Individualizing salvage regimens

Increased drug levels could overcome resistance and improve the response to a salvage regimen. We evaluate the inhibitory quotient (IQ, Ctrough/protein-binding corrected IC50) as a predictor of virological response in 52 patients included in two dual protease inhibitor (PI)-based salvage regimens. The HIV-RNA level decrease at 12 weeks was greater in patients who achieved an IQ greater than 1. The IQ could be useful to improve the virological response to a dual PI salvage regimen. The emergence of resistance to protease inhibitors (PI) represents a significant barrier to the successful long-term suppression of viral replication [1]. Nevertheless, resistance should be considered as an increase in the inhibitory concentration to a level higher than that of wild-type viruses. This fact opens the possibility of overcoming resistance by increasing drug exposure [2]. Recently, the use of the Ctrough/IC50 ratio has been suggested as a useful marker of antiretroviral efficacy. The aim of this study was to analyse the clinical usefulness of the Ctrough/IC50 as a predictor of virological response in two different dual PI regimens. For the study, we selected patients who met the following criteria: (i) being on therapy with either ritonavir/indinavir, or with nelfinavir/saquinavir, twice a day and after meals; (ii) to have phenotypic testing and pharmacokinetic determinations at baseline; and (iii) a follow-up that lasted at least 6 months. All the patients received two nucleoside analogues with the combination of ritonavir and indinavir, 100/800 mg (14 patients), or 400/400 mg (11 patients), or the combination of nelfinavir and saquinavir, 1250/1000 mg twice a day (27 patients). Phenotypic resistance analysis was performed using a recombinant virus assay [3] (Antivirogram, Virco, Mechelen, Belgium). This analysis expressed the IC50 (ng/ml) of a particular drug when tested with patient-derived recombinant virus isolates, in comparison with the mean IC50 (ng/ml) of the same drug obtained when tested with a reference wild-type virus isolate (IIIB/LAI). In this reference isolate, the mean IC50 of indinavir was 0.0137 μM (84.09 ng/ml), the mean IC50 of ritonavir was 0.0232 μM (16.77 ng/ml), the mean IC50 of nelfinavir was 0.0112 μM (6.35 ng/ml), and the mean IC50 of saquinavir was 0.0063 μM (4.22 ng/ml) [4]. Plasma drug levels were measured after at least 15 days on therapy (mean 21 days). On the day of measurement, after overnight fasting, blood samples were collected in heparinized tubes just before the ingestion of the dual PI combination. Indinavir, ritonavir, nelfinavir, and saquinavir concentrations in plasma were measured using a validated high performance liquid chromatography assay [5]. In every patient, the inhibitory quotient (IQ) was calculated by dividing the trough level by the IC50 of the isolate (expressed in ng/ml). As IC50 values were calculated in vitro, they were adjusted for in-vivo protein binding. Therefore, IC50 values were mathematically corrected by multiplying for a correction factor (2 for indinavir; 33 for saquinavir; and 50 for nelfinavir) [4,6]. A total of 52 HIV-infected patients were included. Most patients had had extensive previous antiretroviral exposure (median time 56 months), with PI experience in all cases (median time, 27 months). Of note was the fact that 87 and 68% of those included in the ritonavir/indinavir combination had previously failed an indinavir or ritonavir-containing regimen, respectively, whereas 75% of those in the nelfinavir/saquinavir combination had received saquinavir therapy. At baseline, the median CD4 cell count and plasma HIV-RNA level were 232 × 106/l and 4.2 log10 copies/ml. The median change in the IC50 was a 25.4-fold increase in indinavir, 108.5 in ritonavir, 14.5 in nelfinavir, and 2.5 in saquinavir; and this change was directly correlated with the previous time on PI therapy (r = 0.67; P = 0.001). After 3 and 6 months, there was a median HIV load decrease of 1.25 and 1.05 log10 copies/ml. Overall, the median trough levels of indinavir and ritonavir were 1552 ng/ml [interquartile range (IQR) 619–2572] and 125 ng/ml (IQR 92–412), respectively, whereas the nelfinavir and saquinavir Ctrough values were 2480 ng/ml (100–1001) and 160 ng/ml (50–1350), respectively. The median protein-binding corrected IQ of indinavir was 0.42 (0.06–5.01). For nelfinavir and saquinavir, the wild-type IC50 was 6.35 and 4.22 ng/ml, and the median protein-binding corrected IC50 of resistant isolates was 3407 and 274 ng/ml, respectively. Therefore, the nelfinavir IQ was 0.85 (0.06–14.74) and the saquinavir IQ was 0.3 (0.009–6.39) (Fig. 1).Fig. 1.: (a) IC 50 of resistant isolates (ng/ml) to indinavir, and (b) nelfinavir, in comparison with the IC 50 of reference wild-type isolate, and median trough levels reached. (c) HIV-RNA decrease at the 3rd month according to the inhibitory quotient of ritonavir/indinavir, and (d) nelfinavir/saquinavir. □ Indinavir; ▓ ritonavir; ▪ nelfinavir; ░ saquinavir.There was a linear relationship between increasing IQ values and a higher HIV-RNA level decrease. Therefore, for indinavir-treated patients, the mean HIV decrease was −0.83 and −1.2 log10 copies/ml with an IQ below or above 1, respectively (P = 0.09). Similar results were obtained for ritonavir-treated patients, with a mean decrease of −1.2 log10 copies/ml with an IQ greater than 1. For nelfinavir, the mean HIV-RNA level decrease was −0.51 log10 copies/ml with an IQ less than 1 and −1.68 log10 copies/ml with an IQ greater than 1 (P = 0.04). In saquinavir-treated patients, the rate of decrease was −0.92 and −1.1 log10 copies/ml (P = 0.03) (Fig. 1). This study suggests that it is possible to improve the rate of response by reaching higher drug levels than the IC50. In other words, it demonstrated that the use of available phenotypic testing and the knowledge of the Ctrough reached in every patient may be a useful tool in salvage regimens. However, the difficulties with this lie in the uncertainty as to what levels of drugs, relative to in-vitro inhibitory values, are required to achieve optimal efficacy, the lack of standardization of pharmacokinetic data in PI, and the variability of drug levels among patients [6]. Therefore, we can not establish an ideal IQ to be reached in a dual PI combination. Kempf et al. [7] found a better response to the combination of ritonavir and indinavir for those patients having an indinavir IQ greater than 2, although they used an IQ obtained through a virtual phenotype. Molla et al. [8] also suggested that an IQ greater than 15 with the combination of ritonavir/lopinavir was strongly associated with a better virological response in PI-experienced patients. This study demonstrates that an IQ greater than 1 was associated with significant responses to nelfinavir, saquinavir and indinavir. As most of our patients had a low IQ, as expected in heavily PI-experienced patients, we can not rule out the possibility that a higher IQ could be associated with better virological responses. A priori, it could be an easy approach to reach the highest possible level in the most resistant strains, but this objective could be unrealistic if toxicity to PI is related to the overall drug exposure [9,10]. Of note is the fact that the virological response was observed in an important percentage of patients reaching an IQ of less than 1 and therefore with plasma drug levels below the IC50. Although this fact could be attributed to the simultaneous use of other drugs in the regimen, the intracellular action of PI may be responsible, in spite of subtherapeutic levels, or the antiviral action of drug metabolites, as shown with nelfinavir [7]. In summary, this preliminary study proved the usefulness of the IQ as a predictor of virological response to a salvage regimen with a dual PI combination in multi-PI-experienced patients. Acknowledgements The authors wo would like to thank Virco for their support in performing phenotypic resistance testing.