Study of prognosis in Waldenstrom's macroglobulinemia: a proposal for a simple binary classification with clinical and investigational utility

Abstract

Abstract Prognostic evaluation of Waldenstrom's macroglobulinemia (WM) is unreliable, few studies considered prognostic factors in WM and only one was derived from a multivariate analysis. One hundred forty-four retrospective, previously untreated patients with clinically overt WM were studied to learn whether overall survival was related to any of the various clinical features presented at diagnosis. Patients were homogeneously treated with intermittent doses of chlorambucil for as long as this showed an effect on the monoclonal component. The population was randomly subdivided into a 90-patient exploratory sample, on whom investigation would be conducted, and in a 54-patient test sample, on whom the results would be validated. In the exploratory sample univariate analysis identified the following parameters as the most important for prognosis: age (< or > or = 70 years), platelet count (< or > or = 120 x 10(9)/L), presence or absence of an abnormal number of red blood cells in the urine, hemoglobin concentration (< or > or = 9 g/dL), erythrocyte sedimentation rate (< or > or = 110 mm at first hour), presence or absence of cryoglobulinemia and of weight loss. Cox multivariate analysis showed that only hemoglobin, age, weight loss, and cryoglobulinemia independently affected survival. These four clinical variables were also shown to be able to discriminate survival significantly in the test sample. Moreover, it was possible to demonstrate (both in the exploratory and the test sample) that clear-cut, albeit dichotomic, survival discrimination can be reached with the presence at diagnosis of either no more than one, or any two or more, of these four prognosticators. These simple clinical criteria could be the basis of an initial binary, prognostic classification of WM, which could help in differentiating therapy according to the severity of the disease, and in properly designing future clinical trials.