Summary. In chronic lymphocytic leukaemia (CLL), stage A was defined in 1979 as the best prognostic group of the three-stage (A, B, C) classification which was thereafter adopted by an International Workshop on CLL. Recently, the question of whether or not all stage A patients should be treated was raised by authors who described potential harmful effects of treatment among stage A patients and heterogeneity of stage A with respect to survival and disease progression.
In this work, we studied the natural history of stage A-CLL in order to determine on which grounds a decision to treat should be made. We performed a prognostic study on 309 untreated stage A-CLL patients enrolled in a randomized clinical trial from 1980 to 1985. Our aim was to segregate a subgroup with a low probability of disease progression and death. We used three endpoints: overall survival (50 deaths), disease progression to stage B or C (78 events) and disease progression to stage C (39 events). Multivariate analyses, using Cox's model, selected nine variables as having a prognostic value for at least one endpoint, namely age, sex, general symptoms, involved axillary lymph node, splenomegaly, haemoglobin, platelets, lymphocyte count and bone-marrow infiltration. By considering the three endpoints jointly, we obtained a proposed definition of smouldering CLL based on four variables, namely haemoglobin level, lymphocyte count, the number of enlarged areas and bone-marrow infiltration. However, we did not succeed in better defining smouldering CLL than in a previous proposal, based on haemoglobin level and lymphocyte count. Moreover, whatever the definition used, including other previous works, disease progression within 5 years was still observed for about 10% of patients with smouldering CLL. Future attempts to define smouldering CLL could perhaps benefit from using biological markers.