The ideal surgical reconstruction of the aortic root in patients with complex endocarditis is controversial. We compared the short- and long-term outcomes between mechanical valves, bioprostheses, and homografts. We identified all patients undergoing an operation for active complex aortic endocarditis at our institution between 2003 and 2017. We grouped patients according to those who received a mechanical valve, bioprosthesis, or homograft. We used multiple logistic regression and proportional hazards models. To minimize confounding by indication, we used marginal risk adjustment to simulate that every patient would undergo (contrary to fact) all 3 operations. Of 159 patients with complex active endocarditis, 48 (30.2%) had a valve plus patch reconstruction, and 85 (53.4%) had a root replacement. Of all, 50 (31.5%) had a mechanical valve, 56 (35.2%) had a bioprosthesis, and 53 (33.3%) had a homograft. The groups were similar in age, sex, body mass index, comorbid conditions, organism, abscess location, and mitral involvement (all P > .05). However, patients receiving mechanical reconstructions were more likely to have native valve endocarditis (46% vs 37.5% vs 17%; P= .005) and less likely to undergo root replacement (32% vs 28.6% vs 100%; P < .001). Marginal risk-adjusted operative mortality was lowest for mechanical valves (4.8%) and highest for homografts (16.9%; P= .041). Long-term survival after root replacement was worse with homografts than with mechanical valve conduits (adjusted hazard ratio, 2.9; P= .045). In patients with complex endocarditis, mechanical valves are associated with similar, if not better, short- and long-term outcomes compared with homografts, even after adjusting for important baseline characteristics and limiting the analysis to root replacements only.