S4 File -

Introduction
High HHV-8 viral load (VL) in Kaposi Sarcoma (KS) has been associated with Severe Immune Reconstitution Inflammatory Syndrome (Severe-IRIS-KS), which can occur after initiating cART, and leads to high mortality, particularly in patients with pulmonary involvement. We investigate if valganciclovir (as an anti-HHV-8 agent) initiated before cART reduces the mortality associated with Severe-IRIS-KS and the incidence of Severe-IRIS-KS.
Methods
Open-label parallel-group randomized clinical trial in AIDS cART naïve patients with disseminated KS (DKS) as defined by at least two of the following: pulmonary, lymph-node, or gastrointestinal involvement, lymphedema, or ≥30 skin lesions. In the experimental group (EG), patients received valganciclovir 900 mg BID four weeks before cART and continued until week 48; in the control group (CG), cART was initiated on week 0. Non-severe-IRIS-KS was defined as: an increase in the number of lesions plus a decrease of ≥one log10 HIV-VL, or an increase of ≥50cells/mm3 or ≥2-fold in baseline CD4+cells. Severe-IRIS-KS was defined as abrupt clinical worsening of KS lesions and/or fever after ruling out another infection following cART initiation, and at least three of the following: thrombocytopenia, anemia, hyponatremia, or hypoalbuminemia.
Results
40 patients were randomized and 37 completed the study. In the ITT analysis, at 48 weeks, total mortality was the same in both groups (3/20), severe-IRIS-KS attributable mortality was 0/20 in the EG, compared with 3/20 in the CG (p = 0.09), similar to the per-protocol analysis: 0/18 in the EG, and 3/19 in the control group (p = 0.09). The crude incidence rate of severe-IRIS-KS was four patients developed a total of 12 episodes of Severe-IRIS-KS in the CG and two patients developed one episode each in the EG. Mortality in patients with pulmonary KS was nil in the EG (0/5) compared with 3/4 in the CG (P = 0.048). No difference was found between groups in the number of non-S-IRIS-KS events. Among survivors at week 48, 82% achieved >80% remission.
Conclusions
Although mortality attributable to KS was lower in the EG the difference was not statistically significant.