Discussion

This study focused on the epidemiological description of CHD and the analysis of risk factors associated with the presentation of this congenital anomaly in 44 hospitals in Bogotá and Cali between 2002 and 2020.

The prevalence of CHD in this study was 19.36 per 10 000 LB (95% CI 18.23–20.55). This is considerably lower than other European countries reported in the EUROCAT, such as Spain, which had a prevalence of 79.37 per 10 000 live births between 2008 and 2017. ^{Reference Kinsner-Ovaskainen, Morris, Garne, Loane and Lanzoni9} Such variation could be due to the under-reporting and diagnosis of CHD cases in developing countries like Colombia (8). Although the prevalence was higher than some Latin American countries such as Chile, whose prevalence was 14.76 per 10 000 during the period 2017–2019. ^{Reference Orioli, Dolk and Lopez Camelo4}

According to Tassinari et al, during the period 2001–2014, in Bogotá the prevalence of CHD was 15.1 per 10 000 LB (95% CI 13.94–16.36), ^{Reference Tassinari, Martínez-Vernaza and Erazo-Morera8} and 11.55 (95% CI 11.01–15.08) in Cali, ^{Reference Orioli, Dolk and Lopez Camelo4} less than what was found in our study between 2015 and 2020; being 35.94 per 10 000 LB (95% CI 32.68–39.44). We found an increasing trend over the years regarding CHD prevalence, which is presumably due to the implementation of a vigilance programme as PREVERDEC. ^{Reference Tassinari, Martínez-Vernaza and Erazo-Morera8} However, the most frequent isolated CHD was non-specified, revealing the need to improve the prenatal and postnatal diagnostic mechanisms, optimising the specificity in the notification of this malformation. ^{Reference Eckersley, Sadler, Parry, Finucane and Gentles14}

Regarding birth size and weight, we found significant differences between cases and controls (p < 0.01) (Table 3). In spite of that result, these variables are not attributable as CHD risk factors, as they are evident after birth.

Several studies have shown that maternal age > 35 years is a risk factor for CHD, ^{Reference Persson, Razaz, Edstedt Bonamy, Villamor and Cnattingius11,Reference Persson, Cnattingius and Villamor15,Reference Miller, Riehle-Colarusso, Siffel, Frías and Correa16} this result was also found by us, especially for those aged > 45 years, excluding cases with Down syndrome (OR: 5.05; 95% CI 1.54–16.58). However, the findings differ to some extent in comparison to studies carried out in Europe, where this association was not statistically significant. ^{Reference Best and Rankin17–Reference Dolk, McCullough, Callaghan and Beyerlein19} We found that 82.4% (n = 28) of syndromic cases in mothers ≥ 36 years old correspond to Down syndrome, suggesting the relationship between these two variables. ^{Reference Bruckner, Singh, Lelong and Khoshnood20} The impact of this result on public health is high in spite of increasing the probability of CHD presentation, because in 2017 in Colombia, 10.9% (n = 71 712) of births were mothers over 35 years of age. ^{Reference Acevedo and Valle21}

A previous cohort study by Joinau-Zoulovits et al. showed that paternal age greater than 35 years increases the risk of developing CHD by 16%; ^{Reference Joinau-Zoulovits, Bertille, Cohen and Khoshnood22} while our study found that this risk occurs from an age > 45 years (OR: 1.64; 95% CI 1.12–2.38). This meta-analysis only considers paternal age without maternal age association for CHD presentation. We found that 58.3% (n = 21) of paternal > 45 years have a partner > 35 years, which increases the risk of presenting CHD as previously found; however, other explanations could be considered.

We found that gestational age ≤ 34 weeks was associated with increasing the risk for CHD (OR: 2,88; 95% CI 2.55–3.24). Similar results were obtained by Steurer et al for ≤ 38 weeks ^{Reference Steurer, Baer and Keller23} and by Chu et al for ≤ 30 weeks. ^{Reference Chu, Li, Kosinski, Hornik and Hill24} This could be attributable to instances of patent ductus arteriosus, which may arise as a result of prematurity. Even so, low gestational age cannot be addressed as a risk factor for CHD since it is evident after cardiac organogenesis, even though gestational time is an important factor in development. ^{Reference Persson, Cnattingius and Villamor15,Reference Cnota, Gupta, Michelfelder and Ittenbach25} Still, it is recommended that each premature newborn have an ideal surveillance and evaluation of possible congenital anomalies.

Additionally, we did not find any association between the sex of the newborn and the risk of developing CHD (OR: 0.99; 95% CI 0.87–1.14). However, there is heterogeneity in terms of the results obtained in other studies, since the sex of the newborn was a risk factor, whether it is male, ^{Reference Persson, Cnattingius and Villamor15} woman, ^{Reference Persson, Razaz, Edstedt Bonamy, Villamor and Cnattingius11} or both. ^{Reference Øyen, Diaz and Leirgul18} Although the results vary, it is important to note that differences are reported in terms of prognosis and mortality in patients with CHD according to sex, being higher for females. ^{Reference Aubry and Demian26} Regarding maternal BMI, different studies have reported an increased risk of presenting CHD for BMI < 18Kg/m², ^{Reference Yuan, Liu and Zhu27} when being overweight or obese, ^{Reference Persson, Razaz, Edstedt Bonamy, Villamor and Cnattingius11,Reference Persson, Cnattingius and Villamor15,Reference Liu, Ding and Yang28} similar to our results when the BMI ≥ 25Kg/m² (OR: 1.24; 95% CI 1.03–1.48). Yet, Dolk et al ^{Reference Dolk, McCullough, Callaghan and Beyerlein19} found no association for maternal BMI.

In this study, it is important to recall the association found between pre-pregnancy diabetes and the risk of presenting CHD (OR: 7.00; 95% CI 3.20–15.34), as demonstrated in different studies. ^{Reference Jenkins, Correa and Feinstein10,Reference Øyen, Diaz and Leirgul18,Reference Dolk, McCullough, Callaghan and Beyerlein19,Reference Chen, Yang and Chen29} Risk association for transposition of great arteries, septal defects and outflow tract defects was also observed. ^{Reference Jenkins, Correa and Feinstein10} A possible mechanism that explains the relationship between diabetes and the presentation of CHD was proposed by Engineer et al. Where pregestational diabetic patients develop oxidative stress induced by hyperglycaemia, increasing reactive oxygen species and free radicals, allowing cell damage and alteration in gene expression during cardiac morphogenesis. ^{Reference Engineer and Greco30} Some works try to reduce the risk through exercise, although it has not been shown as a protective factor in humans. ^{Reference Dolk, McCullough, Callaghan and Beyerlein19} On the other hand, in the same study by Dolk et al ^{Reference Dolk, McCullough, Callaghan and Beyerlein19} gestational diabetes was not a risk factor for CHD, according to our findings (OR: 0.83 95% CI 0.55–1.26); however, it was considered a risk factor for other authors. ^{Reference Wu, Liu and Sun12,Reference Chen, Yang and Chen29}

Different studies demonstrate that FA consumption in the pre-pregnancy period and during any trimester of pregnancy. ^{Reference Jenkins, Correa and Feinstein10,Reference Feng, Wang and Chen31} Though our findings highlight the importance of pre-pregnancy FA consumption for the prevention of CHD (OR: 0.20; 95% CI 0.05–0.81) and during the first trimester of pregnancy (OR: 0.43; 95% CI 0.38–0.50) allowing optimal levels of FA prior to the period of cardiac morphogenesis. In addition, this finding is supported by the risk of presenting CHD that is reported by the use of FA antagonists (OR: 7.7; 95% CI 2.8–21.7). ^{Reference Hernández-Díaz and Mitchell32} In Colombia, the resolution 3280 of 2018 of the Ministry of Health and Social Protection establishes the supplementation of FA and micronutrients 0.4mg/day in the gestational period, and for those mothers with history of births with congenital defects, a dose of 4mg/day has to be initiated 3 months before pregnancy. ³³

It is important to recognise the role of FA in the context of cardiac development, in that it is required for cell maintenance and division, as well as for the construction and repair of DNA. ^{Reference Czeizel, Dudás, Vereczkey and Bánhidy34} Therefore, a daily intake of 400 μg is recommended through diet or using supplements. ^{Reference Czeizel, Dudás, Vereczkey and Bánhidy34} In Colombia for 2011, 89 unintended pregnancies were estimated per 1 000 women of reproductive age, ^{Reference Prada, Singh, Remez and Villarreal35} so there is a public health problem regarding late multivitamin supplementation. Thus, the fortification of rice with FA is an important prevention strategy, for 2017 it was calculated that 35% of the rice was fortified. ^{Reference Rozo36}

The meta-analysis carried out by Yu et al showed that the low SS (OR: 1.05; 95% CI 1.01–1.09) and the mother’s low educational level (OR: 1.11; 95% CI 1.03–1.21) are risk factors for the presentation of CHD, ^{Reference Yu, Feng and Yang37} which agrees with our findings for low SS (OR: 1.27; 95% CI 1.04–1.55) and mother’s low educational level (OR: 1.22; 95% CI 1.01–1.48), respectively, similar findings in other study. ^{Reference Peyvandi, Baer and Chambers38} On the contrary, Dolk et al did not find a significant association for any SS, but the mother’s low educational level was associated with CHD (OR: 1.63; 95% CI 1.13–2.34). ^{Reference Dolk, McCullough, Callaghan and Beyerlein19}

In developing countries, low SS is a risk factor for CHD since it represents having less access to health services and education about contraceptive mechanisms and the preconception and prenatal care that must be taken. ^{Reference Adegbosin, Zhou, Wang, Stantic and Sun39} Additionally other factors are linked to early age pregnancies and thus with risk behaviours such as folic acid deficiency. Likewise, maternal malnutrition and exposure to teratogenic agents is reflected in low SS due to the scarcity of resources to live in optimal environments, ^{Reference Bronberg, Groisman, Bidondo, Barbero and Liascovich40} factors that are possibly associated with the appearance of CHD, still other explanations can be considered.

In Colombia, a mandatory maternal-perinatal health care route is established for all pregnant women and its costs are covered by the health system. It includes two prenatal ultrasounds, one between weeks 11 and 14 to evaluate aneuploidies, and another between weeks 18 and 24 to evaluate structural abnormalities. Factors such as late diagnosis of pregnancy, the socio-economic conditions of patients, and the lack of adherence to prenatal programmes influence the fact that this route is not fulfilled for some pregnant women. As we have previously described, prenatal detection of CHD is around 30% in Bogotá and Cali, and this also tends to increase under-reporting. ^{Reference García, Imbachí, Hurtado, Gracia and Zarante41}

The study’s main limitations are incomplete records within the surveillance and monitoring system, affecting the sample space for the different analyzes, as well as the memory bias that occurs in mothers on behalf it is a retrospective study.