Protease inhibitors have had a dramatic impact on the management and natural history of HIV disease [1,2]. In controlled clinical studies [3-5], patients who initiate combination therapy with a potent protease inhibitor and at least one new nucleoside reverse transcriptase inhibitor (NRTI) often achieve durable suppression of viral replication. The efficacy of these therapies observed in clinical trials may not translate directly into clinical practice, partly because clinical trials may select for a healthier, more motivated and possibly more adherent patient population . Similarly, because patients who experience virological failure in a clinical trial setting often withdraw from the study prematurely, the relationship between the HIV RNA and CD4 response in experiencing virological failure has not been well described. Therefore, to describe the effectiveness of protease inhibitor therapy in clinical practice, and to determine the relationship between baseline characteristics and outcome, an analysis of patients seen at the San Francisco General Hospital AIDS Clinic was performed. HIV-1 RNA levels and CD4 cell counts over the first year of therapy with a potent protease inhibitor-based regimen (indinavir, ritonavir or nelfinavir) were analysed. Additionally, the response to second-line therapy for patients who failed their initial regimen (‚salvage therapy‚) was studied in a subset of patients.
The University of California, San Francisco AIDS program at San Francisco General Hospital (SFGH) is an urban, university-based public hospital clinic that provides comprehensive primary care to HIV-infected adults. The majority of patients (99%) are uninsured or are covered by Medicaid or Medicare. In 1996 and 1997, the majority of patients in the clinic were men (86%) and Caucasian (56%).
Patients were identified through an administrative database that records outpatient visits. The names of all patients seen at least three times by the same clinician between March 1996 and September 1997 were identified. Medical records were reviewed to identify those who were eligible for this study. Patients were included in this analysis if they received at least 16 weeks of continuous therapy with an antiretroviral regimen containing indinavir, ritonavir or nelfinavir. To allow at least 48 weeks of subsequent follow-up, we analysed patients who initiated therapy before November 1997. Patients were excluded if they received protease inhibitor monotherapy or had less than 16 weeks of follow-up for any reason.
Because of its limited in-vivo antiviral potency, we did not evaluate the virological effectiveness of the hard-gel formulation of saquinavir (Invirase) . Patients were eligible, however, if they received saquinavir hard-gel capsule and later switched to an indinavir, ritonavir or nelfinavir-based regimen. No patient in our clinic received the soft-gel formulation of saquinavir (Fortovase) during the study period.
The virological effectiveness of second-line therapy after virological failure of the first regimen (‚salvage therapy‚) was also evaluated. Patients were included in this analysis if they experienced virological failure with their initial protease inhibitor-based regimen (defined by an HIV RNA level above 500 copies/ml after week 16), and switched to a second regimen. To qualify as salvage therapy, the second regimen had to contain a protease inhibitor that the patient had not previously taken. To evaluate at least 24 weeks of salvage therapy, we analysed patients who initiated their salvage regimen before April 1998. Patients were excluded if they discontinued their salvage regimen for any reason before week 8.
Two research associates independently reviewed the medical records of all patients who received at least 16 continuous weeks of indinavir, ritonavir, nelfinavir or ritonavir/saquinavir. The following data was obtained during the initial review: date of birth, sex and length of previous exposure to each individual antiretroviral agent. Once patients were identified, their medical records were reviewed every 3-4 months to November 1998. Plasma HIV RNA assays were performed using a branched DNA (bDNA) assay (Chiron Corp., Emeryville, CA, USA). Before 1 July 1996, HIV RNA tests were performed using an earlier version of the bDNA assay (Quantiplex HIV bDNA version 1.0; lower limit of quantification 10000 copies/ml). If these assays were not available, HIV RNA results from an experimental reverse transcriptase-polymerase chain reaction (RT-PCR) assay were used (Immuno-Diagnostic Laboratories, San Leandro, California, USA). After 1 July 1996, all HIV RNA determinations were performed with the bDNA assay, version 2.0 (lower limit of quantification 500 copies/ml). After March 1998, all HIV RNA samples below the level of quantification with version 2.0 were re-analysed with version 3.0 (lower limit of quantification 50 copies/ml). Routine CD4 T cell phenotyping was performed at the SFGH central laboratory.
The primary outcome measured was plasma HIV RNA response at week 48 of therapy (±4 weeks). For patients who did not have an HIV RNA measure at week 48, the first determination after week 48 was used. Treatment failure was defined as an HIV RNA level greater than 500 copies/ml at week 48. In the analysis of response to salvage therapy, the primary outcome measured was the plasma HIV RNA level 24 weeks (±4 weeks) after initiating salvage therapy; failure was defined as an HIV RNA level of over 500 copies/ml at week 24. The bivariate relationship between categorical risk factors and virological outcome was tested for each analysis using chi-squared and Fisher‚s exact tests. The independent association of each predictor variable with virological outcome was examined using a multiple logistic regression model. Only variables from the bivariate analysis that had a P-value less than 0.2 were considered in the logistic model. To limit the number of variables in the logistic regression model, the CD4 and HIV RNA strata were collapsed into three categories. Crude and adjusted odds ratios with 95% confidence intervals are reported for each variable included in the analysis.
To determine the effect of virological failure on the CD4 cell response, patients with an evaluable baseline HIV RNA level were categorized into four groups on the basis of their HIV RNA response over time: complete responders (HIV RNA <500 at week 48), partial responders (HIV RNA >500 copies RNA/ml but at least 1.0 log copies/ml below baseline), transient responders (an initial HIV RNA decrease of >1.0 log copies/ml followed by a rebound to within 1.0 log copies/ml of baseline) and non-responders (no HIV RNA >1.0 log copies/ml below baseline). HIV RNA and CD4 cell counts were evaluated at weeks 6 (±3 weeks), 12 (±3 weeks), 24 (±6 weeks), 36 (±6 weeks) and 48 (±6 weeks). Pairwise comparisons between groups at each time point were analysed using Wilcoxon rank sum. All statistical analyses were performed using SAS version 6.12 (SAS Institute, Cary, NC, USA).
We identified a total of 856 HIV-positive patients were seen at least three times by 14 SFGH AIDS Program clinicians between March 1996 and September 1997. Of these patients, 440 initiated a regimen containing indinavir, ritonavir or nelfinavir before 1 October 1997. A total of 103 patients were excluded from this analysis because they stopped therapy before week 16 due to intolerance (n=41), were lost to follow-up before week 16 (n=17), died before week 16 (n=12), received protease inhibitor monotherapy (n=13) or had limited source documentation describing when protease inhibitor therapy was initiated (n=20).
A total of 337 patients met the eligibility criteria. The median baseline CD4 cell count and baseline HIV RNA level were 143 cells/mm3 and 4.64 log10 copies/ml, respectively. Only 47 patients (13.9%) were treatment naïve at the time indinavir, ritonavir or nelfinavir was initiated.
After 48 weeks of treatment, 170 patients(50.4%) had a successful outcome (HIV RNA <500 copies/ml), whereas the remaining 167 patients (49.6%) experienced virological failure (see Table 1). If patients who were excluded because they died or discontinued therapy as a result of intolerance before week 16 are classified as failures (n=53), the percentage of patients achieving a successful outcome declines to 43.6%.
Patients meeting a definition of virological success versus failure did not differ significantly by sex, mean age or protease inhibitor used. Treatment-experienced patients who added a protease inhibitor to a stable pre-existing NRTI regimen were more likely to fail protease inhibitor therapy than patients who initiated at least one new NRTI (P=0.01). Notably, 40 of the 47 (85%) treatment-naïve patients who initiated combination therapy with a protease inhibitor had a successful outcome.
Of the 337 patients eligible for this analysis, 68 (20.1%) had been treated with saquinavir hard-gel capsule before initiating a more potent protease inhibitor. Within the saquinavir-experienced group, 25 (36.8%) had a successful outcome 48 weeks after switching to a regimen containing indinavir, ritonavir, nelfinavir or ritonavir/saquinavir. Within the saquinavir-naïve group, 53.9% had a successful outcome (see Table 1).
To examine the relationship between baseline CD4 or HIV RNA levels and outcome, patients were categorized into several CD4 and HIV RNA strata (see Table 1). Using a chi-squared test for trend analysis, the relationship between each of these predictors and outcome was highly significant (P<0.001 for both predictors). Although there was a trend suggesting that the risk of failure was higher in patients with a CD4 cell count between 200 and 400 cells/mm3 than those with a CD4 cell count higher than 400 cells/mm3, this did not reach statistical significance. As the baseline CD4 cell count decreased below 200 cells/mm3, the risk of virological failure increased significantly. There was a similar relationship between baseline HIV RNA level and outcome, with an increase in the risk of failure particularly evident as the baseline RNA level increased above 4.5 log10 copies/ml (31622 copies/ml).
Complete baseline data (HIV RNA, CD4 cell count, and previous antiretroviral therapy) were available on 300 patients. In multivariate logistic regression analysis, independent predictors of failure were high baseline HIV RNA level, low baseline CD4 cell count, and failure to initiate a new NRTI when protease inhibitor therapy was initiated (see Table 2). Whereas in the bivariate analysis there was an incremental increase in the risk of virological failure as the baseline CD4 cell count decreased, in multivariate analysis that controlled for other factors there was little increase in the risk of virological failure until the baseline CD4 cell count fell below 100 cells/mm3. Similarly, when controlling for other baseline characteristics, the risk of virological failure increased as the baseline HIV RNA increased above 4.0 log10 copies/ml; this became more pronounced as the level increased above 4.5 log10 copies/ml.
A total of 137 patients with an HIV RNA level of less than 500 copies RNA/ml (bDNA, version 2.0) after at least 48 weeks of therapy had plasma available for analysis with an ultrasensitive assay (bDNA, version 3.0; lower limit of quantitation 50 copies RNA/ml).
At the time of the analysis, patients had been on a protease inhibitor-based regimen for a median of 93 weeks (interquartile range 75-108 weeks). Of the 137 patients evaluated with the ultrasensitive assay, 91 (66%) had a plasma HIV RNA of less than 50 copies RNA/ml.
CD4 cell count response
At week 48 of therapy, the median increase in CD4 cell count was 153 cells/mm3 in the virological success group (interquartile range 52-263) and 75 cells/mm3 in the virological failure group (interquartile range 17-147). The median change in CD4 cell count was similar for both groups at weeks 6, 12, and 24, with the difference becoming significant only at weeks 36 and 48 (P=0.005 at week 36 and P=0.001 at week 48; Wilcoxon rank sum).
The relationship between the HIV RNA and CD4 response over 48 weeks of therapy is illustrated in Fig. 1(a and b). On the basis of the HIV RNA response over time, patients were classified as complete responders (n=156), partial responders (n=62), transient responders (n=56) and non-responders (n=27). As shown in Fig. 1(a), there was a median increase in CD4 cell count at week 48 for all groups (median change of 153, 114, 69.5 and 15 cells/mm3 for the complete responders, partial responders, transient responders and non-responders, respectively). Notably, the CD4 response in the complete versus partial responders was similar at all time points. The CD4 response appeared to be sustained to week 48 for patients who had a transient HIV RNA response.
Response to salvage therapy
Of the 167 patients who failed their initial regimen, 99 switched to a second-line protease inhibitor-based salvage regimen and had at least 24 weeks of follow-up. Before the switch, patients were failing either indinavir (n=73), ritonavir (n=23) or nelfinavir (n=3)-based regimens. The median CD4 cell count and HIV RNA level at the time of the switch were 196 cells/mm3 (interquartile range 96-265 cells/mm3) and 4.46 log10 copies RNA/ml (interquartile range 4.04-4.94 log10 copies RNA/ml). Notably, the median CD4 cell count before the initial protease inhibitor-based regimen was significantly lower (89 cells/mm3; interquartile range 30-162 cells/mm3) than the CD4 cell count at the time of salvage therapy. All patients initiated a regimen containing at least one protease inhibitor that they had not received in the past; 27 patients were non-nucleoside reverse transcriptase inhibitor (NNRTI)-naïve and initiated a regimen containing an NNRTI.
After 24 weeks of salvage therapy, 22 patients (22.2%) had a successful outcome (plasma HIV RNA <500 copies/ml) (see Table 3). In a multivariate model, independent predictors of failure included: baseline HIV RNA greater than 4.0 log10 copies/ml at the time of the switch and failure to initiate a NNRTI (nevirapine, delavirdine or efavirenz) as part of the salvage regimen (see Table 4). Patients experiencing a successful response to salvage therapy had a median CD4 cell increase of 65 cells/mm3 at week 24 whereas patients who did not have a successful response had a median CD4 cell increase of 18 cells/mm3 (P=0.01; Wilcoxon rank sum).
The recommended goal of antiretroviral therapy is to maintain the suppression of viral replication below the level of quantitation using the most sensitive HIV RNA assay available [8-10]. At SFGH only 50.4% of patients who received at least 16 weeks of therapy with a potent protease inhibitor-containing regimen achieved this goal through the first year of follow-up using a standard assay (<500 copies/ml). This is in contrast to the 70-90% success rates commonly observed in clinical trials [4,5], and is similar to success rates seen in patients with advanced disease  and in other clinical cohorts . Notably, a large proportion of patients experiencing virological failure (defined as a detectable HIV RNA) still had a durable decrease in HIV RNA levels of at least 1.0 log copies RNA/ml below baseline.
Much of the virological failure observed in this study can be attributed to the suboptimal use of antiretroviral agents. As had been previously recommended , many patients in our clinic received sequential monotherapy with NRTIs before the availability of protease inhibitors. The use of antiretroviral agents in this manner may result in the development of broad cross-resistance. The addition of a protease inhibitor to an existing NRTI regimen in which resistance has developed may be similar to using the protease inhibitor as monotherapy [8-10]. Similar trends were observed in a phase II clinical trial of indinavir, in which approximately 90% of patients who initiated zidovudine, lamivudine and indinavir simultaneously had evidence of durable viral suppression. In contrast, when indinavir was added to zidovudine and lamivudine in a sequential manner, only 40% of the patients had a successful outcome with the same three-drug regimen .
The first protease inhibitor to be approved for use by the US Food and Drug Administration was saquinavir hard-gel capsule (saquinavir-hgc; Invirase). Because of the limited oral bioavailability of the initial hard-gel formulation, it was difficult to achieve optimal serum concentrations of saquinavir. Virological response to saquinavir-hgc is relatively limited . As a result, the drug has never been a recommended first line option [8-10,13,14]. Nonetheless, a significant number of patients used saquinavir-hgc before switching to a more potent protease inhibitor. Preliminary observations indicated that saquinavir may confer cross-resistance to other protease inhibitors , although this remains controversial . After controlling for other baseline characteristics (HIV RNA, CD4 cell count and previous NRTI therapy), saquinavir-hgc was a weak predictor of subsequent drug failure, suggesting that saquinavir may indeed confer cross-resistance to other protease inhibitors.
The optimal time at which therapy should be initiated is controversial [10,17]. In this study, disease stage, as defined by either the baseline CD4 cell count or HIV RNA level was an independent predictor of virological outcome. Controlling for other factors, a baseline HIV RNA greater than 4.5 log10 copies/ml was strongly predictive of subsequent virological failure. Similarly, a very low baseline CD4 T cell count (<100 cells/mm3) was predictive of failure. Assuming that the goal of therapy is to achieve durable viral suppression, these observations support recommendations that therapy be initiated before the development of advanced HIV disease. Whether therapy can be safely delayed until these potential threshold values (HIV RNA level 4.5 log10 copies/ml; CD4 cell count 100 cells/mm3) remains unknown, but deferring therapy beyond the level at which it is currently recommended [8,9] may still result in durable viral suppression. It is important to emphasize that the outcome measured in this study was virological; early intervention may still be necessary to ensure the preservation of immune function. Larger studies, with clinical endpoints, are necessary to determine the optimal time to begin combination therapy.
As has been reported in other observational studies , patients experiencing virological failure in this study generally had a sustained CD4 cell count response. There was no clear difference in the CD4 cell response between patients who had a complete virological response (HIV RNA <500 copies RNA/ml) and those who had a durable but partial response (HIV RNA >500 copies RNA/ml but at least 1 log below baseline to 48 weeks). Furthermore, those patients who had an initial virological response and later rebounded back towards baseline maintained a stable increase in CD4 cell count to one year of follow-up. Factors determining the interval between virological failure and CD4 depletion remain to be determined, but will probably depend on the duration of virological failure, new steady state viral load and type of regimen the patient is receiving. Longer-term follow-up will be necessary to examine these issues.
The treatment of patients experiencing virological failure on a protease inhibitor-containing regimen is undergoing intense investigation. Recognizing the lack of prospective data, current guidelines recommend the use of aggressive regimens containing two protease inhibitors (ritonavir/saquinavir or nelfinavir/saquinavir) and two NRTIs, using drugs to which the patient is naïve whenever possible . In this study, overall response rates were low for patients who switched to the standard regimen of two protease inhibitors and two NRTIs, even when NRTI therapy was modified at the time of the switch. Successful salvage therapy was, however, common if patients switched to a protease inhibitor regimen containing a NNRTI. Considering the lack of cross-resistance between NNRTIs and the other classes of antiretroviral agents, the potent response to an NNRTI-containing regimen is perhaps not surprising. Larger prospective studies, with pharmacokinetic sub-studies, are necessary to evaluate the role of aggressive regimens containing protease inhibitors and an NNRTI.
Largely on the basis of theoretical considerations, current guidelines recommend switching to a salvage regimen as soon as virological failure is determined. Continuing therapy in the face of a failing regimen may select for further resistance and broad-cross resistance, thus limiting the options for effective salvage therapy. In this study, the delay between the onset of virological failure and switching to a salvage regimen was not measured. A lower plasma HIV RNA at the time of the switch (<104 log10 copies/ml) was strongly associated with an improved outcome, supporting recommendations that switching to a salvage regimen should be considered as soon as virological failure is confirmed [8,9].
This observational study has several limitations. The patients studied reflect an urban, public hospital AIDS clinic population treated between 1996 and 1998, and may not be representative of all patient populations. Also, the clinical implication of failure defined solely by viral RNA determinations is unknown and remains controversial. Although the magnitude of treatment-mediated reduction in plasma HIV RNA is predictive of clinical outcome , the clinical implication of viral RNA rebound remains unclear. Similarly, our definition of a successful outcome does not account for potential long-term toxicities of therapy [20-22]. Finally, factors other than those analysed in this observational study probably contribute to the long-term virological activity of combination therapy. Adherence to the medication regimen, drug absorption, drug metabolism, the presence of tissue sanctuaries for viral replication and increased target cell availability have all been proposed as possible determinants of virological outcome [23,24]. Their contribution to virological failure could not be addressed in this study.
The observations made in this study have several clinical implications for the management of antiretroviral therapy. First, combination therapy should be initiated simultaneously, as is currently recommended [8-10]. Second, initiating therapy late in HIV disease is associated with an increased risk of virological failure. The risk of virological failure increases significantly as the baseline CD4 T cell count falls below 100 cells/mm3 or the HIV RNA level increases above 4.0-4.5 log10copies/ml. Patients who initiate therapy beyond these levels may benefit from more aggressive regimens than two NRTIs and a protease inhibitor. Prospective studies evaluating four-drug regimens in advanced disease are underway. Third, successful salvage therapy after the failure of an indinavir- or ritonavir-based regimen is uncommon in clinical practice. In this study, successful salvage therapy was associated with using a new class of drugs to which the patient was naïve (i.e. NNRTIs) and switching when the plasma HIV RNA was low. Finally, although virological failure in this group was common, immunological failure (defined by declining CD4 cell count) was rare; most patients in this study experienced a sustained increase in CD4 cells over the first year. The clinical implication of virological failure remains unknown.
We would like to acknowledge Dr Nina Simonds, Dr George Beatty and Shayna Chin for their assistance in reviewing the medical records.
1. Mouton Y, Alfandari S, Valette M, et al
.: Impact of protease inhibitors on AIDS-defining events and hospitalizations in 10 French AIDS reference centres. AIDS
2. Palella FJ, Delaney KM, Moorman AC, et al
.: Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. N Engl J Med
3. Hammer SM, Squires KE, Hughes MD, et al
.: A controlled trial of two nucleoside analogues plus indinavir in persons with human immunodeficiency virus infections and CD4 cell counts of 200 per cubic millimeter or less. N Engl J Med
4. Gulick R, Mellors J, Havlir D, et al
.: Simultaneous vs. sequential initiation of therapy with indinavir, zidovudine and lamivudine for HIV-1 infection: 100 week follow-up. JAMA
5. Saag M, Knowles M, Chang Y, Chapman S, Clendeninn NJ, Viracept Cooperative Study Group: Durable effect of Viracept (nelfinavir mesylate, NFV) in triple combination therapy. Programs and Abstracts of the 37th International Conference on Antimicrobial Agents and Chemotherapy
. Toronto, Canada 1997 [abstract I-101].
6. Brook RH; Lohr KN: Efficacy, effectiveness, variations, and quality. Boundary-crossing research. Med Care
7. Collier AC, Coombs RW, Schoenfeld DA, et al
.: Treatment of human immunodeficiency virus infection with saquinavir, zidovudine, and zalcitabine. N Engl J Med
8. US Department of Health and Human Services and Henry J. Kaiser Family Foundation: Guidelines for the use of antiretroviral agents in HIV infected adults and adolescents. MMWR
9. Carpenter CC, Fischl M, Hammer SM, et al
.: Antiretroviral therapy for HIV infection in 1998: updated recommendations of the International AIDS Society - USA Panel. JAMA 280
10. Gazzard B, Moyle G and the BHIVA Guidelines Writing Committee: 1998 Revision to the British HIV Association guidelines for antiretroviral treatment of HIV seropositive individuals. Lancet
11. Fatkenheuer G, Theisen A, Rockstroh J, et al
.: Virological treatment failure of protease inhibitor therapy in an unselected cohort of HIV-infected patients. AIDS
12. Sande MA, Carpenter CC, Cobbs CG, Holmes KK, Sanford JP: Antiretroviral therapy for adult HIV-infected patients. Recommendations from a State-of-the-Art Conference. National Institute of Allergy and Infectious Diseases State-of-the-Art Panel on Anti-Retroviral Therapy for Adult HIV-Infected Patients. JAMA
13. Carpenter CC, Fischl M, Hammer SM, et al
.: Antiretroviral therapy for HIV infection in 1996. Recommendations of an international panel. JAMA
14. Carpenter CC, Fischl M, Hammer SM, et al
. for the International AIDS Society-USA: Antiretroviral therapy for HIV infection in 1997: updated recommendations of the International AIDS Society - USA Panel. JAMA
15. Para MF, Collier A, Coombs R for the ACTG 333 Study Team: ACTG 333: antiviral effects of switching from saquinavir hard capsule (SQVhc) to saquinavir soft gelatin capsule (SQVsgc) vs. switching to indinavir (IDV) after prior saquinavir. Program and Abstracts of the Infectious Diseases Society of America 35th Annual Meeting
. San Francisco, 13-16 September, 1997 [abstract 21].
16. Roberts NA, Craig JC, Sheldon J: Resistance and cross-resistance with saquinavir and other HIV protease inhibitors: theory and practice. AIDS
17. Levy JA: Caution: should we be treating HIV infection early? Lancet
18. Kaufmann D, Panteleo G, Sudre P, Telenti A: CD4 T cell count in HIV-1 individuals remaining viremic with highly active antiretroviral therapy (HAART): Swiss HIV Cohort Study [letter]. Lancet
19. Marschner IC, Collier AC, Coombs RW, et al
.: Use of changes in plasma levels of human immunodeficiency virus type 1 RNA to assess the clinical benefit of antiretroviral therapy. J Infect Dis
20. Carr A, Samaras K, Burton S, Law M, Freund J, Chisholm DJ, Cooper DA: A syndrome of peripheral lipodystrophy, hyperlipidaemia and insulin resistance in patients receiving HIV protease inhibitors. AIDS
21. Lo JC, Mulligan K, Tai VW, et al
.: Buffalo hump in men with HIV-1 infection. Lancet
22. Miller KD, Jones E, Yanovsk JA, et al
.: Visceral abdominal fat accumulation associated with use of indinavir. Lancet
23. Hirsch MS, Conway B, D‚Aquila RT, et al
., for the International AIDS Society - USA Panel: Antiretroviral drug resistance testing in adults with HIV infection: implications for clinical management. JAMA
24. Havlir DV, Marschner IC, Hirsch MS, et al
.: Maintenance antiretroviral therapies in HIV-infected subjects with undetectable plasma HIV RNA after triple-drug therapy. N Engl J Med
Keywords:© 1999 Lippincott Williams & Wilkins, Inc.
Protease inhibitors; viral load; CD4; combination therapy; observational study