Bisphosphonate effectiveness in patients with cirrhosis: An emulated clinical trial.

Abstract

Falls and fractures are common and morbid for patients with cirrhosis. Bisphosphonates are recommended for the prevention of fractures for people with osteoporosis cirrhosis; however, data supporting effectiveness in cirrhosis are lacking. We sought to emulate a clinical trial of bisphosphonates in cirrhosis. We used national Medicare data (2008-2020) to examine the 5-year risk of fractures in patients who did or did not receive bisphosphonates with a new-user design among people diagnosed with cirrhosis and osteoporosis. We balanced treated and untreated with inverse probability of treatment weighting, evaluated intention-to-treat and as-treated effects, and examined both control exposures (statin use) and outcomes (decompensation) to test causal relationships. There were 253 and 20,888 new users and non-users of bisphosphonates, respectively. The median age was 74 years. The most common bisphosphonate used was alendronate (73.6%). Bisphosphonates significantly reduced fractures overall (27.5% vs. 33.0%, p = 0.0004) in the intention-to-treat analysis, particularly for people <65 years (sHR 0.56) old, men (sHR 0.64) and those with non-alcohol related liver disease (sHR 0.85). Though there were fewer arm (20.7% vs. 26.4%, p < 0.0001) and femur (28.9% vs. 31.2%, p = 0.005), there were more spinal (25.8% vs. 19.0%), rib (40.0% vs. 32.2%) and skull (10.1% vs. 8.7%) fractures. In the as-treated analysis, cumulative bisphosphonate exposure significantly reduced fractures, sHR 0.95 95% CI (0.91, 0.98). Treatment was inconsistent; bisphosphonate users spent 29.9% person-years of follow-up on the drug. In a nationally representative cohort of elderly patients with cirrhosis, bisphosphonates reduced fractures overall. Efforts to increase uptake and drug continuation are needed.