the domestic environment was conducted by interview. Thereafter, % of breast-fed 73.0 34.8b 0.001
each child was visited at home twice a week by a field worker in order to collect children
information on any illness of the child and the family. Any child who developed Severity of rotavirus
diarrhea was assessed clinically and treated appropriately. Samples of feces were diarrhea
collected from the child during the diarrheal episode. The child was followed up Mild (%) 32.9 3.3c 0.001
daily until cessation of diarrhea. Moderate (%) 45.1 40.0 c
In this cohort, 351 of 385 children had a total of 1,299 episodes of diarrhea Severe (%) 22.0 56.7c
between May 2002 and May 2004. Samples collected from these children during Mean Vesikari score 7.6 (3.4) 11.0 (2.5)c 0.001
this period were analyzed for the comparative study. (SD)
Patients from the hospital. The study was carried out at the Christian Medical Place of treatment
College, a 1,800-bed hospital in Vellore, India. Vellore is an urban area with a Home (%) 13.4 0
population of 400,000 and is estimated to have approximately 42,000 children Clinic (%) 79.3 0
less than 5 years of age. All children under 5 years of age presenting to the Hospital (%) 7.3 100
hospital with acute gastroenteritis and requiring hospitalization for rehydration Mortality rate (%) 1.2 0 0.28
for at least 6 h were enrolled in the study. A total of 343 children less than 5 years a
Episodes requiring a hospital visit.
of age, hospitalized for acute diarrhea between January 2002 and December b
Data for 69 children.
2003, were screened for rotavirus. c
Data for 90 children.
Informed consent was obtained from the parents of all children enrolled in the
community- and hospital-based studies. The studies were approved by the guide-
lines of the Research Committee of the Christian Medical College, Vellore,
India. The severity of diarrhea was assessed using the Vesikari scoring system (41).
All samples from the hospital and the community were transported on ice Information was collected on duration of diarrhea, maximum number of stools
chests to the laboratory and were processed immediately for the detection of passed per day, duration and peak frequency of vomiting, degree of fever,
rotavirus. Four 2-g aliquots were stored at 70C for strain characterization and presence and severity of dehydration, and treatment. To permit calculation of
future testing. this score, the episode was considered mild for a score of 5, moderately severe
Rotavirus screening. Samples were screened for rotavirus VP6 antigen by a for a score of 6 to 10, and severe for a score of 10.
latex agglutination assay (Meridian Diagnostics, Inc.) until September 2002 (2% For calculating the severity scores for children in the community, a modified
of samples from the community and 39% from the hospital). Thereafter, all Vesikari scoring system was followed. Since accurate temperature measurements
samples were examined for group A rotavirus by an enzyme-linked immunosor- were not possible in the field, temperatures were recorded as normal, low-grade
bent assay (Rota IDEIA; DakoCytomation Ltd., United Kingdom) to detect VP6 fever, and high-grade fever as reported by the caregivers.
antigen following the manufacturers instructions. Statistical analysis. Data were entered in Microsoft Access and analyzed using
RNA extraction and cDNA synthesis. Viral RNA was extracted from 200 l of SPSS v. 9.0. Chi square and Students t tests were performed to determine the
the 10% of rotavirus-positive fecal suspension in balanced salt solution, using significance of differences observed between the two different groups of patients.
guanidine isothiocyanate-silica according to the method described by Boom et al.
(4). It was then eluted in 50 l of RNase-free distilled water containing 40 units
of RNase inhibitor (RNasin; Ambion, United Kingdom).
RESULTS
cDNA was generated from 40 l of the extracted RNA by reverse transcription
Rotavirus diarrhea. A total of 351 children in the commu-
in the presence of random primers (hexamers) [Pd(N)6; Pharmacia Biotech,
United Kingdom], using 400 units of Moloney murine leukemia virus reverse nity cohort presented with 1,299 episodes of diarrhea between
transcriptase (Invitrogen, Life Technologies, United Kingdom). May 2002 and May 2004. The mean age of the group at the end
Genotyping PCR. The resulting cDNA was the template for both VP7- and of this monitoring period was 17 4.3 months. There were 175
VP4-specific typing PCRs, using the oligonucleotide primers described earlier males (49.9%) and 176 females (50.1%). Diarrheal samples
(13, 15, 17, 19). Primers complementary to conserved regions of VP7 were used
for the first-round VP7 PCR to amplify an 881-bp region of the VP7 gene (19).
were collected during 88.7% of the episodes of diarrhea. The
The second-round typing PCR was a multiplex PCR and incorporated the 3- age range of children presenting with diarrhea was 0 to 24
end-labeled primer (VP7 Reverse) and the G-type-specific primers G1, G2, G3, months (interquartile range [IQR], 3 to 10 months, indicating
G4, G8, G9, and G10 (19). that 50% of the cases occurred between 3 and 10 months).
Primers Con2 and Con3 were used for the first-round PCRs to amplify an
Rotavirus was detected in 82/1,152 (7.1%) episodes for which
876-bp fragment of the VP4 gene (17). The second-round PCR included primers
specific for genotypes P[4], P[6], P[8], P[9], P[10], and P[11] and the consensus samples were obtained.
primer Con3 (19). In the hospital, 343 children were admitted with diarrhea
Assessment of severity. Diarrhea was defined as the passage of three watery from January 2002 to December 2003. Rotavirus was detected
stools in a 24-h period. In children less than 6 months of age, a change in number in 94 children, accounting for 27.4% of all diarrhea cases. The
or consistency of stools reported by the mother was considered indicative of
diarrhea. An episode was defined as at least 1 day of diarrhea, preceded and
age range of children presenting with diarrhea to the hospital
followed by 2 or more days without diarrhea (32). The episode was considered to was 0 to 60 months (IQR, 5 to 12 months). In seven cases, none
have ended on the day bowel movements returned to normal. of whom were rotavirus positive, age was not recorded on the
2470 BANERJEE ET AL. J. CLIN. MICROBIOL.

TABLE 2. Distribution of community and TABLE 4. Genotyping of rotaviruses from

hospital rotavirus infections by agea children in the community
No. of No. of No. (%) of children with indicated genotype
No. (%) of RV- No. (%) of RV- P type
diarrheal diarrheal
Age (mo) positive cases in positive cases in G1 G2 G9 G10 Mixed UT a Total
episodes in episodes in
community hospital
community hospital
P4 10 11 2 23 (28.0)
05 522 25 (4.8) 85 11 (12.9) P6 1 1 (1.2)
611 431 39 (9.0) 123 52 (42.3) P8 11 1 1 13 (15.9)
1217 157 18 (11.5) 71 19 (26.7) P11 7 7 (8.5)
18 and above 42 0 (0.0) 57 12 (21.1) Mixed 2 1 1 4 (4.9)
Unknown 0 0 7 0 UT 7 2 5 6 2 12 34 (41.5)

Total 1,152 82 (7.1) 343 94 (27.4) Total 30 (36.5) 13 (15.9) 6 (7.3) 14 (17.1) 6 (7.3) 13 (15.9) 82
a a
RV, rotavirus. UT, untypeable.

patient charts. Demographic, virological, and clinical data of
children in the community and the hospital with rotavirus
diarrhea are shown in Table 1.
The median age of detection of rotavirus in children from
the community was 7.5 months (IQR, 4.7 to 10.2) compared to
10.0 months (IQR, 7.5 to 12.5) in the hospital (P 0.007). A
larger proportion of those admitted in the hospital were male
children (63.8%) than were those in the community (53.7%).
However, this difference did not reach statistical significance
(P 0.15). A greater proportion of children in the community
were breast-fed than were those admitted to the hospital (73%
versus 34.8%) (P 0.001) at the time they presented with
gastroenteritis.
In the community, 32.9% of the episodes of rotavirus diar-
rhea were mild, 45.1% were moderate, and 22.0% were severe.
During these episodes, 13.4% of children were treated at
home, 79.3% were taken to the study clinic, and 7.3% required
hospitalization. Not all severe cases required admission, be-
cause they received supervised oral rehydration at the clinic. In
the community, six young children with severe dehydration
were admitted (mean age, 4.3 months; range, 0 to 10 months;
four of six breast-fed). In contrast, only 3.3% of rotavirus
diarrheal cases admitted in the hospital were mild, 40.0% were
moderate, and 56.7% were severe (P 0.001). The mild cases
were admitted to the hospital to alleviate parental anxiety. The
mean Vesikari score was also higher in the hospital than in the
community (11.0 2.5 versus 7.6 3.4) (P 0.001). There
was one rotavirus-associated death in the community (1.2%)
and no documented mortality in the hospital.
The distribution of rotavirus infection in the community and
hospital by age is given in Table 2. The largest proportion of
rotavirus diarrhea was noted in the 12- to 17-month age group
in the community and in the 6- to 11-month age group in the
hospital (11.5% and 42.3%, respectively). The smallest propor-
tion of rotavirus diarrhea was noted in the 0- to 6-month age
group in both settings. The difference in severity across age
groups was not found to be significant in both the community
(P 0.3) and the hospital (P 0.6) (Table 3). Severity scores
could not be determined for four children admitted to the
hospital because they were discharged before the visit of the
study medical officer.
Temporal distribution. No significant seasonal trends were
observed in the community as well as in the hospital. However,
there appeared to be a peak of rotavirus diarrhea in the
months of July to September in 2003, corresponding to the
rainy season.
Genotyping. All of the 82 positive episode samples collected
from the community were available in sufficient quantity for
VP7 genotyping. G and P typing results could be obtained for
84.2% and 58.5% of the samples, respectively. The G types
identified in children with diarrhea from the community, in
order of frequency, were G1 (36.5%), G10 (17.1%), G2
(15.9%), and G9 (7.3%). The P types identified were P[4]
(28.0%), P[8] (15.9%), P[11] (8.5%), and P[6] (1.2%) (Table
4). In children requiring admission from the community, the G
types were G10 (50%), mixed infections with G1 and G9 and
with G1 and G10, and untypeable strains.
For the hospital samples, G and P typing results could be
obtained on 79.7% and 65.9% of the samples, respectively. The
most common G type was G1 (46.8%), followed by G9 (19.1%)
and G2 (8.5%) (see Table 6). The most common P type was
P[8] (55.2%). Other P types included P[4] (5.3%), P[6] and
P[11] (2.2%), and P[10] (1.1%) (Table 5).

TABLE 3. Relationship of age with severity of rotavirus

diarrhea in the community and the hospital TABLE 5. Genotyping of rotaviruses from children in the hospital
No. of children No. of children No. of severe No. (%) of children with indicated genotype
No. (%) of P type
with rotavirus with rotavirus cases/total
Age (mo) severe cases G1 G2 G9 G10 Mixed UT a Total
diarrhea in diarrhea in no. of cases in
in community
community hospital hospital (%)a
P4 5 5 (5.3)
05 25 6 (24) 11 5/10 (50.0) P6 1 1 2 (2.2)
611 39 6 (15.4) 52 31/50 (62.0) P8 36 13 3 52 (55.2)
1217 18 6 (33.3) 19 11/19 (57.9) P10 1 1 (1.1)
18 and above 0 0 12 4/11 (36.4) P11 1 1 2 (2.2)
Mixed
Total 82 18 (22.0) 94 51/90 (56.7) UT 8 3 4 17 32 (34.0)
a
Severity scores could not be determined for four children admitted to the Total 44 (46.8) 8 (8.5) 18 (19.1) 1 (1.1) 4 (4.3) 19 (20.2) 94
hospital because they were discharged before the visit of the study medical
a
officer. UT, untypeable.
VOL. 44, 2006 COMMUNITY- AND HOSPITAL-BASED ROTAVIRUS STUDY IN INDIA 2471

TABLE 6. Major G-P combinations in the community and the hospital and relationship with age and severity of diarrheaa
No. of children with
Median age (mo) of children in: Median severity score (IQR)b in:
RV strain RV diarrhea in:

Community Hospital Community Hospital Community Hospital

G1P4 10 0 8.3 6.5 (5.57.5)

G1P8 11 36 6.0 9.5 5.0 (3.07.0) 10.5 (9.012.0)
G2P4 11 5 11.8 9.5 6.0 (3.09.0) 12.0 (10.513.5)
G9P6 0 1 5.0 7.0
G9P8 1 13 5.1 10.0 7.0 12.0 (10.513.5)
G10P11 7 1 0.4 12 6.0 (3.09.0) 9.0
Mixed 6 4 9.2 9.5 6.0 (2.010.0) 15.0 (12.018.0)
G untyped 1 2 12.2 23.5 9.0 12.0 (0)
P untyped 22 7 7.6 8.0 6.5 (4.58.5) 11.0 (9.013.0)
G and P untyped 13 25 7.9 10.0 7.0 (4.59.5) 11.0 (9.212.7)

Total 82 94 7.5 10.0 6.0 (3.58.5) 11.0 (9.013.0)

a
RV, rotavirus.
b
IQR values represent the upper and lower limits for the months in which 50% of cases occurred.

The most frequent G-P combination in the community was
G1P[8] (13.4%; 95% confidence interval [CI], 6.0 to 20.8%)
and G2P[4] (13.4%; 95% CI, 6.0 to 20.8%), followed by G1P[4]
(12.2%; 95% CI, 5.1 to 19.3%), G10P[11] (8.5%; 95% CI, 2.5
to 14.5%), and G9P[8] (1.2%; 95% CI, 1.2 to 3.6%). Mixed
infections with more than one G type accounted for 7.3% of
rotavirus infections. Twelve samples (14.6%) remained un-
typeable for both VP7 and VP4 (Table 4) despite being posi-
tive in a VP6 PCR.
The most common G-P combinations seen in the hospital
were G1P[8] (38.3%; 95% CI, 28.5 to 48.1%), G9P[8] (18%;
95% CI, 10.2 to 25.8%), and G2P[4] (5.3%; 95% CI, 0.8 to
9.8%) (Table 5). One case each of infection by G9P[6] and
G10P[11] was also seen. Mixed infection was seen in four
samples, accounting for 4.3% of all rotavirus infections. Sev-
enteen samples remained untypeable (18.1%) for both VP7
and VP4 (Table 5) despite being positive in a VP6 PCR. The
dual infections in the community and hospital were G1-G2,
G1-G9, G1-G10, and G2-G10 combinations. One instance of
triple infection was seen in the hospital setting with G1-G2-G9.
Table 6 summarizes the major G-P combinations in the
community and the hospital. The distribution of the G-P types
among the community was significantly different from that
among the hospital (P 0.001). The age distribution and
severity scores of various rotavirus G-P types are also tabulated
in Table 6 for both hospital and community children. No sig-
nificant difference was noted among the various types. Overall,
four mixed infections were severe.
Eight children in the community-based cohort had a repeat
episode of rotavirus diarrhea during this period of follow-up.
Three children were reinfected with the same G type of rota-
virus during the second episode, two with G1, and one with a
G10 strain. The durations between infections in these children
were 16, 24, and 60 weeks. In the remaining five children, the
durations between infections were 3, 6, 16, 24, and 28 weeks.
Mixed infections were seen in three children during the second
episode, and two were severe.
DISCUSSION
This study is a comparison of rotavirus diarrhea in two dif-
ferent settings within the same geographical location during
overlapping 2-year periods. There were 82 rotavirus-associated
diarrheal episodes in the community and 94 rotavirus-associ-
ated diarrheal admissions to the hospital which were subjected
to comparative analysis.
Rotavirus-infected children admitted to the hospital were
older than their counterparts in the community (P 0.007).
This difference may have been due to the closer monitoring of
the community children, resulting in the early identification of
symptomatic children and prompt treatment. It is also possible
that the severity of rotavirus is greater in older children, and
this accounted for the older age of the children admitted to the
hospital.
A significantly smaller proportion of children who presented
with rotavirus diarrhea in the hospital were breast-fed than
were children from the community (P 0.001). A relationship
between breast-feeding and rotavirus diarrhea has been dem-
onstrated, with a twofold-greater risk of rotavirus diarrhea in
non-breast-fed children than in those who were exclusively
breast-fed (27). Lactadherin in human milk is believed to in-
terfere with rotavirus replication (34). The increased replica-
tion in non-breast-fed children may partially explain the
greater severity of rotavirus diarrhea in the children admitted
to the hospital. However, in the absence of information on the
socioeconomic status of hospitalized children and cultural be-
liefs related to breast-feeding, the lower levels of breast-feed-
ing in children requiring hospitalization cannot be explained,
except possibly by their older age.
A recent review of 24 community-based studies and 72 hos-
pital-based studies indicated that rotavirus accounted for a
median of 8.1% and 21.3% of diarrheal episodes in the two
settings, respectively (37). The percentage of diarrheal epi-
sodes attributable to rotavirus in our study conformed to this
trend, with rotavirus accounting for 7.1% of diarrheal episodes
in the community and 27.4% in the hospital (P 0.0001). The
diarrhea seen in the hospital was more severe (56.7% versus
22.0%), and the mean Vesikari score was also significantly
higher for children in the hospital than for children in the
community (P 0.001). This difference may be due to an
inherent referral bias, as more-severe cases need admission in
the hospital. Economic considerations may also be responsible
for this referral bias. While it is possible in other settings that
2472 BANERJEE ET AL.
severe cases in the community have greater mortality rates and
do not reach the caregiver because of an economic disadvan-
tage, we did not find any significant difference in rotavirus-
associated mortality between the two groups (P 0.28). The
pooled mortality rate of all rotavirus diarrheal episodes was
0.57% in our study and is consistent with the published esti-
mates of rotavirus-associated mortality of 0.34% (37). Despite
being a large referral hospital, free treatment is given to chil-
dren and economic reasons may not play a significant role in
determining health-seeking behavior of patients in this setting.
In developed countries, the most prevalent rotavirus strains
causing childhood diarrhea are G1 to G4 and G9 (21). In
contrast, there is significant diversity of circulating rotavirus
strains in India, though G1 and G2 are the most commonly
isolated strains (21, 25). We found that the G1 strain accounted
for 36.5% of episodes of rotavirus diarrhea in the community and
46.8% in the hospital, showing that it still continues to be the
dominant strain. The most common G-P combinations in the
community were G1P[8] and G2P[4], each accounting for 13.4%
of all rotavirus diarrheal episodes. In comparison, G1P[8] ac-
counted for 38.3% of all episodes of diarrhea in the hospital
setting. A recent review of global rotavirus genotypes states
that G1P[8] accounts for 70% of rotavirus infections in North
America, Australia, and Europe, whereas it represents about
30% of infections in South America and Asia and about 23%
in Africa (42).
A significant finding in this study was the high percentage of
G1P[4] strains in the community, accounting for 12.2% of
rotavirus diarrhea. This strain was not detected in hospitalized
children. In a previous hospital-based study from Vellore, this
strain comprised 11% of the total rotavirus strains (22). It has
been reported in 14% of isolates from Argentina (1, 16) and
may be an emerging natural reassortant strain.
Rotavirus serotype G9 is recognized as the most widespread
of the emerging genotypes, representing 4.1% of global rota-
virus infections and accounting for up to 70% of rotavirus
infections in recent reports (26, 30, 42, 43, 48). In our study,
this strain was identified in 7.3% of rotavirus diarrhea cases
from the community and 19.1% of cases from the hospital,
where it was the second most common strain.
The G10P[11] strain is a bovine strain which has been iden-
tified as a cause of asymptomatic neonatal infection in India
(10). Recent studies from India identified this strain in infants
and young children as a cause of diarrhea (20). In our study,
the G10P[11] strain was detected in 8.5% of rotavirus diarrhea
cases in the community and in 1.1% of those in the hospital.
However, the G10 genotype was the second most common G
type in children from the community, accounting for 17.1% of
all rotavirus diarrhea cases. The G10 strains are suspected to
have arisen in the human population through interspecies
transmission or animal-human rotavirus reassortment (7, 45).
In Indian communities, there is close proximity of humans and
animals, facilitating the potential for zoonotic transmission.
The cohort is based in an urban slum setting, and the differ-
ence of G10 prevalence between the community and the hos-
pital children may have been due to differences in socioeco-
nomic status between the two populations. It is possible that
the children from the community cohort belong to a lower
socioeconomic status, making them more prone to acquire
animal strains.
J. CLIN. MICROBIOL.
The rates of untypeable strains were 14.6% in the commu-
nity children and 18.1% in the hospital children. Genotyping of
G types has been more successful than that of P types in all
Indian studies published so far, and the results presented here
are comparable with previously published data (23). Due to the
constant accumulation of point mutations through genetic drift
and to the emergence of novel genotypes, possibly zoonotic
transmission and subsequent reassortment, the reagents and
methods used for genotyping require close monitoring and
updating. We have developed methods and oligonucleotide
primers which were used to overcome failures to type G9, G10,
and P rotavirus strains (11, 23). In other studies, the detection
rates of nontypeable strains have ranged from 6.5% to 16%
(12, 31).
There have been two previous epidemiological studies of
rotavirus diarrhea from the years 1983 to 1985 and 1995 to
1998 in Vellore (5, 22). A comparison of the data for the
various rotavirus strains isolated from previous studies and the
combined community and hospital data from the current study
shows an interesting pattern emerging. The contributions of
various strains as causative agents of diarrhea in these three
time periods (1983 to 1985, 1995 to 1998, and 2002 to 2004) are
as follows: G1, 32.6%, 39.6%, and 42%; G2, 6.5%, 19%, and
11.9%; G3, 7%, 0.8%, and 0%; G4, 30%, 23.8%, and 0%; G9,
0%, 3.9%, and 13.6%; and G10, 0%, 0%, and 8.5%. Based on
these data, there appears to have been a gradual disappear-
ance of G3 and G4 strains and an emergence of G9 and G10
strains.
The impact of the type of circulating rotavirus strains in the
community and the hospital on vaccine development cannot be
overemphasized. Current vaccine development efforts are tar-
geted at the six most common types of rotavirus, G1 to G4, G9,
and P[8]. A live attenuated vaccine based on a human strain,
Rotarix, has been on trial in four countries and has shown
efficacy against severe rotavirus gastroenteritis caused by G1
and non-G1 types, including the emerging G9 type (B. De Vos,
A. C. Linhares, G. Ruiz-Palacios, L. Guerrero, B. Salinas, I.
Perez-Schael, Abstr. 8th Intl. Symp. Double-Stranded RNA
Viruses, Lucca, Italy, abstr. RT-4, 2003), although protection
against the G9 type may have been due to all G9 strains being
combined with P[8]. It has currently been licensed for use in
Mexico (9). In our study, the G1 strain constituted 37% of all
rotavirus diarrhea cases in the community and 47% of cases in
the hospital. Unless there is a significant heterotypic response
to this strain, it appears that this vaccine will fail to protect a
majority of the children in this population. Another candidate
vaccine, Rotateq, is a reassortant pentavalent vaccine with a
backbone of bovine rotavirus and surface proteins of human
serotypes G1, G2, G3, G4, and P[8] (35). Preliminary data
from a Finnish study suggest that there was about 75% pro-
tection for overall rotavirus infection (35). This vaccine has
been licensed for use in the United States. The presence of a
significant proportion of G9 and G10 rotavirus strains in the
community would theoretically make this vaccine effective in
only 55% of affected children.
The observation that rotavirus strains infecting neonates are
usually asymptomatic and that neonatal infection protects
against subsequent rotavirus infection has led to the interest in
these strains as putative vaccine candidates (3). In this context,
it is important to evaluate the safety of the neonatal strain I
VOL. 44, 2006 COMMUNITY- AND HOSPITAL-BASED ROTAVIRUS STUDY IN INDIA
321, a G10P[11] strain (7). We have previously noted that this
strain causes symptomatic diarrhea in neonates and older chil-
dren, and sequencing of VP7, VP6, VP4, and NSP4 has shown
significant homology between circulating G10P[11] strains and
I 321 strains (20). In this study, G10P[11] was responsible for
8.5% of rotavirus diarrhea cases in the community and 1.1% in
the hospital. In addition, the median severity score for G10
infection was in the moderately severe range (community, 6;
and hospital, 9 [as shown in Table 6]). In light of these findings,
it is imperative that stringent safety trials are conducted prior
to the use of this strain as a vaccine.
This comparative study of community and hospitalized chil-
dren with rotavirus diarrhea has several important implica-
tions. It has clearly shown that there are distinct differences in
the ages of infection and severities of symptoms in these two
settings. Variations in genotypes suggest a possible difference
in genotype-specific severity, possible zoonotic spread, and the
emergence of reassortant viruses. The broad diversity of rota-
virus strains also indicates that unless a vaccine induces a broad
heterotypic response, there may be a limited protective re-
sponse and the composition of the vaccine may need to be
changed frequently to reflect the changes in circulating viruses.
ACKNOWLEDGMENT
This work was supported by the Wellcome Trust under the Trilateral
Cooperative Initiative for Research in Infectious Diseases in the De-
veloping World (grant number 063144).
REFERENCES
1. Arguelles, M. H., G. A. Villegas, A. Castello, A. Abrami, P. D. Ghiringhelli,
L. Semorile, and G. Glikmann. 2000. VP7 and VP4 genotyping of human
group A rotavirus in Buenos Aires, Argentina. J. Clin. Microbiol. 38:252
259.
2. Arista, S., E. Vizzi, D. Ferraro, A. Cascio, and R. Di Stefano. 1997. Distri-
bution of VP7 serotypes and VP4 genotypes among rotavirus strains recov-
ered from Italian children with diarrhea. Arch. Virol. 142:20652071.
3. Bhan, M. K., J. F. Lew, S. Sazawal, B. K. Das, J. R. Gentsch, and R. I. Glass.
1993. Protection conferred by neonatal rotavirus infection against subse-
quent rotavirus diarrhea. J. Infect. Dis. 168:282287.
4. Boom, R., C. J. Sol, M. M. Salimans, C. L. Jansen, P. M. Wertheim-van
Dillen, and J. van der Noordaa. 1990. Rapid and simple method for purifi-
cation of nucleic acids. J. Clin. Microbiol. 28:495503.
5. Brown, D. W. G., M. M. Mathan, M. Mathew, R. Martin, G. M. Beards, and
V. I. Mathan. 1988. Rotavirus epidemiology in Vellore, South India: group,
subgroup, serotype, and electrophoretype. J. Clin. Microbiol. 26:24102414.
6. Cunliffe, N. A., J. S. Gondwe, R. L. Broadhead, M. E. Molyneux, P. A. Woods,
J. S. Bresee, R. I. Glass, J. R. Gentsch, and C. A. Hart. 1999. Rotavirus G
and P types in children with acute diarrhea in Blantyre, Malawi, from 1997
to 1998: predominance of novel P[6]G8 strains. J. Med. Virol. 57:308312.
7. Das, M., S. J. Dunn, G. N. Woode, H. B. Greenberg, and C. D. Rao. 1993.
Both surface proteins (VP4 and VP7) of an asymptomatic neonatal rotavirus
strain (I321) have high levels of sequence identity with the homologous
proteins of a serotype 10 bovine rotavirus. Virology 194:374379.
8. Reference deleted.
9. De Vos, B., T. Vesikari, A. C. Linhares, B. Salinas, I. Perez-Schael, G. M.
Ruiz-Palacios, L. Guerrero Mde, K. B. Phua, A. Delem, and K. Hardt. 2004.
A rotavirus vaccine for prophylaxis of infants against rotavirus gastroenteritis.
Pediatr. Infect. Dis. J. 23:S179S182.
10. Dunn, S. J., H. B. Greenberg, R. L. Ward, O. Nakagomi, J. W. Burns, P. T.
Vo, K. A. Pax, M. Das, K. Gowda, and C. D. Rao. 1993. Serotypic and
genotypic characterization of human serotype 10 rotaviruses from asymp-
tomatic neonates. J. Clin. Microbiol. 31:165169.
11. Estes, M. 1996. Rotaviruses and their replication, p. 16251655. In D. M.
Knipe, B. N. Fields, and P. M. Howley (ed.), Fields virology, 3rd ed. Lip-
pincott-Raven Publishers, Philadelphia, Pa.
12. Fischer, T. K., H. Steinsland, K. Molbak, R. Ca, J. R. Gentsch, P. Valentiner-
Branth, P. Aaby, and H. Sommerfelt. 2000. Genotype profiles of rotavirus
strains from children in a suburban community in Guinea-Bissau, western
Africa. J. Clin. Microbiol. 38:264267.
13. Gentsch, J. R., R. I. Glass, P. Woods, V. Gouvea, M. Gorziglia, J. Flores,
B. K. Das, and M. K. Bhan. 1992. Identification of group A rotavirus gene 4
types by polymerase chain reaction. J. Clin. Microbiol. 30:13651373.
2473
14. Gentsch, J. R., P. A. Woods, M. Ramachandran, B. K. Das, J. P. Leite, A.
Alfieri, R. Kumar, M. K. Bhan, and R. I. Glass. 1996. Review of G and P
typing results from a global collection of rotavirus strains: implications for
vaccine development. J. Infect. Dis. 174(Suppl. 1):S30S36.
15. Gouvea, V., R. I. Glass, P. Woods, K. Taniguchi, H. F. Clark, B. Forrester,
and Z. Y. Fang. 1990. Polymerase chain reaction amplification and typing of
rotavirus nucleic acid from stool specimens. J. Clin. Microbiol. 28:276282.
16. Husain, M., P. Seth, L. Dar, and S. Broor. 1996. Classification of rotavirus
into G and P types with specimens from children with acute diarrhea in New
Delhi, India. J. Clin. Microbiol. 34:15921594.
17. Iturriza-Gomara, M., J. Green, D. W. Brown, U. Desselberger, and J. J.
Gray. 1999. Comparison of specific and random priming in the reverse
transcriptase polymerase chain reaction for genotyping group A rotaviruses.
J. Virol. Methods 78:93103.
18. Iturriza-Gomara, M., J. Green, D. W. Brown, M. Ramsay, U. Desselberger,
and J. J. Gray. 2000. Molecular epidemiology of human group A rotavirus
infections in the United Kingdom between 1995 and 1998. J. Clin. Microbiol.
38:43944401.
19. Iturriza-Gomara, M., G. Kang, and J. Gray. 2004. Rotavirus genotyping:
keeping up with an evolving population of human rotaviruses. J. Clin. Virol.
31:259265.
20. Iturriza Gomara, M., G. Kang, A. Mammen, A. K. Jana, M. Abraham, U.
Desselberger, D. Brown, and J. Gray. 2004. Characterization of G10P[11]
rotaviruses causing acute gastroenteritis in neonates and infants in Vellore,
India. J. Clin. Microbiol. 42:25412547.
21. Jain, V., U. D. Parashar, R. I. Glass, and M. K. Bhan. 2001. Epidemiology
of rotavirus in India. Indian J. Pediatr. 68:855862.
22. Kang, G., J. Green, C. I. Gallimore, and D. W. Brown. 2002. Molecular
epidemiology of rotaviral infection in South Indian children with acute
diarrhea from 19951996 to 19981999. J. Med. Virol. 67:101105.
23. Kang, G., S. D. Kelkar, S. D. Chitambar, P. Ray, and T. Naik. 2005. Epi-
demiological profile of rotaviral infection in India: challenges for the 21st
century. J. Infect. Dis. 192(Suppl. 1):S120S126.
24. Kapikian, A. Z., Y. Hoshino, and R. M. Chanock. 2001. Rotaviruses, p. 1787
1833. In B. N. Fields and P. M. Howley (ed.), Fields virology, 4th ed., vol. 2.
Lippincott-Raven Publishers, Philadelphia, Pa.
25. Khetawat, D., P. Dutta, S. K. Bhattacharya, and S. Chakrabarti. 2002.
Distribution of rotavirus VP7 genotypes among children suffering from
watery diarrhea in Kolkata, India. Virus Res. 87:3140.
26. Kirkwood, C., N. Bogdanovic-Sakran, E. Palombo, P. Masendycz, H. Bugg,
G. Barnes, and R. Bishop. 2003. Genetic and antigenic characterization of
rotavirus serotype G9 strains isolated in Australia between 1997 and 2001.
J. Clin. Microbiol. 41:36493654.
27. Kurugol, Z., S. Geylani, Y. Karaca, F. Umay, S. Erensoy, F. Vardar, M. Bak,
I. Yaprak, F. Ozkinay, and C. Ozkinay. 2003. Rotavirus gastroenteritis
among children under five years of age in Izmir, Turkey. Turk. J. Pediatr.
45:290294.
28. Leite, J. P., A. A. Alfieri, P. A. Woods, R. I. Glass, and J. R. Gentsch. 1996.
Rotavirus G and P types circulating in Brazil: characterization by RT-PCR,
probe hybridization, and sequence analysis. Arch. Virol. 141:23652374.
29. Martella, V., M. Ciarlet, K. Banyai, E. Lorusso, A. Cavalli, M. Corrente, G.
Elia, S. Arista, M. Camero, C. Desario, N. Decaro, A. Lavazza, and C.
Buonavoglia. 2006. Identification of a novel VP4 genotype carried by a
serotype G5 porcine rotavirus strain. Virology 346:301311.
30. Masendycz, P., N. Bogdanovic-Sakran, C. Kirkwood, R. Bishop, and G.
Barnes. 2001. Report of the Australian Rotavirus Surveillance Program,
2000/2001. Commun. Dis. Intell. 25:143146.
31. Maunula, L., and C. H. van Bonsdorff. 1995. Rotavirus serotypes and elec-
tropherotypes in Finland from 1986 to 1990. Arch. Virol. 140:877890.
32. Morris, S. S., S. N. Cousens, C. F. Lanata, and B. R. Kirkwood. 1994.
Diarrhoeadefining the episode. Int. J. Epidemiol. 23:617623.
33. Murphy, T. V., P. M. Gargiullo, M. S. Massoudi, D. B. Nelson, A. O.
Jumaan, C. A. Okoro, L. R. Zanardi, S. Setia, E. Fair, C. W. LeBaron, M.
Wharton, and J. R. Livengood. 2001. Intussusception among infants given an
oral rotavirus vaccine. N. Engl. J. Med. 344:564572.
34. Newburg, D. S., J. A. Peterson, G. M. Ruiz-Palacios, D. O. Matson, A. L.
Morrow, J. Shults, M. L. Guerrero, P. Chaturvedi, S. O. Newburg, C. D.
Scallan, M. R. Taylor, R. L. Ceriani, and L. K. Pickering. 1998. Role of
human-milk lactadherin in protection against symptomatic rotavirus infec-
tion. Lancet 351:11601164.
35. Orellana, C. 2003. Rotavirus vaccine shows promise. Lancet Infect. Dis.
3:396.
36. Parashar, U. D., C. J. Gibson, J. S. Bresse, and R. I. Glass. 2006. Rotavirus
and severe childhood diarrhea. Emerg. Infect. Dis. 12:304306.
37. Parashar, U. D., E. G. Hummelman, J. S. Bresee, M. A. Miller, and R. I.
Glass. 2003. Global illness and deaths caused by rotavirus disease in chil-
dren. Emerg. Infect. Dis. 9:565572.
38. Perez-Schael, I., M. J. Guntinas, M. Perez, V. Pagone, A. M. Rojas, R.
Gonzalez, W. Cunto, Y. Hoshino, and A. Z. Kapikian. 1997. Efficacy of the
rhesus rotavirus-based quadrivalent vaccine in infants and young children in
Venezuela. N. Engl. J. Med. 337:11811187.
39. Ramachandran, M., B. K. Das, A. Vij, R. Kumar, S. S. Bhambal, N. Kesari,
2474 BANERJEE ET AL.
H. Rawat, L. Bahl, S. Thakur, P. A. Woods, R. I. Glass, M. K. Bhan, and J. R.
Gentsch. 1996. Unusual diversity of human rotavirus G and P genotypes in
India. J. Clin. Microbiol. 34:436439.
40. Ruiz-Palacios, G. M., I. Perez-Schael, F. R. Velazquez, H. Abate, T.
Breuer, S. C. Clemens, B. Cheuvart, F. Espinoza, P. Gillard, B. L. Innis,
Y. Cervantes, A. C. Linhares, P. Lopez, M. Macias-Parra, E. Ortega-
Barria, V. Richardson, D. M. Rivera-Medina, L. Rivera, B. Salinas, N.
Pavia-Ruz, J. Salmeron, R. Ruttimann, J. C. Tinoco, P. Rubio, E. Nunez,
M. L. Guerrero, J. P. Yarzabal, S. Damaso, N. Tornieporth, X. Saez-
Llorens, R. F. Vergara, T. Vesikari, A. Bouckenooghe, R. Clemens, B. De
Vos, and M. ORyan. 2006. Safety and efficacy of an attenuated vaccine
against severe rotavirus gastroenteritis. N. Engl. J. Med. 354:1122.
41. Ruuska, T., and T. Vesikari. 1990. Rotavirus disease in Finnish children: use
of numerical scores for clinical severity of diarrhoeal episodes. Scand. J. In-
fect. Dis. 22:259267.
42. Santos, N., and Y. Hoshino. 2005. Global distribution of rotavirus serotypes/
genotypes and its implication for the development and implementation of an
effective rotavirus vaccine. Rev. Med. Virol. 15:2956.
43. Sung, Y. L., Y. F. Huang, C. F. Huang, K. T. Chen, H. Y. Chen, G. H. Lee,
B. R. Nong, C. S. Lin, C. M. Chuang, F. C. Liu, C. S. Chen, C. C. Lin, K. S.
Hsieh, and Y. C. Liu. 2004. Emergence of G9 serotype rotavirus as a major
cause of infectious gastroenteritis in southern Taiwan. J. Microbiol. Immu-
nol. Infect. 37:322326.
44. Unicomb, L. E., G. Podder, J. R. Gentsch, P. A. Woods, K. Z. Hasan, A. S.
J. CLIN. MICROBIOL.
Faruque, M. J. Albert, and R. I. Glass. 1999. Evidence of high-frequency
genomic reassortment of group A rotavirus strains in Bangladesh: emer-
gence of type G9 in 1995. J. Clin. Microbiol. 37:18851891.
45. Urasawa, S., A. Hasegawa, T. Urasawa, K. Taniguchi, F. Wakasugi, H.
Suzuki, S. Inouye, B. Pongprot, J. Supawadee, S. Suprasert, et al. 1992.
Antigenic and genetic analyses of human rotaviruses in Chiang Mai, Thai-
land: evidence for a close relationship between human and animal rotavi-
ruses. J Infect. Dis. 166:227234.
46. Varshney, B., M. R. Jagannath, R. R. Vethanayagam, S. Kodhandharaman,
H. V. Jagannath, K. Gowda, D. K. Singh, and C. D. Rao. 2002. Prevalence of,
and antigenic variation in, serotype G10 rotaviruses and detection of sero-
type G3 strains in diarrheic calves: implications for the origin of G10P11 or
P11 type reassortant asymptomatic strains in newborn children in India.
Arch. Virol. 147:143165.
47. Vesikari, T., D. O. Matson, P. Dennehy, P. Van Damme, M. Santosham, Z.
Rodriguez, M. J. Dallas, J. F. Heyse, M. G. Goveia, S. B. Black, H. R.
Shinefield, C. D. Christie, S. Ylitalo, R. F. Itzler, M. L. Coia, M. T. Onorato,
B. A. Adeyi, G. S. Marshall, L. Gothefors, D. Campens, A. Karvonen, J. P.
Watt, K. L. OBrien, M. J. DiNubile, H. F. Clark, J. W. Boslego, P. A. Offit,
and P. M. Heaton. 2006. Safety and efficacy of a pentavalent human-bovine
(WC3) reassortant rotavirus vaccine. N. Engl. J. Med. 354:2333.
48. Zhou, Y., L. Li, B. Kim, K. Kaneshi, S. Nishimura, T. Kuroiwa, T. Nishimura,
K. Sugita, Y. Ueda, S. Nakaya, and H. Ushijima. 2000. Rotavirus infection in
children in Japan. Pediatr. Int. 42:428439.