Medication-Related Adverse Events and Discordancies in Cystatin C–Based vs Serum Creatinine–Based Estimated Glomerular Filtration Rate in Patients With Cancer

IMPORTANCE: Serum creatinine–based estimated glomerular filtration rate (eGFRcr) may overestimate the glomerular filtration rate (GFR) in patients with cancer. Cystatin C–based eGFR (eGFRcys) is an alternative marker of GFR. OBJECTIVE: To determine whether the therapeutic drug levels and adverse events (AEs) associated with renally cleared medications were higher in patients with cancer whose eGFRcys was more than 30% lower than their eGFRcr. DESIGN, SETTING, AND PARTICIPANTS: This cohort study analyzed adult patients with cancer at 2 major academic cancer centers in Boston, Massachusetts. These patients had their creatinine and cystatin C measured on the same day between May 2010 and January 2022. The date of the first simultaneous eGFRcr and eGFRcys measurement was considered to be the baseline date. EXPOSURE: The primary exposure was eGFR discordance, defined as an eGFRcys that was more than 30% lower than the eGFRcr. MAIN OUTCOMES AND MEASURES: The primary outcome was risk of the following medication-related AEs within 90 days of the baseline date: (1) supratherapeutic vancomycin trough level greater than 30 μg/mL, (2) trimethoprim-sulfamethoxazole–related hyperkalemia (>5.5 mEq/L), (3) baclofen toxic effect, and (4) supratherapeutic digoxin level (>2.0 ng/mL). For the secondary outcome, a multivariable Cox proportional hazards regression model was used to compare 30-day survival of those with vs without eGFR discordance. RESULTS: A total of 1869 adult patients with cancer (mean [SD] age, 66 [14] years; 948 males [51%]) had simultaneous eGFRcys and eGFRcr measurement. There were 543 patients (29%) with an eGFRcys that was more than 30% lower than their eGFRcr. Patients with an eGFRcys that was more than 30% lower than their eGFRcr were more likely to experience medication-related AEs compared with patients with concordant eGFRs (defined as eGFRcys within 30% of eGFRcr), including vancomycin levels greater than 30 μg/mL (43 of 179 [24%] vs 7 of 77 [9%]; P = .01), trimethoprim-sulfamethoxazole–related hyperkalemia (29 of 129 [22%] vs 11 of 92 [12%]; P = .07), baclofen toxic effects (5 of 19 [26%] vs 0 of 11; P = .19), and supratherapeutic digoxin levels (7 of 24 [29%] vs 0 of 10; P = .08). The adjusted odds ratio for vancomycin levels more than 30 μg/mL was 2.59 (95% CI, 1.08-7.03; P = .04). Patients with an eGFRcys more than 30% lower than their eGFRcr had an increased 30-day mortality (adjusted hazard ratio, 1.98; 95% CI, 1.26-3.11; P = .003). CONCLUSIONS AND RELEVANCE: Results of this study suggest that among patients with cancer with simultaneous assessment of eGFRcys and eGFRcr, supratherapeutic drug levels and medication-related AEs occurred more commonly in those with an eGFRcys more than 30% lower than their eGFRcr. Future prospective studies are needed to improve and personalize GFR estimation and medication dosing in patients with cancer.