‘Fit for surgery’; retrospective analysis of the relationship between cardiorespiratory fitness and postoperative outcome

Student thesis: Doctoral Thesis


Introduction: The aging population is a major concern for healthcare providers and the number of surgical procedures performed is increasing each year. The ‘high-risk’patient accounts for 13% of surgical cases but contributes to over 80% of postoperative deaths. Evidence suggests that cardiorespiratory fitness (CRF) may be an independent predictor of postoperative outcome. However, this relationship requires further understanding and optimisation to better inform patient care.

Aims: The overarching objective was to explore the ‘potential’ relationship between CRF and post operative outcome (morbidity and survival) in patients undergoing major intra-abdominal surgery. Three aims were established to: 1) Improve the detection and interpretation of CRF, 2) Explore novel thresholds of CRF predictive of postoperative outcome, and 3) Enhance patient management using exercise.

Hypotheses: It was hypothesised that: 1) Natural variation (biological and analytical noise) is present in markers of CRF and thus impacts upon patient fitness stratification, 2) CRF is impaired in diseased patients and can predict postoperative outcomes, 3) Preoperative CRF is lower in females which may translate into inferior postoperative outcomes over males, 4) Preoperative exercise training is well tolerated and associated with objective cardiopulmonary improvement.

Methodology: Study 1 –In a two-armed experiment, natural variation was calculated for CRF in a young, healthy population. Subsequent values of natural variation were retrospectively applied to an anonymised database of patients who underwent preoperative cardiopulmonary exercise testing (PCPET) before colorectal surgery, to re-appraise fitness stratification. Study 2 –A retrospective cross-sectional analysis of patients (n=124) with abdominal aortic aneurysm (AAA) was conducted to compare CRF with that of a matched apparently healthy 3cohort, and to examine the association between impaired CRF and postoperative outcome. Study 3 –In a large cohort of patients (n=640) who underwent PCPET prior to colorectal surgery, firstly, the association between impaired CRF and postoperative outcome was investigated and compared with traditional cardiovascular disease (CVD) risk factors. A subsequent comparative analysis was conducted to investigate sex-differences in preoperative CRF and postoperative outcomes to re-appraise risk stratification. Study 4 –A case-report was conducted describing a 70-year-oldhigh-risk female patient with a complicated medical history, who required major thoraco-abdominal surgery. A preoperative supervised 10-week high intensity interval training (HIIT) exercise intervention was conducted, and its ability to improve perioperative risk stratification evaluated.

Results: Study 1–Natural variation was present in measures of CRF and accounted for up to ± 19%, 13%, and 10% for oxygen consumption at anaerobic threshold (V̇O2-AT),peak oxygen consumption(V̇O2 peak), and ventilatory equivalent of carbon dioxide at anaerobic threshold (V̇E/V̇CO2-AT) respectively. A theoretical potential for up to 60% of patients to have indeterminate fitness existed if natural variation was not considered. Study 2 –Patients with AAA undergoing PCPET highlighted impaired CRF compared to age adjusted/sex-matched sedentary controls. Values of<13.1 mL O2.min-1.kg-1for V̇O2 peak and ≥34 for V̇E/V̇CO2-AT were independent predictors of postoperative mortality at 2-years. Study 3 –Being ‘unfit’ defined by preoperative CRF (V̇O2 peak <14.3mL kg-1 min-1and > 34 for V̇E/V̇CO2-AT) identified a five-fold greater 1-year mortality rate and was a stronger predictor than traditional CVD risk factors in a large cohort of patients undergoing colorectal surgery. Female patients exhibited lower preoperative CRF, and more were stratified ‘high risk’, however postoperative outcomes were equivalent to males. Consequently, females demonstrated lower threshold values of CRF than male counterparts and the application of sex-specific thresholds improved the prediction of postoperative mortality. Study 4 –10 weeks of HIIT proved well tolerated and conferred impressive gain in CRF (27 and 36% for V̇O2-AT and V̇O2 peak respectively) which exceeded sources of variation and positively changed perioperative risk stratification in a high-risk patient prior to major thoraco-abdominal surgery.

Discussion: The overarching premise that CRF is related to postoperative outcome in patients undergoing intra-abdominal surgery is strongly supported. CRF was impaired relative to similarly aged apparently healthy people prior to major surgery and being unfit was a stronger predictor of mortality than traditional CVD risk factors. Furthermore,a 3 to 5-fold greater risk of postoperative mortality occurred in patients undergoing vascular and colorectal surgery if stratified unfit. This work has demonstrated: 1) Improved detection and interpretation of CRF, however unlike previous work the use of V̇O2-AT is not supported, the consequence of experimental ‘noise’ (mostly biological variation) that requires consideration when interpreting PCPET results. 2). Novel threshold values of CRF in specific patient cohorts undergoing surgery improved mortality prediction, and importantly, patients should be stratified by sex as females are more sensitive to CRF, with lower values yet equivalent postoperative outcomes. 3) Preoperative CRF was objectively improved using a short duration HIIT intervention that was well tolerated in the high-risk patient and enabled fitness to transcend stratification boundaries.

Conclusion: Preoperative CRF better predicted postoperative outcome than traditional CVD risk factors, and PCPET should be considered a principal component of surgical risk assessment. This work advances the potential to use CRF to predict postoperative outcome to help clinicians better direct care provision and advance patient outcomes following major surgery.
Date of Award2022
Original languageEnglish
SupervisorDamian Bailey (Supervisor) & Richard G. Davies (Supervisor)

Cite this