Comparative short-term safety and efficacy of hypnotics: A quantitative risk-benefit analysis

Several professional societies have provided recommendations for prescribing medications for insomnia. None has provided an integrative analysis that concurrently quantifies safety and efficacy (e.g., risk-benefit ratios). This represents an important gap for informing clinician decision-making. Accordingly, the aim of the present review is to provide such an analysis for five classes of sleep-promoting medications. Adverse event data values were extracted from the most recent FDA-approved package inserts and converted to an integer before being placebo-adjusted and standardized as a rate per 1000 (AEr). Efficacy data, pre-to-post self-reported data for active and placebo conditions were acquired from pivotal trials identified in "white papers" and systematic reviews/meta-analyses. Weighted effect sizes were calculated for subjective sleep latency, wake time after sleep onset and total sleep time, and then were averaged by medication class for each sleep continuity variable. Overall efficacy was represented by a single variable, SWT (sleep latency + wake time after sleep onset + total sleep time). Risk-benefit was represented using a simple ratio value. For safety, it was found that melatonin receptor agonists had the lowest adverse event rate (AEr = 43.1), and non-benzodiazepine benzodiazepine receptor agonists had the highest rate (AEr = 255.0). For efficacy, it was found that the pre-to-post placebo adjusted effect sizes were largest for benzodiazepines (effect size = 1.94) and smallest for melatonin receptor agonists (effect size = 0.109). For risk-benefit, histamine antagonist had the most favourable profile (risk-benefit = 69.5), while melatonin receptor agonist had the least favourable profile (risk-benefit = 395.7). Overall, the combined metric for risk-benefit suggests that treatment with a histamine antagonist is optimal and potentially represents the best first-line therapy for the medical management of insomnia.