Paola Soares Fernandes ·Larissa Rosario Magalh?es ·Tainara Rita Pezzini ·Edige Felipe de Sousa Santos ·Mauricio Giusti Calderon

Abstract Background Congenital heart diseases (CHD) are the most common type of birth defects,affecting millions of newborns every year;no prevalence data are available in S?o Paulo State,Brazil’s most populous state.The objective is to identify trends in prevalence and risk factors for CHD in S?o Paulo State,Brazil.Methods We developed a population-based study to estimate the CHD trend of prevalence in recent years,stratified by maternal age and geographical clusters,using all cases of CHD identified by the Live Births Information System (SINASC—Sistema de Informa??o Sobre Nascidos Vivos) from January 1,2005,to December 31,2018.We calculated the trends of prevalence,the regression coefficient (β),the annual percent change (APC),and 95% confidence interval using the Prais—Winsten regression model,with the Durbin—Watson test.Results We found 10,594 cases of CHD among 8,536,101 live births (LB),a prevalence of 12.4/10,000 LB.There was no difference in the sex distribution;they are primarily Caucasian (60.2%),75.2% born at term,and 74.4% weight ＞ 2500 g,66.9% of births were by cesarean section.S?o Paulo State presented an increasing CHD trend of prevalence (APC=18.9%).The highest CHD prevalence rate was in mothers aged ≥ 35 years (22.2/10,000 LB).There were 12,271 specific congenital heart defects among 10,594 patients (1.16 CHD/patient).Atrial septal defect has the highest number of cases (3835),with a prevalence of 4.49/10,000 LB,corresponding to 31.3% of all CHD.Conclusion CHD had an increasing prevalence trend in recent years,being highest in S?o Paulo City and ≥ 35-year mothers.

Keywords Congenital abnormalities ·Epidemiology ·Heart defects ·Interrupted time series ·Prevalence ·Public health

Introduction

The definition of congenital heart disease (CHD) is a functional or structural abnormality of the heart or great vessels at birth.These defects occur in the fetus as it develops in the uterus during pregnancy [1].There are several subtypes,such as atrial septal defect (ASD),ventricular septal defect(VSD),tetralogy of Fallot,patent ductus arteriosus,and pulmonary stenosis [2,3].The manifestations of congenital heart disease are variable and may occur soon after birth,childhood,or adolescence [4].

The heart is the organ most frequently affected by congenital anomalies globally,reaching millions of newborns every year[3,5,6].The world’s most accepted estimate of CHD incidence is 8 per 1000 live births (LB) [5,7].Approximately,20%—30%of CHD cases can be attributed to genetic or environmental cause [8,9].However,there are still uncertainties regarding the risk factors for 70% to 80% of the remaining cases [3].

Approximately,30,000 children are born with CHD per year,1% of all live births in Brazil [4,10],and about 80%need cardiac surgery,half of them in the first year of life [4,11].Congenital malformations are the second leading cause of mortality in children under one year of age [4,12].CHD is the most frequent and has the highest mortality in the first year of life,being the second cause of death within 30 days of life in Brazil [4].There are no data regarding CHD’s type frequency and prevalence trends in S?o Paulo State,Brazil.

This study aims to identify trends in prevalence and risk factors of congenital heart disease in a middle-income country like Brazil using the S?o Paulo State population database.

Methods

Study design

This is a population-based study with time trend [13],following Strengthening the Reporting of Observational Studies in Epidemiology [14] guidelines,using official microdata of all cases of congenital heart disease identified by the Live Births Information System (SINASC—Sistema de Informa??o Sobre Nascidos Vivos) in S?o Paulo state,Brazil,from January 1,2005,to December 31,2018,and using data from the Unified Health System Department of Informatics [15].

S?o Paulo State was chosen,for being the most populous state (41,262,199 in 2010) of Brazil [16],with a high birth rate (610,000/year) [15] and where the completeness of public data is more reliable [17,18].

Study population

We included all LB of mothers residing in S?o Paulo State,Brazil,whose information in the field of congenital malformations of the Live Births Declaration (LBD) was completed with CHD.We used the International Classification of Diseases 10th Edition [19] (ICD-10) from Q200 to Q249 identifying the cases of CHD at the LBD among all LB in the study period.

Outcomes

To describe the outcome of this study,sociodemographic and clinical explanatory variables were selected,such as gender,race/color,gestational age,birth weight,age,maternal education and occupation,type of pregnancy,type of delivery,and the number of prenatal consultations.SINASC provided the number of live births in the S?o Paulo State,and to develop the statistical analysis,they were stratified,year by year,according to territorial clusters and maternal age group,according to the methodology described by Calderon et al.[20].

Data source

Microdata were extracted from Unified Health System Department of Informatics’s (Departamento de Informática do Sistema único de Saúde,DATASUS) file transfer service,using TABNET and TABWIN programs;the files were converted from.dbf to.xls,and thus exported to Excel,on which variables were categorized.To minimize possible discrepancies,the data were collected by two different researchers independently.This database has been validated and previously used for public health and perinatal epidemiologic research and surveillance [20].

Bias

We used only the LB population to obtain the prevalence rates,and the proportion of maternal age range did not differ in the linear regression model by clusters in S?o Paulo State (

P

＞0.05).With this result,all the analyses can be used for comparison to others studies.

We identified some limitations in data collection,recording,and transmission inherent to the notification of health services,where there are fields with an incomplete filling,missing data,or ignored content.

Statistical analysis

Congenital heart disease prevalence rates were calculated for 10,000 LB by maternal age group and territorial clusters.For prevalence,in addition to the global period(2005—2018),we used two consecutive 7-year intervals(2005—2011 and 2012—2018).

For trend analysis,we used the Prais—Winsten regression model following the methodology proposed by Antunes and Cardoso [13],which allowed first-order autocorrelation corrections to be performed on values organized by time.Thus,we estimated the values of the angular coefficient (

β

) and respective probability (

p

),considering a 95% confidence interval (95% CI) significance level.

The data modeling process includes transforming prevalence rates into a base 10 logarithmic function using the Durbin—Watson test to measure the existence of the firstorder autocorrelation of the time series composed of the annual coefficients to verify that the correlation was compatible with the random regression residuals hypothesis[13].

The results of the logarithmic rates (

β

) of the Prais—Winsten regression allowed us to estimate the annual percentage change (APC) according to the methodology suggested by Antunes and Waldman [21].This procedure made it possible to classify the CHD prevalence rates as increasing,decreasing,or stationary.The trend is stationary when the coefficient was not significantly different from zero (

p

＞ 0.05) [13].

We use the 3rd-order central moving average technique to visualize trends and reduce white noise in time-series graphs [22] (Figs.1,2) and the Stata 15.1 statistical software (College Station,TX,2018) for all analyzes.

Data availability

The microdata used in this study are publicly available by SINASC,using data from DATASUS,maintained by the Ministry of Health of Brazil [15].DATASUS provides open public access to these data for any purpose (http://datas us.saude.gov.br/nasci dos-vivos-desde-1994).

Results

From 2005 to 2018,among 8,536,101 live births,10,594 cases of congenital heart disease were reported in S?o Paulo State,a prevalence of 12.4/10,000 LB (95% CI 12.2—12.6).There was no statistical difference in the distribution between the sexes;they are primarily Caucasian(60.2%);however,the Asian population presented the higher proportional prevalence (33.87/10,000 Asians).Although the majority of the patients were born at term,weighing ≥ 2500 g,proportionally,CHD affects more premature infants (49/10,000 from 28 to 31 weeks) and those with very low birth weight (56.29/10,000 with ＜1500 g).Despite 95.5% of single pregnancies,we observed a higher proportional CHD prevalence distribution in twins (22.81/10,000 twins),and 66.9% of births were by cesarean section.According to the maternal age group,25.3% of the mothers were high-risk pregnant women(22.15/10,000≥35 years),and 54.7% had between 8 and 11 years of schooling,but there is a higher CHD proportional prevalence in mothers with 12 or more schooling years (19.58/10,000 ≥ 12 years of schooling years)(Table 1).

The rate of CHD in S?o Paulo State had increased from 2.78/10,000 in 2005 to 21.56/10,000 in 2018 (overall,12.38 cases/10,000 LB;95% CI 12.2—12.6).The highest and lowest CHD case number and prevalence occurred in 2017 (1365,22.3/10,000) and 2005 (172,2.8/10,000).The highest and lowest annual CHD prevalence occurred in S?o Paulo City-SPC (27.1/10,000;95% CI 26.5—27.8) and in Taubaté Administrative Region-TAR (3.8 cases/10,000;95% CI 3.2—4.3).The highest and lowest CHD case numbers were in S?o Paulo Metropolitan Region (8508) and Baixada Santista Metropolitan Region-BSMR (169) (Supplementary Table 1,Fig.1).

Fig.1 Congenital heart diseases prevalence trend by cluster per 10,000 live births in S?o Paulo State,Brazil from 2005 to 2018.SPMR S?o Paulo Metropolitan Region,SPC S?o Paulo City,BSMR Baixada Santista Metropolitan Region,TAR Taubaté Administrative Region,CSC Central South Cluster,CRC Campinas Region Cluster,CNC Central North Cluster,NWE Northwest Cluster,SPS S?o Paulo State,w/o without

From 2005 to 2018,S?o Paulo State presented an increasing CHD prevalence trend,with an APC of 18.9%.In S?o Paulo State,the CHD prevalence trend increased in the first period (2005—2011),an APC of 29.1%,and was considered stationary in the second period (2012—2018).From 2005 to 2018 and 2005—2011,the highest APC was in SPC,24.6% and 50.7%,respectively;from 2012 to 2018,the highest APC was in BSMR (17%).Central North Cluster(CNC) had the lowest APC from 2005 to 2018 (4.7%) and 2011—2018 (3.9%),and Central South Cluster (CSC) from 2005 to 2011 (0.7%).We identified that BSMR was the only cluster with an increasing CHD trend of prevalence both in the global period 2005—2018 (APC=21.2%) and in the twointerval periods,2005—2011 (APC=27.4%) and 2012—2018(APC=17%) (Table 2).

In young mothers (＜ 20 years),the LB number with CHD increased by 1.081% from 2005 to 2018,whereas in older mothers (≥ 30 years),there was an increase of 1.018% in the same period.We identified a total number of 1103 and 5251 cases of LB with CHD in young and older mothers,respectively.The highest CHD prevalence rate was in mothers aged ≥ 35 years (22.2/10,000;95% CI 21.3—23.0).The overall CHD prevalence rate increased over the course of the study period (2005—2018) in all maternal age groups (Supplementary Table 2,Fig.2).The highest CHD prevalence rate increased in mothers aged 15—19 years (APC=24.6%;95% CI 13.6—36.6) (Table 3).

Fig.2 Congenital heart diseases prevalence trend by maternal age group per 10,000 live births in S?o Paulo State,Brazil from 2005 to 2018

There were 12,271 specific congenital heart defects among 10,594 patients (1.16 CHD/patient).ASD (ICD-10:Q211) has the highest number of cases (3835),with a prevalence of 4.49/10,000 (95% CI 4.35—4.63),corresponding to 31.3% of all CHD,followed by unspecified congenital malformation of heart (ICD-10:Q249) with 2338 (19.1%) cases and prevalence of 2.74/10,000 (95% CI 2.63—2.85) and VSD(ICD-10:Q210) with 1664 (13.6%) cases and prevalence of 1.95/10,000 (95% CI 1.86—2.04).The other group is formed by grouping 26 CHDs,all with ＜ 1% of proportional result,counting in total 1015 (8.3%) cases.Detailed CHD distribution is in Table 4.

Table 1 Maternal and infant sociodemographic characteristics of S?o Paulo State residents,Brazil from 2005 to 2018

Table 1 (continued)

Source:Live Births Information System (SINASC).Data:Unified Health System Department of Informatics (DATASUS,http://www.datasus.saude.gov.br).Ministry of Health,Brazil. live birth, congenital heart diseases,confidence interval

Table 2 Prais—Winsten regression model for trends in congenital heart disease prevalence by clusters in S?o Paulo State,Brazil from 2005 to 2018

Bold letters for significant results (＜ 0.05).Source:Live Births Information System (SINASC).Data:Unified Health System Department of Informatics (DATASUS,http://www.datas us.saude.gov.br ).Ministry of Health,Brazil. S?o Paulo Metropolitan Region,S?o Paulo City, Baixada Santista Metropolitan Region,Taubaté Administrative Region, Central South Cluster,Campinas Region Cluster, Central North Cluster,Northwest Cluster, S?o Paulo State, annual percentage change,confidence interval

Table 3 Prais—Winsten regression model for trends in congenital heart disease prevalence by maternal age in S?o Paulo State,Brazil from 2005 to 2018

Bold letters for significant results (＜0.05).Source:Live Births Information System (SINASC).Data:Unified Health System Department of Informatics (DATASUS,http://www.datas us.saude.gov.br).Ministry of Health,Brazil. annual percentage change,confidence interval

Table 4 Frequency of diagnostics,proportion and prevalence of congenital heart disease,by International Classification of Diseases-10th in S?o Paulo State,Brazil from 2005 to 2018 (per 10,000 LB)

Source:Live Births Information System (SINASC).Data:Unified Health System Department of Informatics (DATASUS,http://www.datas us.saude.gov.br).International Classification of Diseases, congenital heart disease,confidence interval, live birth

Discussion

To the best of our knowledge,this is the first populationbased study,using the SINASC-DATASUS database,which used 14 years of data from a validated data source,to describe the recent prevalence trend and characterize the LB population with CHD in S?o Paulo State,Brazil.

The prevalence of CHD in S?o Paulo State from 2005 to 2018 was 12.4 (95% CI 12.2—12.6) cases per 10,000 LB,and it significantly increased by 771.4% from 2005 [2.8(95% CI 2.4—3.2)] to 2018 [21.6 (95% CI 20.4—22.7)] cases per 10,000 LB,with an increasing APC of 18.9% (95% CI 12.9—25.1).This increasing prevalence of CHD in S?o Paulo State is concordant with other studies worldwide [2,3,7,23,24].This situation may be related to the CHD increase detection rates in all age groups,using cardiac and Doppler ultrasonography [24] and the pulse oximetry test in newborns [25].

From 2014 to 2015,it is possible to identify a significant increase in prevalence in the predominant rural clusters of the state,as CSC and CNC.In Northwest Cluster (NWE),we can only identify a significant increase in prevalence from 2017;the possible reasons for this finding are lower population,fewer public health centers,inadequate training of the health staff and early postnatal discharge.

We found no statistical difference in the distribution by sex in CHD;nevertheless,some studies have shown that males have a higher prevalence of CHD [2,23].Some authors found that male is an independent predictor of severe CHD because of the interaction of sex hormones on fetal development [26,27].

We found a higher proportional prevalence in the Asian population (33.87/10,000 born Asians) compared to the other races/ethnic.Since 1908,Brazil has received immigrants of Asian origin,mainly from Japan,while the movement of Chinese immigrants began in 1950.Currently,Brazil has the largest population of Japanese origin outside Japan [16] (about 1.5 million nikkeis).Due to the early language and communication difficulties with Brazilians,Asian immigrants have formed local communities,whether neighborhoods or cities,frequently enabling consanguineous interfamilial marriage,and this is a well-known factor that increases the possibility of expression of genetic diseases,as genetic CHD [28— 30].

From 2005 to 2018,we observed significant results of an increasing CHD prevalence trend in all clusters.In the BSMR,we identified an increasing CHD prevalence trend with APC=21.24% (95% CI 17.87—24.71),the only cluster showing an increasing prevalence trend both in the global period as the two-interval periods.All clusters from the central state to the northwest (CSC,Campinas Region Cluster,CNC,and NWE) presented increasing trends in the global period and stationary trends in the two-interval periods.An explanation for this situation is the lower 2011 prevalence and the higher 2012 prevalence,the inter-interval years(2011—2012),which can be attributed to the inclusion of the pulse oximetry test in newborns recommended since 2011 [25] and incorporated into the Brazilian public health system’s list of procedures in 2014 [31].

In SPC we identified the highest prevalence trend during 2005—2018 (APC=24.6%;95% CI 11.22—28.58)and during the period 2005—2011 (APC=50.71%;95%CI 30.11—74.57).From 2012 to 2018,we identified thehighest APC (APC=16.97%;95% CI 6.96—27.92) in BSMR.The possibility for this higher CHD prevalence in these clusters can be the difference in health and referral systems.Lack of resources,health insurance,screening programs,and referral systems likely lead to an underestimation of true birth prevalence [7].Regional differences can be explained by the different degrees of adherence to the registry and the asymmetry of highly differentiated resources,such as fetal echocardiography and the guidance of cases with prenatal diagnosis (PND) to pediatric cardiology services.Remote regions with a lower population density and low births may have less access to these resources,and thus,a lower number of PNDs for heart diseases [5].

Advanced maternal age is a risk factor for CHD [2,23,32].However,data on the effect of younger maternal age on the disease are limited [2].In Shenzhen,China,a study showed that maternal age ＜25 years reduced CHD risk [33].In addition,as an increasing number of women are postponing motherhood,maternal age has increased in recent decades,consequently causing a higher prevalence of congenital abnormalities at birth [7].Our results confirm that advanced maternal age is a significant risk factor for CHD in S?o Paulo State;once maternal age increases,the CHD prevalence rate also increases.The estimated risk of CHD increased in deliveries to older (≥35 years) women,corroborating the multicenter study by Zhang et al.[2].The highest CHD prevalence was in mothers aged ＞35 years (22.2/10,000 LB;95%CI 21.3—23.0),followed by 30—34 years (14.2/10,000 LB;95% CI 13.6—14.7),25—29 years (10.3/10,000 LB;95% CI 9.8—10.7),20—24 years (9.6/10,000 LB;95% CI 9.2—10.1),15—19 years (8.8/10,000 LB;95% CI 8.3—9.3) and ＜ 14 years(8.6/10,000 LB;95% CI 5.9—11.3).

From 2005 to 2018,the CHD prevalence trend by maternal age was increasing in all age groups.However,in both subperiods,maternal age groups ≤ 14 years and ≥ 35 years have stationary trends.We can also explain this using the lower/higher prevalence value situation in the inter-interval years (2011—2012).In other age groups,the prevalence trend is increasing in the first period and stationary in the second period.For mothers,≤ 14 years,and ≥ 35 years,from 2012 to 2018,the CHD stationary trend may be associated with rigorous prenatal care once those mothers are at higher risk of giving birth to a child with CHD [2].The

Manual of Primary Care

[34] (Ministry of Health,Brazil,2012) recommends that girls from 10 to 14 years must have differentiated prenatal care,childbirth,puerperium,and contraception.Also,prenatal care was recommended by the primary care team,especially for pregnant women under the age of 15 and over 35 years,as this age group should receive high-risk prenatal care.

ASD (ICD-10:Q211) was the most frequent CHD in S?o Paulo State (2005—2018),with 31.3% of cases and prevalence 3.8/10,000 LB,followed by unspecified malformation of the heart (ICD-10:Q249) with 19.1%,and VSD (ICD-10:Q210),with 13.6%.The other 39 CHD studied have a low frequency,ranging from 0.01 to 3.9%.In China,the CHD distribution is similar to S?o Paulo State,being ASD the most frequent,with 67.9%,followed by VSD with 16.4%[2].In Portugal,data show a reverse order of the CHD frequency;with 42%,the VSD is the most common,followed by ASD (15%) [5].A study in Quebec,Canada [24],which evaluated the prevalence of CHD from 1985 to 2000,showed that VSD is the most prevalent CHD in children,with 4.2/1000 children.ASD is the second most prevalent in Canadian’s children;however,it is the CHD with the highest growth rate,ranging from 1.2/1000 children in 1985 to 3.9/1000 children in 2000 [24].Liu et al.’s systematic review and meta-analysis show that ASD is the most marked estimated increase in prevalence among mild lesions,increasing sixfold from 1970—74 to 2010—17 [3].

A possible cause for this ASD increasing prevalence is the not distinguish ICD-10 coding type,which can range from a large septal defect (with atrial or ventricular dilatation) to an ASD of the fossa ovalis (with patent foramen ovale with tiny left-to-right shunt),or ASDs smaller than 5 mm (high rate of spontaneous closure) [6].Furthermore,because an ASD is usually asymptomatic and has murmurs that are often soft,these defects frequently do not lead to early diagnosis or referral [6].We believe that the pulse oximetry screening test and the echocardiography are responsible for the increasing prevalence,mainly in S?o Paulo State’s rural areas,between 2016 and 2018.

There are some limitations to this study.There may be a bias when recording the designations of congenital anomalies in the LBD to the public database.Another limitation is a large number of ICD-10 codes with non-specific terminology for the various diagnoses [3182 (25.93%)],perhaps due to the neonatal care team’s specific lack of knowledge of a prenatal diagnosis or the impossibility of carrying out an adequate postnatal diagnosis or the facility of being able to group several malformations in a single codification.This situation of diagnostic underreporting can be life-threatening in cases of complex malformations[35] and have increased cardiac,neurological,cognitive sequelae as well as social and economic losses [36] in cases of malformations without systemic repercussions.The importance of the national plan for assistance to newborns with heart disease and the pulse oximetry screening test is to modify the lack of diagnosis and improve the health assistance to these newborns.Also,we had no data on stillborn with CHD.

In the coming years,the impact caused by these public policies may be verified both in the increase in the number of diagnoses and consequently in the prevalence,and in the treatment of children with congenital heart disease,especially in predominantly low—middle-income rural regions such as the central-south,central-north and northwestern clusters of S?o Paulo State.It is stated that pulse oximetry screening in low—middle-income areas is a necessary and feasible test [36].

As the prevalence of CHD increases,it will be necessary to create qualified centers for diagnosis and treatment in these rural regions since the detection rate of CHD can be greatly affected by variations in definitions,selection criteria,diagnostic methods,and skills of physicians at the different centers [37,38].

In conclusion,our results show CHD as a public health problem,with an increasing prevalence trend in S?o Paulo State in recent years,being highest in SPC,and ≥ 35 years mothers.The highest overall prevalence was in SPC and the lowest in TAR.We found that ASD (ICD-10:Q211) is the most frequent type of CHD.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s12519-022-00543-3.

Author contributions

FPS designed and conducted the research,collected data,and wrote the manuscript.MLR and PTR conducted the research,collected data,and wrote the manuscript.SSEF analyzed data and conducted critical analysis.CMG designed the research,analyzed data,conducted critical analysis,and wrote the manuscript.All authors read,reviewed,and approved the final manuscript.

Funding

This research did not receive funding from agencies in the public,commercial,or not-for-profit sectors.

Data availability

The data used for this study are administered by the Live Births Information System (SINASC—Sistema de Informa??o sobre Nascidos Vivos),using data from the Unified Health System Department of Informatics (DATASUS—Departamento de Informática do Sistema único de Saúde),maintained by the Ministry of Health of Brazil.DATASUS provides open public access to these data for any purpose (http://www.datas us.saude.gov.br,http://datas us.saude.gov.br/nasci dos-vivos-desde-1994 ).

Declarations

Ethical approval

The present study involves only the description and analysis of secondary population data obtained from the general population census and birth data.No additional information that is not freely accessible was collected.In particular,no information with individual identification was obtained for this study.Ethics approval was not required.

Conflict of interest

No financial or non-financial benefits have been received or will be received from any party related directly or indirectly to the subject of this article.The authors have no conflict of interest to declare.