Clinical outcome, quality of life, and mental health in long-gap esophageal atresia: comparison of gastric sleeve pull-up and delayed primary anastomosis

The study included all children who underwent surgery for EA repair at the University Medical Center Hamburg-Eppendorf between April 2007 and April 2021. The follow-up was based on the protocol of the Esophageal Atresia and Tracheo-Esophageal Fistula Support Federation (KEKS). This includes follow-up visits at the age of 6 months, 1 year, 2 years, 4 years, 6 years, 10 years, 14 years and at age 18. Weight, food intake and reflux symptoms are checked at every visit. Clinical features are reevaluated until adulthood. Patients with missing data, such as long-term follow ups or those who refused to participate in the study, were excluded. The study received ethical approval from the Medical Chamber Hamburg (PV7161) and was registered at ClinicalTrials.gov (NCT04382820). Analysis and reporting were guided by Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) recommendations [23]. The patients were grouped into two main cohorts: (1) primary and (2) secondary anastomosis, while the secondary anastomosis group comprised the subgroups (a) gastric sleeve pull-up and (b) delayed primary anastomosis, which are the two main operating techniques performed for LGEA at our medical center.

Patient data were collected using medical records and included clinical variables such as details regarding perinatal data, birth weight/length, and co-morbidities (i.e., VACTERL). The type and length of esophageal gap were identified by review of initial postnatal chest X-rays: The type of EA was defined according to VOGT criteria. Additionally, LGEA was defined as a gap between the proximal and distal esophageal ends measuring ≥ 3 vertebrae.

Moreover, age and weight at the time of EA corrective surgery, as well as the duration and type of surgery were analyzed. Post-operative data obtained for the study included duration of mechanical ventilation, chest tube, and length of stay in the intensive care unit (ICU). Common complications, both long and short term (dysphagia, reflux, tracheomalacia, strictures, leakage, PPI requirements), were noted. Further, details regarding the number and duration of hospitalizations pertaining to dilatation procedures were recorded.

With respect to weight gain data, feeding regimes before and after anastomosis were analyzed and the functional oral intake scale (FOIS) was used to evaluate oral intake. FOIS consists of a numeric scale quantifying oral intake, ranging from 1 (nothing by mouth) to 7 (full oral diet, no restrictions) [24]. Weight and height measurements were collected and converted into the weight-for-length z-score using The Netherlands Organization for Applied Scientific Research (TNO) growth standards [25]. To assess the presence of reflux symptoms, guardians were sent the ‘GERD-Q’; a questionnaire containing 7 items related to the frequency and severity of symptoms related to reflux disease [26, 27].

A study-specific questionnaire was sent to the families of affected children to assess the psychosocial outcomes like their (Hr)QoL and mental health status. Sociodemographic variables included sex, age, level of care required for the affected child, marital status, number of children, educational qualifications, employment status, and current physical/mental health status.

For descriptive data, frequencies, means, standard deviations, and bivariate tests (Chi-square tests) were used. Differences between groups were calculated using t-tests or Wilcoxon Rank test. Pearson correlations was applied to investigate the bivariate associations between psychosocial outcomes. To account for any known biases, propensity score matching was performed using an optimal matching algorithm with a caliper of 0.2 for gender, age, weight at operation, diagnosis, and revisionary surgery. Multiple linear regression models were used to define predictors of psychosocial outcomes. To indicate the size of the effect, Cohen’s d and Cramer’s V were calculated. Statistical significance was set at p \(\le\) 0.05 (two-tailed). Statistical analyses were conducted using SPSS Statistics 28 and GraphPad Prism 9.

Clinical outcomes of patients who underwent a primary anastomosis surgery were compared to those who underwent a secondary anastomosis surgery for EA. As suspected and shown in Table 1, children with primary anastomosis were significantly older, heavier, and had significantly shorter gap length than patients who underwent secondary repair. As shown in Table 2, primary anastomosis patients experienced significantly shorter operating and anesthesia times, during both the anastomosis repair surgery as well as all subsequent surgeries associated with the EA. However, the length of post-operative stay in the ICU and mechanical ventilation did not differ significantly between the primary and secondary anastomosis groups. Further, children with primary anastomoses had significantly fewer complications like tracheomalacia and reflux symptoms and were quicker to reach full oral feedings. Compared to patients who underwent secondary anastomosis, a primary anastomosis seems to lead to a fewer number of required endoscopic dilation procedures at the esophageal stricture site. Consequently, this results in a shorter cumulative duration of general anesthesia, operating time, and stay in the ICU.

As presented in Table 3, there were no significant differences regarding gap length between the secondary anastomosis subgroups. Additionally, as patients were propensity score matched for gender, age, and weight at the time of surgery, there were no significant differences with regards to these factors (Table 4). As summarized in Table 4, there were no differences regarding clinical outcomes and long-term complications between the two secondary anastomosis subgroups. However, the duration of general anesthesia during the surgery was significantly higher in the gastric sleeve pull-up group than the delayed anastomosis group. When analyzing all atresia-related surgeries and hospitalizations, no significant differences in the cumulative duration of general anesthesia or operating time between these two groups were found. Additionally, there were no significant differences regarding the of number of days spent in ICU or the duration of machine ventilation time post-operatively. Even more, the rates of post-operative complications, such as anastomosis leakage, strictures, re-fistula, tracheomalacia, scoliosis, recurrent infections or reflux (i.e., functional oral intake scale) were similar across both groups. The cumulative number of days spent in the ICU immediately following anastomosis surgery and all stays related to the atresia diagnosis thereafter was referred to as ‘cumulative length of stay in the ICU’, which was significantly higher in the delayed primary anastomosis group. The overall mortality rate was significantly lower in the gastric sleeve pull-up group. However, bearing in mind that patients in the gastric sleeve pull-up group were less frequent pre-term and had a lower rate of cardiac and anorectal malformations could potentially be accounting for this significant difference in mortality. As shown in Fig. 1, there were no long-term significant differences in weight gain between the gastric sleeve pull-up and delayed primary anastomosis group.

Out of 104 families assessed for eligibility and 63 families who received the questionnaires, in total 39 families responded and were included in analysis of the questionnaires. Thus, responder rate was almost comparable to previous studies [16, 17]. There were no significant differences between respondents and non-respondents regarding weight, age, gender, gap length, associated co-morbidities, operation method and complications. Supplementary Fig. 2 shows the CONSORT flow diagram illustrating the total patient cohort for analysis of HRQoL and mental health questionnaires. Table 5 shows the sociodemographic and disease characteristics of the participating parents and their affected child. When comparing primary with secondary anastomosis, as well as delayed primary anastomosis with gastric sleeve pull-up, there were no significant differences. Neither the QoL and mental health of the patient’s parents, nor the parent-reported HRQoL and mental health of the affected child, showed any significant differences in the comparison of these groups. Additionally, the entire cohort was analyzed in comparison to norm data available for each of the standardized questionnaires. Here, mothers reported a significant reduction in their child’s HRQoL and mental health, while fathers only reported a significant reduction in their child’s mental health. Data is summarized in Table 6.

Most limitations of the current study are inherent with the retrospective cross-sectional study design, in particular the small numbers and the lack of randomization. Thus, propensity score matching was used to limit the effects of the most relevant factors [51]. Moreover, due to the cross-sectional nature of the design, the statements regarding the treatment evaluation can only be made cautiously. The rate of non-respondents for the HRQoL and mental health assessment is unfortunately frequent and therefore may affect the results [52]. Even though a condition-specific questionnaire is recommended in addition to the use of a general questionnaire for HRQoL in previous research as a condition-specific questionnaire is generally more sensitive to evaluate clinical differences [53‐56], we did not use it due to the total amount of questionnaires used in the research project, which should be considered as a limitation of our study. Nonetheless, validated standardized instruments were used, which is a strength of the study. Moreover, GERD-Q is often used but has not been validated for children. Thus, reflux may be either over- or underestimated in the current study. Finally, in this study, parent-proxy instead of child-report has been used which may over- or under-estimate certain aspects.