S114 • JID 2005:192 (Suppl 1) • Bahl et al.
S U P P L E M E N T A R T I C L E
Incidence of Severe Rotavirus Diarrhea
in New Delhi, India, and G and P Types
of the Infecting Rotavirus Strains
Rajiv Bahl,1Pratima Ray,1Swati Subodh,1Prashant Shambharkar,1Manju Saxena,1Umesh Parashar,2Jon Gentsch,2
Roger Glass,2and M. K. Bhan,1for the Delhi Rotavirus Study Groupa
1Center for Diarrheal Disease and Nutrition Research, Department of Pediatrics, All India Institute of Medical Sciences, New Delhi, India;
Gastroenteritis Division, Centers for Disease Control and Prevention, Atlanta, Georgia
A total of 62,475 children !5 years old from a defined population of ∼500,000 children and adults from slums
in New Delhi, India, were assessed for 1 year by means of passive surveillance, to identify children who were
hospitalized for diarrhea. The incidence of severe rotavirus diarrhea was estimated, and the G and P types of
the infecting rotavirus strains were determined and were correlated with the clinical severity of diarrhea. Of
584 children who were hospitalized with diarrhea, 137 (23.5%) had rotavirus detected in stool specimens
(incidence of rotavirus diarrhea–associated hospitalizations, 337 hospitalizations/100,000 children !5 years of
age). Most cases of diarrhea (98%) occurred during the first 2 years of life, peaking at 9–11 months of age.
Rotavirus-associated diarrhea occurred year-round but was predominant in winter. Among the strains that
could be G-typed, G1 was the most common serotype, followed by G9 and G2; 10% of cases of diarrhea were
due to mixed G-type infections. Common strains identified in the present surveillance study were PG1,
PG2, PG9, PG1, PG9, and PG3. Children infected with G1 strains had a greater risk ofdeveloping
more-severe cases of diarrhea than did children infected with other rotavirus strains (odds ratio, 2.95; 95%
confidence interval, 1.3–6.67).
Rotavirus is the most common cause of severe diarrhea
in children worldwide . In India, health carefacility–
based studies of children have indicated that rotavirus
infection accounts for ∼26% of all diarrhea-relatedhos-
pitalizations . There are no population-based data
from which to estimate the incidenceofsevererotavirus
diarrhea and related hospitalizations. These data would
be valuable not only to determine sample sizes for the
evaluation of rotavirus vaccines but, also, to define pol-
icy regarding the priority of rotavirus vaccines among
the available interventions in public health programs
aimed at reducing mortality and morbidityamongchil-
dren and the associated health care costs. Also, because
Financial support: Department of Vaccines and Biologicals, World Health
Potential conflicts of interest: none reported.
aMembers of the study group are listed at the end of the text.
Reprints or correspondence: Prof. M. K. Bhan, Dept. of Pediatrics, All India
Institute of Medical Sciences, New Delhi 110029, India (email@example.com).
The Journal of Infectious Diseases
? 2005 by the Infectious Diseases Society of America. All rights reserved.
rotaviruses show unusual genomic diversity in India
and in other developing countries [3–8], continuous
surveillance for different serotypes is important to as-
sess the appropriateness to the local population of the
antigenic makeup of the vaccines under development.
We conducted a longitudinal studyinwhichallchildren
!5 years old from a defined population of ∼500,000
children were monitored after being admitted with di-
arrhea to 6 hospitals, to determine the incidence of se-
vere rotavirus disease and the proportion of G types
and P types among infecting rotavirus strains and their
association with the clinical severity of diarrhea.
MATERIALS AND METHODS
Study population and census.
∼120,000 households (median household size, 4.5 in-
dividuals) of low socioeconomic status fromsouthNew
Delhi, India, was chosen for the present study. A door-
to-door survey was conducted by trainedfield-workers,
to identify the number of children !5 years of age in
the study population. This number was the denomi-
A population of
Incidence of Severe Rotavirus Diarrhea • JID 2005:192 (Suppl 1) • S115
among 62,475 children !5 years of age in New Delhi, India.
Age-specific incidence of severe rotavirus diarrhea
aNo. of cases/100,000 children/year. Corrected for 65% coverage of hos-
pitalizations anticipated at the 6 hospitals included in the surveillance study.
bAll children !60 months of age.
Seasonality of severe rotavirus diarrhea in New Delhi, India
nator used in the calculation of the overall and age-specific
incidences of severe rotavirus diarrhea.
Selection of hospitals for passive surveillance.
a door-to-door survey, information was collected about anyhos-
pitalization that had occurred since the birth of any child ?5
years of age who resided in the household. These data indicated
that ∼65% of hospitalizations of childreninthestudypopulation
were covered by 6 hospitals in the southern part of New Delhi
and that the remaining children were hospitalized in numerous
small nursing homes located in and around the study area. Data
until age 1 year, also suggested that approximately two-thirds of
hospitalized infants were admitted to the following 6 hospitals
in south Delhi (M.K.B., unpublished data): Safdarjang Hospital,
All India Institute of Medical Sciences, Batra Hospital, Majeedia
Hospital, Holy Family Hospital, and Jeevan Hospital. We there-
fore monitored these hospitals during the present study, antic-
ipating that ∼65% of admissions of children from the selected
population would be captured.
Surveillance for identification of hospitalized children with
Surveillance for the identification of children hos-
pitalized for diarrhea was conducted for the calendar year from
1 August 2000 through 31 July 2001. All hospitals selected for
inclusion in the study were visited on all work days by a study
team led by a physician. All children admitted to theemergency
department, diarrhea treatment unit, or pediatrics wardof each
hospital were screened. Children residing in the study area who
were admitted for diarrhea and who were hospitalized for at
least 6 h were included in the study. Theaddressesofthechildren
were confirmed by the parent(s) and were matched with the
address recorded on each child’s case report form. Clinical in-
formation was recorded for and a stool sample was obtained
from each of the children studied.
By means of
presence of rotavirus by use of a monoclonal antibody–based
EIA kit (Premier Rotaclone; Meridian Diagnostics), which was
used according to the manufacturer’s instructions. Of the 137
rotavirus-positive stool specimens tested, 135 were character-
ized for the G and P outer capsid proteins; the remaining 2
stool specimens could not be tested because the amount of the
sample obtained was insufficient. Rotavirus double-stranded
RNA was extracted from stool specimens by use of the standard
glass powder method described elsewhere . G and P types
were determined using seminested reverse-transcription poly-
merase chain reaction performed with the use of primers spe-
cific for G1–G4 and G9 (G type) and for P4, P6, P8, and P11
(P type) [10, 11].
We calculated the incidence of hospitalizations
due to rotavirus diarrhea by using the number of hospitalized
children who had rotavirus-positive stool specimens as the nu-
merator and the total number of children in the relevant age
category as the denominator. The value that was calculated was
adjusted for the 65% coverage of hospitalizations anticipated
to be captured by the 6 hospitals participating in the study.
A 20-point clinical severityscorewascalculatedforeachchild
with diarrhea: mild diarrhea was defined by a score of 1–8;
moderate diarrhea, by a score of 9–14; and severe diarrhea, by
a score of 114, as described elsewhere [12, 13]. The proportion
of children with severe diarrhea (severity score, 114) and the
proportion of children with severe dehydration (according to
World Health Organization criteria) at admission to the hos-
pital were compared for the groups of children infected with
different rotavirus strains. Adjustment for age and sex wasdone
in logistic regression models.
The study was reviewed and approved bytheethical
Stool specimens were screened for the
S116 • JID 2005:192 (Suppl 1) • Bahl et al.
specimens obtained from children from the slums of New Delhi, India, who were hospitalized
for severe rotavirus diarrhea.
Distribution of G and P serotypes of rotavirus strains isolated from 137 stool
Mixed UntypeableTotalP PP P
52 (38.0)16 (11.7)29 (21.2) 32 (23.4) 1 (0.7)
aFor 2 children, no stool specimen was available for testing. For 20 of the remaining 30 stool specimens,
rotavirus RNA could not be detected by PAGE.
Table 3. Association of the G serotype of rotavirus strains with severe diarrhea.
No. (%) of children
with very severe
OR (95% CI)
Adjusted for age
aSeverity score (on the 20-point severity scale), 114.
CI, confidence interval; OR, odds ratio.
review committee of the All IndianInstituteofMedicalSciences
in New Delhi, India.
Overall and age-specific incidences of severe rotavirus
There were 62,475 children !5 years of age in the
study population, with approximately equal numbers of chil-
dren in each 1-year age group up to 5 years of age. A total of
584 children from study households whowerehospitalizedwith
diarrhea were identified during 1 calendar year of surveillance
at the 6 study hospitals. Of these 584 children, 137 (23.5%)
had a stool specimen that tested positive for rotavirus, yielding
an incidence of rotavirus diarrhea–associated hospitalizationof
337 hospitalizations/100,000 children !5 years of age. This in-
cidence indicates that ∼1 in 59 children in this population is
likely to be hospitalized with rotavirus diarrhea between birth
and 5 years of age.
Most (98%) of the children who were hospitalized with ro-
tavirus diarrhea were !2 years of age. Theincidenceofrotavirus
diarrhea was 1270, 534, and 12 cases/100,000 children during
the first, second, and third to fifth years of life, respectively.
The incidence was low during the first 3 months of life, peaked
at age 9–11 months, and decreased sharply after age 18 months
None of the hospitalized children with rotavirus diarrhea
had disease of mild severity, 101 (73.7%) had moderatelysevere
disease, and 36 (26.3%) had severe disease. None of the 137
children died during the study period; 133 (97.1%) were dis-
charged from the hospital, and 4 (2.9%) left the hospitalagainst
Seasonality of rotavirus infection.
tavirus-associated diarrhea occur year-round in Delhi, but we
identified a distinct peak in winter (i.e., from Novemberthrough
February) (figure 1). In contrast, the number of hospitalizations
not associatedwith rotaviruspeakedduringthesummermonths.
During the winter months, G1 and G9 strains were the most
common strains identified, accounting for 45% and 29% of
typeable strains, respectively. G1 (42%) and G2 (36%) strains
were the most common strains identified during July–October,
whereas G9 (33%), G1 (27%), and G2 (27%) strains were most
commonly isolated during the summer months (March–June).
be determined for 105 of the 135 stool specimens tested. For
20 of 30 untypeable specimens, RNAwas notdetectedbyPAGE.
Hospitalizations for ro-
Incidence of Severe Rotavirus Diarrhea • JID 2005:192 (Suppl 1) • S117
at admission to the hospital.
Association of the G serotypes of rotavirus strains with severe dehydration
No. (%) of children
OR (95% CI)
Adjusted for age
CI, confidence interval; OR, odds ratio.
Among the strains that could be G-typed, G1 was the most
common type identified, followed by types G9 and G2. Mixed
G-type infections were identified in 10% of thestoolspecimens.
P-typing was used to indicate the diversity of strains in cir-
culation and to identify the frequency of the 4 most common
strains worldwide—namely, P combined with G1, G3, or
G4; PG2; and P or P combined with G9 (table 2).
Only strains PG1, PG2, PG9, and PG9 were com-
monly isolated in the present study. Other strains that are not
common globally but that were frequently identified during
this surveillance study were PG1 and PG3.
Association of G types with clinical severity of diarrhea.
Compared with all other rotavirus-positive children, a higher
proportion of children infected with the G1 strain had severe
diarrhea, as defined by a severity score of 114 (age- and sex-
adjusted odds ratio [OR], 2.95; 95% confidence interval [CI],
1.3–6.67) (table 3). The mean severity score was also higher
for children infected with G1 strains (difference in means, 0.8;
95% CI, 0–1.5). Similarly, the prevalence of severe dehydration
at admission was significantly higher for children infected with
G1 strains than for children infected with other G-type strains
(age- and sex-adjusted OR, 3.03; 95% CI, 1.11–8.26 (table 4).
We did not find any association of P type with clinical severity
We believe the present study to be the first study from a de-
veloping country to report the population-based incidence of
hospitalization for rotavirus diarrhea in a well-defined catch-
ment area. The high incidence of rotavirus diarrhea requiring
patient hospitalization confirms the immense importance of
rotavirus diarrhea with regard to public health. If the estimates
of incidence obtained from the present study were applied to
the total birth cohort of ∼25 million children in India, then
∼450,000 hospitalizations for rotavirus diarrhea would be ex-
pected to occur annually. An effective rotavirus vaccine, in ad-
dition to preventing 100,000 rotavirus disease–associated deaths
annually, as estimated elsewhere , could prevent most of
these hospitalizations and result in large savings in health care
costs. The findings also reiterate the need for effective oral
rehydration therapy for the prevention of the serious conse-
quences of diarrhea and, in particular, rotavirus diarrhea.
In Europe and North America, rotavirus disease is predomi-
nant in winter; however, previous studies from India have re-
ported that the disease occurs year-round [15–19]. Thosestudies
found either a peak in the occurrence of rotavirus disease in
winter [15, 16], the occurrence of 2 peaks in a year [17, 18], or
no seasonal pattern . The findings of the present study con-
firm that the disease occurs year-round in New Delhi and that
diarrhea to nonrotavirus diarrhea during the winter months of
November through February (average maximum daily temper-
ature, 21?C–24?C) was almost 1:1, it was 11:10 for the summer
months fromMarch throughJune(averagemaximumdailytem-
perature, 36?C–41?C). The predominance of rotavirus diarrhea
in winter may have implications for treatment.
In previous hospital-based studies performed in India, the
most common rotavirus strains had G2 and G9 serotypes .
Although we also found these strains to be common, G1 was
the most common type identified in the present study. G1
rotavirus strains are also, epidemiologically, the most common
circulating strains globally . The findings of the present
study may differ from those of previous studies, because we
assessed only hospitalized children, whereas manypreviousstud-
ies also evaluated outpatients. Furthermore, if G1 strains are
associated with greater clinical severity of disease, as indicated
by the current data, then outpatient-based surveillance is likely
to yield a lower prevalence of infections with strains of the G1
type. Nevertheless, because our observations covered only a 1-
year period, some of the differences in serotype distributions
from previous studies may be attributable to year-to-year vari-
ations. In addition, a significant proportion of strains could not
be G- or P-typed, indicating a need to search for otherserotypes.
In southern India, G8 and, recently, G10 strains have also been
implicated as the cause of gastroenteritis in children [21, 22].
Several studies have attempted to associate the type of in-
S118 • JID 2005:192 (Suppl 1) • Bahl et al.
fecting rotavirus strain with the clinical severity of disease [23–
34]: 7 studies found no association [23–29], 1 study found that
strains in subgroup 2 (G1, G3, and G4 strains) were associated
with greater clinical severity , 1 study reported that G1 and
G3 strains were associated with more-prolonged diarrheal ill-
ness , 2 studies found that G2 strains were associated with
greater clinical severity [32, 33], and 1 study reported that the
P type, rather than the G type, was associated with clinical se-
verity . Our findings indicate that G1 strains cause great-
er clinical severity, both in terms of the overall severity score
and the associated development of severe dehydration. If true,
this finding may be potentially important for the development
of rotavirus vaccines. Candidate vaccines must be shown to
be effective against G1 strains, in addition to other rotavirus
strains, to have a maximal effect on the prevention of deaths
and hospitalizations due to rotavirus diarrhea. Clinical severity
may be influenced not only by viral factors but, also, by such
host factors as nutritional and immune status. However, there
was no association of the infecting rotavirus strain with any of
the host factors assessed.
Some limitations of the present study require consideration.
Surveillance was conducted only for 1 year, and, therefore, we
cannot be certain of the year-to-year variation in strain types.
Also, the study population included residents of the slums of
New Delhi, and the rate of migration in and out of the slums
could be up to 10%–15% during the year. Finally, although we
used a corrective factor of 65% to estimate incidence on the
basis of our survey results, this factor may not have been totally
In conclusion, the findings of the present study imply that
the development of a safe and effective rotavirus vaccine that
is efficacious against multiple serotypes is a public health pri-
ority in India. These data should facilitate trials of candidate
vaccines and should assist in the estimation of the cost-effec-
tiveness of such vaccines.
DELHI ROTAVIRUS STUDY GROUP MEMBERS
The members (hospital affiliation in New Delhi, India) of the
Delhi Rotavirus Study Group are M. Saxena and S. Bhatnagar
(All India Institute of Medical Sciences); N. Gupta, S. Paul, G.
Srivastav, S. Saraf, A. S. Bakshi, and S. Bagai (Batra Hospital);
S. Warsi (Holy Family Hospital); S. Shakdhar(JeevanHospital);
B. Oberoi, P. Dewan, and S. K. Sethi (Majeedia Hospital); and
M. S. Prasad and I. P. Chaudhary (Safdarjang Hospital).
1. De Zoysa I, Feachem RG. Interventions for the control of diarrhoeal
disease among young children: rotavirus and cholera immunization.
Bull World Health Organ 1985;63:569–83.
2. Jain V, Parashar UD, Glass RI, Bhan MK. Epidemiology of rotavirus
in India. Indian J Pediatr 2001;68:855–62.
3. Jain V, Das BK, Bhan MK, Glass RI, Gentsch JR. Indian Strain Sur-
veillance Collaborating Laboratories. Great diversity of group A ro-
tavirus strains and high prevalence of mixed rotavirus infections in In-
dia. J Clin Microbiol 2001;39:3524–9.
4. Das S, Sen A, Uma G, et al. Genomic diversity of group A rotavirus
strains infecting humans in eastern India. J Clin Microbiol 2002;40:
5. Landaeta ME, Dove W, Vinh H, et al. Characterization of rotaviruses
causing diarrhoea in Vietnamese children. Ann Trop Med Parasitol
6. Leite JP, Alfieri AA, Woods PA, Glass RI, Gentsch JR. Rotavirus G and
P types circulating in Brazil: characterization by RT-PCR, probe hy-
bridization, and sequence analysis. Arch Virol 1996;141:2365–74.
7. Ramachandran M, Das BK, Vij A, et al. Unusual diversity of human
rotavirus G and P genotypes in India. J Clin Microbiol 1996;34:436–9.
8. Armah GE, Pager CT, Asmah RH, et al. Prevalence of unusual human
rotavirus strains in Ghanaian children. J Med Virol 2001;63:67–71.
9. World Health Organization/Child Health and Development Division
Immunisation-Linked Vitamin A Supplementation Study Group. Ran-
domised trial toassessbenefitsandsafetyofvitaminAsupplementation
linked to immunisation in early infancy. Lancet 1998;352:1257–63.
10. Gentsch JR, Glass RI, Woods P, et al. Identification of groupArotavirus
gene 4 types by polymerase chain reaction. J Clin Microbiol 1992;30:
11. Das BK, Gentsch JR, Cicirello HG, et al. Characterization of rotavirus
strains from newborns in New Delhi, India. J Clin Microbiol 1994;
12. Ruuska T, Vesikari T. Rotavirus disease in Finnish children: use of
numerical scores for clinical severity of diarrhoeal episodes. Scand
J Infect Dis 1990;22:259–67.
13. Perez-Schael I, Guntinas MJ, Perez M, et al. Efficacy of the rhesus
rotavirus-based quadrivalent vaccine in infants and young children in
Venezuela. N Engl J Med 1997;337:1181–209.
14. Parashar UD, Hummelman EG, Bresee JS, Miller MA, Glass RI. Global
illness and deaths caused by rotavirus disease in children. Emerg Infect
15. Kelkar SD, Purohit SG, Simha KV. Prevalence of rotavirus diarrhoea
among hospitalized children in Pune, India. Indian J Med Res 1999;
16. Chakravarti A, Kumar S, Mittal SK, Broor S. Clinical and epidemio-
logical features of acute gastroenteritis caused by human rotavirussub-
groups. J Diarrhoeal Dis Res 1992;10:21–4.
17. Singh V, Broor S, Mehta S, Mehta SK. Clinical and epidemiological
features of acute gastroenteritis associated with human rotavirus sub-
groups 1 and 2 in northern India. Indian J Gastroenterol 1989;8:23–5.
18. Brown DW, Mathan MM, Mathew M, Martin R, Beards GM, Mathan
VI. Rotavirus epidemiology in Vellore, south India: group, subgroup,
serotype, and electrophoretype. J Clin Microbiol 1988;26:2410–4.
19. Yachha SK, Singh V, Kanwar SS, Mehta S. Epidemiology, subgroups
and serotypes of rotavirus diarrhea in north Indian communities. In-
dian Pediatr 1994;31:27–33.
20. Gentsch JR, Woods PA, Ramachandran M, et al. Review of G and P
typing results from a global collection of strains: implications for vac-
cine development. J Infect Dis 1996;174(Suppl 1):S30–6.
21. Kang G, Green J, Gallimore CI, Brown DW. Molecular epidemiology
of rotaviral infection in south Indian children with acute diarrheafrom
1995–1996 to 1998–1999. J Med Virol 2002;67:101–5.
22. Iturriza Gomara M, Kang G, Mammen A, et al. Characterization of
G10P rotaviruses causing acute gastroenteritis in neonates and in-
fants in Vellore, India. J Clin Microbiol 2004;42:2541–7.
23. Barnes GL, Unicomb L, Bishop RF. Severity of rotavirus infection in
relation to serotype, monotype and electropherotype. J Paediatr Child
24. Raul Velazquez F, Calva JJ, Lourdes Guerrero M, et al. Cohort study
Incidence of Severe Rotavirus Diarrhea • JID 2005:192 (Suppl 1) • S119 Download full-text
of rotavirus serotype patterns in symptomatic and asymptomatic in-
fections in Mexican children. Pediatr Infect Dis J 1993;12:54–61.
25. Armah GE, Hori H, Anyanful A, et al. Human rotavirus subgroups
and severity of associated diarrhoea in Ghana. Afr J Health Sci 1995;2:
26. Polanco-Marin G, Gonzalez-Losa Mdel R, Rodriguez-Angulo E, Man-
zano-Cabrera L, Camara-Mejia J, Puerto-Solis M. Clinical manifes-
tations of the rotavirus infection and his relation with the electro-
pherotypes and serotypes detected during 1998 and 1999 in Merida,
Yucatan, Mexico. J Clin Virol 2003;27:242–6.
27. Pitson GA, Grimwood K, Coulson BS, et al. Comparison between
children treated at home and those requiring hospital admission for
rotavirus and other enteric pathogens associated with acute diarrhea
in Melbourne, Australia. J Clin Microbiol 1986;24:395–9.
28. Georges-Courbot MC, Beraud AM, Beards GM, et al. Subgroups, se-
rotypes, and electrophoretypes of rotavirus isolated from children in
Bangui, Central African Republic. J Clin Microbiol 1988;26:668–71.
29. Steele AD, Bos P, Alexander JJ. Clinical features of acute infantile
gastroenteritis associated with human rotavirus subgroups I and II. J
Clin Microbiol 1988;26:2647–9.
30. Uhnoo I, Svensson L. Clinical and epidemiological features of acute
infantile gastroenteritis associated with human rotavirus subgroups 1
and 2. J Clin Microbiol 1986;23:551–5.
31. Flores J, Taniguchi K, Green K, et al. Relative frequencies of rotavirus
serotypes 1, 2, 3, and 4 in Venezuelan infants with gastroenteritis. J
Clin Microbiol 1988;26:2092–5.
32. Bern C, Unicomb L, Gentsch JR, et al. Rotavirus diarrhea in Bangla-
deshi children: correlation of disease severity with serotypes. J Clin
33. Cascio A, Vizzi E, AlaimoC, AristaS. RotavirusgastroenteritisinItalian
children: can severity of symptoms be related to the infecting virus?
Clin Infect Dis 2001;32:1126–32.
34. Mota-Hernandez F, Calva JJ, Gutierrez-Camacho C, et al. Rotavirus
diarrhea severity is related to the VP4 type in Mexican children. J Clin