Background and purpose - In-hospital death following total hip arthroplasty (THA) is related to comorbidity. The long-term effect of comorbidity on all-cause mortality is, however, unknown for this group of patients and it was investigated in this study. Patients and methods - We used data from the Swedish Hip Arthroplasty Register, linked to the National Patient Register from the National Board of Health and Welfare, for patients operated on with THA in 1999-2012. We identified 120,836 THAs that could be included in the study. We evaluated the predictive power of the Charlson and Elixhauser comorbidity indices on mortality, using concordance indices calculated after 5, 8, and 14 years after THA. Results - All comorbidity indices performed poorly as predictors, in fact worse than a base model with age and sex only. Elixhauser was, however, the least bad choice and it predicted mortality with concordance indices 0.59, 0.58, and 0.56 for 5, 8, and 14 years after THA. Interpretation - Comorbidity indices are poor predictors of long-term mortality after THA.