Prevalence and Predictors of Immunosuppression Nonadherence Following Liver Transplantation

Abstract

Introduction: Nonadherence to immunosuppression (IS) medications following liver transplantation (LT) increases risk of graft failure and death. We assessed the prevalence of IS nonadherence and the predictive utility of six psychosocial risk factors for post-LT nonadherence. Method: Patients meeting the following study inclusion criteria at two transplant centers participated in the study: primary LT recipient, 6 to 24 months post-LT, age ≥ 18 years old, and English-speaking. A telephone interview included questions about sociodemographics, physical health, and IS medication adherence in the past 6 months. Patients were categorized as adherent (missed IS meds “never” or “no more than a few times per month”) or nonadherent (missed IS meds “once per week” or “more than once per week”). In addition, patients reported the frequency of IS holidays of 24h, 48h, and 72h. Potential pre-LT predictors were gathered from medical records and included lab MELD at time of LT, time since LT, and the presence/absence of 6 hypothesized pre-LT psychosocial risk factors based on a priori criteria: history of mood disorder, passive coping style, prior medication nonadherence, history of substance abuse/dependence, limited social support, and social support instability. Results: Two hundred thirty-six patients participated in the study. Mean age was 53 (±11) years, 61% were male, 86% were White, 68% were married, and 37% had attended college. Mean lab MELD at LT was 21.2 (±8) and mean time since LT was 14.4 (±5) months. Eighty-two patients (35%) were classified as nonadherent. Thirty-four patients (14%) reported changing how they took IS medication without first consulting with transplant provider. Seventy-one patients (30%) reported at least one 24 hr IS holiday, 38 (16%) reported going 48 hrs without IS medications, and 23 (10%) reported at least one 72 hr IS holiday in the past 6 months. Univariate analyses showed that, compared to adherent patients, nonadherent patients were more likely to have longer time since LT (13.6 vs.15.6 months, t=2.8, P=0.006), be male (56% vs. 70%, P=0.02), view their overall health as unchanged or worse since LT (10% vs. 27%, P=0.003), and have the following in the 24 months prior to LT: mood disorder (43% vs. 64%, P=0.002), medication nonadherence (28% vs. 51%, P< 0.001), limited social support (27% vs. 47%, P=0.002), and social support instability (18% vs. 36%, P=0.001). However, in the multivariate logistic regression model, IS nonadherence was predicted by male gender (OR=0.43, 95% CI=0.22, 0.85), longer time since LT (OR=1.08, CI=1.02, 1.14), pre-LT mood disorder history (OR=0.45, CI=0.25, 0.84), and limited social support at time of LT (OR=0.45, CI=0.22, 0.91). Nineteen percent of patients with 0-1 pre-LT psychosocial risk factors were classified as nonadherent, compared to 32% with 2-3 risk factors and 60% with ≥4 risk factors (χ2=24.4, P< 0.001). Conclusions: Rates of IS nonadherence and drug holidays in the first two years after LT are unacceptably high. Some pre-LT psychosocial factors, including a history of mood disorder and limited social support, increase the risk of post-LT nonadherence in this population. Interventions should target modifiable risk factors both before and immediately after LT.