The antiviral effect of ritonavir and saquinavir incombination amongst HIV-infected adults: results from a community-based study

Objective: To characterize the antiviral effect and predictors of response to ritonavir and saquinavir-based antiretroviral combination therapy.

Design: Intent-to-treat analysis with suppression of plasma viral load to levels below 2.7 log¹⁰ copies/ml as the main outcome measure.

Patients: All adult HIV-positive individuals in the province of British Columbia who started taking ritonavir and saquinavir (each at 600 mg twice daily) in combination from 1 September 1996 to 28 February 1997, with a minimum of two plasma viral load measurements, one at baseline and one after the initiation of therapy.

Results: A total of 58 participants were prescribed ritonavir and saquinavir. The median plasma viral load at entry was 4.80 log₁₀ copies/ml (interquartile range, 4.51-5.15 log₁₀ copies/ml). A total of 29 (50%) subjects demonstrated a decrease in plasma viral load to levels below 2.7 log₁₀ copies/ml. This level of suppression was associated with higher baseline CD4 cell counts (P = 0.022) and no prior exposure to protease inhibitors (P = 0.001). After controlling for baseline CD4 cell count and plasma viral load, participants naive to protease inhibitors were almost seven times (odds ratio, 6.99; 95% confidence interval, 1.85-26.39; P = 0.004) more likely to suppress their plasma viral load to below 2.7 log₁₀ copies/ml than those who had previously used protease inhibitors.

Conclusion: Our analysis demonstrates that a ritonavir and saquinavir-based combination can produce a substantial decrease in plasma viral load with half of the participants decreasing their plasma viral load to below the limit of quantification of the assay. This response, however, is seriously compromised by prior exposure to protease inhibitors.

Introduction

The introduction of protease inhibitors to the anti-retroviral armamentarium has produced dramatic results in both surrogate marker and clinical endpoint trials [^1-5]. Ritonavir (RTV) and saquinavir (SQV) both demonstrate powerful antiretroviral effects in vitro [^6,7]. The potent effect of RTV on plasma viral load, CD4 cell counts and clinical endpoints has been verified in clinical trials [^3-5,8]. In contrast, SQV's poor bioavail-ability has limited the magnitude and durability of its antiretroviral effect [⁹]. Pharmacokinetic data indicate that RTV therapy can substantially increase SQV serum levels due to RTV's inhibition of the cytochrome P450-3A metabolic pathway [^10-12]. Preliminary reports have suggested that the RTV and SQV combination can produce a profound suppression of viral replication and substantial increases in CD4 cell counts [^11,13-16]. The beneficial pharmacokinetic interaction, the lack of overlapping toxicities, and varying mechanisms of action makes this a potentially attractive combination. There is, however, very little available data on either the safety or the efficacy of two protease inhibitor combinations. We have examined our experience with RTV and SQV-based combinations for antiretroviral experienced, HIV-positive individuals in the community setting.

Methods

Under the mandate of the British Columbia Ministry of Health, the British Columbia Centre for Excellence in HIV/AIDS (referred to here as the Centre) has developed a province-wide program, referred to as the HIV/AIDS Drug Treatment Program, to distribute antiretroviral agents at no cost to eligible HIV-infected individuals. Three protease inhibitors, SQV, indinavir and RTV, are currently available to program participants.

A physician enrolling an HIV-positive individual into the Centre's HIV/AIDS Drug Treatment Program must complete a drug request enrollment form. The enrollment request form acts as a legal prescription and compiles information on the HIV-positive applicant's address and enrolling physician, past HIV-specific drug history, CD4 cell counts, current drug request, and, since June 1996, plasma viral load determinations. Each request was reviewed by the Centre's consultants to ensure that it met the Centre's established therapeutic guidelines [¹⁷]. Approved prescriptions were renewed every 2 months. At the time of the initial refill a participant was asked to complete an enrollment survey and program consent form, while their physician was asked to complete a clinical staging form. Participant surveys and clinical staging forms were completed annually. The clinical staging form recorded participant-specific information on HIV/AIDS-related conditions according to the World Health Organization clinical staging system [¹⁸]. The diagnosis of AIDS was defined according to the 1987 Centers for Disease Control and Prevention revised case definition, which did not incorporate CD4 cell counts [¹⁹].

The Centre distributed antiretroviral drugs based on specific guidelines generated by the Therapeutic Guidelines Committee. Since June 1996, the Centre's therapeutic guidelines have recommended that anti-retroviral therapy-naive individuals with plasma viral loads over 5.0 log₁₀ copies/ml receive three drugs, whereas those with plasma viral loads from 3.7 log₁₀ to 5.0 log₁₀ copies/ml should receive two drugs. A change to a new triple drug regimen was recommended if the plasma viral load was ≥ 4.3 log₁₀ copies/ml whilst on therapy. Plasma viral load monitoring was encouraged 1 month after starting or changing therapy and quarterly thereafter. These recommendations have been in place since June 1996. Plasma viral load (Amplicor; Roche Molecular Systems, Nutley, New Jersey, USA) monitoring was offered free of charge to all program participants.

Our analysis was based on HIV-positive adult men and women in the province of British Columbia who started taking combination therapy including RTV and SQV (600 mg twice daily of each) during the period 1 September 1996 to 28 February 1997. Only subjects with a minimum of two plasma viral load measurements, one at baseline and one following initiation of therapy, were included in the efficacy analysis. Baseline clinical information, including primary AIDS diagnosis, plasma viral load determination, CD4 cell count, gender, and age, as well as previous experience with antiretroviral therapy was obtained directly from the HIV/AIDS Drug Treatment Program records. Data on side-effects and causes of early termination were gathered through chart review. The primary outcome in this analysis was a decrease in plasma viral load to levels below 2.7 log₁₀ copies/ml (Amplicor) at any time during the study period.

Statistical analysis

Statistical comparisons, performed on an intent-to-treat basis, were conducted using distribution-free methods [²⁰]. Categorical variables and ordinal and skewed continuous variables were compared with Mantel-Haenszel and Wilcoxon rank-sum tests, respectively. Fisher's exact test was used for 2 × 2 contingency tables in which any of the expected cell frequencies was less than 5.0. Multivariate logistic regression was used to identify baseline predictors of the plasma viral load suppression below 2.7 log₁₀ copies/ml.

Results

A total of 58 individuals were started on RTV and SQV-based combination therapy in the province of British Columbia during the period 1 September 1996 to 28 February 1997. The overall median time on RTV and SQV therapy was 5.3 months [interquartile range (IQR), 3.5-6.2 months]. No-one in this group was lost to follow-up during the study period.

The median plasma viral load for the 58 study subjects at entry was 4.8 log₁₀ copies/ml (IQR, 4.51-5.15 log₁₀ copies/ml). A moderately strong negative correlation between baseline plasma viral load and CD4 cell count determination (Spearman's ρ correlation = -0.55; P < 0.001) was observed in this group of individuals. The median CD4 cell count at baseline was 140 × 10⁶/l (IQR, 40-270 × 10⁶/l). Subjects with AIDS at baseline had significantly higher median plasma viral load measurements at entry than those without a diagnosis of AIDS (5.11 log₁₀ versus 4.69 log₁₀ copies/ml; P < 0.025).

The third antiretroviral drug at baseline was lamivudine (22 subjects, 38%), stavudine (19 subjects, 33%), didanosine (nine subjects, 16%), zalcitabine (one subject, 2%), and zidovudine (one subject, 2%). A total of six (10%) subjects were on RTV and SQV alone.

All subjects had previous experience with antiretrovirals at baseline. The median time on antiretroviral therapy prior to initiating therapy with RTV-SQV was 42 months (IQR, 21-51 months). Prior to the start of RTV-SQV-based combination therapy, patients were on the following antiretroviral therapy: 10 (17%) subjects were on stavudine-lamivudine; 10 (17%) were on zidovudine-lamivudine; nine (16%) were on triple combination therapy with indinavir; seven (12%) were on triple combination therapy with SQV; six (10%) were on triple combination therapy with stavudine-didanosine-lamivudine; three (5 %) were on triple combination therapy with RTV; three (5%) were on stavudine-didanosine; and 10 (17%) were on other regimens.

At baseline, 36 (62%) subjects had no prior history of protease inhibitor use. Amongst the 52 participants who were prescribed a nucleoside reverse transcriptase inhibitor (NRTI) in conjunction with RTV and SQV, 20 (38%) were taking a new NRTI to which they had never been previously exposed. Fourteen (70%) of these 20 participants were prescribed stavudine as their new NRTI. Nine (64%) of these 14 participants were on zidovudine-lamivudine prior to taking RTV-SQV-stavudine.

Of the 58 participants included in the analysis, seven (12%) had discontinued therapy (two deaths, three because of nausea or vomiting with or without diarrhea, and two because of increase in plasma viral load after starting therapy). Reported side-effects that did not lead to discontinuation of therapy included insomnia, rash, hypertension, increased triglycerides, light-headedness, and orthohypotension. Each of these occurred in one participant, with the exception of light-headedness, which occurred in two individuals.

Nausea or vomiting with or without diarrhea was also reported for 24 (47%) other participants and did not lead to discontinuation of therapy.

Figs 1 and 2 show the baseline (the median of the last plasma viral load and CD4 cell counts obtained before starting RTV-SQV), and also the median of the latest values obtained within each time period after starting RTV-SQV-based combination therapy. Median plasma viral load decreased from 4.79 log₁₀ copies/ml (IQR, 4.51-5.15 log₁₀ copies/ml) at baseline to below 2.7 log₁₀ copies/ml [IQR, < 2.7 log₁₀ (non-quantifiable) to 3.32 log₁₀ copies/ml] at weeks 21-28. Median CD4 cell count increased from 140 × 10⁶/l (IQR, 40-270 × 10⁶/l) at baseline to 260 × 10⁶/l (IQR, 80-520 × 10⁶/l) at weeks 21-28.

Fig. 1

Fig. 2

Table 1 compares baseline factors associated with a subsequent decrease in plasma viral load below 2.7 log₁₀ copies/ml at any time during the study period. A total of 29 (50%) subjects demonstrated a decrease in plasma viral load to levels below 2.7 log₁₀ copies/ml and the median time for them to reach this level of suppression was 9.4 weeks (IQR, 6.4-16.6 weeks). The proportion of subjects with viral load decreasing below 2.7 log₁₀ copies/ml did not differ with respect to age (P = 1.0), gender (P = 1.0), having AIDS (P = 0.063), or using a new NRTI with the treatment regimen (P = 0.826). Baseline CD4 cell count (P = 0.022), but not plasma viral load (P = 0.431), was statistically associated with a drop in plasma viral load below 2.7 log₁₀ copies/ml. The median baseline CD4 cell count was 70 × 10⁶/l for those subjects who did not decrease plasma load below 2.7 log₁₀ copies/ml and 205 × 10⁶/l for those who did. Amongst individuals naive to protease inhibitors at entry, 24 (67%) dropped below 2.7 log₁₀ copies/ml, whereas only five (23%) of those previously exposed to protease inhibitors ever reached a viral load below 2.7 log₁₀ copies/ml (P = 0.001).

Table 1

A final multivariate model was constructed for baseline factors associated with a subsequent drop in plasma viral load below 2.7 log₁₀ copies/ml after initiating RTV and SQV-based combination therapy. After controlling for baseline plasma viral load and CD4 cell count, a subsequent decrease in plasma viral load below 2.7 log₁₀ copies/ml was associated with being naive to protease inhibitors at baseline (P = 0.004). The adjusted risk ratio from this model indicated that study subjects naive to protease inhibitors at baseline were almost seven times (odds ratio, 6.99; 95% confidence interval, 1.85-26.39) more likely than those previously exposed to decrease plasma viral load below 2.7 log₁₀ copies/ml.

Finally, we repeated all relevant analyses to determine whether increased duration of therapy with any antiretroviral agent predicted virological failure. After controlling for baseline plasma viral load and CD4 cell count we found only the previous duration of protease inhibitor therapy to be associated with decreasing plasma viral load below 2.7 log₁₀ copies/ml (P = 0.023). In this multivariate model each month of previous protease inhibitor use was associated with a 30% (odds ratio, 0.70; 95% confidence interval, 0.51-0.95) decrease in the likelihood of ever decreasing plasma viral load below 2.7 log₁₀ copies/ml during the study period.

Discussion

This observational study describes the short-term virological effect in 58 subjects treated with RTV and SQV. Our results demonstrate that RTV and SQV-based combination therapy suppressed plasma viral load to below the lower limit of quantification (2.7 log₁₀ copies/ml) in half of the study participants. No prior exposure to protease inhibitors enhanced the likelihood of suppressing plasma viral load to this level by almost sevenfold.

The extent of plasma viral load suppression described here is in keeping with that found in two recent pilot studies. Arvieux et al. [¹⁴] found that 60% of protease inhibitor-naive subjects had plasma viral loads below 2.7 log₁₀ copies/ml after 8 weeks of therapy when RTV (600 mg every 12 h) and SQV (400 mg every 12 h) were added to a stable regimen of zidovudine and lamivudine. In addition, Sun et al. [¹⁶] demonstrated that the median viral load was at the limit of quantification of the assay (< 2.3 log₁₀ copies/ml) after 24 weeks of therapy with two dosing regimens (RTV 400 mg and SQV 400 mg every 12 h, or RTV 600 mg and SQV 400 mg every 12 h).

Prior exposure to protease inhibitors seriously compromised the antiviral effect of RTV and SQV-based combination therapy in our study. This is consistent with data demonstrating extensive cross-resistance within the protease inhibitor class [²¹]. After a year of therapy with SQV, patients who were changed to indinavir only decreased their viral load by 0.58 log₁₀ copies/ml [²²]. However, significant sensitivity to RTV or indinavir was shown among patients pretreated for 32-35 weeks with SQV. After 4-8 weeks of therapy, 100% of the subjects who added RTV to their regimen had a decline in viral load (mean, -1.72 log₁₀ copies/ml), while 79% of those who changed from SQV to indinavir had a decline in viral load (mean, -1.57 log₁₀ copies/ml) [²³]. Variations in response to RTV and indinavir following SQV therapy may be due to the varying degrees of effective exposure to SQV as a result of its variable and often poor bioavailability. It is clear, however, that any substantial previous exposure to, or more clearly yet, virological failure with a protease inhibitor will ultimately limit response to RTV and SQV-based combinations. However, in the presence of advanced immunosuppression the virologic response to this regimen may be unpredictable even in persons who have not received prior protease inhibitor therapy [²⁴].

Compliance with combinations including RTV has often been difficult secondary to the frequent occurrence of side-effects. Within our study, even though 47% of individuals complained of nausea, vomiting or diarrhea on initiating therapy, only three participants discontinued therapy as a result of these side-effects. Two individuals discontinued therapy because of virological failure. In an attempt to ameliorate this problem we used both an escalating dose regimen and liberal use of anti-emetics and anti-diarrheal agents [²⁵]. Overall, this relatively high rate of side-effects could potentially affect subject compliance with these regimes.

There are several features of our study that should be emphasized. First, our study was carried out within a province-wide treatment program where all individuals had access to combination antiretroviral therapy free of cost. We are confident, therefore, that our results were not influenced by access to therapy-related issues, which have often compromised the interpretation of similar population-based studies. Second, no participants were lost to follow-up in this population-based analysis. This very high rate of follow-up was attributable to the centralized distribution of antiviral medications and compares favorably with other observational studies of this nature [²⁶].

Combination therapy with a two- or three-drug anti-retroviral regimen has been demonstrated to be more effective than monotherapy in both surrogate marker and clinical endpoint trials [^1,27,28]. Powerful triple drug regimens including two nucleosides plus a protease inhibitor or a non-nucleoside reverse transcriptase inhibitor are now the preferred therapy for most patients [²⁹]. Given the limited number of antiretrovirals currently available, and the significant cross-resistance among currently available protease inhibitors, the use of a two protease inhibitor combination may allow patients to maximize the use of protease inhibitors, while minimizing the use of other agents that could be used in future regimens.

In summary, our results demonstrate that combination antiretroviral therapy with RTV and SQV suppresses plasma viral load to below 2.7 log₁₀ copies/ml in half of antiretroviral-experienced participants. Subjects without prior exposure to protease inhibitors and with higher baseline CD4 cell counts were more likely to decrease their viral load to below the limit of detection of our assay. Our findings support the use of RTV and SQV in combination, although prior exposure to protease inhibitors will severely compromise response to this combination therapy.

Acknowledgements

The authors thank R. Dewletian, B. Devlin, D. Hamann-Trou, M. Reginaldo, D. Campbell, E. Ferris, N. Gataric, R. Gomes, and K. Hsu for their research and administrative assistance.