Objective: This study aimed to identify situational and behavioral factors associated with condomless anal sex without on-demand PrEP in the open-label extension (OLE) study of the ANRS-IPERGAY trial. Methods: Univariable and multivariable modified Poisson regressions with a generalized estimating equation (GEE) were used. The attributable risk percentage for each explanatory variable and for condomless anal sex without PrEP was calculated. Results: In the OLE, 19% of anal intercourses were unprotected (i.e. no PrEP or condom). Of these, 85% were attributable to sexual intercourse with main partners and 47% with HIV-negative partners. The following factors were positively associated with condomless anal sex without PrEP: a depressive episode in the previous 12 months [aR (95% CI), P-value: 1.49 (1.02--2.17), 0.039], a higher number of sexual intercourses during the previous 4 weeks [1.01 [1.002--1.02], 0.014], and sexual intercourses under the influence of alcohol [1.45 (1.10--1.92), 0.008]. By contrast, condomless anal sex without PrEP was less frequent during sexual intercourses with known casual, unknown casual and multiple partners [0.20 (0.14--0.30), <0.001; 0.10 (0.05--0.20), <0.001; 0.11 (0.05--0.29), <0.001, respectively], as well as with HIV+ partners with an undetectable viral load and HIV+ partners with a detectable/unknown viral load or unknown serology status [0.57 (0.38--0.86), 0.007; 0.52 (0.32--0.87), 0.012, respectively]. Conclusion: Choosing to have condomless anal sex without PrEP depends primarily on the sexual partner's characteristics (level of intimacy, serological status). This reflects a form of rationality in HIV risk management. However, our results raise questions about the true efficacy of managing HIV risk using this approach.