Estimating Gestational Age from Ultrasound: External Validation of the NICHD Formula with Comparison to the Hadlock Regression.

Objective To externally validate the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) formula developed from the National Fetal Growth Studies-Singletons and compare with 1984 Hadlock regression in a general obstetrical population. Study Design Cross-sectional study of nonanomalous singletons with a crown-rump length (CRL) and >= 1 additional ultrasound (US) with complete fetal biometrics. CRL established the referent estimated due date to calculate the error at every examination from both formulas. Error was the difference between the CRL-derived gestational age (GA) and each method's predicted GA. Comparisons were also made in three GA intervals: 1 (14 (0/7) -20 (6/7) ), 2 (21 (0/7) -28 (6/7) ), and 3 (>= 29 (0/7) ). Odds ratios evaluated the likelihood of errors outside the prespecified (+/-) day ranges. Repeated measures analysis of variance and generalized estimating equations controlled multiple US in the same patient. Results A total of 6,043 patients produced 16,904 USs for evaluation. The NICHD formula yielded significantly smaller mean errors in all GA ranges compared with the Hadlock formula ( p < 0.01). In interval 3, the NICHD formula had significantly lower odds of discerning examinations outside the prespecified error range (odds ratio: 1.27). Conclusion The NICHD formula is a valid estimate of estimating GA in a general obstetrical population and was superior to the Hadlock formula, most notably in the third trimester.