Worldwide, diarrhea represents a major health problem and a common cause of morbidity and mortality in children. Rotavirus constitutes the single most important cause of acute dehydrating viral gastroenteritis. However, other viral agents such as adenovirus, calicivirus, and astrovirus are also associated with gastroenteritis in childhood. Some studies have found that astrovirus is the second most common cause of viral gastroenteritis in infants and young children (1–4) while others have stressed the importance of adenovirus (5). Specific adenovirus serotypes, 40 and 41, have been associated with acute gastrointestinal disease in children while other serotypes have a less clear relationship to clinical viral gastroenteritis (6).
Previous studies have shown that viral agents such as rotavirus, astrovirus, and adenovirus account for a large portion of both community acquired and nosocomial gastroenteritis in hospitalized pediatric patients in developed countries (1,4,7–10). However, these hospital-based studies have not systemically examined the degree of asymptomatic shedding of these agents to determine their true association with disease and possible role of asymptomatic shedding in transmission.
The development of enzyme immunoassays for agents of viral disease has enabled rapid detection of the virus and identification of infection. Awareness of the importance of astroviruses in pediatric gastroenteritis has increased in recent years with advent of sensitive and specific diagnostic tools such as ELISA tests and RT-PCR assays. Previous surveys using mainly electron microscopy as a diagnostic tool estimated much lower prevalence rates of astrovirus (11).
The purpose of this study was to determine the relative prevalence of astrovirus, rotavirus, and adenovirus among hospitalized children less than 5 years of age and to examine the association of infection with symptomatic gastroenteritis to assess the burden of gastrointestinal illness due to each agent.
MATERIALS AND METHODS
From July 1998 to June 2000, fecal samples were collected twice weekly from children less than 5 years of age hospitalized in Lucile Salter Packard Children's Hospital at Stanford, California, U.S.A. Lucile Salter Packard Children's Hospital serves as both a regional tertiary care center and a local primary care hospital. Samples were collected on two wards. One ward is designated as a general pediatric ward while the other is mainly used for pre-and postoperative pediatric solid organ transplant patients. Both wards admit a broad age range of patients from newborn to older adolescents.
Stool samples were requested from all patients less than 5 years of age on both wards twice weekly. Fecal specimens were stored at -70°C until examination. All fecal samples were examined for astrovirus, rotavirus and adenovirus using enzyme immunoassay tests. Investigators were blinded to clinical data during performance of ELISA tests. Rotavirus was detected using a commercial ELISA kit (Rotaclone, Meridian Diagnostics). Adenovirus was detected using commercial ELISA kit (Adenoclone, Meridian Diagnostics) with subsequent determination of serotypes 40 and 41 by Adenoclone 40,41 (Meridian Diagnostics). Positive and negative controls included in each kit were used. Astrovirus antigen was detected by a commercial ELISA (IDEIA™, DAKO Diagnostics). We found that results of the commercial assay for astrovirus were in complete concordance with our previously developed ELISA (1) when tested against a panel of 20 known positive and 20 known negative stools.
Between July 1997 and June 2000, a separate passive surveillance program was performed for inpatients and outpatients at Packard Children's Hospital. An aliquot of each stool submitted for microbiologic analysis was saved and tested as described above by ELISA for rotavirus and astrovirus.
Astroviruses were serotyped by a previously described neutralization assay after culture on Caco 2 cells (1).
Clinical data were obtained by review of medical records. All studies were approved by the Stanford University Institutional Review Board.
Gastroenteritis was defined as greater than three stools of looser than usual consistency per 24 hours and/or vomiting more than twice daily above patient's baseline as determined by parental histories and nursing observations. Nosocomial infections were defined as occurring when either serial stool samples changed from negative to positive for viral antigen during hospitalization or when symptoms developed greater than 72 hours after admission.
Statistical analysis was performed using the Minitab program (Addison Wesley Publishing, Reading, MA).
A total of 480 specimens were collected from 309 hospitalized pediatric subjects (age range: 1 month–60 months). This represented a sampling of 34% of the eligible patients who were admitted during this time period. Twenty-two percent of the patients were immunocompromised. Twenty-one percent of the subjects (62 children) had symptoms (vomiting and/or diarrhea) attributable to gastroenteritis at some point during their hospitalization (Table 1). Of the children with gastroenteritis symptoms, 57% (35 subjects) had symptoms on admission but only half of them (17 subjects) were admitted primarily because of their gastroenteritis symptoms. Five percent (17 patients) of all admissions in the studied group were primarily because of dehydration secondary to gastroenteritis. Forty three percent of the symptomatic children developed symptoms more than 72 hours after admission.
Seventeen percent of the subjects had viral antigen detected in the stools at some point. One child had the simultaneous presence of astrovirus and rotavirus. One child had both astrovirus and a nonenteric adenovirus.
Astrovirus was detected in 5.5% (16 of 309) of the hospitalized children. Most (75%) of the infected children were symptomatic and the association of astrovirus antigen shedding and clinical gastroenteritis was statistically significant (Table 1). Viral shedding was usually detected for less than one week but one transplant patient had symptomatic shedding for a prolonged period (four weeks). At least six of the infections were hospital acquired in that their initial stool sample was negative for astrovirus antigen. Three astrovirus infected patients were admitted with a primary admitting diagnosis of gastroenteritis/dehydration. A number of astrovirus infected patients had a variety of viral co-infections: one rotavirus, three respiratory syncytial virus, and two CMV. Six patients were immunocompromised (5 solid organ transplant and one with AIDS). Three episodes of astrovirus infection in transplant patients were associated with organ rejection. The mean age of patients infected with astrovirus was 7.5 months, significantly younger than patients infected with rotavirus or adenovirus (Fig. 1). Six astrovirus strains were isolated from stool samples in tissue culture and serotyped as previously described (1). All six were serotype 1 strains.
Rotavirus antigen was detected in the stools of 21 children. Ninety percent of these children had symptomatic gastroenteritis. In fact, half of them were admitted because of dehydrating gastroenteritis (Table 1). As in astrovirus infection, rotavirus shedding was strongly correlated with symptomatic disease. Only two asymptomatic patients were identified. Shedding was generally detected for one week or less and no examples of prolonged shedding were identified. Three infections were hospital acquired. Co-infections included one astrovirus, three RSV, one influenza B, and one adenovirus. Three rotavirus positive patients were immunocompromised. The mean age of rotavirus positive patients (15 months, range: 1.5–43 months) was significantly higher than that of astrovirus infected infants (Fig. 1).
Adenovirus antigen was detected in stools from 14 patients. When the adenovirus positive stools were tested by an EIA which detects only enteric serotypes 40 and 41 that are associated with gastroenteritis, only three were positive. Only the serotype 40/41 positive patients had any symptoms of viral gastroenteritis and none had gastroenteritis as an admitting diagnosis. There was no association between shedding of nonenteric adenovirus and symptomatic gastroenteritis. Eight of the nonenteric adenovirus positive children were admitted with respiratory infections. Six of the fourteen adenovirus infections were hospital acquired and three patients had intermittent fecal shedding of adenovirus antigen. These three patients had viral shedding for three weeks or longer. Two of these three patients were immunocompromised. Concomitant viral infections in subjects with nonenteric adenovirus infection included two RSV, one astrovirus, and two CMV.
Seasonality of Rotavirus and Astrovirus
Because our data suggest that the great majority of astrovirus and rotavirus infections are symptomatic we examined data from a passive surveillance of stools from children with gastroenteritis symptoms. This surveillance consists of testing for rotavirus and astrovirus antigens on all samples submitted to the hospital laboratory for gastroenteritis. Data include a summation of three years of rotavirus/astrovirus surveillance. Data include a summation of three years of rotavirus/astrovirus surveillance (July 1997–June 2000) shown in Figure 2 as monthly totals. The data show both viruses have remarkably parallel incidence rates and peak in the winter months (January–March).
We have previously shown that astrovirus and rotavirus are frequently present in stools of children with symptomatic gastroenteritis in our hospital (1). The present study was designed to establish the prevalence of both symptomatic and asymptomatic gastrointestinal viral infection of hospitalized children over nearly two years. We were able to obtain stool samples from approximately one third of the patients admitted. Many of the admitted patients who were not sampled were admitted for 24 hours or less. Nevertheless, our study population may have been somewhat biased due to the relative ease of collecting stool samples from children with diarrhea.
We observed that the great majority of rotavirus and astrovirus infections detected by ELISA were symptomatic. While adenovirus shedding was relatively common, most adenoviruses were not types 40 or 41 and there was little associated gastroenteritis.
Recently there has been evidence that human caliciviruses may be important agents of pediatric gastroenteritis (12,13). Unfortunately, due to the large antigenic differences in strains of calicivirus, a simple ELISA test capable of detecting most infections was not available to us. Thus we were unable to obtain data comparable to our results for the other viruses for this potentially important emerging pathogen.
Our use of ELISA as a detection method likely had some influence on our results. Others have shown polymerase chain reaction based methods may be more sensitive, particularly among asymptomatic patients (14,15). Thus our results may underestimate the frequency of viral infections, particularly in asymptomatic children.
Although astrovirus and rotavirus were both causes of gastroenteritis in hospitalized children, closer examination of the differences between these two groups is of interest. Astrovirus infected a significantly younger cohort of patients (Fig. 1), an observation we and others have previously made (1,4). The reason for this difference is not known although one might speculate that maternal immunity declines more rapidly for astrovirus or that there is earlier and more frequent exposure to astrovirus. Despite infecting older children, rotavirus disease tended to be more severe. Children with rotavirus infection were significantly more likely to have been admitted with dehydration than those with astrovirus (Table 1). Half of the children found to have rotavirus infection in our study were admitted primarily for dehydrating acute gastroenteritis. In contrast, less than one fourth of the astrovirus infected patients were admitted mainly for gastroenteritis.
Astrovirus infections were more likely to be hospital acquired and were significantly more commonly found in immunocompromised patients. Six astrovirus infections were nosocomial and six infections occurred in transplant/AIDS patients. Several previous reports have noted a fairly high prevalence of astrovirus infection among immunocompromised patients including adult AIDS patients, adult bone marrow transplant patients, and pediatric transplant and oncology patients (15–19). Astrovirus infections in immunocompromised children tended to be more symptomatic and prolonged than those in normal children. A previous report associated astrovirus infection with prolonged diarrhea in malnourished children in Bangladesh (20). Three episodes of astrovirus infection in liver transplant recipients in our series were associated with organ rejection episodes. It is unclear whether this may have been due to generalized stimulation of the immune system, malabsorption of immunosuppressive medications, or some other mechanism. We previously identified patients with short bowel syndrome as being at high risk for symptomatic astrovirus infection (1). The current study could not address that question, as pediatric surgical wards were not included in this study.
The relatively low nosocomial infection rate for rotavirus in our study (only three cases observed) is somewhat surprising given the ease of rotavirus transmission in pediatric populations. We hypothesize that the more dramatic symptomatology of rotavirus patients and/or the ready availability of a rapid and specific diagnostic test from the clinical laboratory may have lead to more prompt and efficient isolation of rotavirus patients from the ward populations.
The seasonal variation in rotavirus and astrovirus were markedly similar. Previous reports have described the increased incidence of each virus in winter months in temperate climates (1,21–23) but this is the first study to simultaneously track both infections in the same facility. The remarkable similarity in the curves suggests that similar climatic/epidemiologic factors affect the transmission of both infections in contrast to year-round infection pattern reported for other agents of viral gastroenteritis such as enteric adenovirus (3). These data suggest that a portion of the winter gastroenteritis burden frequently ascribed to rotavirus in epidemiologic studies is due astrovirus infection.
In summary, both rotavirus and astrovirus infections cause large morbidity in pediatric population. Approximately 5% of all admissions were due directly to viral gastroenteritis and a similar number of children developed disease after admission. Rotaviruses tended to afflict older infants in the general population while the effects of astrovirus infection were more important in younger infants, especially those with serious underlying diseases. Both infections are perpetuated by efficient transmission in both the hospital and the community, particularly in the winter months.
1. Shastri S, Doane AM, Gonzales J, Upadhyayula U, Bass DM. Prevalence of astroviruses in a children's hospital. J Clin Microbiol 1998; 36:2571–4.
2. Herrmann JE, Taylor DN, Echeverria P, Blacklow NR. Astroviruses as a cause of gastroenteritis in children. N Engl J Med 1991; 324:1757–60.
3. Qiao H, Nilsson M, Abreu ER, Hedlund KO, Johansen K, Zaori G, et al. Viral diarrhea in children in Beijing, China. J Med Virol 1999; 57:390–6.
4. Dennehy PH, Nelson SH, Spangenberger S, Noel JS, Monroe, SS, Glass RI. A prospective case-control study of the role of astrovirus in acute diarrhea among hospitalized young children. J Infect Dis 2001; 184:10–15
5. Bates PR, Bailey AS, Wood DJ, Morris DJ, Couriel JM. Comparative epidemiology ofrotavirus, subgenus F (types 40 and 41) adenovirus and astrovirus gastroenteritis in children. J Med Virol 1993; 39:224–8.
6. Gary Jr., GW, Hierholzer JC, Black RE. Characteristics of noncultivable adenoviruses associated with diarrhea in infants: a new subgroup of human adenoviruses. J Clin Microbiol 1979; 10:96–103.
7. Bennet R, Hedlund KO, Ehrnst A, Eriksson M. Nosocomial gastroenteritis in two infant wards over 26 months. Acta Paediatr 1995; 84:667–71.
8. Esahli H, Brebiick K, Bennet R, Ehrnst A, Eriksson M, Hedlund KO. Astroviruses as a cause of nosocomial outbreaks of infant diarrhea. Pediatr Infect Dis J 1991; 10:511–5.
9. Ford-Jones EL, Mindorff CM, Gold R, Petric M. The incidence of viral-associated diarrhea after admission to a pediatric hospital. Am J Epidemiol 1990; 131:711–8.
10. Lam BC, Tam J, Ng MH; Yeung CY. Nosocomial gastroenteritis in paediatric patients. J Hosp Infect 1989; 14:351–5.
11. Glass RI, Noel J, Mitchell D, Herrmann JE, Blacklow NR, Pickering LK, et al. The changing epidemiology of astrovirus-associated gastroenteritis: a review. Arch Virol Suppl 1996; 12:287–300.
12. Pang XL, Honma S, Nakata S, Vesikari T. Human caliciviruses in acute gastroenteritis of young children in the community. J Infect Dis 2000; 181S:1–2
13. Steele A, Phillips D, Smit J, Peenze TK, Jiang X. Snow mountain-like virus identified in young children with winter vomiting disease in South Africa. J Diarrhoeal Dis Res 1997 15:177–182.
14. Mitchell DK, Monroe SS, Jiang X, Matson DO, Glass RI, Pickering LK. Virologic features of an astrovirus diarrhea outbreak in a day care center revealed by reverse transcriptase-polymerase chain reaction. J Infect Dis 1995; 172:1437–44.
15. Cubitt WD, Mitchell DK, Carter MJ, Willcocks MM, Holzel H. Application of electronmicroscopy, enzyme immunoassay, and RT-PCR to monitor an outbreak of astrovirus type 1 in a paediatric bone marrow transplant unit. J Med Virol 1999; 57:313–21.
16. Bjorkholm M, Celsing F, Runarsson G, Waldenstrom J. Successful intravenous immunoglobulin therapy for severe and persistent astrovirus gastroenteritis after fludarabine treatment in a patient with Waldenstrom's macroglobulinemia. Int J Hematol 1995; 62:117–20.
17. Cox GJ, Matsui SM, Lo RS, Hinds M, Bowden RA, Hackman RC, et al. Etiology and outcome of diarrhea after marrow transplantation: a prospective study. Gastroenterology 1994; 107:1398–407.
18. Coppo P, Scieux C, Ferchal F, Clauvel J, Lassoued K. Astrovirus enteritis in a chronic lymphocytic leukemia patient treated with tludarabine monophosphate. Ann Hematol. 2000; 79:43–5.
19. Grohmann GS, Glass RI, Pereira HG, Monroe SS, Hightower AW, Weber R, et a1. Enteric viruses and diarrhea in HIV-infected patients. Enteric Opportunistic Infections Working Group. N Engl J Med 1993; 329:14–20.
20. Unicomb LE, Banu NN, Azim T, Islam A, Bardhan PK, Faruque AS, et al. Astrovirus infection in association with acute, persistent and nosocomial diarrhea in Bangladesh. Pediatr Infect Dis J 1998; 17:611–4.
21. Sakamoto T, Negishi H, Wang QH, Akihara S, Kim B, Nishimura S, et al. Molecular epidemiology of astroviruses in Japan from 1995 to 1998 by reverse transcription-polymerase chain reaction with serotype-specific primers (1 to 8). J Med Virol 2000; 61:326–31.
22. Palombo EA, Bishop RF. Annual incidence, serotype distribution, and genetic diversity of human astrovirus isolates from hospitalized children in Melbourne, Australia. J Clin Microbiol 1996; 34:1750–3.
23. Mustafa H, Palombo EA, Bishop RF. Epidemiology of astrovirus infection in young children hospitalized with acute gastroenteritis in Melbourne, Australia, over a period of four consecutive years, 1995 to 1998. J Clin Microbiol 2000; 38:1058–62.