Introduction
Combination antiretroviral therapy reduces HIV RNA levels to < 200 copies/ml in the majority of HIV-infected patients [1,2]. Many individuals, however, do not achieve sustained virologic suppression below this threshold because of inadequate regimen potency or adherence, pharmacokinetics or genetic factors [3,4]. The importance of maintaining antiviral potency to sustain virologic suppression was emphasized by the rapid virologic rebound observed in three clinical studies among patients with virologic suppression who discontinued one or two drugs of a three to four drug regimen [5-7].
One approach to improve long-term rates of virologic suppression is to increase the antiviral potency of the regimen. Regimen potency can be increased by adding new drugs to a regimen or by intensifying drug exposure through alternative mechanisms. Antiviral activity of didanosine (ddI) can be enhanced by the chemotherapeutic agent hydroxyurea (HU) [8]. In contrast to HIV therapeutic agents targeted at viral enzymes, HU inhibits cellular enzymes, which may be less susceptible to developing resistance. HU inhibits ribonucleotide reductase, producing reduced levels of intracellular pools of dATP, thereby enhancing the competitive potency of the active metabolite of ddI, ddATP. HU also decreases cellular activation, although the magnitude and clinical benefit of this effect are not known.
Clinical studies document the superior antiviral activity of ddI plus HU compared with regimens containing only ddI [9-12]. To determine whether enhancing antiviral potency would improve long-term rates of viral suppression, subjects who had achieved virologic suppression while receiving indinavir (IDV) + zidovudine (ZDV) + lamivudine (3TC) were randomized to an IDV regimen containing ddI + stavudine (d4T) with or without HU or to continue on their current regimen. Because of the potential toxicity associated with increased drug exposure in patients with intensified therapy, both drug toxicity and virologic rebound were included in our definition of the primary study endpoint of treatment failure.
Methods
Patient population
Eligible subjects were at least 13 years of age, had documented HIV infection and had been treated for at least 6 months with a regimen of IDV + ZDV (or d4T) + 3TC. Subjects with a history of treatment with an HIV protease inhibitor other than IDV and subjects with a treatment history including both ddI and d4T were excluded. All subjects were required to have HIV RNA < 200 copies/ml confirmed within 50 days of study entry. Subjects were also required to have CD4 cell count of > 100 × 106 cells/l within 90 days prior to beginning their IDV regimen and to have CD4 cell count > 200 × 106 cells/l within 60 days of study entry.
Laboratory exclusion values included hemoglobin < 9.1 g/d for men and 8.9 g/d for women, absolute neutrophil count < 1000 × 106 cells/l, platelet count < 75 × 109 cells/l, serum glutamate oxaloacetate transaminase and serum glutamate pyruvate transaminase over three times the upper limit of normal, and serum amylase > 1.5 times upper limit of normal. Other exclusion criteria included a history of grade 2 peripheral neuropathy, current breast-feeding and documented or suspected hepatitis within 90 days prior to study entry.
Study design
This was a prospective, randomized, multicenter AIDS Clinical Trials Group (ACTG)-sponsored study, designated ACTG 5025. Approval by governing institutional review boards (IRB) was obtained by all participating sites, and all subjects signed IRB-approved consent forms. The study was designed as a rollover study of ACTG 343, results of which have been published [5].
Eligible subjects were randomized to receive either a regimen containing IDV + ddI + d4T + HU, IDV + ddI + d4T, or IDV + ZDV (or d4T) + 3TC. Randomization was stratified by prior participation or not in ACTG 343. ZDV and 3TC were administered in a combination tablet, Combivir, containing 300 mg ZDV and 150 mg 3TC. Subjects intolerant of ZDV were permitted to substitute d4T. IDV, ZDV and d4T were administered open label at the FDA-approved doses. Subjects who weighed ≤ 60 kg received 400 mg ddI once daily; those weighing < 60 kg received 250 mg once daily. Subjects were instructed to take ddI at least 30 min after a snack or 2 h after a meal and at least 1 h before or 2 h after IDV dosing. HU (600 mg twice daily) or a matching placebo was given to subjects randomized to the arms containing ddI + d4T.
Subjects had complete blood counts monitored every 4 weeks and were evaluated every 8 weeks with clinical and laboratory assessments, including CD4 cell counts, liver function tests and amylase. T cell subsets were performed at ACTG-certified laboratories. After study closure, stored plasma samples were evaluated for the presence of hepatitis B surface antigen. Hepatitis C antibody was used to define evidence of hepatitis C.
Dose reductions in HU were required for hematologic toxicities. For grade 1 anemia, thrombocytopenia, or neutropenia, HU was reduced to 300 mg three times daily. For recurrent grade 2 abnormalities, HU or placebo was held until the toxicity was grade 1 or less, and the HU dosage was then reduced to 300 mg twice daily. Intolerance at this dose was defined as a toxicity endpoint. Subjects with pancreatitis were defined as a study toxicity endpoint and were permitted to substitute ZDV + 3TC for ddI + d4T + HU. Subjects with asymptomatic elevations in amylase or lipase (over twice upper limit of normal for 2 weeks) while receiving ddI + d4T were also considered a study endpoint. These subjects were required to discontinue these two nucleosides and HU (or placebo) and permitted to resume ZDV + 3TC.
Study endpoints
The primary study endpoint was either confirmed HIV RNA > 200 copies/ml or a treatment-limiting toxicity, whichever occurred first. HIV RNA values were quantified at a central laboratory using the Roche Amplicore ultrasensitive assay (Roche Molecular Systems, Nutley, NewJersey, USA). The limit of detection of the assay was 50 copies/ml. Treatment-limiting toxicity endpoints included those with protocol-mandated study drug discontinuation and those who found the toxicity intolerable even though the protocol did not require treatment discontinuation.
Study termination
The study was prematurely terminated on September 22, 1999, based on the recommendations of an ACTG study monitoring committee review of the interim analysis, which showed a significantly higher treatment failure rate in the HU-containing arm compared with the IDV + ZDV + 3TC arm. In the interim review, the protocol-specified Lan DeMets/O'Brien-Fleming critical stopping value was crossed for the pairwise comparisons between the HU-containing arm and each of the other two arms.
Pancreatitis case control study
To address risk factors for pancreatitis, a case-control study was performed after the study was terminated. Between two and four controls were matched for each of the seven patients with clinical pancreatitis by treatment arm and study site. Matched logistic regression models were used to assess univariable associations between the risk of pancreatitis and baseline variables [13].
Statistics
The primary analyses were intention-to-treat where outcome was assessed based on initial randomization. The time to treatment failure was compared among the study arms using Kaplan-Meier curves and a log rank test stratified by prior participation in ACTG 343.
Secondary analyses compared treatments by time to virologic failure. This analysis was intention-to-treat, for which subjects who discontinued treatment were not censored. Differences in CD4 cell counts between baseline and week 4 and between baseline and week 24 were compared across arms with the non-parametric Kruskal-Wallis test [14]. Additional analyses compared treatments by time to virologic failure with toxicity endpoints censored and by time to toxicity endpoint with virologic endpoints censored. These latter competing risk analyses assessed the relative extent to which treatment differences in the composite primary endpoint are explained by virologic versus toxicity events.
Cox regression models were used to assess prognostic factors for the progression to the primary endpoint, the virologic endpoint and the toxicity endpoints [15]. For all tests, a P value of 0.05 was used as the cutoff to judge statistical significance. This critical value is slightly liberal, as early stopping can spuriously increase the strength of significance.
Results
Between November 1998 and July 1999, 207 subjects were randomized to treatment. Five subjects had baseline HIV RNA > 200 copies/ml, leaving an eligible study population of 202 subjects. The median follow-up time for the study was 40 weeks. Seven subjects in the IDV + ddI + d4T + HU arm and no subjects in the other two arms went off study prior to the completion of the study. Five of these seven subjects had a primary endpoint before going off study, and two subjects withdrew without having an endpoint.
Table 1 describes the composition of the study population. The median prior exposure to IDV was 86 weeks, and the median CD4 cell count was 770 × 106 cells/l. HIV RNA was < 50 copies/ml at both pre-entry and entry measurements for 83%. The treatment arms were balanced at baseline except for hepatitis B surface antigenemia and history of injection drug use. Eight subjects in the IDV + ZDV + 3TC arm and one subject each in the other two arms tested positive for hepatitis B surface antigen (P = 0.005). There was an 18% history of injection drug use in the HU-containing arm compared with 10% in the other arms (P = 0.053).
Treatment failure
Treatment failure occurred in 22/68 (32.4%) of subjects randomized to the IDV + ddI + d4T + HU arm, in 12/68 (17.6%) in the IDV + ddI + d4T arm, and in 5/66 (7.6%) in the IDV + ZDV + 3TC arm (Table 2). The Kaplan-Meier estimates of the time to treatment failure are shown in Fig. 1. Subjects in the HU arm experienced treatment failure more rapidly than those in the IDV + ZDV + 3TC arm (P < 0.001) or the IDV + ddI + d4T arm (P = 0.032). The relative risk (RR) of failure of those in the HU arm was 5.25 [95% confidence interval (CI), 1.99-13.90] compared with the IDV + ZDV + 3TC arm and 2.12 (95% CI, 1.05-4.30), compared with the IDV + ddI + d4T arm.
Baseline predictors of the primary study endpoint were evaluated in a Cox multivariate model. HIV RNA > 50 copies/ml (RR, 2.21; 95% CI, 1.07-4.55;P = 0.031); higher white blood cell count (RR, 1.40 per 109 cells/l; 95% CI, 1.14-1.71;P = 0.0014); prior IDV use (RR, 2.18 per 24 weeks; 95% CI, 1.26-3.78;P = 0.0053); and hepatitis C antibody positivity (RR, 3.20; 95% CI, 1.50-6.81;P = 0.0026) were associated with greater risk of treatment failure. When treatment arm was added to this model, the covariate-adjusted RR for failure was 3.58 (95% CI, 1.33-9.67) for the IDV + ddI + d4T + HU arm compared with the IDV + ZDV + 3TC arm (P = 0.012). There was no difference between the IDV + ddI + d4T and IDV + ZDV + 3TC arms (P = 0.27). The adjusted RR for failure in the IDV + ddI + d4T + HU arm was 1.93 (95% CI, 0.89-4.17) compared with the IDV + ddI + d4T arm (P = 0.096).
Treatment failure was triggered by virologic rebound in 16 subjects and by toxicity in 23 subjects (Table 2). In the competing risks analysis of the time to a virologic endpoint with censoring by toxicity endpoints, there were no significant differences among the three study arms (Fig. 2). In contrast, subjects randomized to the HU arm progressed to a toxicity endpoint more quickly than those in the IDV + ZDV + 3TC arm (RR, 8.72; 95% CI, 1.99-38.2;P < 0.001) or in the IDV + d4T + ddI arm (RR, 2.77; 95% CI, 1.08-7.51;P = 0.028). These data indicate that the differences in the primary study endpoint rate of treatment failure was primarily a result of differences in toxicity and not virologic failure. In a univariate analysis of baseline predictors of toxicity-related treatment discontinuation, hepatitis C was associated with an increased risk of toxicity (RR, 3.61; 95% CI, 1.41-9.22;P = 0.0041).
In the secondary intention-to-treat analysis of virologic failure, where toxicity endpoints were not censored, rates of virologic failure were higher in the HU-containing arm compared with the other two arms. The RR of virologic failure of IDV + d4T + ddI + HU was 1.31 (P = 0.045), compared with the IDV + d4T + ddI arm and 3.66 (P = 0.007) compared with the IDV + ZDV + 3TC arm.
Treatment toxicity and deaths
The most frequently observed treatment-limiting toxicity was pancreatitis (Table 3). Three subjects each in the IDV + ddI + d4T + HU and the IDV + ddI + d4T arms developed pancreatitis. Additionally, one subject in the HU-containing arm developed pancreatitis 4 weeks after discontinuing study medications because of elevated liver function tests and two subjects in the HU-containing arm discontinued medication for asymptomatic elevation of amylase. The estimated pancreatitis rate was 9.57/100 person-years (95% CI, 2.97-22.24) in the HU-containing arm and 6.55/100 person-years (95% CI, 1.63-16.98) for the IDV + ddI + d4T arm. The pancreatitis rate pooled over both arms containing IDV + ddI + d4T was 7.99/100 person-years (95% CI, 3.43-15.45). For the IDV + ZDV + 3TC arm, the estimated pancreatitis rate was 0/100 person years (95% CI, 0-4.18). Four subjects (three receiving IDV + ddI + d4T + HU and one receiving IDV + ddI + d4T) discontinued medications for elevation in liver function tests. Treatment was discontinued for hematologic toxicity (anemia) in only one subject, who was in the HU-containing arm.
Three patients died during the study, all randomized to the HU-containing arm. Pancreatitis was contributory to outcome in two of these. The first subject was a 49-year-old obese male with a baseline CD4 cell count of 1073 × 106 cells/l and a baseline triglyceride of 1205 mg/dl. On study week 12, the patient complained of abdominal pain; his amylase was within normal limits. On study week 14, the patient developed abdominal pain and elevated lipase and was hospitalized with a diagnosis of pancreatitis. The patient's clinical course was complicated by a pancreatic pseudocyst, ascites and polymicrobial peritonitis. One day after attempted surgical drainage of the necrotizing pancreatic pseudocyst, the patient died.
The second patient was a 51-year-old male with a history of elevated triglycerides and a baseline non-fasting triglyceride of 770 mg/dl. On study week 9, the patient developed epigastric pain and was treated for esophageal reflux and costochondritis. On study week 14, he was hospitalized with acute pancreatitis. His lipase was 2653 U/l, amylase was within normal limits and the anion gap was 26. The patient's pancreatitis symptoms improved with discontinuation of study treatment but his liver function deteriorated, and he died of Escherichia coli sepsis. Autopsy showed hemorrhagic pancreatitis and microvesicular steatosis in the liver, consistent with nucleoside analogue toxicity.
The third subject was a 48-year-old male with a history of cerebrovascular disease. On study week 16, he developed a left middle cerebral artery stroke. His hospital course was complicated by acute pancreatitis and by a myocardial infarction. He died after developing an aspiration pneumonia.
Pancreatitis case-control study
Seven subjects were diagnosed with clinical pancreatitis (Table 4). In one of these, the primary reason for treatment discontinuation was elevation in liver function tests, which preceded the development of pancreatitis. Among the seven, the median time on study treatment until the development of pancreatitis was 12 weeks (range, 4-16). Clinical presentation was characterized by abdominal pain (five), nausea (six), and vomiting (five). In four patients, lipase was elevated and amylase was within normal limits. In one, only amylase was elevated, and in two, both amylase and lipase were elevated. Pancreatitis was complicated by pseudocyst formation in one subject, who ultimately died. In surviving patients, the median time to resolution of symptoms was 6 weeks (range, 1-14).
In the analyses aimed at identifying risk factors for the development of pancreatitis, none of the following factors was associated with the development of pancreatitis: age, weight, gender, injection drug use, history of gallstones, history of alcohol abuse, history of lipid abnormality, hepatitis B surface antigen reactivity, hepatitis C antibody positivity or baseline elevation in non-fasting triglycerides or cholesterol. Elevations in non-fasting triglycerides were present in four of the seven subjects with clinical pancreatitis and in 7 of 16 controls. Two of the seven subjects with pancreatitis (28.6%) tested positive for hepatitis C antibody compared with 5 of 23 controls (21.7%). None of the subjects who died had evidence of hepatitis B or C infection.
CD4 cell count analysis
Subjects randomized to the HU arm had significant reductions in CD4 cell counts compared with the other two treatment arms at week 4. The median CD4 cell count was decreased by 101 × 106 cells/l for the HU-treated subjects, increased by 6 × 106 cells/l for the IDV + ZDV + 3TC arm, and increased by 8 × 106 cells/l for the IDV + ddI + d4T arm (Fig. 3). The differences were significant between the HU arm and the IDV + ZDV + 3TC arm (P < 0.001) and the IDV + ddI + d4T arm (P < 0.001). The changes in percentage of CD4 cell counts at 24 weeks were small: -1, -1 and 1 for the HU arm, IDV + ddI + d4T arm and IDV + ZDV + 3TC arm, respectively. There was no association between the decline in absolute CD4 cell counts and HIV-related clinical events in the HU arm.
Discussion
Current approaches to antiretroviral therapy emphasize the importance of achieving and maintaining virologic suppression [16,17]. Lower levels of HIV RNA are associated with slower rates of HIV disease progression and decreased likelihood of emergence of drug-resistant virus populations [18-20]. In patients with virologic suppression receiving standard antiretroviral therapy, increasing antiviral potency could be beneficial by reducing the frequency of virologic rebound but could also prove detrimental by increasing toxicity above an acceptable threshold. In this study, therapy intensification was associated with a worse outcome, namely toxicity. Although the comparison of tolerance between IDV + ddI + D4T + HU and patients preselected for tolerance for IDV + ZDV + 3TC was biased to favor the latter, excess toxicity and deaths observed in the HU-containing arm posed a risk that far outweighed long-term improvements of virologic suppression.
The frequency and magnitude of the toxicities observed in the HU-containing arm were unanticipated. HU has been utilized for more than 30 years as a chemotherapeutic agent, is often administered for years and has not been associated with life-threatening toxicity [21]. In published studies, the most common toxicities observed in HIV-infected patients receiving HU are hematologic: anemia, thrombocytopenia and neutropenia [9-12,22,23]. In addition, inhibition of leukocyte proliferation reduces CD4 cell counts or blunts CD4 cell responses to new therapy. In one randomized study comparing subjects initiating ddI + d4T with and without HU, CD4 cell count increases were 28 and 107 × 106 cells/l, respectively [11]. In the present study, the HU-containing arm was associated with a mean decrease of 100 × 106 cells/l. However, the CD4 cell percentages did not diminish, and HIV-associated events were not more frequent. In the present study, non-hematologic toxicities were the most serious and frequent adverse events that occurred in the HU arm.
Why was the observed toxicity so great in the HU arm in this study? The dose of HU utilized may have been partially responsible. At the time this study was designed, pilot data suggested some benefit to higher HU doses. Although up to 2 g HU in combination with ddI have been administered in short pilot studies in HIV-infected patients, the 1200 mg daily dosage utilized in this study was higher than the daily dosage of 1 g used in HIV trials of longer duration. Either a direct effect of HU itself or higher levels of exposure to the ddI active metabolite may have contributed to the observed toxicity.
Pancreatitis was the most commonly occurring toxicity in this study. Pancreatitis occurred more frequently in the ddI-containing arms, and the rates were higher than those previously reported. Across ddI studies at currently utilized doses, rates of pancreatitis range from 1 to 7%, with estimated mortality of 0.3%[24]. In the HU-containing arm of this study, augmenting intracellular exposure to the active ddI metabolite may have contributed to the frequency and severity of this toxicity. In a review of pancreatitis across 20 ACTG studies, the highest rates of pancreatitis were observed in studies using daily doses of ddI of 500-750 mg. For example, in ACTG 118 the incidence rates were 14.3 (95% CI, 9.7-21.1) and 11.5 (95% CI, 7.7-17.5) for the 750 and 500 mg daily dosage of ddI, respectively [25]. While these data argue that increased drug exposure increases pancreatitis risk, it is difficult to compare rates among studies. Many confounding variables are present, such as stage of HIV disease, which is itself associated with a higher risk of pancreatitis in the absence of drug therapy, and use of concomitant medications, such as pentamidine [26].
In addition to ddI use, risk factors for pancreatitis include history of pancreatitis, alcohol abuse, obesity, elevated triglycerides, cholelithiasis and other medications known to cause pancreatitis. In our case control study, we were unable to identify other factors that predisposed patients to pancreatitis. One hypothesis to explain higher pancreatitis rates in this study is that potent antiretroviral therapy induced metabolic and lipid perturbations that increased pancreatitis risk. However, our analysis to identify metabolic and other risk factors may have been limited by our small study size and our evaluation of lipid parameters, which included only non-fasting cholesterol and triglycerides.
Pancreatitis was present in all three subjects who died in the study. In the first subject, complications of pancreatitis led to a protracted clinical course characterized by development of a pancreatic pseudocyst and abscess that required surgical drainage. These complications are rare but are associated with a high mortality rate [27]. The second patient developed pancreatitis on study week 14 but also had unexplained elevations in liver function tests. His clinical pancreatitis improved with discontinuation of study medications, but he developed an elevated anion gap acidosis and progressive liver failure. Autopsy findings were consistent with hepatic steatosis attributable to nucleotide toxicity. Although tissue samples were not suitable for electron microscopic examination to document mitochondrial toxicity, the patient's clinical course was more consistent with this diagnosis than pancreatitis alone [28]. Recent reports emphasize the presence of pancreatitis in this syndrome.
Monitoring patients for the development of pancreatitis is an essential aspect in caring for patients receiving ddI. Continued administration of ddI in the presence of pancreatitis can worsen outcome. Frequent clinical assessments and measurement of amylase were included in this study and resulted in discontinuation of treatment for unexplained elevations in amylase. Even with this level of monitoring, the pancreatitis diagnosis may have been delayed if clinicians relied on measurements of amylase only. In four subjects, including the two who died, significant elevations in serum lipase preceded rises in serum amylase. Based on the data, monitoring for pancreatitis should include not only patient education and rigorous clinical evaluation but also measurement of both amylase and lipase.
In addition to pancreatitis, neuropathy and elevated liver function tests were reasons for treatment discontinuation. Both ddI and d4T increase the risk for neuropathy, and a recent study has demonstrated that HU, when added to ddI and d4T, is also associated with increased risk of neuropathy [29]. Hepatic steatosis and fatal hepatic injury have been reported rarely with nucleoside treatment, although milder forms of this toxicity have recently been reported and may explain abnormalities in this study [30,31]. Underlying hepatitis also increases the risk for liver dysfunction and drug toxicity, and HU may produce hepatic injury in HIV-infected patients receiving potent antiretroviral therapy [32].
Findings in this study emphasize the importance of balancing drug toxicity with antiviral efficacy in the management of HIV-infected patients. They also underscore the importance of careful monitoring for drug toxicity. Since this study was completed, another randomized trial comparing efavirenz + ddI + d4T with or without HU was prematurely discontinued because of excess toxicity in the HU-containing arm [33]. Results of this and other clinical trials should be used to reassess all trials of HU. In the setting of salvage therapy, the risk-benefit ratio may sometimes favor the use of HU, which may augment the activity of other drugs, such as tenofovir. In countries with limited resources, HU may provide one of the few acceptable therapeutic options because of its low cost. Regarding future studies of therapeutic intensification, this strategy should continue to be evaluated in a research setting, but with more tolerable regimens.
Acknowledgments
We thank all the patients participating in the study; the study coordinators and site investigators; the study pharmacist Ana Martinez; the data manager Carol Suckow; Bristol-Myers Squibb for donating ddI, d4T and HU; Glaxo Wellcome for donating Combivir (zidovudine and lamivudine); Merck and Co. for donating indinavir. We also appreciate the assistance of Bryna Block, Kim Schulze and Colin Reed in manuscript preparation.
References
1. Gulick RM, Mellors JW, Havlir D. et al. Treatment with indinavir, zidovudine, and lamivudine in adults with human immunodeficiency virus infection and prior antiretroviral therapy. N Engl J Med 1997, 337: 734-739.
2. Staszewski S, Morales-Ramirez J, Tashima KT. et al. Efavirenz plus zidovudine and lamivudine, efavirenz plus indinavir, and indinavir plus zidovudine and lamivudine in the treatment of HIV-1 infection in adults. Study 006 Team. N Engl J Med 1999, 341: 1865-1873.
3. Paterson DL, Swindells S, Mohr J. et al. Adherence to protease inhibitor therapy and outcomes in patients with HIV infection. Ann Intern Med 2000, 133: 21-30.
4. O'Brien T, McDermott D, Ioannidis J. et al. Effect of chemokine receptor gene polymorphisms on the response to potent antiretroviral therapy. AIDS 2000, 14: 821-826.
5. Havlir DV, Marschner IC, Hirsch MS. et al. Maintenance antiretroviral therapies in HIV infected patients with undetectable plasma HIV RNA after triple-drug therapy. AIDS Clinical Trials Group Study 343 Team. N Engl J Med 1998, 339: 1261-1268.
6. Pialoux G, Raffi F, Brun-Vezinet F. et al. A randomized trial of three maintenance regimens given after three months of induction therapy with zidovudine, lamivudine, and indinavir in previously untreated HIV-1-infected patients. Trilège (Agence Nationale de Recherches sur le SIDA 072) Study Team. N Engl J Med 1998, 339: 1269-1276.
7. Reijers MH, Weverling GJ, Jurriaans S. et al. Maintenance therapy after quadruple induction therapy in HIV-1 infected individuals: Amsterdam Duration of Antiretroviral Medication (ADAM) study. Lancet 1998, 352: 185-190.
8. Lori F, Malykh A, Cara A. et al. Hydroxyurea as an inhibitor of human immunodeficiency virus-type 1 replication. Science 1994, 266: 801-805.
9. Montaner JS, Zala C, Conway B. et al. A pilot study of hydroxyurea among patients with advanced human immunodeficiency virus (HIV) disease receiving chronic didanosine therapy: Canadian HIV trials network protocol 080. J Infect Dis 1997, 175: 801-806.
10. Biron F, Lucht F, Peyramond D. et al. Anti-HIV activity of the combination of didanosine and hydroxyurea in HIV-1-infected individuals. J Acquir Immune Defic Syndr Hum Retrovirol 1995, 10: 36-40.
11. Rutschmann OT, Opravil M, Iten A. et al. A placebo-controlled trial of didanosine plus stavudine, with and without hydroxyurea, for HIV infection. The Swiss HIV Cohort Study. AIDS 1998, 12: F71-F77.
12. Foli A, Lori F, Maserati R, Tinelli C, Minoli L, Lisziewicz J. Hydroxyurea and didanosine is a more potent combination than hydroxyurea and zidovudine. Antiviral Therapy 1997, 2: 31-38.
13. Hosmer DW, Lemeshow S. Applied Logistic Regression. New York: Wiley; 1989.
14. Agresti A. Categorical Data Analysis. New York: Wiley; 1990.
15. Kalbfleisch JD, Prentice RL. The Statistical Analysis of Failure Time Data. New York: Wiley, 1980.
16. US Department of Health and Human Services and the Henry J Kaiser Family Foundation.
Guidelines for the use of antiretroviral agents in HIV-infected adults and adolescents. 2000,
http://www.hivatis.org/trtgdlns.html. 17. Carpenter CCJ, Cooper DA, Fischl MA. et al. Antiretroviral therapy in adults - updated recommendations of the International AIDS Society-USA Panel. JAMA 2000, 283: 381-390.
18. Coffin JM. HIV population dynamics in vivo: implications for genetic variation, pathogenesis, and therapy. Science 1995, 267: 483-489.
19. Mellors JW, Muñoz A, Giorgi JV. et al. Plasma viral load and CD4+ lymphocytes as prognostic markers of HIV-1 infection. Ann Intern Med 1997, 126: 946-954.
20. Mellors JW, Rinaldo CR Jr, Gupta P, White RM, Todd JA, Kingsley LA. Prognosis in HIV-1 infection predicted by the quantity of virus in plasma. [Published erratum appears in Science 1997, 275:14.] Science 1996, 272:1167-1170.
21. Donehower RC. An overview of the clinical experience with hydroxyurea. Semin Oncol 1992, 19: 11-19.
22. Hellinger JA, Iwane MK, Smith JJ. et al. A randomized study of the safety and antiretroviral activity of hydroxyurea combined with didanosine in persons infected with human immunodeficiency virus type 1. J Infec Dis 2000, 181: 540-547.
23. Giacca M, Zanussi S, Comar M. et al. Treatment of human immunodeficiency virus infection with hydroxyurea: virologic and clinical evaluation. J Infect Dis 1996, 174: 204-209.
24. Bristol-Myers Squibb. Didanosine Package Insert. Princeton, NJ: Bristol-Myers Squibb; 2000.
25. Murphy R, Reisler R, Liou S et al. Pancreatitis incidence rates of patients treated in 20 adult antiretroviral treatment trials. XIII International Conference on AIDS. Durban, July, 2000 [WePeB4255].
26. Dassopoulos T, Ehrenpreis ED. Acute pancreatitis in human immunodeficiency virus-infected patients: a review. Am J Med 1999, 107: 78-84.
27. Baron TH, Morgan DE. Acute necrotizing pancreatitis. [Published erratum appears in N Engl J Med 1999, 341:460.] N Engl J Med 1999, 340:1412-1417.
28. Miller KD, Cameron M, Wood LV, Dalakas MC, Kovacs JA. Lactic acidosis and hepatic steatosis associated with use of stavudine: Report of four cases. Ann Intern Med 2000, 133: 192-196.
29. Moore RD, Wong WM, Keruly JC, McArthur JC. Incidence of neuropathy in HIV-infected patients on monotherapy versus those on combination therapy with didanosine, stavudine and hydroxyurea. AIDS 2000, 14: 273-278.
30. Weissman SB, Sinclair GI, Green CL, Fissell WH. Hydroxyurea-induced hepatitis in human immunodeficiency virus-positive patients. Clin Infect Dis 1999, 29: 223-224.
31. Lonergan JT, Behling C, Pfander H, Hassanein TI, Mathews WC. Hyperlactatemia and hepatic abnormalities in 10 human immunodeficiency virus-infected patients receiving nucleoside analogue combination regimens. Clin Infect Dis 2000, 31: 162-166.
32. Sulkowski MS, Thomas DL, Chaisson RE, Moore RD. Hepatotoxicity associated with antiretroviral therapy in adults infected with human immunodeficiency virus and the role of hepatitis C or B virus infection. JAMA 2000, 283: 74-80.
33. Murphy R, Katlama C, Autran B et al. The effects of hydroxyurea or placebo combined with efavirenz, didanosine, and stavudine in treatment naive and experienced patients: preliminary 24 week results from the 3d study. XIII International Conference on AIDS. Durban, July, 2000 [WeOrB603].
Appendix
The following individuals and institutions participated in the performance of this trial: Laura Stiffler, G.-Y. Kim and William G. Powderly (Washington University, St Louis); Deb Heise, Beth Zwickl and Jean Decker (Indiana University, Indianapolis); Michael F. Para, Charlotte Mills, Kathy Watson and Laura Laughlin (Ohio State University, Columbus); Charles Van der Horst (University of North Carolina, Chapel Hill); Beck A. Royer, Sheryl Storey, Lynn Houseworth and N. Jeanne Conley (University of Washington, Seattle); Henry Balfour (University of Minnesota, Minneapolis); Margaret A. Fischl, Ernesto G. Scerpella, Allan Rodriguez and Roberto Monroig (University of Miami, Coral Gables); Judith Feinberg (University of Cincinnati, Cincinnati); Linda Meixner (University of California at San Diego); Beverly Putnam, Nancy Madinger and Graham Ray (University of Colorado, Denver); Ilene Wiggins, Dorcas Baker, Megan Ossing and Denise A. Wright (Johns Hopkins University, Baltimore); Dennis Israelski (San Mateo County AIDS Program, San Mateo); Jane Norris (Willow Clinic, Menlo Park); Pat Cain (Stanford University, Stanford); Debbie Slamowitz (San Mateo Clinic, San Mateo); Barbara Gripshover, Ann Conrad, Angela Davidson and Michael Banchy (Case Western University, Cleveland); Mary Albrecht and Helen Fitch (Beth Israel Deaconess Medical Center, Boston); Teri Flynn and Amy Sbrolla (Massachusetts General Hospital, Boston); Juan J. L. Lertora, Mark A. Beilke, David M. Mushatt and Russell A. Strada (Tulane University School of Medicine, New Orleans); Nancy Hanks and Lyle Oshita (University of Hawaii, John A. Burns School of Medicine, Honolulu); Scott Souza (Pharmacy Department, Queen's Medical Center, Honolulu); Anita Caballero, Gloria Carrera and Michael J. Borucki (University of Texas, Galveston); Mario Guerrero (Harbor-UCLA Medical Center, Los Angeles); Mary Shoemaker and Richard Reichman (University of Rochester, Rochester); Donna Mildvan, Ron D'Amico, Sanjiv Shah and Steven Nowling (Mount Sinai Medical Center, New York); Betty McCulloch and Steve Yoon (University of Alabama, Birmingham); Cynthia Leissinger (Tulane University, New Orleans). Cited Here...
© 2001 Lippincott Williams & Wilkins, Inc.