Higher-dose gabapentinoids and the risk of adverse events in older adults with CKD: a population-based cohort study

Rationale & Objective — Gabapentinoids are an opioid substitute whose elimination by the kidneys is reduced as kidney function declines. To inform their safe prescribing in older adults with chronic kidney disease (CKD), we examined the 30-day risk of serious adverse events according to the prescribed starting dose.

Study Design — Population-based cohort study.

Setting & Participants — 74,084 older adults (64% women; median age 79 [interquartile range 73-85]) with CKD (defined as an estimated glomerular filtration rate [eGFR] <60 mL/min/1.73m2 while not receiving dialysis) and a newly prescribed a gabapentinoid between 2008-2020 in Ontario, Canada.

Exposure — Higher-dose gabapentinoids (gabapentin >300 mg/day or pregabalin >75 mg/day) vs lower-dose gabapentinoids (gabapentin ≤300 mg/day or pregabalin ≤75 mg/day).

Outcomes — The primary composite outcome was the 30-day risk of a hospital visit with encephalopathy, a fall, a fracture, or a hospitalization with respiratory depression.

Analytical Approach — Comparison groups were balanced on indicators of baseline health using inverse probability of treatment weighting using propensity score analysis that generated a pseudo-sample for the reference group with a distribution of measured covariates similar to the exposed group. Weighted risk ratios (RR) were estimated using modified Poisson regression, and weighted risk differences (RD) estimated using binomial regression. Pre-specified subgroup analyses were conducted by eGFR category and type of gabapentinoid.

Results — Among 74,084 patients were identified with CKD and a new prescription for gabapentin or pregabalin 41% started at >300 mg/day and >75 mg/day, respectively. From this set of patients a weighted study population with a size of 61,367 was generated. Patients who started at a higher dose had a 30-day risk of the primary outcome higher than the risk among with patients who started at lower dose. Within the weighted population, the numbers of events for higher vs. lower dose were 585/30,660 [1.9%] vs 462/30,707 [1.5%], respectively. The weighted RR was 1.27 [95% CI, 1.13 to 1.42] and the weighted RD was 0.40% [95% CI, 0.21% to 0.60%]). In subgroup analyses, neither multiplicative nor additive interactions were statistically significant.

Limitations — Residual confounding.

Conclusions — In this population-based study, starting a gabapentinoid at a higher vs lower dose was associated with a slightly higher risk of a hospital stay with encephalopathy, a fall, or a fracture, or hospitalization with respiratory depression. If verified, these risks should be balanced against the benefits of using a higher-dose gabapentinoid.