Prognostic impact of renal failure recovery in patients with newly diagnosed multiple myeloma

Background: Renal failure (RF) is a common complication in patients with newly diagnosed multiple myeloma (NDMM). Aim: To evaluate the frequency of RF in NDMM patients, and the prognostic impact of its reversibility. Material and Methods: A retrospective study evaluating demographic and clinical characteristics of 154 consecutive patients with NDMM was carried out. Estimated glomerular filtration rate (eGFR) was calculated at the beginning and at the end of the induction therapy. In addition, we evaluated renal responses (RR) according to the International Myeloma Working Group (IMWG) criteria. The induction regimen was based on thalidomide in all cases. Results: Fifty-three patients had RF (34.4%). Complete renal response (RR) was achieved in 51%. Three years overall survival in patients without RF, with RF and complete RR, and patients with RF and any other RR, was 66, 47 and 13%, respectively. Median survival was 53, 27 and 6 months, respectively (p < 0.01). In the multivariate analysis, RF and hypercalcemia were independent predictors of a worse outcome. Conclusions: Achieving a complete RR in patients with NDMM, is associated with a better survival. (Rev Med Chile 2019; 147: 1374-1381)

R enal failure (RF) is a well-known complication in patients with multiple myeloma (MM). RF occurs in 20% to 50% of patients at diagnosis 1,2 . Approximately, 10% will require renal replacement therapy (RRT) 3 . The RF in MM patients has historically been defined arbitrarily by the "CRAB" criteria (hypercalcemia, renal impairment, anemia and bone lesions) of the International Myeloma Working Group (IMWG), which was previously defined as 2 mg/ dL or greater creatinine levels. These criteria were updated in 2014, and the current definition includes creatinine clearance < 40 ml/min 4 . The criteria to define renal response (RR) are also specific for this disease 5,6 and will be described later in this article. It is important to consider that RF determines worse short-term prognosis, decreasing survival of these patients 7 . The introduction of the so-called "novel agents", such as bortezomib, thalidomide or lenalidomide, improved the survival of patients with MM and RF, even reversing this condition 8 .
There are no studies that correlate the different degrees of reversibility of RF by IMWG criteria with the survival impact of patients in Chile or Latin America.
The aim was to evaluate the frequency of RF in MM at diagnosis, its reversibility with first-line treatment, and the impact of this response on survival.

Material and Methods
A retrospective, descriptive, single-center study was performed. Demographic and clinical characteristics of patients with MM consecutively diagnosed in our institution between 2013 and 2017 were evaluated. The frequency of RF and need for renal replacement therapy (RRT) were also evaluated. In addition, reversibility with first-line treatment defined by IMWG criteria was registered. The induction regimen was based on thalidomide in all cases. Renal failure was defined as a glomerular filtration rate estimated by MDRD of < 40 ml/min, as recommended by the IMWG in 2014. This calculation incorporates creatinine, age, sex and ethnicity. Renal response (RR) was defined according to IMWG criteria. Complete RR (CR) was defined as the sustained improvement (at least two months) of creatinine clearance (CrCl) or estimated glomerular filtration rate (eGFR) from baseline < 50 ml/min to greater than or equal to 60 ml/min. Partial RR (PR) was defined as sustained CrCl improvement from baseline less than 15 ml/min at a value between 30-59 ml/ min. Minimal RR was defined as sustained CrCl improvement from baseline < 15 ml/min at a value between 15 and 29 ml/min. Alternatively, if the basal level is 15 to 29 ml/min, minimal RR was defined as an improvement at 30-59 ml/min. No response (NR) was recorded if none of these criteria were met 6 .

Statistical analysis
A descriptive statistical analysis was performed. Analytical statistics were performed with Student's t test, chi square or Fisher exact test as appropriate. Overall survival (OS) was calculated with Kaplan Meier curves, and comparisons were made using log rank tests. Uni-and multivariate analyses were calculated by the Cox regression test.
The corresponding Ethics Committee approved the present study.

Results
Of a total of 154 consecutive patients diagnosed with MM, 53 (34.4%) fulfilled the RF definition, and 16 (10.3%) required RRT.
The main characteristics of the cohort are presented in Table 1. Induction regimens were as follow: cyclophosphamide, dexamethasone and thalidomide (CTD) in 48%. Melphalan, prednisone and thalidomide (MPT) was used as induction in The median eGFR estimated by MDRD was 20 ml/min (± 4.47 with 95% CI) at diagnosis and 58 ml/min (± 10.2 with 95% CI) at the end of the induction (Figure 1).  The results regarding renal responses of patients in whom data were obtained (52 of the 53 patients) are shown in Table 2.
Three of the patients (19%) became RRT independent.
The 3-year OS of patients without RF at diagnosis was 66% vs. 33% for those with RF (p <0.0001) with a median survival of 53 vs 15 months, respectively.
Among patients with RF with vs. without RRT, the median survival was 9 vs 22 months (p = 0.05) with a 3-year OS of 23% and 38%, respectively ( Figure 2). When divided into 3 groups, including pa-tients without renal failure, patients with RF and complete RR, and patients with RF and any other RR, the 3-year OS was 66, 47 and 13%, respectively (p < 0.001), and the median survival was 53, 27 and 6 months, respectively ( Figure 3). Univariate analysis by Cox regression showed that renal failure (p < 0,001), LDH above the normal range (p = 0,075), ISS III stage (p = 0,006), hypercalcemia (p < 0,001) and age over 60 years (p = 0,026) were related to a worse survival. Multivariate analysis showed that RF and hypercalcemia at diagnosis were independent prognostics factors ( Table 3).

Discussion
We obtained a frequency of RF close to 40%, which is higher to that described in the literature. As expected, our cohort of patients with RF was associated with adverse parameters, such as elevated LDH, hypercalcemia, anemia, advance ISS stage and light chain MM.
It must be highlighted that the incidence of RF depends on the definition used, which, as we have previously pointed, has changed over time.
Using the old definition of creatinine > 2 mg/ dL, Dimopoulos et al. reported a RF frequency of 21% 9 , and Kyle et al. reported a frequency of 19% 10 . Knudsen et al. reported a 29% incidence using a definition of renal creatinine of > 1.3 mg/ dL 11 . Others have reported greater incidences, such as Park 12 , who described a 31% RF incidence based on a definition of CrCl < 60 ml/min. The equation used for RF is also variable. Similar to our group, currently, international studies use MDRD to measure RF. However, for acute RF, the RIFFLE or AKIN classification should be used instead. Very few studies in MM actually use these classifications 13 .
The frequency of patients who required RRT was also similar to the 10% classically described; as expected, this factor conferred a worse prognosis. The median survival of patients requiring RRT for life is less than 2 years 14,15 , and it was 9 months in our cohort. Only three (19%) patients became independent of RRT. The RRT independence rate classically reported is 22% 16 , which is similar to that achieved in our cohort. However, recent studies using novel agents, particularly bortezomib, reported a 44% RRT independence rate from 2008 to 2014 17  Prior to the novel agents era, the survival of these patients with RF due to MM was very poor, ranging from 4 to 19.5 months of median survival (11), and the condition was mainly associated with early mortality (3). We reported a 15-month median survival, which is similar to that period. This could be explained by the fact that there was a suboptimal treatment in our cohort, with no patients receiving bortezomib 19 nor high dose of dexamethasone 20 , some of them receiving doublets rather than triplets 21 , and with very few patients undergoing transplant 22 .
Thalidomide, an immunomodulatory drug (IMiD), is currently available in the Chilean public system. It does not require adjustment for renal failure and achieves RR rates between 55 and 75%, which is similar to the 57% achieved in our cohort 23,24 . Lenalidomide is also a current option, but it should be noted that it requires adjustment for renal failure. Therefore, lenalidomide is not the first choice in these cases 25 .
Proteasome inhibitors are the drugs of choice in RF in MM patients, especially bortezomib, which is the drug with the most robust evidence. It does not require dose adjustment and is usually administered with high doses of dexamethasone and thalidomide or cyclophosphamide. It can also be used in older patients in conjunction with melphalan; however, the latter requires dose adjustment. Bortezomib is associated with high rates of RR and longer  16,19,26 . Moreover, in a Swedish study, bortezomib was an independent favorable factor in the improvement of RF 27 . Reversibility in RF improves survival in MM patients. Knudsen et al. reported 775 patients with MM, in whom reversibility was noted in 50%, representing an important prognostic factor for survival in this group 11 . Silva et al. 28 reported a median OS of 40,2 months in patients who achieve major RR vs 24,8 months in no responders. Our study clearly revealed better survival in patients who achieved complete RR vs any other RR. Gonsalves et al. 15 also reported this difference, with a reported median survival of 112 month in those patients without RF, 56 months in those with RF but who showed RR, and 33 months in those without RR. These data must encourage clinicians to apply the IMWG RR criteria in daily practice.
It seems that achieving RR is crucial, but the chosen drug for this purpose is also important. Dimopoulos et al. reported complete or partial RR of 77% with regimens based on bortezomib, 55% based on immunomodulators such as thalidomide and 43% with lenalidomide 20 . Similarly, Roussou et al. 29 showed 79% improvement in RF with IMiDs vs 94% with bortezomib.
Half of our patients achieved complete RR, which seems relevant considering that no patient received the recommended drug bortezomib. However, based on all the current data, it becomes clear that bortezomib-based regimens are more effective, and, as we demonstrated in our study, higher levels of RR are associated with higher survival rates.
Acute RF is mainly produced by the formation and subsequent obstruction and inflammation of the distal tubules by light chain casts. This condition is attributed to a high tumor load 30 . A limitation of this study (besides its retrospective nature) is that very few patients underwent kidney biopsy. Acute RF was assumed in all these cases given the fact that there was no previous RF history. Indeed, some authors have shown that up to 15% of RF in MM is due to other causes, such as amyloidosis or monoclonal immunoglobulin deposit disease 31 . However, we can assume that the majority had a high tumor burden given that 98% presented with B2 microglobulin levels above the normal range, and these factors are correlated 32 . Also, the current recommendations of performing renal biopsy in MM patients are limited to patients with lower levels of paraprotein and no classical clinical course 5,6 .

Conclusion
Achieving complete RR according to the IMWG criteria in patients with newly diagnosed MM is positively related to increased survival. Therefore, prompt and effective treatment is crucial.
Although not common among the clinical practices in our country, we encourage use of the IMWG RR criteria given its usefulness in prognostically assessing these patients.
Larger and prospective studies are needed to corroborate our data.