Esophageal Atresia in Asia: Regional Disparities in Incidence, Congenital Anomalies, Mortality, and Morbidity – A Systematic Narrative Review

ABSTRACT

Background:

Esophageal atresia (EA), with or without tracheoesophageal fistula, remains associated with substantial early mortality and long-term morbidity. Reported incidence and outcomes vary across Asia, but interpretation is complicated by differences in case capture, referral pathways, diagnostic capacity, and neonatal intensive care availability.

Methods:

We performed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020-aligned systematic narrative review of observational studies from Asian countries published up to March 2024 (PubMed and Google Scholar). Studies reporting incidence, associated anomalies, morbidity, or mortality were included. Findings were synthesized descriptively and presented stratified by study base (population-based vs. hospital-based) and resource setting (World Bank income proxy). Risk of bias was assessed using a generic domain-based framework (D1–D6).

Results:

Thirty-four studies (3304 infants) were included. Six population-based studies reported incidence ranging from 6.17 to 20.30 per 100,000 live births. Hospital-based reports documented 921 associated anomalies, most commonly gastrointestinal (36%) and cardiovascular (27%). Mortality across 19 cohorts ranged from 2% to 83%. Morbidity profiles differed by setting: in higher-resource cohorts, reported events were dominated by gastroesophageal reflux and anastomotic stricture, whereas in lower-resource cohorts, septicemia (47% of events) and wound infection (10%) were prominent.

Conclusion:

EA incidence in Asia is within international ranges, but published outcomes vary widely and should be interpreted descriptively given referral and ascertainment differences, variable anomaly detection, heterogeneous definitions, and limited follow-up. Priorities include strengthening prenatal detection and referral, improving neonatal stabilization and safe transport, standardizing perioperative care with infection prevention, and establishing feasible long-term follow-up supported by a pragmatic action framework.

INTRODUCTION

Esophageal atresia (EA) is a life-threatening congenital anomaly that exemplifies the dependence of neonatal surgical outcomes on health system capacity. Advances in neonatal intensive care, anesthesia, and pediatric surgery have transformed EA into a largely survivable condition in high-income settings, where survival now exceeds 90%. In contrast, mortality in many low- and middle-income countries (LMICs) remains above 30%–50%, with some hospital-based series reporting mortality >70%.[1,2] These disparities largely reflect differences in early recognition, referral pathways, neonatal transport, access to intensive care, infection prevention, and continuity of follow-up rather than intrinsic differences in disease biology.

Asia represents a particularly heterogeneous region, encompassing countries with advanced neonatal surgical systems alongside settings where neonatal surgical care remains severely constrained. Interpretation of EA outcomes in the region is complicated by referral dependency and underreporting of early neonatal deaths prior to admission or surgery, which can inflate apparent survival in hospital-based cohorts.[3] Meaningful synthesis therefore requires explicit consideration of health system context.

Globally, EA occurs in approximately 0.4–3.6 per 10,000 live births, although reported incidence varies widely across regions.[2] Population-based studies from Asia demonstrate substantial variation in reported incidence, likely reflecting differences in congenital anomaly surveillance systems, registry coverage, and prenatal detection practices rather than true epidemiologic variation.

EA is frequently associated with other congenital anomalies, with 40%–60% of infants affected, most commonly involving the cardiovascular and gastrointestinal systems. Multisystem patterns such as VACTERL association are well recognized.[4,5] The presence of major associated anomalies, particularly congenital heart disease, is a key predictor of mortality.[6] Reported anomaly profiles vary between studies, reflecting differences in diagnostic work-up rather than underlying disease patterns.

Beyond survival, EA is increasingly recognized as a chronic condition with substantial long-term morbidity. More than half of survivors experience gastrointestinal or respiratory complications, including gastroesophageal reflux, anastomotic stricture, dysphagia, and recurrent respiratory infections, in cohorts with structured follow-up.[7] In lower-resource Asian settings, long-term outcomes are rarely reported due to limited follow-up infrastructure, leading to underestimation of the full disease burden.

Despite the growing literature, few reviews have focused on EA in Asia while explicitly accounting for differences in surveillance systems, referral pathways, and health system capacity. To address this gap, we conducted a systematic narrative review of EA in Asia with three objectives: (1) to synthesize available evidence into a pragmatic, health system-oriented framework highlighting actionable priorities for reducing EA-related mortality and morbidity in LMIC Asian settings, (2) to summarize population-based incidence estimates while accounting for differences in surveillance coverage and case ascertainment, and (3) to describe associated congenital anomalies, mortality, and morbidity reported in hospital-based cohorts, stratified by resource setting to support system-level interpretation. This approach aims to provide a policy-relevant synthesis to inform neonatal surgical planning and quality improvement across Asia.

METHODS

This systematic narrative review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines. Literature searches were conducted in PubMed and Google Scholar from January 2014 to March 2024, representing the most recent 10-year period prior to the search. The search used combinations of Medical Subject Headings and free-text terms related to EA, tracheoesophageal fistula, congenital anomalies, morbidity, mortality, incidence, prevalence, and Asia. Reference lists of relevant articles were also screened to identify additional eligible studies.

We included observational studies from Asian countries (cohort studies, cross-sectional studies, and case series including ≥10 infants) that reported at least one of the following outcomes: population-based incidence or prevalence, associated congenital anomalies, morbidity, or mortality in infants with EA. Studies were limited to those published in English with accessible full texts to allow consistent evaluation of eligibility and data extraction across reviewers using publicly searchable databases. Reviews, editorials, letters, case reports, randomized trials, non-Asian studies, and studies without extractable EA-specific data were excluded. After removal of duplicates, titles and abstracts were screened, followed by full-text assessment of potentially eligible articles. Discrepancies in study selection were resolved through discussion. The study selection process is summarized in the PRISMA flow diagram [Figure 1].

Figure 1.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses flowchart of the study

Data were extracted using a standardized template capturing country or subregion, study design, study period, study setting (population-based surveillance/registry versus hospital-based cohort), sample size, and extractable numerators and denominators for incidence, associated anomalies, morbidity, and mortality. For incidence studies, the number of EA cases and the corresponding total number of births in the source population were recorded.

To support stratified interpretation, studies were categorized according to study base (population-based versus hospital-based) and resource context using the World Bank income classification as a proxy for health-system resources. Countries were grouped into higher-resource settings (high-income and upper-middle-income countries) and lower-resource settings (lower-middle-income and low-income countries). Where reported, morbidity outcomes were also categorized according to follow-up type, distinguishing perioperative or early postoperative complications from outcomes reported during longer-term follow-up.

Risk of bias was assessed using a prespecified generic domain-based framework designed for heterogeneous observational evidence. Each study was rated as low risk, some concerns, or high risk across six domains: (D1) selection and referral processes, (D2) incomplete case ascertainment including early neonatal deaths, (D3) differential detection of associated anomalies, (D4) outcome definition and measurement, (D5) missing follow-up or attrition, and (D6) confounding or incomplete case-mix reporting. This framework was selected because the included evidence base comprised both population-based surveillance studies and hospital-based case series, making the consistent application of a single design-specific risk-of-bias tool difficult. These domains were chosen to capture the principal threats to comparability across settings, including referral bias, underascertainment of early deaths, variable diagnostic capacity, and incomplete long-term follow-up. Study-level assessments are summarized in Figure 2 with detailed ratings provided in Supplementary Table 1.

Figure 2.

Risk-of-bias summary[8]

Supplementary Table 1.

Risk-of-bias assessment of included studies using the generic domain-based framework (D1–D6) with overall judgment

Because of substantial heterogeneity in study design, outcome definitions, follow-up duration, and reporting practices, findings were synthesized narratively using ranges and descriptive summaries. Where aggregated counts were presented, these were interpreted descriptively rather than as pooled estimates or causal comparisons between the settings.

RESULTS

Study characteristics

A total of 34 studies including 3304 infants with EA met the eligibility criteria and were included in this review. The included studies varied substantially in design, surveillance coverage, and health system context. To improve interpretability, findings are presented separately for population-based studies, which contributed incidence estimates, and hospital-based cohorts, which reported associated anomalies, mortality, and morbidity. Hospital-based studies accounted for the majority of included evidence and primarily originated from tertiary referral centers. Population-based studies were derived from national or regional congenital anomaly surveillance systems and birth registries. The characteristics of included studies and outcome reporting are summarized in Tables 1-5.

Table 1.

Congenital-related anomalies associated with esophageal atresia

Anomaly type	Number of cases (%)
Gastrointestinal	331 (35.94)
Cardiovascular	250 (27.14)
Musculoskeletal	64 (6.95)
Genitourinary	41 (4.45)
Genetic	12 (1.30)
Neurological	5 (0.54)
Pulmonary	12 (1.30)
Unspecified VACTERL	121 (13.14)
Unspecified non-VACTERL	85 (9.23)
Total	921 (100)

Percentages represent the distribution of reported anomaly events across included studies. As individual infants may have multiple anomalies, these values do not represent patient-level prevalence

Table 5.

Population-based incidence studies

Study	Country	Region	Resource setting	Study base	Cases (n)	Births/reported rate	Incidence per 100,000
Faal et al. (2018)[29]	Iran	Western Asia	Higher-resource	Population-based	1	6000 births	16.67
Lee et al. (2023)[12]	Korea	East Asia	Higher-resource	Population-based	470	2,703,822 births	17.38
Sirichamratsakul et al. (2023)[17]	Thailand	Southeast Asia	Higher-resource	Population-based	216	0.88 per 10,000	8.80
Song et al. (2022)[30]	China	East Asia	Higher-resource	Population-based	8	1.65 per 10,000	16.50
Ajimamatov (2023)[31]	Uzbekistan	Central Asia	Lower-resource	Population-based	1	1 in 16,219 births	6.17
Yang et al. (2022)[32]	China	East Asia	Higher-resource	Population-based	30	147,817 births	20.30

Congenital-related anomalies

Across the included cohorts, a total of 921 associated anomalies were reported [Table 1]. These counts represent event-level totals rather than patient-level prevalence, as individual infants may have more than one anomaly, and most studies did not report the proportion of infants with at least one associated anomaly.

Among reported anomaly events, gastrointestinal anomalies were most frequently reported (331, 36%), followed by cardiovascular anomalies (250, 27%). Syndromic or multisystem patterns were also commonly reported, including unspecified VACTERL (121, 13%) and unspecified non-VACTERL multisystem patterns (85, 9%). Musculoskeletal (64, 7%) and genitourinary system anomalies (41, 4%).

Anomaly detection depends on the intensity of diagnostic evaluation (e.g. routine echocardiography, genetic testing, or multisystem screening), differences in reported anomaly distributions between studies should be interpreted primarily as differences in detection and reporting practices rather than definitive differences in underlying epidemiology [Table 1].

Mortality (hospital-based cohorts)

Nineteen studies reported mortality outcomes [Table 2a and b]. Mortality varied widely across cohorts, ranging from 2.2% (Zhang, 2020, China) to 83% (Chowdhury, 2020, Bangladesh) [Table 2a and b]. In higher-resource settings, multiple cohorts reported mortality in the single-digit to low-teen range, including 8% (Al-Jahdali, 2018, UAE), 8% (Kazemian, 2022, Iran), 5% (Shahri, 2021, Iran), 11% (Chiang, 2023, Singapore), 12% (Lee, 2023, Korea), and 14% (Sirichamratsakul, 2023, Thailand) [Table 2a]. Very low-reported mortality was also observed in selected cohorts, including 3% (Okuyama, 2018, South Korea) and 2% (Zhang, 2020, China), although these findings should be interpreted in relation to study design and case-mix [Table 2a].

Table 2a.

Mortality in higher resource settings

Study	Country	Resource setting	Sample size (n)	Deaths (n)	Mortality (%)
Al-Jahdali et al. (2018)[9]	UAE	Higher-resource	50	4	8.0
Chiang et al. (2023)[7]	Singapore	Higher-resource	65	7	10.8
Kazemian et al. (2022)[10]	Iran	Higher-resource	233	19	8.2
Mottaghi et al. (2021)[11]	Iran	Higher-resource	38	2	5.3
Lee et al. (2023)[12]	Korea	Higher-resource	470	56	11.9
Lu et al. (2021)[13]	Taiwan	Higher-resource	86	14	16.3
Murase et al. (2015)[14]	Japan	Higher-resource	35	3	8.6
Muto et al. (2022)[15]	Japan	Higher-resource	32	4	12.5
Okuyama et al. (2018)[16]	South Korea	Higher-resource	62	2	3.2
Sirichamratsakul et al. (2023)[17]	Thailand	Higher-resource	216	30	13.9
Syed et al. (2020)[18]	Saudi Arabia	Higher-resource	12	4	33.3
Zhang et al. (2020)[19]	China	Higher-resource	92	2	2.2

Table 2b.

Lower-resource settings

Study	Country	Resource setting	Sample size (n)	Deaths (n)	Mortality (%)
Ahmed and Tareen (2018)[20]	Pakistan	Lower-resource	30	5	16.7
Chowdhary et al. (2020)[21]	India	Lower-resource	25	6	24.0
Chowdhury et al. (2020)[22]	Bangladesh	Lower-resource	18	15	83.3
Gupta and Bhatnagar (2016 )[23]	India	Lower-resource	430	296	68.8
Hatti et al. (2020)[24]	India	Lower-resource	61	12	19.7
Jagdeesh Prasad (2017)[25]	India	Lower-resource	53	21	39.6
Saleem et al. (2020)[26]	Pakistan	Lower-resource	18	3	16.7

In contrast, several cohorts from lower-resource settings reported substantially higher mortality, including 24% (Chowdhary, 2020, India), 40% (Jagdeesh Prasad, 2017, India), 69% (Gupta, 2018, India), and 83% (Chowdhury, 2020, Bangladesh) [Table 2b]. These differences are consistent with variation in referral timing, access to neonatal intensive care, perioperative support, and infection control infrastructure.

Mortality definitions and observation windows were not standardized across studies, with some reporting inhospital mortality, while others reported perioperative or unspecified postoperative mortality [Supplementary Table 2]. Consequently, direct quantitative comparison between studies should be interpreted cautiously.

Supplementary Table 2.

Mortality definitions and observation windows across included hospital-based studies

Study	Country	Resource setting	Mortality definition reported	Observation window/time frame	Notes
Al-Jahdali, 2018	UAE	Higher-resource	In-hospital mortality	During hospital admission	Mortality reported for neonatal surgical admission
Chiang, 2023	Singapore	Higher-resource	Postoperative mortality	During follow-up after EA repair	Long-term cohort with structured follow-up
Kazemian, 2022	Iran	Higher-resource	Early mortality	Not clearly specified	Reported as early deaths during treatment period
Shahri, 2021	Iran	Higher-resource	In-hospital mortality	During admission	Based on hospital cohort
Lee, 2023	Korea	Higher-resource	Mortality among surgical cases	Not specified	Derived from national surgical dataset
Lu, 2021	Taiwan	Higher-resource	Mortality reported in cohort	Not clearly defined	Focus primarily on postoperative morbidity
Murase, 2015	Japan	Higher-resource	Postoperative mortality	Not clearly defined	Study focused on stricture prevention
Muto, 2022	Japan	Higher-resource	Early postoperative mortality	Not clearly specified	Patients with congenital heart disease included
Okuyama, 2018	South Korea	Higher-resource	Surgical mortality	Perioperative	Thoracoscopic EA repair cohort
Sirichamratsakul, 2023	Thailand	Higher-resource	Mortality among EA cases	Not specified	Registry-based congenital anomaly study
Syed, 2020	Saudi Arabia	Higher-resource	Postoperative mortality	Early postoperative period	Focus on early complications
Zhang, 2020	China	Higher-resource	Surgical mortality	Perioperative	Thoracoscopic repair cohort
Ahmed, 2018	Pakistan	Lower-resource	Mortality among treated EA patients	Not clearly defined	Hospital cohort
Chowdhary, 2020	India	Lower-resource	Mortality in EA with ARM	During hospital management	Mixed congenital anomaly cohort
Chowdhury, 2020	Bangladesh	Lower-resource	Mortality among pediatric surgical patients	Not specified	EA subgroup extracted
Gupta, 2018	India	Lower-resource	In-hospital mortality	During admission	High-volume tertiary center
Hatti, 2020	India	Lower-resource	Early mortality	Perioperative	Study focused on long-gap EA
Jagdeesh, Prasad, 2017	India	Lower-resource	Mortality during treatment	Hospital stay	District hospital cohort
Saleem, 2020	Pakistan	Lower-resource	Mortality reported in surgical series	Not clearly defined	Esophageal replacement cohort

EA: Esophageal atresia, ARM: Anorectal malformation

Morbidity and follow-up dependent outcome capture

Eleven studies reported morbidity outcomes in sufficient detail for extraction and were stratified according to resource setting and follow-up type [Table 3a and b]. A descriptive distribution of reported morbidity events within each resource group is presented in Table 4.

Table 3b.

Morbidity across studies in lower-resource settings

Study (year)	Country	Resource setting	Follow-up type	Preoperative morbidity (n)	Intraoperative morbidity (n)	Postoperative morbidity (n)
Ahmed and Tareen, 2018[20]	Pakistan	Lower-resource	Includes follow-up	NR	NR	Early: Sepsis+pneumonia=13; anastomotic leak=3; aspiration pneumonia=1; pneumothorax=1; late: Regurgitation=6; dyspnea=2; stricture=4; GERD=2; tracheomalacia=1; hiatus hernia=1
Gupta, 2018[28]	India	Lower-resource	Includes follow-up	NR	NR	Wound infection=11; pneumonia=11; wound dehiscence=9; septicemia=7; minor leak=3; major leak=1; GERD=3; readmission: Pneumonia=11; stricture=4; septicemia=2; tracheomalacia=1; minor leak=1; major leak=1
Hatti et al., 2020[24]	India	Lower-resource	Perioperative only	NR	NR	Culture-positive sepsis=9 + 4; anastomotic leak=4 + 2
Jagdeesh Prasad, 2017[25]	India	Lower-resource	Perioperative only	NR	NR	Septicemia=18; Leak=5

GERD: Gastroesophageal reflux disease, TEF: Tracheoesophageal fistula, NR: Not reported

Table 3a.

Morbidity across studies in higher-resource settings

Study (year)	Country	Resource setting	Follow-up type	Preoperative morbidity (n)	Intraoperative morbidity (n)	Postoperative morbidity (n)
Chiang et al. (2023)[7]	Singapore	Higher-resource	Includes follow-up	NR	NR	Anastomotic stricture=20; leak=4; recurrent TEF=3; GERD=28; dysphagia=20; esophagitis=4; chronic cough=19; pneumonia=19; recurrent RTI=23; asthma/BHR=13; bronchiectasis=3; tracheomalacia=22
Hattori et al., 2024[27]	Japan	Higher-resource	Includes follow-up	NR	NR	Anastomotic leak=5; recurrent fistula=2; stricture=8; fundoplication=7; scoliosis=3; chest wall asymmetry=5; rib deformity=7
Lu et al., 2021[13]	Taiwan	Higher-resource	Includes follow-up	NR	NR	GERD=39; stricture (2 years postoperative)=22
Murase et al., 2015[14]	Japan	Higher-resource	Includes follow-up	NR	NR	Anastomotic leak=2; GER=15
Okuyama et al., 2018[16]	Japan	Higher-resource	Includes follow-up	NR	NR	Stricture=1; fundoplication (GERD)=2
Syed et al., 2020[18]	Saudi Arabia	Higher-resource	Includes follow-up	NR	NR	GERD=6; septicemia=2; leak=2; lung collapse=1
Zhang et al. (2020)[19]	China	Higher-resource	Includes follow-up	NR	NR	Severe pneumonia=8 + 23 (Groups A/B); poor wound healing=2 + 14; chest wall deformity=0 + 11; pneumothorax=8 + 7; leak=15+13; stenosis=19+23; GERD=4 + 2

TEF: Tracheoesophageal fistula, GERD: Gastroesophageal reflux disease, BHR: Bronchial hyperresponsiveness, RTI: Respiratory tract infection, NR: Not reported

Table 4.

Aggregated morbidity summary

Morbidity category	Higher-resource, n (%)	Lower-resource, n (%)
Gastroesophageal reflux	94 (26.4)	5 (4.6)
Anastomotic stricture	93 (26.2)	8 (7.4)
Pneumonia	50 (14.0)	12 (11.1)
Anastomotic leak	41 (11.5)	18 (16.7)
Septicemia	2 (0.6)	51 (47.2)
Tracheomalacia	22 (6.2)	2 (1.9)
Recurrent respiratory tract infection	23 (6.5)	0
Pneumothorax	15 (4.2)	1 (0.9)
Wound infection	0	11 (10.2)
Poor wound healing	16 (4.5)	0
Total morbidity events	356 (100.0)	108 (100.0)

Percentages use the total number of reported morbidity events within each resource group as the denominator (higher-resource 356 events; lower-resource 108 events). Counts are derived from Table 3 and reflect reported events rather than patient-level risk; distributions are sensitive to heterogeneity in definitions, follow-up duration, and surveillance intensity

In higher-resource settings, the most frequently reported morbidity categories were gastroesophageal reflux (94 events, 26%) and anastomotic stricture (93 events, 26%), followed by pneumonia (50 events, 14%) and anastomotic leak (41 events, 11%). Respiratory sequelae such as recurrent respiratory tract infection (23 events, 7%) and tracheomalacia (22 events, 6%) were also reported [Table 4].

In lower-resource settings, reported morbidity events were dominated by septicemia (51 events, 47%), followed by anastomotic leak (18 events, 17%), pneumonia (12 events, 11%), and wound infection (11 events, 10%) [Table 4].

These aggregated proportions represent descriptive distributions of reported morbidity events rather than pooled prevalence estimates, because follow-up duration, surveillance intensity, and outcome definitions varied substantially across studies.

Population-based incidence

Six population-based studies reported incidence estimates for EA in Asian settings [Table 5]. Reported incidence ranged from 6.17 to 20.30 per 100,000 live births.

The lowest estimate was reported from Uzbekistan (6.17 per 100,000), while higher estimates were reported from China (20.30 per 100,000) and Korea (17.38 per 100,000). Intermediate estimates were reported from Iran (16.67 per 100,000) and China (16.50 per 100,000), while Thailand reported 8.80 per 100,000 live births.

Several incidence estimates were derived from small case counts in surveillance datasets, which may produce unstable rate estimates and should therefore be interpreted cautiously.

Risk-of-bias summary

Risk-of-bias assessment identified recurrent concerns related to selection and referral processes in hospital-based cohorts, incomplete capture of early neonatal deaths prior to referral or surgery, and differential detection of associated anomalies due to variation in diagnostic work-up. Long-term morbidity outcomes were additionally affected by loss to follow-up, which may lead to underestimation of later complications. Overall risk-of-bias judgments across domains are summarized in Figure 2, with study-level assessments presented in Supplementary Table 1.

DISCUSSION

This systematic narrative review synthesized 34 studies including 3304 infants with EA across Asian settings, separating population-based incidence estimates from hospital-based outcome cohorts to improve interpretability. Population-based incidence ranged from 6.17 to 20.30 per 100,000 live births, indicating substantial variation across countries and surveillance systems. Similar variability has been documented in other population-based studies, where reported incidence closely reflects registry coverage, inclusion criteria, and diagnostic ascertainment rather than true epidemiologic differences.[2,33] The wide range observed across Asian settings therefore likely reflects differences in surveillance completeness, prenatal detection, and case definition rather than biological variation. In addition, several incidence estimates were derived from small case counts within surveillance datasets, which may produce unstable rate estimates and should therefore be interpreted cautiously.

In hospital-based reporting, associated anomalies were common. Gastrointestinal anomalies comprised 36% of all reported associated anomalies, followed by cardiovascular anomalies at 27%, while multisystem patterns such as VACTERL association and other multisystem anomalies together accounted for more than 22% of reported anomalies. As anomaly counts were extracted as reported events across studies, these distributions represent the frequency of reported anomaly types rather than the prevalence of anomalies at the patient level. These distributions are consistent with findings from multiple regional cohorts, where 40%–60% of infants with EA have at least one associated anomaly, and cardiac defects are most frequently identified.[4,5,34] These findings reaffirm that EA is frequently part of a broader congenital disease spectrum and that the apparent anomaly profile is highly dependent on the intensity and completeness of diagnostic evaluation.

A key observation of this review was the wide range of reported mortality and morbidity across Asian settings, which should be interpreted in relation to health system factors affecting case capture, perioperative support, and follow-up rather than as intrinsic differences between the regions. Mortality in hospital-based cohorts ranged from 2% to 83%. Higher-resource settings consistently reported single-digit to low-teen mortality, comparable to outcomes reported in long-established tertiary centers where survival exceeds 85%–90%.[1] By contrast, substantially higher mortality was reported in several lower-resource cohorts, particularly from South Asia. Interpretation of mortality differences should also consider variation in reported mortality definitions and observation periods across studies, including inhospital, perioperative, and unspecified postoperative mortality.

Similar patterns have been documented in other LMIC settings, where mortality frequently exceeds 40%–50% and is strongly associated with delayed presentation, low birth weight, prematurity, and sepsis.[6,35,36] Importantly, early neonatal deaths prior to referral or surgery are often underrepresented in hospital-based series, potentially inflating apparent survival in referral centers and limiting cross-setting comparability.[3]

Morbidity profiles also differed across settings but were strongly influenced by follow-up duration and surveillance intensity. In higher-resource settings, reported morbidity was dominated by longer-term gastrointestinal and respiratory sequelae, including gastroesophageal reflux, anastomotic stricture, dysphagia, and recurrent respiratory tract infections. Long-term follow-up data from Singapore demonstrated that more than 60% of survivors experienced ongoing gastrointestinal or respiratory morbidity,[7] consistent with international reviews highlighting the chronic disease burden among survivors.[37] In contrast, lower-resource cohorts primarily reported early postoperative complications, with septicemia and wound infection accounting for a large proportion of morbidity events. This pattern is consistent with reports from neonatal surgical literature in lower-resource settings. As follow-up duration and surveillance intensity differed across studies, the reported morbidity distributions likely reflect differences in outcome capture rather than directly comparable complication rates.[38,39]

The observed burden of early infectious morbidity and high mortality in lower-resource cohorts supports prioritizing pragmatic system improvements across the continuum of care. Delayed presentation beyond 24–48 h of life has been consistently associated with increased mortality in EA and other neonatal surgical emergencies.[20,35] Strengthening referral pathways, neonatal transport, and early stabilization is therefore likely to yield substantial survival gains.

Standardization of perioperative care and infection prevention is another high-impact target. Sepsis remains the leading cause of neonatal surgical mortality in many Asian and African settings, accounting for up to 40%–50% of deaths.[40,41] Even modest improvements in perioperative protocols and neonatal intensive care access have been associated with dramatic reductions in mortality over time.[33]

Pragmatic action framework for low- and middle-income Asian settings

To facilitate translation into practice, we propose a pragmatic action framework for LMIC Asian settings, summarized in Figure 3. The framework integrates antenatal detection, neonatal stabilization and safe transport, standardized perioperative care, and longitudinal follow-up within enabling health-system structures.

Figure 3.

Action framework to reduce esophageal atresia morbidity and mortality in the Global South. EA: Esophageal atresia

Similar system-oriented interventions have been associated with reductions in anastomotic leak and mortality, including targeted perioperative adjuncts and improved neonatal intensive care support. The framework is intentionally modular, allowing phased implementation prioritizing early stabilization, infection prevention, and safe referral before expansion to comprehensive long-term follow-up. These findings should therefore be interpreted primarily as descriptive patterns across heterogeneous health system contexts rather than direct quantitative comparisons between the settings.

Study limitations

Several limitations constrain inference from the available evidence. Most included studies were hospital-based tertiary series and therefore vulnerable to referral and selection bias, with undercapture of early neonatal deaths prior to referral or surgery. Detection bias is likely in associated anomaly reporting due to variable diagnostic work-up, particularly for cardiac and genetic anomalies. Outcome definitions were heterogeneous, with inconsistent mortality time windows and follow-up duration, and long-term morbidity was susceptible to loss to follow-up. Incomplete reporting of key case-mix variables such as gestational age, birth weight, and disease severity further limits risk adjustment. Associated anomaly counts were also reported as aggregated events across studies; the proportion of infants with at least one associated anomaly could not be consistently determined. These limitations support descriptive synthesis and system-level interpretation rather than causal attribution of outcome differences between the settings.

Future directions

Future research should prioritize prospective, standardized multicenter registries capturing all diagnosed EA cases, including early deaths, with harmonized definitions for mortality and morbidity. Implementation research evaluating feasible bundles for neonatal stabilization, infection prevention, and safe transport in lower-resource settings is urgently needed, alongside sustainable follow-up programs to quantify long-term gastrointestinal and respiratory morbidity.

CONCLUSION

This systematic narrative review synthesized 34 Asian studies including 3304 infants with EA, separating population-based incidence estimates from hospital-based outcome cohorts. Population-based studies reported incidence ranging from 6.17 to 20.30 per 100,000 live births. Hospital-based reports showed frequent associated anomalies, most commonly gastrointestinal (36%) and cardiovascular (27%), and wide variation in mortality (2% to 83%), alongside substantial postoperative morbidity; higher-resource cohorts more often reported longer-term gastrointestinal and respiratory sequelae, while lower-resource cohorts reported a high proportion of early infectious complications including septicemia and wound infection.

These findings should be interpreted descriptively given referral and selection processes in tertiary series, incomplete capture of early deaths, variable anomaly detection, heterogeneous outcome definitions, and limited follow-up. Priorities to reduce preventable mortality and morbidity include strengthening prenatal detection and referral networks, improving neonatal stabilization and safe transport, standardizing perioperative care with infection prevention, and ensuring accessible long-term follow-up. The proposed action framework [Figure 3] summarizes pragmatic steps for implementation in resource-constrained Asian settings.

Conflicts of interest

There are no conflicts of interest.

Funding Statement

Nil.