Benzodiazepines and Z-drugs in Sweden : pharmacoepidemiological studies on patterns of use and risks

<p dir="ltr">Benzodiazepines and related Z-drugs (BZDR) are effective for the short-term treatment of symptoms like anxiety and insomnia, but their long-term use is discouraged by clinical guidelines. Despite the concerns about adverse effects of BZDR treatment, there existed limited knowledge specifically on associations between BZDR treatment and risks of subsequent infections and alcohol- and drug-related problems. This thesis, therefore, leveraged the unique collection of nationwide Swedish registers to examine the predictors of BZDR prescription patterns and assess the associations of BZDR treatment with risks of serious infections and substance-related problems.</p><p dir="ltr">Study I assessed BZDR use patterns among 930,465 new users in 2007-2013. We found that the proportion of individuals progressing to long-term BZDR users (>6 months) was substantial and increased with age, from 21% in those who were 0-17 years at BZDR initiation to 57% in individuals aged 265 years. Further, we identified four distinct five-year BZDR use trajectories, including the 'discontinued' trajectory that showed eventual cessation of BZDR use from the third year after the treatment start, and the 'decreasing', 'slow decreasing' and 'maintained' trajectories that indicated continued BZDR use up to five years from the initiation. More BZDR recipients belonged to the 'discontinued' trajectory compared to other trajectories, but the proportions of discontinued users reduced with age. Across all age groups, initiating with multiple BZDRs and having the concurrent use of other medicines were linked with higher risks of subsequent long-term BZDR use (compared with short-term use) and belonging to the 'decreasing', 'slow decreasing' and 'maintained' trajectory memberships (compared with developing the 'discontinued' trajectory). The study highlighted the importance of providing support to prescribers in making evidenced-based decisions when starting and managing BZDR treatment for patients of all ages.</p><p dir="ltr">Study II assessed the association between BZDR treatment and the risk of serious infections among individuals under 65 years. Infection outcomes were defined by the diagnosis from in- and/or out-patient care or underlying cause of death. We constructed three cohorts, including a demographically matched cohort of 713,896 pairs of incident BZDR users and nonusers matched on sex, birth year and month, and county of residence, a co-twin control cohort of 9197 BZDR users and 9298 unexposed co-twins (co-multiples), and an active comparator cohort of 434,900 BZDR users and 428,074 selective serotonin reuptake inhibitor users. In the demographically matched cohort, BZDR initiation was associated with an 83% increase in the risk of any serious infection. This association persisted, although attenuated, in the co-twin control cohort and active comparator cohort. Further, we found a dose-response relationship between BZDR cumulative dosage and an elevated risk of infections. Future research is called for examining the potential mechanisms underlying the observed association of BZDR treatment and increased risks of serious infections, given such insights may inform safer treatment strategies.</p><p dir="ltr">Study III assessed the long-term risk of substance-related problems following BZDR initiation in Swedish residents who were aged 10 or older and had no prior BZDR prescriptions 2007. The outcomes included incident alcohol-related problems and drug-related problems, defined by alcohol and drug use disorders, poisoning, suspected criminal offenses, and related deaths. The demographically matched cohort consisted of 960,430 pairs of incident BZDR users and nonusers matched on sex, year and month of birth and residence county, and the co-twin control cohort included 12048 BZDR exposed twins and 12,579 unexposed co-twins (co-multiples). Over a 14-year follow-up, BZDR treatment initiation was linked to a slight yet consistent increase in the risks of alcohol-related and drug-related problems. These associations persisted in the co-twin control comparison and all sensitivity analyses.</p><p dir="ltr">Study IV subsequently examined the concurrent risk of substance-related problems using a within-individual design. We compared risks of acute alcohol- or drug-related problems during the period on active BZDR treatment to that during the period off BZDR treatment, among individuals born before 1997 who received the first BZDR prescription during 2007-2020. Periods on BZDR treatment were associated with a 9% increase in the risk of alcohol-related problems and a 16% increase in the risk of drug-related problems, compared to periods off BZDR treatment. Additionally, the analyses for specific subgroups of outcomes showed the strongest association for unintentional drug overdose. When we considered timing of these associations by further splitting the on-treatment and off-treatment periods, the observed increased risks of alcohol- and drug-related problems persisted within the first 91 days and beyond 91 days after BZDR treatment start. The findings suggest substance-related risks in patients undergoing BZDR treatment should be monitored.</p><p dir="ltr">In conclusion, the thesis has explored long-term use of prescribed BZDR and the dynamic use trajectories across different age groups, assessed the associations between BZDR treatment and subsequent infection risks, and examined the relationship between BZDR treatment and risks of substance-related problems. The findings may prompt future research and inform clinical practice on BZDR prescribing.</p><h3>List of scientific papers</h3><p dir="ltr">I. Isomura, K.*, <b>Wang, X.</b>*, Chang, Z., Hellner, C., Hasselström, J., Ekheden, I., ... & Sidorchuk, A. (2023). Factors associated with long-term benzodiazepine and Z-drug use across the lifespan and 5-year temporal trajectories among incident users: a Swedish nationwide register-based study. European Journal of Clinical Pharmacology, 79(8), 1091-1105. <a href="https://doi.org/10.1007/s00228-023-03515-2" rel="noreferrer" target="_blank">https://doi.org/10.1007/s00228-023-03515-2</a></p><p dir="ltr">II. <b>Wang, X.</b>, Isomura, K., Lichtenstein, P., Kuja-Halkola, R., D'Onofrio, B. M., Brikell, I., ... & Sidorchuk, A. (2024). Incident benzodiazepine and Z-Drug use and subsequent risk of serious infections. CNS drugs, 38(10), 827-838. <a href="https://doi.org/10.1007/s40263-024-01108-w" rel="noreferrer" target="_blank">https://doi.org/10.1007/s40263-024-01108-w</a></p><p dir="ltr">III. <b>Wang, X.</b>, Chang, Z., Molero, Y., Isomura, K., de la Cruz, L. F., Lichtenstein, P., ... & Sidorchuk, A. Incident benzodiazepine and Z- drug use and subsequent risk of alcohol-and drug-related problems: a nationwide matched cohort study with co-twin comparison. Journal of Psychopharmacology. [Accepted]</p><p dir="ltr">IV. <b>Wang, X.</b>, Molero, Y., Chang, Z., Isomura, K., Lichtenstein, P., Larsson, H., ... & Sidorchuk, A. Association of benzodiazepine and Z-drug use with alcohol- and drug-related problems: a within-individual comparison. [Manuscript]</p><p dir="ltr">*Joint first authorship.</p>