Patterns and risk factors of birth defects in Bangladesh: evidence from a neonatal intensive care unit Authors: Syeda Humaida Hasan, Humayra Akter, Jannatul Ferdous, Tamanna Zahur, Kamrun Nahar, Jagadish C. Das, Tahmina Banu Published: 2026 ABSTRACT IntroductionBirth defects constitute substantial contributor to neonatal mortality and morbidity particularly in low and middle income countries. In Bangladesh, limited screening and low awareness contribute to delayed diagnoses and inadequate care. This study aimed to delineate the patterns of birth defects and identify associated risk factors to provide baseline data in a tertiary care neonatal intensive care unit in Bangladesh to support future surveillance, inform policy development, and guide preventive strategies.MethodsThis hospital based prospective observational study was conducted in the Department of Neonatology at Chittagong Medical College Hospital, Bangladesh, from August to December 2020. Neonates with birth defects were identified through detailed clinical examination and classified according to the International Classification of Diseases, 10th Revision, Clinical Modification. Relevant data were collected using a predesigned questionnaire. Data were analyzed in SPSS version 28. Logistic regression was done to identify perinatal risk factors. A p-value < 0.05 was considered statistically significant.ResultsAmong 5,223 admitted neonates, 183 (3.5%) had birth defects. Musculoskeletal anomalies were most common (67 cases), followed by cleft lip/palate (30), circulatory (29), chromosomal (24), eye, ear, face and neck (23) anomalies. Among these neonates, 56 (30.6%) died, with a 1.59 fold higher mortality risk compared to those without defects (95% CI: 1.15–2.18; p = 0.005). Maternal hypertension (AOR: 3.92; 95% CI: 2.56–5.99; p < 0.001), diabetes mellitus (AOR: 4.15; 95% CI: 2.37–7.27; p < 0.001), polyhydramnios (AOR: 7.33; 95% CI: 2.63–20.41; p < 0.001), and oligohydramnios (AOR: 10.19; 95% CI: 3.68–28.23; p < 0.001) were strongly associated with birth defects.ConclusionsThe study highlights the significant burden of birth defects and associated neonatal mortality. Musculoskeletal anomalies were the most frequently observed defects. Maternal hypertension and diabetes mellitus were identified as strong predictors, warranting targeted antenatal screening and timely intervention. Routine neonatal examination at community level as well as in hospital is recommended to facilitate early detection and proper referral of birth defects. Introduction Birth defects are a global concern, but their impact in terms of burden and mortality is disproportionately severe in low-and middle-income countries (LMICs) [1, 2]. They account for approximately 8% of global under five deaths and 10% of neonatal deaths representing a significant cause of child mortality, with 94% of serious birth defects and 95% of related deaths occurring in LMICs [3]. The incidence and pattern of birth defects vary across regions, and this variation is influenced by economic status, maternal healthcare, education, and surveillance capacity. The World Health Organization (WHO) estimates the prevalence of birth defects at 47.2, 55.7, and 64.2 per 1,000 live births in developed, middle-income, and low-income countries, respectively highlighting a clear gradient of increasing burden with decreasing income level [4]. Congenital anomalies or birth defects, are structural or functional abnormalities that arise during intrauterine life and may be identified prenatally, at birth, or later in life. “Birth defects,” “congenital abnormalities,” and “congenital malformations” are often used interchangeably [5]. Globally, congenital anomalies are responsible for an estimated 295,000 neonatal deaths annually, and more than 50% of disability-adjusted life years (DALYs) lost each year could be prevented with timely surgical or other appropriate interventions [6]. As under-five mortality declines in LMICs due to improved control of infectious diseases, birth defects are emerging as a significant contributor to child mortality in these settings [7, 8]. Strengthening birth defect surveillance and management has become increasingly important for further reductions in neonatal and infant mortality [3]. Despite this high burden, current estimates likely represent only the tip of the iceberg. The true burden remains substantially underestimated in many LMICs, including Bangladesh, where no national surveillance system for birth defect screening or population-based registry currently exists. Where registries do exist in LMICs, detection is predominantly facility-based and clinician dependent, with very few community-based surveillance systems in place [9]. In Bangladesh, previous hospital-based studies have described the pattern of birth defects; however, most were limited to specific anomaly groups or departments and lacked comprehensive risk factor and outcome analysis. A 12-year retrospective study conducted at the study center reported gastrointestinal and genitourinary anomalies as the most common among pediatric surgical patients, but excluded non-surgical anomalies and did not use standardized ICD-10 classification [10]. A more recent study using ICD-10 classification identified predominantly musculoskeletal and central nervous system anomalies, although multivariate risk factor analysis was not performed [11]. Overall, the existing literature is characterized by methodological limitations, inconsistent use of standardized classification systems, and absence of concurrent risk factor analysis. Studies reporting birth prevalence of congenital anomalies in LMIC settings are further subject to significant methodological biases including sampling, information, observer, and reporting biases, collectively contributing to systematic underestimation of the true burden [12]. There is therefore a significant research gap regarding a comprehensive assessment of the patterns, associated risk factors, and clinical outcomes of birth defects in Bangladesh. Although hospital-based data cannot replace population-based surveillance, as many neonates with birth defects are born in and referred to tertiary care hospitals, data from such facilities may partially reflect current scenario. The World Health Organization recommends hospital based surveillance in high volume centers as an appropriate initial approach in settings where population based surveillance has not yet been established [13]. Therefore, this study aimed to describe the patterns and associated perinatal risk factors of birth defects, assess their clinical outcomes, and generate baseline evidence to support future surveillance and preventive strategies in Bangladesh. Methods Study area This hospital-based, prospective observational study was conducted in the Department of Neonatology, Chittagong Medical College Hospital (CMCH) from August 2020 to December 2020. This department serves as a referral center for the entire Chattogram division and handles an average of 10,000 to 11,000 neonatal admissions per year with a high bed occupancy rate, ensuring a large and diverse sample. The average number of admitted neonates is 150–180 per day in 100 beds facility. The Chattogram division comprises a total area of 33,909.00 square kilometers. The unit provides advanced neonatal care, including neonatal intensive care services (Level III) for critically ill neonates referred to this center from private Neonatal Intensive Care Units (NICUs), district hospitals or remote areas across the division [14]. Study population The study population comprised neonates admitted with birth defect to the “Neonatology Department” during study period. Consecutive sampling technique was adopted to minimize selection bias, ensuring inclusion of all eligible cases presenting to this tertiary referral. Both inborn and outborn neonates were included. Stillbirths were excluded from the study as no systematic data on stillbirths with birth defects were available within the study setting. Neonates with congenital heart diseases (classified under circulatory system anomalies) were included in the study only if confirmed by echocardiography. Although many admitted neonates had suspected congenital heart disease based on clinical features, and bedside oxygen saturation measurements, only confirmed cases were considered to ensure diagnostic accuracy. All enrolled neonates were observed until definitive clinical outcome. Process of data collection During the study period, information on all neonatal admission was documented at admission and updated at discharge. Each neonate was registered only once at the time of first admission and repeat admissions were excluded to avoid duplication. For neonates with birth defect, detailed information were collected using a pre-designed questionnaire. Detailed demographic and clinical information including age, gender, gestational age, birth weight, residence and mode of delivery was obtained. Maternal information including age, parity, underlying medical conditions, and previous obstetric history was also recorded from the parents or caregivers. Each newborn underwent a thorough clinical examination performed by attending physician at the time of admission. Clinical findings were documented and photographs of birth defect were captured after obtaining informed consent from caregivers. All cases were subsequently reevaluated by a consultant pediatrician or neonatologist for confirmation. The diagnoses of birth defects were made in accordance with the International Classification of Diseases, 10th revision, Clinical Modification (ICD-10-CM), Chapter XVII [15]. The birth defects were categorized by body system, and eleven systems were identified: nervous system (Q00–Q07); eye, ear, face, and neck (Q10–Q18); circulatory system (Q20–Q28); respiratory system (Q30–Q34); cleft lip and cleft palate (Q35–Q37); digestive system (Q38–Q45); genital system (Q50–Q56); urinary system (Q60–Q64); musculoskeletal system (Q65–Q79); other congenital anomalies (Q80–Q89); and chromosomal abnormalities (Q90–Q99) [9]. Newborns with two or more ICD-10 codes were considered to have multiple congenital anomalies, except for established sequences, chromosomal disorders, and single-gene defects [16, 17]. The number of reported defects did not correspond to the number of affected neonates, as individual neonates could present with multiple defects involving more than one organ system. For statistical analysis, each neonate was considered a single unit, even if they were presented with multiple congenital anomalies. Therefore, the analysis was based on the number of neonates with congenital anomalies rather than the total number of individual birth defects identified. In contrast, individual defects were used illustrate the distribution and co-occurrence patterns of anomalies. After initial resuscitation, neonates underwent appropriate laboratory and imaging investigations whenever feasible. They were closely monitored until a outcome (discharge, referral, or death) in neonatology department. Neonates diagnosed with birth defects required multidisciplinary management depending on the type and severity of the anomaly. Most cases required referral to and consultation with relevant specialties for definitive evaluation and treatment planning. Congenital anomalies involving the nervous system (Q00–Q07) were referred to the Department of Neurosurgery. Defects affecting the digestive system (Q38–Q45), genital organs (Q50–Q56), urinary system (Q60–Q64), and selected musculoskeletal anomalies (Q65–Q79) were referred to the Department of Pediatric Surgery. Cleft lip and palate cases were referred to Orthodontics for feeding prosthesis until they were fit for corrective surgery. Selected neonates required consultation from otolaryngology, ophthalmology, orthopedics, and pediatric cardiology services based on clinical findings. Conceptual framework The conceptual framework illustrates the relationship between maternal and neonatal factors and the occurrence of birth defects among neonates. The dependent variable is the presence of birth defects (Yes/No) (Fig. 1). Independent variables are grouped into maternal and neonatal characteristics. Selected variables from baseline characteristics and maternal factors, including birth weight, gestational age, multiple gestation, maternal hypertension, maternal diabetes mellitus, polyhydramnios, and oligohydramnios were included in the multivariable logistic regression model. These factors were selected based on their established association with congenital anomalies in published literature, biological plausibility, recognized clinical relevance, and the availability and completeness of data across both neonates with and without birth defects. Several important maternal factors, such as maternal age, parity, consanguinity, previous obstetric history, and periconceptional folic acid supplementation, were excluded from the regression analysis. These variables were recorded only among neonates with birth defects and were therefore not available in neonates without birth defect, limiting their usefulness for comparative modeling. Fig. 1Conceptual framework of the study. Illustration of the relationship between maternal and neonatal factors (independent variables) and occurrence of birth defects in neonates (dependent variable) Conceptual framework of the study. Illustration of the relationship between maternal and neonatal factors (independent variables) and occurrence of birth defects in neonates (dependent variable) Data analysis Data were analyzed using SPSS version 28. Qualitative variables were summarized as frequencies and percentages and compared using the Pearson Chi-square (χ²) test. Logistic regression analysis was performed to assess perinatal risk factors associated with birth defects. Results were expressed as odds ratios (ORs) with 95% confidence intervals (CIs). A p-value of < 0.05 was considered statistically significant. The goodness of fit of the logistic regression model was assessed using the Hosmer-Lemeshow test. Visual representations included bar diagrams and UpSet plots. System wise distributions of birth defect were visualized using bar diagrams and multisystem involvement was depicted using upset plots. UpSet plot is a matrix based visualization method in which the rows represent the individual sets and the columns represent their intersections, allowing multiple relationships to be explored and customized in ways not supported by traditional set visualization approaches [18]. Results Out of 5,223 admitted neonates during the study period, 183 were identified with birth defects, yielding a prevalence of 3.5%. Among these neonates with birth defects, the majority were male (61.7%), born at term (79.8%), had normal birth weight (51.9%), were delivered per vaginal route (57.4%), and were primarily from non-urban or other districts of Chattogram (73.8%). However, there were no statistically significant differences between neonates with and without birth defects in terms of gender, gestational age, birth weight, mode of delivery, or locality (p > 0.05 for all variables), indicating that these perinatal factors were not significantly associated with the presence of congenital anomalies (Table 1). Of the 183 neonates with birth defects, 56 died, reflecting a mortality rate of 30.6%. The mortality risk among neonates with birth defects was 1.59 times higher (95% CI: 1.15–2.18) compared to those without birth defects, and this difference was statistically significant (p = 0.005). Table 1Comparison of demographic profile of admitted newborns with birth defect and without birth defectVariableTotal(n = 5223)No birth defect(n = 5040)Birth defect(n = 183)P valueN (%)N (%)Gender Male33133200 (96.7%)113 (3.4%)0.63 Female19101840 (96.5%)70 (3.7%)Gestational Age Preterm12281192 (97.1%)36 (2.9%)0.21 Term39953848 (96.3%)147 (3.7%)Birth Weight LBW23112226 (96.3%)85 (3.7%)0.638 NBW28532758 (96.7%)95 (3.3%) HBW5956 (94.9%)03 (5.1%)Mode of Delivery Pervaginal32273122 (96.7%)105 (3.3%)0.339 LSCS19651889 (96.1%)76 (3.9%) Instrumental3129 (93.5%)02 (6.5%)Locality Ctg Urban14341386 (96.7%)48 (3.3%)0.705 Non urban and other districts37893654 (96.4%)135 (3.6%)Demographic profile of neonates with birth defects and without birth defects. Values are presented as number (N) and percentage (%) Table 1Comparison of demographic profile of admitted newborns with birth defect and without birth defectVariableTotal(n = 5223)No birth defect(n = 5040)Birth defect(n = 183)P valueN (%)N (%)Gender Male33133200 (96.7%)113 (3.4%)0.63 Female19101840 (96.5%)70 (3.7%)Gestational Age Preterm12281192 (97.1%)36 (2.9%)0.21 Term39953848 (96.3%)147 (3.7%)Birth Weight LBW23112226 (96.3%)85 (3.7%)0.638 NBW28532758 (96.7%)95 (3.3%) HBW5956 (94.9%)03 (5.1%)Mode of Delivery Pervaginal32273122 (96.7%)105 (3.3%)0.339 LSCS19651889 (96.1%)76 (3.9%) Instrumental3129 (93.5%)02 (6.5%)Locality Ctg Urban14341386 (96.7%)48 (3.3%)0.705 Non urban and other districts37893654 (96.4%)135 (3.6%)Demographic profile of neonates with birth defects and without birth defects. Values are presented as number (N) and percentage (%) Comparison of demographic profile of admitted newborns with birth defect and without birth defect Demographic profile of neonates with birth defects and without birth defects. Values are presented as number (N) and percentage (%) The bar diagram provides a visual illustration of the pattern and frequency of single and multiple congenital anomalies across various organ systems (Fig. 2). Among the 183 identified cases, musculoskeletal anomalies were the most frequently observed overall, particularly in the group with multiple anomalies (39 cases), compared to 28 cases in those with isolated musculoskeletal involvement. Among single-system anomalies, circulatory defects were the second most common (21 cases), followed by cleft lip and cleft palate (19) and anomalies involving nervous system (15). In contrast, within the multiple anomaly group, anomalies involving eye, ear, face and neck were more frequent (19 cases) followed by genital anomalies (14 cases) and cleft lip and palate (11 cases). Fig. 2Systemic distribution of single and multiple anomaly. Illustration of the distribution and frequency of birth defect according to organ system. For each system, bars represent total, single, and multiple anomalies Systemic distribution of single and multiple anomaly. Illustration of the distribution and frequency of birth defect according to organ system. For each system, bars represent total, single, and multiple anomalies Birth defects were classified according to the ICD-10 system, identifying 11 broad categories of defects. The most common category was musculoskeletal deformities (Q65-Q79), followed by cleft lip and cleft palate (Q35-Q37), and anomalies related to the circulatory system (Q20-Q27). Among individual anomalies, Down syndrome (Q90) was the most frequent (21 cases), followed by talipes equinovarus (Q66) with 17 cases, and cleft palate with cleft lip (Q37), which had 15 cases (Table 2). Table 2Distribution of birth defects by system according to ICD-10 classificationSystemIndividual anomaliesICD codeNumber of casesNervous System (n = 20)MicrocephalyQ0203Congenital hydrocephalousQ0308Spina bifidaQ0509Eye, ear, face and neck(n = 23)Congenital ectropionQ10.101MicrophthalmousQ11.201Congenital cataractQ12.001Corneal opacityQ13.302Accessory auricle/Preauricular tagQ17.001MicrotiaQ17.209Misplaced earQ17.405Webbing of neckQ18.302MicrostomiaQ18.501Circulatory system(n = 28)Discordant ventriculoarterial connectionQ20.302Ventricular septal defectQ21.006Atrial septal defectQ21.109Complete AV canal defectQ21.2301Tetrology of FellotQ21.305DextrocardiaQ2404Congenital vascular malformationQ2701Respiratory system (n = 1)Choanal atresiaQ30.001Cleft lip and cleft palate(n = 30)Cleft palateQ3512Cleft lipQ3603Cleft palate with cleft lipQ3715Digestive system(n = 13)AnkyloglossiaQ38.101Salivary mucocele/RanulaQ38.401Congenital absence, atresia and stenosis of small intestineQ4104Anorectal malformationQ42.303Hirschsprung’s diseaseQ43.101Congenital malformation of intestinal fixationQ43.303Genital organs (n = 21)Undescended testicleQ53.206Hypoplasia of penisQ55.6202HypospadiasQ5409Indeterminate sexQ5604Urinary system (n = 4)Polycystic kidneyQ61.101Multicystic kidneyQ61.401HydronephrosisQ6201EpispadiasQ6401Musculoskeletal system(n = 67)Congenital deformities of hipQ6504Congenital talipes equinovarusQ6617Vertical talusQ66.803Congenital facial asymmetryQ6701DolicocephalyQ67.201Congenital deformity of kneeQ68.201PolydactylyQ6907Congenital absence of hand and fingersQ71.302Split handQ71.601Congenital absence of foot and toesQ72.301Congenital shortening of lower limbQ72.8101Arthrogryposis multiplexQ74.306CraniosynostosisQ75.001Hypertelorism of orbit2Q75.202AchondroplasiaQ77.402Osteogenesis imperfectaQ78.001Congenital diaphragmatic herniaQ79.008EventrationQ79.101ExomphalosQ79.206Amniotic band syndromeQ79.801Other congenital malformations(n = 18)Congenital icthyosisQ8004Epidermolysis bullosaQ8103Aplasia cutis congenitaQ84.801Pierre Robin sequenceQ87.008Noonan syndromeQ87.101Conjoined twinsQ89.401Chromosomal abnormalities(n = 24)Down syndromeQ9021Edwards syndromeQ9102Turner’s syndromeQ9601Frequency and distribution of birth defects classified according to ICD-10. Numbers may not sum to the total due to overlapping anomalies Table 2Distribution of birth defects by system according to ICD-10 classificationSystemIndividual anomaliesICD codeNumber of casesNervous System (n = 20)MicrocephalyQ0203Congenital hydrocephalousQ0308Spina bifidaQ0509Eye, ear, face and neck(n = 23)Congenital ectropionQ10.101MicrophthalmousQ11.201Congenital cataractQ12.001Corneal opacityQ13.302Accessory auricle/Preauricular tagQ17.001MicrotiaQ17.209Misplaced earQ17.405Webbing of neckQ18.302MicrostomiaQ18.501Circulatory system(n = 28)Discordant ventriculoarterial connectionQ20.302Ventricular septal defectQ21.006Atrial septal defectQ21.109Complete AV canal defectQ21.2301Tetrology of FellotQ21.305DextrocardiaQ2404Congenital vascular malformationQ2701Respiratory system (n = 1)Choanal atresiaQ30.001Cleft lip and cleft palate(n = 30)Cleft palateQ3512Cleft lipQ3603Cleft palate with cleft lipQ3715Digestive system(n = 13)AnkyloglossiaQ38.101Salivary mucocele/RanulaQ38.401Congenital absence, atresia and stenosis of small intestineQ4104Anorectal malformationQ42.303Hirschsprung’s diseaseQ43.101Congenital malformation of intestinal fixationQ43.303Genital organs (n = 21)Undescended testicleQ53.206Hypoplasia of penisQ55.6202HypospadiasQ5409Indeterminate sexQ5604Urinary system (n = 4)Polycystic kidneyQ61.101Multicystic kidneyQ61.401HydronephrosisQ6201EpispadiasQ6401Musculoskeletal system(n = 67)Congenital deformities of hipQ6504Congenital talipes equinovarusQ6617Vertical talusQ66.803Congenital facial asymmetryQ6701DolicocephalyQ67.201Congenital deformity of kneeQ68.201PolydactylyQ6907Congenital absence of hand and fingersQ71.302Split handQ71.601Congenital absence of foot and toesQ72.301Congenital shortening of lower limbQ72.8101Arthrogryposis multiplexQ74.306CraniosynostosisQ75.001Hypertelorism of orbit2Q75.202AchondroplasiaQ77.402Osteogenesis imperfectaQ78.001Congenital diaphragmatic herniaQ79.008EventrationQ79.101ExomphalosQ79.206Amniotic band syndromeQ79.801Other congenital malformations(n = 18)Congenital icthyosisQ8004Epidermolysis bullosaQ8103Aplasia cutis congenitaQ84.801Pierre Robin sequenceQ87.008Noonan syndromeQ87.101Conjoined twinsQ89.401Chromosomal abnormalities(n = 24)Down syndromeQ9021Edwards syndromeQ9102Turner’s syndromeQ9601Frequency and distribution of birth defects classified according to ICD-10. Numbers may not sum to the total due to overlapping anomalies Distribution of birth defects by system according to ICD-10 classification Eye, ear, face and neck (n = 23) Circulatory system (n = 28) Cleft lip and cleft palate (n = 30) Digestive system (n = 13) Musculoskeletal system (n = 67) Other congenital malformations (n = 18) Chromosomal abnormalities (n = 24) Frequency and distribution of birth defects classified according to ICD-10. Numbers may not sum to the total due to overlapping anomalies A total of 41 neonates presented with multiple congenital anomalies; however, only 38 demonstrated multisystem involvement and were included in the UpSet plot analysis. This UpSet plot visually represents the frequency and co-occurrence of congenital anomalies among these 38 neonates. The black dots at the base of the plot represent the specific systems involved in each combination, while vertical bars show the number of neonates with each combination. Horizontal bars on the left indicate the total frequency of anomalies in each individual system across the cohort. The musculoskeletal system was the most frequently affected, appearing in a substantial number of neonates, often in combination with other systems, highlighting its prominent role in multisystem birth defects. Among the 38 neonates analyzed, 28 patients (approximately 74%) exhibited anomalies involving two organ systems, while the remaining 10 neonates (26%) had more complex presentations involving three or more systems, suggestive of more severe or syndromic congenital anomalies (Fig. 3). Fig. 3An UpSet plot analysis of multiple congenital anomalies in 38 patients. UpSet plot visually represents the frequency and co-occurrence of congenital anomalies among 38 neonates with multisystem involvement. Each dot combination represents a unique overlap of affected organ systems. Vertical bar height indicates the number of neonates with specific combination, and horizontal bars on the left show the total frequency of anomalies in each system An UpSet plot analysis of multiple congenital anomalies in 38 patients. UpSet plot visually represents the frequency and co-occurrence of congenital anomalies among 38 neonates with multisystem involvement. Each dot combination represents a unique overlap of affected organ systems. Vertical bar height indicates the number of neonates with specific combination, and horizontal bars on the left show the total frequency of anomalies in each system Maternal characteristics including age, parity, consanguinity, medical conditions, and detailed obstetric history were analyzed primarily in relation to single versus multiple congenital anomaly groups (Table 3). Selected subgroup associations based on anomaly types were also explored where relevant. Polyhydramnios was found exclusively in neonates with multiple congenital anomalies (100%, p = 0.001), indicating a statistically significant correlation, though based on a small number of cases. Although not reaching statistical significance, a history of previous abortion was more commonly reported across all anomaly groups. Similarly, instances of consanguinity and prior neonatal death were more frequent in the multiple anomaly group, suggesting potential trends that warrant further study in larger cohorts. Most variables showed no significant association when comparing single and multiple anomaly groups (Table 3). However, further analysis across specific anomaly subgroups based on ICD classifications revealed a few notable associations. Maternal age over 35 years was strongly linked to chromosomal anomalies (p = 0.009), with an adjusted odds ratio (AOR) of 4.16 (95% CI: 1.42–12.22), identifying it as a substantial risk factor. Multiparity was also notably higher in the chromosomal anomaly group (87%) and demonstrated a significant association (p = 0.025) with an adjusted odds ratio (AOR) of 4.23 (95% CI: 1.20–14.79). Within the musculoskeletal anomaly group, 19.6% of mothers reported a previous neonatal death, a statistically significant finding (p = 0.026) (AOR 2.86, 95% CI: 1.13–7.2). Additionally, 75.4% of neonates with birth defects were born to mothers who did not receive folic acid during perconceptional period. Folic acid supplementation among neonates with neural tube defects (NTDs) versus those with other anomalies were further analyzed. No significant association was found between folic acid supplementation and neurological defects (p = 0.854) (AOR 0.90, 95% CI: 0.31–2.67). Table 3Risk factors associated with single and multiple anomaliesRisk FactorsTotal (n = 183)Single (n = 142)Multiple (n = 41)P valueMaternal Age< 20 yrs4236 (85.7%)06 (14.3%)0.32520–35 yrs12291 (74.6%)31 (25.4%)> 35 yrs1915 (78.9%)04 (21.1%)ParityPrimi6555 (84.6%)10 (15.4%)0.091Multi11887 (73.7%)31 (26.3%)Consanguinity of marriage1811 (61.1%)07 (38.9%)0.077Previous history of affected newborn0403 (75%)01 (25%)0.90Previous abortion3224 (75%)08 (25%)0.70Previous still birth0202 (100%)0 (0%)0.45Previous neonatal death2113 (61.9%)08 (38.1%)0.07Maternal diabetes mellitus1409 (64.3%)05 (35.7%)0.21Maternal hypertension2618 (69.2%)08 (30.8%)0.27Oligohydramnios0504 (80%)01 (20%)0.90Polyhydramnios030 (0%)03 (100%)0.001Maternal hypothyroidism0201 (50%)01 (50%)0.35No periconceptional folic acid supplementation138111 (80.4%)27 (19.6%)0.11Indigenous drug use0303 (100%)0 (0%)0.35Maternal risk factors associated with single versus multiple anomalies. Numbers are presented as n (%) and P values indicate statistical significance of differences betweeen the two groups Table 3Risk factors associated with single and multiple anomaliesRisk FactorsTotal (n = 183)Single (n = 142)Multiple (n = 41)P valueMaternal Age< 20 yrs4236 (85.7%)06 (14.3%)0.32520–35 yrs12291 (74.6%)31 (25.4%)> 35 yrs1915 (78.9%)04 (21.1%)ParityPrimi6555 (84.6%)10 (15.4%)0.091Multi11887 (73.7%)31 (26.3%)Consanguinity of marriage1811 (61.1%)07 (38.9%)0.077Previous history of affected newborn0403 (75%)01 (25%)0.90Previous abortion3224 (75%)08 (25%)0.70Previous still birth0202 (100%)0 (0%)0.45Previous neonatal death2113 (61.9%)08 (38.1%)0.07Maternal diabetes mellitus1409 (64.3%)05 (35.7%)0.21Maternal hypertension2618 (69.2%)08 (30.8%)0.27Oligohydramnios0504 (80%)01 (20%)0.90Polyhydramnios030 (0%)03 (100%)0.001Maternal hypothyroidism0201 (50%)01 (50%)0.35No periconceptional folic acid supplementation138111 (80.4%)27 (19.6%)0.11Indigenous drug use0303 (100%)0 (0%)0.35Maternal risk factors associated with single versus multiple anomalies. Numbers are presented as n (%) and P values indicate statistical significance of differences betweeen the two groups Risk factors associated with single and multiple anomalies Maternal risk factors associated with single versus multiple anomalies. Numbers are presented as n (%) and P values indicate statistical significance of differences betweeen the two groups Table 4 presents the results of univariate and multivariate logistic regression analyses assessing maternal and perinatal risk factors associated with birth defects. The analysis included 5,223 live births, with 183 cases identified as having congenital anomalies. Significant associations were found for several maternal conditions. Maternal hypertension was strongly associated with birth defects (AOR: 3.92, 95% CI: 2.56–5.99, p < 0.001), indicating nearly fourfold increased odds compared to normotensive mothers. Maternal diabetes showed a similar association (AOR: 4.15, 95% CI: 2.37–7.27, p < 0.001). Abnormalities in amniotic fluid volume were also significant predictors, with polyhydramnios (AOR: 7.33, 95% CI: 2.63–20.41) and oligohydramnios (AOR: 10.19, 95% CI: 3.68–28.23) both showing strong associations (p < 0.001). In contrast, multiple gestation, preterm birth, and variations in birth weight were not significantly associated with birth defects in the multivariate model. Table 4Logistic regression analysis of risk factors for birth defectsRisk FactorNo birth defect n = 5040Birth defectn = 183Univariate OR95% CI,p valueMultivariate OR95% CI,p valueMaternal Hypertension No4757 (96.9%)154 (3.1%)RefRef Yes283 (90.7%)29 (9.3%)3.172.09–4.79, p < 0.0013.922.56–5.99, p < 0.001Maternal Diabetes No4902 (96.7%)167 (3.3%)RefRef Yes138 (89.6%)16 (10.4%)3.401.98–5.84, p < 0.0014.152.37–7.27, p < 0.001Polyhydramnios No5016 (96.6%)178 (3.4%)RefRef Yes24 (82.8%)5 (17.2%)5.872.21–15.57,p < 0.0017.332.63–20.41, p < 0.001Oligohydramnios No5023 (96.6%)178 (3.4%)RefRef Yes17 (77.3%)05 (22.7%)8.303.03–22.75, p < 0.00110.193.68–28.23, p < 0.001Multiple Gestation No4842 (96.5%)175 (3.5%)RefRef Yes196 (96.1%)08 (3.9%)1.130.55–2.33, p 0.7421.840.87–3.93, p 0.113Gestational age Term3848 (96.3%)147 (3.7%)RefRef Preterm1192 (97.1%)36 (2.9%)0.7910.546–1.145, p 0.2130.7540.514–1.11, p 0.15Birth weight NBW2758 (96.7%)95 (3.3%)RefRef LBW2226 (96.3%)85 (3.7%)1.110.82–1.49, p 0.491.280.89–1.84, p 0.172 HBW56 (94.9%)3 (5.1%)1.560.48–5.06, p 0.460.920.27–3.13, p 0.89Maternal and perinatal risk factors associated with birth defects were analyzed using univariate and multivariate logistic regression. Data are presented as odds ratio (OR) with 95% confidence intervals (Cl) and P values. "Ref" indicates the reference category Table 4Logistic regression analysis of risk factors for birth defectsRisk FactorNo birth defect n = 5040Birth defectn = 183Univariate OR95% CI,p valueMultivariate OR95% CI,p valueMaternal Hypertension No4757 (96.9%)154 (3.1%)RefRef Yes283 (90.7%)29 (9.3%)3.172.09–4.79, p < 0.0013.922.56–5.99, p < 0.001Maternal Diabetes No4902 (96.7%)167 (3.3%)RefRef Yes138 (89.6%)16 (10.4%)3.401.98–5.84, p < 0.0014.152.37–7.27, p < 0.001Polyhydramnios No5016 (96.6%)178 (3.4%)RefRef Yes24 (82.8%)5 (17.2%)5.872.21–15.57,p < 0.0017.332.63–20.41, p < 0.001Oligohydramnios No5023 (96.6%)178 (3.4%)RefRef Yes17 (77.3%)05 (22.7%)8.303.03–22.75, p < 0.00110.193.68–28.23, p < 0.001Multiple Gestation No4842 (96.5%)175 (3.5%)RefRef Yes196 (96.1%)08 (3.9%)1.130.55–2.33, p 0.7421.840.87–3.93, p 0.113Gestational age Term3848 (96.3%)147 (3.7%)RefRef Preterm1192 (97.1%)36 (2.9%)0.7910.546–1.145, p 0.2130.7540.514–1.11, p 0.15Birth weight NBW2758 (96.7%)95 (3.3%)RefRef LBW2226 (96.3%)85 (3.7%)1.110.82–1.49, p 0.491.280.89–1.84, p 0.172 HBW56 (94.9%)3 (5.1%)1.560.48–5.06, p 0.460.920.27–3.13, p 0.89Maternal and perinatal risk factors associated with birth defects were analyzed using univariate and multivariate logistic regression. Data are presented as odds ratio (OR) with 95% confidence intervals (Cl) and P values. "Ref" indicates the reference category Logistic regression analysis of risk factors for birth defects Maternal and perinatal risk factors associated with birth defects were analyzed using univariate and multivariate logistic regression. Data are presented as odds ratio (OR) with 95% confidence intervals (Cl) and P values. "Ref" indicates the reference category Discussion Many healthcare systems, particularly in LMICs like Bangladesh, are unprepared to cope with the growing incidence of congenital anomalies [19], largely due to lack of awareness and inadequate diagnostic capabilities [20, 21]. The majority of newborns are not screened for visible birth defects, especially those delivered outside formal healthcare settings [22]. Against this background, this study was conducted to address a critical gap in the existing literature on birth defects in Bangladesh, where previous hospital-based studies have been limited by inconsistent classification systems and absence of concurrent risk factor analysis. By comprehensively investigating the prevalence, patterns, and risk factors of birth defects among neonates admitted to a government tertiary NICU, the present study provides a more complete epidemiological picture than prior investigations. The findings further emphasize the need to improve hospital-based monitoring and develop nationwide surveillance systems. The prevalence of birth defects in the current study was 3.5%. As the denominator comprises admitted neonates rather than all live births, the reported prevalence reflects a hospital-based estimate and should be interpreted accordingly. The prevalence is comparable to other hospital-based studies conducted in Bangladesh. Ahmed et al. conducted a study at a private NICU in the same city and reported a prevalence of 4.38% [22], while two other studies conducted in Obstetrics and Gynaecology departments reported prevalence rates of 4.34% [23] and 3.68% [24] respectively. The marginal differences are likely due to variations in the denominator, admission criteria, and institutional setting. A higher prevalence of 8.9% was reported by Verma et al. in a seven-year retrospective study conducted in a tertiary NICU in northern India [25], which may reflect differences in population characteristics and regional genetic predispositions. A comparably elevated prevalence of 7.0% was reported from a five-year prospective study of singleton pregnancies in the Middle East [26]. The higher figures observed in both studies may be partly explained by their longer study periods, which allow for greater case accumulation and more comprehensive detection of birth defects compared with shorter duration studies. The mortality rate among neonates with birth defects in our study was 30.6%, This is notably higher than the 5.67% reported by Ahmed et al. in a private NICU setting in Chattogram [22], which may reflect differences in case mix and referral patterns, as our public tertiary neonatal unit receives a disproportionately higher burden of critically ill and complex cases, often beyond its designated bed capacity. Amin et al. reported a mortality rate of 17% among neonates with birth defects in a cross sectional study [27]. It is worth mentioning that, both of these studies were conducted several years prior to the present research. Our results further demonstrate that neonates with birth defects had a significantly higher risk of death (OR = 1.59; 95% CI: 1.15–2.18) compared with those without birth defects. As comparable estimates are currently lacking from Bangladesh, comparison was made with a study conducted in Uganda. This LMIC also reported a similar directional association (OR = 1.81; 95% CI: 0.95–3.46) [28]. We did not assess individual determinants of neonatal mortality. Previous studies have identified cesarean delivery, advanced maternal age, consanguinity, and low birth weight as factors significantly associated with anomaly-related death [29], warranting further investigation in future studies from our setting. Beyond mortality outcomes, the distribution of anomaly subtypes in our study revealed several distinct patterns. Musculoskeletal anomalies were the most predominant group in our study, both as isolated defects and as part of multiple anomalies. Talipes equinovarus was the most common anomaly, followed by polydactyly, congenital diaphragmatic hernia, and arthrogryposis. Talipes equinovarus was also reported as the most frequent musculoskeletal anomaly by Akter et al. [11], Habib et al. [23], and Amin et al. [27] in Bangladesh. Sinha et al. similarly observed a comparable pattern in northern India [30]. The higher proportion of musculoskeletal anomalies in hospital-based studies may be explained by their relatively easier clinical detectability during routine clinical examination. Among other frequently identified anomaly groups, Congenital heart disease (CHD) was also one of the prevalent anomalies in our study. CHD is increasingly emerging as one of the most frequently detected congenital anomalies in our country [22, 27, 31] and abroad [4, 32, 33]. Studies conducted in Chattogram by Ahmed et al. [22], as well as in other regions of the country by Amin et al. [27] and Afroze et al. [31], also reported similar findings. Consistently, Verma et al. in India identified CHD as the most common anomaly [25]. Globally, a marked rise in the prevalence of congenital heart disease has also been reported [34]. This rising trend may reflect improved detection due to increased availability of and growing expertise in neonatal echocardiography. We believe that the actual percentage of congenital heart disease is likely higher than reported in this study, as only echocardiograhically confirmed cases were included. Thirty-nine suspected cases were excluded because bedside echocardiography was not available in our facility and some neonates were too unstable to be referred for the investigations. In our study, only 13 birth defects were attributed to the digestive system. In contrast, Kabir et al. [10] and Rouf et al. [35], in surgical department-based studies, reported gastrointestinal anomalies as a substantial proportion of cases, with anorectal malformation being the most common. This discrepancy likely reflects speciality-based case concentration and underscores the need for interdepartmental collaboration to determine the true prevalence. In our study, 20 congenital anomalies involved the nervous system, mainly presenting as spina bifida and hydrocephalus; however, this category was not among the top three most frequent groups. Similar findings have been reported by Akter et al. [11], a more recent study that also documented cases of anencephaly in addition to these anomalies. Ahmed et al. observed only a few such cases, with hydrocephalus being the sole reported neural anomaly [22]. In contrast, earlier obstetric-based studies conducted by Fatema et al. [24] and Parveen et al. [36], reported nervous system anomalies as the most prevalent category. These studies primarily relied on clinical or visual examination, and it was not specified whether mothers were admitted with prior knowledge of the anomaly via imaging. Taken together, these findings suggest a temporal shift in the distribution pattern of birth defects over time, potentially related to gradual improvements in socioeconomic conditions, maternal health, and perinatal care. Twenty-four chromosomal anomalies were identified in the present study, with Down syndrome being the most common, also ranking as the most frequently identified individual anomaly overall. Ahmed et al. [22] identified Down syndrome as the third most common anomaly in their series, and Akter et al. [11] documented 17 chromosomal anomaly cases in their cohort. Advanced maternal age and multiparity were significantly associated with chromosomal anomalies in the present study, consistent with findings reported in previous study [2]. With advancing maternal age, the risk of errors during chromosomal division particularly meiotic non-disjunction increases, resulting in gametes with an abnormal chromosomal complement [37]. An estimated 15–30% of infants born with a birth defect will have more than one malformation [38]. In our study, 20.7% neonates with birth defects presented with multisystem involvement corresponding to a rate of 7.2 per 1,000 neonates. Multisystem involvement, often termed multiple congenital anomalies (MCA), may arise secondary to a single primary defect, as seen in certain sequences. This category generally excludes the conditions that are explained by single-gene mutation or chromosomal abnormalities [16, 39, 40]. The reported prevalence of MCA varies across studies; Amin et al. [27] reported an MCA rate of 27%. While Sinha et al. [30] reported a 25% prevalence in an Indian cohort, with musculoskeletal and nervous system anomalies being the most frequent combination. Total 8804 cases with two or more major congenital anomalies in different organ systems were identified in EUROCAT multiple congenital anomaly algorithm. They analyzed 1,386 combinations of two anomalies, of which six pairs remained classified as new associations [40]. Among neonates with two system affected, musculoskeletal anomalies most commonly co-existed with genital anomalies or with anomalies of eye, ear, face and neck as well as circulatory defects. As cleft palate is not an essential criterion of Pierre Robin sequence and has a seperate ICD code, such cases were included as MCA in our study to ensure consistency with anomaly-based coding [15, 41]. Among neonates with involvement of three systems, anomalies involving musculoskeletal and genital anomalies, cleft lip and palate, and anomalies of the eye, ear, face and neck were more common. Two neonates with Down syndrome were included in the MCA group because talipes equinovarus and congenital ichthyosis are not typical association of the syndrome. Maternal hypertension (OR = 3.92, 95% CI: 2.56–5.99, p < 0.001) and maternal diabetes (OR = 4.15, 95% CI: 2.37–7.27, p < 0.001) were each associated with an approximately fourfold increased risk of birth defects in offspring. Afroze et al. [31] and Hussain et al. [42] similarly reported significant associations between maternal diabetes and congenital anomalies in Bangladesh, with comparable effect sizes (OR: 4.47, 95% CI: 1.09–18.41, p = 0.038 and OR: 4.43, 95% CI: 1.15–17.07, p = 0.03, respectively). Offsprings of women with pregestational diabetes mellitus (PGDM) had a higher risk of congenital heart defects (RR = 3.46; 95% CI: 2.77–4.32). Gestational diabetes mellitus (GDM) and PGDM was also associated with increased risks of congenital anomalies across multiple systems [43]. Exposure to hypertensive disorders of pregnancy such as pre-eclampsia, gestational hypertension, and chronic hypertension may increase the risk of birth defects particularly congenital heart disease with chronic hypertension markedly elevating the odds of renal, limb, and orofacial malformations, particularly when combined with superimposed preeclampsia [44, 45]. Polyhydramnios (AOR: 7.33, 95% CI: 2.63–20.41, p < 0.001) and oligohydramnios (AOR: 10.19, 95% CI: 3.68–28.23, p < 0.001) both emerged as significant independent predictors of congenital anomalies. Amniotic fluid volume abnormalities are well-documented across multiple anomaly groups. Oligohydramnios is most strongly associated with urogenital anomalies, while polyhydramnios shows high association factors for craniospinal, abdominal, and cardiovascular defects [46]. However, these associations were based on small case numbers (n = 5 each), and the wide confidence intervals reflect statistical instability; findings should therefore be interpreted cautiously and confirmed in larger prospective studies. We further acknowledge that these conditions may in some instances represent clinical manifestations of underlying structural anomalies rather than purely antecedent risk factors, and this bidirectional relationship should also be considered when interpreting our findings. This study did not demonstrate any significant connection between prematurity, low birth weight, or multiple gestation and the occurrence of birth defects. Preconceptional folic acid is a well-established protective factor against congenital anomalies [1]. Most mothers in our country seek their first antenatal care late [47]. This may explain why 75% of the mothers in this study did not receive folic acid during the first trimester or in the periconceptional period. Consequently, folate deficiency could not be identified as an independent risk factor for neural tube defects in our study; Furthermore, data on folic acid use among neonates without anomalies were not available, so the overall protective effect of folic acid could not be assessed in the full cohort. Afroze et al. [31] and Hussain et al. [42] have reported that lack of folic acid supplementation significantly increased the risk of birth defects in Bangladesh suggesting that this association likely exists in our population but could not be captured within the constraints of our study design. The WHO recommended that LMICs should identify one or more priority birth defects and begin national initiatives for integrating screening, diagnosis, and management of the prioritised conditions into routine health services [8]. Therefore, this study may guide us in understanding the regional perspective, as geographical differences in the pattern of congenital anomalies are already well recognised. Moreover, identification of modifiable risk factors such as maternal hypertension, diabetes mellitus, and abnormal amniotic fluid volume warrants better implementation of antenatal care and improved management of high-risk pregnancies. Limitation This was a single-center, hospital-based study with a relatively short study period, which may limit the generalizability of the findings. Moreover, the study captured only hospital-based outcomes and neonatal mortality within the study setting. Neonates were not followed after discharge or referral; therefore, post-discharge and community or home deaths could not be assessed. In addition, some important maternal variables were not available for the comparison group, which limited their inclusion of these variables in the multivariable regression analysis and may have introduced residual confounding. Conclusion The burden of birth defect and associated mortality was significant in this study. Musculoskeletal anomalies were the most frequently observed defects, with notable multisystem involvement. Maternal illness, particularly diabetes and hypertension, made a substantial contribution to the occurance of birth defects and need to be addressed more rigorously through targeted antenatal screening and counselling. Our study findings suggest that as an LMIC, Bangladesh should prioritize targeted screening for congenital anomalies. It is high time to initiate birth defect surveillance through clinical examination of neonates, at community levels in addition to hospitals, to ensure early detection and timely referral of visible birth defects.