Abstract
Aim
Smoking is known to impair wound healing and to increase the risk of peri-operative adverse events and is associated with orthopaedic infection and fracture non-union. Understanding the magnitude of the causal effect on orthopaedic infection recurrence may improve pre-operative patient counselling.
Methods
Four prospectively-collected datasets including 1173 participants treated in European centres between 2003 and 2021, followed up to 12 months after surgery for clinically diagnosed orthopaedic infections, were included in logistic regression modelling with Inverse Probability of Treatment Weighting for current smoking status [1–3]. Host factors including age, gender and ASA score were included as potential confounding variables, interacting through surgical treatment as a collider variable in a pre-specified structural causal model informed by clinical experience. The definition of infection recurrence was identical and ascertained separately from baseline factors in three contributing cohorts. A subset of 669 participants with positive histology, microbiology or a sinus at the time of surgery, were analysed separately.
Results
Participants were 64% male, with a median age of 60 years (range 18–95); 16% of participants experienced treatment failure by 12 months. 1171 of 1173 participants had current smoking status recorded. As expected for the European population, current smoking was less frequent in older participants (Table 1). There was no baseline association between Charlson score or ASA score and smoking status (p=0.9, p=1, Chi squared test). The estimated adjusted odds ratio for treatment failure at 12 months, resulting from current smoking at the time of surgery, was 1.37 for all participants (95% CI 0.75 to 2.50) and 1.53 for participants with recorded confirmatory criteria (95% CI 1.14 to 6.37).
Conclusions
Smoking contributes to infection recurrence, particularly in people with unequivocal evidence of osteomyelitis or PJI. People awaiting surgery for orthopaedic infection should be supported to cease smoking, not only to reduce anaesthetic risk, but to improve treatment outcomes. Limitations of this study include unmeasured socioeconomic confounding and social desirability bias resulting in uncertainty in true smoking status, resulting in underestimated effect size.
