Prognostic relevance of soluble CD23 levels in CLL.

Wereadwith interest thepaperdealingwith theprognostic relevance of a scoring system based on clinical and biological parameters in early B-cell chronic lymphocytic leukemia.While thedeterminationofparameters that can serve as indicators of disease progression in B-CLL patients remains a subject for debate, the prognostic value of soluble CD23 levels is useful in CLL. When compared to lactate dehydrogenase, albumin and b2 microglobulin levels, only sCD23 is higher than the normal value for Stage A patients, and signi®cantly so for those Stage A patients who experience disease progression. Recent data from a series of 101 B-CLL patients followed-up in a single centre (Centre Hospitalier Universitaire de Nancy, France) con®rm the clinical value of sCD23 determination. Patients (61 males; 40 females) with a median age of 62 years were classi®ed at diagnosis into Stage A (n=66), Stage B (n=15) and Stage C (n=20). Disease evolution was evaluated according to recommendations from the National Cancer Institute and immunophenotyping (Smlg, CD5, CD22, CD23 FMC7) gave a score of 5 in 48%of patients, 4 in 36% and 3 in 16%. Soluble CD23 levels (sCD23) were determined using the Bender ELISA kit (Medsystems, San Diego, USA), which speci®es a normal range of 0 ± 100 U/ml. Calibration tests conducted with samples from blood donors were never higher than 50 U/ml. Disease progression was observed in 23 patients, including nine of the 47 Stage A patients, and overall survival was 10.4 years. In all cases, sCD23 was higher than the reference value of 100 U/ml and correlated with disease stage (P50.0001), lymphocytosis and b2m level (P50.001). Levels of sCD23 were signi®cantly higher (P50.0001) in patients who experienced disease progression (8342+1390 U/ml) than in those who remained stable (2804+484 U/ml). Furthermore, when patients were separated into two groups according to the median sCD23 of 4493 U/ml, the occurrence of disease progression/death was signi®cantly higher in those patients with sCD23 higher than this median (P50.004). Among Stage A patients, neither disease progression nor death were observed when sCD23 levels remained below this median value, anddi€erences in sCD23 levels between those Stage Apatients who experienced disease progression and those who did not were highly signi®cant (P50.0001). While our values for sCD23 do not appear to be comparable with those reported by Leotard et al., the di€erences are likely to be explained by the di€erent kits used for sCD23 determination. We tested sera (20 blood donors and 18 B-cell lymphoproliferative diseases) using both the Medsystems kit and the Medgenix kit (used in the Leotard study), and obtained a good correlation (r=0.92) with comparable cut-o€ values. Regardless of the technique involved in determining sCD23 levels, it seems clear that this parameter is relevant to the assessment of disease prognosis, at diagnosis. In our series, sCD23 was the only parameter that showed signi®cant di€erences between Binet Stage A, B, and C patients and our data con®rm the likelihood that soluble CD23 levels are strongly associated with disease progression and shorter survival. Determining levels of soluble CD23 is, therefore, of important predictive value in identifying the subset of Stage A CLL patients who are likely to experience disease progression and, thus, likely to bene®t from more intensive therapy.