Abstract

The disease burden of rotavirus diarrhea in Vietnam was assessed by surveillance of children <5 years old who were hospitalized for diarrhea at 3 centers in the north and 3 centers in the south. Rotavirus was identified in 56% (range, 47%–60%) of the 5768 patients surveyed between July 1998 and June 2000. G-typing of the first 224 strains indicated that only 2% were nontypeable, 9% were in mixed infections, and the remainder were of the common serotypes G1, G2, G3, G4, and G9. In Vietnam, diarrhea accounts for 9880 deaths per year, which is ∼15% of all deaths among children <5 years old, or 6.5 deaths per 1000 children. If even 50% of these diarrhea-related deaths in Vietnam were due to rotavirus, the number would represent 4%–8% of all deaths among children <5 years old, 2700–5400 rotavirus-related deaths per year, and 1 death per 280–560 children during the first 5 years of life. Thus, the disease burden of rotavirus in Vietnam is substantial, and programs to encourage the use of oral rehydration should be encouraged while efforts to develop vaccines continue

Rotavirus is the most common cause of severe diarrhea in children worldwide, and vaccines that are being developed and tested potentially could prevent much of the morbidity and mortality due to this disease. The greatest burden of disease falls on children in developing countries, where rotavirus accounts for an estimated 600,000–800,000 deaths per year or 5% of the deaths among children <5 years old [1]. Despite the global recognition of this problem, many countries are unaware or uncertain of the importance of rotavirus diarrhea in their own setting. Consequently, since 1997, 3 international workshops that addressed the potential use of rotavirus vaccines for developing countries have each recommended that countries that might consider using rotavirus vaccines in the future should begin surveillance for the disease today [2, 3]. This surveillance could identify peculiarities and important features in the local epidemiology of the disease, yield baseline estimates of the burden of severe rotavirus diarrhea, create an awareness of rotavirus disease among pediatricians and policy makers, and provide experience on the most efficient ways to monitor the impact of a rotavirus vaccine program in the future should this be seen as a priority

Several differences in the epidemiology of rotavirus between tropical developing countries and developed countries in temperate climates could have direct major implications on future strategies to administer rotavirus vaccine. First, in tropical countries, rotavirus occurs year round, but winter peaks and summer troughs of infection are seen in developed countries [4]. It has been suggested that this difference in seasonality could explain the younger median age of illness among children in developing countries, compared with that among children in developed countries (6–8 vs. 14–18 months, respectively) [1]. Children in industrialized countries who are born at the end of the winter peak of rotavirus infection must wait a full year until they are next exposed to rotavirus, whereas children in the tropics can be exposed any day of the year. A practical consequence of the younger age of severe disease is that infants in developing countries may have to be immunized earlier, perhaps when they receive bacille Calmette-Guérin vaccine, to get the full benefit of the vaccine

Vietnam is a developing country of 76 million people with a large birth cohort (1.52 million births/year), high mortality among children <5 years old (42.2 deaths/1000 per year, of which 15.4% are related to diarrheal disease [5, 6]), and a per capita income of $374. The country stretches >1400 km from the tropical areas in the south to the subtropics in the north

We initiated simple hospital-based rotavirus surveillance in South and North Vietnam, to assess the burden of rotavirus disease and to determine whether differences in the seasonality of disease might be reflected in the age distribution of the illness. A subset of rotavirus strains was genotyped to identify common strains and unusual variants. Last, we estimated the mortality associated with rotavirus diarrhea by linking the mean detection rate of rotavirus for hospitalized children with national data on diarrheal disease–related mortality among children <5 years old

Materials and Methods

Purpose and surveillance sitesSurveillance was conducted to assess the disease burden of rotavirus in Vietnam and to determine whether there are any regional epidemiologic differences in the disease that might reflect the major climatic differences between North and South Vietnam. The north has a subtropical climate with 4 seasons and a moderately cold winter during December through February; the south has a tropical climate with a rainy season from September to December and a dry season for the rest of the year. A common surveillance protocol was implemented at 6 major pediatric hospitals in 4 principal cities of Vietnam: in the north, St. Paul's Hospital and the National Children's Hospital in Hanoi and the Children's Hospital in Haiphong, and in the south, the General Hospital in Khanh Hoa (Nha Trang) and both the General Children's Hospitals No. 1 and No. 2 in Ho Chi Minh City (figure 1). Surveillance was initiated as a pilot study in the north in February 1998, and activity began at all sites in July 1998. Here, we report data from the first 2 years, July 1998 through June 2000, although the surveillance activity is continuing

Figure 1

Map of Vietnam, indicating location of surveillance sites in the north (Hanoi and Haiphong) and south (Ho Chi Minh City and Nha Trang [Khan Hoa])

Figure 1

Map of Vietnam, indicating location of surveillance sites in the north (Hanoi and Haiphong) and south (Ho Chi Minh City and Nha Trang [Khan Hoa])

Surveillance population and study protocolAll children <5 years old with diarrhea admitted to one of the surveillance hospitals were enrolled in the study. For each patient, demographic data and information on clinical presentation was recorded on a standardized questionnaire completed by a trained health worker specifically assigned to the project. Symptoms of gastroenteritis (vomiting, diarrhea, fever, and dehydration) were recorded individually and were coded on a 20-point scale of severity used in previous rotavirus vaccine trials [7]. Dehydration was assessed for severity and was categorized as none, mild to moderate (3%–9%), or severe (⩾10%), according to World Health Organization criteria [8]. Fecal specimens were obtained from each patient within 24 h of admission, were frozen at −20°C, and were transported frozen for testing at the Biochemistry and Enterovirus Laboratories of the National Institute of Hygiene and Epidemiology in Hanoi and the Enterovirus Laboratory of the Pasteur Institute in Ho Chi Minh City. In Ho Chi Minh City, surveillance was discontinued for 1 month, December 1998, for administrative reasons, and in Hanoi, no specimens were collected in August 1998 and April 1999 because of a change in staff

Laboratory studiesFecal specimens were screened for rotavirus by use of a monoclonal antibody–based EIA kit (Premier Rotaclone; Meridian Diagnostics), which was used according to the manufacturer's instructions. The hospitals that are involved do not routinely screen for any enteric pathogens, and no other agents were screened for in this study

To examine the diversity of rotavirus strains circulating in Vietnam, a collection of rotavirus-positive specimens, ⩽5 per site each month, were selected for characterization of the G and P outer capsid proteins. Genotyping was done by reverse-transcription polymerase chain reaction [9,11] on RNA extracted from Freon-treated fecal suspensions (30%), using the glass powder method of Boom et al., as modified by Gentsch et al. [12]. An attempt was made to G-type all strains, and a subset of these were P-typed. Strains are referred to by the G and P designation of their 2 outer capsid proteins

AnalysisFor analysis, we compiled monthly data on detection rates of rotavirus at each site. Differences in the distribution of symptoms among patients with diarrhea caused by rotavirus versus other pathogens were compared and were tested for significance by use of a χ2 test. We anticipated that differences in climate between North and South might alter the age distribution of cases. Therefore, we compared the cumulative frequency distribution of the ages of patients in each region

Results

Incidence of rotavirusIn total, 5768 fecal specimens were collected from July 1998 through June 2000 from the 6 sentinel hospitals and were screened for rotavirus (table 1). Of these, 3247 (56%) were determined to be rotavirus positive by EIA and detection rates did not vary greatly by sites (range, 47%–60%) or by geographic location (north vs. south). Nearly twice as many boys (64%) as girls (36%) were admitted to the hospital for rotavirus diarrhea. This ratio reflected the same male:female ratio seen for all diarrheal admissions (data not shown)

Table 1

Detection rates of rotavirus among children <5 years old who were hospitalized for diarrhea in 6 sentinel hospitals in Vietnam, July 1998 to June 2000

Table 1

Detection rates of rotavirus among children <5 years old who were hospitalized for diarrhea in 6 sentinel hospitals in Vietnam, July 1998 to June 2000

Symptoms of patients with diarrheal diseaseThe presenting symptoms of children hospitalized for diarrhea were analyzed for the subset of 2473 children for whom full information was available. Data were excluded if all fields on the questionnaire were not completed. With one exception, the symptoms of patients with and without rotavirus did not differ (table 2). Vomiting, however, was significantly more common among patients with rotavirus diarrhea than among those with diarrhea due to other causes (68.2% vs. 58.3%); this difference was reflected in a marginally higher mean severity score (13.5 vs. 12.9, P<.01) [7] but not in a difference in the severity of dehydration [8]

Table 2

Sentinel hospital surveillance in Vietnam: comparison of symptoms of a subset of 2473 patients <5 years old hospitalized for diarrhea with and without rotavirus

Table 2

Sentinel hospital surveillance in Vietnam: comparison of symptoms of a subset of 2473 patients <5 years old hospitalized for diarrhea with and without rotavirus

Seasonality of rotavirus infectionThe seasonality of rotavirus diarrhea suggests a modest difference between children hospitalized in the north and those hospitalized in the south (figure 2). At 2 sites with full data in the north, rotavirus was detected at rates above the annual median for 5 consecutive months in winter or spring and below the median in the summer and fall. This pattern was not present in the south, where rotavirus occurred year round with less seasonal variation. Of note, rotavirus was detected in all sites every month of screening and was never absent

Figure 2

Monthly detection of rotavirus among children with diarrhea admitted to 3 sentinel hospitals in North Vietnam (Hanoi and Haiphong) and 3 in South Vietnam (Khan Hoa and Ho Chi Minh City) from July 1998 through June 1999. Solid line indicates the median detection rate at each site. Left panel Hospitals in the north; right panel hospitals in the south

Figure 2

Monthly detection of rotavirus among children with diarrhea admitted to 3 sentinel hospitals in North Vietnam (Hanoi and Haiphong) and 3 in South Vietnam (Khan Hoa and Ho Chi Minh City) from July 1998 through June 1999. Solid line indicates the median detection rate at each site. Left panel Hospitals in the north; right panel hospitals in the south

We hypothesized that the age distribution of children with rotavirus might be younger in the south, where rotavirus is in constant circulation with little seasonal variation, than in the north, where there are more seasonal differences (figure 3). The seasonality of rotavirus differed somewhat between North and South Vietnam, so we tested this hypothesis by plotting the cumulative frequency distribution of the ages of children hospitalized in each setting. The age distribution of children was the same in both areas, but the curves provided several additional insights. First, few (12%–14%) of the surveyed children developed rotavirus diarrhea in the first 6 months of life, and 56%–58% were hospitalized after their first birthday

Figure 3

Comparison of the age distribution of patients with rotavirus diarrhea in North and South Vietnam

Figure 3

Comparison of the age distribution of patients with rotavirus diarrhea in North and South Vietnam

Strain characterizationA sample of 224 rotavirus strains collected from the 6 locations during the first year of surveillance were selected for G-genotyping (table 3). An effort was made to test 5 specimens per month from each hospital, but this procedure was postponed during the second year because of changes in personnel. Ninety-eight percent of strains could be G-typed; among these, G2 was the most common (53%) and was followed by types G4 (17%), G1 (13%), and G3 (3.6%). Mixed infections were identified in 9% of the specimens. The distribution of strains varied markedly by location, and the most common strain, G2, represented 81% of the rotaviruses circulating in Ho Chi Minh City but only 27% of those in Khanh Hoa, where G1 was the predominant type (38%) in single infections. G4 was the second most common strain in Hanoi (28%) but was less common in the southern cities of Khanh Hoa (7.3%) and Ho Chi Minh City (2.4%)

Table 3

Distribution of rotavirus G genotypes by regions of Vietnam, July 1998 to June 1999

Table 3

Distribution of rotavirus G genotypes by regions of Vietnam, July 1998 to June 1999

P-typing was used to indicate the diversity of strains in circulation and to identify the frequency of the 4 most common strains worldwide, namely, P[8] combined with G1, G3, G4, and P[4],G2 (table 4). Only 3 of the 5 common strains were identified (P[8],G9 was present as a mixed infection), representing 74% of the strains in circulation. In addition, 2 strains with P[6] were present, which were combined with G4 (n=12) and G3 (n=1). Fourteen of the 140 specimens tested could not be P-typed, and a mixed infection was present in 7 samples

Table 4

Distribution of P[genotypes] of 140 strains from Vietnamese children <5 years old

Table 4

Distribution of P[genotypes] of 140 strains from Vietnamese children <5 years old

Estimates of rotavirus mortalityInformation on mortality in children <5 years old and on diarrheal mortality was obtained from published national sources [4, 5] for 1992. In 1992, mortality among children <5 years old was 42.2 per 1000 per year, of which 15.4% deaths (6.5/1000 per year) were attributed to diarrheal diseases. In the absence of diagnostic information on the etiology of diarrhea at the time of death or on other contributing causes of death associated with diarrhea (e.g., malnutrition), we estimated that the fraction of diarrheal deaths attributable to rotavirus would be about the same as the percentage of cases of rotavirus among children hospitalized for severe diarrhea, for whom a diagnosis of rotavirus was specifically sought. Rotavirus was present in a mean of 56% of hospitalized patients, which, when extrapolated to deaths, would lead to an estimate of 3.6 rotavirus-associated deaths per 1000 children per year or ∼5472 deaths per year for a birth cohort of 1.52 million children in 1998 and 1999. Even if rotavirus were the true etiology in only half of these deaths, the figures are substantial and would suggest that rotavirus is associated with 4.3%–8.5% of reported deaths in Vietnam in children <5 years old (15.4% of deaths are due to nonrotavirus-related diarrhea, and 56% are due to rotavirus). In other words, a Vietnamese child has 1 chance in ∼300 of dying from rotavirus before the age of 5 years, which translates to 3.6 rotavirus deaths per 1000 children per year)

Discussion

Our data clearly demonstrate that the burden of rotavirus diarrhea in Vietnamese children is substantial. Rotavirus was detected in 56% of all children <5 years old who were hospitalized for severe diarrhea, and there were relatively small differences among the 6 sentinel hospitals. Although we do not have the specific etiologic cause of death for diarrhea, national data indicate that ∼15.4% of all deaths, or 6.5 deaths/1000 per year, among children <5 years old are due to diarrhea, and rotavirus must make a substantial contribution. If rotavirus were associated with one-quarter to one-half of these diarrheal deaths, we would estimate that, in Vietnam, 4%–8% of all deaths each year among children <5 years old would be due to rotavirus diarrhea and that ∼1 Vietnamese child in 300 would die of rotavirus diarrhea before the age of 5 years

Several unanticipated epidemiologic features of rotavirus in Vietnam were identified in this study. First, we identified differences in the seasonality of rotavirus between the subtropical north, where the disease was seasonal, and the tropical south, where rotavirus occurred year round with a less distinct seasonal pattern. Nonetheless, although we hypothesized that these seasonal differences might lead to the younger age of case patients in the south versus the north, no such differences were observed. Furthermore, the age distribution of hospitalized cases in Vietnam was intermediate between the patterns seen in India [13] and the United States [14], where 80% and 40% of rotavirus hospitalizations, respectively, occur in the first year of life. Consequently, in Vietnam, vaccine delivered on the current childhood immunization schedule at 6, 10, and 14 weeks of age could effectively prevent most cases of severe disease. Last, the genotype diversity of rotavirus strains in Vietnam is not nearly as great as that seen in other developing countries in Asia, such as India and Bangladesh [15, 16]. All but 3 of the 135 specimens that were successfully G-typed contained strains that belonged to 1 of the 5 common G types. The G and P types of only 3% and 10%, respectively, of the specimens could not be determined, and 14% of patients had a mixed infection. These results are similar to those of Nishio et al. [17], which were based on specimens obtained in Vietnam several years earlier

Several limitations of this study need to be considered further. First, we did not examine patients for other enteric pathogens, an effort that would have involved a much greater budget and more laboratory support. Nonetheless, a series of etiologic studies conducted in recent years in Hanoi, in which >1000 hospitalized children <5 years old were tested, provide results consistent with our own (table 4). In these investigations, Shigella species (14.5%), enterotoxigenic Escherichia coli (10%), and Campylobacter species (9%) were the most common bacterial agents, which were followed by cryptosporidium (2.4%) and salmonella (<1%); >60% of patients had no agent identified [18,20]. Earlier studies of rotavirus in Vietnam had lower detection rates, but each was performed in a single location with smaller samples, and some addressed disease in the community, which is of lesser severity than that of hospitalized patients, and rotavirus was less frequently identified in these milder cases

High rates might represent overinterpretation of assays by enthusiastic laboratory workers, but high rates and little variability were detected from all sites. Quality control assessments will be introduced into this study in the future to examine differences in diagnostic capability by site. A greater limitation is our inability to use national statistics to identify the true mortality from diarrheal deaths and the fraction of these deaths that are due specifically to rotavirus. National data for Vietnam are at least consistent with those of other countries with a similar per capita income. Extrapolation of rotavirus detection data from hospitalizations to mortality is more problematic, since children who die are likely to have other contributing causes, such as malnutrition or other infections that might make them particularly debilitated and put them at greater risk. However, even assuming that only half the diarrhea deaths were directly associated with rotavirus leads to estimates that are substantial, which would warrant intervention

This study provides several insights into addressing the disease burden of rotavirus and considering strategies for disease control and prevention. Sentinel hospital surveillance was easy to initiate and already has provided important national data concerning the patterns of disease in diverse geographic regions, features such as the age distribution of cases and the diversity of strains in circulation. More attention needs to be directed to understanding the causes of diarrheal deaths, since this is the key information required to move policy makers to consider rotavirus-specific interventions

Although it might be argued that all cases of rotavirus might be treated with oral rehydration therapy, a diarrheal disease program in place for more than a decade has not prevented the large number of diarrhea-related deaths that continues to occur nor decreased the number of children admitted to hospital for diarrhea. Furthermore, this study identified vomiting as a key clinical feature occurring in a majority of patients with rotavirus, and vomiting has been identified in the past as a common cause of failure of oral rehydration therapy. Improvements in sanitation and in food and water quality are unlikely to alter the incidence of rotavirus diarrhea in small children, so we are left with vaccines as the most likely intervention to prevent disease. This study indicates that, when the next generation of vaccines becomes available, such vaccination administered on the routine schedule for childhood immunizations could have a major impact on the incidence of severe diarrhea diseases in Vietnam and the improvement in child health

Vietnam Rotavirus Surveillance Network

Nguyen Gia Khanh (National Children's Hospital, Hanoi); Nguyen Thi Xuan Hoi (Saint Paul's Hospital, Hanoi); Nguyen Van Hieu and Pham Quoc Vinh (Center for Preventive Medicine, Haiphong); Nguyen Tuan Tu (Haiphong Children's Hospital, Haiphong); Nguyen Manh Tien (General Hospital, Khanh Hoa); Dinh Thi Ngoc Tuyet (Pasteur Institute, Nha Trang); Phan Ngoc Thanh (General Children's Hospital, No. 1, Ho Chi Minh City); and Nguyen Thi Tien (General Children's Hospital No. 2, Ho Chi Minh City)

Acknowledgments

We thank Phung Dac Cam et al. (Division of Enteric Pathogens, Department of Bacteriology and Immunology, National Institute of Hygiene and Epidemiology [NIHE]) for scientific information and their assistance with laboratory equipment and reagents; John O'Connor (Centers for Disease Control and Prevention) for editorial assistance; Nguyen Hanh Phuc (Department of Virology, NIHE) for scientific advice; and Gary West and Dwayne Gubler for administrative support

References

1
Bresee
J
Glass
RI
Ivanoff
B
Gentsch
J
Current status and future priorities for rotavirus vaccine development, evaluation, and implementation in developing countries
Vaccine
 , 
1999
, vol. 
17
 (pg. 
2207
-
22
)
2
World Health Organization
Vaccine research and development: rotavirus vaccines for developing countries
Wkly Epidemiol Rec
 , 
1997
, vol. 
72
 pg. 
35
 
3
World Health Organization
Report of the meeting on future directions for rotavirus vaccine research in developing countries
2000
Geneva
WHO
 
WHO/V&B/00.23
4
Cook
SM
Glass
RI
LeBaron
CW
Ho
MS
Global seasonality of rotavirus infections
Bull World Health Organ
 , 
1990
, vol. 
68
 (pg. 
171
-
7
)
5
Canh
DG
Do
TT
General and diarrhea related mortality rates in Vietnamese children
J Preventive Med, Vietnam
 , 
1993
, vol. 
3
 
(Suppl 12)
pg. 
48
 
6
Canh
DG
Do
TT
Tuan
NA
, et al.  . 
A study of mortality rates due to diarrhea and other causes using Brass and PBT methods
J Preventive Med, Vietnam
 , 
1993
, vol. 
3
 
(Suppl 12)
pg. 
25
 
7
Rennels
MB
Glass
RI
Dennehy
PH
Bernstein
DI
Pichichero
ME
Safety and efficacy of high-dose rhesus-human reassortant rotavirus vaccines: report of the national multicenter trial
Pediatrics
 , 
1996
, vol. 
97
 (pg. 
7
-
13
)
8
World Health Organization (WHO)
A manual for the treatment of diarrhoea
1990
Geneva
WHO
 
WHO/CDD/SER/80.2 Rev 2
9
Gouvea
V
Glass
RI
Woods
P
, et al.  . 
Polymerase chain reaction amplification and typing of rotavirus nucleic acid from stool specimens
J Clin Microbiol
 , 
1990
, vol. 
28
 (pg. 
276
-
82
)
10
Cicirello
HG
, et al.  . 
Characterization of rotavirus strains from newborns in New Delhi, India
J Clin Microbiol
 , 
1994
, vol. 
32
 (pg. 
1820
-
2
)
11
Glass
RI
Woods
P
, et al.  . 
Identification of group A rotavirus gene 4 types by polymerase chain reaction
J Clin Microbiol
 , 
1992
, vol. 
30
 (pg. 
1365
-
73
)
12
Woods
PA
Ramachandran
M
, et al.  . 
Review of G and P typing results from a global collection of strains: implications for vaccine development
J Infect Dis
 , 
1996
, vol. 
174
 
(Suppl 1)
pg. 
S30
 
13
Bhan
MK
Raj
P
Bhandari
N
, et al.  . 
Role of enteric adenoviruses and rotaviruses in mild and severe acute enteritis
Pediatr Infect Dis J
 , 
1988
, vol. 
7
 (pg. 
320
-
3
)
14
Parashar
UD
Holman
RC
Clarke
MJ
Bresee
JS
Glass
RI
Hospitalizations associated with rotavirus diarrhea in the United States, 1993 through 1995: surveillance based on the new ICD-9–CM rotavirus-specific diagnostic code
J Infect Dis
 , 
1997
, vol. 
178
 (pg. 
13
-
7
)
15
Ramachandran
M
Das
BK
Vij
A
, et al.  . 
Unusual diversity of human rotavirus G and P genotypes in India
J Clin Microbiol
 , 
1996
, vol. 
34
 (pg. 
436
-
9
)
16
et al.  . 
Evidence of high-frequency genomic reassortment of group A rotavirus strains in Bangladesh: emergence of type G9 in 1995
J Clin Microbiol
 , 
1999
, vol. 
37
 (pg. 
1885
-
91
)
17
Nishio
O
Matsui
K
Lan
DTP
Ushijima
H
Isomura
S
Rotavirus infection among infants with diarrhea in Vietnam
Pediatr Int
 , 
2000
, vol. 
42
 (pg. 
422
-
4
)
18
Nguyen
TT
Chinh
LH
Cam
PD
Characteristics of faecal specimens in shigellosis
Practical Med
 , 
1988
Vietnam
Ministry of Health
(pg. 
47
-
8
)
19
Van Hoa
N
Cam
PD
Thuy Long
H
Some epidemiological and clinical features of acute diarrhea caused by campylobacter in a rural commune in Hanoi
J Preventive Med, Vietnam
 , 
1999
, vol. 
IX
 
no. 2
 
(40)
20
Dan
LC
Cam
PD
Long
HT
Cryptosporidium as the cause of diarrhea among children in Hanoi
Practical Med
 , 
1999
Vietnam
Ministry of Health

Author notes

a
Members of the Vietnam Rotavirus Surveillance Network are listed at the end of the text.
b
Financial support: World Health Organization Global Programme on Vaccines; Centers for Disease Control and Prevention
The study protocol was approved by the Council of Science and Technology, Ministry of Health, and the protocol for the handling of specimens was reviewed and approved by the National Institute of Hygiene and Epidemiology and the Council of Science and Technology, Ministry of Health, Vietnam