BACKGROUND:

Long-term outcomes in heterotaxy syndrome (HS) are poorly described. Some reports suggest improved survival in the recent era, whereas others do not. We sought to describe long-term outcomes and assess whether outcomes have changed over time.

METHODS:

Patients with HS born between 1985 and 2014 who had cardiac care (except initial palliation) at our institution were divided into 4 birth eras and survival over time was compared. Independent risk factors for mortality were identified by using Cox proportional hazards regression. In patients who underwent surgery, association between surgical pathway (univentricular versus biventricular repair) and mortality after adjusting for baseline confounders was evaluated. A risk stratification model was created by using classification and regression analysis.

RESULTS:

Among 264 patients, 118 (44.7%) had asplenia and 146 (55.3%) had polysplenia syndrome. Overall mortality was 40.2% (n = 106), with median follow-up of 10.2 years (longest 31.5 years). In multivariable analysis, pulmonary vein stenosis, coarctation, univentricular circulation, asplenia phenotype, and at least mild atrioventricular valve regurgitation at presentation were associated with mortality, whereas birth era was not. Among patients who underwent surgery, univentricular repair remained associated with mortality after adjustment. In classification and regression analysis, patients with biventricular circulation (especially those with polysplenia) had lower mortality than those with univentricular circulation.

CONCLUSIONS:

In this large retrospective study of HS, outcomes remain poor and have not improved since the early 1990s. We identified risks factors associated with earlier mortality and found that those with univentricular circulation and totally anomalous pulmonary venous connection had the worst prognosis. Survival was higher in those with biventricular circulation.

What’s Known on This Subject:

Long-term outcomes in heterotaxy syndrome are not well known. Previous reports suggest that midterm outcomes are poor, but long-term outcomes are not well characterized. It is further unknown whether outcomes or associated risk factors have changed in contemporary birth era.

What This Study Adds:

In this study, we show that heterotaxy syndrome carries poor long-term prognosis, survival has not improved in the current era, and risk factors have not changed. These data may inform risk stratification, counseling, and future efforts to improve management strategies.

Long-term outcomes in heterotaxy syndrome (HS) are not well known. Previous reports have included small series of patients or have been focused on specific anatomic or operative subgroups, such as those undergoing Fontan procedure.15  We have previously reported midterm outcomes among a large group of patients with HS born between 1985 and 1997, with a broad range of anatomic variations and surgical management.6  In this group, 5-year survival was poor at 53%, and independent risk factors for mortality included obstructed totally anomalous pulmonary venous connections, mild or greater atrioventricular valve regurgitation at presentation, and a common atrioventricular canal defect. The long-term survival in this group of patients, however, has not been characterized. Specifically, it is unknown whether the survival curve of patients with HS plateaus after the first several years of life or whether the risk of late mortality persists. Furthermore, although some reports suggest improved survival with current surgical and transcatheter management strategies,79  others have failed to show similar results.2,3,5  Importantly, the sample size in these reports is relatively small, and follow-up time is short. We, therefore, aimed to describe the long-term outcomes in patients with HS and to assess whether outcomes and risk factors for mortality have changed in the current era.

We conducted a retrospective database review to identify all patients with HS as defined below who met the following inclusion criteria: (1) birth year 1985 through 2014; (2) presentation to Boston Children’s Hospital (BCH) by age 6 months; (3) residence in the primary referral area for BCH, which includes New England and upstate New York; and (4) all surgery at BCH with the exception of early palliation with a pulmonary artery band or systemic-to-pulmonary arterial shunt. Patient demographics, clinical history, imaging and catheterization data, and surgical details were abstracted from the medical records. The Social Security Death Index was queried to ascertain patient deaths not in the medical record. The Committee on Clinical Investigation at BCH approved this study.

The diagnosis of HS was established on the basis of the definition proposed by Van Praagh et al10 : “an abnormal symmetry of certain viscera and veins (lungs, liver, vena cavae) and situs discordance between the various organs, as well as between the various segments of the heart.” Characteristic anatomic features as set forth by Van Praagh et al10  and modified by Lin et al11  were used for reference (Supplemental Table 5), and patients were classified as having a phenotype most consistent with “asplenia syndrome” or “polysplenia syndrome” based on commonly observed anatomic patterns as previously described.6 

Patients were divided into 4 eras based on birth year: 1985 to 1991, 1992 to 1998, 1999 to 2007, and 2008 to 2014. Descriptive statistics are shown as median (interquartile range [IQR]) or n (%). The primary outcome was defined as survival, and follow-up time was defined as time from birth to death or last contact date. Mortality incidence rates (number deaths divided by total follow-up time) were calculated for each birth era, the overall cohort, and the subgroup of patients who underwent surgical intervention. Kaplan–Meier analysis was used to estimate the overall hazard for death and survival probabilities by the 4 birth eras, and the distributions were compared by using the log-rank test and the Wilcoxon–Gehan test. On the basis of review of the Kaplan–Meier analyses, an additional cutoff of birth year 1991 was chosen, and mortality incidence rates were also calculated for those born after 1991 versus those born in or before 1991. Poisson regression was used to estimate mortality incidence rates and compare differences in mortality incidence by age cohort. A conditional analysis, based on one-month survivors, was performed in a similar manner. In addition, for each birth era, the rate of surgeries or catheterizations was defined as the number of total surgeries or catheterizations divided by total number of patient follow-up years. Poisson regression modeling was used to assess whether the rates of intervention differed by birth era.

Risk factors for mortality were identified by using Cox proportional hazards regression, with examination of interactions of birth era and each candidate risk factor to determine if the risk factor set differs by birth era. Independent factors significant at the 0.20 level in the univariate analysis were candidate factors in a stepwise procedure to construct a multivariable Cox regression model. A P value of .15 was required for entry, and a P value <.05 was required for inclusion in the final model. Analyses were performed for the entire cohort of patients and separately for the patients who underwent at least one surgical intervention to determine if the latter group had a different set of risk factors compared with the entire cohort. In this latter group, a multivariable Cox regression model was also constructed to assess the association between surgical pathway (univentricular versus biventricular repair) and mortality after adjusting for baseline confounders. Classification and regression (CART) analysis based on a recursive partitioning algorithm was also performed to create a risk stratification model for predictors associated with death, identifying variables and partition points that optimally separated low-risk versus high-risk of death. All analyses were conducted by using SAS version 9.4 (SAS Institute, Inc, Cary, NC) and R version 3.5.0.

A total of 264 patients (49% female) met inclusion criteria, 118 (44.7%) with asplenia and 146 (55.3%) with polysplenia phenotypes. Median follow-up time among survivors was 10.2 years (IQR: 4.2–18), with the longest follow-up of 31.5 years. Table 1 describes the clinical and anatomic characteristics of the entire cohort and stratified by birth era. Notable changes during the study period included an increase in the rate of prenatal diagnosis to nearly 90% at the latest birth era, earlier overall age at diagnosis, decreased prevalence of ventricular dysfunction at presentation, and increased frequency of biventricular repair.

TABLE 1

Clinical and Anatomic Characteristics

Overall N = 264Birth Year 1985–1991, n = 64Birth Year 1992–1998, n = 59Birth Year 1999–2007, n = 65Birth Year 2008–2014, n = 76P
Female sex, n (%) 130 (49) 26 (41) 38 (64) 36 (55) 30 (39) .01 
Median (IQR) follow-up, y 4.7 (0.7–15.1) 5.1 (0.2–24.0) 10.8 (0.2–19.8) 10.2 (0.9–14.2) 3.0 (0.9–5.2) .005 
Age at diagnosis (n = 258)      .01 
 ≤1 d, n (%) 214 (83) 49 (77) 41 (72) 57 (91) 67 (91) — 
 2–30 d, n (%) 34 (13) 13 (20) 13 (23) 3 (5) 5 (7) — 
 >30 d, n (%) 10 (4) 2 (3) 3 (5) 3 (5) 2 (3) — 
Fetal diagnosis (n = 258), n (%) 181 (70) 30 (47) 33 (58) 52 (83) 66 (89) <.001 
Asplenia phenotype, n (%) 118 (45) 37 (58) 25 (42) 21 (32) 35 (46) .03 
Heart block at presentation, n (%) 18 (7) 2 (3) 6 (10) 5 (8) 5 (7) .46 
Pulmonary venous anatomy, n (%)      .06 
 Normal connection and drainage 128 (49) 25 (39) 28 (48) 35 (54) 40 (53) — 
 Totally anomalous connection 80 (30) 25 (39) 15 (25) 14 (22) 26 (34) — 
 Partially anomalous connection 13 (5) 2 (3) 2 (3) 3 (5) 6 (8) — 
 Anomalous drainage only 43 (16) 12 (19) 14 (24) 13 (20) 4 (5) — 
Pulmonary vein stenosis at any time, n (%) 70 (27) 25 (39) 13 (22) 14 (22) 18 (24) .09 
Atrioventricular canal anatomy, n (%)      .03 
 Normal 98 (37) 19 (30) 19 (32) 30 (46) 30 (40) — 
 Complete defect 143 (54) 39 (61) 29 (49) 32 (49) 43 (57) — 
 Transitional or partial defect 23 (9) 6 (9) 11 (19) 3 (5) 3 (4) — 
At least mild atrioventricular valve regurgitation at presentation, n (%) 140 (53) 34 (53) 34 (58) 28 (43) 44 (58) .29 
Ventricular anatomy (n = 263), n (%)      .38 
 Normal 143 (54) 27 (42) 34 (59) 40 (62) 42 (55) — 
 Hypoplastic RV and/or single LV 24 (9) 9 (14) 4 (7) 5 (8) 6 (8) — 
 Hypoplastic LV and/or single RV 96 (37) 28 (44) 20 (35) 20 (31) 28 (37) — 
Ventricular function at presentation (n = 194), n (%)      .03 
 Normal 158 (81) 12 (67) 36 (78) 50 (89) 60 (81) — 
 Mild dysfunction 18 (9) 1 (6) 5 (11) 2 (4) 10 (14) — 
 Moderate dysfunction 9 (5) 3 (17) 1 (2) 2 (4) 3 (4) — 
 Severe dysfunction 9 (5) 2 (11) 4 (9) 2 (4) 1 (1) — 
Coarctation of the aorta, n (%) 27 (11) 9 (14) 3 (5) 6 (10) 9 (13) .35 
Median age at first surgery, d (IQR) 9 (4–83) 9.5 (3.5–8) 6 (2–99.5) 7.5 (5–86) 14 (6–70) .19 
Type of surgery, n (%)      .12 
 No surgery 39 (15) 12 (19) 7 (12) 11 (17) 9 (12) — 
 Univentricular palliation 138 (52) 38 (59) 35 (59) 32 (49) 33 (43) — 
 Biventricular repair 87 (33) 14 (22) 17 (29) 22 (34) 34 (45) — 
Final circulation, n (%)      .002 
 Univentricular 148 (56) 48 (75) 34 (58) 32 (49) 34 (45) — 
 Biventricular 116 (44) 16 (25) 25 (42) 33 (51) 42 (55) — 
Overall N = 264Birth Year 1985–1991, n = 64Birth Year 1992–1998, n = 59Birth Year 1999–2007, n = 65Birth Year 2008–2014, n = 76P
Female sex, n (%) 130 (49) 26 (41) 38 (64) 36 (55) 30 (39) .01 
Median (IQR) follow-up, y 4.7 (0.7–15.1) 5.1 (0.2–24.0) 10.8 (0.2–19.8) 10.2 (0.9–14.2) 3.0 (0.9–5.2) .005 
Age at diagnosis (n = 258)      .01 
 ≤1 d, n (%) 214 (83) 49 (77) 41 (72) 57 (91) 67 (91) — 
 2–30 d, n (%) 34 (13) 13 (20) 13 (23) 3 (5) 5 (7) — 
 >30 d, n (%) 10 (4) 2 (3) 3 (5) 3 (5) 2 (3) — 
Fetal diagnosis (n = 258), n (%) 181 (70) 30 (47) 33 (58) 52 (83) 66 (89) <.001 
Asplenia phenotype, n (%) 118 (45) 37 (58) 25 (42) 21 (32) 35 (46) .03 
Heart block at presentation, n (%) 18 (7) 2 (3) 6 (10) 5 (8) 5 (7) .46 
Pulmonary venous anatomy, n (%)      .06 
 Normal connection and drainage 128 (49) 25 (39) 28 (48) 35 (54) 40 (53) — 
 Totally anomalous connection 80 (30) 25 (39) 15 (25) 14 (22) 26 (34) — 
 Partially anomalous connection 13 (5) 2 (3) 2 (3) 3 (5) 6 (8) — 
 Anomalous drainage only 43 (16) 12 (19) 14 (24) 13 (20) 4 (5) — 
Pulmonary vein stenosis at any time, n (%) 70 (27) 25 (39) 13 (22) 14 (22) 18 (24) .09 
Atrioventricular canal anatomy, n (%)      .03 
 Normal 98 (37) 19 (30) 19 (32) 30 (46) 30 (40) — 
 Complete defect 143 (54) 39 (61) 29 (49) 32 (49) 43 (57) — 
 Transitional or partial defect 23 (9) 6 (9) 11 (19) 3 (5) 3 (4) — 
At least mild atrioventricular valve regurgitation at presentation, n (%) 140 (53) 34 (53) 34 (58) 28 (43) 44 (58) .29 
Ventricular anatomy (n = 263), n (%)      .38 
 Normal 143 (54) 27 (42) 34 (59) 40 (62) 42 (55) — 
 Hypoplastic RV and/or single LV 24 (9) 9 (14) 4 (7) 5 (8) 6 (8) — 
 Hypoplastic LV and/or single RV 96 (37) 28 (44) 20 (35) 20 (31) 28 (37) — 
Ventricular function at presentation (n = 194), n (%)      .03 
 Normal 158 (81) 12 (67) 36 (78) 50 (89) 60 (81) — 
 Mild dysfunction 18 (9) 1 (6) 5 (11) 2 (4) 10 (14) — 
 Moderate dysfunction 9 (5) 3 (17) 1 (2) 2 (4) 3 (4) — 
 Severe dysfunction 9 (5) 2 (11) 4 (9) 2 (4) 1 (1) — 
Coarctation of the aorta, n (%) 27 (11) 9 (14) 3 (5) 6 (10) 9 (13) .35 
Median age at first surgery, d (IQR) 9 (4–83) 9.5 (3.5–8) 6 (2–99.5) 7.5 (5–86) 14 (6–70) .19 
Type of surgery, n (%)      .12 
 No surgery 39 (15) 12 (19) 7 (12) 11 (17) 9 (12) — 
 Univentricular palliation 138 (52) 38 (59) 35 (59) 32 (49) 33 (43) — 
 Biventricular repair 87 (33) 14 (22) 17 (29) 22 (34) 34 (45) — 
Final circulation, n (%)      .002 
 Univentricular 148 (56) 48 (75) 34 (58) 32 (49) 34 (45) — 
 Biventricular 116 (44) 16 (25) 25 (42) 33 (51) 42 (55) — 

LV, left ventricle; RV, right ventricle; —, not applicable.

Overall mortality was 40.2% (106 of 264 patients) during a follow-up period of 2166 patient-years. For the entire cohort, 5-, 8-, 15-, and 20-year survival rates were 65%, 64%, 61%, and 54%, respectively (Fig 1). Median survival (half of the cohort deceased by) was 21.2 years. Among nonsurvivors, the median age at death was 0.47 years (IQR: 0.05–2.24), with 12% occurring in the first 30 days of life, and 24% died of noncardiac causes (Table 2). As shown in Fig 2, overall mortality was similar across birth eras (P = .116). When comparing the latter 3 birth eras (1992 onwards) to the 1985 to 1991 birth era, overall survival was better after 1991 (P = .025) (Fig 3A), but there was no difference in survival when patients who did not undergo surgery were excluded (P = .714) (Fig 3B). Rates of surgery were similar across the birth eras (P = .23). Heart transplants were rare, with only 5 patients in the entire population undergoing transplant.

FIGURE 1

Kaplan–Meier survival analysis of the entire study cohort (n = 264) with 95% confidence intervals.

FIGURE 1

Kaplan–Meier survival analysis of the entire study cohort (n = 264) with 95% confidence intervals.

TABLE 2

Mode of Death

Overall N = 106Birth Year 1985–1991, n = 38Birth Year 1992–1998, n = 25Birth Year 1999–2007, n = 19Birth Year 2008–2014, n = 24
Noncardiac, n (%) 25 (24) 6 (16) 7 (28) 4 (21) 8 (33) 
Cardiac, n (%) 20 (19) 9 (24) 4 (16) 4 (21) 3 (13) 
 Pulmonary vein stenosis 8 (3) 5 (13) 1 (4) 2 (11) 
 Nonpulmonary vein stenosis 12 (5) 4 (11) 3 (12) 2 (11) 3 (14) 
Peri-intervention complication, n (%) 19 (7) 3 (8) 6 (24) 4 (21) 6 (25) 
Unknown, n (%) 38 (36) 17 (45) 8 (32) 7 (37) 6 (25) 
No intervention offered, n (%) 4 (2) 3 (8) 1 (4) 
Overall N = 106Birth Year 1985–1991, n = 38Birth Year 1992–1998, n = 25Birth Year 1999–2007, n = 19Birth Year 2008–2014, n = 24
Noncardiac, n (%) 25 (24) 6 (16) 7 (28) 4 (21) 8 (33) 
Cardiac, n (%) 20 (19) 9 (24) 4 (16) 4 (21) 3 (13) 
 Pulmonary vein stenosis 8 (3) 5 (13) 1 (4) 2 (11) 
 Nonpulmonary vein stenosis 12 (5) 4 (11) 3 (12) 2 (11) 3 (14) 
Peri-intervention complication, n (%) 19 (7) 3 (8) 6 (24) 4 (21) 6 (25) 
Unknown, n (%) 38 (36) 17 (45) 8 (32) 7 (37) 6 (25) 
No intervention offered, n (%) 4 (2) 3 (8) 1 (4) 
FIGURE 2

Kaplan–Meier analysis of overall survival by birth era.

FIGURE 2

Kaplan–Meier analysis of overall survival by birth era.

FIGURE 3

Kaplan–Meier analysis of overall survival by birth year 1985 to 1991 vs 1992 to 2014 (A); overall survival by birth year 1985 to 1991 vs 1992 to 2014 among patients who underwent surgery (B); 1-month survival by birth era (C).

FIGURE 3

Kaplan–Meier analysis of overall survival by birth year 1985 to 1991 vs 1992 to 2014 (A); overall survival by birth year 1985 to 1991 vs 1992 to 2014 among patients who underwent surgery (B); 1-month survival by birth era (C).

In contrast to overall mortality, 30-day mortality significantly improved over birth eras (Fig 3C), but this difference did not persist after 30 days. The hazard of death was highest in infancy and early childhood but continued through life, with a slow increase in risk in late childhood and young adulthood (Fig 4). Mortality incidence was 0.56, 1.01, and 2.15 deaths per 100 patient-years for ages 5 to <10, 10 to <15, and ≥15 years, respectively; the rate in the oldest age group differed from that in the 5 to <10-year age group (P = .045). Although fetal diagnosis was more common in the more recent 2 birth eras, fetal diagnosis was not associated with lower mortality (P = .57).

FIGURE 4

Instantaneous hazard of death in the study cohort (n = 264).

FIGURE 4

Instantaneous hazard of death in the study cohort (n = 264).

In univariate analysis, several anatomic, surgical, and functional variables were associated with a higher mortality risk in the cohort (Supplemental Table 6). In multivariable analysis (Table 3), the following variables were independently associated with a higher risk of shorter time to death: development of pulmonary vein stenosis, coarctation of the aorta, univentricular circulation, asplenia phenotype, and at least mild atrioventricular valve regurgitation at presentation.

TABLE 3

Multivariable Cox Regression Model of Factors Associated With Death

Hazard Ratio95% CIPPseudo R2
All patients    0.29 
 Pulmonary vein stenosis during follow-up 2.71 1.76–4.17 <.001 — 
 Coarctation of aorta 2.40 1.33–4.36 .004 — 
 Asplenia phenotype 1.73 1.07–2.79 .025 — 
 Univentricular circulation 2.39 1.39–4.13 .002 — 
 At least mild atrioventricular valve regurgitation at presentation 1.66 1.09–2.51 .017 — 
Patients with ≥1 surgery    0.24 
 Pulmonary vein stenosis during follow-up 2.91 1.85–4.56 <.001 — 
 Univentricular circulation 2.14 1.19–3.82 .011 — 
 At least mild atrioventricular valve regurgitation at presentation 1.73 1.11–2.69 .016 — 
 Asplenia phenotype 1.67 1.02–2.74 .044 — 
Hazard Ratio95% CIPPseudo R2
All patients    0.29 
 Pulmonary vein stenosis during follow-up 2.71 1.76–4.17 <.001 — 
 Coarctation of aorta 2.40 1.33–4.36 .004 — 
 Asplenia phenotype 1.73 1.07–2.79 .025 — 
 Univentricular circulation 2.39 1.39–4.13 .002 — 
 At least mild atrioventricular valve regurgitation at presentation 1.66 1.09–2.51 .017 — 
Patients with ≥1 surgery    0.24 
 Pulmonary vein stenosis during follow-up 2.91 1.85–4.56 <.001 — 
 Univentricular circulation 2.14 1.19–3.82 .011 — 
 At least mild atrioventricular valve regurgitation at presentation 1.73 1.11–2.69 .016 — 
 Asplenia phenotype 1.67 1.02–2.74 .044 — 

CI, confidence interval.

When restricted to patients who underwent surgical intervention (Table 3), coarctation of the aorta no longer remained in the model, but the remaining independent predictors were unchanged, including univentricular circulation. Within this group, after adjusting for baseline confounders that differed between the univentricular and biventricular repair groups, univentricular repair remained an independent predictor of shorter time to death. Other independent predictors included heart block, pulmonary vein stenosis before surgery, at least mild atrioventricular valve regurgitation before surgery, asplenia phenotype, and totally anomalous pulmonary venous connection (Table 4).

TABLE 4

Multivariable Cox Regression Model of Factors Associated With Death in Patients Who Underwent Surgical Intervention, Adjusting for Univentricular Versus Biventricular Repair Groups (Pseudo R2 0.25)

Hazard Ratio95% CIP
Type of surgery    
 Biventricular repair Reference — — 
 Univentricular repair 2.52 1.35–4.70 .004 
Heart block 3.2 1.19–8.63 .021 
Totally anomalous pulmonary venous connection 1.73 1.03–2.88 .037 
Asplenia phenotype 1.78 1.01–3.13 .002 
At least mild atrioventricular valve regurgitation before surgery 2.08 1.31–3.31 .002 
Pulmonary vein stenosis before surgery 2.73 1.45–5.13 <.002 
Hazard Ratio95% CIP
Type of surgery    
 Biventricular repair Reference — — 
 Univentricular repair 2.52 1.35–4.70 .004 
Heart block 3.2 1.19–8.63 .021 
Totally anomalous pulmonary venous connection 1.73 1.03–2.88 .037 
Asplenia phenotype 1.78 1.01–3.13 .002 
At least mild atrioventricular valve regurgitation before surgery 2.08 1.31–3.31 .002 
Pulmonary vein stenosis before surgery 2.73 1.45–5.13 <.002 

CI, confidence interval; —, not applicable.

Results of the CART analysis for the patients are shown in Fig 5. The final model had a positive predictive value of 61% and a negative predictive value of 84%, with an accuracy of 72%. By using univentricular and biventricular circulation alone, the accuracy was 69%, suggesting a small improvement in accuracy using the entire tree. Based on this analysis, the highest risk of mortality was in patients who had a univentricular circulation and totally anomalous pulmonary venous connection (71% mortality), intermediate mortality risk was observed in those with univentricular circulation and normal pulmonary venous connections (46%) and those with asplenia phenotype and biventricular circulation (42%), and the lowest risk was found in patients with polysplenia phenotype and a biventricular circulation (11% mortality).

FIGURE 5

CART model for death. Red box represents subgroup with highest mortality risk. Blue box represents subgroup with lowest mortality risk. TAPVC, totally anomalous pulmonary venous connection.

FIGURE 5

CART model for death. Red box represents subgroup with highest mortality risk. Blue box represents subgroup with lowest mortality risk. TAPVC, totally anomalous pulmonary venous connection.

This is one the largest studies to date evaluating long-term outcomes in patients with HS treated at a single institution for up to 3 decades. We found that outcomes in this cohort remain poor, with an overall mortality of 40% and median age at death of 0.47 years. Furthermore, despite significant improvement in neonatal survival over time and substantial increase in rates of fetal diagnosis, overall outcomes have not improved appreciably in the more recent birth eras, except as compared with the earliest era (1985 to 1991). Our analysis identified several factors independently associated with survival and established the highest and lowest risk categories based on CART analysis, information that may help with risk stratification and patient counseling.

HS has long been known to be a challenging disease with suboptimal short- and midterm outcomes.46,12  Although some recent reports have suggested improved outcomes over time,79,13,14  other studies have not found improved survival in the current era.15,16  These discrepancies among reported outcomes in HS may, at least in part, be due to small sample sizes and short follow-up times in previous series. To our knowledge, there is only one published report that evaluated change in outcomes over time, but this was a meta-analysis that had pooled data from multiple studies from several institutions with varied inclusion criteria.17  As such, differences in outcome may have reflected differences in patient selection and institutional practice. With our data, which include a relatively large cohort of patients treated at a single institution, we provide important context for patient counseling in the current era in which fetal diagnosis nears 90% and we counter the prevailing wisdom that patients with HS may be faring better with current surgical and interventional techniques. Furthermore, although, as in other long-term follow-up studies in pediatrics,18,19  precise cause of death was not available in one third of our patients, the proportion of those who died of noncardiac causes was substantial, highlighting the importance of identifying and treating noncardiac morbidities in this population.

A recent report from our institution showed a 23% overall mortality among live-born infants with HS, significantly lower than the data presented in this study.20  However, that study had a smaller sample size because it was focused on patients with prenatally diagnosed HS. Furthermore, rates of prenatal termination were high, especially in the asplenia cohort, which likely resulted in a selection bias toward lower-risk patients in the live-born group. By including a larger cohort of patients with HS, both prenatally and postnatally diagnosed, the data presented in this study provide a more complete picture of risk among infants born with this challenging diagnosis and should inform prenatal counseling and postnatal risk stratification.

In the CART analysis shown in Fig 5, we stratified patients into low (polysplenia syndrome and biventricular circulation), high (univentricular circulation and totally anomalous pulmonary venous connection), and intermediate risks of mortality based on a few simple criteria. This flowchart may be particularly useful for fetal counseling and early postnatal care. Although fetal diagnosis did not confer a survival benefit in our cohort, it does allow families to prepare for postnatal care and other implications of the diagnosis, and, with these data, may allow for more nuanced decision-making around pregnancy and postnatal interventions, such as pursuing a biventricular reconstruction strategy.

Risk factors for poor outcome in our cohort were similar to those found in other studies in which asplenia phenotype, atrioventricular valve regurgitation, pulmonary venous stenosis, and univentricular circulation have all been associated with poor outcomes.1517,20  These parameters increase the risk of poor surgical outcomes, adverse postoperative hemodynamics, long-term cardiac complications,21  and noncardiac complications such as infection.22  Although univentricular circulation has been previously shown to be associated with poor outcomes in HS,16,17,23  ours is the first study to show that univentricular repair remains independently associated with poor outcomes even after adjustment for baseline confounders such as ventricular hypoplasia. This suggests that in patients with borderline anatomy, pursuing a biventricular reconstruction strategy may provide a survival benefit despite the complexities of the repair. However, this hypothesis awaits confirmation through further studies. Additional studies are also needed to help risk stratify patients who may benefit most from this approach.

Notable limitations of this study include those inherent to all retrospective studies. Although we cannot exclude the possibility of referral bias, we mitigated such bias by restricting our patient cohort to those who lived in our referral area and obtained their surgical care primarily at our institution. Despite all efforts, some patients were lost to follow-up, and their outcomes are unknown, which may have affected our results. However, the use of a national death index in addition to institutional electronic medical records should have minimized ascertainment bias. In terms of risk factors for poor outcome, although we attempted to correct for baseline confounders in the univentricular versus biventricular circulation groups, we cannot assess how many patients with borderline anatomy or a need for complex intracardiac reconstructions were sent to a biventricular repair and what their outcomes were like. Further studies are needed to answer this important question.

In this large retrospective study of patients with HS, outcomes were poor and did not improve in recent birth eras. We identified several risk factors associated with earlier mortality and stratified patients into low-, intermediate-, and high-risk groups based on baseline characteristics, anatomy, and surgical strategy. These data may assist with risk stratification and patient counseling, especially in the context of fetal diagnosis. Further studies are needed to evaluate whether patients with borderline anatomy may benefit from a biventricular surgical approach, despite the complexity of repair. These data also highlight the critical need for breakthrough treatments for pulmonary vein stenosis.

We thank Minmin Lu, MS, for her statistical programming related to this project.

Drs Banka and Geva conceptualized and designed the study and drafted the initial manuscript; Drs Adar and Schaetzle collected data; Dr Sleeper conducted the data analyses; Dr Emani provided expert insights; and all authors reviewed and revised the manuscript and approved the final manuscript as submitted and agreed to be accountable for all aspects of the work.

Dr Banka’s current affiliation is Merck & Co, Inc, Rahway, NJ.

Dr Adar’s current affiliation is Schneider Children’s Medical Center, Petah Tivka, Israel.

Dr Schaetzle’s current affiliation is Kardiologie Praxis Futuro, Liestal, Switzerland.

FUNDING: Funded in part by the Higgins Family Noninvasive Research Fund.

     
  • BCH

    Boston Children’s Hospital

  •  
  • CART

    classification and regression

  •  
  • HS

    heterotaxy syndrome

  •  
  • IQR

    interquartile range

1
Atz
AM
,
Cohen
MS
,
Sleeper
LA
, et al;
Investigators of the Pediatric Heart Network
.
Functional state of patients with heterotaxy syndrome following the Fontan operation
.
Cardiol Young
.
2007
;
17
(
suppl 2
):
44
53
2
Bartz
PJ
,
Driscoll
DJ
,
Dearani
JA
, et al
.
Early and late results of the modified fontan operation for heterotaxy syndrome 30 years of experience in 142 patients
.
J Am Coll Cardiol
.
2006
;
48
(
11
):
2301
2305
3
Cohen
MS
,
Schultz
AH
,
Tian
ZY
, et al
.
Heterotaxy syndrome with functional single ventricle: does prenatal diagnosis improve survival?
Ann Thorac Surg
.
2006
;
82
(
5
):
1629
1636
4
Eronen
MP
,
Aittomäki
KA
,
Kajantie
EO
,
Sairanen
HI
.
Outcome of left atrial isomerism at a single institution
.
Pediatr Cardiol
.
2012
;
33
(
4
):
596
600
5
Eronen
MP
,
Aittomäki
KA
,
Kajantie
EO
,
Sairanen
HI
,
Pesonen
EJ
.
The outcome of patients with right atrial isomerism is poor
.
Pediatr Cardiol
.
2013
;
34
(
2
):
302
307
6
Foerster
SR
,
Gauvreau
K
,
McElhinney
DB
,
Geva
T
.
Importance of totally anomalous pulmonary venous connection and postoperative pulmonary vein stenosis in outcomes of heterotaxy syndrome
.
Pediatr Cardiol
.
2008
;
29
(
3
):
536
544
7
Morales
DL
,
Braud
BE
,
Booth
JH
, et al
.
Heterotaxy patients with total anomalous pulmonary venous return: improving surgical results
.
Ann Thorac Surg
.
2006
;
82
(
5
):
1621
1627–1628
8
Ota
N
,
Fujimoto
Y
,
Murata
M
, et al
.
Improving outcomes of the surgical management of right atrial isomerism
.
Ann Thorac Surg
.
2012
;
93
(
3
):
832
838–839
9
Anagnostopoulos
PV
,
Pearl
JM
,
Octave
C
, et al
.
Improved current era outcomes in patients with heterotaxy syndromes
.
Eur J Cardiothorac Surg
.
2009
;
35
(
5
):
871
877–878
10
Van Praagh
S
,
Santini
F
,
Sanders
SP
. Cardiac Malpositions with Special Emphasis on Visceral Heterotaxy (Asplenia and Polysplenia Syndromes). In:
Keane
JF
,
Lock
JE
,
Fyler
DC
, eds.
Nadas’ Pediatric Cardiology
.
Philadelphia, PA
:
Hanley & Belfus
;
1992
:
589
608
11
Lin
AE
,
Krikov
S
,
Riehle-Colarusso
T
, et al;
National Birth Defects Prevention Study
.
Laterality defects in the national birth defects prevention study (1998-2007): birth prevalence and descriptive epidemiology
.
Am J Med Genet A
.
2014
;
164A
(
10
):
2581
2591
12
Ota
N
,
Fujimoto
Y
,
Hirose
K
, et al
.
Improving results of atrioventricular valve repair in challenging patients with heterotaxy syndrome
.
Cardiol Young
.
2010
;
20
(
1
):
60
65
13
Baban
A
,
Cantarutti
N
,
Adorisio
R
, et al
.
Long-term survival and phenotypic spectrum in heterotaxy syndrome: a 25-year follow-up experience
.
Int J Cardiol
.
2018
;
268
:
100
105
14
Chen
W
,
Ma
L
,
Cui
H
, et al
.
Early- and middle-term surgical outcomes in patients with heterotaxy syndrome
.
Cardiology
.
2016
;
133
(
3
):
141
146
15
Khan
MS
,
Bryant
R
 III
,
Kim
SH
, et al
.
Contemporary outcomes of surgical repair of total anomalous pulmonary venous connection in patients with heterotaxy syndrome
.
Ann Thorac Surg
.
2015
;
99
(
6
):
2134
2139–2140
16
Alsoufi
B
,
McCracken
C
,
Schlosser
B
, et al
.
Outcomes of multistage palliation of infants with functional single ventricle and heterotaxy syndrome
.
J Thorac Cardiovasc Surg
.
2016
;
151
(
5
):
1369
1377.e2
17
Loomba
RS
,
Nijhawan
K
,
Anderson
R
.
Impact of era, type of isomerism, and ventricular morphology on survival in heterotaxy: implications for therapeutic management
.
World J Pediatr Congenit Heart Surg
.
2016
;
7
(
1
):
54
62
18
Nollert
G
,
Fischlein
T
,
Bouterwek
S
,
Böhmer
C
,
Klinner
W
,
Reichart
B
.
Long-term survival in patients with repair of tetralogy of Fallot: 36-year follow-up of 490 survivors of the first year after surgical repair
.
J Am Coll Cardiol
.
1997
;
30
(
5
):
1374
1383
19
Hjortshøj
CMS
,
Kempny
A
,
Jensen
AS
, et al
.
Past and current cause-specific mortality in Eisenmenger syndrome
.
Eur Heart J
.
2017
;
38
(
26
):
2060
2067
20
Escobar-Diaz
MC
,
Friedman
K
,
Salem
Y
, et al
.
Perinatal and infant outcomes of prenatal diagnosis of heterotaxy syndrome (asplenia and polysplenia)
.
Am J Cardiol
.
2014
;
114
(
4
):
612
617
21
Niu
MC
,
Dickerson
HA
,
Moore
JA
, et al
.
Heterotaxy syndrome and associated arrhythmias in pediatric patients
.
Heart Rhythm
.
2018
;
15
(
4
):
548
554
22
Chiu
SN
,
Shao
PL
,
Wang
JK
, et al
.
Severe bacterial infection in patients with heterotaxy syndrome
.
J Pediatr
.
2014
;
164
(
1
):
99
104.e1
23
Makhija
Z
,
Marwah
A
,
Mishra
S
,
Kumar
J
,
Goel
A
,
Sharma
R
.
Biventricular repair in heterotaxy patients
.
World J Pediatr Congenit Heart Surg
.
2015
;
6
(
2
):
195
202

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