Assessment of GFR in Patients with Cancer: A Statement from the American Society of Onco-Nephrology

Abstract

Accurate assessment of GFR is crucial to guiding drug eligibility, dosing of systemic therapy, and minimizing the risks of both undertreatment and toxicity in patients with cancer. Up to 32% of patients with cancer have baseline CKD, and both malignancy and treatment may cause kidney injury and subsequent CKD. To date, there has been lack of guidance to standardize approaches to GFR estimation in the cancer population. In this two-part statement from the American Society of Onco-Nephrology, we present key messages for estimation of GFR in patients with cancer, including the choice of GFR estimating equation, use of race and body surface area adjustment, and anticancer drug dose-adjustment in the setting of CKD. These key messages are based on a systematic review of studies assessing GFR estimating equations using serum creatinine and cystatin C in patients with cancer, against a measured GFR comparator. The preponderance of current data involving validated GFR estimating equations involves the CKD Epidemiology Collaboration (CKD-EPI) equations, with 2508 patients in whom CKD-EPI using serum creatinine and cystatin C was assessed (eight studies) and 15,349 in whom CKD-EPI with serum creatinine was assessed (22 studies). The former may have improved performance metrics and be less susceptible to shortfalls of eGFR using serum creatinine alone. Since included studies were moderate quality or lower, the American Society of Onco-Nephrology Position Committee rated the certainty of evidence as low. Additional studies are needed to assess the accuracy of other validated eGFR equations in patients with cancer. Given the importance of accurate and timely eGFR assessment, we advocate for the use of validated GFR estimating equations incorporating both serum creatinine and cystatin C in patients with cancer. Measurement of GFR via exogenous filtration markers should be considered in patients with cancer for whom eGFR results in borderline eligibility for therapies or clinical trials.

Figure 1

Summary of GFR estimating equations assessed in studies identified by systematic literature review. ^aStudies reporting use of standardized serum creatinine assay and GFR measurement on the basis of plasma or urinary clearance are included in the above table (a full table of all identified studies is included in the Supplemental Table 5 and a list of literature search terms in included in Supplemental Appendix E). ^bColor-coding reflects GFR estimating equation inclusion for assessment within each study: CG: Cockcrocft-Gault (yellow), Jeliffe (orange); Wright (light brown); MDRD: Modification of Diet in Renal Disease equation (brown); CKD-EPI: CKD Epidemiology Collaboration–Cr: creatinine (light green); Cys: cystatin C (medium green); Cr-Cys: creatinine and cystatin C (dark green); Janowitz/Cambridge GFR version 1 (dark blue); Cambridge GFR version 1 (light blue). ^cSolid tumors. ^dMix of solid tumors and hematology cancers. For studies reporting P30, other reported metrics were omitted. For studies not reporting P30, we focused on mean or median percentage errors and mean or median absolute percentage errors. Mean percentage errors were calculated from the proportional differences between eGFR and measured GFR expressed as %. Negative and positive values reflect over and underestimate of the measured GFR, respectively. Mean absolute percentage error were calculated from the proportional differences between eGFR and measured GFR expressed as %, but do not incorporate positive or negative signs. MdnAD; median absolute difference calculated from the subtraction of eGFR-measured GFR expressed as ml/min per 1.73 m². Variations of reported metrics were described when available in the original paper as follows: 95% confidence interval (separated by to), interquartile range (separated by hyphen), and standard deviation (within parenthesis). P30 is the percentage of estimates that differed by more than 30% from the measured GFR. AUC, area under the curve; CamGFRv1, Cambridge GFR version 1; CamGFRv2, Cambridge GFR version 2; CG, Cockcrocft-Gault; CKD-EPI, CKD Epidemiology Collaboration; HSCT, hematopoietic stem cell transplantation; MAPE, mean absolute percentage error; MdnAD, median absolute difference; MDRD, modification of diet in renal disease; mGFR, measured GFR; MPE, mean percentage error; P, plasmatic clearance; U, urinary clearance; UN, unknown.

Figure 2

Recommendations for initial anticancer drug dose adjustment from the International Consensus Guideline for Anticancer Drug Dosing in Kidney Dysfunctiona. ^aReproduced with permission from ref. 55. ^bProposed dosing under specific circumstances (see Anticancer Drug Dosing in Kidney Dysfunction's Drug specific recommendations for individual drug details). AUC, area under the curve; eGFR, eGFR via the CKD Epidemiology Collaboration equation; KRT, kidney replacement therapy.