Abstract Variables from cardiopulmonary exercise testing (CPET) has been used to predict negative prognostic markers in oesophagectomy. We investigate the predictive value of anaerobic threshold (AT) ≥11 mL/kg/min in predicting length of stay (LOS) in patients having Ivor Lewis oesophagectomy (ILO). We also correlate AT with pulmonary complications (pneumonia), anastomotic leak rate and survival in our outcome analysis. A retrospective analysis of a prospectively maintained database was carried out at a regional centre in the UK, evaluating patients undergoing ILO for oesophageal cancer (OC) between May 2016 and December 2019. Data was collected using the hospitals electronic system containing clinic letters, anaesthetic assessment including CPET, pathology reports and information regarding death of the patient. Patients were divided into AT <11 mg/kg/min (LAT) and those with AT ≥11 mg/kg/min (HAT). χ2 test of association were used for categorical variables and Mann–Whitney test to compare two sets of non-parametric data. A p < 0.05 was considered significant. Kaplan–Meier curve was used for survival analysis. 153 patients underwent ILO (median age 67, 86% males). 70 from this cohort underwent CPET testing. Median AT was 10.5 (7.4–11.7). There was a statistically significant negative correlation between AT and LOS (p = 0.03). Median LOS for patients with LAT was 15.5 days (11–225) and HAT was 13.5 days (10–19.75) with no significant differences in LOS when comparing LOS between LAT and HAT (p = 0.09). There was no difference in anastomotic leaks between LAT and HAT (χ2 test) (p = 0.15). Overall survival estimated using Kaplan–Meier curves failed to demonstrate a statistical difference between LAT and HAT (median 28.5 vs 30.5) (p = 0.71). Our analysis shows a negative correlation between AT and LOS. Although LOS, anastomotic leaks, and survival were quantitatively better in the HAT group, it failed to reach statistical significance. This suggests that CPET variables are more beneficial as adjuncts to help shared decision making but not to predict post-operative outcomes. AT does not predict LOS in patients undergoing ILO.