Epidemiological trends of isolated and non-isolated central nervous system congenital malformations in live births in a middle-income setting

Abstract Objectives This study aimed to analyze, in the São Paulo state of Brazil, time trends in prevalence, neonatal mortality, and neonatal lethality of central nervous system congenital malformations (CNS-CM) between 2004 and 2015. Methods Population-based study of all live births with gestational age ≥22 weeks and/or birthweight ≥400 g from mothers living in São Paulo State, during 2004–2015. CNS-CM was defined by the presence of International Classification Disease 10th edition codes Q00–Q07 in the death and/or live birth certificates. CNS-CM was classified as isolated (only Q00–Q07 codes), and non-isolated (with congenital anomalies codes nonrelated to CNS-CM). CNS-CM associated neonatal death was defined as death between 0 and 27 days after birth in infants with CNS-CM. CNS-CM prevalence, neonatal mortality, and lethality rates were calculated, and their annual trends were analyzed by Prais-Winsten Model. The annual percent change (APC) with 95% confidence interval (95%CI) was obtained. Results 7,237,628 live births were included in the study and CNS-CM were reported in 7526 (0.1%). CNS-CM associated neonatal deaths occurred in 2935 (39.0%). Isolated CNS-CM and non-isolated CNS-CM were found respectively in 5475 and 2051 livebirths, with 1525 (28%) and 1410 (69%) neonatal deaths. CNS-CM prevalence and neonatal lethality were stationary, however neonatal mortality decreased (APC −1.66; 95%CI −3.09 to −0.21) during the study. For isolated CNS-CM, prevalence, neonatal mortality, and lethality decreased over the period. For non-isolated CNS-CM, the prevalence increased, neonatal mortality was stationary, and lethality decreased during the period. The median time of CNS-CM associated neonatal deaths was 18 h after birth. Conclusions During a 12-year period in São Paulo State, Brazil, neonatal mortality of infants with CNS-CM in general and with isolated CNS-CM showed a decreasing pattern. Nevertheless CNS-CM mortality remained elevated, mostly in the first day after birth.

Brazil has several socioeconomic inequities [14], and CNS-CM mortality rates during the first year after birth varied among Brazilian states between 4.1 and 7.0 per 10,000 live births during 2000-2014.São Paulo is among the states with the lowest rates of deaths [15].São Paulo State is the richest, most populous, and urbanized subnational Brazilian entity, with good socioeconomic indicators [14].São Paulo State's vital statistics system captures 99.5% of civil record data [16].There are some reports on CNS-CM mortality during the first year after birth in São Paulo State [8,15], however population-based studies that specifically address the CNS-CM prevalence, neonatal mortality and lethality in the state are sparse.
Considering that understanding CNS-CM epidemiology in a specific region is essential to provide consistent information to guide public health policies, the objectives of this study were to analyze the time trends in prevalence, neonatal mortality, and neonatal lethality of CNS-CM in the São Paulo state of Brazil, between 2004 and 2015.

Methods
This is a population-based study of all live births from mothers residing in São Paulo State, Brazil, between 2004 and 2015, after exclusion of live births with birthweight <400 g, gestational age <22 weeks and unknown birthweight and gestational age.
The study used the databases of live births and deaths provided by the "State Data Analysis System Foundation" (SEADE Foundation).SEADE Foundation is responsible for the civil records in São Paulo.The Foundation receives monthly all the information related to births and infant deaths (up to 365 days after birth) filled in by physicians in the Live Birth and Death Certificates, respectively.In Live Birth and Death Certificates, notification and coding of congenital anomalies is mandatory [17].Based on live birth and death information, the SEADE Foundation makes the deterministic linkage from death to live birth records in order to identify birth information from all infants who died within 365 days after birth [18].For this study, the datasets on live births and on neonatal deaths linked to livebirth information were reorganized by integrating common variables between both datasets [19].
The CNS-CM definition was based on the International Classification of Diseases, 10 th revision (ICD-10, WHO) codes and considered the presence of codes Q00 to Q07 found in any line of the death and/ or live birth certificates.As live births may present multiple CNS-CM codes registered in the Birth and/or Death Certificates, a hierarchy was created to transform the CNS-CM multiple diagnoses into a single one.In this way, each liveborn infant would have one main CNS-CM.The hierarchy was sequentially based on eight ICD-10 CNS-CM subgroups, visibility of CNS-CM at birth, the expected frequency of specific CNS-CM [2,3,20], and on grouping of related CNS-CM.According to these criteria, the following subgroups were considered for the hierarchical organization: (1) Anencephaly and similar malformations, (2) Encephalocele, (3) Spina bifida and/or Arnold Chiari Syndrome, (4) Congenital hydrocephalus, (5) Microcephaly and (6) Other.ICD-10 codes considered for each subgroup are available in Supplementary Table 1.
CNS-CM associated neonatal death was defined as death occurring between 0 and 27 days after birth with an ICD-10 code of Q00-Q07 reported in the Birth and/or Death Certificate.
Prevalence, neonatal mortality, and lethality rates of CNS-CM (global, isolated, and non-isolated CNS-CM) were calculated and presented as mean and 95% confidence interval (95%CI).Prevalence was calculated by the number of live births with CNS-CM per 10,000 live births; neonatal mortality was obtained by the number of CNS-CM associated neonatal deaths per 10,000 live births; and lethality was calculated by the number of CNS-CM associated neonatal deaths per 100 live births with CNS-CM.Prevalence, neonatal mortality, and lethality annual trends were analyzed by the Prais-Winsten Model, and their annual percent change (APC) with 95%CI were calculated.
Kaplan-Meier estimator was applied to identify the time from birth to death for each CNS-CM group during the 12-year study period.
All procedures were done using Stata 17V R (StataCorp LLC, Texas, USA).The study was approved by the Ethics Committee of Universidade Federal de São Paulo, under the number 4.055.489,and by the Board of Directors of Fundac¸ão SEADE.

Results
During the 12-year period, there were 7,317,611 live births in São Paulo State, Brazil.Among them, 7,237,628 were included in the study.CNS-CM were reported in 7526 (0.1%) live births, with 2935 (39.0%)CNS-CM associated neonatal deaths (Figure 1).
Among CNS-CM live births, 335 were identified with multiple CNS-CM codes and classified by according to the hierarchical procedure (Supplementary Table 1).Most infants with CNS-CM had an isolated malformation (Figure 1).Among those with nonisolated malformations, anomalies in the musculoskeletal system (26.4%)were most frequent, followed by other congenital malformations (25.5%) and by circulatory system malformations (11.0%) (Supplementary Table 2).
Regarding the subgroups, the prevalence of spina bifida and/or Arnold Chiari syndrome was 3.1 per 10,000 live births, followed by congenital hydrocephalus (2.7 per 10,000 live births), and anencephaly and similar malformations (1.9 per 10,000 live births) (Supplementary Table 1).Subgroups' data on prevalence, mortality and lethality during the study period are available on Supplementary Table 3.
Regarding neonatal characteristics, infants of the non-isolated CNS-CM subgroup, compared to the isolated CNS-CM subgroup had lower gestational age and birthweight and a higher frequency of males and 1st and 5th minute Apgar scores lower than 7 (Supplementary Table 4).Comparing maternal and neonatal characteristics between infants with isolated and non-isolated CNS-CM who died during the neonatal period or were alive with 27 days after birth, CNS-CM associated neonatal deaths, in both groups, had less prenatal care visits, lower gestational age and birthweight, higher frequency of vaginal delivery, and higher frequency of 1st and 5th minute Apgar scores lower than 7 (Table 1).

Discussion
This population-based study evaluated temporal trends of CNS-CM prevalence, neonatal mortality and lethality during a 12-year period in São Paulo State, Brazil.CNS-CM prevalence was 10.4 per 10,000 live births, with a stationary pattern throughout the period.Neonatal mortality was 4.1 per 10,000 live births and decreased during the studied period, but lethality was 39% and stationary.Isolated CNS-CM occurred in 72.7% of the cases and its prevalence, neonatal mortality and lethality decreased along the years.For non-isolated CNS-CM, prevalence increased over time, with stationary neonatal mortality and decreasing neonatal lethality.Among CNS-CM cases,  spina bifida and/or Arnold-Chiari syndrome were the most frequent anomalies.The CNS-CM associated neonatal deaths occurred mostly within the first day after birth.
During the last 20 years, the Brazilian government adopted several strategies for the prevention and management of congenital malformations [17].Among them, maternal folic acid fortification, efforts to improve notification and training of health professionals to provide timely diagnosis, and adequate initial care have the highest potential to impact CNS-CM outcomes [17,21,22].
Maternal folic acid supplementation has been a mandatory public health strategy in Brazil since 2004 [21].According to a hospital-based study from the ECLAMC database, folic acid supplementation contributed to a 40-60% reduction in neural tube defects in the country between 1982 and 2007 [20].A Brazilian population study presented a 6.3% reduction in spina bifida rates comparing 2001-2004 with 2006-2010 [22].In the United States, as well, the prevalence of neural tube diseases during the post-fortification period was stable [23].Similarly, in São Paulo State between 2004 and 2015, neural tube defects had a stationary pattern.Considering neural tube defects as the most prevalent CNS-CM subgroup, this finding may have contributed to the stationary pattern of CNS-CM prevalence showed in the present study.
CNS-CM prevalence was 10.4 per 10,000 live births in São Paulo State, slightly lower than 11.8 per 10,000 live births in the Brazilian Southeast, reported by a population-based study from 2001 and 2018 that included the period in which the Zika virus outbreak occurred in the country [8].The Zika virus outbreak significantly increased the rates of microcephaly in Brazilian states, especially in the Northeast region of the country [17], and may explain the difference observed in both studies.Furthermore, besides Zika, other viral infections are a significant cause of morbidity among infants, particularly in low-and middleincome countries [24].
Considering other low and middle-income countries around the globe, CNS-CM prevalence is variable according to the time-period included, definition of the studied congenital anomalies, quality ascertainment, and study type (hospital or population-based study) [7].The main CNS-CM prevalence rates found in low and middle-income countries are presented in Supplementary Table 5 [5-7, 25,26].Despite the stationary trend in the CNS-CM prevalence in general, São Paulo State showed an increasing prevalence of non-isolated CNS-CM during the study period.This finding possibly results from genetic causes of congenital anomalies not associated with folic acid supplementation, exposure to environmental contamination [27], and from other factors related to socioeconomic variables, in addition to improvements in diagnosis and notification [3,17].Increased qualification of health professionals [28] and availability of adequate equipment may have contributed to a higher sensitivity of the screening methods for less evident malformations [3,12].Therefore, advances in diagnosis and improved notification [3,17] possibly resulted in the diagnosis and reporting of minor anomalies associated with CNS-CM and/or in adequate reporting of complex multiple malformations [8,28], increasing prevalence of non-isolated CNS-CM.
From 2004 to 2015, CNS-CM neonatal mortality decreased around 1.7% per year in São Paulo State.This finding suggests the success of interventions to reduce neonatal mortality rates adopted by some lowand middle-income countries [29], such as improved access to surgical care [30], advances in neonatal and surgical techniques [21], and medical team and hospital staff qualification in treating infants with these anomalies [31].Despite improvements, the higher mortality rates reported for the non-isolated CNS-CM may reflect the impact of additional anomalies, considered as an independent predictor of lower survival during the first years after birth [32].Multiple congenital anomalies are associated with a high in-hospital mortality rate [33].
In the present study, CNS-CM neonatal mortality was associated with unfavorable maternal socioeconomical factors, lower gestational age, lower birthweight, and perinatal asphyxia.These variables were also observed as associated to other causes of neonatal mortality in São Paulo State [34,35].Difficulties of accessing specialized healthcare services may have contributed to the mortality in this group of infants [36].
The first few days and the first week after birth are critical for the survival of infants with CNS-CM [37].According to the International Clearinghouse for Birth Defects Surveillance and Research (2001-2012), newborns with spina bifida have mortality rates of up to 7.3% in the first day after birth and up to 10% during the neonatal period [38].In the present study, 50% of CNS-CM associated neonatal deaths occurred during the first 18h after birth.These deaths may have resulted from births outside referral hospitals [37].Accessibility to centers with optimal pediatric surgical care is difficult in middle income countries [30].In São Paulo State, the richest Brazilian State, health inequities have been reported, with discrepancies in access to qualified health care [39].
This study has some limitations.The database was provided by SEADE Foundation after linkage and anonymization, a process which relies on a time-consuming manual component; therefore, the most recent year available for the study was 2015.The database is originated from live birth and death certificates and there is a risk of information bias since it depends on diagnoses notification.Furthermore, the use of epidemiological data derived from civil records (birth and death certificates) does not include information on any clinical and/or laboratory antenatal, delivery care and post-natal maternal and neonatal data, not allowing the discrimination of the underlying cause of infants' death.Finally, the database used in the present study provided epidemiological information regarding live births with CNS-CM.Thus, the comparison with data from studies that include information on stillbirths and termination of pregnancy may be misleading.During 2010 and 2017, termination of pregnancy occurred in 38 to 41% of cases of fetal CNS-CM in countries where this practice is allowed [4,12].Consequently, in countries where termination of pregnancy is illegal, the congenital anomalies mortality rates appear to be higher [13].Despite these limitations, this study is one of the first population-based evaluations of the temporal trends of CNS-CM prevalence and neonatal deaths in a middle-income country.The results have the potential to contribute to the strategic planning of neonatal care for infants with CNS-CM in the State of São Paulo.However, further research is necessary to identify CNS-CM prevalence and outcome variation across the state.In Brazil, the reporting of congenital anomalies is mandatory; however, there is a clear spatiotemporal heterogeneity in anomaly reporting at the national level and, in most cases, this variability can be attributed to underreporting or incorrect registration of certain types of anomalies.Therefore, since 2019, the country has expanded its strategy for monitoring priority anomalies and has a dedicated technical area for the structuring and implementation of active surveillance of congenital anomalies in Brazil [17].
In conclusion, during a 12-year period in São Paulo State, Brazil, neonatal mortality rate of infants with THE JOURNAL OF MATERNAL-FETAL & NEONATAL MEDICINE CNS-CM in general and with isolated CNS-CM showed a decreasing pattern, suggesting some success in the public health strategies for the management of congenital anomalies adopted by the State.However, early mortality is still high and may be attributed to the inequality in access to quality healthcare services within the State.03748-7, however with no role in any step of the study and report (including study design, analysis and interpretation of data, in the writing of the report or in the decision to submit the paper for publication), which was authors' responsibility.

Table 1 .
Maternal and neonatal characteristics of isolated and non-isolated CNS-CM for infants alive at 27 days after birth and CNS-CM associated neonatal deaths.
CNS-CM: congenital malformation of central nervous system; n: total number available for the variable.p ¼ Pearson's chi-square test � p ¼ Two sample t-test.