Real-world renal function among patients with multiple myeloma in the United States

Background
 Multiple myeloma (MM) is the second most common hematologic malignancy and is associated with significant patient burden. Renal impairment has been shown to affect up to 50% of patients (pts) with MM. Renal impairment is an independent predictor of poor survival outcomes for pts with MM, with a median survival of approximately half that of pts without renal impairment.
 Aims
 To assess change in renal function by treatment line and drug class among pts with MM and renal impairment (defined as eGFR <50 mL/min/1.73 m2 using the Modification of Diet in Renal Disease equation [eGFR-MDRD]), and to assess real-world outcomes by baseline renal status, renal response, and drug class.
 Methods
 Using the nationwide Flatiron Health EHR-derived de-identified database based in the United States, we identified MM pts diagnosed and treated between January 2011 and November 2019 who received ≥1 line of MM therapy and had information on race. We assessed the distribution of pts by eGFR-MDRD level at the start of the first and second line of therapy and the distribution of therapy use (ie, proteasome inhibitor [PI], immunomodulatory drug [IMiD], monoclonal antibody [mAb]). We also evaluated overall survival (OS) by treatment line, stratified by eGFR at the start of each treatment line. Complete renal response (CRR) was assessed in pts with eGFR-MDRD <50 mL/min/1.73 m2; using International Myeloma Working Group recommendations, responders were defined as pts with ≥1 eGFR measurement ≥60 mL/min/1.73 m2 during the treatment line. Logistic regression models were used to examine the association between treatment class and complete renal responder status adjusted for other treatment classes received, age, sex, race, practice type, year of therapy line, and cytogenetic risk. Cox proportional hazard models were used to examine OS by treatment and renal response, adjusted for other treatment classes received, age, sex, race, practice type, year of therapy line, and cytogenetic risk. Flatiron Health, Inc., did not participate in the analysis of this data.
 Results
 For the 6990 pts included in the analysis, the mean age at start of initial therapy was 68 years, 46% were female, and 17% were Black. At start of treatment lines 1 and 2, approximately 25% had eGFR-MDRD <50 mL/min/1.73 m2. At treatment initiation, pts with renal impairment were older (71 vs 67 years) and were more likely to present with ISS stage III at diagnosis (38% vs 10%). Pts with renal impairment at the start of each treatment line exhibited decreased OS. Among pts with renal impairment, pts with PI use in first (adjusted odds ratio [aOR] 1.36; 95% CI, 1.04-1.77) or second line (aOR 2.09; 95% CI, 1.38-3.16) were significantly more likely to have a CRR than those without PI use. Pts with IMiD use in first (aOR 2.14; 95% CI, 1.68-2.72) and second (aOR 1.61; 95% CI, 1.10-2.36) line were significantly more likely to have a CRR than those without IMiD use. When classified by both PI and IMiD use, pts with both PI and IMiD use were significantly more likely to have a CRR than those without use of either in the first (aOR 2.35; 95% CI, 1.54-3.60) and second line (aOR 3.89; 95% CI, 1.71-8.86). Pts with PI and IMiD use as well as a CRR also had greater OS in the first (adjusted hazard ratio [aHR] 0.52; 95% CI, 0.37-0.73) and second line (aHR 0.53; 95% CI, 0.32-0.88) vs pts without either PI or IMiD use and no CRR. The use of available mAbs in line 1 or line 2 was not significantly associated with renal response. Lower mAb use, especially in earlier treatment lines, prevented further analyses by mAb combination therapies.
 Conclusion
 In this study, MM pts with renal impairment were more likely to be older and to present with ISS stage III at diagnosis. Pts with decreased renal function exhibited decreased OS compared with non-renally impaired pts. Pts who were treated with combination therapy that included PIs and IMiDs together in early treatment lines were more likely to have a CRR and OS was prolonged among these pts, highlighting the benefits of combination therapy. These data suggest that treatment inducing a renal response may result in improved outcomes. Future investigations with larger datasets may improve the understanding of the prognostic value of renal impairment and optimal combination treatment regimens for these pts.
 
 
 
 Mikhael: Amgen, Celgene, GSK, Janssen, Karyopharm, Sanofi, Takeda: Honoraria. Singh:Sanofi-Genzyme: Current Employment. Rice:Sanofi: Current Employment.


eGFR-MDRD was calculated from creatinine lab values using the following equation: 175 × (creatinine mg/ dL) −1.154 × (age) −0.203 × (0.742 if female) × (1.212 if Black/ African American) [8][9][10] . The distribution of patients was assessed by eGFR-MDRD level (<50 and ≥50 mL/min/ 1.73 m 2 ) at the start of first-(1 L) and second-line (2 L) therapy. Overall survival (OS) was evaluated by treatment line, stratified by eGFR-MDRD level at the start of the treatment line. Renal response was assessed in patients with eGFR-MDRD < 50 mL/min/1.73 m 2 at the start of treatment, who had ≥1 eGFR-MDRD measurement during the treatment line; using International Myeloma Working Group recommendations, patients with complete renal response (CRR) were defined as patients with ≥1 eGFR-MDRD measurement ≥60 mL/min/1.73 m 2 during the treatment line.
Logistic regression models were used to examine the association between treatment class and CRR status. These models were adjusted for the following clinically relevant variables, as they were potential confounders of the examined associations: other treatment classes received, age, sex, race, practice type, year of therapy line, and cytogenetic risk. Kaplan-Meier analyses and Cox proportional hazard models were used to examine OS from the start of 1 L therapy and 2 L therapy by baseline renal status and by both treatment class and renal response. Cox proportional hazard models were adjusted, as described above. The data that support the findings of this study have been originated by Flatiron Health, Inc. These de-identified data may be made available upon request, and are subject to a license agreement with Flatiron Health; interested researchers should contact DataAccess@flatiron.com to determine licensing terms. Flatiron Health, Inc, did not participate in the analysis of this data. Patient characteristics overall and by eGFR-MDRD level (<50 and ≥50 mL/min/1.73 m 2 ) at the start of 1 L therapy are shown in Supplemental Table 1. The mean age at start of initial therapy was 68.3 years, 45.8% of patients were female, and 17.2% were Black/African American. One quarter (n = 1722) of patients at the start of initial therapy had eGFR-MDRD < 50 mL/min/1.73 m 2 (including 10.5% with eGFR-MDRD < 30 mL/min/1.73 m 2 ), approximately half (n = 3465) had eGFR-MDRD ≥ 50 mL/min/1.73 m 2 , and the remainder had unknown levels. On average, patients with eGFR-MDRD < 50 mL/min/1.73 m 2 at the start of initial therapy were older (71.3 vs 67.3 years) and more likely to be diagnosed at stage III (37.7% vs 10.3%) compared with patients with eGFR-MDRD ≥ 50 mL/min/1.73 m 2 .
Among all patients with eGFR-MDRD < 50 mL/min/ 1.73 m 2 at the start of the treatment line and ≥1 eGFR-MDRD measurement during the treatment line (N = 1682 in 1 L and 920 in 2 L), 38.5% of patients achieved a CRR during 1 L therapy, but only 19% achieved a CRR during 2 L therapy. The majority of these patients with renal impairment had 1 L and 2 L use of PI (77% and 64%, respectively). These patients were significantly more likely to have a CRR compared with those without PI use in both 1 L (adjusted odds ratio [aOR], 1.36; 95% CI, 1.04-1.77) and 2 L therapy (aOR, 2.09; 95% CI, 1.38-3.16) in multivariable-adjusted models.
Approximately half of patients with renal impairment had IMiD use in 1 L and 2 L therapy. In both treatment lines, patients with IMiD use were significantly more likely to have a CRR compared with those without IMiD use in multivariable models, particularly in 1 L therapy (aOR, 2.14; 95% CI, 1.68-2.72).
When we examined PI and IMiD use in combination together among patients with renal impairment, 36% of patients had combination PI and IMiD treatment, 42% used PIs but not IMiDs, 14% used IMiDs but not PIs, and 8% used neither in 1 L therapy. In 2 L, 26% of patients had combination PI and IMiD treatment, 38% used PIs but not IMiDs, 26% used IMiDs but not PIs, and 10% used neither. In 1 L and 2 L therapy, patients with both PI and IMiD use were significantly more likely to have a CRR compared with those without use of either treatment class (1 L aOR, 2.35 [95% CI, 1.54-3.60]; 2 L aOR, 3.89 [95% CI, 1.71-8.86]). In 2 L therapy, patients with PI use but no IMiD use also were significantly more likely to have a CRR compared with those without use of either (aOR, 2.54; 95% CI, 1.15-5.59; Fig. 1). Patients with eGFR-MDRD < 30 mL/min/1.73 m 2 with both PI and IMiD use were significantly more likely to have a CRR compared with those without the use of either treatment class in 1 L (aOR, 5.46; 95% CI, 1.77-16.8), but there was no association in 2 L (aOR, 0.50; 95% CI, 0.06-4.17).
Low mAb use in early treatment lines during the database time frame (through November 2019) prevented most analyses of this treatment class. Few patients (<1%) were treated with mAb frontline and~10% had mAb use as 2 L therapy. We did not observe a significant difference in CRR by mAb use in 2 L (22% vs 19%; aOR, 1.51; 95% CI, 0.84-2.74). Due to small sample sizes, we were unable to examine mAb combination therapies.
Patients receiving combination PI and IMiD treatment who had a CRR had significantly better OS compared with patients without use of either treatment, who did not have a CRR in both frontline (aHR, 0.52; 95% CI, 0.37-0.73) and second line (aHR, 0.53; 95% CI, 0.32-0.88; Supplemental Table 3). This relationship was not observed for patients with eGFR-MDRD < 30 mL/min/1.73 m 2 ; however, due to the relatively smaller sample size, we cannot make definitive conclusions for this subgroup. Logistic regression models were used to examine the association between treatment class (i.e., PI and IMiD) and CRR status adjusted for other treatment classes received, age, sex, race, practice type, year of therapy line, and cytogenetic risk. *P < 0.05, **P < 0.01, ***P < 0.001 vs no PI or IMiD use from multivariable models. Multivariable models adjusted for other treatment classes received, age, sex, race, practice type, year of therapy line, and cytogenetic risk.
In this study, MM patients with renal impairment were more likely to be older and to present with ISS stage III at diagnosis, which has been reported in a previous phase III study 11 . Patients with renal impairment exhibited decreased OS compared with non-renally impaired patients. Patients treated with combination therapy that included PIs and IMiDs together in early treatment lines were more likely to have a CRR, supporting the known benefits of triplet vs doublet therapy. In addition, OS was prolonged among patients with PI and IMiD combination therapy who achieved a CRR, highlighting the benefits of both combination therapy and renal response. Moreover, these data underscore the importance of providing patients the best therapy for their MM.
This study has some limitations. We were unable to discern whether renal impairment was due to MM or other causes. Though we adjusted for several baseline characteristics, including age and practice type, we may not have fully accounted for factors influencing receipt of triplet regimens. In addition, as few patients received mAbs or had very low renal function (eGFR-MDRD < 30 mL/min/1.73 m 2 ) in early treatment lines, power for analyses in these subgroups was limited. The primary strengths of this study are the use of repeated eGFR-MDRD assessments and adjustment for several potential confounders. MM patients with renal insufficiency have been shown to have poorer outcomes and continue to have unmet medical need. Overall, these data suggest that using combination therapy early with the goal of inducing a renal response in these patients is not only feasible, but also may result in improved outcomes. Future investigations with larger datasets that include newer agents may improve the understanding of the optimal combination treatment regimens for these patients.