Abstract The reliability of daily unconjugated plasma estriol (E 3 ) as a first-line test for fetal surveillance in the management of diabetic pregnancies was assessed. Seventy consecutive insulin-dependent diabetic women admitted in late pregnancy were followed up with daily 8 am unconjugated plasma E 3 assays and weekly antepartum fetal heart rate testing (AFHRT). There were no perinatal deaths, and only two neonates developed the respiratory distress syndrome. The mean gestational age at delivery was 38 weeks. Pregnancies continued until elective delivery at 38 weeks and development of a mature amniotic fluid lecithin/sphingomyelin (L/S) ratio (n = 40) unless spontaneous labor began earlier (n = 20) or early delivery was indicated for either maternal reasons (n = 3) or presumed fetal distress (n = 7). Intervention for fetal distress took place when plasma E 3 levels had fallen by 40% or more from the highest mean of three consecutive preceding values (significant drop) and both a nonstress test (NST) and a contraction stress test (CST) were abnormal irrespective of the L/S ratio (n = 2) or, in the presence of a mature L/S ratio, when either plasma E 3 levels had fallen significantly or AFHRT was abnormal (n = 5). Five patients without a mature L/S ratio and a significant fall in plasma E 3 with a reactive NST were not delivered. Only 14 (1.2%) of the 1,180 plasma E 3 assays in 12 (17%) of the 70 patients represented a significant drop and fetal distress was documented in two of these 12 pregnancies, the only two in which both NST and CST were abnormal. In 18 of the 70 patients in whom 317 NSTs were performed, 26 NSTs (8.2%) were nonreactive initially, but 18 of these 26 NSTs were either reactive when repeated within hours or followed by negative CSTs. Documented fetal distress was found to occur 1 day after a reactive NST and a normal plasma E 3 but was indicated in time for successful intervention by both tests. These data indicate that: (1) daily unconjugated plasma E 3 assays constitute a reliable first-line test of fetal status in diabetic pregnancies, (2) a single daily plasma E 3 at 8 a.m. suffices, (3) the 40% limit for a significant drop appears to be sensitive enough to detect fetal distress, (4) AFHRT, if done daily, may be used in lieu of daily plasma E 3 determinations, and (5) both plasma E 3 assays and AFHRT render false abnormal results. By combining both biophysical and biochemical testing, it is possible to virtually eliminate the need for unwarranted intervention and unnecessary prematurity.