Clinical and Pharmacokinetic Data Support Once-Daily Low-Dose Boosted Saquinavir (1,200 Milligrams Saquinavir with 100 Milligrams Ritonavir) in Treatment-Naive or Limited Protease Inhibitor-Experienced Human Immunodeficiency Virus-Infected Patients

Abstract

We evaluated the plasma and intracellular pharmacokinetics, clinical efficacy, and safety of once-daily low-dose boosted saquinavir (SQVr; 1,200 of saquinavir [SQV] with 100 mg of ritonavir) plus two nucleotide reverse transcriptase inhibitors in treatment-naive or limited protease inhibitor (PI)-experienced human immunodeficiency virus (HIV)-infected patients. A prospective study without entry restrictions on the plasma HIV-RNA (VL) or CD4 cell count was carried out. Plasma and intracellular SQV levels were measured by high-performance liquid chromatography. Efficacy was evaluated by an intention-to-treat analysis; treatment failure was defined as virological failure (a VL of >50 copies/ml after 24 weeks or a confirmed rebound to >50 copies/ml) or interruption for any reason. A total of 151 patients were included in the study (106 of them either had never received PI or had no previous virological failure on PIs) and could be characterized as follows: previous C3 stage, 28.9%; injection-drug users, 69.1%; subjects with chronic viral hepatitis, 53%; and subjects with cirrhosis, 10%. The median baseline CD4 level was 184/μl, and the median VL was 4.8 log₁₀ copies/ml. Median C_max, area under the concentration-time curve from 0 to 24 h, and C_min plasma and intracellular SQV levels were 3,672 and 10,105 ng/ml, 34,283 and 99,535 ng·h/ml, and 359 and 1,062 ng/ml, respectively. The efficacy as determined by intention to treat at 52 weeks was 69.7% (96% in the on-treatment analysis), with similar results regardless of the baseline VL and CD4 counts. Only five patients had virological failure despite adequate C_min levels, but with a poor adherence (the only variable related to virological failure). Adverse events caused the withdrawal of the treatment in four patients (2.6%). In conclusion, given the pharmacokinetic profile, efficacy, and tolerability of this regimen, once-daily low-dose SQVr may be considered a treatment option in treatment-naive or limited PI-experienced HIV-infected patients, with the additional benefit of being currently the least-expensive PI-based regimen available.

Saquinavir (SQV) was the first protease inhibitor (PI) approved for the treatment of human immunodeficiency virus (HIV) infection. Its low bioavailability and short elimination half-life forced its administration at high doses every 8 h.

The addition of low doses of ritonavir highly improves SQV bioavailability by inhibiting the intestinal P-glycoprotein (P-gp) and first-pass metabolism by the cytochrome P450 enzyme system in the liver and gut wall, although not affecting its elimination rate (14, 18, 20, 23, 28, 29, 32, 35, 36, 37). The first pharmacokinetic study of low-dose ritonavir-boosted SQV (SQVr), dosed as 1,600 mg of SQV and 200 mg of ritonavir (1,600/200)/day, was published in 1999 (24). A few months later it was reported that the SQV C_max and area under the concentration-time curve from 0 to 24 h (AUC_0-24) increased proportionally to the dose, but with similar trough levels írrespective of the dose of SQV administered when different dosages of SQVr ranging from 1,200/100 to 1,800/100 mg/day were compared (22).

SQVr is currently licensed as a twice-daily regimen at a dosage of 1,000/100 mg. The efficacy of this dosage has been proved as part of highly active antiretroviral therapy (HAART), although the required pill burden and dosing frequency are disadvantageous factors for adherence (15, 33). Since nonresistant viruses require lower PI concentrations in plasma to be inhibited (33) and a relationship among dose, exposure, and response to SQV has been demonstrated (18), we decided to evaluate the clinical efficacy of the lowest pharmacokinetically assayed SQVr dosage (1,200/100 mg once daily) in treatment-naive or limited PI-experienced HIV-infected patients on the basis of plasma pharmacokinetic data but also assessing intracellular concentrations (the true effect compartment of SQV).

Provided the efficacy of this lower SQVr dose is confirmed, it would be a profitable option with additional advantages such as less toxicity, a lower pill burden with, probably, a better compliance, and lower cost.

MATERIALS AND METHODS

Study population and design.

From January 2000 to November 2004, every treatment-naive or limited PI-experienced HIV-infected patient older than 18 years treated at the Infectious Diseases Department of Hospital Universitario Virgen del Rocio and initiating a regimen consisting of SQVr at 1,200/100 mg once daily plus two nucleoside or nucleotide reverse transcriptase inhibitors (NRTIs) was enrolled in this observational, prospective, single-arm, open-label study. NRTIs prescribed as part of HAART were selected by the responsible physicians on the basis of previous antiretroviral treatments and/or genotypic resistance testing. No entry restrictions were made regarding plasma HIV-1 RNA, CD4 cell count, illegal drug or methadone use, social circumstances, opportunistic infections at any time, alterations in transaminases, or other laboratory parameters. The only exclusion criteria were history of previous antiretroviral treatments and/or genotypic resistance testing, suggesting a high resistance to SQVr, or the concomitant use of drugs with potentially adverse interactions with SQV pharmacokinetics, such as rifampin. All patients gave informed consent, and the study was approved by the local Ethics Committee. The patients' inclusion was censored in November 2004 to allow a minimum of 12 months of follow-up.

Follow-up, assessments, and endpoints.

Patient assessment was performed at baseline, after the first month on treatment, and every 3 months thereafter and included adverse effects, biochemical profiles, hematologic counts, flow cytometric counts of CD4/μl, and plasma HIV-1-RNA levels (VL) measured by PCR (lower detection limit, 50 copies/ml; Amplicor HIV-1 Monitor test, version 1.0 [Roche Diagnostic Systems]).

The primary clinical endpoint of the study was the percentage of subjects with therapeutic success at month 12. Efficacy data were analyzed by intention-to-treat (noncomplete/missing equals failure), considering treatment failure as either treatment interruption whatever the reason (adverse events [AEs], death, or loss to follow-up) or virological failure, defined as inability to suppress the VL to <50 copies/ml after 24 weeks on treatment or a confirmed VL of >50 copies/ml in patients who had previously achieved a viral suppression to <50 copies/ml or had an undetectable viral load at inclusion. If confirmed, the time of the first measurement meeting the failure criteria was selected as the time of failure. Patients missing two consecutive scheduled visits were considered lost to follow-up.

Secondary outcomes included virological efficacy according to on-treatment (OT) analysis, changes in CD4 cell counts, and incidence of AEs. The changes in serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) from pretreatment levels to the highest level during treatment were categorized via a standardized toxicity-grade scale, modified from that used by the AIDS Clinical Trials Group (1). Patients with pretreatment serum AST and ALT levels within normal range (AST < 35 IU/liter and ALT < 31 IU/liter) were classified according to the changes observed with respect to the upper limit of normal (ULN): grade 0, <1.25 × ULN; grade 1, (1.25 to 2.5) × ULN; grade 2, (2.6 to 5) × ULN; grade 3, (5.1 to 10) × ULN; and grade 4, >10 × ULN. In patients with chronic viral hepatitis or cirrhosis, toxicity was classified according to changes relative to baseline values rather than ULN: grade 0, <1.25 × baseline; grade 1, (1.25 to 2.5) × baseline; grade 2, (2.6 to 3.5) × baseline; grade 3, (3.6 to 5) × baseline; and grade 4, >5 × baseline. Self-reported adherence was investigated at each study visit.

Blood sampling and determination of SQV concentrations.

Patients were instructed to always take SQVr in the morning, after a highly caloric breakfast. Pharmacokinetic data were obtained from 103 unselected patients; for 26 of these patients, we obtained a 24-h pharmacokinetic profile with blood samples collected predose (baseline) and at 2, 4, 6, 8, 12, and 24 h after a supervised drug intake, whereas from each of the remaining 77 patients a single sample was withdrawn at 24 ± 0.5 h postdose (C_min). The pharmacokinetic studies were performed at steady state, once the patients had been on treatment for at least 1 month.

At each time point, blood samples were withdrawn and collected in three 7-ml EDTA-tubes. Plasma was immediately separated from one of the three EDTA-tubes after centrifugation (10 min, 3,000 rpm) and divided into aliquots in two 2-ml Eppendorf tubes. Peripheral blood mononuclear cells (PBMC) were extracted by density cushion centrifugation with Ficoll-Paque according to Boyum's technique (8-10) from the remaining two 7-ml EDTA-tubes. Briefly, 14 ml of whole blood, diluted (1/1 [vol/vol]) in 150 mM phosphate-buffered saline (pH 7.4,), were carefully layered onto a Ficoll gradient and then centrifuged (20 min, 2,000 rpm). PBMC were aspirated with a Pasteur pipette at the resulting interface and washed twice in phosphate-buffered saline before collecting the resulting cell pellets in two 2-ml Eppendorf tubes. PBMC aliquots were weighed, and their volumes were calculated as equal to weight/density. Since the density of mononuclear cells is <1.077 and that of plasma is 1.030 (9, 10), the aliquot's weight was equalized with its volume. Intracellular associated concentrations of SQV were assimilated to intracellular concentrations. Samples were frozen at −70°C until assay.

SQV (RO-31-8959/000 MRS), pure substance, was obtained from Roche Diagnostics GmbH (Mannheim, Germany) and used as an internal standard. Concentrations of SQV were determined by using high-performance liquid chromatography with an UV detector (Gilson Medical Electronics, Inc., Middleton, WI). The analytical column was a Chrompack Inertsil ODS-2 C₁₈ column (5 μm; 150 × 4.6 mm). The volume injected was 40 μl, and the flow rate was 1.5 ml/min. The mobile phase consisted of a mixture of acetonitrile and 50 mM KH₂PO₄ adjusted to pH 5.63 with 50 mM Na₂HPO₄ (50:50 [vol/vol]). Chromatographic analysis was performed at room temperature with isocratic conditions at a wavelength of 215 nm. The assays were found to be linear and were validated over a concentration range of 0.1 to 15 μg/ml. The lower limit of quantification was 40 ng/ml. Recovery of SQV from human plasma was 98.2% ± 2.66%. The mean intra- and interassay coefficients of variation were 3.1 and 5%, respectively. Our laboratory participates in an external quality assurance program (KKGT, The Netherlands).

Pharmacokinetic analysis.

The SQV pharmacokinetic parameters were determined by using WinNonLin program (version 3.1; Pharsight Corp., California) according to a noncompartmental method from plasma and PBMC-associated drug concentration-time data. The terminal log-linear period was determined on the basis of the last two points (12 to 24 h).

Statistical analysis.

Descriptive statistics were used for demographic, epidemiological, and clinical data; prior antiretroviral treatments; CD4 cell count; VL values; and pharmacokinetic variables. The correlations between SQV plasma and intracellular concentrations were assessed by the Spearman's rank correlation coefficients and linear regression. The SQV concentrations in plasma were compared among different groups by using the Mann-Whitney U test and the Kruskal-Wallis test. The relationships between virological failure and different variables were assessed by the chi-square test for qualitative variables and by the Spearman's rank correlation coefficients for quantitative variables. The variables tested by univariate analysis as predictors of virological failure were baseline VL below or above 100.000 copies/ml, baseline CD4 counts below or above 200/μl, previous treatment failure on PIs, plasma and intracellular SQV C_min levels, drug addiction, and adherence. All variables with a P value of <0.1 in the univariate analysis (baseline VL, baseline CD4 count, previous treatment failure on PIs, and adherence) were included in the multivariate analysis. The results were expressed as median (range). Statistical calculations were performed with the Statistical Product and Service Solutions for Windows (12.0 version; SPSS, Chicago, IL).

(This study was presented in part at the 10th European AIDS Conference/EACS, Dublin, Ireland, 17 to 20 November 2005, abstr. PE4.1/5.)

RESULTS

Baseline patient characteristics.

A total of 151 patients were included in the study (120 male, 31 female). The baseline characteristics are summarized in Table 1. Among the 35 PI-experienced patients with previous virological failure while on PI, 9 had a genotypic resistance test available, in which mutations at the protease gene were detected in 4 cases, but with none of them conferring a reduced susceptibility to SQV. No genotypic resistance tests were available in the remaining 26 patients due to several reasons: the technique was not ready for use at the moment the patients were included, amplification was not possible in cases with a VL of <1,000 copies/ml, some patients had been without any treatment for a long period and so the test was not expected to add relevant data, and in some cases the test had not been requested. Forty-eight patients had also received non-NRTIs, and eighteen of them demonstrated previous virological failure in response to this group of drugs.

TABLE 1.

Patient characteristics at inclusion

Patient parameter	Value
No. of patients	151
Median age in yr (range)	37 (20-70)
No. male (%)	120 (79.6)
Median wt in kg (range)	65 (40-101)
No. (%) of patients with risk factor for HIV
Intravenous drug use	104 (68.9)
Hetero/homosexual contacts	44 (29.1)
Blood product transfusion	2 (1.3)
Unknown	1 (0.7)
Median CD4/μl (range)	186 (1-988)
Median HIV-RNA log10 copies/ml (range)	4.8 (<1.69-6.50)
No. (%) of subjects with:
Previous C stage (CDC)	57 (37.8)
Chronic viral hepatitis	80 (53)
Cirrhosis	15 (10)
No. (%) of patients with:
No prior treatment (naive)	33 (21.8)
PI unexperienced	25 (16.5)
PI experienced/previous failure on PIs	83 (54.9)/35 (23.2)
Median total cholesterol level in mg/dl (range)	157 (54-306)
Median triglyceride level in mg/dl (range)	124 (45-1,191)

Eighty-nine patients started this regimen with SQV hard gel capsule (SQV-hgc), while 62 patients started with SQV soft gel capsule (SQV-sgc) but switched to SQV-hgc later on. The combinations of NRTIs used with SQVr consisted of didanosine and either lamivudine (72 patients), tenofovir (12 patients), or abacavir (1 patient) and thymidine analogues (zidovudine or stavudine) together with lamivudine (34 patients) or didanosine (23 patients). Other NRTI combinations were used in nine patients. The median follow-up was 21 months (range, 2 to 66 months).

Plasma and intracellular pharmacokinetics of SQV.

The pharmacokinetic study was carried out in 102 unselected patients. In 26 of these patients (21 male, 5 female; median weight, 67 kg [range, 48 to 89 kg]) a 24-h profile was determined in both plasma and intracellular samples. The remaining 76 patients (61 male, 15 female; median weight, 64 kg [range, 48 to 101 kg]) were only assessed for plasma and intracellular trough levels. Median and interquartile range SQV plasma and intracellular concentration-versus-time curves are represented in Fig. 1, and pharmacokinetic data are summarized in Table 2. The intracellular SQV levels were higher than those in plasma, resulting in median intracellular drug accumulation ratios of 2.75 (range, 0.98 to 7.17), 2.75 (0.9 to 6.15), and 3.01 (0.5 to 24) for C_max, AUC_0-24, and C_min, respectively (Table 2). A significant relationship was found between plasma and intracellular values for both C_max and C_min (r = 0.231 [P = 0.01] and r = 0.429 [P = 0.03], respectively), whereas the relationship between both compartments regarding the total SQV exposure (AUC_0-24) showed just a borderline significance (r = 0.349; P = 0.07). The median t_1/2 (β) values of SQV in plasma and cells were 8.2 h (2.4 to 20.7 h) and 12.5 h (3.5 to 37 h), respectively. No significant differences in SQV levels were found regarding gender, weight, or the presence of chronic hepatitis and/or cirrhosis.

FIG. 1.

Median and interquartile range SQV plasma and intracellular concentration-versus-time curves after administration of 1,200 mg in combination with 100 mg of ritonavir.

TABLE 2.

Pharmacokinetic data for SQV in plasma and cellular compartment^a

SQV	Median and CV pharmacokinetic data
	Cmax		Cmin		AUC0-24		t1/2 β
	ng/ml (range)	CV (%)	ng/ml (range)	CV (%)	ng·h/ml (range)	CV (%)	h (range)	CV (%)
Plasma
n = 26	3,672 (999-9,832)	49.7	431 (62-848)	60.7	34,283 (12,645-70,389)	42.3	8.25 (2.4-20.7)	49.1
n = 102			350 (62-1,119)	65.2
Intracellular
n = 26	10,105 (2,212-19,946)	48.6	996 (401-5,730)	90.0	99,535 (26,595-283,824)	57.0	12.5 (3.5-37)	62.9
n = 102			1,060 (358-6,714)	87.4
Ic/P ratio
n = 26	2.75 (0.98-7.15)	57.5	2.39 (0.5-24)	124.6	2.75 (0.9-6.7)	52.5	1.5 (0.5-6.5)	98.0
n = 102			3.01 (0.5-24)	93.3

^aA 24-h pharmacokinetic profile was obtained from each of 26 patients, and only trough levels were obtained from 72 additional patients. The Ic/P is the ratio of the intracellular to the plasma concentration. CV, coefficient of variation. n, No. of patients.

Efficacy and safety.

At 52 weeks, the efficacy was 69.5% (95% confidence interval [CI₉₅] = 65.5 to 73.5%) by ITT and 96% (CI₉₅ = 92.5 to 99.5%) by OT analysis, with similar efficacy when patients were stratified according to baseline VL levels (above and below 100,000 copies/ml) and CD4 counts (above and below 200/μl) (Fig. 2). Forty-six patients (30.4%) failed treatment before completing 12 months of follow-up because of treatment dropout in 15 patients (9.9%), loss to follow-up in 12 patients (7.9%), AEs in 4 patients, death in 3 patients (due to cirrhosis, bacterial pneumonia, and an unknown reason in a patient with Buerguer's disease), hepatitis A in 1 patient, a suicide attempt in 1 patient, simplification in 1 patient, and social circumstances (imprisonment, move, and lack of social support) in 4 patients. Virological failure was the cause of treatment failure in only five patients (3.3%) who had received HAART during a period ranging from 7 to 69 weeks before inclusion and were PI experienced with previous failure on nelfinavir in two cases and indinavir in one case. Two of them were considered failure at 12 weeks since VL did not decrease ≥1 log with respect to baseline, while the other three showed an initial response but had not achieved a VL of <50 copies/ml after 24 weeks of treatment. The SQV C_min was available in four of these five patients, with values ranging from 154 to 840 ng/ml, and a 24-h profile was available for three of them (Table 3). Previous dropouts were common in three of these cases, and irregular adherence was self-reported at each study visit in all five patients. Genotypic resistance testing could be obtained in four of them: one patient showed failure with a wild-type virus, one patient had only NRTI-associated mutations, and the remaining two patients had isolates showing the following mutations: D67N, M184V, T215F, L10I, G48V, I54L, L63P, A71T, V82A, and L90M in the first case and M184V, T74P, F53L, L63P, and L90M in the second case.

FIG. 2.

Proportion of patients without protocol-defined treatment failure through week 104 (ITT) in the overall cohort (A) and according to baseline HIV-RNA (B) and CD4 cell count (C).

TABLE 3.

Pharmacokinetic data for SQV in plasma and cellular compartments for the three patients with virological failure in whom a complete 24-h profile was available (n = 3)^a

SQV	Median (range)
	Cmax (ng/ml)	Cmin (ng/ml)	AUC0-24 (ng·h/ml)	t1/2 β (h)
Plasma	3,320 (2,840-3,511)	383.95 (154-848)	35,308 (25,254-25,254)	8.51 (4.9-17.3)
Intracellular	15,198.50 (14,016-16,381)	1,217.30 (606-5,730)	132,158.00 (116,478-147,838)	10.909 (6.3-15.6)
Ic/P ratio	4.46 (3.99-4.93)	3.014 (1.4-10.1)	3.53 (3.3-3.8)	1.5 (0.5-6.5)

^aThe Ic/P ratio is the ratio of the intracellular to the plasma concentration.

Treatment with SQVr was followed by a rapid reduction in VL, exceeding 2 and 3 loga at 1 and 3 months, respectively. A progressive increase in CD4 counts was also evident after initiating treatment, with a median raise of 130 at 6 months and 179 at 12 months (Fig. 3). The variables associated with virological failure in the univariate analysis were baseline VL values of >100,000 copies/ml, baseline CD4 counts of <200/μl, previous treatment failure on PIs, and nonadherence. No relationship was found between virological outcome and either plasma or intracellular SQV levels. The only variable associated with virological failure in the multivariate analysis was adherence (odds ratio = 47.31 [CI₉₅ = 5.40 to 413.8]). Seventy-one patients remained on follow-up at week 104, and the estimated probability of continuing on treatment with undetectable VL was 56.1% (CI₉₅ = 48.1 to 64.1). Only two patients presented virological failure during the second year.

FIG. 3.

Median changes and IQR in plasma HIV RNA and CD4 cell counts from baseline to month 12.

The most frequently reported AEs were nausea, vomiting, and/or abdominal discomfort (grade 1 to 2) in eight patients (5.3%), grade 1 to 2 diarrhea in seven patients (4.6%), and epigastralgia and pyrosis in five patients (3.3%) and 1 patient (0.6%), respectively. Lipodystrophy appeared in five cases (3.3%). These AEs caused the withdrawal of the treatment in four patients (2.6%). No symptoms of methadone withdrawal were observed. No signs or symptoms of hematological toxicity were recorded in any patient. The levels of glucose, urea, and creatinine in plasma did not significantly differ from baseline to month 12 of follow-up. Eight patients had liver-associated toxicity as defined above, being grade 3 or 4 in only two cases. None of these patients had symptoms of acute liver disease, and the alterations observed in the laboratory hepatic parameters were transient and improved without treatment discontinuation in every case. The median increase in total cholesterol was 17 mg/dl (range, −110 to 121) after 52 weeks on treatment, while fasting triglycerides levels remained stable during the whole study, with a median change of −3 mg/dl (range, −272 to 509).

DISCUSSION

Although once-daily dosing has not been approved yet for SQVr, different once-daily dosing schemes have been studied in recent years, both in healthy volunteers and in HIV-infected patients, with SQVr at 1,600/100 mg/day being the most frequently assessed regimen. Other than one earlier report of ours on this SQVr dose (27), the present study is the first study to date to demonstrate the clinical efficacy of the 1,200/100 mg once-daily dosage together with two NRTIs.

Our study's starting point was the pharmacokinetic study published by Kilby et al. in 2000 on different once-daily doses of SQVr, in which the values of C_max and AUC_0-24 observed with doses of 1,600/100 mg were higher than those with 1,200/100 mg. However, trough levels (C_min) were similar with both doses (median of 484 ng/ml [range, 201 to 1,080 ng/ml] versus a median of 490 ng/ml [range, 171 to 1,230 ng/ml]), and thus a similar efficacy would be expected (with the lower dose), but requiring a lower pill burden with 1,200/100 mg daily (22). As expected, this 1,200/100-mg SQVr regimen has yielded C_max and AUC_0-24 values that are considerably lower than those observed with higher doses of once-daily SQVr, whereas the C_min values are within the range of previously reported levels (6, 11, 13, 21, 22, 24, 27, 34), being higher than the IC₉₅ value (25 ng/ml) for wild HIV-1 isolates and the estimated trough level required to obtain the half-maximal antiviral response (50% effective concentration = 50 ng/ml). In addition, the SQV AUC_0-24 values in our patients were >20,000 ng/ml·h, the cutoff at which 85% of the maximum antiviral effect is expected in patients without protease gene mutations (17, 18). The observed interindividual variability in SQV concentrations, both in plasma and intracellularly, was considerable, although within the range of other studies on SQV and other PIs (3, 4, 5, 6, 7, 11, 13, 17, 19, 21, 25, 26, 30, 31, 34).

The median SQV intracellular/plasma concentration ratios were 2.75 for C_max and AUC_0-24 and 3 for C_min, which could be contributing to the clinical efficacy observed in our series, with a more prolonged t_1/2 in cells than in plasma. All of these values are concurrent with previously reported data (17, 26) and suggest that intracellular SQV levels may be sufficient even when concentrations in plasma are below the minimum effective concentration. No significant differences could be found between plasma SQV levels and gender, weight, or the presence of hepatitis and/or cirrhosis, except for a slight trend toward higher SQV C_min levels in cirrhotic patients.

The efficacy in the ITT analysis in our series (69.5%) is lower than that reported with the 1,600/100-mg once-daily regimen and similar to that observed in the MaxCmin2 trial with a 1,000/100 twice-daily dose (2, 12, 13, 16, 26). However, only five patients (3.3%) had virological failure despite adequate C_min levels but with a confirmed poor adherence. In fact, adherence was the only variable related to virological failure in the multivariate analysis. In the OT analysis, the efficacy increased to 96%, with similar results independent of the baseline VL and CD4 counts.

It must be taken into account that the reasons for treatment failure in 80% of our patients (37 of 46) were dropout, loss to follow-up, or other causes not related to the antiretroviral treatment itself but mainly due to the fact that the only entry restriction in our study was previous resistance to SQV. This situation mirrors the usual clinical practice setting and is very different from a clinical trial in which patients are carefully selected.

The lack of a relationship between SQV C_min and its long-term efficacy (Tables 2 and 3), despite the wide range of concentrations observed, suggests that the levels observed were in the plateau portion of the dose-response curve. Thus, the observed plasma and intracellular SQV levels support the efficacy of low-dose SQVr (1,200/100 mg) given once daily in treatment-naive or limited PI-experienced HIV-infected patients. Moreover, the tolerance of this regimen was excellent, with no grade 3 or 4 AEs, a low incidence of hepatotoxicity, no interactions with methadone, and only 2.6% patients interrupting treatment due to light or moderate AEs. All of these factors are worth considering in a population such as ours, with a high rate of hepatitis C virus coinfection and concomitant use of methadone.

The pharmacokinetic and clinical results of our study demonstrate that SQVr (1,200/100 mg) given once daily plus two NRTIs is an optimal treatment scheme in patients with no SQV resistance-associated protease mutations, with the additional benefit of being the lease-expensive PI regimen currently available.

Since a new 500-mg formulation of SQV is already available, carrying a significant reduction in the daily pill burden that will probably favor a better adherence, a regimen of SQVr (1,500/100) once daily would be expected to have a similar virological efficacy without adding much toxicity, although this is an issue that needs to be addressed in further studies.