INTRODUCTION: Azathioprine has variable efficacy in inflammatory bowel disease. Previous studies suggested that either neutropenia, an increase in the mean corpuscular volume, the assessment of thiopurine methyl-transferase activity or erythrocyte 6-thioguanine values might predict the treatment response. However, due to the conflicting results of the preceding studies there are yet no established laboratory values which allow an estimation of the clinical response. PATIENTS AND METHODS: 45 patients with Crohn's disease and 39 patients with ulcerative colitis were enrolled in this retrospective evaluation. After a minimum of six months therapy with azathioprine patients in remission were compared with those who did not achieve a stable remission with respect to the number of leucocytes, lymphocytes, neutrophil granulocytes and the mean corpuscular volume. RESULTS: Patients who went into remission during treatment with azathioprine displayed significantly lower leukocyte counts if compared to patients who were not in remission (p = 0.004 in Crohn's disease and 0.003 in ulcerative colitis). A similar tendency was also observed with respect to the granulocyte count (p = 0.007 in Crohn's disease and 0.004 in ulcerative colitis). The mean corpuscular volume did not correlate with the response to purine analogues. DISCUSSION: The absolute leukocyte count and the percentage of granulocytes seem to predict the response to purine analogues in inflammatory bowel disease and possibly offers a feasible and cost effective diagnostic tool for the assessment of therapeutic efficacy. Subsequent preferably prospective studies should aim to define the optimal cut-off value for the leukocyte count.