Cirrhosis is a risk factor for total hip arthroplasty for avascular necrosis: A Danish nationwide cohort study

Deleuran, Thomas; Overgaard, Søren; Vilstrup, Hendrik; Jepsen, Peter

The main strength of this study was its population-based design and complete follow-up. One possible limitation was the validity of our data sources. NPR diagnoses for cirrhosis have previously been validated with biopsy or clinical evaluation as the gold standard. Cirrhosis was confi rmed in 85% of patients with this diagnosis in the NPR (Vestberg et al. 1997) , and the bias introduced by misclassifying individuals without cirrhosis as cirrhosis patients, or vice versa, would cause us to underestimate the true strength of the association between cirrhosis and AVN. The indication for hip arthroplasty in the DHR has been validated, showing that the positive predictive value for AVN was 99% (Pedersen et al. 2004 ). Still, it is possible that orthopedic surgeons are reluctant to perform arthroplasties in patients with cirrhosis, who have a high risk of postoperative complications (Deleuran et al. 2014 ). However, such a bias would cause us to underestimate the risk of Hung et al. (2011) found an HR for AVN of 2.5 in a cohort of patients with cirrhosis (primarily due to viral hepatitis) relative to an age-matched cohort of hospitalized patients without cirrhosis. These hospitalized patients without cirrhosis had a high prevalence of conditions that predispose to AVN, but the reported associations were only adjusted for conditions recorded at the index hospitalization. Such incomplete recording of confounders may have resulted in residual confounding, which would have caused Hung et al. to underestimate the true strength of the association.