Background: Mortality from tuberculosis is high even in the chemotherapy era. Our study determines the predictor role of lymphocyte count in the mortality of non-HIV tuberculous patients. Materials and Methods: This case- control study was performed retrospectively in a university hospital, which is a secondary and referral center for tuberculosis in Iran. All documented pulmonary tuberculosis cases that had died while being hospitalized in TB wards during the year 2002 were enrolled. Equal number of documented tuberculosis who had been discharged from hospital with good conditions were also enrolled by simple randomized selection. All pertinent data including the first documented hematologic indices of the cases were gathered and analyzed using X2 , logistic regression, and nonparametric tests. Results: During the last year, 33 patients died from tuberculosis with an average age of 53 ± 16.5 yr. [11(33%) were female]. The mean lymphocyte percent in CBCs taken from the case group was 15.5 ± 10.2 comparing to 27.1 ± 9.8 for the control group. The frequency of lymphocyte counts below 1000 was 11 (37%) for the cases and 5 (15%) for the controls. The frequency of lymphocyte counts below 15% was 21 (67%) in the case group compared to 3 (9%) for the control group. Both results showed significant differences between the two groups (P = 0001). The odds ratio for total lymphocyte count deficiency was 3.5 and the odds ratio for < 15% was 27. Conclusion: This study revealed that lymphocytopenia may be used as a proper measure for determining mortality risk in non-HIV pulmonary TB patients. However, it seems necessary to confirm this new finding with introspective studies using broader sample sizes. (Tanaffos 2003; 2(7): 25-31)