Comment on: Multimodal prehabilitation and perioperative immune function in patients undergoing abdominal cancer surgery

Correspondence to: Imen Ben Ismail (email: imen_bi@yahoo.fr)

University of Tunis el Manar

Faculty of Medicine of Tunis

Tunis

Tunisia

_____

BJS, https://doi.org/10.1093/bjs/znag070, published 06 June 2026

_____

We commend Jacobs et al.¹ for the first mechanistic evaluation of prehabilitation on perioperative immune function, but two design features temper the causal interpretation offered.

First, oesophageal and rectal pathways had already transitioned to prehabilitation before the study period, so postoperative comparisons were restricted to bladder cancer, where no significant difference from control was observed at any time point. The principal mechanistic conclusion that prehabilitation modulates basal immune function, therefore, relies predominantly on uncontrolled within-patient comparisons rather than contemporaneous between-group comparisons, a limitation inherent to stepped-wedge designs². Changes occurring during the several-week interval before surgery, including recovery from neoadjuvant chemoradiotherapy (46.1% of the cohort) and other time-dependent factors, cannot readily be separated from a specific effect of prehabilitation.

Second, ex vivo cytokine production is reported as absolute LPS-stimulated values from whole blood, without adjustment for circulating leucocyte counts, which fluctuated substantially perioperatively. Reporting cytokine production per monocyte, or after leucocyte-count adjustment, would better distinguish altered cellular responsiveness from altered cell abundance.³ Moreover, without a predefined washout period before blood sampling, acute exercise-induced leukocytosis and myokine-driven IL-6 release⁴ may have contributed to the observed differences.

Finally, laboratory personnel were unblinded to allocation, and correction for multiple testing was not applied⁵; some nominally significant findings should therefore be interpreted cautiously. Leucocyte-normalised data, standardised sampling relative to exercise, and replication within concurrently controlled cohorts would substantially strengthen this important hypothesis.