One has to appreciate the messiness that results from science and research during the initial phase of evidence gathering. Just after the letter in NEJM describing the lack of resurgence of microcephaly cases in Brazil in 2016, the U.S. CDC published data from their Zika Pregnancy Registry, showing in a paper published on 4th April in the Morbidity and Mortality Weekly Report that Zika-associated birth defects occurred in 51 of 972 (5.2%) completed pregnancies in women with possible recent Zika infection in the U.S.

Screen capture of the MMWR report on 4th April 2017.

This report made it into most of the major newspapers – with alarming headlines – that same day.

Screen capture from the New York Times report.

The language used sounds a little clunky, but this is because the denominator includes pregnant women with “serological evidence of unspecified flavivirus infection” (i.e. a false-positive result might have been conferred by recent dengue infection). In the subset of pregnant women where recent Zika infection was actually confirmed, 24 of 250 (9.6%) such pregnancies resulted in birth defects.

The data and results are also available online, more or less updated regularly, here. As of 28th March this year, there were 56 live-born infants with birth defects and 7 pregnancy losses with birth defects out of 1,311 completed pregnancies reported to the U.S. Zika Pregnancy Registry.

Screen capture from the U.S. Zika Pregnancy Registry, showing the outcomes of completed pregnancies as of 28th March 2017.

How do we square the different conflicting results? Despite its inherent biases (i.e. pregnant women with symptoms are more likely to be tested than pregnant women without symptoms, and the Zika serological test – unless a plaque reduction neutralization test is performed – is notoriously nonspecific), the U.S. Zika Pregnancy Registry represents the best collected large dataset on Zika-infected pregnant women to date. The results are of significant concern and we will have to wait for local results from our much smaller birth cohort to be released, even as efforts are made to contain the latest Zika clusters.

How about Brazil and its lack of resurgence of Zika-associated microcephaly in 2016? That is another mystery that needs an explanation. There is a pre-print paper in bioRxiv which includes an author from the Health Secretariat in Rio de Janeiro Brazil that reported a 15% drop in live births in Brazil between September and December 2016, although this alone is insufficient to account for findings in the NEJM letter.