Time‐dependent risk of falls associated with hypnotic use in hospitalized patients

AIM: Falls are a common and serious problem among hospitalized patients. While hypnotics are a known risk factor, it remains unclear whether fall risk varies by time of day. This study examined the time-dependent risk of falls associated with different classes of hypnotics. METHODS: We conducted a retrospective cohort study of patients aged ≥20 years who had their first admission to our hospital and were discharged between April 2016 and March 2018. Time-to-fall was analyzed using stepwise Cox proportional hazards models across four time periods (0:00-5:59, 6:00-11:59, 12:00-17:59, 18:00-23:59). Covariates included hypnotic use (Z-drugs, short-acting benzodiazepines, long-acting benzodiazepines, ramelteon, suvorexant, trazodone), age, sex, body mass index (BMI), length of stay, activities of daily living score, emergency admission, and surgery. RESULTS: Among the 22,458 patients included in the study, 268 falls occurred, with 73, 65, 56, and 74 in each time period, respectively. Association between hypnotic use and fall risk varied by time. Z-drugs were associated with falls during 18:00-23:59 (adjusted hazard ratio [aHR] 2.65, 95% confidence interval [CI] 1.29-5.45). Short-acting benzodiazepines were associated with falls during 00:00-05:59 (aHR 2.06, 95% CI 1.07-3.97). Long-acting benzodiazepines were associated with falls during 06:00-11:59 (aHR 3.37, 95% CI 1.20-9.26). No significant associations were observed for ramelteon, suvorexant, or trazodone. CONCLUSIONS: Fall risk associated with hypnotic use is not uniform throughout the day. Short-acting benzodiazepines were linked to evening and nighttime falls, while long-acting benzodiazepines increased morning fall risk. These findings highlighted the need to consider time-specific fall risks when prescribing hypnotics.