A proportion of HIV-patients does not normally restore their CD4 counts despite virological response to HAART. Those whose CD4 counts persistently remain closed to the critical threshold for opportunistic infections deserve special interest. To study the risk factors, the long-term CD4 counts evolution, and the risk of death of patients who persistently maintain low CD4 counts, despite virological response to HAART, within a multicenter, hospital-based cohort study. A total of 147 patients were selected from CoRIS-MD and classified into a "Low-Group" or a "High-Group", depending on their CD4 counts after two-years of effective HAART (threshold 250 cells/microL). Associated risk factors were analysed by logistic regression, the CD4 dynamics were evaluated over a total period of 7.70 years (IQR, 6.70-9.00), and mortality was estimated by Cox proportional hazard. A total of 40 patients (27%) were classified into the "Low-Group". The odds ratio for this group increased with age, being 4.56 (2.23-9.33) for over 40, and was also higher among IDU, 3.63 (1.04-12.68). Six years thereafter, among these patients, only a 30% exceeded 350 CD4 cells/microL and a 12% exceeded 500 CD4 cells/microL. Furthermore, the "Low-Group" had a death rate of 2.42 per 100 persons/year (95%CI, 1.01-5.81), although once adjusted by age the estimates were no longer significant [4.14 (0.87-19.72)]. Our results suggest that those HIV patients who have not overcome the critical threshold of 250 CD4 cells/microL after a two years period of virologically effective HAART do persist with the aforementioned failure of CD4 restoration for a much longer time.