A placebo-controlled trial of didanosine plus stavudine, with and without hydroxyurea, for HIV infection. The Swiss HIV Cohort Study

OBJECTIVE: To explore the short-term effects on surrogate markers for HIV progression of didanosine (ddl) plus stavudine (d4T), with or without hydroxyurea. DESIGN: Randomized, double-blinded, prospective study. SETTING: Swiss HIV Cohort Study. PATIENTS: A total of 144 patients (75% antiretroviral-naive) were studied (mean baseline HIV-1 RNA, 4.53 log10 copies/ml; mean CD4 cell count, 370 x 10(6)/l). INTERVENTION: Patients received ddl (200 mg twice daily) plus d4T (40 mg twice daily), with additional hydroxyurea (500 mg twice daily) or placebo. MAIN OUTCOME MEASURES: The primary endpoint was a reduction of viraemia below 200 copies/ml after 12 weeks. At that time, patients who did not reach the primary endpoint were withdrawn in the hydroxyurea arm, whereas patients in the placebo group had the option of adding hydroxyurea to ddl and d4T. All patients were followed until week 24. RESULTS: After 12 weeks, 54% of the patients randomized to hydroxyurea had viraemia below 200 copies/ml, compared with 28% on placebo (P < 0.001). Using an ultrasensitive assay with a limit of detection of 20 copies/ml, 19% of patients receiving hydroxyurea had viraemia levels below 20 copies/ml, compared with 8% on placebo (P = 0.05). Mean decrease in HIV-1 RNA was 2.3 and 1.7 log10 copies/ml for hydroxyurea and placebo groups, respectively (P = 0.001). Hydroxyurea was found to induce lymphopenia (-124 x 10(6)/l). Increase in CD4 cell counts was +28 x 10(6)/l during hydroxyurea treatment compared with +107 x 10(6)/l on placebo (P = 0.001). CONCLUSIONS: Hydroxyurea improved the antiviral activity of d4T and ddl over a 12-week period, but was associated with a smaller increase in CD4 cell counts due to hydroxyurea-induced lymphopenia.     

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 x 10(6) to 418 x 10(6) cells/l (range, 241-537 x 10(6) 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.     

Value of HIV-1 viral load and CD4 lymphocyte count as determinants of progression to AIDS and survival

BACKGROUND: HIV-1 viral load is regarded as a better surrogate marker for progression and death than CD4+ cell counts. Both markers are analysed in a cohort of patients with unknown seroconversion date and advanced HIV infection. PATIENTS AND METHODS: Retrospective cohort analysis of 421 patients, most on antiretroviral therapy, with a median initial CD4+ cell count of 209 x 10(6)/l and a median initial viral load of 4.7 log copies/ml. One thousand two hundred and eighty-six samples were analysed. Univariate and bivariate analysis were performed with initial and sequential CD4+ cell counts and viral load values to estimate the risk of progression and death by Cox regression models. RESULTS: After a median follow up of 763 days, 124 patients developed AIDS and 117 died. Relative risks of progression related to the group that maintained viral load values always < 35,000 copies/ml were: 5-fold (95% CI: 1.4-17.0; p < 0.05) for patients with any viral load value > 35,000 copies/ml but always < 200,000 copies/ml; and 13.6 fold (95% CI: 5.4-34.2; p < 0.0001) for patients who could not maintain viral load < 200.000 copies/ml. CD4+ counts = 100 x 10(6)/l and viral load = 220,000 copies/ml were the threshold values that best fitted to estimate the probability of survival by a bivariate analysis. CONCLUSIONS: The maintenance of sequential viral load values < 35.000 copies/ml is associated with a lower risk of progression. The maintenance of sequential viral load values < 150,000 copies/ml is associated with higher short-term survival rates.     

Once-daily administration of didanosine in combination with anti-retroviral zidovudine in previously untreated patients]

25 HIV-infected antiretroviral-naive adults were included in a 24-week study to evaluate the efficacy and the tolerability of a zidovudine/didanosine combination therapy in which didanosine was administered once daily (200 mg if weight < 60 kg, 300 mg if weight > 60 kg) and zidovudine twice daily (500 mg/day if weight < 90 kg, 600 mg/day if weight > 90 kg). 5 patients discontinued their treatment early: 3 had poor compliance and 2 presented adverse events. Evaluation of treatment efficacy was based on CD4+ T cell enumeration and HIV RNA level quantitation in plasma (NASBA). Baseline values were 278 CD4+/mm3 and 5.42 log RNA copies/ml. Mean changes from baseline were +102 CD4+/mm3 and -2.14 log RNA copies/ml at week 8 and +156 CD4+/mm3 and -2.07 log RNA copies/ml at week 24. HIV RNA in plasma was lower than the detection limit (2.60 log RNA copies/ml) in 55% of patients at week 8 and in 30% at week 24. No major adverse events such as neuropathy or pancreatitis were observed. Once-daily administration of didanosine in combination with twice-daily administration of zidovudine is a well tolerated regimen that appears to be as effective ad the conventional zidovudine/didanosine combination regimen.     

Quantitative HIV-1 RNA as a marker of clinical stability and survival in a cohort of 302 patients with a mean CD4 cell count of 300 x 10(6)/l

OBJECTIVE: To analyse plasma HIV-1 RNA levels as a marker of clinical stability and survival in a cohort of HIV-infected patients whose time of seroconversion is unknown. DESIGN: Retrospective cohort study. SETTING: Retrovirology laboratory and AIDS Unit in a teaching hospital. PATIENTS: A total of 916 samples from 302 patients, most on antiretroviral therapy, were analysed. Mean initial CD4 cell counts and HIV-1 RNA were 299 x 10(6)/l (range: 0-1600) and 134,261 copies/ml (range: < 200-4,300,000), respectively. Sixty-six cases had been diagnosed previously with AIDS. METHODS: Analysis of progression to AIDS and survival, according to initial and longitudinal viral load (VL) and CD4 cell count measurements was performed by Kaplan-Meier test. Relative risks were calculated by Cox's proportional hazards model. RESULTS: During a mean follow-up of 444 +/- 309 days, 29 patients developed AIDS and 21 died. Relative risk (RR) of progression related to the group with VL < 35,000 was: 10.4 when CD4 > or = 250 x 10(6)/l and VL > or = 35,000 (P = 0.001); and 45.3 when CD4 < 250 x 10(6)/l and VL > or = 35,000 (P < 0.0001). Cumulative probability of progression was: 0%, 0% and 12.3%, at the first, second and third year respectively, for patients with all their sequential VL determinations < 60,000; and 13.3%, 34.7% and 79.3% for patients who did not maintain VL values always < 60,000 (RR = 23; P < 0.0001). The minimum value of VL that reached statistical significance for the survival analysis was 100,000 copies/ml (P < 0.0001). CONCLUSIONS: VL > or = or < 35,000 is a better discriminant for progression than a CD4 cell count > or = or < 250 x 10(6)/l. Sequential VL determinations < 60,000 are associated with a better prognosis.     

Evaluation of guidelines for initiation of highly active antiretroviral therapy in a longitudinal cohort of HIV-infected individuals

OBJECTIVES: Expert panels have developed several guidelines for initiating highly active antiretroviral therapy (HAART) in patients with HIV infection. To evaluate these guidelines, we simulated their application in a cohort of HIV-infected patients established and followed before HAART was available, and determined how long such patients survived without disease progression in the absence of HAART. METHODS: Longitudinal data was used that had been collected from 1982 to 1995 on a prospective cohort of 133 homosexual men with known or closely approximated dates of HIV-1 seroconversion and negligible antiretroviral exposure. The main definition of disease progression was CD4 cell count < or = 300x10(6)/l or development of clinical AIDS diagnosis within 12 months. RESULTS: The mean number of years between the recommended initiation of therapy and when disease progression occurred in the absence of HAART were as follows: initiation of treatment at first visit, 4.81 years [median, 3.78 years; interquartile range (IQR), 1.85-6.59 years]; CD4 cell count <500x10(6)/l or serum RNA >5000 copies/ml (at least 10000 copies/ml fresh plasma), 4.35 years (median, 3.22 years; IQR, 1.56-6.19 years); CD4 cells <500x10(6)/l or serum RNA >20000 copies/ml (at least 40000 copies/ml fresh plasma), 3.61 years (median, 2.70 years; IQR, 1.40-5.11 years); and CD4 cells <500x10(6)/l, 2.72 years (median, 2.17 years; IQR, 0.81-4.25 years). The percentage of patients who had disease progression before HAART would have been recommended was 0.8, 1.6, 3.2 and 13.6% with each of these four approaches, respectively. CONCLUSIONS: Implementation of recommended treatment guidelines will result in a substantial proportion of patients being treated for long periods before immunologic or clinical disease progression would have occurred in the absence of HAART. These findings should be considered in the clinical care of HIV-infected patients and in future recommendations for the initiation of HAART.     

The relative prognostic value of plasma HIV RNA levels and CD4 lymphocyte counts in advanced HIV infection

OBJECTIVE: It has been suggested that the plasma HIV RNA level is a better predictor of AIDS and death than the CD4 lymphocyte count. We assessed whether the prognostic value of plasma virus levels was different according to the CD4 count. DESIGN: Prospective cohort study of HIV-infected patients followed for a median of 2.91 years (range, 0.02-4.54). SETTING: Department of Infectious Diseases at Rigshospitalet, Copenhagen, Denmark. PARTICIPANTS: A group of 255 HIV-infected individuals with an initial measurement of CD4 lymphocyte count and plasma HIV RNA. MAIN OUTCOME MEASURE: Survival time. RESULTS: The plasma HIV RNA (median 101410 copies/ml; range (range 200-7200000) and the CD4 lymphocyte count (median 250 cells x 10(6)/l; range 1-1247) were negatively correlated (Pearson r = -0.53; P < 0.00001). Of the 255 patients, 110 died during follow-up. Overall, a higher HIV RNA level was associated with increased risk of death, but the association was smaller in patients with lower CD4 lymphocyte counts (test for interaction P < 0.0001). In patients with CD4 count below 50 cells x 10(6)/l the association between HIV RNA and risk of death was not statistically significant (relative hazard per 10-fold higher HIV RNA level was 1.53; P = 0.11; adjusted for age and CD4 count) while that between the CD4 count and risk of death was highly significant (relative hazard per 50% lower CD4 count 1.38; P = 0.005; adjusted for age and HIV RNA level). CONCLUSIONS: Patients were relatively lightly treated with antiretroviral drugs both before and during this study. In this situation, it appears that the HIV RNA level has a relatively weak association with risk of death in patients with advanced HIV infection and that the CD4 lymphocyte count is probably more useful in assessing prognosis.     

A pilot study of hydroxyurea among patients with advanced human immunodeficiency virus (HIV) disease receiving chronic didanosine therapy: Canadian HIV trials network protocol 080

To assess the in vivo short-term antiretroviral effect of hydroxyurea in human immunodeficiency virus (HIV)-infected persons chronically treated with didanosine (ddI), 26 patients with CD4 cell counts between 100 and 350 were enrolled in a 12-week, open-label pilot study and randomly assigned to receive 500 or 1000 mg/day hydroxyurea. Clinical status, laboratory toxicities, CD4 lymphocyte count, and HIV RNA plasma virus load were assessed weekly. Median declines from baseline of 0.02 and 0.63 log10 HIV-1 RNA copies/mL of plasma were observed for the 500- and 1000-mg/day groups, respectively (P = .02). CD4 cell counts did not change significantly with the addition of hydroxyurea; however, a small but statistically significant decrease in counts was observed during the washout phase. Both doses of hydroxyurea were well-tolerated. These results demonstrate a substantial decrease in plasma virus load when 1000 mg of hydroxyurea is administered over 1 month as adjunctive therapy to ddI among HIV-infected persons with 100-350 CD4 cells/mm3.     

Treatment history and baseline viral load, but not viral tropism or CCR-5 genotype, influence prolonged antiviral efficacy of highly active antiretroviral treatment

BACKGROUND: The efficacy of highly active antiretroviral treatment (HAART) in HIV-1 disease may vary between nucleoside-naive and experienced patients as well as between patients with different viral phenotypes and in different stages of disease. OBJECTIVE: To investigate variables of importance for successful long-term viral suppression by analysing virological, clinical and immunological characteristics at initiation of protease inhibitor treatment on suppression of HIV RNA over 1 year. DESIGN: An open, non-randomized, observational clinical study. SETTING: Venh?lsan, Department of Dermatovenereology, S?der Hospital, Stockholm, Sweden. PATIENTS: A total of 147 unselected advanced patients with known HIV-1 infection for a mean of 7 years, of whom 37% had AIDS and who started treatment with a protease inhibitor during 1996. INTERVENTIONS: All patients received HAART with at least two nucleoside analogues in combination with either indinavir (81%) or ritonavir (19%). The majority (77%) had been previously treated with nucleoside analogues for a mean of 39 months. MEASUREMENTS: CD4+ lymphocyte count, plasma HIV-1 RNA, viral phenotype and HIV-1 coreceptor CCR-5 genotype at baseline. Viral load and CD4+ lymphocyte count were determined every 3 months. RESULTS: Patients were analysed on an intention-to-treat basis. The mean CD4+ lymphocyte count at baseline was 170 x 10(6)/l and the median viral load was 68 600 copies/ml. Heterozygosity for the delta32 deletion of the CCR-5 gene (delta32/wt) was found in 27%. MT-2 positive virus (syncytium-inducing) was isolated in 46%. Logistic regression revealed that nucleoside analogue experience and baseline log10 HIV-1 RNA were the only factors independently related to plasma HIV-1 RNA levels below 500 copies/ml after 1 year of treatment, which was found in 69%. CONCLUSION: The virological outcome after 1 year of HAART was strongly correlated to prior treatment history and baseline viral load, whereas CD4+ lymphocyte count, CCR-5 genotype and viral biological phenotype had less influence. The long-term antiviral efficacy of HAART was lowest in individuals with previous nucleoside analogue treatment and a high baseline viral load. In these individuals an even more aggressive treatment should be considered.     

Changes in CD4+ cell count and the risk of opportunistic infection or death after highly active antiretroviral treatment. Groupe d'Epidémiologie Clinique du SIDA en Aquitaine

OBJECTIVE: To study the relationship between the CD4+ cell response after initiation of protease inhibitors and the occurrence of opportunistic infections and survival. DESIGN: Prospective observational cohort study. METHODS: HIV-1-seropositive subjects followed-up in HIV centres of Bordeaux University Hospital, Southwest France who were prescribed at least one available protease inhibitor between January and December 1996 were included in this analysis. A Cox model estimated the independent effect of baseline covariates and CD4+ cell response, considered as a time-dependent covariate, on the occurrence of new AIDS-defining opportunistic infection, new AIDS-defining events, new AIDS-defining opportunistic infection or death. RESULTS: A total of 556 HIV-positive patients were prescribed at least one protease inhibitor: 34% saquinavir, 52% indinavir, and 14% ritonavir. Median CD4+ cell count at baseline was 95 x 10(6)/l and mean plasma HIV RNA was 5.0 log10 copies/ml. After a median follow-up of 230 days, 65 patients experienced a new episode of opportunistic infection, 79 patients experienced at least one AIDS-defining event, and 24 had died. On average, the increase in CD4+ cell count was 42 x 10(6)/l (SD, 74) after a median of 49 days. In the multivariate analysis of opportunistic infection or death, each 50% higher CD4+ cell count at baseline was associated with a 23% reduction [95% confidence interval (CI), 14-30] of risk. Each 50% increase in CD4+ cell count during follow-up was associated with a 9% reduction (95% CI, 2-15) of risk, adjusted for the presence of AIDS prior to protease inhibitor therapy (hazard ratio, 3.76 versus absence of AIDS; P < 0.01) and haemoglobin level (hazard ratio, 0.48 if > 11 g/dl versus <11 g/dl; P < 0.01). CONCLUSION: Our results show, at least indirectly, how protease inhibitors might produce clinical stabilization. This result may be due to improved functionality of CD4+ cells in patients started on protease inhibitors.     

ALL R-87 protocol in the treatment of children with acute lymphoblastic leukaemia in early bone marrow relapse

Seventy-three children with acute lymphoblastic leukaemia (ALL) in first bone marrow (BM) relapse, occurring within 30 months from complete remission (CR), were enrolled in an Italian cooperative study (ALL R-87 protocol). This treatment programme consisted of an induction phase with intermediate-dose cytarabine (IDARA-C) plus idarubicin (IDA) and prednisone (PDN), followed by a multidrug consolidation therapy and bone marrow transplant (BMT). 55/73 children achieved CR (75.3%); 15 (20.5%) failed to respond and three (4.2%) died during induction. The response rate was significantly higher for children with a first CR duration > or = 12 months (P=0.0005) and for those with a white blood cell (WBC) count at relapse < 20 x 10(9)/l (P=0.004). The estimated disease-free survival (DFS +/- SE) at 82 months was 0.18 +/- 0.05 for all responders, and 0.70 +/- 0.14 for allotransplanted patients versus 0.05 +/- 0.05 for those autografted (P=0.001). The estimated probabilities of survival +/- SE and event-free survival (EFS +/- SE) at 83 months were 0.16 +/- 0.07 and 0.13 +/- 0.04, respectively. for all enrolled children. Univariate analysis showed that age < 10 years at initial diagnosis and B-lineage immunophenotype favourably influenced both DFS (P=0.001) and EFS probabilities (P=0.0014 and P=0.012, respectively), whereas a first CR duration > or = 12 months and a WBC count at relapse < 20 x 10(9)/l were associated only with a better EFS rate (P=0.026 and P=0.004, respectively). Our results show the efficacy of the IDA plus IDARA-C schedule used in the ALL R-87 protocol in high-risk relapsed ALL children. Allogeneic BMT proved effective for patients with an HLA sibling donor. In a multivariate analysis, age > or = 10 years at initial diagnosis (P=0.016) and WBC count at relapse > or = 20 x 10(9)/l (P=0.048) were independently associated with a worse disease outcome.     

Effects of antihypertensive treatment on vasopressin secretion and on its osmoregulation in moderate hypertension

BACKGROUND: Changes in plasma osmolality and arterial pressure can affect the secretion of vasopressin (AVP). OBJECTIVE: To investigate the effect of a drug-induced lowering of the arterial pressure on the plasma concentration of AVP and on its osmoregulation in moderately severe uncomplicated hypertensives. DESIGN AND METHODS: A group of 33 moderate uncomplicated and untreated essential hypertensives of both sexes (mean age 48 +/- 1 years, average arterial pressure 171 +/- 3/108 +/- 2 mmHg) was studied. We measured AVP and other plasma and urine variables in 21 of them before and after administration of a hypertonic NaCl solution (100 mmol NaCl in 50 ml). Antihypertensive treatment with a single drug or, if necessary, with a combination of drugs was initiated for eight of these subjects and hypertonic saline administration was repeated after 1 month of treatment. The hypertonic stimulus was administered to the other 12 subjects after acute lowering of the arterial pressure by continuous intravenous infusion either of 0.3 mg clonidine in 100 ml (n = 6) or of 50 mg sodium nitroprusside in 250 ml (n = 6). RESULTS: Administration of hypertonic saline to untreated hypertensives increased their AVP level from 1.6 +/- 0.28 to 5.4 +/- 0.7 pg/ml (n = 21, P < 0.01). Their mean arterial pressure was lowered after pharmacological treatment for 1 month (n = 8) from 125 +/- 2 to 101 +/- 2 mmHg; their baseline AVP level remained unchanged (1.2 +/- 0.21 versus 0.9 +/- 0.25 pg/ml); after hypertonic saline had been administered to hypertensives with lowered arterial pressures, their AVP level increased to 6.0 +/- 1.03 pg/ml (P < 0.01). The AVP level in subjects whose MAP had been lowered acutely by administration of clonidine (n = 6) or of sodium nitroprusside (n = 6; on the average, from 132 +/- 3 to 110 +/- 4 mmHg) increased concurrently from 1.6 +/- 0.63 to 3.4 +/- 0.7 pg/ml (P < 0.05); after administration of the hypertonic saline the AVP level increased to 10.8 +/- 2.22 pg/ml (P < 0.01). This stimulated value was significantly (P < 0.01) higher than that observed after hypertonic saline had been administered to untreated hypertensives (5.4 +/- 0.7 pg/ml). CONCLUSIONS: Acute lowering of the arterial pressure in moderate essential hypertension appears to facilitate the secretion and osmoregulation of AVP. On the other hand, during prolonged antihypertensive treatment, baroreflex regulation of the secretion of AVP appears to be set at a lower operating point, thus exerting the same influence on the release of AVP as it did before antihypertensive treatment.     

Organochlorine pesticide-induced oxidative stress and immune suppression in rats

OBJECTIVE: To investigate the different responses to antiretroviral treatment including zidovudine, of patients harbouring HIV with primary resistance to zidovudine, serum viral load, and CD4+ cell counts, for 24 weeks in a group of antiretroviral-naive patients with the codon 215 mutation of the HIV pol gene and in a control group at the start of treatment. DESIGN: A case-control retrospective study (1989-1996). METHOD: Nineteen out of 210 patients previously studied harboured the codon 215 mutation, had a self-reported compliance with treatment, a minimum follow-up of 24 weeks, and were treated with zidovudine alone or in combination with other nucleoside analogues. These patients were matched with 19 patients with wild-type strains at entry by initial CD4+ cell counts, clinical status, and antiretroviral treatment. RESULTS: During the first 12 weeks, CD4+ cell counts increased (76+/-26 and 64+/-26 x 10(6)/l in wild-type and mutant virus-infected groups, respectively), decreasing slightly until week 24, although no significant differences were found between the two groups studied. Serum viral load decreased in both groups (change in serum viral load of 0.80+/-0.11 log10 and 0.87+/-0.26 log10 copies/ml, wild-type and mutant virus-infected, respectively), although no significant differences were found between groups. CONCLUSION: No significant differences were found between patients with the primary mutation to zidovudine and control patients harbouring wild-type virus in terms of short-term response measured by serum viral load and CD4+ cell counts.     

Virological treatment failure of protease inhibitor therapy in an unselected cohort of HIV-infected patients

OBJECTIVE: To determine the rate of virological treatment failure with protease inhibitor therapy in unselected patients and to assess underlying risk factors. DESIGN AND SETTING: Retrospective study in two German tertiary care treatment centres. PATIENTS: A total of 198 HIV-infected patients treated with protease inhibitors in 1996. MAIN OUTCOME MEASURES: Levels of HIV RNA 1-6 months after start of treatment; definition of treatment failure of < 1 log10 reduction in plasma HIV RNA within 6 months after starting protease inhibitor therapy; multivariate analysis of risk factors for treatment failures. RESULTS: A total of 226 treatment episodes with protease inhibitors were evaluable (saquinavir, 83; ritonavir, 47; indinavir, 96). The rate of virological treatment failure was 44% (saquinavir, 64%; ritonavir, 38%; indinavir, 30%). In a multivariate analysis, the following independent risk factors for virological failure were found: CD4 cell count, pretreatment with antiretroviral drugs (number), and protease inhibitor (compound). The relative risk reduction for each CD4 cell count increase was 0.997 (P = 0.012), 2.64 for pretreatment with one or two drugs versus no drug (P = 0.05), 2.97 for pretreatment with more than two drugs versus no drug (P = 0.05), and 4.62 for treatment with saquinavir versus indinavir (P = 0.001). CONCLUSION: An unexpectedly high rate of virological treatment failure of protease inhibitor therapy was found in an unselected cohort of HIV-infected patients. Response to antiretroviral combination therapy in normal clinical practice may considerably differ from results of randomized clinical trials. Further studies are warranted to find optimal treatment strategies for both initial and salvage therapy.