Background and purpose - Body mass index (BMI) outside the normal range possibly affects the perioperative morbidity and mortality following total hip arthroplasty (THA) and total knee arthroplasty (TKA) in traditional care programs. We determined perioperative morbidity and mortality in such patients who were operated with the fast-track methodology and compared the levels with those in patients with normal BMI. Patients and methods - This was a prospective observational study involving 13,730 procedures (7,194 THA and 6,536 TKA operations) performed in a standardized fast-track setting. Complete 90-day follow-up was achieved using national registries and review of medical records. Patients were grouped according to BMI as being underweight, of normal weight, overweight, obese, very obese, and morbidly obese. Results - Median length of stay (LOS) was 2 (IQR: 2-3) days in all BMI groups. 30-day re-admission rates were around 6% for both THA (6.1%) and TKA (5.9%), without any statistically significant differences between BMI groups in univariate analysis (p > 0.4), but there was a trend of a protective effect of overweight for both THA (p = 0.1) and TKA (p = 0.06). 90-day re-admission rates increased to 8.6% for THA and 8.3% for TKA, which was similar among BMI groups, but there was a trend of lower rates in overweight and obese TKA patients (p = 0.08 and p = 0.06, respectively). When we adjusted for preoperative comorbidity, high BMI in THA patients (very obese and morbidly obese patients only) was associated with a LOS of >4 days (p = 0.001), but not with re-admission. No such relationship existed for TKA. Interpretation - A fast-track setting resulted in similar length of hospital stay and re-admission rates regardless of BMI, except for very obese and morbidly obese THA patients.