Infectious Complications After Esophagectomy

Each infectious episode occurring within 28 days after esophagectomy will be retrospectively identified by review of hospitalization records. The anatomical site of infection (e.g., pulmonary, pleural, mediastinal, urinary, abdominal, surgical wound, bloodstream) will be determined based on the description found in medical documentation.

The diagnosis of infection will rely on retrospective analysis of recorded clinical signs (e.g., fever), laboratory data (e.g., leukocytosis, elevated CRP), imaging reports (e.g., CT scan, chest X-ray), and physician assessmentsdocumented in the medical file.

The unit of measure is : number of infectious episodes per participant. The data will be reported as the total number of episodes and their distribution by anatomical site, summarized per participant and overall using descriptive statistics (median [IQR], counts, percentages).

For each postoperative infectious episode occurring within 28 days after esophagectomy, the presence or absence of specific pathogens will be retrospectively extracted from hospitalization reports using a structured checklist.

The pathogens are not recorded as free-text entries but as binary variables (Yes/No) based on the presence of documentation in medical reports (microbiology results, discharge summaries, etc.).

The predefined list includes :

• Gram-positive cocci : Streptococcus spp., Enterococcus spp., E. faecium, Staphylococcus aureus (including MRSA)
• Gram-negative bacteria : Haemophilus spp., Enterobacteriaceae (groups 1-3), non-fermenting bacilli (P. aeruginosa, A. baumannii, S. maltophilia)
• Resistance profiles : ESBL-producing Enterobacteriaceae, carbapenemase-producing organisms
• Fungi : Yeasts
• Polymicrobial vs. monomicrobial infections Only pathogens explicitly mentioned in the clinical documentation (culture or PCR results) are recorded. The data are collected usi

This outcome aims to explore the association between clinical and surgical factors and the risk of developing at least one infectious episode within 28 days following esophagectomy.

Postoperative infections will be identified retrospectively based on hospitalization records. A multivariable logistic regression analysis will be conducted, with infection occurrence (Yes/No) as the dependent variable.

Covariates will include :

• Demographic characteristics : age and sex
• Lifestyle factors : tobacco use and alcohol consumption (>2 drinks/day)
• Nutritional status and general severity : Body Mass Index (BMI), IGS2 score at ICU admission
• Comorbidities : cardiovascular disease (hypertension, myocardial infarction, congestive heart failure), chronic pulmonary disease, cerebrovascular disease (including stroke or transient ischemic attack), peripheral vascular disease (including aortic aneurysm >6 cm), diabetes (complicated or not), renal failure, liver disease (moderate or severe), dementi

Association between postoperative infection and the occurrence of major complications (sepsis, septic shock, AKI, ARDS, or death), as assessed by logistic regression analyses of hospitalization records.

Description

This outcome assesses the relationship between the occurrence of postoperative infection (within 28 days of esophagectomy) and the development of the following complications :

• Sepsis (according to Sepsis-3)
• Septic shock (according to Sepsis-3)
• Acute kidney injury (AKI according to KDIGO)
• Acute respiratory distress syndrome (ARDS according to the Berlin's definition)
• Mortality All events will be collected retrospectively from hospitalization records. A logistic regression will be used for each outcome separately to estimate the effect of infection (Yes/No) on the likelihood of each complication.

Unit of measure : Odds Ratio (OR) with 95% Confidence Interval for each complication

Length of hospital stay (in days, from surgery to discharge) will be extracted from hospitalization records.

A linear regression model will be used to assess whether the occurrence of a postoperative infection is associated with a prolonged hospital stay, adjusted for relevant covariates.

Unit of measure : Median difference in number of days, with 95% Confidence Interval