Inflammation and postoperative complications after major surgery: a bi-directional two-sample Mendelian randomization study

Background Observational studies have reported an association between inflammation and postoperative complications but it is unclear whether these associations are causal. It is also unknown whether postoperative outcomes share a causal architecture with chronic, all-cause disease. Methods We performed bi-directional two-sample Mendelian randomization to investigate potential causal effects of 19 genetically-proxied inflammatory markers on postoperative acute kidney injury, atrial fibrillation (AF), delirium, myocardial infarction, stroke and surgical site infection, and their all-cause equivalents. Genetic instruments for inflammatory markers were sourced from nine GWAS of up to 204,402 European participants with outcome data derived from UK Biobank. Results The primary postoperative analysis showed a protective effect of down-regulated IL-6 signalling on stroke risk (OR (95% CI) 0.27 (0.11--0.69), p=0.006). However, in the all-cause analysis a causal effect on stroke was not present (OR (95% CI) 1.14 (0.75--1.24), p=0.78), whilst a robust protective effect was seen for down-regulated IL-6 with AF across all three instruments studied (all p<0.009). In postoperative and all-cause analyses, genome-wide variants for CRP showed a protective effect on delirium that was not present in cis-restricted analyses. Conclusions We found evidence supporting a potential causal role for IL-6 signalling in perioperative stroke. However, the divergence in IL-6 effects between postoperative and all-cause outcomes suggests that the inflammatory architecture of acute postoperative complications may differ from chronic disease states. Furthermore, our findings suggest previously reported associations between CRP and delirium likely represent horizontal pleiotropy rather than direct causation. Future work should interrogate local tissue responses and the immediate perioperative period.