Does the Prescription of Anxiolytic and Hypnotic Drugs Increase Mortality in Older Adults?

Abstract

Listen

Translate article

To the Editor: Anxiolytics and hypnotics are widely prescribed for older adults, but they may cause severe side effects (e.g., falls, altered consciousness, confusion) and even increase mortality.1, 2 The objective of this study was to investigate the relationship between the prescription of anxiolytics and hypnotics and mortality in old people followed for 2 to 4 years. The Système d'Information sur la Perte d'Autonomie Fonctionnelle (SIPAF) Study included 2,350 French recipients of a supplementary pension fund aged 70 and older. Trained nurses interviewed participants at home between 2008 and 2010. Information was collected in a standardized geriatric assessment on sociodemographic characteristics, chronic diseases (including history of anxiety or depression), difficulties in activities of daily living, cognitive impairment (Mini-Mental State Examination score ≤ 26), social support, use of healthcare resources, and self-perceived health. Drugs prescribed were determined from prescriptions that participants had at home and identified using the Anatomical Therapeutic Chemical Classification System. Prescription of anxiolytics or hypnotics (PAH) was defined as the prescription of a benzodiazepine or benzodiazepine-related medication (N05BA, N05CD, N05CF, N03AE01, M03BX07), meprobamate (N05BC01), meprobamate with aceprometazine (N05CX01), or an anxiolytic or hypnotic with anticholinergic properties (R06AA09, R06AD01, N05BB01). Mortality was recorded until June 2012 (median follow-up 2.8 years). The effect of PAH on mortality was assessed using a Cox proportional hazard model with inverse probability of treatment weighting (IPTW). An independent ethics committee approved the research protocol (permission 060313). Analyses were performed using Stata version 13.0 (Stata Corp., College Station, TX). The mean age of the participants was 83 ± 8; 60% were female. Participants received on average 6 ± 3 drugs per day, and 677 (30%) had a PAH. More precisely, 456 (20%) had a prescription for an anxiolytic and 264 (12%) for a hypnotic. Zolpidem (19%), lorazepam (17%), bromazepam (16%), and zopiclone (14%) were the most frequent PAH. During follow-up, 368 (16%) participants died. Participants with a PAH were more likely to die during follow-up (19%) than those without (14%, hazard ratio (HR) = 1.42, 95% confidence interval (CI) = 1.14–1.76). The propensity score (or probability of treatment) included variables related to the PAH and to mortality (sex, age, economic status, polypharmacy, self-perceived health, history of anxiety or depression, dependency, social isolation, marital status, body mass index ≤ 18.5 kg/m2 or weight loss, cognitive impairment, hospitalization in the past 6 months, respiratory disease, cancer). IPTW has improved data comparability between treated and untreated people (Figure 1). After IPTW adjustment, the association between PAH and mortality was not significant (HR = 1.08, 95% CI = 0.84–1.40). The comprehensive geriatric assessment, using validated instruments, enabled a number of confounding factors that could modify the association between PAH and mortality to be taken into account. Using propensity score with a weighting strategy allowed the analyses to be conducted in the whole study sample, whereas matching strategies often restrict the study population. Furthermore, examination of participants' prescriptions at home allowed recall bias to be minimized. The fact that the association between PAH and mortality was not statistically significant in multivariate analysis suggests that the factors included in the propensity score mainly explained the association observed in bivariate analysis. Analysis of the literature shows that the association between PAH and mortality is more likely to be found in studies including younger adults2-4 than in studies conducted only in older adults, which often fail to reach significance.5-9 Different patterns of prescriptions according to age could explain this difference; anxiolytics and hypnotics are prescribed to treat psychiatric disorders but also to overcome age-related problems (morbidity, social isolation, stressful life event) in older adult. This result also raises questions about the choice of vital status as the outcome. All-cause mortality may not be a specific-enough outcome to study the effect of PAH, especially in aging populations in which it is difficult to determine the effect of one risk factor among many others, including multiple comorbid conditions. Although PAH was associated with all-cause mortality in bivariate analysis, this association did not persist when considering potential confounders. This result should be considered with caution, and this relationship should be further explored in studies distinguishing different causes of death, especially deaths due to falls that are more likely to be related to PAH.