rrent bacterial pneumonia. Similarly, an AIDS-defining event was considered only if the patient had not experienced that particular AIDS-related opportunistic infection or neoplasm before. 3 / 15 Risk of Severe Non-AIDS Events among Immunological Non-Responders Statistical analysis Uni- and multivariable logistic regression models were conducted to identify variables associated with immunologic non response at 1 year after the initiation of HAART. Predictors of non AIDS-related severe event occurring after year 1 were assessed using Kaplan-Meier estimates and adjusted Cox proportional BIRB-796 cost hazard regression analysis. Follow-up accrued from the first viral load <50 copies/ml measured after 915 months of treatment and was right censored in case of VL >50 copies/ml or missing for 180 days, HAART discontinuation, loss to follow-up or death. Patients were categorized as INR or immunological responders if their CD4+ T-cell count was <200 or 200 cells/l after 1 year of HAART, respectively. The following covariates were tested: immunological response at year 1, age, gender, risk factor for HIV acquisition, previous nADE, previous AIDS-defining event, hepatitis B and C co-infection, pre-HAART CD4+ T-cell count, CD4+ count change at year 1, type of HAART and occurrence of AIDS-defining event during the follow-up. Two multivariable models were performed: in the first one, all variables were adjusted for age and gender. In the second model, estimates were adjusted for all the variables significantly associated with the outcome in the first model plus age, gender and CD4+ T-cell count percentage change. When more than one event had occurred in a single patient, the time to the event was defined as the time between baseline and the first occurrence of any of the events considered. The same models and covariates were used in the secondary analysis exploring predictors of the composite outcome. All statistical analyses were performed using SAS 9.2 statistical software. All P-value presented are two sided and a P-value <0.05 indicated conventional statistical significance. Ethics Statement Patients included in the MASTER cohort study provide, at enrolment, written informed consent to include their clinical and biological data in the MASTER database for scientific purposes. The data are anonymized and the database is hosted in the "Fondazione MISI's" headquarter, in compliance with current Italian regulations. The study was approved by the Ethical Committee of the Hospital "Spedali Civili", Brescia and those of PubMed ID:http://www.ncbi.nlm.nih.gov/pubmed/19778700 the following Institutions: University Hospital of Ferrara, Ferrara; AO 
Papa Giovanni XXIII, Bergamo; Ospedale Policlinico–University of Bari, Bari; San Gerardo de’ Tintori” Hospital, Monza; Hospital of Cremona, Cremona; “Santa Maria Annunziata” Hospital, Firenze; “Sacro Cuore” Catholic University, Roma. Results Patient characteristics Among 1869 patients initiating HAART with a CD4+ T-cell count <200 cells/l in the MASTER Cohort, 164 patients were not included because followed for <12 months, 477 because did not have two consecutive HIV plasma viral load <50 copies/ml within the first year of treatment and 7 because lacked of CD4+ T-cell count determination at month 12. The 4 / 15 Risk of Severe Non-AIDS PubMed ID:http://www.ncbi.nlm.nih.gov/pubmed/19778211,3601242,2373292,2299688 Events among Immunological Non-Responders remaining 1221 patients were included and studied over a median of 3 years, accounting for a total of 4,708 person-years of follow-up. The median age at treatment initiation was 41.2 years; the youngest patie