Body mass index is used by the World Health Organization to classify obesity. While obesity influences the onset of arthritis and type-2 diabetes, its effect on implant survival is still open to debate, with conflicting results from clinical and registry studies, as well as meta-analyses. Other known factors such as gender or diabetes status could ponderate or mask the effect of BMI on implant survival.

Our hypothesis was BMI influenced hip and knee arthroplasty survival, when results were made independent of gender and diabetes status.

A registry study was designed on 30733 Total Hip Arthroplasties (THA), 28483 Total Knee Arthroplasties (TKA), 3754 Uni compartmental Knee Arthroplasties (UKA) and 649 Hinged Knee arthroplasties (HK), from 01/01/2003 to 31/12/2015. Mean follow-up was 5.5 years. Diabetes status was added to the model. Each arthroplasty survival was tested for age at implantation, gender, diabetes status, implant characteristics and specifically BMI, taking into account gender and diabetes status.

Gender had a strong influence on arthroplasty results. Age also influenced arthroplasty survival, especially aseptic loosening; a young age would lower implant survival. Diabetes had an influence in hip survival, but its influence on septic loosenings in TKA wasn’t proven (p=0.065). A mobile liner and/or a cruciate retaining knee were factors increasing the risk of revision. Weight influenced THA survival, especially aseptic loosening, but didn’t have a measurable effect in any other arthroplasty. BMI was not found to influence any arthroplasty survival, whatever the endpoint, when diabetes and gender were taken into account.

Gender, age and diabetes influenced survival of the lower limb arthroplasties, whereas BMI did not. Only weight did influence THA results and should be used instead of BMI.

Studies on arthroplasty survival should systematically mention gender and diabetes status and beware of potential group incomparability.

III, cohort study.