BACKGROUND AND OBJECTIVES:

Presence of a syndrome (or association) is predictive of poor survival in esophageal atresia (EA). However, most reports rely on historical patient outcomes, limiting their usefulness when estimating risk for neonates born today. We hypothesized improved syndromic EA survival due to advances in neonatal care.

METHODS:

A retrospective single-center review of survival in 626 consecutive patients with EA from 1980 to 2017 was performed. Data were collected for recognized risk factors: preterm delivery; birth weight <1500 g; major cardiac disease; vertebral defects, anal atresia, cardiac defects, tracheoesophageal fistula, renal anomalies, and limb abnormalities (VACTERL); and non-VACTERL syndromes. Cox proportional hazards regression models were used to evaluate temporal trends in survival with respect to year of birth and syndromic EA.

RESULTS:

Overall, 87% of 626 patients with EA survived, ranging from 82% in the 1980s to 91% in the 2010s. After adjusting for confounders, syndromic EA survival did not improve during the study, with no association found between year of birth and survival (hazard ratio [HR] 0.98, 95% confidence interval [CI]: 0.95–1.01). Aside from lethal non-VACTERL syndromes, patients with nonlethal non-VACTERL syndromes (HR 6.85, 95% CI: 3.50–13.41) and VACTERL syndrome (HR 3.02, 95% CI: 1.66–5.49) had a higher risk of death than those with nonsyndromic EA.

CONCLUSIONS:

Survival of patients with syndromic EA has not improved, and patients with non-VACTERL syndromes have the highest risk of death. Importantly, this is independent of syndrome lethality, birth weight, and cardiac disease. This contemporary survival assessment will enable more accurate perinatal counseling of parents of patients with syndromic EA.

What’s Known on This Subject:

The presence of a syndrome (or association) is reported to predict poor survival in patients with esophageal atresia (EA). However, most reports rely on historical patient outcomes, limiting their usefulness when estimating risk for neonates born today.

What This Study Adds:

This study reveals survival of patients with syndromic EA remains poor and non–VACTERL syndromes have the highest associated risk of death. This contemporary survival assessment will enable more accurate perinatal counseling of parents of patients with syndromic EA.

The relationship between esophageal atresia (EA), with or without an associated tracheoesophageal fistula (TEF), and syndromes or associations is well described. Often termed syndromic EA, this subgroup is widely held to have poorer outcomes compared with patients born with nonsyndromic EA.1,2  Indeed, in historical case series, survival rates of 30% or less are reported in patients with EA and syndromes such as trisomy 21 and colobomas, heart defects, atresia of the nasal choanae, restriction of development, genitourinary abnormalities, and ear and hearing abnormalities (CHARGE syndrome).35  However, these survival estimates are derived from EA case series dating back to the 1960s, an era with expertise and philosophies for the treatment of patients with syndromic EA management that are vastly different from today.

Therefore, this begs the following question: what is the expected survival for patients born today with syndromic EA? We hypothesized that contemporary survival of syndromic EA would have improved from the dire estimates cited above on the basis of 3 observations. Firstly, previous studies have revealed that overall outcome for EA has improved during the late 20th century, which is attributed to advances in neonatal intensive, anesthetic, and surgical care.68  However, these studies often do not include or detail the outcomes specific to patients with syndromic EA. Secondly, survival has increased in subgroups of patients with EA previously considered at higher risk, for example, patients with low birth weight, major cardiac disease, or vertebral defects, anal atresia, cardiac defects, tracheoesophageal fistula, renal anomalies, and limb abnormalities (VACTERL) association.9,10  Finally, overall survival has also improved for these and other anomalies or syndromes associated with EA (eg, trisomy 21 and CHARGE syndrome).1114 

To address our hypothesis, in this study, we aimed to compare the survival of patients with EA over a period of almost 40 years with specific reference to syndromic EA. In addition, in the study, we aimed to investigate the association between the presence of a syndrome and other risk factors and death.

A retrospective cohort study was undertaken of all patients with EA treated as a neonate at The Royal Children’s Hospital (RCH), Melbourne, from January 1980 to December 2017, inclusive. During the study period, RCH was the principal site of neonatal EA treatment of the Australian states of Victoria and Tasmania, serving a population of 6 million. Patients were identified by using the Nate Myers Oesophageal Atresia Database, which prospectively maintains records for all live-born patients with EA managed at RCH since 1948. Additional data were obtained from 2 institutional databases: Cardiobase (Magnus Medical Software, Victoria, Australia), and Operating Room Management Information System (iSoft, Banbury, United Kingdom).

The study population comprised 626 consecutive patients with EA. No eligible patients were excluded. These patients were divided into 4 epochs: 1980–1989 (n = 147), 1990–1999 (n = 156), 2000–2009 (n = 168), and 2010–2017 (n = 155) (Table 1).

TABLE 1

Patient Characteristics by Epochs

Overall (N = 626), n (%)1980s (n = 147), n (%)1990s (n = 156), n (%)2000s (n = 168), n (%)2010s (n = 155), n (%)
Survival      
 Alive 543 (87) 120 (82) 134 (86) 148 (88) 141 (91) 
 Deceased 83 (13) 27 (18) 22 (14) 20 (12) 14 (9) 
EA type      
 EA with distal TEF 526 (84) 124 (84) 129 (83) 140 (83) 133 (86) 
 EA only 42 (7) 4 (3) 17 (11) 11 (7) 10 (6) 
 TEF only 33 (5) 14 (10) 6 (4) 6 (4) 7 (5) 
 EA with proximal TEF 13 (2) 3 (2) 2 (1) 5 (3) 3 (2) 
 EA with double TEF 12 (2) 2 (1) 2 (1) 6 (4) 2 (1) 
Syndromic group      
 Nonsyndromic 431 (69) 101 (69) 106 (68) 118 (70) 106 (68) 
 VACTERL 136 (22) 25 (17) 33 (21) 39 (23) 39 (25) 
 Nonlethal non-VACTERL 47 (8) 15 (10) 12 (8) 11 (7) 9 (6) 
 Lethal non-VACTERL 12 (2) 6 (4) 5 (3) 0 (0) 1 (1) 
Gestationa      
 Preterm 224 (36) 50 (34) 54 (35) 64 (38) 56 (36) 
 Term 395 (63) 92 (63) 100 (64) 104 (62) 99 (64) 
Sex      
 Male 378 (60) 87 (59) 85 (54) 106 (63) 100 (65) 
 Female 248 (40) 60 (41) 71 (46) 62 (37) 55(35) 
Birth weight, gb      
 ≥1500 565 (90) 132 (90) 139 (89) 149 (89) 145 (94) 
 <1500 53 (8) 11 (7) 14 (9) 19 (11) 9 (6) 
Cardiac diseasec      
 Severe 77 (12) 17 (12) 20 (13) 24 (14) 16 (10) 
 Nonsevere 155 (25) 26 (18) 38 (24) 45 (27) 46 (30) 
 None 388 (62) 103 (70) 98 (63) 99 (59) 88 (57) 
Overall (N = 626), n (%)1980s (n = 147), n (%)1990s (n = 156), n (%)2000s (n = 168), n (%)2010s (n = 155), n (%)
Survival      
 Alive 543 (87) 120 (82) 134 (86) 148 (88) 141 (91) 
 Deceased 83 (13) 27 (18) 22 (14) 20 (12) 14 (9) 
EA type      
 EA with distal TEF 526 (84) 124 (84) 129 (83) 140 (83) 133 (86) 
 EA only 42 (7) 4 (3) 17 (11) 11 (7) 10 (6) 
 TEF only 33 (5) 14 (10) 6 (4) 6 (4) 7 (5) 
 EA with proximal TEF 13 (2) 3 (2) 2 (1) 5 (3) 3 (2) 
 EA with double TEF 12 (2) 2 (1) 2 (1) 6 (4) 2 (1) 
Syndromic group      
 Nonsyndromic 431 (69) 101 (69) 106 (68) 118 (70) 106 (68) 
 VACTERL 136 (22) 25 (17) 33 (21) 39 (23) 39 (25) 
 Nonlethal non-VACTERL 47 (8) 15 (10) 12 (8) 11 (7) 9 (6) 
 Lethal non-VACTERL 12 (2) 6 (4) 5 (3) 0 (0) 1 (1) 
Gestationa      
 Preterm 224 (36) 50 (34) 54 (35) 64 (38) 56 (36) 
 Term 395 (63) 92 (63) 100 (64) 104 (62) 99 (64) 
Sex      
 Male 378 (60) 87 (59) 85 (54) 106 (63) 100 (65) 
 Female 248 (40) 60 (41) 71 (46) 62 (37) 55(35) 
Birth weight, gb      
 ≥1500 565 (90) 132 (90) 139 (89) 149 (89) 145 (94) 
 <1500 53 (8) 11 (7) 14 (9) 19 (11) 9 (6) 
Cardiac diseasec      
 Severe 77 (12) 17 (12) 20 (13) 24 (14) 16 (10) 
 Nonsevere 155 (25) 26 (18) 38 (24) 45 (27) 46 (30) 
 None 388 (62) 103 (70) 98 (63) 99 (59) 88 (57) 
a

Missing gestation, n = 7.

b

Missing birth weight, n = 8.

c

Missing cardiac disease status, n = 6.

To enable valid comparisons with key related literature, the 1994 definition by Spitz et al15  was adopted for major cardiac disease, namely “either cyanotic congenital heart disease that required palliative or corrective surgery or noncyanotic congenital heart disease that required medical or surgical treatment for cardiac failure. This excluded patent ductus arteriosus, unless it required surgical ligation.” Current consensus definitions were used for VACTERL association and preterm gestation (<37 completed weeks’ gestation; International Classification of Diseases, 10th Revision codes O60 and P07.3).16,17 

Regarding cause of death, late deaths were defined as those occurring after primary discharge.18  Deaths recorded as sudden infant death syndrome (SIDS) follow international consensus definitions, namely “the sudden unexpected death of an infant <1 year of age, with onset of the fatal episode apparently occurring during sleep, that remains unexplained after a thorough investigation, including performance of a complete autopsy and review of the circumstances of death and the clinical history.”19  For the purposes of our cohort analyses, deaths due to SIDS were grouped together with those sudden, unexpected, and otherwise unexplained deaths in children ≥1 year of age.

The primary outcome was patient survival. Data from the Nate Myers Oesophageal Atresia Database were cross-checked against state death records to ensure late and out-of-hospital deaths were also captured. Data were censored on the final cross-checking (May 10, 2018). Data were collected for each risk factor identified through literature review of EA prognostic classification systems, namely the following: preterm delivery, birth weight <1500 g, major cardiac disease, VACTERL association, and non-VACTERL syndromes, including chromosomal abnormalities.15,2022  These risk factors were considered to be potential confounders for the survival analyses outlined below.

Cox proportional hazards regression models were constructed to evaluate temporal trends in survival, adjusting for the aforementioned potential confounders. Temporal trends were tested by using year of birth as a continuous variable. A sensitivity analysis was performed, with year of birth as a categorical variable (1980–1989, 1990–1999, 2000–2009, 2010–2017). A second Cox proportional hazards regression model was developed to investigate temporal trends, specifically in the syndromic EA group (ie, restricted to the 195 syndromic cases). Finally, a separate Cox proportional hazards regression model was developed to investigate the association between syndromic EA subgroups and other risk factors, and death. Cases of lethal non-VACTERL EA (n = 12) were excluded from all Cox proportional hazards models because of the lethality (median [interquartile range (IQR)] time to death: 3.5 days [0–16 days]). The assumption of proportionality was evaluated by using a log–log plot of survival and tested by using Schoenfeld residuals; this was met in all models. Hazard ratios (HRs) and 95% confidence intervals (CIs) are reported. The Stata statistics package (version 14.2; Stata Corp, College Station, TX) was used. A value of P < .05 was considered significant.

This study was approved by the RCH Human Research Ethics Committee (DA014-2014-02).

A total of 626 patients with EA were identified during the study period. There were 378 (60%) boys and 248 (40%) girls. The median gestational age was 38 (IQR 35–39) weeks. The median birth weight was 2670 (IQR 2046–3145) g. Overall mortality was 13% (83 of 626). The characteristics of the patient cohorts by epoch are summarized in Table 1. Thirty-one percent (195 of 626) of patients with EA were syndromic: 21% (136 of 626) VACTERL association, 8% (47 of 626) nonlethal non-VACTERL, and 2% (12 of 626) lethal non-VACTERL; see Table 2 for specific diagnoses.

TABLE 2

Survival in Non-VACTERL Syndromic EA (n = 59)

1980s1990s2000s2010s
AllSurvival, n (%)AllSurvival, n (%)AllSurvival, n (%)AllSurvival, n (%)
Lethal         
 Trisomy 18 0 (0) 15 0 (0)  0 (0) 
Nonlethal 15 6 (40) 12 9 (75) 11 6 (55) 7 (78) 
 Trisomy 21 0 (0) 3 (100) 1 (33) 1 (50) 
 CHARGE syndrome 3 (50) 2 (50) 2 (50) 1 (50) 
 Tetralogy of Fallot 3 (100) 2 (100) 1 (100) 1 (100) 
 Fanconi anemia 0 (0) — 0 (0) — 
 Pierre Robin syndrome 0 (0) — — — 
 Goldenhar syndrome — 1 (50) 1 (100) — 
 Klippel-Feil syndrome — 1 (100) — — 
 Feingold syndrome — — 1 (100) 1 (100) 
 Type 3 Duane syndrome — — — 1 (100) 
 Apert syndrome — — — 1 (100) 
 MDM — — — 1 (100) 
1980s1990s2000s2010s
AllSurvival, n (%)AllSurvival, n (%)AllSurvival, n (%)AllSurvival, n (%)
Lethal         
 Trisomy 18 0 (0) 15 0 (0)  0 (0) 
Nonlethal 15 6 (40) 12 9 (75) 11 6 (55) 7 (78) 
 Trisomy 21 0 (0) 3 (100) 1 (33) 1 (50) 
 CHARGE syndrome 3 (50) 2 (50) 2 (50) 1 (50) 
 Tetralogy of Fallot 3 (100) 2 (100) 1 (100) 1 (100) 
 Fanconi anemia 0 (0) — 0 (0) — 
 Pierre Robin syndrome 0 (0) — — — 
 Goldenhar syndrome — 1 (50) 1 (100) — 
 Klippel-Feil syndrome — 1 (100) — — 
 Feingold syndrome — — 1 (100) 1 (100) 
 Type 3 Duane syndrome — — — 1 (100) 
 Apert syndrome — — — 1 (100) 
 MDM — — — 1 (100) 

There were no cases of trisomy 13 in any of the epochs examined. MDM, mandibular dysostosis microcephaly (EFTUD2 gene mutation); —, not applicable.

Overall, 543 of 626 (87%) patients with EA were alive at the time of data collection. Table 3 details the survival by epochs, as well as the number of deaths that occurred in the first 3 months of life and after 1 year of age. All cases of lethal non-VACTERL syndromic EA were in live-born patients with EA and trisomy 18. At the time of data collection, 140 of 195 (72%) patients with syndromic EA were alive. The survival rates for specific syndromes in each cohort are summarized (Table 2).

TABLE 3

Survival Details by Time Epochs (n = 626)

1980s, n (%)1990s, n (%)2000s, n (%)2010s, n (%)
Alive 120 of 147 (82) 134 of 156 (86) 148 of 168 (88) 141 of 155 (91) 
 All syndromic 26 of 46 (57) 34 of 50 (68) 39 of 50 (78) 41 of 49 (84) 
  VACTERL 20 of 25 (80) 25 of 33 (76) 33 of 39 (85) 34 of 39 (87) 
  Nonlethal non-VACTERL 6 of 15 (40) 9 of 12 (75) 6 of 11 (55) 7 of 9 (78) 
  Lethal non-VACTERL 0 of 6 (0) 0 of 5 (0) — 0 of 1 (0) 
Deceased 27 of 147 (18) 22 of 156 (14) 20 of 168 (12) 14 of 155 (9) 
 Death in the first 3 mo 21 of 27 (78) 18 of 22 (82) 12 of 20 (60) 8 of 14 (57) 
 Death after 1 y of age 3 of 27 (11) 2 of 22 (9) 3 of 20 (15) 2 of 14 (14) 
1980s, n (%)1990s, n (%)2000s, n (%)2010s, n (%)
Alive 120 of 147 (82) 134 of 156 (86) 148 of 168 (88) 141 of 155 (91) 
 All syndromic 26 of 46 (57) 34 of 50 (68) 39 of 50 (78) 41 of 49 (84) 
  VACTERL 20 of 25 (80) 25 of 33 (76) 33 of 39 (85) 34 of 39 (87) 
  Nonlethal non-VACTERL 6 of 15 (40) 9 of 12 (75) 6 of 11 (55) 7 of 9 (78) 
  Lethal non-VACTERL 0 of 6 (0) 0 of 5 (0) — 0 of 1 (0) 
Deceased 27 of 147 (18) 22 of 156 (14) 20 of 168 (12) 14 of 155 (9) 
 Death in the first 3 mo 21 of 27 (78) 18 of 22 (82) 12 of 20 (60) 8 of 14 (57) 
 Death after 1 y of age 3 of 27 (11) 2 of 22 (9) 3 of 20 (15) 2 of 14 (14) 

—, no cases.

A Cox proportional hazards regression model was used to examine temporal trends in survival of patients with EA over the study period. Considering all patients with EA in the cohort, there was no change in survival over the study period in either univariate analysis (HR 0.99, 95% CI: 0.97–1.01, P = .47), or after adjusting for the confounders specified previously (HR 1.00, 95% CI: 0.98–1.02, P = .92; Table 4). In sensitivity analysis, year of birth (as a categorical variable) was also not associated with survival (Supplemental Table 8).

TABLE 4

Multivariable Model Investigating Temporal Trends in Mortality in EA (n = 598)

VariableHR (95% CI)P
Year of birth 1.00 (0.98–1.02) .92 
Syndromic group   
 Nonsyndromic Reference — 
 VACTERL 3.03 (1.67–5.50) <.001 
 Nonlethal non-VACTERL 7.40 (3.72–14.72) <.001 
Gestation period   
 Term Reference — 
 Preterm (<37 wk) 2.38 (1.36–4.16) .002 
Birth weight, g   
 ≥1500 Reference — 
 <1500 2.41 (1.31–4.43) .005 
Major cardiac disease   
 None Reference — 
 Nonsevere 0.53 (0.25–1.08) .08 
 Severe 1.13 (0.61–2.09) .69 
VariableHR (95% CI)P
Year of birth 1.00 (0.98–1.02) .92 
Syndromic group   
 Nonsyndromic Reference — 
 VACTERL 3.03 (1.67–5.50) <.001 
 Nonlethal non-VACTERL 7.40 (3.72–14.72) <.001 
Gestation period   
 Term Reference — 
 Preterm (<37 wk) 2.38 (1.36–4.16) .002 
Birth weight, g   
 ≥1500 Reference — 
 <1500 2.41 (1.31–4.43) .005 
Major cardiac disease   
 None Reference — 
 Nonsevere 0.53 (0.25–1.08) .08 
 Severe 1.13 (0.61–2.09) .69 

—, not applicable.

Separate analysis was used to examine for temporal trends in the survival of patients with syndromic EA as a specific subgroup. This subgroup analysis revealed no significant difference in survival of patients with syndromic EA over the study period, including after adjusting for confounders (HR 0.98, 95% CI: 0.95–1.01, P = .16; Table 5). In sensitivity analysis, year of birth (as a categorical variable), was also not associated with survival of syndromic EA (Supplemental Table 9). Together, these results confirm no change in survival of patients with syndromic EA over the study period. Kaplan-Meier survival estimates are revealed in Fig 1.

TABLE 5

Multivariable Model Investigating Temporal Trends in Syndromic EA (Restricted to Syndromic EA Group Only, n = 177)

VariableHR (95% CI)P
Year of birth 0.98 (0.95–1.01) .16 
Gestation period   
 Term Reference — 
 Preterm (<37 wk) 1.64 (0.83–3.22) .15 
Birth weight, g   
 ≥1500 Reference — 
 <1500 1.66 (0.67–4.08) .27 
Major cardiac disease   
 None Reference — 
 Nonsevere 0.56 (0.26–1.27) .17 
 Severe 0.81 (0.39–1.66) .69 
VariableHR (95% CI)P
Year of birth 0.98 (0.95–1.01) .16 
Gestation period   
 Term Reference — 
 Preterm (<37 wk) 1.64 (0.83–3.22) .15 
Birth weight, g   
 ≥1500 Reference — 
 <1500 1.66 (0.67–4.08) .27 
Major cardiac disease   
 None Reference — 
 Nonsevere 0.56 (0.26–1.27) .17 
 Severe 0.81 (0.39–1.66) .69 

—, not applicable.

FIGURE 1

Kaplan-Meier survival estimates for patients with EA, compared by syndromic groups. The “no syndrome” group number at risk at time = 0 excludes n = 1 patient because of death at time = 0, that is, death on the first day of life. Cases of lethal non-VACTERL EA (n = 12) were excluded from survival analyses and so are not depicted here. Vertical lines indicate censored patients.

FIGURE 1

Kaplan-Meier survival estimates for patients with EA, compared by syndromic groups. The “no syndrome” group number at risk at time = 0 excludes n = 1 patient because of death at time = 0, that is, death on the first day of life. Cases of lethal non-VACTERL EA (n = 12) were excluded from survival analyses and so are not depicted here. Vertical lines indicate censored patients.

Close modal

Survival was reduced in all syndromic EA subgroups compared with their nonsyndromic EA counterparts (Table 6, Fig 1). Excluding lethal syndromes, patients with nonlethal non-VACTERL syndromic EA were at greatest risk, with an almost sevenfold increased risk of death compared with patients with nonsyndromic EA (HR 6.85, 95% CI: 3.50–13.41, P < .001). Patients with EA and VACTERL association were also at increased risk of death (HR 3.02, 95% CI: 1.66–5.49, P < .001; Table 6).

TABLE 6

Multivariable Model Investigating Association Between Syndromic Groups (and Other Covariates) and Mortality (n = 598)

VariableHR (95% CI)P
Syndromic group   
 Nonsyndromic Reference — 
 VACTERL 3.02 (1.66–5.49) <.001 
 Nonlethal non-VACTERL 6.85 (3.50–13.41) <.001 
Gestation period   
 Term Reference — 
 Preterm (<37 wk) 2.37 (1.36–4.15) .002 
Birth weight, g   
 ≥1500 Reference — 
 <1500 2.41 (1.31–4.42) .005 
Major cardiac disease   
 None Reference — 
 Nonsevere 0.52 (0.26–1.07) .077 
 Severe 1.13 (0.61–2.09) .69 
VariableHR (95% CI)P
Syndromic group   
 Nonsyndromic Reference — 
 VACTERL 3.02 (1.66–5.49) <.001 
 Nonlethal non-VACTERL 6.85 (3.50–13.41) <.001 
Gestation period   
 Term Reference — 
 Preterm (<37 wk) 2.37 (1.36–4.15) .002 
Birth weight, g   
 ≥1500 Reference — 
 <1500 2.41 (1.31–4.42) .005 
Major cardiac disease   
 None Reference — 
 Nonsevere 0.52 (0.26–1.07) .077 
 Severe 1.13 (0.61–2.09) .69 

—, not applicable.

Other factors significantly associated with reduced survival were birth weight <1500 g (HR 2.41, 95% CI: 1.31–4.42, P = .005) and preterm gestation (HR 2.37, 95% CI: 1.36–4.15, P = .002). Within this EA cohort, survival was not significantly reduced in patients with EA and major cardiac disease (HR 1.13, 95% CI: 0.61–2.09, P = .69; Table 6).

The cause of death was recorded in 79 of 83 (95%) of deceased patients with EA. The 4 patients for whom no cause of death was recorded were all “late” deaths, having all died after primary discharge. Causes of death were categorized in relation to key stages of EA surgical management, syndromic status, and pertinent underlying diagnoses (Fig 2). See also Supplemental Figs 36 for similar categorization of cause of death in patients with EA for each of the 4 study epochs.

FIGURE 2

Cause of EA death in relation to timing of surgical management, syndromic status, and diagnoses. “Late death” denotes death occurred after primary discharge; “palliated” indicates documented redirection of care to palliative measures and/or withdrawal of life-maintaining intensive care; SIDS is categorized here together with unexpected and unexplained deaths at >1 year of age. CNS, central nervous system.

FIGURE 2

Cause of EA death in relation to timing of surgical management, syndromic status, and diagnoses. “Late death” denotes death occurred after primary discharge; “palliated” indicates documented redirection of care to palliative measures and/or withdrawal of life-maintaining intensive care; SIDS is categorized here together with unexpected and unexplained deaths at >1 year of age. CNS, central nervous system.

Close modal

In our EA cohort, death was subsequent to a redirection of clinical care to palliative measures in 43 of 79 (54%) of cases with recorded cause of death (Fig 2). The proportion of deaths after palliation in each epoch ranged from 68% (17 of 25) in the 1980s to 43% (9 of 21) in the 1990s (Table 7). The majority (33 of 43 [77%]) of palliated patients with EA were also syndromic patients, and the majority of these patients (22 of 33 [67%]) were redirected to palliated measures before any surgery was performed (Table 7, Fig 2). Postoperatively, palliation for associated anomalies was associated with 4 (3 syndromic, 1 nonsyndromic) deaths after non-EA surgery, 5 (2 syndromic, 3 nonsyndromic) deaths after TEF ligation only, and 8 (6 syndromic, 2 nonsyndromic) deaths after both TEF ligation and esophageal anastomosis.

TABLE 7

Causes of EA Death in Relation to Time Epoch and Syndromic Status (n = 79)

1980s, n (%)1990s, n (%)2000s, n (%)2010s, n (%)
Palliated 17 of 25 (68) 9 of 21 (43) 10 of 20 (50) 6 of 13 (46) 
 Syndromic 14 of 17 (82) 9 of 9 (100) 5 of 10 (50) 4 of 6 (67) 
 Nonsyndromic 3 of 17 (18) — 5 of 10 (50) 2 of 6 (33) 
Postoperative complication 2 of 25 (8) 9 of 21 (43) 5 of 20 (25) 2 of 13 (15) 
 Syndromic 1 of 2 (50) 2 of 9 (22) 2 of 5 (40) 1 of 2 (50) 
 Nonsyndromic 1 of 2 (50) 7 of 9 (88) 3 of 5 (60) 1 of 2 (50) 
SIDS or unexplained ≥1 y 1 of 25 (4) — 1 of 20 (5) 3 of 13 (23) 
 Syndromic — — 1 of 1 (100) 2 of 3 (67) 
 Nonsyndromic 1 of 1 (100) — — 1 of 3 (33) 
Other late deatha 5 of 25 (25) 2 of 21 (10) 3 of 20 (15) 2 of 13 (15) 
 Syndromic 4 of 5 (80) — 2 of 3 (67) 1 of 2 (50) 
 Nonsyndromic 1 of 5 (20) 2 of 2 (100) 1 of 3 (33) 1 of 2 (50) 
1980s, n (%)1990s, n (%)2000s, n (%)2010s, n (%)
Palliated 17 of 25 (68) 9 of 21 (43) 10 of 20 (50) 6 of 13 (46) 
 Syndromic 14 of 17 (82) 9 of 9 (100) 5 of 10 (50) 4 of 6 (67) 
 Nonsyndromic 3 of 17 (18) — 5 of 10 (50) 2 of 6 (33) 
Postoperative complication 2 of 25 (8) 9 of 21 (43) 5 of 20 (25) 2 of 13 (15) 
 Syndromic 1 of 2 (50) 2 of 9 (22) 2 of 5 (40) 1 of 2 (50) 
 Nonsyndromic 1 of 2 (50) 7 of 9 (88) 3 of 5 (60) 1 of 2 (50) 
SIDS or unexplained ≥1 y 1 of 25 (4) — 1 of 20 (5) 3 of 13 (23) 
 Syndromic — — 1 of 1 (100) 2 of 3 (67) 
 Nonsyndromic 1 of 1 (100) — — 1 of 3 (33) 
Other late deatha 5 of 25 (25) 2 of 21 (10) 3 of 20 (15) 2 of 13 (15) 
 Syndromic 4 of 5 (80) — 2 of 3 (67) 1 of 2 (50) 
 Nonsyndromic 1 of 5 (20) 2 of 2 (100) 1 of 3 (33) 1 of 2 (50) 

Cause of death was recorded for 79 of 83 EA deaths, with all 4 deaths without a recorded cause occurring after primary discharge (ie, late deaths); 2 of 79 deaths due to complications of cardiac anomalies are not detailed in this table. —, no cases.

a

Other late deaths reported in this table represent an aggregate of late respiratory, cardiac, hematologic, and trauma deaths.

Surgical complications were another important cause of death, recorded in 19 of 79 (24%) patients with EA (Table 7). Of these, 5 of 19 (26%) related to complications of cardiac surgery, 13 of 19 (68%) to complications of EA surgery (either TEF ligation or esophageal anastomosis), and 1 of 19 (5%) to complications after other gastrointestinal surgery. The timing of these postoperative complications in relation to key stages of EA surgical management and underlying associated diagnoses are provided in Fig 2, with additional epoch-specific data provided in Supplemental Figs 36.

Late deaths, that is, those occurring after primary discharge, accounted for 17 of 79 (25%) of EA deaths. These included 5 cases of SIDS or otherwise unexpected and unexplained deaths in children ≥1 year of age. Three cases of SIDS were in infants previously diagnosed clinically with tracheomalacia, although none were considered to have severe tracheomalacia. Six further late deaths were due to respiratory causes, with all but one of these late respiratory deaths in patients with syndromic EA. All 6 of these patients were documented to have concomitant tracheomalacia, including 2 patients with severe tracheomalacia, of which 1 had undergone aortopexy and tracheostomy. One late respiratory death resulted from a catastrophic choking episode, with the remaining 5 deaths resulting from acute respiratory infections with or without associated aspiration. In addition, 3 late deaths resulted from circulatory causes, and there were single cases of late death for each of the following: hematologic, gastrointestinal, and trauma-related causes (Table 7, Fig 2).

With this large single-center study, we provide a contemporary estimate of survival in patients with EA, of direct relevance to antenatal counseling and individual patient care. Contrary to expectation, survival of patients with syndromic EA has not improved over the last 4 decades. This is striking and unexpected, given that concurrent advances in neonatal and surgical care have improved the overall survival of patients with these diagnoses and other high-risk neonates over the same period.1114,20 

One explanation for this paradox is additive, or even synergistic, morbidity: whereas neither EA nor a given syndrome would ordinarily be lethal in isolation, the combined morbidity of EA and the syndrome(s) is sufficient to increase the risk of death. Support for this explanation is provided by 2 series revealing that improvements in contemporary survival in either major cardiac disease or trisomy 21 are muted by the presence of any additional genetic, chromosomal, or major congenital structural abnormality.21,22  A similar phenomenon is evident in the EA survival data reported by Lopez et al9 . These authors observed improved contemporary outcomes in patients with EA and either low birth weight (<1500 g) or major cardiac disease (Spitz group II), but no improvement in survival for patients with EA and both low birth weight and major cardiac disease (Spitz group III).9  Another possible explanation is that, although we have a relatively large sample size, we may still be underpowered to detect differences in survival over time, particularly in the syndromic groups with smaller numbers. With this persisting poor prognosis in mind, earlier identification of patients with syndromic EA and earlier institution of a multidisciplinary team approach to management and follow-up may provide an avenue toward an improved prognosis for patients with syndromic EA.

We have also found that the greatest risk of death in our cohort occurred in patients with nonlethal non-VACTERL syndromes, which include trisomy 21 and CHARGE syndrome. This noteworthy finding contradicts the expectations and hypothesis outlined above that patients born today with EA and such syndromic comorbidities would have better survival than previously reported cohorts.35  We have considered whether this sustained poor survival in patients with EA and nonlethal non-VACTERL syndromes could be explained by the confounding influence of comorbidities. However, the related finding that major cardiac disease did not increase mortality in our EA cohort speaks against this alternative explanation. This finding also supports the concept of synergistic morbidity, as previously discussed.

In the current study, the risk of death was increased threefold in patients with EA and VACTERL association. Although this associated risk of death was less than for the aforementioned nonlethal non-VACTERL group, VACTERL association was the commonest diagnosis in the “syndromic” EA group, making the overall impact of this diagnosis on EA survival still of significance. The increased risk of death for patients with EA and VACTERL revealed here is seemingly at odds with the findings of Keckler et al,10  in which the authors concluded that VACTERL anomalies “appear to have little impact on [EA] survival.” However, Keckler et al10  measured the prevalence of VACTERL anomalies in patients with EA, rather than cases fulfilling the diagnostic criteria of VACTERL association.16  Moreover, the authors did not perform analyses for associations between survival and VACTERL anomalies. Instead, they reported survival rates without further qualification within a much smaller study population of 112 patients with EA.10 

Another distinction between this and previous EA survival studies is the relative influence of major cardiac disease on survival. This study has not revealed any significant differences for risk of death in patients with EA and major cardiac disease. This is contrary to other published studies, including recent studies by Yamoto et al,23  which aimed to update the clinically used Spitz classification, and Turner et al,24  which also validated the Spitz model. Although this may represent a true measure of risk of death, it is possible that a smaller proportion of patients with EA and major cardiac disease in our cohort may influence these differences. Additionally, a smaller proportion of patients with EA and low birth weight has resulted in fewer Spitz group III patients in our cohort, further skewing the influence of major cardiac disease on survival rates in this higher risk group.

This study has its limitations, most notably the biases inherent in its retrospective study design, albeit with data extracted from an internationally recognized, prospectively maintained database. For example, the presence of respiratory disease was not collected in the registry used in this study. This potential confounding factor is emphasized in several EA prognostic classifications,2527  although the Spitz criteria also overlooked respiratory disease.9,15  These limitations notwithstanding, several biases common to retrospective studies have been reduced by our study’s relatively large sample size, given the rarity of this condition and completeness of data. Selection bias is not considered likely in this study. Inclusion of all patients with EA at the principal state neonatal EA treatment center throughout the 38-year study period has ensured a largely unselected population.

Whereas patients with EA not surviving to transfer to our tertiary center were unable to be accounted for, state birth defect registry data indicate no more than 10% of all EA patients recorded in our state population died prenatally from 1983 to 2014.2831  These aggregate state birth defect registry data do not permit direct differentiation between termination rates for EA cases with or without accompanying syndrome. We can, however, discern that fewer terminations were performed in our population for related syndromes in the 2013–2014 reporting period when compared to 2007–2009. During the period 2013–2014, with a total of 57 715 reported births of ≥20 weeks’ gestation, there were 106 terminations performed for trisomy 21, 40 for trisomy 18, and <20 terminations for trisomy 13.31  This compares to 2007–2009, during which 145 809 reported births of ≥20 weeks’ gestation were reported, and 524 terminations were performed for trisomy 21, 212 for trisomy 18, and 76 for trisomy 13.30  We have considered whether accompanying trends in terminations for relevant antenatally diagnosed conditions may have influenced EA survival trends in our cohort. For example, if fetuses with EA that would have previously been terminated because of the presence of important antenatally diagnosed conditions (eg, trisomy 21, trisomy 18, trisomy 13) are now surviving to birth, might this increase the number of patients with EA at risk of postnatal death and so mask improved survival in patients with EA? On balance, we do not consider changes in termination rates to have notably biased our EA survival analyses. First, the proportions of patients in our cohort with VACTERL, nonlethal non-VACTERL syndromic or severe cardiac disease do not vary significantly between epochs. Second, our EA survival analyses will not have been skewed by variations in lethal non-VACTERL syndromic case numbers because these cases were excluded a priori from such trend analyses.

Attrition bias has been limited by prospective entry and ongoing maintenance of the Nate Myers Oesophageal Atresia Database records, including accurate survival data. In addition, the ability to cross-check hospital and state death records has reduced the likelihood of missed deaths. Finally, given the binary nature of survival outcomes (alive or deceased), information bias is unlikely to have skewed the results of this particular retrospective cohort study.

In this study, we have highlighted the significantly increased risk of death in syndromic EA. Despite many other advances in intensive and surgical care of complex neonates, the contemporary survival of patients with syndromic EA has not improved and remains poor. Patients with EA and non-VACTERL syndromes are at the greatest increased risk of death, and this risk estimation is sustained when lethal conditions are excluded and is independent of other risk factors. Other risk factors for death identified in this study were VACTERL association, preterm gestation, and birth weight <1500 g, but major cardiac disease was not a significant risk factor in our population. It is hoped that this study will inform accurate perinatal counseling and decision-making for practitioners and parents in the setting of present-day care of patients with syndromic EA.

We acknowledge the patients and pediatric surgery and newborn intensive care staff of RCH, Melbourne, represented in this study. The Nate Myers Oesophageal Atresia Database is maintained by funds from the Oesophageal Atresia Research Association, whose fundraising also supports the availability of specialty nurses and other patient and parent support initiatives.

Dr Tan Tanny interpreted the data and drafted and edited the manuscript; Dr Beck interpreted the data and reviewed the manuscript; Dr King reviewed and edited the manuscript; Ms Hawley and Ms Brooks maintained the database from which patients with esophageal atresia were identified, assisted in data acquisition, and reviewed the manuscript; Dr McLeod assisted in study design and reviewed the manuscript in its earlier and final draft iterations; Dr Hutson reviewed and edited the manuscript; Dr Teague conceptualized and designed the study, performed the data acquisition, interpreted the data, and reviewed, drafted, and edited the manuscript throughout its draft iterations; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

FUNDING: Dr Tan Tanny is supported by a National Health and Medical Research Council Medical Research Postgraduate Scholarship (1168142), a Clifford Family PhD Scholarship, and an Australian Government Research Training Program Scholarship. Dr Teague’s and Dr King’s positions as academic pediatric surgeons are supported by The Royal Children’s Hospital Foundation.

CHARGE syndrome

colobomas, heart defects, atresia of the nasal choanae, restriction of development, genitourinary abnormalities, and ear and hearing abnormalities

CI

confidence interval

EA

esophageal atresia

HR

hazard ratio

IQR

interquartile range

RCH

The Royal Children’s Hospital

SIDS

sudden infant death syndrome

TEF

tracheoesophageal fistula

VACTERL

vertebral defects, anal atresia, cardiac defects, tracheoesophageal fistula, renal anomalies, and limb abnormalities

1
Chittmittrapap
S
,
Spitz
L
,
Kiely
EM
,
Brereton
RJ
.
Oesophageal atresia and associated anomalies
.
Arch Dis Child
.
1989
;
64
(
3
):
364
368
2
Leonard
H
,
Barrett
AM
,
Scott
JE
,
Wren
C
.
The influence of congenital heart disease on survival of infants with oesophageal atresia
.
Arch Dis Child Fetal Neonatal Ed
.
2001
;
85
(
3
):
F204
F206
3
Beasley
SW
,
Allen
M
,
Myers
N
.
The effects of Down syndrome and other chromosomal abnormalities on survival and management in oesophageal atresia
.
Pediatr Surg Int
.
1997
;
12
(
8
):
550
551
4
Kutiyanawala
M
,
Wyse
RK
,
Brereton
RJ
, et al
.
CHARGE and esophageal atresia
.
J Pediatr Surg
.
1992
;
27
(
5
):
558
560
5
Beasley
SW
.
Influence of associated anomalies on the management of oesophageal atresia
.
Indian J Pediatr
.
1996
;
63
(
6
):
743
749
6
Engum
SA
,
Grosfeld
JL
,
West
KW
,
Rescorla
FJ
,
Scherer
LR
 III
.
Analysis of morbidity and mortality in 227 cases of esophageal atresia and/or tracheoesophageal fistula over two decades
.
Arch Surg
.
1995
;
130
(
5
):
502
508–509
7
Koivusalo
AI
,
Pakarinen
MP
,
Rintala
RJ
.
Modern outcomes of oesophageal atresia: single centre experience over the last twenty years
.
J Pediatr Surg
.
2013
;
48
(
2
):
297
303
8
Lilja
HE
,
Wester
T
.
Outcome in neonates with esophageal atresia treated over the last 20 years
.
Pediatr Surg Int
.
2008
;
24
(
5
):
531
536
9
Lopez
PJ
,
Keys
C
,
Pierro
A
, et al
.
Oesophageal atresia: improved outcome in high-risk groups?
J Pediatr Surg
.
2006
;
41
(
2
):
331
334
10
Keckler
SJ
,
St Peter
SD
,
Valusek
PA
, et al
.
VACTERL anomalies in patients with esophageal atresia: an updated delineation of the spectrum and review of the literature
.
Pediatr Surg Int
.
2007
;
23
(
4
):
309
313
11
Issekutz
KA
,
Graham
JM
 Jr.
,
Prasad
C
,
Smith
IM
,
Blake
KD
.
An epidemiological analysis of CHARGE syndrome: preliminary results from a Canadian study
.
Am J Med Genet A
.
2005
;
133A
(
3
):
309
317
12
Irving
C
,
Basu
A
,
Richmond
S
,
Burn
J
,
Wren
C
.
Twenty-year trends in prevalence and survival of Down syndrome
.
Eur J Hum Genet
.
2008
;
16
(
11
):
1336
1340
13
Kucik
JE
,
Shin
M
,
Siffel
C
,
Marengo
L
,
Correa
A
;
Congenital Anomaly Multistate Prevalence and Survival Collaborative
.
Trends in survival among children with Down syndrome in 10 regions of the United States
.
Pediatrics
.
2013
;
131
(
1
). Available at: www.pediatrics.org/cgi/content/full/131/1/e27
14
Oster
ME
,
Lee
KA
,
Honein
MA
,
Riehle-Colarusso
T
,
Shin
M
,
Correa
A
.
Temporal trends in survival among infants with critical congenital heart defects
.
Pediatrics
.
2013
;
131
(
5
). Available at: www.pediatrics.org/cgi/content/full/131/5/e1502
15
Spitz
L
,
Kiely
EM
,
Morecroft
JA
,
Drake
DP
.
Oesophageal atresia: at-risk groups for the 1990s
.
J Pediatr Surg
.
1994
;
29
(
6
):
723
725
16
Solomon
BD
.
VACTERL/VATER association
.
Orphanet J Rare Dis
.
2011
;
6
:
56
17
World Health Organization
.
International Statistical Classification of Diseases and Related Health Problems, 10th Revision
, 2nd ed.
Geneva, Switzerland
:
World Health Organization
;
2004
18
Tan Tanny
SP
,
Fearon
E
,
Hawley
A
, et al
.
Predictors of mortality after primary discharge from hospital in patients with Esophageal atresia
.
J Pediatr
.
2020
;
219
:
70
75
19
Krous
HF
,
Beckwith
JB
,
Byard
RW
, et al
.
Sudden infant death syndrome and unclassified sudden infant deaths: a definitional and diagnostic approach
.
Pediatrics
.
2004
;
114
(
1
):
234
238
20
Doyle
LW
;
Victorian Infant Collaborative Study Group
.
Evaluation of neonatal intensive care for extremely low birth weight infants in Victoria over two decades: I. Effectiveness
.
Pediatrics
.
2004
;
113
(
3 pt 1
):
505
509
21
Rankin
J
,
Tennant
PWG
,
Bythell
M
,
Pearce
MS
.
Predictors of survival in children born with Down syndrome: a registry-based study
.
Pediatrics
.
2012
;
129
(
6
). Available at: www.pediatrics.org/cgi/content/full/129/6/e1373
22
Wren
C
,
Irving
CA
,
Griffiths
JA
, et al
.
Mortality in infants with cardiovascular malformations
.
Eur J Pediatr
.
2012
;
171
(
2
):
281
287
23
Yamoto
M
,
Nomura
A
,
Fukumoto
K
, et al
.
New prognostic classification and managements in infants with esophageal atresia
.
Pediatr Surg Int
.
2018
;
34
(
10
):
1019
1026
24
Turner
B
,
Dasgupta
R
,
Brindle
ME
.
A contemporary prediction rule for esophageal atresia (EA) and tracheo-esophageal fistula (TEF)
.
J Pediatr Surg
.
2014
;
49
(
12
):
1758
1761
25
Waterston
DJ
,
Carter
RE
,
Aberdeen
E
.
Oesophageal atresia: tracheo-oesophageal fistula. A study of survival in 218 infants
.
Lancet
.
1962
;
1
(
7234
):
819
822
26
Poenaru
D
,
Laberge
JM
,
Neilson
IR
,
Guttman
FM
.
A new prognostic classification for esophageal atresia
.
Surgery
.
1993
;
113
(
4
):
426
432
27
Yagyu
M
,
Gitter
H
,
Richter
B
,
Booss
D
.
Esophageal atresia in Bremen, Germany–evaluation of preoperative risk classification in esophageal atresia
.
J Pediatr Surg
.
2000
;
35
(
4
):
584
587
28
Riley
M
,
Halliday
J
.
Birth Defects in Victoria 1983 - 1998
.
Melbourne, Australia
:
Victorian Perinatal Data Collection Unit
;
2000
29
Riley
M
,
Halliday
J
.
Birth Defects in Victoria 2005 - 2006
.
Melbourne, Australia
:
Victorian Perinatal Data Collection Unit
;
2008
30
Victorian Congenital Anomalies Register
.
Congenital Anomalies in Victoria 2007 - 2009
.
Melbourne, Australia
:
Consultative Council on Obstetric and Paediatric Mortality and Morbbidity (CCOPMM)
;
2017
31
Victorian Congenital Anomalies Register
.
Congenital Anomalies in Victoria 2013 - 2014
.
Melbourne, Australia
:
Consultative Council on Obstetric and Paediatric Mortality and Morbidity (CCOPMM)
;
2017

Competing Interests

POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.

FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.

Supplementary data