Document resource
Background The widespread adoption of cabotegravir (CAB)/rilpivirine (RPV) long-acting (LA) in clinical practice may be limited not only by organizational challenges but also by clinical uncertainties due to limited real-world evidence in selected key-subgroups of people with HIV (PWH).Methods PWH enrolled in the ICONA Cohort who started CAB/RPV with HIV-RNA <50 cp/ml and with ≥1 follow-up (FU) were included. Aim: to estimate the risk of treatment failure (TF, defined as CAB/RPV discontinuation for any cause, TD, or virological failure, VF50, 2 HIV-RNA >50 cp/ml or a single >1000 cp/ml followed by ART change), according to selected exposures of interest: PWH aged ≥60 years (vs <60), females (vs males), migrants (vs Italy-born), PWH without availability of genotype resistance test (GRT) before CAB/RPV start, PWH with previous HBV infection (HBcAb pos vs HBcAb neg), body mass index (BMI) ≥30kg/m 2 (vs <30kg/m2).TF was compared across exposure groups using standard survival analysis (Kaplan-Meier (KM) curves) and multivariable Cox regression models stratified by center. Covariate adjustment for each exposure was guided by a directed acyclic graph. Secondary analysis of TD and VF50 were also performed, VF was also evaluated as 2 HIV-RNA >200 cp/ml or one >1000 cp/ml followed by ART change (VF200).Results Overall, 920 PWH started CAB/RPV LA ( table 1): 11.6% females, 10.8% ≥60 years, 7.4% with BMI ≥30 kg/m2, 11.7% non-Italians, 17.9% anti-HBc pos and 8.7% without previous GRT available (% excludes missing data).In a median FU of 21 months (interquartile range, IQR, 10-30) 104 TF occurred (12 VF50 and 92 TD); the estimated cumulative probability of TF was 9.2% (95% CI 7.4-11.4%) at 1-yr and 12.4% (10.2%-15.0%) at 2-yrs.KM curves showed comparable cumulative probabilities of TF across the selected exposures (figure 1). In the adjusted Cox regression model, no evidence for a difference between the groups in TF was observed (females vs males, adjusted hazard rate, aHR, 1.02, 95% CI 0.52-1.99; migrants vs Italian-born, 0.90, 0.47-1.69; obese vs BMI<30 kg/m2, 1.33, 0.65-2.74; previous GRT available vs not, 0.49, 0.21-1.16; age ≥60 ys vs <60, 0.82; 0.41-1.64; AntiHBc pos vs AntiHBc neg, 1.14, 0.66-1.97) (table 2).Estimated cumulative probability of TD, mainly driven by toxicity/adverse events (69.3%), was 9.2% (7.4-11.4%) at 1-yr and 12.2% (10.0-14.8%) at 2-yrs. Similarly, no evidence for a difference in aHR of TD was detected.Estimated cumulative probability of VF50 was 1.5% (0.8-2.6%) at 1-yr and 11/12 VF occurred <1yr from CAB/RPV start. The estimated 1-yr probability of VF200 was 0.7% (0.3-1.6%).Conclusions In this real-world Italian study, rate of TF of CAB/RPV was 12.4% at 2 years, with no evidence of differences across the key groups evaluated, supporting a broader use of this strategy across heterogeneous groups, while acknowledging the low prevalence of key exposures; residual confounding and potential selection bias cannot be excluded.Abstract OC74 Figure 1 and Table 1–2Estimated probability of treatment failure by key exposure groups