RESUMEN
PURPOSE: To evaluate which reference curve (RC)-Snijders, Intergrowth 21st (IG21) and World Health Organization (WHO)-is more accurate for microcephaly diagnosis. METHODS: Retrospective cohort study with more than 30,000 exams in more than 11,000 women. Microcephaly was confirmed by a neonatologist at birth and positive predictive values (PPVs) and misdiagnosis were assessed. RESULTS: A total of 71 cases were confirmed as microcephaly at birth. IG21 and Snijders PPVs showed to be more significant over WHO's (p < 0.001), without difference between them (p = 0.39). All RC were superimposed and did not show significant difference. When evaluated in different fragments, three trends were observed (until 30 weeks, between 30 and 36 and after 36 weeks of gestational age), with the latter interval showing a significant difference between IG21 and WHO (p = 0.0079). Conversely, WHO exhibited only one misdiagnosis, a much lower rate than Snijders, who missed eight cases and IG21, nine. CONCLUSION: WHO's RC appears to misdiagnose fewer cases, which could be useful for a population screening, while IG21's RC presented a more significant PPV, being more useful for a more precise final diagnosis in reference centers.