Abnormal Serum Free Light Chain Ratios Are Associated with Earlier Time to First Treatment and Poor Survival in Patients with Chronic Lymphocytic Leukaemia

Introduction: Serum free light chains (sFLC) have prognostic significance in plasma cell disorders. In B-cell chronic lymphocytic leukemia (CLL), a small study found 8/18 (44%) of patients to have abnormal FLC ratios but no assessment of prognostic value was published. The aim of the present study was to determine whether abnormal serum FLC concentrations are indicative of a poor prognosis in CLL patients.

Methods: Sera were analysed from 381 previously diagnosed CLL patients (Stage A 307; B 30; C 26; 18 missing; male: Female Ratio 1.6:1, mean age 71 (29–98)) with samples taken before their first treatment (303) or after treatment (78). The study was approved by the Birmingham Heart of England NHS Trust Review Board. Patients were described using the Binet staging system and measured for prognostic markers including CD38, Zap70, mutational status, β2M and FLC. Kaplan Meier survival curves and Cox proportional hazards regression (age, sex, CD38, Zap 70, mutational status, β2M and sFLC) were calculated using SPSS v14.

Results: 147/381 (39%) patient sera had abnormal sFLC ratios. Kaplan Meier analysis of all deaths showed abnormal ratios were significantly associated with worse survival (n=350, p<0.001). Analysis of deaths attributed to CLL (n=30) also indicated that an abnormal FLC ratio was predictive of shorter survival (p=0.001). However, for deaths not attributed to CLL (n=32), the FLC ratio was not significantly predictive of outcome (p=0.112). For Cox regression analysis (n=228) of deaths attributed to CLL only, three significant, independent, prognostic factors were identified: CD38 (p<0.001), abnormal ratio (p<0.001) and Stage (p=0.027). Analysis of the untreated patient population (n=303), using Kaplan Meier analysis of time to first treatment, found that an abnormal lambda ratio (p=0.04) but not an abnormal kappa ratio (p=0.443) predicted earlier treatment. For patients with an abnormal lambda ratio, the mean time to first treatment was 38 months earlier than those patients with a normal ratio. Cox regression analysis (n=171) of time to first treatment, found 4 significant, independent factors predicting earlier treatment: Zap70 (p<0.001), Age (p<0.001), abnormal sFLC ratio (p=0.001) and Stage (p=0.027).

Conclusions: As shown in other monoclonal gammopathies, abnormal sFLC ratios were associated with poorer outcomes in patients with CLL. Furthermore, in an untreated population, patients with an abnormal lambda sFLC ratio required earlier treatment, indicating a pathological mechanism which is as yet unclear but which warrants further investigation.