With improved survival of very low birth weight infants, the incidence of bronchopulmonary dysplasia has significantly increased. Pulse oximetry appears to be an adequate alternative to transcutaneous PO2, for continuous arterial oxygen saturation (SaO2) monitoring in neonates; however, its usefulness has not been very well documented in very low birth weight infants. We studied 68 patients with birth weight less than 1,250 g; 44 neonates had respiratory distress syndrome and 24 had bronchopulmonary dysplasia. Using a Nellcor N-100 pulse oximeter, we compared transcutaneous oxygen saturation with simultaneous arterial samples analyzed for SaO2 (range 78% to 100%) using an IL 282 co-oximeter. Fetal hemoglobin was measured in 66 patients. We also evaluated the accuracy of transcutaneous PO2 in reflecting arterial PO2 in patients with bronchopulmonary dysplasia. Over a wide range of PO2, PCO2, pH, heart rate, BP, hematocrit, and fetal hemoglobin, linear regression analysis revealed a close correlation between pulse oximeter values and co-oximeter measured SaO2 in patients with acute (r = .88, Y = 19.41 + 0.79X) and chronic (r = .90, Y = 9.72 + 0.92X) disease. Regression analysis of transcutaneous v arterial PO2 in infants with bronchopulmonary dysplasia showed an r value of .78. In addition, in these patients with chronic disease, the mean difference between pulse oximeter SaO2 and co-oximeter measured SaO2 was 2.7 +/- 1.9% (SD); whereas the mean difference between transcutaneous and arterial PO2 was -14 +/- 10.7 mm Hg. Our findings indicate that pulse oximetry can be used reliably in very low birth weight infants with acute and chronic lung disease, for SaO2 values greater than 78%.(ABSTRACT TRUNCATED AT 250 WORDS)