Background Benzodiazepine use can potentially cause confusion and delays in mental processes. These well-known side effects appear to be linked to an increased risk of being diagnosed with dementia. Objective To evaluate the possibility of an association between benzodiazepine and dementia. Setting Korean healthcare database from 2002 to 2013. Methods Sequence symmetry analysis was conducted to investigate whether benzodiazepine use increases the risk of dementia or not. We defined exposure as new benzodiazepine users and outcome as new diagnosis of dementia (ICD-10: F00-03, G30, and G318). Benzodiazepines were categorized into two groups (long-acting and short-acting) based on the duration of action. Antidepressants, opioid analgesic, and statin were used as active comparators to rule out any possible non-causal interpretations of our results. The time-trend adjusted sequence ratio (ASR) with 95% confidence intervals (CI) was measured to identify possible associations. Main outcome measure Adjusted sequence ratio. Results Benzodiazepine users were shown to be associated with dementia [benzodiazepine: 4212 pairs, ASR = 2.27 (95% CI 2.11-2.44)]. In addition, long-acting benzodiazepines had a higher ASR than that of short-acting benzodiazepines [long-acting: 3972 pairs, ASR = 2.22 (95% CI 2.06-2.39] and [short-acting: 5213 pairs, ASR = 1.88 (95% CI 1.77-2.00)]. However, our SSA found no duration-response relationship. Conclusion Our signal detection suggests that there is a possible association between benzodiazepines and dementia. Additionally, it proposes that persons receiving long-acting benzodiazepines are at a higher risk of developing dementia than those receiving short-acting benzodiazepines. Further studies are recommended to confirm whether this epidemiological association is a causal effect or not.
Keywords: Benzodiazepines; Data-mining; Dementia; Sequence symmetry analysis (SSA); South Korean health care database.