Portal de Periódicos da CAPES

Twenty years of boosting antiretroviral agents

2015; Lippincott Williams & Wilkins; Volume: 29; Issue: 17 Linguagem: Inglês

10.1097/qad.0000000000000800

1473-5571

Marta Boffito, David Back, José M. Gatell,

HIV Research and Treatment

Nearly 20 years after the demonstration that saquinavir plasma concentrations could be boosted by ritonavir coadministration [1–3], a relevant question may be whether boosting antiretroviral agents is still necessary nowadays and when. Importantly, more than 90% of patients who are adherent to their combination antiretroviral therapy (cART) achieve and maintain an undetectable viral load, and have a life expectancy approaching the general population [4]. This can be explained by the succession of pivotal advances during the past 20 years, such as the introduction of effective third agents (nonnucleoside reverse transcriptase inhibitors, NNRTI [5], and ritonavir-boosted protease inhibitors, PI/r [6]), by the introduction of fixed-dose combinations (FDCs) containing two or three different active compounds, and the use of strand transfer integrase inhibitors, starting with raltegravir, then elvitegravir and dolutegravir [4]. Also the currently available integrase inhibitor, elvitegravir, needs boosting with ritonavir or cobicistat. Cobicistat was initially specifically developed to boost elvitegravir but it can also boost atazanavir or darunavir, and coformulations containing cobicistat and these protease inhibitors are becoming available in clinical practice. Contrary to ritonavir, cobicistat does not have any intrinsic antiretroviral activity. Both ritonavir and cobicistat inhibit the cytochrome P450–3A4 (CYP3A4) enzyme, thereby reducing the metabolism of concomitantly administered protease inhibitors (or elvitegravir) and leading to a change in their pharmacokinetic parameters. The boosting of drug bioavailability and enhancing of drug exposure in a patient prevents or overcomes resistance and allows less frequent dosing, potentially improving adherence [7]. This is important for drugs that are substrates of CYP3A4 and undergo extensive first-pass metabolism and/or hepatic clearance. Nowadays, efforts are still warranted to try to optimize the response to cART by ensuring high levels of adherence and improvement of tolerability and this can be achieved by better tolerated and safer nucleosides/nucleotides reverse transcriptase inhibitors (NRTIs) (e.g. tenofovir alafenamide to potentially replace tenofovir fumarate) [8]; intramuscularly administered long-acting agents (now studied in phase II/III clinical trials) [9]; ‘less-drug’ regimens (e.g. biotherapies); and, finally, ‘booster-free’ regimens as preferred drug combinations to improve tolerability and reduce the risk for drug–drug interactions (DDI), when possible. When ritonavir was used at doses of 600 mg twice daily as a treating protease inhibitor, it was shown to be associated with important adverse effects [10]. The low doses used today in clinical practice are much safer and remarkably better tolerated. However, ritonavir doses of 100 mg once or twice daily have been shown, both in healthy volunteers and in HIV-infected patients, to be associated with significant increases in total cholesterol, low-density lipoprotein cholesterol and triglycerides, and decreases in high-density lipoprotein cholesterol in a concentration-dependent manner [11–13]. Importantly, when using PI/r, it is often difficult to isolate the ritonavir-related adverse effects from the active protease inhibitor. However, when comparing unboosted atazanavir (400 mg once daily, not licensed in Europe) to atazanavir/ritonavir 300/100 mg, atazanavir is shown not to be a ‘lipid friendly’ agent any longer [14,15]. Small but significant decreases in fasting triglycerides and total cholesterol were observed both in naive patients at 48/96 weeks [16,17] and also in an induction maintenance strategy starting with atazanavir/ritonavir and then switching to unboosted atazanavir 9 months following cART initiation, when plasma viral load became undetectable [18]. Furthermore, although the effects of cobicistat alone have not been studied in detail, a study comparing atazanavir/cobicistat versus atazanavir/ritonavir has shown similar virologic suppression and adverse event rates through 144 weeks. Particularly, severe adverse events leading to study drug discontinuations were 7% (identical in both the study arms). Increase in total fasting triglycerides and total cholesterol were small but significantly higher in the ritonavir arm. Conversely, bilirubin elevation was higher in the cobicistat arm [19]. The same pharmacokinetic characteristics that make ritonavir an effective booster also make it an important perpetrator of DDI, as it inhibits the activity of CYP3A and the multidrug efflux transporter P-glycoprotein (P-gp). CYP3A4 is responsible for the oxidative metabolism of approximately 40–50% of available medications and is often involved in clinically significant DDI [20]. Similarly, P-gp inhibition may result in clinically important DDI (e.g. with digoxin, fexofenadine and others) [21,22]. Interestingly, ritonavir also exerts an inductive effect on drug-metabolizing enzymes including CYP1A2, 2B6, 2C9, 2C19 and uridine diphosphate–glucuronosyltransferase [23]. DDI involving these metabolic pathways may cause suboptimal responses to treatment as they lead to decreased drug systemic exposures. Cobiscitat is also a perpetrator of DDI, as it is a potent inhibitor of CYP3A4 [24] and a moderate inhibitor of CYP2D6 [25]. It also inhibits intestinal transport proteins (e.g. P-gp), increasing the overall absorption of several antiretrovirals, including atazanavir, darunavir and tenofovir. Consequently, both ritonavir and cobicistat have similar potential for DDI. However, cobicistat has not shown any inductive effects on drug metabolic pathways [26]. Moreover, both ritonavir and cobicistat have been shown to cause serum creatinine increases with corresponding estimated glomerular filtration rate (GFR) decreases. The rapid onset and magnitude of these changes seem to be consistent with a ‘drug-interaction’ effect of cobicistat or ritonavir altering creatinine secretion, as changes in actual GFR were not observed during a study conducted with cobicistat in healthy volunteers [27]. Inhibition of creatinine renal tubular secretion by ritonavir and cobicistat is probably caused by the inhibition that these agents exert on the organic cation transporters expressed in the basolateral membrane of the renal proximal tubules (e.g. OCT2) and multidrug and toxin extrusion efflux protein 1 [28,29]. Although the changes in serum creatinine may not be of clinical relevance, the inhibition of these transmembrane transporters could be the cause of DDI and increases in concentrations of renally excreted drugs like tenofovir in the proximal tubule, and therefore nephrotoxicity [30]. For the same reasons, the advent of the new NRTI tenofovir alafenamide will be associated with the use of different doses of the drug: 10 mg with PI/r or protease inhibitor/cobicistat and elvitegravir/cobicistat and 25 mg with other antiretroviral agents like NNRTIs or nonboosted strand transfer integrase inhibitors [31]. When considering our current antiretrovirals, a relevant question may be whether boosted agents are still necessary to treat people living with HIV. For initial therapy in antiretroviral-naïve patients, there are nowadays numerous options that do not involve the administration of ritonavir or cobicistat and thereby avoid many DDI and tolerance issues associated with the booster. The majority of recommended first-line cART in the major guidelines [32–34] are booster free [13 (54%) of the 24 regimens recommended as preferred or alternatives in one or more of these major guidelines], unless there is the need to treat individuals with primary transmitted resistance. In this case, PI/r-containing regimens may still be necessary to ensure virological control because of their high genetic barrier to resistance [35] and plasma concentrations high enough after boosting to overcome low/intermediate levels of resistance. However, in some countries, PI/r are still extensively used as first-line third agents, in combination with two NRTIs, mostly in patients presenting with low CD4+ cell counts or high plasma viral load, because of the perception of a greater potency (although this has never been demonstrated). Indeed, randomized clinical trials have demonstrated the contrary in terms of a better virological and immunological outcome of efavirenz-based regimens as opposed to PI/r regimens [36,37]. PI/r may also be preferred when a physician decides to initiate cART without waiting for a baseline resistance test result. However, today in the Western world, results of a genotypic resistance tests are available approximately 3–6 days following patients’ blood collection. Subanalysis of the CASTLE [38] and ARTEMIS [39] studies has demonstrated advantages of atazanavir/ritonavir or darunavir/ritonavir versus lopinavir/ritonavir regimens in advanced patients mainly because of better tolerance. Moreover, cobicistat-boosted elvitegravir has only been able to demonstrate a marginal benefit in fasting triglycerides (when compared with atazanavir/ritonavir) [40] and total cholesterol (when compared with efavirenz) [41]. Cobicistat was also associated with more nausea and less central nervous system side-effects when compared with efavirenz. In both the studies, and for the reasons discussed earlier, cobicistat was associated with a small but significant increase in plasma creatinine (and a decrease in estimated GFR) that became evident at 2 to 4 weeks after initiation and remained stable thereafter [40,41]. Antiretroviral agent switches are today well tolerated and frequent, as long as the new regimen is able to maintain virological suppression. PI/r-containing regimens have been the target of several switching clinical trials in virologically suppressed patients [42–45], wherein new simpler unboosted agents (in the absence of resistance mutations at baseline and/or previous episodes of virological failure) have been studied. Most of these have been able to demonstrate a noninferior outcome and some improvement in selected laboratory parameters, such as lipid profiles [42–45]. The FDC elvitegravir boosted by cobicistat, tenofovir (disoproxil fumarate) and emtricitabine has demonstrated noninferiority in virologically suppressed patients switching from PI/r [46] or NNRTI-containing regimens [47]. Apart from the advantage of being a single tablet, no further benefits were clearly demonstrated except some marginal improvements in fasting triglycerides and total cholesterol mainly among those subgroups receiving lopinavir/ritonavir or efavirenz at baseline and randomized to be switched to elvitegravir/cobicistat. Dolutegravir is also being tested in a switching strategy both in a broad population receiving an NNRTI or a PI/r-containing regimen (SSAT056 and STRIIVING, https://clinicaltrials.gov) and in a more selected population older than 50 years or with high cardiovascular risk receiving a PI/r-containing regimen (NEAT022/SSAT060), and as part of TRIUMEQ (abacavir/lamivudine/dolutegravir FDC; ViiV Healthcare) in HIV-infected individuals over the age of 65 years (SSAT064, https://clinicaltrials.gov). Finally, third, fourth or further-line therapies are becoming relatively rare nowadays in the Western world; yet, patients in these clinical scenarios require a minimum of two fully active drugs (ideally three) [48]. In these settings, PI/r (and maybe in the future protease inhibitor/cobicistat) are still required. In resource-limited settings, where resistance tests are rarely available, the hope of being able to recommend a first-line regimen A (booster free), followed by regimens B and C in case of successive virological/therapeutic failures, would be ideal but remains a matter of discussion with regimens B and C most likely containing ritonavir or cobicistat-boosted protease inhibitors along with retaining some fully or partially active NRTIs [49,50]. In summary, cART containing a boosting agent is associated with a range of limitations related to tolerance and DDI and may be avoided in many scenarios. However, we recognize that drug costs, the introduction of generics and the preferences of prescribing physicians and patients, including preferences for single-tablet regimens are crucial elements in deciding which cART is recommended. Acknowledgements Conflicts of interest M.B. has received travel and research grants from and has been an adviser for Janssen, Roche, ViiV, Bristol-Myers Squibb, Merck Sharp & Dohme, Gilead, Teva, Cipla and Mylan. D.B. has received travel and research grants from and has been an adviser for Janssen, ViiV, Bristol-Myers Squibb, Merck Sharp & Dohme, Gilead, Teva and Abbivie. J.M.G. has received travel and research grants from and has been an adviser for Janssen, ViiV, Bristol-Myers Squibb, Merck Sharp & Dohme, Gilead and Abbivie.