64 The Impact of Thoracotomy and Choice of Analgesic Modality on Post-Operative Intensive Care Requirements After Oesophagectomy

Abstract Aim Oesophagectomy is a high-risk operation with major morbidity. Operation type and analgesic modality may influence outcomes. Non-epidural regional anaesthesia (RA) is potentially safer than epidural and theorised to minimise systemic hypotension and vasoconstrictor requirements. We aimed to assess whether surgical approach and analgesic technique influence vasoconstrictor use and ICU admission after oesophagectomy. Method This single-site retrospective observational study included all patients undergoing oesophagectomy between 2018-2023. Patients were grouped by analgesic modality: intrathecal morphine (ITM) only; epidural; or non-epidural RA, and by operation type: thoracic or non-thoracic. Groups were compared for ICU admission odds and total noradrenaline use using multivariable logistic regression and Mann-Whitney U-tests. Results We identified 254 patients who underwent thoracic (n=159) or transhiatal oesophagectomy (n=95). 121 received ITM only; 133 had RA (44 epidural, 89 non-epidural). Eighty-five patients required elective or emergency ICU admission. Adjusted for age and analgesic modality, thoracic oesophagectomy was associated with higher ICU admission odds (OR=3.02, 95% CI:1.39-6.74, p<0.01). Among patients not receiving epidural, thoracic oesophagectomy still conferred increased ICU admission risk (OR=2.84, 95% CI:1.26-6.44, p<0.05). Non-epidural RA showed a non-significant trend toward reduced ICU admission versus epidural (OR=0.50, 95% CI:0.23-1.09, p=0.081). Median (Q1-Q3) total noradrenaline requirement was 1.7mg (0.0-16.3). No significant differences in vasoconstrictor use or length of ICU stay were seen by operation type (pvasoconstrictor=0.65, pICU=0.12) or analgesic modality (pvasoconstrictor=0.62, pICU=0.40). Conclusions Thoracic oesophagectomy increases odds of post-operative ICU admission. Non-epidural RA trended toward reduced ICU admission but was not significant. Our single-centre retrospective design and unmeasured confounders limit interpretation.