Chronic Kidney Disease Impairs Predictive Ability of Serum Free Light Chain Ratio for Clonal Bone Marrow Plasma Cells in MGUS Patients

Introduction:

Various models, including the Mayo Clinic model and iStopMM, aid in predicting which patients with monoclonal gammopathy of undetermined significance (MGUS) are likely to have or develop smoldering (SMM) or multiple myeloma (MM). Both models incorporate serum free light chain ratio (SFLCR) as a variable. Renal impairment is known to affect serum free light chains by raising free kappa, and may lessen the predictive ability of SFLCR for clonal bone marrow plasma cell (BMPC) percentage. Further, patients with CKD and kappa-involved MGUS may receive unnecessary bone marrow biopsies due to higher scores on current models, while those with lambda-involved disease may be underdiagnosed due to normalization of SFLCR.

Methods:

We performed a retrospective chart review of all patients who had a listed diagnosis of a monoclonal gammopathy and underwent bone marrow biopsy at Thomas Jefferson University Hospital and affiliate hospitals between January 1^st, 2017 and December 31^st, 2023. Eligible patients had a diagnosis of MGUS or suspected MGUS based on an abnormal SFLCR. Patients were excluded if they met any criteria for SMM, MM, or a lymphoproliferative disorder before the bone marrow biopsy was performed, if documentation was not available before and after the biopsy, and if key lab variables were not available. Clonal BMPCs had to be reported based on histopathology and immunohistochemistry (patients with only flow cytometry were excluded), and those with polyclonal plasma cells only were considered to have BMPCs of zero.

Results:

The full study population included 372 patients (43.5% female) with a median age of 70 years at the time of bone marrow biopsy. 93.5% of patients had a diagnosis of MGUS, while 6.5% had abnormal SFLCs without evidence of serum or urine M protein. Kappa was involved in 64.2% of patients. Among patients with MGUS, 89.1% had a single clone while the remainder had 2 or more clones. Overall, 46.0% of patients had at least stage 3 CKD based on GFR < 60 mL/min/1.73m² at the time of biopsy. Regarding risk factors, 17.2% had M spike ≥ 1.5 g/dL, 29.0% had abnormal SFLCR, 33.9% had a non-IgG isotype, and 11.0% had immunoparesis. Regarding outcomes after biopsy, 65.3% were confirmed to have MGUS, 4.3% had no plasma cell dyscrasia, 22.6% had SMM, 4.6% had Waldenstrom macroglobulinemia/lymphoplasmacytic lymphoma (WM/LPL), and 1.3% had MM. In total, 29.6% of patients had ≥ 10% BMPCs (including LPL cells). Variables associated with a significantly greater likelihood of having ≥ 10% BMPCs included abnormal SFLCR, M spike ≥ 1.5 g/dL, and immunoparesis, but not non-IgG isotype.

Next, we excluded patients with biclonal gammopathies involving both kappa and lambda to compare patients by affected light chain and presence of CKD (N=345). In non-CKD patients, abnormal SFLCR was significantly associated with ≥ 10% BMPCs (37.6%, vs 18.2% for normal SFLCR, P=.010), however no significant association was seen in CKD patients (29.8% vs 23.3%, P=.550). Evaluating by involved light chain, in non-CKD kappa-involved patients an abnormal SFLCR was associated with a nonsignificant trend toward greater likelihood of ≥ 10% BMPCs in kappa (35.5%, vs 13.3% for normal K/L, P=.140) and lambda patients had a significant association (46.2% vs 20.0%, P=.031). In CKD patients, while there was an association between abnormal SFLCR and ≥ 10% BMPCs in kappa patients (27.1%, vs 0.0% for normal SFLCR, P=.035), there was no association in lambda patients (37.9% vs 34.5%, P=1.0). In CKD patients with no Mayo risk factors (low-risk), zero kappa patients had ≥ 10% BMPCs, while 26.7% of lambda patients did. And among CKD patients whose only Mayo risk factor was abnormal SFLCR, 24.1% of kappa patients had ≥ 10% BMPCs while 40.0% of lambda patients did.

Conclusions:

Our retrospective study found that while an abnormal SFLCR in MGUS patients was associated with subsequently having ≥ 10% BMPCs, the association was lost in patients with CKD, who typically have excretion-related increases in free kappa. Most notably, there was no observed difference in probability of ≥ 10% BMPCs between normal vs abnormal SFLCR among CKD patients with lambda-involved MGUS, and these patients had a high rate (34.5%) of ≥ 10% BMPCs even with a normal SFLCR, likely due to false normalization of K/L. Clinicians using the Mayo Clinic or iStopMM models must consider the presence of CKD when assessing the utility of SFLCR for predicting BMPC percentage in MGUS patients.

Rhoades:Novartis: Consultancy; Spark Therapeutics: Research Funding; Veralox Therapeutics: Research Funding.