Association and birth prevalence of microcephaly attributable to Zika virus infection among infants in Paraíba, Brazil, in 2015–16: a case-control study

In 2015, the number of infants born with microcephaly increased in Paraíba, Brazil, after a suspected Zika virus outbreak. We did a retrospective case-control investigation to assess the association of microcephaly and Zika virus. We enrolled cases reported to the national database for microcephaly and born between Aug 1, 2015, and Feb 1, 2016, on the basis of their birth head circumference and total body length. We identified controls from the national birth registry and matched them to cases by location, aiming to enrol a minimum of two controls per case. Mothers of both cases and controls were asked about demographics, exposures, and illnesses and infants were measured at a follow-up visit 1–7 months after birth. We took blood samples from mothers and infants and classified those containing Zika virus IgM and neutralising antibodies as evidence of recent infection. We calculated prevalence of microcephaly and odds ratios (ORs) using a conditional logistic regression model with maximum penalised conditional likelihood, and combined these ORs with exposure probability estimates to determine the attributable risk. We enrolled 164 of 706 infants with complete information reported with microcephaly at birth, of whom we classified 91 (55%) as having microcephaly on the basis of their birth measurements, 36 (22%) as small, 21 (13%) as disproportionate, and 16 (10%) as not having microcephaly. 43 (26%) of the 164 infants had microcephaly at follow-up for an estimated prevalence of 5·9 per 1000 livebirths. We enrolled 114 control infants matched to the 43 infants classified as having microcephaly at follow-up. Infants with microcephaly at follow-up were more likely than control infants to be younger (OR 0·5, 95% CI 0·4–0·7), have recent Zika virus infection (21·9, 7·0–109·3), or a mother with Zika-like symptoms in the first trimester (6·2, 2·8–15·4). Once Zika virus infection and infant age were controlled for, we found no significant association between microcephaly and maternal demographics, medications, toxins, or other infections. Based on the presence of Zika virus antibodies in infants, we concluded that 35–87% of microcephaly occurring during the time of our investigation in northeast Brazil was attributable to Zika virus. We estimate 2–5 infants per 1000 livebirths in Paraíba had microcephaly attributable to Zika virus. Time of exposure to Zika virus and evidence of infection in the infants were the only risk factors associated with microcephaly. This in