Treatment response and durability of a double protease inhibitor therapy with saquinavir and ritonavir in an observational cohort of HIV-1-infected individuals

Objective: To evaluate treatment response, durability and tolerance of a four-drug regimen including saquinavir and ritonavir in combination with either zidovudine/lamivudine or stavudine/lamivudine.

Design: Observational cohort of HIV-positive individuals.

Methods: Viral load, CD4+ and CD8+ T lymphocyte counts were assessed at intervals of 1-3 months in subjects commencing therapy between July 1996 and November 1996. Adverse events were evaluated as well as risk factors for therapeutic failures.

Results: A group of 56 male patients were included and followed for 48 weeks. Of these, 66% had already taken a protease inhibitor. Viral load dropped by a median 1.98 log₁₀ HIV RNA copies/ml from baseline (interquartile range: 1.49-2.46) and became undetectable (< 400 copies/ml) in 68% of patients. Response varied: 9% were non-responders (HIV RNA reduction < 0.5 log₁₀ copies/ml) and 23% were incomplete responders (nadir of HIV RNA > 400 copies/ml). After 48 weeks, viral load remained undetectable in 49%. Median CD4+ T lymphocyte count increased from 191 × 10⁶ to 418 × 10⁶ cells/l (range, 241-537 × 10⁶ cells/l). Although protease inhibitor and nucleoside pretreatment selected for drug-resistant viral mutants, only the protease inhibitor experience was identified as a risk factor for therapeutic failure. Adverse events occurred in 73% of patients and led to a change of therapy in 9%.

Conclusion: Despite advanced HIV disease and pretreatment with multiple antiretroviral drugs, a strong initial treatment response to this drug regimen was observed. However, virological failure occurred in 51% of patients after 48 weeks and frequent adverse events complicated therapy. © 1998 Lippincott Williams & Wilkins

Introduction

Recent investigations of combination antiretroviral therapy in patients infected with HIV-1 showed that a protease inhibitor in combination with two reverse transcriptase inhibitors was superior to treatment regimens consisting of only two nucleoside analogues or one protease inhibitor and one reverse transciptase inhibitor [^1-3]. Despite the initial substantial reduction of plasma viraemia by those drug regimens, the treatment effect is frequently only transient. A rapid emergence of drug resistance appears to be responsible for treatment failures and limits the durability of the therapeutic response [^4-11]. One strategy to reduce the risk of drug resistance is the concurrent application of two agents that inhibit the same step in the viral life cycle. The combination of ritonavir and saquinavir appears to be favourable because of the absence of significant cross-resistance [^12,13]. Moreover, pharmacological interactions between ritonavir and saquinavir alter the elimination kinetics of saquinavir and increase maximum saquinavir plasma concentrations up to 30-fold, which may also improve therapeutic efficacy [¹⁴]. Higher protease inhibitor concentrations also have been shown to reduce the risk of drug resistance [⁶].

We analysed treatment response, durability and safety of a four-drug regimen including the two protease inhibitors saquinavir and ritonavir in combination with two different regimens of two reverse transcriptase inhibitors in an observational cohort of HIV-1-positive patients. Moreover, we sought to define risk factors for treatment failures. Most patients were, in contrast to other studies, pretreated with multiple nucleoside regimens and two-thirds were protease inhibitor experienced.

Materials and methods

Study design and patients

All HIV-1-positive patients starting therapy with saquinavir, ritonavir and either zidovudine and lamivudine or stavudine and lamivudine between July and November 1996 in the collaborating centres were considered for this analysis. Data were prospectively collected in an observational database. Patients were enrolled independently of their Centers for Disease Control and Prevention (CDC) stages or pretreatments with nucleosides or protease inhibitors. They were not eligible if they stopped treatment within the first treatment month because of adverse events (n = 3) or if baseline values of viral load or CD4+ cell counts were missing (n = 6).

Study medication

The nucleoside analogue regimen consisted either of a combination of zidovudine 250 mg plus lamivudine 150 mg or stavudine 40 mg plus lamivudine 150 mg twice daily. The dose of ritonavir was 600 mg and saquinavir 400 mg, both twice daily. If patients were unable to tolerate the full dose of ritonavir, it was reduced to 400 mg twice daily and the saquinavir dose raised to 600 mg twice daily.

Laboratory assessment

Plasma samples were assessed at intervals of 1-3 months. Viral load was measured by the Amplicor HIV Monitor test (Roche Molecular Systems, Branchburg, New Jersey, USA). Circulating CD4+ and CD8+ T lymphocyte counts were determined by three-colour flow cytometry using CD45PerCP, CD3-FITC, CD4-PE and CD8-PE monoclonal antibodies (Becton-Dickinson, San Jose, California, USA), whole blood lysis (FACS lysing solution; Becton-Dickinson) and analysis on EPICS XL flow cytometer (Coulter Electronics, Hialeah, Florida, USA) in accordance with the CDC guidelines [¹⁵].

Evaluation

Treatment success was analysed according to viral load changes. Responders showed a viral load reduction of at least 0.5 log₁₀ HIV RNA copies/ml, whereas viral load in non-responders remained unchanged by treatment. Responders were grouped into complete and incomplete responders depending on the magnitude of viral load reduction. Undetectable values (< 400 RNA copies/ml) were observed in complete responders, whereas nadir > 400 RNA copies/ml defined the group of incomplete responders. A new viral load increase during the therapy of > 1 log₁₀ RNA copies/ml from nadir was defined as recurrence and, therefore, as therapeutic failure. Additional parameters of treatment efficacy were CD4+ T lymphocyte changes and the occurrence of HIV-1 related events.

Risk factors for therapeutic failures

The following potential risk factors for treatment failures were analysed: CDC classification, baseline viral load, baseline CD4+ and CD8+ T lymphocyte count, concurrent nucleoside regimens and the protease inhibitor or nucleoside pretreatment as well as the baseline resistance to protease inhibitors and nucleoside analogues.

Resistance genotyping of HIV protease and reverse transcriptase genes

Viral RNA isolated from plasma underwent reverse transcription (RT) and amplification by a nested polymerase chain reaction (PCR). Subsequent DNA templates were used for sequencing. The resultant PCR product was cleaned to remove excess nucleotide bases, enzyme and buffer components, and a dye-primer cycle sequencing reaction was performed. The product was electrophoresed on an ABI Prism (Perkin Elmer Corp., Foster City, California, USA) DNA sequencer to determine the DNA sequence. The sequence was then edited as both DNA and amino acid sequence compared with a standard sequence, with annotation to identify sites and consensus amino acid changes that result in resistance.

Assessment of adverse events

Patients were seen every 4 weeks for a clinic visit after commencing the double protease therapy. Standard haematology and blood chemistry tests including those for liver enzymes were performed on at least 3-monthly intervals. Fasting triglyceride and cholesterol levels were measured in 41 patients after 6 months of treatment.

Statistical analysis

Demographic parameters are summarized as means and standard deviations (SD). CD4+ and CD8+ T lymphocyte counts as well as log-transformed plasma HIV RNA are shown in medians and interquartile ranges. The correlation between potential risk factors for therapeutic failures, such as the CDC stage and the prior drug experience, was tested by a chi-square test. A Mann-Whitney-U test was used to detect associations between baseline CD4+ and CD8+ T lymphocyte counts, plasma HIV RNA and the treatment success. A two-sided P value < 0.05 was considered statistically significant. Statistical tests were performed with the SPSS for Windows software package V 6.1 (SPSS, Chicago, Illinois, USA).

Results

The study included 56 male HIV-positive patients of which 14% were classified CDC stage A, 43% stage B and 43% stage C. Forty patients (71%) continued therapy for at least 48 weeks. Sixteen individuals discontinued therapy, 11 owing to treatment failures, three because of adverse events and two for both reasons. Two patients with treatment failure had disease progression and died during the observation period. One patient developed non-Hodgkin's lymphoma and another one tuberculosis. Demographic data and laboratory values at baseline are shown in Table 1.

Table 1

Sixteen patients (29%) received a nucleoside regimen including zidovudine. A higher number of patients (40, 71%) were treated with stavudine and lamivudine because of their prior extensive zidovudine experience. No significant differences of baseline parameters were noted between patients on the two different nucleoside regimens. Full-dose ritonavir (600 mg twice daily) was only tolerated by 33% and the others received the low-dose ritonavir regimen (400 mg twice daily).

Thirty-four per cent of subjects had not taken a protease inhibitor previously, of which 9% were completely treatment naive. Of the 66% of subjects who had taken a protease inhibitor previously, 50% were pretreated with saquinavir, 14% with ritonavir and 2% with indinavir. The median duration of the protease inhibitor pretreatment was 22 weeks (interquartile range, 13-26). In patients who were pre-treated with saquinavir or ritonavir, the second protease inhibitor was added to increase antiretroviral efficacy. One patient pretreated with indinavir was switched to saquinavir and ritonavir because of indinavir intolerance. During an average of 9 years after seroconversion, patients have been pretreated with multiple nucleosides for several months to years (Table 1).

Evaluation of the therapeutic response

Viral load dropped from baseline values by a median of 1.98 log₁₀ RNA copies (interquartile range, 1.49-2.46) (Fig. 1). CD4+ lymphocytes increased from the baseline value of 191 × 10⁶ cells/l to 418 × 10⁶ cells/l (range, 241-537 × 10⁶ cells/l). Between week 0 and week 24, CD4+ T lymphocyte counts increased by 125 × 10⁶ cells/l and between weeks 24 and 48 by an additional 102 × 10⁶ cells/l. Initial CD4+ and CD8+ T lymphocyte counts were not associated with the degree of CD4+ cell increase. CD8+ T lymphocyte counts increased rapidly from already elevated baseline of 906 × 10⁶ cells/l (range, 650-1256 × 10⁶ cells/l) to 1265 × 10⁶ cells/l (range, 1008-1565 × 10⁶ cells/l) at 48 weeks.

Fig. 1

In all, 91% of patients had a significant viral load reduction (> 0.5 log₁₀ copies/ml), of which 68% reached undetectable values (Fig. 2). Of those patients, 50% achieved undetectable values within 12 weeks and 18% after that time. A recurrence of viral load of > 1 log₁₀ RNA copies/ml was observed in 16 patients (29%); in 11 this occurred within 6 months. Nine per cent of patients were non-responders, with no significant changes of viral load at all. After 48 weeks, viral load was still undetectable in 49% of patients.

Fig. 2

In patients with incomplete treatment responses (23%), viral load showed a trend to return to baseline values (P = 0.06, chi-square test). No significant differences in treatment response were observed between the two nucleoside groups and the low- and high-dose ritonavir group as measured by viral load change and CD4+ and CD8+ T lymphocyte changes.

Risk factors

Treatment failures were less frequent among treatment-naive and nucleoside-pretreated patients compared with protease inhibitor-experienced patients at 24 weeks (15% versus 43%, P = 0.04, chi-square). However, after 48 weeks no significant difference between the two groups could be detected anymore (37% versus 59%, P = 0.12, chi-square). The analysis of each protease inhibitor group revealed an increasing failure rate in ritonavir-pretreated patients from week 24 to week 48 compared with protease inhibitor-naive patients (43% versus 15% at week 24, P > 0.05, chi-square; 86% versus 37% at week 48, P = 0.03, chi-square). Although the failure rate of saquinavir-pre-treated patients was also higher than in protease inhibitor-naive patients, the difference was only statistically significant at week 24 (43% versus 15% at week 24, P = 0.04, chi-square; 54% versus 37% at week 48; P > 0.05, chi-square).

The sequence of the HIV protease gene was available in 17 patients, of whom eight subjects were pretreated with saquinavir and two with ritonavir. Of these, 16 patients had virus with at least one point mutation. Mutations were found with the following frequencies: ^L63^P in 14 patients, ^A71^V/T in seven, ^L10^I in five, ^L90^M in five, ^V82^A in three, ^I54^V in two, and ^M36^I and ^G48^V in one patient. Multiple mutations were detected in 10 individuals, of whom eight were pretreated with pro-tease inhibitor. Each of the three non-responders showed multiple mutations (^G48^V/^L63^P/^V71^I/^V82^A, ^L10^I/^L63^P and ^L10^I/^A71^T/^V77^I). Three of the eight saquinavir-experienced patients had mutation ^L90^Mand both ritonavir-pretreated subjects had mutation ^V82^A. Therapy failed in all patients with those mutations. In contrast, mutations at position 63 and 71 were found in five individuals who were successfully treated during a period of 48 weeks.

Baseline viral load, CD4+ or CD8+ T lymphocyte counts and CDC classifications were not correlated with treatment failures. Also the prior nucleoside analogue experience was not a significant risk factor (Fig. 3). An analysis of RT gene drug-resistance mutations was performed in 40 patients from plasma samples stored at baseline. The frequency of mutations at position 41 and 215 conferring resistance to zidovudine were 40% and 57%, respectively. Mutations at position 184 and 75 associated with lamivudine and stavudine resistance were detected in 59% and 2.5%, respectively. Pre-existing RT gene drug-resistance mutations were not associated with treatment failure.

Fig. 3

Adverse events

Symptoms and signs of adverse events were frequent, occurring in 73% of patients (Table 2): 48% of patients suffered from diarrhoea and 21% from nausea. Gastrointestinal adverse events were treated sympto matically and antiretroviral therapy continued whenever possible. Nevertheless, five patients stopped treatment as a result of severe diarrhoea. Less frequently, lipodystrophy (n = 17, 30%) and a transient rash (n = 2) were observed. Lipodystrophy occurred usually after several months of treatment and was clinically detected by a loss of fatty tissue from extremities and face. No change of weight from baseline to 48 weeks was observed, with a median weight of 73 kg (range, 68-79). However, in some patients, a transient increase of weight in the first few months after commencing the treatment was noted.

Table 2

In 96% of patients, at least one liver enzyme was increased, but elevations were usually mild. Ranges observed were alanine aminotransferase (ALT) 31-286 U/l (normal value, < 30 U/l), alkaline phosphatase 101-207 U/l (normal value, < 100 U/l), gamma-glutamyl transferase 36-321 U/l (normal value, < 35 U/l). Triglycerides were elevated in 85% of patients (range, 2.1-16 mmol/l; normal value, 2.0 mmol/l) and cholesterol in 68% of patients (range, 5.6-12.7 mmol/l; normal value, 5.5 mmol/l). In 16% of patients, increased creatinine phosphokinase levels were seen (range, 140-358 U/l; normal value, 120 U/l).

Discussion

Protease inhibitors have improved considerably the efficacy of HIV therapy [^1-3,16,17]. Nevertheless, the rapid emergence of drug resistance to all available antiviral agents remains a serious problem [^4-12]. This study indicates that despite a four-drug regimen inhibiting two different steps in the viral life cycle, only 50% of patients have a durable treatment response after 48 weeks.

We identified pretreatment with protease inhibitors as the only significant risk factor for therapeutic failures. Patients with saquinavir experience demonstrated a poorer initial treatment response. In contrast, ritonavir-experienced patients failed more frequently at 48 weeks, which suggests that adding saquinavir to a ritonavir therapy may be less favourable than adding ritonavir to a saquinavir therapy.

Our observation of protease mutation ^L90^M in three of eight saquinavir-pretreated patients and mutation ^V82^A in ritonavir-pretreated patients indicates that previous therapy with one protease inhibitor selected for drug-resistant viral mutants. The pre-existing drug resistance may be the reason for the observed higher failure rate at 24 weeks in protease inhibitor-pretreated patients.

Many patients were switched from a previous zidovudine-containing regimen to stavudine/lamivudine because of their extensive zidovudine experience. It was not surprising that 57% of virus isolates in those patients were resistant to zidovudine. The high percentage of lamivudine-pretreated patients resulted in a similar frequency of lamivudine-resistant virus. However, there was no correlation of the nucleoside resistance at baseline with the treatment response, suggesting that the protease inhibitor regimen may be more crucial for a favourable treatment response than the nucleoside regimen. Consequently, a similar treatment response for stavudine/lamivudine and zidovudine/lamivudine was noted.

In 23% of our patients, a substantial reduction of viral load occurred, but this did not reach undetectable values. These patients were shown to have a trend towards a higher risk of viral load increase, which indicates that complete viral suppression should be the first therapeutic aim.

A high rate of drug intolerance was observed, leading to a change of treatment in 9% of patients. The inhibition of the hepatic metabolism of saquinavir by ritonavir can lead to a high saquinavir concentration [¹⁴], which may be associated with more adverse events. Moreover, a high subject-to-subject variability of serum saquinavir concentrations, when saquinavir is combined with ritonavir, makes drug concentrations unpredictable and would argue for routine assessment of plasma drug concentrations, especially in patients receiving combination therapy [¹⁴]. Individually adjusting the dose schedule may help to reduce side effects [¹⁸]. Of special concern was the high rate of lipodystrophy, which has been recently recognized as an important adverse event associated with protease inhibitor therapy. It commonly occurs after 9-12 months of treatment [^19-21]. The changes in fat metabolism causing lipodystrophy may also be responsible for the elevated cholesterol and triglyceride levels, which could lead to vascular problems during a long-term protease inhibitor therapy.

In conclusion, a drug regimen consisting of saquinavir and ritonavir with two nucleosides showed a durable effect in 50% of patients. Drug resistance and the high rate of adverse events, especially lipodystrophy, may limit the long-term application of this drug regimen.

Acknowledgements

The study was presented in part at the 37th Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC) in Toronto 1997 and the 5th Conference on Retroviruses and Opportunistic Infections in Chicago 1998.