Heavily experienced people with HIV receiving raltegravir achieved and maintained high levels (77.8%) of virologic suppression after nearly two years of follow-up, similar to previous reports [28, 31, 40]. The SALIR-E study found that 73% of treatment-experienced patients on raltegravir had undetectable PVL at week 206 . Similarly, BENCHMRK trials found 77% of virologic suppression in the raltegravir arm, which was maintained after five years of follow-up, this randomized clinical trial, also identified a superiority of raltegravir over placebo . Nevertheless, our survival analysis showed similar virologic suppression in patients receiving raltegravir or other salvage regimens based on darunavir/ritonavir, maraviroc or etravirine. Also, Buchacz et. al.  found comparable virologic and clinical outcomes (76 and 63%, p=0.51) in raltegravir-based and raltegravir-sparing regimens (with etravirine, maraviroc, enfuvirtide, or elvitegravir).
Our results may be different in part because the subjects were heavily experienced patients with virologic failure and their ART rescue options were somewhat limited by resistance data. In the overall study population, Raltegravir was preferred for sicker patients with a higher PVL, higher proportion of CD4 cell count below 100 cells/ml, and significant proportion of GSS<3 than patients receiving other salvage regimens. Other characteristics that were also higher in raltegravir recipients were the proportion of patients with three or more past treatment failures, alcohol use, cases of anemia, and malnutrition. These characteristics have been associated with an increased risk of virologic failure [41–44].
After matching, the outcomes of raltegravir recipients did not appear to differ from those treated with other third-line drugs, as toxicity rates, treatment interruption, virologic failure, loss of follow-up, and mortality were comparable. Raltegravir was well tolerated, confirming previous safety reports in treatment-experienced patients [21, 28–30], nevertheless adverse events were similar in both groups. Few patients developed integrase mutations at key residues, as noted elsewhere, N155H has been shown to be the primary/major pathway causing resistance to raltegravir. This may be particularly important, as it has been shown that only one mutation is sufficient to observe virologic failure in highly experienced patients [45, 46]. This substitution also reduces susceptibility to elvitegravir but does not affect dolutegravir nor bictegravir alone [47, 48]. We found no explanation for the larger, but non-significant proportion of patients with treatment interruption in the raltegravir group before and after PS matching. Possible reasons remained unknown, as this is not usually coded in medical records available to researchers. However these data raises concerns about unmeasured factors that may affect patient adherence to ART as undiagnosed depression, unreported or unrecognized illicit drug use, or other social reasons [49, 50] that increase risk of mortality or treatment failure [45, 49]. Overall mortality was very low, similar to BENCHMRK and SALIR-E studies, but our analyzes were limited by a relatively small number of participants.
In conclusion, our findings show that virological response is possible in more than 80% of heavily experienced patients who start salvage therapy. They also suggest that raltegravir-based salvage regimens are as safe and effective as those based on other ARV classes, despite the higher disease severity. These results are particularly important given the limited number of third-line drugs for salvage regimens available in Brazil.