Surfactant administration to neonates with respiratory distress syndrome reduces mortality, but the effect on the incidence of BPD is controversial. We hypothesized that the increased survival of patients at high risk for BPD masks the effect of surfactant administration on the incidence of BPD. To test this hypothesis, logistic regression analyses using pretreatment variables reported to be associated with mortality were conducted in the placebo patients enrolled in the prevention and rescue trials with surfactant(Survanta) reported by Hoekstra, et al., (Pediatrics 88:10, 1991) and Liechty, et al., (Pediatrics 88:19, 1991). The logistic regression models were applied to the surfactant groups to identify those neonates expected to die if surfactant had not been administered. The incidence of BPD (oxygen requirement at 28 days) was compared among the patients in the placebo groups as well as the patients expected to die and those expected to survive in the surfactant groups (Table). (* p < 0.05, expected to die versus both placebo and expected to survive groups; † p < 0.05, expected to survive versus placebo group; χ2). In both prevention and rescue studies, the group of neonates expected to die had a higher incidence of BPD than patients in placebo groups and those expected to survive in surfactant groups. In the prevention studies, the incidence of BPD in expected to survive patients in the surfactant group was comparable to that of the placebo group. Surfactant administered as a rescue therapy decreased the incidence of BPD in those patients expected to survive compared to the placebo group. The reported lack of effect of surfactant administration on the incidence of BPD is therefore due to the increased survival of a group of patients at high risk for BPD.

Table 1
Full size table