BACKGROUND: In the SMART study, HIV-viral-hepatitis-coinfected persons were, compared with HIV-monoinfected persons, at higher risk of non-AIDS death if randomized to the antiretroviral therapy (ART) interruption strategy. We hypothesized that a marker of liver fibrosis, hyaluronic acid (HA), would be predictive of development of non-AIDS-related outcomes in coinfected participants in the SMART study. METHODS: All participants positive for HCV RNA or hepatitis B surface antigen (HBsAg) and with stored plasma samples were included (n=675). Plasma samples were tested for HA (normal range 0-75 ng/ml) at baseline and months 6, 12 and 24 during follow-up in the drug conservation (DC; interrupt ART until CD4(+) T-cell count <250) group and the viral suppression (VS; continued use of ART) group. Time to non-AIDS death was investigated using Kaplan-Meier analysis. RESULTS: Overall, 52 (31 in DC and 21 in VS) coinfected participants died during follow-up. Coinfected participants who were randomized to the DC group with baseline HA>75 ng/ml had a cumulative risk of non-AIDS death of 24.6% after 36 months of follow-up compared with 9.3% for participants randomized to the VS group (P=0.005), while the cumulative risk for coinfected participants with HA ≤ 75 ng/ml was 4.1% (DC) and 4.7% (VS; P=0.76). The change in HA from baseline to month 24 was 8.3 ng/ml and 4.7 ng/ml in the DC and VS group (P=0.56), respectively. CONCLUSIONS: Interruption of ART was particularly unsafe in HIV-hepatitis-coinfected individuals if plasma HA was increased. HA changed very little during follow-up and was not influenced by differences in CD4(+) T-cell count or HIV viral load.