A statistical model to determine the optimal bone mineral density (BMD) screening schedule in liver transplant recipients

Background

The American Association for the Study of Liver Diseases (AASLD) and the American Society of Transplantation (AST) 2012 practice guidelines based on expert opinion provide surveillance frequency guidelines for dual-energy X-ray absorptiometry (DXA) bone mineral density (BMD) scans after liver transplant (LT). Since then, post-transplant immunosuppression has changed. This study aims to use statistical models to identify the optimal DXA surveillance frequency after LT.

Materials and Methods

This study retrospectively evaluated 402 LT recipients followed for up to 8 years post-LT, each with at least one pre-LT and two post-LT DXA scans. Linear mixed-effects (LME) models using random slopes and intercepts were used to identify whether BMD decline was linear or non-linear (quadratic) in time. Multivariate LME models were used to model the decline in femoral neck (FN), total hip (TH), and L1-L4 BMD decline after LT adjusting for demographic and clinical variables.

Results

Males had higher pre-LT BMD than females (P<0.001 at all sites). The rate of BMD loss was fastest at the FN, and faster in patients with normal pre-LT BMD than in patients with low pre-LT BMD (osteopenia/osteoporosis). The model predicted that there would be a significant FN BMD decrease in patients with normal pre-LT BMD after approximately 481 days post-LT (1 year, 3 months).

Conclusion

LT recipients with normal pre-LT BMD should have DXA scans ∼481 days post-LT. LT recipients with low BMD pre-LT who were more likely to be treated with bisphosphonates did not need annual DXA screening.