Browse the corpus

Mortality from Congenital Zika Syndrome - Nationwide Cohort Study in Brazil. BACKGROUND: Prenatal exposure to Zika virus has potential teratogenic effects, with a wide spectrum of clinical presentation referred to as congenital Zika syndrome. Data on survival among children with congenital Zika syndrome are limited. METHODS: In this population-based cohort study, we used linked, routinely collected data in Brazil, from January 2015 through December 2018, to estimate mortality among live-born children with congenital Zika syndrome as compared with those without the syndrome. Kaplan-Meier curves and survival models were assessed with adjustment for confounding and with stratification according to gestational age, birth weight, and status of being small for gestational age. RESULTS: A total of 11,481,215 live-born children were followed to 36 months of age. The mortality rate was 52.6 deaths (95% confidence interval [CI], 47.6 to 58.0) per 1000 person-years among live-born children with congenital Zika syndrome, as compared with 5.6 deaths (95% CI, 5.6 to 5.7) per 1000 person-years among those without the syndrome. The mortality rate ratio among live-born children with congenital Zika syndrome, as compared with those without the syndrome, was 11.3 (95% CI, 10.2 to 12.4). Among infants born before 32 weeks of gestation or with a birth weight of less than 1500 g, the risks of death were similar regardless of congenital Zika syndrome status. Among infants born at term, those with congenital Zika syndrome were 14.3 times (95% CI, 12.4 to 16.4) as likely to die as those without the syndrome (mortality rate, 38.4 vs. 2.7 deaths per 1000 person-years). Among infants with a birth weight of 2500 g or greater, those with congenital Zika syndrome were 12.9 times (95% CI, 10.9 to 15.3) as likely to die as those without the syndrome (mortality rate, 32.6 vs. 2.5 deaths per 1000 person-years). The burden of congenital anomalies, diseases of the nervous system, and infectious diseases as recorded causes of deaths was higher among live-born children with congenital Zika syndrome than among those without the syndrome. CONCLUSIONS: The risk of death was higher among live-born children with congenital Zika syndrome than among those without the syndrome and persisted throughout the first 3 years of life. (Funded by the Ministry of Health of Brazil and others.).

We conducted a retrospective population-based cohort study, including all singletons live births in Brazil from January 1, 2015 to December 31, 2018. These live births were followed-up from birth until December 31, 2018, death, or age of 36 months. We obtained information about CZS from the Public Health Event Record 13 , which registered information from all cases with suspected microcephaly and/or central nervous system (CNS) alterations possibly associated with congenital ZIKV infection since 2015. From the Public Health Event Record, we retained data on final disease classification (confirmed or excluded), maternal rash during pregnancy and live birth head circumference. The Live Births Information System-SINASC 14 , an information system with 100% coverage of the Brazilian territory 15 , records data from the Declaration of Live Births, a legal document completed by the health worker who assists the delivery. From this system, we obtained information about the mother (maternal age, education, marital status, ethnicity); the pregnancy (antenatal appointments, length of gestation, multiple fetuses); and the newborn (birth weight, sex).

n of Live Births, a legal document completed by the health worker who assists the delivery. From this system, we obtained information about the mother (maternal age, education, marital status, ethnicity); the pregnancy (antenatal appointments, length of gestation, multiple fetuses); and the newborn (birth weight, sex). Death-related information was obtained from the Mortality Information System, which records the Death Certificates. The death certificate is a legal document that must be completed by the physician responsible for clinical care, an assistant or another practician from the institution who attests to the cause of death. In cases when the death occurs without medical assistance, the death certificate will be provided by a pathologist. We obtained information on the date and the underlying cause of death by International Classification of Diseases/ICD 10. As of 2011, Mortality Information System was estimated to cover 96.1% of deaths in Brazil 16 .

n the death occurs without medical assistance, the death certificate will be provided by a pathologist. We obtained information on the date and the underlying cause of death by International Classification of Diseases/ICD 10. As of 2011, Mortality Information System was estimated to cover 96.1% of deaths in Brazil 16 . Live births records from Live Births Information System were linked with the Public Health Event Record and Mortality Information System. Name, date of birth or age, and maternal residence were used as matching variables. The linkage was performed with the Center of Data and Knowledge Integration for Health -RL-Record Linkage; a novel record-linkage tool developed to link large-scale administrative datasets at Center of Data and Knowledge Integration for Health applying the combination of indexing and searching algorithms approach 17 . All data were extracted in 2020 and made available by the Brazilian Ministry of Health. Linkage procedures were conducted at Center of Data and Knowledge Integration for Health in a strict data protection environment and according to ethical and legal rules 18 . We included all singleton live births who contributed records during the study period. We excluded records registered in the Public Health Event Record as suspected of CZS but ruled out after epidemiological investigation and those classified as cases under investigation or inconclusive.

Live births records from Live Births Information System were linked with the Public Health Event Record and Mortality Information System. Name, date of birth or age, and maternal residence were used as matching variables. The linkage was performed with the Center of Data and Knowledge Integration for Health -RL-Record Linkage; a novel record-linkage tool developed to link large-scale administrative datasets at Center of Data and Knowledge Integration for Health applying the combination of indexing and searching algorithms approach 17 . All data were extracted in 2020 and made available by the Brazilian Ministry of Health. Linkage procedures were conducted at Center of Data and Knowledge Integration for Health in a strict data protection environment and according to ethical and legal rules 18 . We included all singleton live births who contributed records during the study period. We excluded records registered in the Public Health Event Record as suspected of CZS but ruled out after epidemiological investigation and those classified as cases under investigation or inconclusive. In Brazil, live births who meet one or more of the following criteria should be reported and investigated as suspected cases of CZS: 1) microcephaly defined as an head circumference more than 2 standard deviations below the mean for age and sex (according to INTERGROWTH 21st standards 19 for those born with less than 37 gestation weeks, or World Health Organization (WHO) standards for those born with 37 gestation weeks or more); 2) craniofacial disproportion (microcrania in relation to the face); CNS changes suggestive of congenital infection detected from neuro-imaging tests (accepted imaging were cranial computed tomography, brain magnetic resonance, or transfontanellar ultrasound); two or more neurological, visual or auditory manifestations; 3) newborns or fetuses from mothers who reported a fever and/or skin rash during pregnancy, likely or confirmed for ZIKV infection, regardless of the identification of congenital malformations at birth 20,21 . After notification, all suspected cases were investigated by the epidemiological surveillance teams and classified as confirmed, probable, inconclusive or excluded cases (Figure S2).

h during pregnancy, likely or confirmed for ZIKV infection, regardless of the identification of congenital malformations at birth 20,21 . After notification, all suspected cases were investigated by the epidemiological surveillance teams and classified as confirmed, probable, inconclusive or excluded cases (Figure S2). Suspected cases were considered confirmed when they had signs and symptoms consistent with CZS (Table S2), and laboratory evidence of ZIKV infection (from molecular or serological testing) or their mother reported fever and/or rash during pregnancy. Probable cases showed clinical changes compatible with CZS, and tested negative for other congenital infections, but the specific laboratory diagnosis for ZIKV infection was not available, and the mother was asymptomatic during pregnancy. Suspected cases were considered excluded if they had compatible clinical symptoms that, after clinical and epidemiological investigation, were attributed to having another cause, for example, microcephaly related to restricted intrauterine growth or genetic diseases. Other cases were inconclusive due to insufficient information for proper classification or remained under investigation 20 .

symptoms that, after clinical and epidemiological investigation, were attributed to having another cause, for example, microcephaly related to restricted intrauterine growth or genetic diseases. Other cases were inconclusive due to insufficient information for proper classification or remained under investigation 20 . We defined CZS cases as all live births classified in the Public Health Event Record as confirmed or probable cases and had a register linked with a Live Births Information System record. We then classified live births with CZS into two categories: with microcephaly (head circumference more than 2 standard deviations below the mean for age and sex) or without microcephaly, according to INTERGROWTH 21st standards. We also classified CZS cases according to the time of maternal rash during pregnancy (first trimester, second trimester, third trimester, or no report of rash during pregnancy). Live births that died during the study period were identified by linking the Live Births Information System with Mortality Information System.

We also classified CZS cases according to the time of maternal rash during pregnancy (first trimester, second trimester, third trimester, or no report of rash during pregnancy). Live births that died during the study period were identified by linking the Live Births Information System with Mortality Information System. Descriptive statistics are presented for maternal sociodemographic data and newborn characteristics. Mortality rates (deaths/1,000-person-year (PY) and crude hazard ratios (HRs) with 95% CIs comparing live births with CZS to live births without CZS were estimated using Cox proportional hazards models. We also conducted a sensitivity analysis with confirmed cases only. Kaplan-Meier curves were plotted, and we compared live births with and without CZS in total, CZS according to head circumference (classified as with microcephaly and without microcephalic) and CZS according to maternal rash during pregnancy compared with live births without CZS. Finally, we fitted penalized Cox proportional hazards regression models using restricted maximum likelihood with frailty terms corresponding to random effects from a Gaussian distribution to account for within-cluster (maternal residency region) homogeneity in outcomes 22 . We used these analyses to obtain the adjusted HR. The adjusted models were controlled for maternal age, education, marital status, ethnicity, number of prenatal appointments, newborn sex and year of birth and stratified by gestational age (< 32, 32-36, ≥37 weeks), birth weight (<1500, 1500-2499, ≥2500 grams) and small for gestational age (smaller than the 10th centile according to the INTERGROWTH 21st standards). For the time scale in our survival analyses, we used age attained within the study, neonatal mortality (up to 27 days), infant mortality (up to 364 days) and mortality up to 36 months. Data analyses were performed in Stata version 15.0.

al age (smaller than the 10th centile according to the INTERGROWTH 21st standards). For the time scale in our survival analyses, we used age attained within the study, neonatal mortality (up to 27 days), infant mortality (up to 364 days) and mortality up to 36 months. Data analyses were performed in Stata version 15.0. This study analyzed de-identified data and was approved by the Institute of Collective Health, Federal University of Bahia Research Ethics Committee (CAAE registration no. 70745617 2 0000 5030).

We conducted a retrospective population-based cohort study, including all singletons live births in Brazil from January 1, 2015 to December 31, 2018. These live births were followed-up from birth until December 31, 2018, death, or age of 36 months.

We obtained information about CZS from the Public Health Event Record 13 , which registered information from all cases with suspected microcephaly and/or central nervous system (CNS) alterations possibly associated with congenital ZIKV infection since 2015. From the Public Health Event Record, we retained data on final disease classification (confirmed or excluded), maternal rash during pregnancy and live birth head circumference. The Live Births Information System-SINASC 14 , an information system with 100% coverage of the Brazilian territory 15 , records data from the Declaration of Live Births, a legal document completed by the health worker who assists the delivery. From this system, we obtained information about the mother (maternal age, education, marital status, ethnicity); the pregnancy (antenatal appointments, length of gestation, multiple fetuses); and the newborn (birth weight, sex). Death-related information was obtained from the Mortality Information System, which records the Death Certificates. The death certificate is a legal document that must be completed by the physician responsible for clinical care, an assistant or another practician from the institution who attests to the cause of death. In cases when the death occurs without medical assistance, the death certificate will be provided by a pathologist. We obtained information on the date and the underlying cause of death by International Classification of Diseases/ICD 10. As of 2011, Mortality Information System was estimated to cover 96.1% of deaths in Brazil 16 .

Live births records from Live Births Information System were linked with the Public Health Event Record and Mortality Information System. Name, date of birth or age, and maternal residence were used as matching variables. The linkage was performed with the Center of Data and Knowledge Integration for Health -RL-Record Linkage; a novel record-linkage tool developed to link large-scale administrative datasets at Center of Data and Knowledge Integration for Health applying the combination of indexing and searching algorithms approach 17 . All data were extracted in 2020 and made available by the Brazilian Ministry of Health. Linkage procedures were conducted at Center of Data and Knowledge Integration for Health in a strict data protection environment and according to ethical and legal rules 18 .

We included all singleton live births who contributed records during the study period. We excluded records registered in the Public Health Event Record as suspected of CZS but ruled out after epidemiological investigation and those classified as cases under investigation or inconclusive. In Brazil, live births who meet one or more of the following criteria should be reported and investigated as suspected cases of CZS: 1) microcephaly defined as an head circumference more than 2 standard deviations below the mean for age and sex (according to INTERGROWTH 21st standards 19 for those born with less than 37 gestation weeks, or World Health Organization (WHO) standards for those born with 37 gestation weeks or more); 2) craniofacial disproportion (microcrania in relation to the face); CNS changes suggestive of congenital infection detected from neuro-imaging tests (accepted imaging were cranial computed tomography, brain magnetic resonance, or transfontanellar ultrasound); two or more neurological, visual or auditory manifestations; 3) newborns or fetuses from mothers who reported a fever and/or skin rash during pregnancy, likely or confirmed for ZIKV infection, regardless of the identification of congenital malformations at birth 20,21 . After notification, all suspected cases were investigated by the epidemiological surveillance teams and classified as confirmed, probable, inconclusive or excluded cases (Figure S2).

Descriptive statistics are presented for maternal sociodemographic data and newborn characteristics. Mortality rates (deaths/1,000-person-year (PY) and crude hazard ratios (HRs) with 95% CIs comparing live births with CZS to live births without CZS were estimated using Cox proportional hazards models. We also conducted a sensitivity analysis with confirmed cases only. Kaplan-Meier curves were plotted, and we compared live births with and without CZS in total, CZS according to head circumference (classified as with microcephaly and without microcephalic) and CZS according to maternal rash during pregnancy compared with live births without CZS. Finally, we fitted penalized Cox proportional hazards regression models using restricted maximum likelihood with frailty terms corresponding to random effects from a Gaussian distribution to account for within-cluster (maternal residency region) homogeneity in outcomes 22 . We used these analyses to obtain the adjusted HR. The adjusted models were controlled for maternal age, education, marital status, ethnicity, number of prenatal appointments, newborn sex and year of birth and stratified by gestational age (< 32, 32-36, ≥37 weeks), birth weight (<1500, 1500-2499, ≥2500 grams) and small for gestational age (smaller than the 10th centile according to the INTERGROWTH 21st standards). For the time scale in our survival analyses, we used age attained within the study, neonatal mortality (up to 27 days), infant mortality (up to 364 days) and mortality up to 36 months. Data analyses were performed in Stata version 15.0.

We followed 11,737,554 live births from birth up to 36 months (mean 23 [range 0–36]) (Figure S1). The characteristics of our study population are reported in Table 1. In general, live births with CZS (n=3,308) had mothers who were younger and less educated. Nearly 20% of the live births with CZS were born preterm, 36% were low birth weight (LBW), and 37% were small for gestational age (SGA) compared with 10.2% preterm births, 7.3% LBW and 6.9% SGA among live births without CZS (n=11,477,907). By the end of the study period, 398 live births who met the study case definition criteria of CZS and 120,629 live births without CZS had died. The overall mortality rate up to 36 months in live births with CZS was 11.3 times higher (95% CI, 10.2-12.4) than in live births without CZS. The increased risk of death persisted throughout the observation period with no suggestion of attenuation by the age of three years in children with CZS. The highest mortality rate ratio, 21.9 (95% CI, 17.3-27.6), was found after the first year of life when the mortality in children without CZS was 0.7 deaths/1000 PY (Table 2). The analysis restricted to confirmed CZS cases showed similar results (Table S1).

of attenuation by the age of three years in children with CZS. The highest mortality rate ratio, 21.9 (95% CI, 17.3-27.6), was found after the first year of life when the mortality in children without CZS was 0.7 deaths/1000 PY (Table 2). The analysis restricted to confirmed CZS cases showed similar results (Table S1). The likelihood of death in the study population from birth up to 36 months is seen in Figure 1. Among live births with CZS the probability of death was greater than in live births without CZS, and the difference continued to increase during the study period. Information on head circumference was available for 82.6% (2,733/3,308) of live births with CZS: 66.0% had microcephaly at birth. Among live births with CZS and complete information on maternal rash, 51.2% (1,364/2,663) reported this symptom during pregnancy, with 62.2% of exanthems reported in the first trimester, 21.5% the second, 8.7% the third and 7.6% with unspecified trimester. The association of CZS on mortality up to 36 months did not materially differ for live births with and without microcephaly or for whether mothers reported a rash during pregnancy and the timing of any rash. However, numbers were relatively small, and we may have lacked the power to identify sub-group differences (Figure 1).

sociation of CZS on mortality up to 36 months did not materially differ for live births with and without microcephaly or for whether mothers reported a rash during pregnancy and the timing of any rash. However, numbers were relatively small, and we may have lacked the power to identify sub-group differences (Figure 1). Mortality rates differ across gestational age, birth weight and small for gestational age categories. Live births with and without CZS had similar absolute mortality rates if they were born before 32 weeks or weighed less than 1500g. If they were born after 32 weeks, live births with CZS were more likely to die than live births without CZS, and the highest ratios were observed among children born term and with normal birth weight; in these groups, live births with CZS were 14.3 (95% CI, 12.4-16.4) and 12.9 (95% CI, 10.9-15.3) times more likely to die when compared with their counterpart live births without CZS, respectively. Small for gestational age live births with CZS were almost five times more likely to die, while appropriate for gestational age children were nine times more likely to die than their counterpart live births without CZS (Figure 2).

likely to die when compared with their counterpart live births without CZS, respectively. Small for gestational age live births with CZS were almost five times more likely to die, while appropriate for gestational age children were nine times more likely to die than their counterpart live births without CZS (Figure 2). Among infants with CZS, the causes of deaths coded under ICD-10 certain infectious and parasitic diseases (A00-B99), diseases of the nervous system (G00-G99), and congenital malformations (Q00-Q99) were two times more common than among infant without CZS. The leading causes were sepsis, unspecified organism; hydrocephalus unspecified; and microcephaly, respectively. After the first year of life, the nervous system diseases and congenital anomalies continue to be leading chapters. However, for the G00-G99 chapter, the most common cause of death in this age group was cerebral palsy. The circulatory system diseases (I00-I99) were responsible for 58% more deaths in the CZS group than live birth without CZS (Figure 3). The causes of deaths were cardiomyopathy, other cardiac arrhythmias, and heart failure.

Analyses of Brazilian national, registry-based data showed that the mortality rates up to 3 years old were more than eleven times greater in live births with CZS than live births without CZS overall. Among live births with CZS, mortality risk by microcephaly status and maternal reported rash did not materially differ. The risk of death among the smallest infants was similar regardless of CZS status; however, term and normal birth weight live births with CZS were over 12 times more likely to die than their counterparts without CZS. Finally, we observed that conditions classified as congenital malformations, diseases of the nervous system and certain infectious diseases were more common causes of deaths among live births with CZS than those without CZS. Data on mortality associated with CZS are scant. In a series of confirmed CZS cases in Brazil, mortality risk up to 8 days was estimated at 41.1 per 1000 live births 23 . In a population-based surveillance study conducted among all infants and fetuses with congenital abnormalities potentially related to Zika virus infection in the USA, the neonatal mortality risk was 45.9 per 1000 live births 24 . A similar risk was observed in our study. However, only a longer follow-up could have revealed that after the neonatal period the absolute mortality rates in the live births with CZS did not decrease as dramatically over time as they did among live births without CZS.

l mortality risk was 45.9 per 1000 live births 24 . A similar risk was observed in our study. However, only a longer follow-up could have revealed that after the neonatal period the absolute mortality rates in the live births with CZS did not decrease as dramatically over time as they did among live births without CZS. Our understanding of the impact of CZS on the brain is still emerging; the effects of CNS dysfunction with or without microcephaly are expected to result in a wide variety of outcomes 4,25 . Although some studies have shown that the most severe phenotype appears to be associated with exposure during the first trimester 25,26 , neither early exposure (using rash as a proxy) nor microcephaly status showed difference in the mortality risk, potentially due to the small number of events. Therefore, the role of head circumference and timing of maternal symptoms of ZIKV infection as risk factors for death among children with CZS cannot be fulled assessed.

early exposure (using rash as a proxy) nor microcephaly status showed difference in the mortality risk, potentially due to the small number of events. Therefore, the role of head circumference and timing of maternal symptoms of ZIKV infection as risk factors for death among children with CZS cannot be fulled assessed. Previous studies have shown that live births with CZS had greater frequencies of LBW 27 and SGA 28,29 , features compatible with higher child mortality. However, the absolute mortality risks in the smallest infants did not differ according to CZS status. Moreover, in the CZS group, term normal-weight infants, who would have had a high chance of thriving without the impairments resulting from CZS, are at strikingly elevated mortality risk. Children with CZS have multiple neurological complications and long-term sequelae, such as cerebral palsy that was one of the main causes of deaths identified in this study and epilepsy, estimated in 67% in this group 30 , that confer an increased risk of death. However, a better understanding of the causal mortality chain is needed.

ZS have multiple neurological complications and long-term sequelae, such as cerebral palsy that was one of the main causes of deaths identified in this study and epilepsy, estimated in 67% in this group 30 , that confer an increased risk of death. However, a better understanding of the causal mortality chain is needed. A strength of our study was the large sample size, including all confirmed and probable CZS cases notified in the country. We also included a population-representative comparison group and were able to control for confounding. The sensitivity analyses (including only confirmed cases) showed consistency of our findings. There are, however, limitations. First, the present study was based on registry data, and relevant clinical data were not available. Second, at the beginning of the epidemic, the health services network did not have specific diagnostic tests for ZIKV infections. Therefore, there may have been underreporting in the Public Health Event Record, mainly among those fetuses prenatally exposed to ZIKV, but without detectable malformations at birth. Third, the linkage process could have introduced classification bias due to a linkage error. However, if an error occurred in the linkage that enabled the exposure assessment (Public Health Event Record) it would have likely underestimated the measure of association. If the error occurred in the linkage that enabled the outcome assessment (Live Births Information System - Mortality Information System) it is probably non-differential and unlikely to introduced bias in the measure of association, although the absolute measures of risk may be underestimated. There was a slight variation in data completeness by CZS status. However, data on all variables were more than 80% complete in both groups.

ity Information System) it is probably non-differential and unlikely to introduced bias in the measure of association, although the absolute measures of risk may be underestimated. There was a slight variation in data completeness by CZS status. However, data on all variables were more than 80% complete in both groups. Our findings draw attention to the importance of primary prevention of infection in women (of childbearing age and pregnant women) against Aedes aegypti bites. Disclosure forms provided by the authors are available with the full text of this article at NEJM.org.