Prevalence and Predictors of Renal Impairment Among Patients with Multiple Myeloma (MM): An Analysis of Oncology Clinic Electronic Health Records Linked to Commercial Claims in the United States

Background

Although incurable, MM is responsive to treatment, and the prognosis of MM patients has improved dramatically over the past two decades. This has led to increased recognition of the importance of effective long-term management of the classic clinicopathological features of MM. Renal impairment (RI) is a defining hallmark of active myeloma. While it is known that RI is associated with poor survival and can complicate drug dosing and limit treatment options, patterns of renal function and risk factors of RI in MM are poorly characterized. This study seeks to characterize the natural history and risk factors of RI in MM using a unique real-world dataset that links patient-level electronic health records (EHR) to healthcare insurance claims.

Methods

The Oncology Services Comprehensive Electronic Records (OSCER) database contains EHR data from community and hospital-affiliated oncology clinics in the U.S. The MarketScan (MS) database contains healthcare claims information from large employers, managed care organizations, Medicare, and Medicaid programs. This study used a subset of MM patients with overlapping data from OSCER and MS, allowing the analysis of EHR-based serum creatinine and claims-based comorbidity data in the same patient.

Patients ≥ 18 with a MM diagnosis between 2011 and February 28, 2016 in an OSCER clinic; had at least 6 months of continuous MS benefits coverage; no MM diagnosis in MS in the 3 months prior to OSCER MM diagnosis; and no evidence of another malignancy prior to MM diagnosis were included in the study. Index date was the date of OSCER MM diagnosis and patients were followed up until February 28, 2017. Baseline (BL) patient characteristics and potential risk factors of RI, including demographics, BMI, ECOG, ISS stage, and comorbidities (e.g., hypertension, diabetes, hypercalcemia, etc.) were assessed in the 6 months prior to index. Intravenous bisphosphonate (IV BP) use was assessed during follow-up. Renal function was assessed using the MDRD eGFR method or CKD stage diagnosis codes, with RI defined in this study as eGFR < 60 or stage 3+ CKD. BL eGFR was assessed using the serum creatinine measurement closest to the index in the interval 1 month before and 1 month after index.

The prevalence of RI in the year following diagnosis was calculated based on the average number of at-risk patients during this period. Change in renal function for up to 4 years post-diagnosis was assessed by BL renal function and was characterized based on the worst recorded estimate of renal function during this period. A multivariate Cox model was used to assess the impact of BL risk factors and IV BP use during follow-up on progression to RI in patients without RI at BL (eGFR ≥ 60 and CKD diagnosis). Follow-up IV BP use (dose count) was treated as a time-dependent variable.

Results

The median number of eGFR values recorded per patient during follow-up was 18 (range: 7-36). Among newly diagnosed MM patients (n = 625), 69% (95% CI: 65-73%) experienced ≥ 1 episodes of eGFR < 60 ml/min or stage 3+ CKD diagnosis (RI) within 1 year of diagnosis, with 16% (13-20%) experiencing an episode of eGFR 15-29 or stage 4 CKD and 15% (11-18%) experiencing an episode of eGFR < 15 or stage 5 CKD.

Among patients without RI at diagnosis (n = 281), 37% (31-43%) subsequently experienced at least one episode of RI during follow-up (Figure 1). In addition, 26% (20-32%) of patients with BL eGFR 30-59 or stage 3 CKD experienced progression to an episode of eGFR < 30 or stage 4+ CKD, and 31% (20-42%) of patients with BL eGFR 15-29 or stage 4 CKD progressed to an episode of eGFR < 15 or stage 5 CKD.

In our multivariate Cox model, use of the IV BP zoledronic acid (ZA) during follow-up; ECOG; baseline eGFR; known predisposing conditions, including autoimmune diseases, vascular conditions, and infections of the kidney were found to be associated with higher risk for experiencing an episode of RI in patients with high BL renal function (Figure 2). In our sample, 12 doses of ZA was associated with a 2.4-fold (1.2-4.8) increase in risk of RI. A baseline ECOG of 1 vs. 0 was also associated with a 2.4-fold (1.1-5.2) increase in risk of RI.

Conclusions

Using a unique EHR-insurance claims linked data source, we found that two-thirds of patients experienced at least one episode of renal impairment within the first year of MM diagnosis. Repeated exposure to zoledronic acid and poor performance status may increase the risk of RI in patients without renal impairment at diagnosis.

View largeDownload slide

View largeDownload slide

Close modal