Background Since it is possible to identify the subgroups of RSV, A-subtype and B-subtype, there are findings indicating that the subtype may influence severity of RSV infection. Our study was designed to assess the hypothesis that A-subtype infections were more severe than B-subtype infections among hospitalized children. Patients All medical records of patients hospitalized with RSV infection between March 1990 and March 1993 were reviewed. A total of 107 children with proven RSV infection were identified. Methods Nasal waste specimens for culture were obtained from infants with suspected RSV infection. Subtype determination was done on frozen virus cultures. The following risk factors were defined: age less than or equal to 3 months, weight < 5 kg, prematurity and underlying cardiac or respiratory disease and immune deficiency. To analyse the relationship between risk factors, subtype and severity a multivariate analysis was performed. Severity was measured by clinical observations as following: pH, PCO2, SaO(2), oxygen supplementation, history of apnea and length of hospital stay. Main results Of the enrolled patients 11 had underlying disease and 17 were premature. The age range was 1 week to 4.2 years, median 3.5 months. 46 children were younger than 3 months, 33 had a weight of less than 5 kg. The isolates of 84 children were typeable: 63 isolates were subtype A and 21 subtype B. Underlying disease and prematurity were associated with SaO(2) < 87% (p =0.003) and oxygen supplementation (p = 0.017). A weight of less than 5 kg was correlated with a PCO2 greater than or equal to 50 mmHg. The RSV subtype was not significantly correlated with severity. Conclusions RSV infection even in very young children is predominantly influenced by underlying disease, prematurity and weight. The RSV subtype was no independent risk factor for an increased morbidity in this retrospective study. Therefore, in our opinion, RSV subtype is less meaningful to predict the severity of RSV infection than known risk factors.