Unveiling the Hidden Burden: An Umbrella Review of Congenital Anomalies among Newborns in Low- and Middle-Income Countries Free

Congenital anomalies (CAs), as defined by the WHO, are structural or functional abnormalities present at birth, including metabolic disorders. They contribute significantly to mortality in the first month of life and can result in long-term disabilities, affecting individuals, families, and healthcare systems, particularly in developing countries [1]. In 2010, the WHO reported that approximately 270,000 neonatal deaths worldwide were caused by CAs. Among these anomalies, neural tube defects were identified as the most severe and prevalent [2].

The EUROCAT network reported a perinatal death rate of 9.2 per 10,000 births due to CAs. In Italy, the Ministry of Health estimates that 25,000 out of 500,000 annual births have at least one CA, contributing to about 25% of infant mortality. In the USA, the CDC reports that approximately 3.3% of live births are affected by severe birth defects [3]. CAs in low- and middle-income countries (LMICs) account for 94% of global cases and are a leading cause of under-five mortality. Sub-Saharan Africa and Southern and Central Asia bear 80% of this burden [3, 4]. Developed countries have reduced CAs through prenatal counseling, screening, and pregnancy termination options. In contrast, resource-limited countries often lack organized prenatal screening and diagnosis systems. This deficiency, along with insufficient data on prevalence and outcomes, hampers recognition of CAs as a significant public health challenge [5].

Approximately 40–60% of the risk factors associated with CAs remain uncertain, but certain factors have been identified, including genetic factors, environmental factors, and multifactorial inheritance [6]. Environmental factors during embryogenesis that have been linked to CAs include maternal infection, age, drug use, and exposure to substances like caffeine and nicotine. Additionally, maternal nutritional status, health, exposure to hazardous waste, and alcohol consumption also play a role [7]. Additionally, maternal chromosomal abnormalities can lead to genetic disorders, and single gene mutations also contribute to genetically related defects [8].

However, it is widely acknowledged that many birth defects can be prevented through simple measures, such as taking folic acid supplements, practicing infection prevention, and avoiding certain behaviors like smoking, alcohol consumption, and the use of certain drugs during pregnancy [9‒11]. The WHO, in collaboration with various organizations, has underscored the importance of primary prevention and improving the health of children with CAs through recommended strategies including developing expertise in preventive measures, strengthening registration and surveillance systems, and enhancing research on the causes and diagnosis of CAs [1, 2]. Early identification of maternal and neonatal risk factors and accurate assessment of the prevalence of CAs in a given population are crucial for estimating their impact, highlighting the need for prevention, and facilitating the development of public health policies and treatment services [5].

Despite the recommendations made, CAs continue to be a significant public health issue, particularly in LMICs [1, 2, 12]. To this date, seven systematic reviews and meta-analyses (SRMs) have revealed inconsistent prevalence rates of CAs among newborns in these countries, ranging from 2.3 [13] to 23.5 [14] per 1,000 births with varying degrees of quality score in LMICs. Notably, the quality scores of the studies that were included in these SRMs varied. Likewise, there was inconclusive reporting on the risk factors and patterns of CAs across these studies [14‒16]. The heterogeneity of information on CAs can complicate intervention and decision-making for clinicians and users. This umbrella review aims to consolidate findings from systematic reviews on CAs into a comprehensive document for better comparison. To the authors’ knowledge, it is the first review focusing on the prevalence, patterns, and associated factors of CAs in LMICs. The evidence presented will assist clinicians and child health policymakers in designing evidence-based public health responses, ultimately aiming to reduce CAs and improve clinical practices, supporting the goal of reducing preventable neonatal mortality by 2030.

The authors conducted a thorough review based on the systematic methods described in the umbrella review [17], focusing on the prevalence, patterns, and associated factors of CAs among newborns in LMIC. We have registered the protocol in PROSPERO and obtained a registration ID of CRD42023484573. Furthermore, to ensure consistency and adherence to standards, the 2020 PRISMA checklist was utilized during the review process [18].

The objective of the review was to combine systematic review and meta-analysis (SRM) studies in order to get a single pooled estimate of the prevalence, pattern and associated factors of CAs among newborns in LMIC. What is the pooled prevalence of CA and its patterns among newborns in LMIC, and what are the key determinants of CAs among newborns in LMIC?

A systematic search was conducted on PubMed, Embase, HINARI, Science Direct, CINAHL, Scopus, PsycINFO, Web of Science, and the Cochrane Library by authors A.B.Z. and A.W.A. Using keywords and MeSH terms, the search employed “AND” and “OR” Boolean operators. Reference lists of included SRM studies were also reviewed to identify additional relevant studies. The search period spanned September 1 to December 31, 2024, without publication date restrictions. Full articles were assessed for eligibility, with disagreements resolved through discussion. The strategy combined medical subject terms, keywords, and free text.

Key terms included “Congenital anomaly AND prevalence AND associated factors AND meta-analysis.” Search string included (“Congenital anomaly”[MeSH Terms] OR (“congenital anomaly”[All Fields] OR (“congenital malformation”) OR (“birth defects”) OR (“congenital malformations”) AND (“meta-analysis” [Publication Type] OR “meta-analysis as topic”[MeSH Terms] OR “meta-analysis”[All Fields]) AND (“proportion”) OR (“prevalence”) AND (“pattern”) AND (“determinants”) OR (“risk factors”) OR (“associated factors”) AND (“newborns”).

Publications until 31 December 2024 were eligible for inclusion. The following predefined criteria were considered for a study to be regarded as a systematic review with meta-analysis: (a) it presented a defined literature search strategy, (b) it appraised its included studies using a relevant tool, and (c) it followed a standard approach in pooling studies and providing summary estimates. Additionally, publication in the English language was a further inclusion criterion.

Studies were excluded due to any of the following reasons: (a) no report on either the prevalence or determinants of CAs for this study, (b) published other than English language, and (c) article without a full text and abstract. Moreover, literature reviews that did not have a defined research question and search strategy or a defined process of selecting articles were excluded.

Duplicates were removed using EndNote version 20 software [19]. Three authors (A.B.Z., T.A.K., and H.S.N.) conducted screening and data extraction. Initially, studies were assessed by titles and abstracts, followed by full-text evaluations to identify eligible studies. Data were extracted using a predefined template, capturing the first author’s name, sample size, study location, publication year, number of studies, birth prevalence of CAs, and associated risk factors. Prevalence rates were standardized to per 1,000 births. Key variables evaluated included supplemental folic acid use, maternal illness history, drug use during pregnancy, and khat chewing.

The Assessment of Multiple Systematic Reviews (AMSTAR) tool was used to evaluate the methodological quality of SRM studies, as recommended by the WHO due to its reliability and validation. AMSTAR consists of 11 items that assess methodological rigor, with each item scored as “yes,” “no,” “cannot answer,” or “not applicable.” The maximum score is 11, with scores categorized as follows: 0–4 indicates low quality, 5–8 indicates moderate quality, and 9–11 indicates high quality [20], respectively. Therefore, two authors (M.A.B. and H.S.N.) assessed the risk of methodological bias observed in each article and the results were checked by the principal investigator (A.B.Z.) and any discrepancies were resolved by discussion with the review group.

Both SRM methodologies were employed for the analysis using Stata version 17 software [21]. Quantitative reviews established the overall pooled prevalence, patterns, and potential risk factors of CAs in LMICs. Heterogeneity among studies was assessed using Cochran’s Q statistic and I² test, with scores of 0, 25, 50, and 75% indicating no, low, moderate, and high heterogeneity, respectively [22]. Due to observed moderate heterogeneity, a DerSimonian-Laird random-effects model estimated the pooled effect size. Publication bias was examined using funnel plots, Egger’s test, and Begg’s test, with a p value under 0.05 considered significant. Subgroup and sensitivity analyses were conducted to explore study characteristics and their influence on results. Meta-regression analyzed potential heterogeneity sources.

A total of 67 studies on the prevalence, patterns, and associated factors of CAs in LMICs were identified. After removing 26 duplicates, 22 articles were deemed irrelevant based on title and abstract assessments. The remaining 19 articles were reviewed in full text for eligibility according to predetermined criteria. Of these, 5 articles were excluded due to language issues, and 7 were excluded for not reporting the outcome of interest. Ultimately, 7 SRM studies were included in this umbrella review (Fig. 1). Among the eligible studies, 7 were used to assess the overall prevalence of CAs, 4 for the pooled prevalence of different subtypes (patterns), and 3 for investigating associated risk factors.

In the current umbrella review, we have included 7 SRM studies [13‒16, 23‒25] for determining the pooled prevalence of CAs with a total of 171 primary studies providing a sample size of 4,763,268 newborns in LMIC. Four studies were conducted in middle-income countries [13, 15, 23, 24] whereas the rest three were conducted in low-income countries [14, 16, 25]. The sample size ranged from 95,755 in Ethiopia [25] to 1,386,408 in Iran [15]. The highest pooled prevalence was 23.5 per 1,000 in Africa [14] while the lowest was 2.3 per 1,000 births in Iran [13], which is highly inconsistent and showing us the need of conducting umbrella review for better decision-making. The number of primary studies included by the SRM that used in the current umbrella review ranged from 10 [25] to 55 [24]. Regarding year of publication, two studies were published in 2017 [13, 15]; the other two studies in 2018 [23, 24], a single review study in 2020 [16], and the rest two studies were published in 2023 [14, 25]. Regarding the quality assessment, three SRM studies included in the umbrella review were used NOS [15, 16, 25], two studies used STROBE [13, 23], and the rest two studies used JBI [14, 24] and they all included only those primary research articles which had a moderate and high quality score. Four published studies from three middle-income countries [13, 15, 23] and one from low income [16] were used for analysis of pooled prevalence of the subtype of CAs. For determining the pooled odds ratio of the associated factors of CA three SRM studies were used [14, 16, 25]. Online supplementary Table 1 (for all online suppl. material, see https://doi.org/10.1159/000543832) summarizes the characteristics of the SRMs included in this umbrella review.

Authors conducted the appraisal independently, using a standardized form [26] and it was found that ranged from 9 to 11, with a mean score of 10 points, indicating an overall high quality (online suppl. Table 2).

A combined analysis of 7 SRM studies, with 4 studies conducted in middle-income countries and three in low-income countries, indicated that the current study pooled prevalence of CAs was 15 per 1,000 births (95% CI: 9.0, 21.0) (Fig. 2). However, significant heterogeneity was observed among the studies, as evidenced by an I² statistic of 73.8% and a p value of less than 0.001 for Cochran’s Q test.

A subgroup analysis was conducted based on the World Bank's economic classification of the countries where the included studies were conducted. The analysis found that the highest prevalence of CAs was in low-income countries, with a pooled prevalence of 18 per 1,000 births (95% CI: 8.0, 27.0). Middle-income countries had a slightly lower prevalence of 14 per 1,000 births (95% CI: 6.0, 22.0).

Additionally, based on sample size indicated a prevalence of 10 per 1,000 newborns (95% CI: 1.00, 24.0) among studies with fewer than 700,000 samples. Further subgroup analyses were also conducted based on the number of primary studies included and publication year (Fig. 3).

In this umbrella review, an analysis was performed to identify the sources of heterogeneity by examining several factors: the number of primary studies included (p = 0.982), publication year (p = 0.541), sample size (p = 0.277), and study countries (p = 0.541). The results indicated that none of these factors were statistically significant as sources of heterogeneity.

A sensitivity analysis was conducted using the random-effects model to evaluate the impact of individual studies on the overall estimate of CAs. The findings indicated that no single study significantly influenced the overall estimate of CAs among newborns in LMICs (online suppl. Fig. 1).

To assess the potential presence of publication bias among the studies, the researchers visually examined the funnel plot and performed Begg’s correlation and Egger’s regression tests, analyzing the p values. The visual inspection of the funnel plot revealed a symmetrical distribution, suggesting an absence of publication bias. Additionally, both Begg’s correlation and Egger’s regression tests showed no statistically significant publication bias, with p values of 0.448 and 0.801, respectively (Fig. 4).

We have included four studies for determining the pattern of CAs [13, 15, 16, 23]. Table 1 summarizes the pooled prevalence of each subtype of CAs.

A total of 4 SRM studies, three from middle-income countries [13, 15, 23] and one from low-income countries [23] were included and the overall pooled prevalence of these disorder was 8 per 1,000 births (95% CI: 0.00–15.00). The smallest and largest sample size belonged to the study by Zahed Pasha et al. [15] on 108,449 newborns and Adane et al. [16] on 646,249 newborns, respectively.

Four studies [13, 15, 16, 23] revealed that the pooled prevalence of these defects was 3 per 1,000 births (95% CI: 1.00–6.00). The largest sample size belonged to the study by Adane et al. [16] on 1,301,516 newborns. The lowest and highest prevalence rate of these defects was 2.98 per 1,000 births [16] and 11.26 per 1,000 births [13], respectively.

Four SRM studies [13, 15, 16, 23] were included and the pooled prevalence rate of these abnormalities was 3 per 1,000 births (95% CI: 0.00–6.00). The smallest sample size and smallest prevalence were reported by Vatankhah et al. [13] and Daliri et al. [23], respectively.

Two studies [13, 23] showed that a pooled prevalence of these defects was 1 per 1,000 births (95% CI: −3.00–4.00). The highest sample size was reported by Vatankhah et al. [13] which was 215,756 newborns.

Among all studies included in this umbrella review, 4 studies involving 1,154,647 newborns reported that the pooled prevalence of gastrointestinal system defects was 2 per 1,000 births (95% CI: −1.00–4.00). The highest and the lowest prevalence of these defects was 6 per 1,000 births [13] and 1 per 1,000 births [15], respectively.

Three studies [15, 16, 23] having a total 25,108,449 study samples showed that the pooled prevalence of these defects was found 1 per 1,000 births (95% CI: 0.98–2.00). The sample size of the included studies ranged from 45,615 newborns [23] and 24,191,670 newborns [16].

Four studies involving 1,306,551 participants revealed that the pooled prevalence of urogenital defects was 4 per 1,000 births (95% CI: −1.00–9.00). The highest and lowest prevalence rate were reported as 16 per 1,000 births [13] and 1 per 1,000 births [16], respectively.

Four studies with a sample size ranged from 69,682 newborns to 570,307 newborns revealed that the overall prevalence of these defects was found to be 1 per 1,000 births (95% CI: −1.00–3.00).

Three studies with a total of 704,569 samples were pooled and obtained a prevalence of 1 per 1,000 births (95% CI: −1.00–4.00).

Various factors including folic acid supplementation, maternal illness, unidentified drug use, and khat chewing were examined for their potential association with CAs. The findings, including pooled odds ratios (OR), confidence intervals (CI), and first author details, are presented in Table 2. Specifically, the associations between maternal folic acid supplementation, maternal illness, and maternal history of medication during pregnancy with the occurrence of anomalies were reported in three studies [14, 16, 25] whereas the association between kchat chewing and CAs were reported by two studies [14, 25]. Therefore, the study findings indicated that newborns whose mothers did not receive folic acid during pregnancy had a 4.18 times higher likelihood of having CAs compared to those who received folic acid (pooled OR, 4.18; 95% CI: 2.35, 6.02). Additionally, the presence of maternal illness during pregnancy was associated with a 2.19 times increased odds of CAs in newborns (pooled OR, 4.93; 95% CI: 3.85, 6.01). Furthermore, newborns born to mothers with a history of unidentified drug use or medication during pregnancy had a 7.54 times higher likelihood of having CAs compared to their counterparts (pooled OR, 2.83; 95% CI: 2.17, 3.49). Additionally, newborn infants born to mothers who chewed khat were 3.5 times more likely to have CAs compared to their counterparts (pooled OR, 3.5; 95% CI: 2.97, 4.03).

This umbrella review is the first to comprehensively investigate the prevalence, patterns, and factors associated with CAs among newborns in LMICs. Despite seven prior SRMs, findings have been inconsistent, complicating policymaker decisions. The primary aim was to summarize existing evidence on these CAs in this context.

The analysis of CAs prevalence among newborns in LMIC yielded a pooled estimate of 15 per 1,000 births (95% CI: 9.00, 21.00). Significant heterogeneity was found, with an I² value of 73.8% and a p value of less than 0.001. Subgroup analyses were conducted to identify sources of this heterogeneity, focusing on study countries, sample size, year of publication, and the number of primary studies.

The current study finding aligns with a study conducted in Britain, which reported a prevalence of 12.9 per 1,000 births [27]. However, our findings indicate a higher prevalence compared to the global report by the World Health Organization (WHO), where low-income and middle-income countries accounted for 94% of CAs cases, corresponding to 6 per 1,000 births [28]. This disparity could be attributed to several factors in LMIC, including limited access to nutritious diets for pregnant women, higher exposure to illnesses and substances such as alcohol and khat, and limited access to healthcare and prenatal screenings [28]. On the other hand, the current study suggest a lower prevalence compared to study conducted in Europe, which reported that 23.9 per 1,000 births [29]. This difference may stem from better birth registry systems, service quality, and government commitment in developed countries. Socioeconomic disparities such as access to healthcare, infrastructure, education, and media coverage of maternal health services could also explain the discrepancy. Additionally, the likelihood of data loss may be higher in LMIC than in Europe, contributing to lower reported prevalence [5, 25]. In addition, the alignment of findings between the current study and the British study can be attributed to methodological similarities and shared risk factors, despite the inherent differences between the two contexts in terms of geography and socio-economic status [27, 30].

This umbrella review found that the most prevalent CAs were musculoskeletal defects, at 8 per 1,000 births, followed by urogenital system defects at 4 per 1,000 births. Central nervous system and cardiovascular defects each had a prevalence of 3 per 1,000 births. This pattern aligns with a study from southern Vietnam, which identified musculoskeletal system defects as the most common birth defects [31]. This could be attributed to the similarity of risk factors for these specific types of birth defects across different countries.

This study found a significant association between various factors and CAs. Mothers who did not receive folic acid supplementation had 4.18 times higher odds of having a child with CAs compared to those who did. This aligns with clinical evidence emphasizing folic acid's role in DNA replication and fetal growth. Supplementation at conception is known to reduce the likelihood of neural tube defects in children [32‒34]. Furthermore, evidence has indicated that adequate intake of folic acid can lower the risk of significant neural tube defects in infants by at least 50% [33]. Hence, there is a need to routinely supplement a folic acid during the preconception and pregnancy period.

Mothers of newborns who have a medical history of illnesses are 4.93 times more likely to have infants with CAs. This finding is consistent with a study conducted in Canada [35]. Therefore, pregnant mothers must receive proper medical treatment to reduce the risk of CAs in their unborn children. To lower CA prevalence, stakeholders should focus on early screening, diagnosis, and treatment of maternal illnesses, along with raising awareness through maternal conferences on the benefits of early health screening. Additionally, infants born to mothers with a history of unidentified drug use during pregnancy are 4.37 times more likely to experience CAs compared to those without a history of drug use. This finding is supported by multiple studies, which indicate that a significant proportion of neonatal congenital abnormalities stem from unidentified medication use during pregnancy [32, 36, 37]. This might be due to many drugs have a teratogenic effect, especially if it is taken during organogenesis time. Thus, it is advisable for pregnant women to consult their healthcare provider before taking any drugs and avoid a habit of taking un-prescribed medications.

Infants born to mothers who chew khat are 3.5 times more likely to have CAs. Khat contains psychoactive substances like cathinone and cathine, which can cross the placental barrier, reducing blood flow to the fetus and potentially causing growth restrictions and other negative outcomes [38, 39]. Hence, pregnant women should avoid a habit of chewing khat to mitigate the risk of CAs in their developing fetus.

This study has several strengths, including comprehensive searches across multiple databases and the involvement of three researchers in the selection process, which reduces bias. A large sample size enhances result accuracy. However, limitations must be acknowledged, such as the restriction to English-language articles and significant overall heterogeneity that was not fully addressed. Additionally, the over-representation of certain studies and a focus on specific CA subtypes from Iran may introduce bias and affect the overall conclusions.

This study provides summarized evidence amid inconclusive systematic reviews and is the first umbrella review in LMIC on CAs in newborns. Its findings can guide clinicians and policymakers in improving health outcomes. Future research should explore additional risk factors through SRMA studies. Strengthening primary healthcare and integrating preconception care and prenatal screening into maternal health services are essential for effective intervention and prevention.

The overall prevalence of CAs in LMIC was determined to be 15 per 1,000 births, which is notably higher than the global report provided by WHO [40]. The most common CAs were musculoskeletal defects, followed by urogenital, central nervous system, and cardiovascular defects. Significant associations were found with maternal illness, unidentified drug use, khat chewing, and lack of folic acid supplementation. These findings emphasize the need for coordinated preconception care and prenatal screening for fetal anomalies. Addressing the burden of CAs requires prioritizing global research on interventions targeting these risk factors.

The authors would like to thank all their colleagues for their unreserved help during the write up of this review, and they would also like to thank all authors of the primary research used on this umbrella review.

Ethical approval is not applicable because this study is based exclusively on published literature.

The authors declare no competing of interests with respect to the finances, authorship, and/or publication of this article.

There was no funding for this research work.

A.B.Z. conceived and designed the protocol. M.A.B., B.D.T., G.Y., M.A., A.W.A., A.B.Z., and T.A.K. wrote the first draft of the protocol. R.N.H., H.S.N., and A.B.Z. designed a search strategy and conducted the quality assessment. R.N.H., H.S.N., and M.A.B. read and approved the final protocol and conducted data extraction and analysis. All authors wrote and developed the protocol, were involved in the final write up of the umbrella review, and read and approved the final draft of the manuscript.

The data sets used during the current umbrella review are freely available and uploaded as a supplementary file.