Risk factors for post-transplant tuberculosis

Background. Post-transplant tuberculosis (post-TxTB) occurs in 12 to 20% of patients in India and results in the death of 20 to 25% of those patients. Prospective studies on post-TxTB are few. Methods. Renal allograft recipients were studied prospectively for 3.1 (0 t o 13.9) median (range) years for incidence, manifestations, risk factors, a nd prognosis for post-TxTB. Kaplan-Meier analysis was used to study the sur vival rates. The extended Cox proportional model for time-dependent covaria tes was used to measure the risk factors when the hazard was nonuniform. Results. Of the 1414 patients considered for inclusion, multiple-transplant subjects (N = 37) and patients who developed pre-transplant TB (pre-TxTB; N = 126) were excluded from the study. The prevalence of post-TxTB was 13.3 % (N = 166). The risk of post-TxTB when on cyclosporine (CsA) therapy was 2 .5 (P = 0.0311) and 1.9 (P = 0.0430) times at less than or equal to6 and le ss than or equal to 12 months, respectively, compared with patients on pred nisolone plus azathioprine (PRED + AZA). The risk of post-TxTB in the prese nce of diabetes mellitus, chronic liver disease, and other co-existing infe ctions [including deep mycoses, cytomegalovirus (CMV), Pneumocystis carinii pneumonia (PCP), nocardia] was 2.2 (P = 0.0011), 1.7 (P = 0.0010) and 2.4 (P < 0.0001) times, respectively. Of the 166 patients with post-TxTB 53 pat ients died, and of those deaths, 17 (32%) were due to post-TxTB; 11 (65%) o f the 17 had co-existing infections. The factors associated with death were HLA mismatches, PRIED + AZA immunosuppression, pre- and post-TxTB, diabete s mellitus, post-transplant diabetes (PTDM), and other co-existing infectio ns. The extended Cox model for death as the outcome variable showed the fol lowing to be significant risk factors: post-TxTB >2 years (P = 0.0036), chr onic liver disease >6 years (P = 0.0457), PTDM >5 years (P = 0.0729), diabe tes mellitus (P = 0.0091), human lymphocyte antigen match less than or equa l to1 antigen (P = 0.0134), two to three antigens (P = 0.0448), and the pre sence of other co-existing infections (P < 0.0001). Conclusions. Cyclosporine therapy is associated with early post-TxTB. Diabe tes mellitus and chronic liver disease are risk factors for post-TxTB. The occurrence of both pre-TxTB and post-TxTB (>2 years) along with hyperglycem ia, liver disease and other co-existing infections are important risk facto rs for death.