Background and Aims: Ivor-Lewis esophagectomy remains one of the most complex oncologic procedures with
overall postoperative complications of around 60%. Over one third of all patients
suffer from complications that require at least endoscopic or other interventional
therapy even in high volume centers.
Study Design: We retrospectively analysed 1000 consecutive patients that underwent Ivor-Lewis esophagectomy
for adenocarcinoma of the esophagus or the esophagogastric junction or squamous cell
carcinoma of the esophagus in our center between January 2016 and June 2023 from a
prospectively maintained database. Preoperative parameters from a prospectively collected
risk questionnaire were evaluated for their significance to predict complications.
Statistical analysis was done with SPSS, data processing and figure plotting was done
with Excel and Graph Pad Prism. Testing for statistical significance between binary
variables was done with Pearson’s Chi-square.
Results: In an exploratory data analysis we identified several factors that significantly
predict major complications defined as>Clavien-Dindo IIIA (COPD, FEV1<80%, Diabetes
mellitus,), anastomotic leakage (elevated pre-operative serum creatinine, history
of percutaneous coronary intervention (PCI)) and pulmonary complications (VC, FEV1<80%,
COPD, congestive heart disease, BMI>30, active or former smoking). We found factors
that are associated with a prolonged hospital stay (ASA Score, histology, elevated
serum creatinine). We constructed a risk score including 7 of these parameters. Based
on the 7-factor risk score we defined a low and a high-risk group. Distribution of
complications over the low- and high-risk group was highly significantly different
with respect to major complications, anastomotic leakage, pulmonary complications,
prolonged hospital stay and in-hospital mortality, respectively. This withstands multivariable
regression analysis where a high-risk score is an independent risk factor for all
five outcome parameters.
Discussion: With our analysis we can identify complication-specific risk factors that might express
patient individual vulnerabilities. Our proposed 7-factor risk score and specifically
its simplification in a low- and high-risk group might be a genuine help in clinical
practice to identify patients at risk ([Fig. 1]
[2]).
Fig. 1 diagram depicting patient selection process resulting a highly standardized cohort.
Fig. 2
Conclusion: Selective assessment of risc factors with the here discovered measures might help
to tailor risk adjusted peri- and postoperative treatment and surveillance routines.
To convert this knowledge into relevant outcome improvement should be further evaluated
in prospective clinical trials.
