Antidepressant medication use before and after kidney transplant: implications for outcomes - a retrospective study

Abstract

We examined a novel database wherein national US transplant registry identifiers were linked to records from a large pharmaceutical claims warehouse (2008-2015) to characterize antidepressant use before and after kidney transplantation, and associations [adjusted hazard ratio (aHR) 95% CI] with death and graft failure. Among 72 054 recipients, 12.6% filled antidepressant medications in the year before transplant, and use was more common among women and patients who were white, unemployed, and had limited functional status. Pre-transplant antidepressant use was associated with 39% higher 1-year mortality (aHR 1.39, 95% CI 1.18-1.64) and 15% higher all-cause graft loss risk (aHR 1.15, 95% CI 1.02-1.30). More than 50% of patients who filled antidepressants pre-transplant continued fill post-transplant. Antidepressant use in the first year after transplant was associated with twofold higher risk of death (aHR 1.94, 95% CI 1.60-2.35), 38% higher risk of death-censored graft failure, and 61% higher risk of all-cause graft failure in the subsequent year. Pre-listing antidepressant use was also associated with increased mortality, but transplantation conferred a survival benefit regardless of prelisting antidepressant use status. While associations may in part reflect underlying behaviors or comorbidities, kidney transplant candidates and recipients treated with antidepressant medications should be monitored and supported to reduce the risk of adverse outcomes.

Clinical depression as an unfavorable prognostic factor following kidney transplantation-How can we explain it?

“…For the most part, the Lentine et al . findings confirm those reported in three previous kidney transplant studies showing significant associations between being clinically depressed and having higher death and DCGF rates . Dobbels et al .…”

“…While recipient comorbidities were individually controlled in the Lentine et al . Cox multivariable models, potential explanations (ii)‐(iv) were either not controlled or poorly controlled (in our opinion). For instance, a distinction between being retired versus of working age but not working may have offered a clear prognostic separation among patients “not working,” and within the large subgroup of patients having no physical limitations, no further subdivision according to a physical activity measure (inactive to very active) was made.…”

“…For the most part, the Lentine et al [1] findings confirm those reported in three previous kidney transplant studies showing significant associations between being clinically depressed and having higher death and DCGF rates [4][5][6]. Dobbels et al [4] reported using USRDS and Medicare claims data that among 47 897 1st kidney transplant recipients (transplanted during 1995-2003), clinical depression (observed in 7.0% of patients) as a binary time-dependent covariate was associated with significantly higher death-with-a functioning graft (DWFG) and DCGF rates during the first 3 years posttransplant (P < 0.001 each, aHR = 2.24 and 1.97, respectively).…”

“…Dobbels et al [4] reported using USRDS and Medicare claims data that among 47 897 1st kidney transplant recipients (transplanted during 1995-2003), clinical depression (observed in 7.0% of patients) as a binary time-dependent covariate was associated with significantly higher death-with-a functioning graft (DWFG) and DCGF rates during the first 3 years posttransplant (P < 0.001 each, aHR = 2.24 and 1.97, respectively). However, none of the potential explanations (i)-(iv) offered by Lentine et al [1] were controlled as explanatory variables in the Dobbels et al [4] Cox models.…”

“…In this issue, Lentine et al . provide a well thought‐out analysis of the prognostic value of antidepressant medication use (as a surrogate measure of clinical depression) on post‐transplant outcome. Specifically, they show that among 72 054 SRTR patients receiving a kidney transplant during 2008–2015, antidepressant use (obtained from pharmaceutical fill records) within 1 year of transplant and during the first year post‐transplant (12.6% and 18.0% of patients, respectively) correlated with a significantly higher death rate during the first and 2nd year post‐transplant ( P < 0.0001 each, aHR = 1.39 and 1.94, respectively).…”