The median CD4 count at baseline was 61 cells/μL (range 0 to 100 cells/μL), and 39% of the patients had a cell count <50 cells/μL. The median HIV viral load was 98 663 HIV-1 RNA copies/mL (range <40 copies/mL to 3.5 × 107 copies/mL). Forty-one per cent of patients either were already receiving or started an antiretroviral treatment at the time of the CMV measurement. Of these, 22% had full viral suppression (<50 copies/mL)

and 71% had a viral load of >200 copies/mL at baseline. The median duration of follow-up was 4.8 years. During the complete follow-up period, CMV end-organ disease occurred in 25 patients (2.2%; retinitis in 19 patients and gastrointestinal diseases in six patients) and other ODs in 183 patients (16%). A total of 246 patients died (22%). The most frequent ODs were Candida oesophagitis (41 Ibrutinib patients; 22%), atypical mycobacterial diseases (23 patients; 13%), Pneumocystis carinii pneumonia (19 patients; 10%), Kaposi’s sarcoma (14 patients; 8%) and non-Hodgkin’s lymphoma (10 patients; 6%). During the first year of follow-up, CMV end-organ disease occurred in 19 patients (1.7%)

and other ODs in 95 patients (8.4%), and 78 patients (6.9%) died. The median times between the CMV DNA measurement and the development of CMV end-organ disease, other ODs and death were 141, 139 and 160 days, respectively. Thirty-four per cent of patients (368 patients) had detectable CMV DNA in plasma at baseline, with a median of 136 copies/mL and a maximum of 38 800 copies/mL. This percentage was stable from 1996 to 2007. Amongst the patients with AZD6244 in vitro a detectable value, 18 (5%) experienced evolution towards CMV end-organ disease. During the first year of follow-up, 83% of the patients who developed CMV end-organ disease had a detectable CMV

DNA value at baseline, with a median positive value of 1990 copies/mL [interquartile range (IQR) 279.5–4332.5 copies/mL]. Of those who developed an OD other than CMV end-organ disease, 42% were CMV DNA-positive (median CMV DNA 179.0 copies/mL; IQR 89.8–1220.0 copies/mL), and of those D-malate dehydrogenase who died, 38% were CMV DNA-positive (median CMV DNA 283.5 copies/mL; IQR 81.0–4117.5 copies/mL). In the group of patients who neither died nor developed CMV end-organ disease or any other OD, 32% had a detectable value, with a median of 125.5 copies/mL (IQR 51.7–740.0 copies/mL). Using time-dependent ROC curves, we assessed the prognostic performance of the CMV DNA value at baseline in predicting our different endpoints. The areas under the curve are shown in Figure 2 for each endpoint, according to the timeframe. The optimal prognostic performance of the CMV DNA value in predicting CMV end-organ disease was achieved at 6 months (AUC 0.8; 95% CI 0.7–0.9). For predicting other ODs, the optimal prognostic performance was achieved at 2 months (AUC 0.8; 95% CI 0.6–0.9) and for mortality it was achieved at 6 months (AUC 0.6; 95% CI 0.5–0.7).