ASSESSING POTENTIALLY INAPPROPRIATE MEDICATION USE IN PATIENTS AT RISK FOR ADVERSE DRUG EFFECTS: SPIRONOLACTONE AS A CASE STUDY
Secora, Alex Michael
MetadataShow full item record
Background: Polypharmacy and potentially inappropriate medication (PIM) use are common and associated with considerable morbidity, yet they are often modifiable risk factors. However, some PIM use is a result of limited information on medication safety across patient kidney function. One such medication, spironolactone, an aldosterone antagonist indicated for heart failure, has been demonstrated in clinical trials to reduce morbidity and mortality among individuals with normal renal function, but its safety in those with chronic kidney disease (CKD) is unclear. Methods: We used longitudinal data from the Atherosclerosis Risk in Communities (ARIC) study to quantify PIM use by estimated glomerular filtration rate (eGFR), and to assess the relationship between polypharmacy, PIM use, and subsequent hospitalization and death in older adults. We used commercial claims data from MarketScan and electronic health record data from the Geisinger Health System to identify predictors of spironolactone initiation among patients with heart failure, and used target trial emulation to characterize the risk of hyperkalemia and acute kidney injury (AKI) with spironolactone use among patients using loop diuretics. Results: Participants in ARIC (N=6,392) with CKD reported more medications than those without CKD (p<0.001), and PIM use based on kidney function was prevalent (36%) among those with eGFR <30 ml/min/1.73m2. More concurrent medications were associated with higher risks of hospitalization and death, but PIM use was not, and there were no differences in the relative risks associated with greater numbers of medications by CKD status. Among patients with incident heart failure in MarketScan (N=22,956) and Geisinger (N=16,547), 7.0% and 9.9% initiated spironolactone within two years, respectively. Patients with eGFR <30 were least likely to initiate spironolactone compared to patients with eGFR 60-89 (meta-analyzed hazard ratio [HR]: 0.61, 95% confidence interval [CI]: 0.44-0.83). In Geisinger patients with heart failure using loop diuretics (N=17,110), spironolactone initiation was associated with increases in hyperkalemia and AKI risk compared to use of loop diuretics alone (HR 1.69 [CI: 1.35-2.10], and HR 1.12 [CI: 1.00-1.26], respectively), with no observed differences in the relative risk of either outcome associated with spironolactone by eGFR. Conclusions: Polypharmacy and PIM use were common, with greater numbers of medications associated with greater risk of hospitalization or death. Spironolactone initiation was uncommon within two years of heart failure diagnosis, and least likely among patients with lower kidney function. The addition of spironolactone to loop diuretics increased the risk of hyperkalemia, and more modestly, AKI. Improved data on medication safety in patients with CKD are needed.