Measurement uncertainty for serum free light chain assays: estimation and implication on result interpretation

To estimate the measurement uncertainty (MU) and reference change value (RCV) for serum free light chain assays (sFLC), and to review their implications on result interpretation.

Data from 6 to 9months of internal QC and up to 3.5years of EQA results were collected retrospectively, on the Roche Modular P analyzer and Dade Behring BN II nephelometer from two independent laboratories. MU was estimated from its components related to imprecision and bias determinations, while RCV was calculated from the estimated MU and published values on biological variation.

The standard uncertainty related to imprecision for the FLC were similar between two instruments, ranging from ~4% to ~8%, so were the uncertainty related to bias, ranging between 24% and 44%. The overall MU with 95% coverage for Fκ, Fλ and their ratio (rFLC) were 55%, 87% and 103% (Modular P), and 49%, 76% and 91% (BNII) respectively. The RCVs with biological variation for Fκ, Fλ and rFLC were comparable between two instruments and averaged 106%, 138% and 173% respectively. The relatively method-independent MU reflected the sFLC limitations. For monoclonal gammopathy detection, the universal rFLC reference limits ±MU gave an effective range of 0.02-3.15 (BNII) (as compared to 0.26-1.65), revealing substantial diagnostic grey zones on both ends. When monitoring disease activity, a minimum change (RCV) of ~100% in FLC concentrations or ~170% in rFLC would be considered significant. Applying MU when interpreting sFLC results puts current diagnostic cut-offs into perspective, and highlights the importance of collating other clinical findings.