Materials and Methods
A national population-based registry of EA, which is coordinated by the National Reference Center for EA located in Lille University Hospital, collects all new cases of neonates with EA who are born in France after Jan. 1, 2008. The registry was approved by the Advisory Committee on information and research in health (Comité consultatif sur le traitement de l’information en matière de recherche dans le domaine de la santé no. 08.297) and by the National Commission on Informatics and Liberties (CNIL no. 908362). A specific questionnaire was created based on previous information available on EA epidemiology and neonatal characteristics and was addressed to all 38 neonatal surgical centers in France. The questionnaire was then completed by the participating centers on a free-willing basis; a clinical research assistant, when necessary, collected the information. A physician and a research assistant checked each questionnaire and double-checked the data that were entered into the database, paying extra attention to avoid any case duplication. In the event of inconsistencies or a lack of information, the corresponding center was contacted to resolve the issue. The exhaustiveness of inclusion was achieved by the voluntary participation of all the neonatal surgery teams and neonatal intensive care units within the national health system organization (tertiary health care centers). The sources of information were checked by 3 methods to enhance exhaustiveness. First, we checked with all centers by phone calls or email messages that the team was enrolling all new patients; second, we performed an audit in 12 centers where our research assistant checked the source files; third, we crosschecked all questionnaires that were received against data that were obtained from the hospital information system database of each participating center. The inclusion criteria in our study were all new liveborn infants with EA in France and its overseas territories who were born between Jan. 1, 2008 and Dec. 31, 2010.
In France, pregnant women generally have an ultrasound scan at 12, 22, and 32 weeks of gestation. When an abnormality is detected, the woman is referred to a regional specialist center where an MRI can be proposed. In the event of a postnatal diagnosis of EA, the neonate is transferred into 1 of the 38 French pediatric surgery departments. Two groups were determined and compared for this particular study. The prenatal group consisted of all cases with a documented antenatal suspicion of EA (sonography ± MRI). The postnatal group consisted of all infants with an EA that was diagnosed after birth.
Data on prenatal information included sonographic signs (hydramnios and small or absent bubble, pouch sign), MRI, amniocentesis, and in utero transfer. EA was classified based on Ladd’s classification, and associated abnormalities were noted. The defect length was measured during the first intervention by the surgeon. The neonatal treatment, which included the delay between birth and diagnosis, birth and first intervention, transfer after birth, type of surgery, and neonatal care, was also recorded. We defined a composite variable of the 1-year outcome that included neonatal morbidity (anastomotic esophageal leaks, recurrent fistula, stenosis) and death. Finally, we compared the outcome of EA types I and III and evaluated the impact of prenatal diagnosis on survival at 1 year in relation to the EA type. Qualitative variables were expressed as frequencies and percentages; continuous variables were expressed as mean, median, and interquartile range. The normality of the distribution was assessed through the Shapiro-Wilk test. Parametric tests were used for normal distribution. In other cases, nonparametric tests were chosen. Chi-square and Fisher exact tests were used to compare the qualitative parameters of the 2 groups; the Mann-Whitney test or the Student t test was used for the continuous variables. The survival was calculated by the Kaplan-Meier method, and the differences between groups were compared with the use of the log-rank test. The survival time was defined as the period between the date of the birth and the date of death during the 1-year follow-up period. A probability value < .05 was considered statistically significant. All analyses were performed with SAS software (version 9.3; SAS Institute Inc, Cary, NC).
Results
Four hundred sixty-nine live births with EA were recorded between Jan. 1, 2008 and Dec. 31, 2010 ( Figure 1 ). One hundred fourteen cases had a prenatal diagnosis (24.3%) that included 37 type I cases (32.7%), 3 type II cases (2.7%), and 73 type III cases (64.6%). Three hundred fifty-five cases (75.7%) had a postnatal diagnosis with a majority of type III cases (93.5%). Figure 2 shows the repartition of pre- and postnatal diagnosis according to the EA type; 82.2 % of EA type I cases were diagnosed prenatally, compared with 17.9% of EA type III ( P < .001). At the 1-year examination, we observed that the 10.6% cases that were lost to follow up had the same rate of early complications and the same length in neonatal unit care as our study population.
In the prenatal diagnosis group, a sonographic pouch sign was found in 33% of cases ( Table 1 ). Hydramnios (90% vs 34%) and small or absent stomachs (71% vs 3%) were observed more frequently in the prenatal diagnosis group ( P < .001). An MRI was performed in 33.3% of cases, and amniocentesis was performed for fetal karyotype in 58.8% of prenatal diagnosis cases. In utero transfer and antenatal specialist counselling (for EA and/or others malformations) were more frequent in the prenatal group compared with the postnatal group (respectively, 59.3% vs 5% and 60.4% vs 11.5%; P < .001).
| Variable | Prenatal diagnosis (n = 114), % | Postnatal diagnosis (n = 355), % | P value |
|---|---|---|---|
| Hydramnios | 90.1 | 34.0 | < .001 |
| Small or absent stomach | 71.0 | 3.0 | < .001 |
| Sonographic pouch sign | 31.0 | 0.0 | < .001 |
| Magnetic resonance imaging | 33.3 | 1.1 | < .001 |
| Amniocentesis | 58.8 | 15.8 | < .001 |
| In utero transfer | 59.3 | 5.0 | < .001 |
| Antenatal specialist counseling | 60.4 | 11.5 | < .001 |
The proportion of infants with an isolated EA was similar in the 2 groups ( Table 2 ). One-half of the infants (50%) in each group had associated anomalies, and 17% of them had vertebral anomalies, anal atresia, cardiac defects, tracheoesophageal fistula and/or EA, renal anomalies, and limb defects association. We observed abnormal karyotype in 4.7% in the prenatal subset and in 8.0% in the postnatal subset ( P = .55). In this population, 1 infant had Down syndrome with severe cardiac malformations and died at day 2 without surgery. Two other infants that had aneuploidies died: 1 of them had Trisomy 18 and died at day 18 after surgery; the other infant had a Fanconi syndrome and died at day 12. A surgical correction was attempted in all others cases of aneuploidies.
| Variable | Diagnosis, % | P value | |
|---|---|---|---|
| Prenatal (n = 114) | Postnatal (n = 355) | ||
| Isolated esophageal atresia | 49.0 | 50.0 | .89 |
| Associated abnormalities | 51.0 | 50.0 | .89 |
| Neurologic | 12.0 | 7.3 | .28 |
| Renal | 15.8 | 24.7 | .16 |
| Cardiac | 53.5 | 54.5 | 1 |
| Limb | 15.5 | 15.3 | 1 |
| Anorectal | 15.5 | 16.4 | 1 |
| Costovertebral | 34.5 | 34.8 | 1 |
| Vertebral anomalies, anal atresia, cardiac defects, tracheoesophageal fistula and/or esophageal atresia, renal anomalies, and limb defects | 16.7 | 16.6 | 1 |
| Coloboma, heart defect, atresia choanae, retarded growth and development, genital hypoplasia, and ear anomalies | 1.7 | 1.4 | — |
| Abnormal karyotype | 4.7 | 8.0 | .55 |
The mean gestational age at delivery was significantly lower in the prenatal group (36 weeks) than that in the postnatal subset (37 weeks; P < .001; Table 3 ). Infants from the prenatal subset had a significantly lower mean birthweight ( P < .001) with a higher rate of small-for-gestational-age infants ( P = .02). The delay between the birth and the diagnosis was significantly shorter in the prenatal subset (0 vs 0.15 days; P < .001), and the rate of transfer after birth was lower (25.6% vs 82.5%; P < .001), with no difference in the delay between the birth and the first intervention. In the postnatal subset, the defect size of EA was smaller, with a lower rate of gastrostomy at birth and a higher rate of standard primary anastomosis ( P < .001). The length of the neonatal hospitalization was higher in the prenatal subset (71 vs 36 days; P < .001) with no difference in ventilation between the 2 groups.
| Variable | Diagnosis, % | P value | |
|---|---|---|---|
| Prenatal (n = 114) | Postnatal (n = 355) | ||
| Gestational age at birth, wk a | 36; 36 (35-38) | 37; 38 (35-40) | < .001 |
| Birthweight, g a | 2313; 2352 (1950-2700) | 2754; 2680 (2040-3100) | < .001 |
| Small for gestational age, % | 49.9 | 36.6 | .02 |
| Apgar score at 5 min ≤6, % | 11.0 | 12.8 | .61 |
| Delay birth: diagnosis, d a | 0 | 0.2; 0 (0-0) | < .001 |
| Delay >24 h, % | 0 | 10 | < .001 |
| Transfer after birth, % | 25.6 | 82.5 | < .001 |
| Delay birth: first intervention, d a | 0.9; 1 (0-1) | 0.9; 1 (0-1) | .07 |
| Defect size, cm a | 2.6; 2.5 (1.5-3.2) | 1.5; 1 (1-2) | < .001 |
| Gastrostomy at birth, % | 45.6 | 11.8 | < .001 |
| Standard primary anastomosis, % | 74.5 | 94.4 | < .001 |
| Age at anastomosis, d a | 26.5; 1 (0-51) | 4.5; 1 (0-1) | .002 |
| Invasive artificial ventilation, d a | 4.6; 3 (1-5) | 5.8; 3 (2-5) | .15 |
| Noninvasive ventilation, d a | 2.3; 0 (0-1) | 3.1; 0 (0-1) | .7 |
| Length of first hospitalization, d a | 71; 47.5 (20-117) | 36; 21 (14-43) | < .001 |
| Exclusive oral alimentation (at first discharge), % | 67.6 | 86.8 | < .001 |
| Complications, % | |||
| Anastomotic leaks | 16.8 | 5.5 | < .001 |
| Recurrent fistula | 4.2 | 4.5 | 1 |
| Stenosis | 31.6 | 18.0 | .04 |
| Readmission, n a | 2.6; 2 (1-4) | 2.4; 2 (1-3) | .27 |
| Total length of hospitalization, d a | 28.7; 16 (6-30) | 20.9; 10 (4-19.5) | .026 |
| Weight at 12 mos, g a | 8176; 8200 (7400-8940) | 8700; 8590 (7900-9440) | .001 |
| Death rate, % | 7.0 | 6.3 | .83 |
| Composite variable, % b | 44.0 | 27.6 | .003 |
a Results presented as mean; median (quartile 1-3)
b Includes morbidity (anastomotic leaks, recurrent fistula or stenosis) and morbidity at 1 year.
Postoperative complications among survivors were frequent in both subsets but higher in prenatal subsets in regards to anastomotic leaks (16.8% vs 5.45%; P < .001), stenosis (31.6% vs 18%; P = .04), and esophageal dilation (29.5 vs 17.3%; P = .01). The number of readmissions was similar between the 2 groups, but the duration of the stay was higher in prenatal subset (28.7 vs 20.9 days; P = .026). Twenty-seven deaths occurred during the first year of life. The prenatal diagnosis did not influence the mortality rate (7% vs 6.3%; P = NS). The composite variable was higher in the prenatal diagnosis subset (44% vs 27.6%; P = .003). There was still no difference in survival rates after the exclusion of aneuploidies.
In regards to the EA type, the defect size, the rate of gastrostomy at birth, and the length of first hospitalization was higher in EA type I subset ( P < .001; Table 4 ). The composite variable was lower in EA type III (28.3%) than in EA type I (58.1%; P < .001). Figure 3 shows the Kaplan-Meier survival curve for EA types I and III according to pre- or postnatal diagnosis. There was no significant difference between prenatal and postnatal curves for EA type I ( P = .34) and III ( P = .34).
| Variable | Esophageal atresia type | P value | |
|---|---|---|---|
| I (n = 45) | III (n = 405) | ||
| Prenatal diagnosis, % | 80 | 18 | < .001 |
| Gestational age at birth, wks a | 35.5; 36 (34-38) | 37.3; 36.7 (35-39) | < .001 |
| Birthweight, g a | 2295; 2200 (1770-2610) | 2542; 2640 (2050-3015) | .002 |
| Defect size, cm a | 3.6; 3.2 (3-4) | 1.5; 1 (1-2) | < .001 |
| Gastrostomy at birth, % | 91.0 | 11.0 | < .001 |
| Standard primary anastomosis, % | 78.0 | 97.8 | < .001 |
| Invasive artificial ventilation, d a | 3.8. 3 (2-4.5) | 5.7; 3 (1-5) | .89 |
| Length of first hospitalization, d a | 128.6; 121 (92-155) | 36.2; 21 (14-43.5) | < .001 |
| Complications, % | |||
| Anastomotic leaks | 34.2 | 5.4 | < .001 |
| Recurrent fistula | 2.6 | 4.5 | 1 |
| Stenosis | 36.8 | 19.4 | .02 |
| Readmission n a | 3.4; 3 (2-4) | 2.3; 2 (1-3) | .002 |
| Total length of hospitalization, d a | 31; 20 (9-32) | 22.9; 10 (4-21) | .02 |
| Weight at 12 months, g a | 8092; 8095 (7300-8930) | 8628; 8590 (7850-9300) | .02 |
| Death rate, % | 13.6 | 5.0 | .03 |
| Composite variable, % b | 58.1 | 28.3 | < .001 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
