Retrospective study of the prognostic role of serum thymidine kinase levelin CLL patients with active disease treated with fludarabine

Background: Previous studies have shown that the serum thymidine kinase (TK ) level can be used to determine prognosis in patients with lymphoprolifera tive diseases, but mainly those with multiple myeloma and non-Hodgkin's lym phoma. In patients with chronic lymphocytic leukemia (CLL), TK levels may p rovide prognostic information independent of stage and other prognostic fac tors, but it is still unclear whether they can be used to predict the respo nse to treatment and length of survival. Patients and methods: To determine whether TK levels can be used to predict response and survival, we retrospectively examined the serum TK level in 1 88 previously treated and untreated patients with active or advanced CLL wh o were then treated with fludarabine alone or in combination with prednison e. The correlation of the TK level with other prognostic features and with outcome was then assessed. Results: Serum TK levels were elevated in 92% of the patients, and the leve ls proved to associate with previous treatment, stage of disease, and other tumor-burden related features (i.e., white blood cell counts, absolute lym phocyte count, bone marrow cellularity). The levels were also directly asso ciated with indicators of tumor cell turnover (i.e., beta2-microglobulin an d lactate dehydrogenase levels). Of particular importance, we found that th e TK level was a significant prognostic indicator of both response to treat ment and survival. Specifically, 83% of patients with a TK level of < 10 U/ L responded (complete and partial response) to treatment with fludarabine, whereas only 45% of patients with a TK level of greater than or equal to 10 U/l responded to treatment (P < 0.01). This difference was maintained when we separately analyzed untreated and previously treated patients, and in p atients divided according to the Binet stage. The TK level also added progn ostic information about response to a predictive model based on the hemoglo bin and, albumin levels and the extent of prior treatment. Of further impor tance, the median survival rate in patients with a TK level of < 10 U/l was 65%, as opposed to a rate of 22% in patients with a TK level of greater th an or equal to 10 U/l (P = 0.000). Conclusions: The serum TK level in CLL patients provides useful prognostic information regarding both response to therapy and length of survival and s hould be used in planning appropriate therapy. In particular, patients with a TK level of greater than or equal to 10 U/l have a poor prognosis and sh ould be considered for aggressive treatment.