Gabepentinoids and Benzodiazepines in Medicare Part D

I have read with interest the observational study regarding risks of adverse outcomes associated with gapapentin and pregabalin use.¹ The authors adeptly analyzed data in the US Renal Data System database, including Medicare Part D prescription drug events (i.e., claims), and identified positive associations of gabapentin and pregabalin exposure with adjusted hazards of first episodes of altered mental status, fall, and fracture requiring either an emergency room visit or hospitalization.

The study suffers from one unappreciated limitation. The authors attempted to adjust for the influence of concomitant use of benzodiazepines. Evidently, <0.5% of all patients in the study cohort used benzodiazepines. Such low utilization is clinically implausible but can be explained by the design of the Medicare Part D benefit during the study era (i.e., in 2011). The Medicare Modernization Act excluded benzodiazepines from Part D coverage between 2006 and 2012.² This exclusion was eliminated by the Patient Protection and Affordable Care Act. The fact that the authors identified any use of benzodiazepines during the study era reflects the provision of “enhanced alternative coverage,” which generally offers a higher monthly premium in exchange for added value (e.g., reduced deductible, coverage of drugs not ordinarily included in the Part D benefit, reduced cost sharing in the coverage gap, etc.).

Benzodiazepine toxicity is positively associated with risks of altered mental status, fall, and fracture.³ Thus, because of inherent limitations of data ascertained from Medicare Part D, unmeasured confounding by benzodiazepine use cannot be discounted. Because of study design, unmeasured confounding by opioid use also cannot be discounted. Ultimately, polypharmacy is common among patients on dialysis,⁴ and therefore, observational studies of the efficacy and safety of single drugs or single classes of drugs should be interpreted cautiously. Gabapentinoid toxicity is an observable event, but the magnitude of the associations reported by Ishida et al.¹ may be biased.

Disclosures

E.D.W. is also an employee of NxStage Medical but reports no conflict of interest with this content.