Outbreak of rotavirus gastroenteritis among premature infants

Clinical manifestations of rotavirus infection in the neonatal intensive care unit

OBJECTIVES

This report describes clinical signs of rotavirus infection (RVI) among neonates admitted to a neonatal intensive care unit (NICU), compares these signs between term and preterm neonates and assesses the seasonal distribution of RVI in the NICU with that of the community.

METHODS

After an initial prevalence study of 28 days, a prospective longitudinal study in the NICU was conducted. During the next 48 months from December 1, 1991 to November 30, 1995, term and preterm NICU patients were evaluated daily for pre-defined deviations in their baseline gastrointestinal status. Fecal specimens of neonates who fulfilled the entry criteria were tested for rotavirus by a monoclonal antibody-based enzyme immunoassay and by immunoelectron microscopy. Demographic and outcome data for these neonates were collected. In addition data assessing RVI in the community were collected during this period.

RESULTS

The prevalence of RVI among NICU patients was 18.4%. Of 194 neonates included in the longitudinal study, 95 had RVI. Neonates with RVI differed from those without RVI with respect to frequent stooling (P = 0.0005), higher percentage of bloody mucoid stools (P = 0.003) and higher percentage of watery stools (P = 0.023). The odds of these three clinical outcomes were approximately 2.5 times higher in neonates with RVI than in neonates without RVI. Among neonates included in the study advanced necrotizing enterocolitis occurred at the same rate (15%) among those with and without RVI. Comparisons between term and preterm neonates with RVI showed that frequent stooling (P = 0.003) and watery stools (P = 0.0001) occurred more often among term neonates, whereas bloody mucoid stools (P = 0.001), abdominal distention (P = 0.03) and intestinal dilatation (P = 0.016) were more common in preterm neonates. The seasonal distribution of RVI in NICU paralleled its distribution in the community.

CONCLUSION

RVI appears prevalent in the NICU setting. An absence of watery stools in a neonate should not preclude consideration of RVI when evaluating gastrointestinal signs among neonates. The clinical spectrum of RVI differs in term and preterm infants.

[Rotavirus neonatal diarrhea]

A retrospective survey was performed among 24 neonatal care units. In the years 1989-1990, outbreaks of rotavirus infections were observed in 58 % and recurred in 46 % of neonatal care units. Six units presented with 3 or more outbreaks. Rapid diagnosis methods always were available and were usually agglutination latex test or ELISA test. Measures were suggested to limit the spread of the outbreaks : wearing of a gown ascribed to each infant (20/0) ; isolation of infected neonates (16/3) ; cohorting infected neonates (11/7) ; systematic stool examination for rotavirus identification in all neonates (9/10) ; use of gloves for handling contaminated stools (6/13). Stopping admissions and systematic stools examination in staff members were never applied.

A prospective study of rotavirus infections in neonatal and maternity wards

The occurrence and symptomatology of rotavirus infections was studied at three maternity wards and one neonatal unit. Rotavirus was identified in 12.7% of 553 infants and 1.3% of 542 mothers at the maternity wards. Infections were more frequent in a mixed obstetric/gynecology ward than in the pure obstetric wards. Only 10% of the infants had symptomatic infections. Subgroups of rotavirus was determined in 41 infants: 22 of subgroup I and 19 of subgroup II, which is the subgroup accounting for the majority of childhood gastroenteritis. Rotavirus was found in faecal samples from 37% of the infants at the neonatal unit during an eight-month survey. A seasonal variation with most infections during colder months was seen. Subgroup determination was possible in 29 cases, 14 subgroup I and 15 subgroup II. Fifteen per cent of the infections demonstrated diarrheal symptoms. No significant difference among other clinical data registered was seen among rotavirus infected compared to the non-infected infants. We conclude that neonatal rotavirus infections occur as an endemic infection at our maternity wards possibly combined with infections due to external sources of virus in mixed wards and neonatal units.

Serotypic characterization of rotaviruses derived from asymptomatic human neonatal infections

Nineteen rotavirus strains derived from asymptomatic neonates (seven from England, five from Australia, two from Venezuela, and five from Sweden) were successfully cultivated in primary African green monkey kidney cell cultures, serotyped by plaque reduction neutralization tests, subgrouped by indirect enzyme-linked immunosorbent assay, and electropherotyped by polyacrylamide gel electrophoresis. All 19 strains were shown to fall into one of the four known human serotypes; serotype 1 (all Venezuelan strains), serotype 2 (all Swedish strains), serotype 3 (all Australian strains), or serotype 4 (all English strains). Hyperimmune guinea pig serum raised against the Venezuelan strain (M37) neutralized not only serotype 1 (strain Wa) but also serotype 4 (strain St. Thomas no. 3) viruses to a similar degree. The English, Australian, and Venezuelan isolates were found to belong to subgroup 2, and the Swedish strains were subgroup 1 viruses. The potential importance of these rotaviruses obtained from neonates as possible vaccine candidates is discussed.

An improved enzyme-linked immunosorbent assay for the detection of rotavirus in faeces of neonates

A direct enzyme-linked immunosorbent assay (EIA) for the detection of rotavirus in neonatal stools was developed. Rabbit antiserum against SA 11 rotavirus was incorporated as both coating and detector antibody, and rotavirus-negative rabbit serum was applied as a coating antibody control to eliminate false positive results. Pretreatment of stools with EDTA was found to increase both the sensitivity and specificity of the assay. This effect was greatest when 0.25 M EDTA (tetrasodium salt) was included in homogenized stool suspensions before the removal of solid debris by centrifugation. By electron microscopy, this EDTA pretreatment appeared to partly uncoat human rotavirus particles in faeces. Potentially suitable solid phase supports and horseradish peroxidase substrates were evaluated in the development of the assay. Screening of stool samples revealed that repeated freezing and thawing of stools eliminated positive EIA reactions. The SA 11 coating antibody compared favourably with a reference coating antiserum prepared against human faecal rotavirus strains. This EIA showed greater sensitivity for rotavirus detection than electron microscopy of stool concentrates prepared by ultracentrifugation, on testing 143 stools from 99 neonates and children. The assay has been applied successfully to detection of rotavirus in stools of neonates containing meconium, smaller amounts of viral antigen than in older children, and lacteal antirotaviral antibody. It is likely to be particularly useful for cross-infection studies in hospital wards and neonatal nurseries.

Clinical range of neonatal rotavirus gastroenteritis

A baby admitted to a special care baby unit with profuse watery diarrhoea was found to have a rotavirus infection. A total of 196 babies were admitted to the unit over the next year. Routine stool samples were taken weekly from all babies and additional samples were taken from all babies who developed clinical signs suggesting sepsis. A total of 76 babies excreted rotavirus; 32 of these developed a diarrhoeal illness, 12 of whom were severely ill with bloody diarrhoea and abdominal distension; and two had perforations. The smaller, sicker babies who stayed in hospital longer were more likely to acquire infection; colostrum did not confer protection against rotavirus or symptomatic infection. The outbreak of rotavirus declined independently of the measures taken to eradicate it. Our findings suggest that neonatal rotavirus infection may occasionally cause severe gastrointestinal problems.