BACKGROUND: The US Zika Pregnancy and Infant Registry (USZPIR) monitors infants born to mothers with confirmed or possible Zika virus infection during pregnancy. The surveillance case definition for Zika-associated birth defects includes microcephaly based on head circumference (HC). METHODS: We assessed birth and follow-up data from infants with birth HC measurements <3rd percentile and birthweight ≥10th percentile to determine possible misclassification of microcephaly. We developed a schema informed by literature review and expert opinion to identify possible HC measurement inaccuracy using HC growth velocity and longitudinal HC measurements between 2 and 12 months of age. Two or more HC measurements were required for assessment. Inaccuracy in birth HC measurement was suspected if growth velocity was >3 cm/month in the first 3 months or HC was consistently >25th percentile during follow-up. RESULTS: Of 6,799 liveborn infants in USZPIR, 351 (5.2%) had Zika-associated birth defects, of which 111 had birth HC measurements <3rd percentile and birthweight ≥10th percentile. Of 84/111 infants with sufficient follow-up, 38/84 (45%) were classified as having possible inaccuracy of birth HC measurement, 19/84 (23%) had HC ≥3rd percentile on follow-up without meeting criteria for possible inaccuracy, and 27/84 (32%) had continued HC <3rd percentile. After excluding possible inaccuracies, the proportion of infants with Zika-associated birth defects including microcephaly decreased from 5.2% to 4.6%. CONCLUSIONS: About one-third of infants in USZPIR with Zika-associated birth defects had only microcephaly, but indications of possible measurement inaccuracy were common. Implementation of this schema in longitudinal studies can reduce misclassification of microcephaly.