Risk of microcephaly after Zika virus infection in Brazil, 2015 to 2016

Abstract Objective To estimate the risk of microcephaly in babies born to women infected by the Zika virus during pregnancy in Brazil in an epidemic between 2015 and 2016. Methods We obtained data on the number of notified and confirmed microcephaly cases in each Brazilian state between November 2015 and October 2016 from the health ministry. For Pernambuco State, one of the hardest hit, weekly data were available from August 2015 to October 2016 for different definitions of microcephaly. The absolute risk of microcephaly was calculated using the average number of live births reported in each state in the corresponding time period between 2012 and 2014 and assuming two infection rates: 10% and 50%. The relative risk was estimated using the reported background frequency of microcephaly in Brazil of 1.98 per 10 000 live births. Findings The estimated absolute risk of a notified microcephaly case varied from 0.03 to 17.1% according to geographical area, the definition of microcephaly used and the infection rate. Assuming a 50% infection rate, there was an 18–127 fold higher probability of microcephaly in children born to mothers with infection during pregnancy compared with children born to mothers without infection during pregnancy in Pernambuco State. For a 10% infection rate, the probability was 88–635 folds higher. Conclusion A large variation in the estimated risk of microcephaly was found in Brazil. Research is needed into possible effect modifiers, reliable measures of Zika virus infection and clear endpoints for congenital malformations.


Background
The Zika virus was initially identified in rhesus monkeys from the Zika forest in Uganda in 1947. 1 However, it was only in 2007 that it was first reported outside Africa and Asia, 2 when an epidemic occurred on Yap Island, in the Federated States of Micronesia. 3 During 2013 and 2014, there was another epidemic in French Polynesia. 4 With the virus's emergence in Brazil in 2015, a new era began. In October 2015, an increase in the number of babies born with microcephaly (referred to as microcephaly cases) was noticed in Recife in north-east Brazil; numbers continued to increase throughout the following months and reached an unprecedented total of 1912 notified cases with microcephaly by 30 April 2016. 5 In the absence of an alternative explanation and because of the temporal clustering observed, it was hypothesized that there was a causal association with Zika virus infection during pregnancy. 6,7 As the evidence accumulated, the Brazilian Government suspected this association early onand declared a national public health emergency on 11 November 2015. 8 Interestingly, after reports of the possible link between Zika virus infection and microcephaly had appeared in north-east Brazil, researchers in French Polynesia reanalysed their data and also observed this association. 9,10 In addition, the possible association was highlighted in November and December 2015 by the Centers for Disease Control and Prevention in the United States of America, the European Centre for Disease Prevention and Control and the World Health Organization (WHO). 7,11,12 On 1 February 2016, WHO declared that the clusters of microcephaly and other neurological disorders in babies constituted a Public Health Emergency of International Concern. 13 The Zika virus's potential for perinatal transmission had already been documented in 2014. 14 By 2016, the Zika virus had been identified in the amniotic fluid of fetuses with microcephaly in Brazil 15 and isolated cases of congenital malformations associated with the virus started to appear in other parts of the world, such as Slovenia 16 and Hawaii, 17,18 among individuals who had travelled to Brazil during early pregnancy. To our knowledge, never before in the history of public health have countries advised their populations to postpone planned pregnancies, as occurred, for example, in Brazil, Colombia, El Salvador and Jamaica. [19][20][21][22] During the first half of 2016, the accumulating evidence was considered strong enough to support an etiological link between Zika virus infection and birth defects. 23 Still, the exact risk of microcephaly and other congenital malformations linked to Zika virus infection during pregnancy remains unknown. The aim of this study was to estimate the risk of microcephaly -the most severe congenital malformation associated with Zika virus infection -in babies born to women in Brazil who were infected during pregnancy by examining the number of live births and the number of microcephaly cases reported in different states across the country. In addition, for Pernambuco State in north-east Brazil, where weekly figures on microcephaly cases were available for different definitions of the condition, we investigated absolute and relative risks in more detail.

Methods
For notification purposes, microcephaly was defined by the Brazilian Ministry of Health (after 8 December 2015) as a head circumference in full-term babies less than 32 cm and, Objective To estimate the risk of microcephaly in babies born to women infected by the Zika virus during pregnancy in Brazil in an epidemic between 2015 and 2016. Methods We obtained data on the number of notified and confirmed microcephaly cases in each Brazilian state between November 2015 and October 2016 from the health ministry. For Pernambuco State, one of the hardest hit, weekly data were available from August 2015 to October 2016 for different definitions of microcephaly. The absolute risk of microcephaly was calculated using the average number of live births reported in each state in the corresponding time period between 2012 and 2014 and assuming two infection rates: 10% and 50%. The relative risk was estimated using the reported background frequency of microcephaly in Brazil of 1.98 per 10 000 live births. Findings The estimated absolute risk of a notified microcephaly case varied from 0.03 to 17.1% according to geographical area, the definition of microcephaly used and the infection rate. Assuming a 50% infection rate, there was an 18-127 fold higher probability of microcephaly in children born to mothers with infection during pregnancy compared with children born to mothers without infection during pregnancy in Pernambuco State. For a 10% infection rate, the probability was 88-635 folds higher. Conclusion A large variation in the estimated risk of microcephaly was found in Brazil. Research is needed into possible effect modifiers, reliable measures of Zika virus infection and clear endpoints for congenital malformations.

Zika-associated microcephaly in Brazil
Thomas Jaenisch et al.
in preterm babies, more than 2 standard deviations below the mean as indicated by the Fenton scale. 24 The sensitivity and specificity of this definition for identifying confirmed microcephaly, as determined by imaging, were reported to be 86.0% and 93.8%, respectively, based on a series of 31 microcephaly cases from 10 states in Brazil. 24 Subsequently, the cases notified using the Brazilian Ministry of Health definition were reclassified using more specific definitions, such as the WHO InterGrowth standards. 25  . Consequently, we also considered the number of predicted confirmed cases, which was derived by multiplying the number of notified cases by the proportion of notified cases referred for confirmation that had been confirmed.
In estimating the risk of microcephaly in a particular state during a specific time period, we used as the denominator the average of the number of live births that occurred in the same period in that state in the three preceding years: 2012, 2013 and 2014. 27 We assumed that the proportion of women infected by the Zika virus during pregnancy lay between 10 and 50%, as suggested by estimates from recent epidemics in the Pacific Islands, and we used the upper and lower bounds of this range to estimate risks. To calculate relative risks, we used the background frequency of microcephaly in Brazil, which was reported to be 1.98 (95% confidence interval, CI: 1.48-2.27) per 10 000 live births between 1982 and 2013. 28 Confidence limits for relative risks were derived by dividing the lower confidence bound of the estimated absolute risk by the upper confidence bound of the observed background frequency and by dividing the upper confidence bound of the estimated risk by the lower bound of the background frequency, respectively. In Pernambuco State, where the expected number of live births during the study period was 171 402, the number of microcephaly cases reported was 5 in 2011, 9 in 2012, 10 in 2013 and 12 in 2014, 6 which corresponds to a considerably lower frequency than the background frequency we used to estimate relative risk.

Results
By assuming the proportion of women infected by the Zika virus during pregnancy was 50%, our estimate of the absolute risk of a notified microcephaly case in a baby born to a woman infected during pregnancy between 8 November 2015 and 15 October 2016 ranged from 0.03% in Santa Catarina State in southern Brazil to 3.42% in Paraíba State in northeast Brazil; the risk was 0.72% in Rio de Janeiro State and 2.51% in Pernambuco State ( Fig. 1). When the proportion of women infected was assumed to be 10%, the corresponding estimated risks were substantially higher: 0.16% in Santa Catarina State, 17.11% in Paraíba State, 3.61% in Rio de Janeiro State and 12.57% in Pernambuco State (Fig. 2). For all estimates, we used the Brazilian Ministry of Health's definition of a notified microcephaly case. In addition, the estimated absolute risk of a predicted confirmed microcephaly case in a baby born to a woman infected during pregnancy, assuming a 50% infection rate, ranged from 0.006% in Paraná State in southern Brazil to 0.99% in Sergipe State in northeast Brazil; the risk was 0.25% in Rio de Janeiro State and 0.54% in Pernambuco State ( Fig. 1). Assuming a 10% infection rate, the corresponding estimated risks were 0.03%, 4.96%, 1.25% and 2.72% in the four states, respectively (Fig. 2). Table 1 lists the estimated absolute risks of notified, predicted confirmed and confirmed cases during the study period in six states: Rio de Janeiro, Pernambuco, Paraiba, Santa Catarina, Sergipe and Paraná.
In Pernambuco State, 2155 microcephaly cases were notified between 1 August 2015 and 15 October 2016, 988 of which satisfied WHO InterGrowth standards for microcephaly. 25 Of the 988 babies, 375 had a head circumference more than 3 standard deviations below the relevant mean and 613 had a circumference between 2 and 3 standard deviations below the mean. Details of whether cases in Pernambuco State met InterGrowth standards were available for each week throughout the course of the epidemic. 5 In the three peak months of the epidemic from October to December 2015 (i.e. in epidemiological weeks 40 to 52, from 4 October 2015 to 2 January 2016), 448 cases were documented; for comparison, 328 cases were documented over 26 weeks before and after the peak (i.e. in epidemiological weeks 31 to 39 in 2015, from 2 August 2015 to 3 October 2015, and in weeks 1 to 17 in 2016, from 3 January 2016 to 30 April 2016). Table 2 lists the estimated absolute risk of notified, predicted confirmed and confirmed cases for different definitions of microcephaly and for infection rates of 10% and 50%. Depending on the definition of microcephaly used, the estimated relative risk of microcephaly in Pernambuco State varied between 18 and 127 assuming a 50% infection rate and between 88 and 635 assuming a 10% infection rate (Fig. 3). This is equivalent to an 18-127 fold (88-635 fold) higher probability of microcephaly in children born to mothers with ZIKV infection during pregnancy compared with children born to mothers without ZIKV infection during pregnancy.

Discussion
We found the estimated risk that a baby born to a woman infected by the Zika virus during pregnancy would have microcephaly varied substantially across Brazil. In particular, the risk was affected by: (i) geographical area; (ii)the definition of microcephaly used; and (iii) the percentage of women assumed to have been infected by the virus during pregnancy. In the epidemics in Yap Island in 2007 and in French Polynesia during 2013 and 2014, the seroprevalence, which reflected the proportion of the population exposed to the Zika virus, was 73% and 50 to 66%, respectively, for outbreaks that lasted 4 and 14 months, respectively. 3  of cross-reacting antibodies was taken into account when interpreting the results, 3,29 cross-reactivity may still have led to an overestimate of the seroprevalence of Zika virus antibodies in these two countries. Consequently, the reported seroprevalence in the Federated States of Micronesia and French Polynesia may be higher than would be expected in the southern states of Brazil, where there is a substantial seasonal variation in viral transmission. On the other hand, the possibility that an epidemic will undergo a stochastic die-out is greater in isolated, small, island populations; 30 the true seroprevalence may, therefore, have been underestimated, assuming the infection reached an equilibrium.
We observed a large variation in the risk of microcephaly between federal states in Brazil: the highest risks occurred in the north-east, whereas lower figures were observed inland and in the south. Similar to many states in Brazil with a moderate risk, the absolute risk of microcephaly linked to Zika virus infection during the first trimester in French Polynesia was also estimated to be around 1% in a retrospective analysis. 9 Some Brazilian states might not yet have reported microcephaly cases because the epidemic occurred late in 2015 or because they only had imported cases. However, the epidemics in Pernambuco and Rio de Janeiro States peaked almost at the same time in the spring of 2015. 25,31 The observation that the estimated risk was substantially higher in north-east Brazil than in Rio de Janeiro State, therefore, merits further attention. The possibility that cofactors or effect modifiers can play a role should be investigated in future studies.
The definitions of microcephaly used in Brazil have changed and more specific criteria based on WHO InterGrowth standards have been adopted. However, the sensitivity of these more specific criteria may be lower. 24 Current estimates of the sensitivity and specificity of different definitions of microcephaly are based on relatively small samples 24 and need to be validated in larger studies. In our study, we focused on the risk of microcephaly. However, the Zika virus may be associated with a wider range of congenital abnormalities and adverse pregnancy outcomes, including placental diseases that depend on gestational age at the time of infection. An interim analysis from Rio de Janeiro found that 29% of 42 women who had a confirmed Zika virus infection during pregnancy had babies with congenital abnormalities that were detected by ultrasound, including one baby with microcephaly. 32 Our study was limited by the fact that the number of microcephaly cases was updated frequently and because definitions of microcephaly changed over time, both of which could have resulted in a substantial variation in the estimated risk of microcephaly. Furthermore, we had to make assumptions about the proportion of women infected by the Fig. 1. Estimated absolute risk of microcephaly in a baby born to a woman

Research
Zika-associated microcephaly in Brazil Thomas Jaenisch et al.
Zika virus during pregnancy, which had a large influence on the estimated risk. Community-based seroprevalence studies of women of child-bearing age are needed to gain a better understanding of the proportion of the population infected over the course of a Zika virus epidemic.
In the absence of robust estimates of absolute and relative risks of microcephaly, cohort studies are urgently needed to determine the risk in pregnant women at different gestational ages. Moreover, microcephaly should not be the only measurement. Future studies should also evaluate the influence of potential cofactors and effect modifiers, given the wide geographical variation in risk we observed. Nevertheless, preliminary estimates of the magnitude and range of absolute and relative risks, such as those reported here, are valuable for designing future cohort studies. ■

Acknowledgements:
Ernesto TA Marques is also affiliated with the Center for Vaccine Research, University of Pittsburgh, Pittsburg, PA, United States of America.

Research
Zika-associated microcephaly in Brazil Thomas Jaenisch et al.