Serum Free Light Chain Ratio Is Inversely Proportional to the Egfr, and Has Been Increasing over the Last Decade

Background: Serum concentrations of FLC depend on plasma cell production and renal clearance. The reference range was defined from 127 healthy donors (21-62 years of age) and 155 patients 51-90 years of age (Katzmann JA et al, 2002, PMID 12194920) – kappa (3.3-19.4 mg/L), lambda (5.7-26.3 mg/L) and FLC ratio (0.26-1.65). The Iceland Screens, Treats, or Prevents MM (iStopMM) study analyzed 6,461 subjects with eGFR < 60 mL/min, proposing new reference intervals for the FLC ratio of 0.46–2.62, 0.48–3.38, and 0.54–3.30 for eGFR 45–59, 30–44, and < 30 mL/ min/1.73 m2 groups, respectively (Long et al, Blood Cancer J, 2022, PMID 36100605). Over the past 20 years the median FLC ratios in different cohorts (n=100-200) suggested that the median FLC ratio has drifted upward over time (Minnema et al, 2022, PMID 36376268). A recent abstract from Thermo Fisher suggested that buildup of alpha-1 antitrypsin (A1AT) in calibrators involved in the Freelite assay falsely increased the kappa and ratio results (Stewart F et al, ADLM, 2024). Our goal was to evaluate serum free light chains in a larger sample size over time, adjusting for eGFR.

Methods: Data used in this study came from Epic Cosmos, a dataset created in collaboration with a community of Epic health systems representing more than 259 million patient records from over 1,548 hospitals and 35.4K clinics from all 50 states and Lebanon. In Cosmos, we created a histogram of the serum free light chains (kappa, lambda, and ratios) across all sites, isolating the two most popular LOINC codes to include. We isolated only patients with lab values with reference ranges for kappa, lambda, and the ratio specifically for the Freelite assay; eliminating Diazyme, Sebia, etc. We eliminated patients with a history of lymphoproliferative disorders as a billed, admitting, or discharge diagnosis based on ICD9 (20*.**, 277.3*) and ICD10 (C8*.**, C90.**, C91.**). As a control we performed similar analyses for hemoglobin (LOINC 718-7) with 715m results. Analysis was performed in Cosmos virtual environment, SQL code for data extraction, Python for analysis.

Results: We focused on two LOINC codes for lambda (33944-0), kappa (36916-5), and the ratio (48378-4). Single Epic sites for the same LOINC code could have values with different units and reference ranges likely reflecting different vendors/equipment over time. Only approximately 3% of FLC values in Cosmos were obtained prior to 2014, and with new Epic sites there has been a steep linear increase in the number of results available since 2016. A small peak below the reference range minimum were thought to be due to initial dilutions in normal samples, but an upper peak at 120-130 mg/L for both kappa and lambda remains unexplained. There was significant variation between sites: among 198 Epic sites, 15 had median lambda & kappa values as much as twice the upper limit of normal. We saw no change in kappa, lambda, or the ratio over time, but individual sites would often have periods of ≈6 months where the median values would vary more than 2.5x standard deviation.

Using CKD-EPI without a race adjustment (Inker LA et al, 2021, PMID 34554658) with the creatinine drawn on the same day as FLC blood test, in patients without a lymphoproliferative disease by diagnosis and excluding patients with monoclonal protein > 0, two associations came out: 1) the serum free light chain ratio is inversely correlated with eGFR, statistically significant for eGFR > 30 mL/min; and 2) the FLC ratio has been increasing over time since 2014-2024, statistically significant for patients with eGFR>60 mL/min, even when adjusting for eGFR. Although the FLC ratio has been increasing over time, most remained within the normal range.

Conclusions: We are the first to report the relationship between free light chain ratio and eGFR. While the FLC ratio has been increasing over time, most median values remain in the reported normal range. Using the Cosmos data set of 259 million patients, there was a slight increase over time in the free light chain ratio over the last 10 years.

Individual sites should be aware of data presented in 2024 by Thermo Fisher regarding protein-protein interactions that could falsely elevate kappa light chain and serum free light chain ratio values.

Acknowledgements: Brendan Joyce at Epic Research for code to incorporate race blind CKD-EPI. Jill Pauli PhD at Thermo Fisher for a better understanding of FLC measurements and their study of protein-protein interactions.