Rationale: Surfactant and improved ventilators have had limited effects on rates of Chronic Lung Disease (CLD). Improving treatments must focus on those at highest risk.

Methods: We developed multivariate models of CLD risk using 152 infants prospectively studied in 5 NICUs from 4/90-6/92. All met the following criteria 1) birthweight (BW) ≤1250 g, 2) gestational age (GA) ≤28 wk 3) mechanically ventilated within 3 hrs and > 24 of the first 48 hr 4) survived to 42 weeks GA. We defined 3 outcomes: 1) 02 use at 30 days (N=135), 2) 02 use at 36 wk (N=126), and 3) abnormal late pulmonary outcome (02 at 42 wks GA or wheezing requiring maintenance drug therapy during the first 2 years) (N=120). Numerous demographic (race, GA, BW), treatment (antenatal steroids, surfactant, ventilator settings) and early physiological risk factors (Apgars, blood gasses, Score for Neonatal Acute Physiology (SNAP)) were assessed in logistic models predicting each outcome.

Results: There were no significant predictors of 02 use at 30 days. High 02 on day 2 predicted 02 use at 36 weeks and African-American race predicted abnormal pulmonary outcome (see table). Mean day 2 02 of infants on oxygen at 36 weeks was 41.7% ±6.1% vs 34.0%±13.1% (±sd). The significance of African-American race was due to its strength as a predictor of the reactive airway component of the outcome. Reanalysis including infants who died prior to 42 weeks yielded the same results for all 02 use outcomes. Multivariate models poorly predicted the outcomes of interest. None achieved an ROC curve area >.75. Many infants with few traditional risk factors had CLD and many with several risk factors survived intact.

Table 1
Full size table

Conclusion: Epidemiological markers of vulnerability (GA, BW, race, gender) and injury (SNAP, blood gasses, vent settings) poorly predict later outcome in this high risk, homogeneous, ventilated cohort. This suggests that later outcomes may be determined more by genetic predisposition and failure of lung repair than by initial lung injury. Further investigation of the contribution of these factors to abnormal lung outcomes is urgently needed.