Biochemical markers in non-Hodgkin's lymphoma stages III and IV and prognosis: a multivariate analysis

H Hagberg,B Glimelius,S Gronowitz,A Killander,C Källander,T Schröder
DOI: https://doi.org/10.1111/j.1600-0609.1984.tb02211.x
Abstract:The prognostic value of different pretreatment laboratory and clinical findings at diagnosis was assessed in a series of 141 patients with generalized non-Hodgkin's lymphoma. Univariate and multivariate survival analysis (Cox's regression model) was performed, using serum analysis of deoxythymidine kinase (S-TK), beta 2-microglobulin, lactic dehydrogenase, alpha 1-acid glycoprotein = orosomucoid (S-alpha 1 AGP), haptoglobin and ferritin. In addition, Hb and the erythrocyte sedimentation rate (ESR) were measured. The clinical variables were age, presence or absence of B-symptoms, histopathology ('low-grade'; 'intermediate grade' and 'high-grade' malignancy) and bone marrow involvement. Of the 8 biochemical markers, all except Hb and the ESR showed a significant relationship to survival. Among the clinical variables, this finding was made for B-symptoms and histopathology. Using a multivariate analysis on all variables, S-TK was found to be the best factor for predicting duration of survival. The only significant additional information was provided by S-alpha 1 AGP. When only the clinical variables were taken into account, it was found that histopathology added significant information to that yielded by B-symptoms in the prediction of the survival time. When the biochemical variables were added to this model, only S-TK was of significant additional prognostic value.
What problem does this paper attempt to address?