Anaïs Couret1, Adeline Gallini2, Mathilde Poncet3, Axel Renoux4, Maryse Lapeyre-Mestre5, Virginie Gardette2. 1. UMR INSERM 1027, INSERM, Toulouse, France; Université Toulouse III Paul Sabatier, Toulouse, France. Electronic address: anaiscouret@orange.fr. 2. UMR INSERM 1027, INSERM, Toulouse, France; Université Toulouse III Paul Sabatier, Toulouse, France; Centre Hospitalier Universitaire de Toulouse, Department of Epidemiology and Public Health, Toulouse, France. 3. Université Toulouse III Paul Sabatier, Toulouse, France; Centre Hospitalier Universitaire de Toulouse, Toulouse, France. 4. Centre Hospitalier Universitaire de Toulouse, Toulouse, France. 5. UMR INSERM 1027, INSERM, Toulouse, France; Université Toulouse III Paul Sabatier, Toulouse, France; Centre Hospitalier Universitaire de Toulouse, Department of Pharmacology, Toulouse, France.
Abstract
OBJECTIVE: To study temporal trends of benzodiazepine exposure among incident Alzheimer's disease and related dementia (ADRD) cohorts between 2011 and 2016. DESIGN: Repeated cross-sectional study. SETTING AND PARTICIPANTS: Three nationwide incident ADRD cohorts (community-dwelling and institutionalized subjects) were identified in 2011, 2013, and 2016 through the French health insurance database. Subjects were followed 4 semesters around ADRD identification [Semester -2 (S-2) to Semester 2 (S2)]. MEASURES: Odds ratios (ORs) for semestrial prevalent exposure, initiation, and adherence to benzodiazepine prescription recommendations (prescription duration <3 months, single reimbursement) were computed using multivariate logistic regressions for each cohort and according to benzodiazepine half-life. RESULTS: Among 262,024 community-dwelling subjects, as compared to 2011, overall benzodiazepine prevalence risk decreased slightly immediately after ADRD identification [S1: aOR2013 = 0.93 (0.91-0.95), aOR2016 = 0.95 (0.93-0.97)] and did not differ during S2. Among 72,013 institutionalized subjects, it increased over time [S2: aOR2013 = 1.16 (1.11-1.21), aOR2016 = 1.26 (1.21-1.32)]. Long half-life benzodiazepine prevalence risk decreased in the 4 semesters among recent cohorts, for both populations [S2: community-dwelling: aOR2013 = 0.77 (0.74-0.79), aOR2016 = 0.61 (0.59-0.64); institutionalized: aOR2013 = 0.74 (0.68-0.80), aOR2016 = 0.58 (0.54-0.63)]. Short half-life benzodiazepine prevalence risk increased [S2: community-dwelling: aOR2013 = 1.13 (1.10-1.16), aOR2016 = 1.22 (1.20-1.25); institutionalized: aOR2013 = 1.26 (1.21-1.32), aOR2016 = 1.44 (1.38-1.50)]. The same patterns were observed for benzodiazepine initiation. Adherence to benzodiazepine prescription recommendations (based on French prescription duration) worsened over years [prescription duration <3 months: aOR2013 = 0.90 (0.86-0.95), aOR2016 = 0.90 (0.85-0.95), single reimbursement: aOR2013 = 0.95 (0.91-1.00), aOR2016 = 0.94 (0.90-0.99)]. CONCLUSIONS AND IMPLICATIONS: Long half-life benzodiazepine exposure was reduced whereas short half-life benzodiazepine exposure increased, and adherence to recommendations worsened (prescription duration longer than 3 months and more than a single reimbursement in recent cohorts). Efforts from prescribers and authorities are required in order to restrict psychotropic exposure among the ADRD population. Further research among institutionalized ADRD subjects could provide useful data to disentangle the effects of changes in prescribing practices and in patients' characteristics.
OBJECTIVE: To study temporal trends of benzodiazepine exposure among incident Alzheimer's disease and related dementia (ADRD) cohorts between 2011 and 2016. DESIGN: Repeated cross-sectional study. SETTING AND PARTICIPANTS: Three nationwide incident ADRD cohorts (community-dwelling and institutionalized subjects) were identified in 2011, 2013, and 2016 through the French health insurance database. Subjects were followed 4 semesters around ADRD identification [Semester -2 (S-2) to Semester 2 (S2)]. MEASURES: Odds ratios (ORs) for semestrial prevalent exposure, initiation, and adherence to benzodiazepine prescription recommendations (prescription duration <3 months, single reimbursement) were computed using multivariate logistic regressions for each cohort and according to benzodiazepine half-life. RESULTS: Among 262,024 community-dwelling subjects, as compared to 2011, overall benzodiazepine prevalence risk decreased slightly immediately after ADRD identification [S1: aOR2013 = 0.93 (0.91-0.95), aOR2016 = 0.95 (0.93-0.97)] and did not differ during S2. Among 72,013 institutionalized subjects, it increased over time [S2: aOR2013 = 1.16 (1.11-1.21), aOR2016 = 1.26 (1.21-1.32)]. Long half-life benzodiazepine prevalence risk decreased in the 4 semesters among recent cohorts, for both populations [S2: community-dwelling: aOR2013 = 0.77 (0.74-0.79), aOR2016 = 0.61 (0.59-0.64); institutionalized: aOR2013 = 0.74 (0.68-0.80), aOR2016 = 0.58 (0.54-0.63)]. Short half-life benzodiazepine prevalence risk increased [S2: community-dwelling: aOR2013 = 1.13 (1.10-1.16), aOR2016 = 1.22 (1.20-1.25); institutionalized: aOR2013 = 1.26 (1.21-1.32), aOR2016 = 1.44 (1.38-1.50)]. The same patterns were observed for benzodiazepine initiation. Adherence to benzodiazepine prescription recommendations (based on French prescription duration) worsened over years [prescription duration <3 months: aOR2013 = 0.90 (0.86-0.95), aOR2016 = 0.90 (0.85-0.95), single reimbursement: aOR2013 = 0.95 (0.91-1.00), aOR2016 = 0.94 (0.90-0.99)]. CONCLUSIONS AND IMPLICATIONS: Long half-life benzodiazepine exposure was reduced whereas short half-life benzodiazepine exposure increased, and adherence to recommendations worsened (prescription duration longer than 3 months and more than a single reimbursement in recent cohorts). Efforts from prescribers and authorities are required in order to restrict psychotropic exposure among the ADRD population. Further research among institutionalized ADRD subjects could provide useful data to disentangle the effects of changes in prescribing practices and in patients' characteristics.