Introduction
Highly active antiretroviral treatments (HAART), combining two nucleoside analogues and a protease inhibitor, induce marked reduction of the viral load to below the limits of detection of plasma HIV-1 RNA by the polymerase chain reaction (PCR) technique and an increase in the CD4 cell count [1]. An at least partial restoration of the immune function, in the absence of prior severe alteration, appears to be reflected by progressive recovery of part of the CD4 T-cell reactivity to memory antigens, although T-lymphocyte activation is decreased [2].
The course of HIV infection is profoundly modified with HAART: opportunistic infections are less frequent and survival is prolonged [3,4] and the number of hospitalizations of seropositive patients for HIV-related conditions is decreased [5,6]. The beneficial effect of protease inhibitor on the reduction of opportunistic infections and mortality is also observed in the most severely immunodepressed patients [7].
These data encourage the conduct of therapeutic trials designed to stop secondary prophylaxis of cytomegalovirus (CMV) infection when the CD4 cell count rises above 100 × 106 cells/l and raises the question of stopping all primary prophylaxis when the absolute CD4 cell count is maintained above 200 × 106 cells/l, with a percentage exceeding 15%, despite recent recommendations of the US Public Health Service - Infectious Disease Society of America (USPHS-IDSA) indicating the need for maintenance of prophylaxis despite a rise in CD4 cells [8].
However, few data are available concerning the type of infections observed during triple combination therapy and the real protective effect provided by the rise in CD4 cell count, especially in patients at the most advanced stages of the disease. The development of CMV retinal infection has been previously reported during the months following introduction of treatment with indinavir or ritonavir, even when the CD4 cell count had sometimes risen to above 100 × 106 cells/l [9,10]. In study H94-247, in severely immunode-pressed patients (mean CD4 count: 21.3 × 106 cells/l) continuing their previous treatment, comparison of the effect of ritonavir versus placebo showed improvement of survival and a decreased incidence of opportunistic infections, although the incidence of CMV infections and Mycobacterium avium complex diseases was not significantly different between the two treatment groups [7].
In order to study the influence of highly active anti-retroviral treatments on the course of the disease, in the most severely immunodepressed patients, we analyzed the onset of AIDS-defining clinical events in patients who had recently started treatment with protease inhibitors associated with nucleoside analogues.
Patients and methods
This retrospective study, conducted in seven University hospitals in the west of France, included all patients who began treatment with ritonavir and/or indinavir between 26 March 1996, the date of release of ritonavir in France, and 31 December 1996. The main evaluation criterion was the development or recurrence of an AIDS-defining clinical event.
Clinical events were censored until 31 January 1997, and the following data were recorded: mode and date of onset of the first clinical signs, criteria and date of diagnosis, treatment received and clinical course. These data were completed by determination of the CD4 and CD8 cell counts and quantification of plasma HIV-1 RNA before introduction of the protease inhibitor (or during the previous 3 months when there was no modification of antiretroviral drugs), and at the time of onset of the opportunistic event (± 15 days).
The main opportunistic infections were diagnosed on the basis of the following criteria: (i) for CMV: ocular fundus examination, completed by retinography in the case of a suspicious lesion, confirmed by PCR on the vitreous humor in doubtful cases, detection of viral DNA in cerebrospinal fluid or presence of the virus, especially in gastrointestinal biopsies. The date adopted to calculate the time to onset of CMV infection corresponded to the date of appearance of the first clinical signs in the case of symptomatic retinitis or the date of the routine ophthalmological examination revealing asymptomatic retinitis. (ii) For M. avium complex, evidence of positive culture from blood or another normally sterile site and for other mycobacteria: direct examination and culture of pathological specimens, gastric fluid collected by intubation or bronchoalveolar lavage fluid, abscess aspiration or pus sample. (iii) For oesophageal candidiasis: endoscopic demonstration of characteristic lesions or clinical history of severe dysphagia associated with buccal thrush, responding to antifungal treatment. The diagnosis of the other AIDS-defining events were based on the CDC 1987 classification report [11], and criteria commonly used in clinic [12].
Plasma HIV1 RNA, expressed as log10 copies/ml, was measured during the study for each patient by the same test. According to the technique used in each laboratory, the level of detection was 3.6 log10 copies/ml, or 2.6 copies/ml with NASBA, 2.7 log10 copies/ml with bDNA test, and 2.3 log10 copies/ml with Amplicor-Roche.
Patients were classified into two groups according to the effect of treatment at the time of onset of the AIDS-defining event. Group 1 included patients with an excellent response to highly active antiretroviral treatment defined by an at least 1.5 log10 reduction of viral load and 100% increase of CD4 cell count compared to the baseline value. Group 2 included all other patients not satisfying the criteria defined for group 1. Patients with a reduction of viral load, but no rise in CD4 cell count, or who obtained a rise in CD4 cells with no significant variation of viral load were classified in group 2.
Because most events occurred early, within 2 months of HAART initiation, and immune functions could not be restored at that time despite CD4 cell increase, AIDS-defining events were reported as early events if they occurred during the first 2 months and late events if they occurred during follow-up after the second month of HAART.
Patients who died during the first 2 weeks after starting treatment or who had received less than 15 days of protease inhibitor during the study period, were not included in the analysis. Incidence rates were calculated by expressing the number of new and recurrent events occurring during the study as the numerator and the overall number of patients exposed to the same risk as the denominator. Categorical analyses were performed using the appropriate χ2 test. Continuous variables were analysed with the Wilcoxon rank sum test.
Results
Opportunistic events
Four hundred and eighty-six patients were included in the study, of whom 334 received indinavir, some of them after intolerance to ritonavir, and 152 received and continued to receive ritonavir. The protease inhibitor was combined with previous treatment or, whenever possible, with one or two new nucleoside analogues. According to criteria defined by the French Medicinal Products Agency, protease inhibitor treatment was prescribed to patients with a CD4 cell count less than 100 × 106 cells/l, up until June 1996, and less than 200 × 106 cells/l, thereafter. A large number of patients were severely immunodepressed: 215 (44.2%) had a CD4 cell count less than 50 × 106 cells/l at the time of introduction of the protease inhibitor. The mean follow-up (± SD) was 184 ± 66 days (range, 36-314).
Fifty clinical events were observed in 46 patients, representing an incidence of 20.4 cases per 100 person-years. The 50 clinical events, summarized in Table 1, consisted of 48 AIDS-defining events, one episode of Varicella-Zoster virus (VZV) acute retinal necrosis, and one case of visceral leishmaniasis, both reflecting severe immunodepression. Eighty-four per cent of patients were males. The mean age (± SD) on inclusion was 37 ± 8 years, and the clinical stage was C3 for 89% of patients and stage B3 for 11% of patients. The mean duration of previous antiretroviral treatment was 34 ± 24 months (range, 0-84). Two patients had never previously received antiretroviral treatment. The mean baseline CD4 cell count (± SD) was 22 ± 28 × 106 cells/l (range, 0-130). Thirty-nine patients (78%) had a CD4 cell count less than 50 × 106 cells/l. The mean plasma HIV1 RNA (± SD) at baseline was 5.41 ± 0.54 log10 copies/ml (range, 4.4-7). In 215 patients who had, at baseline, a CD4 cell count less than 50 × 106 cells/l, the incidence of events was 36.1 cases per 100 person-years whereas in the 271 patients with a CD4 cell count greater than 50 × 106 cells/l it was eight cases per 100 person-years.
The events were a primary episode in 31 cases, which was an incidence of 12.7 cases per 100 person-year and corresponded to a recurrence in 19 cases. Ten of the 19 recurrences occurred despite maintenance therapy, and nine occurred after withdrawal it. Thirteen patients received ritonavir, 33 received indinavir, always associated with two nucleoside analogues, mainly zidovudine + lamivudine or stavudine + lamivudine, except in two patients. The nucleoside analogues were modified at the time of introduction of the protease inhibitor in 25 of the 46 patients. The mean time (± SD) to onset of the 50 events, following introduction of the protease inhibitor, was 62 ± 56 days (range, 7-250). Thirty-four events (68%) occurred during the first 2 months of treatment, which corresponded to an incidence of 42 cases per 100 person-years, including 24 primary episodes (with an incidence of 29.6 cases per 100 person-years). The number of events decreased over the following months.
Eighteen of the 34 clinical events observed during the first 2 months of treatment occurred despite a marked effect of treatment on laboratory parameters (patients of group 1): 11 CMV infections, five mycobacterial infections including three M. avium infections, one Cryptococcus neoformans infection, one case of VZV acute retinitis. These infections corresponded to recurrences in seven patients. All seven patients were receiving maintenance therapy (Table 1). Fifteen events observed during the first 2 months occurred in the absence of any effect of treatment (patients of group 2) and one event in a patient who could not be classified.
Sixteen infections were observed later, more than 2 months after introduction of the protease inhibitors. Six out of the 16 events occurred despite a marked effect of treatment on laboratory parameters (patients of group 1): one CMV infection, three mycobacterial infections (including one M. avium and one Mycobacterium szulgai infections), one episode of oesophageal candidiasis and one cryptococcal infection. These infections corresponded to a recurrence in five of the six patients. Each of the five patients had stopped their maintenance therapy several weeks before relapse. Nine out of the 16 events occurred after 2 months in group 2 patients, and one event occurred in one patient who could not be classified.
At the time of onset of the clinical event, the mean CD4 cell count (± SD) was 74 ± 71 × 106 cells/l (range, 3-260), and the mean gain compared with the count at inclusion was 51 ± 56 × 106 cells/l (range, -10 to +210) (P < 0.0001). In 13 patients, the CD4 cell count was greater than 100 × 106 cells/l. The mean viral load (± SD) was 4.2 ± 1.3 copies/ml (range, < 2.6 to 7), and the mean reduction was 1.6 ± 1.3 (range, from 4.4 log10 copies/ml decrease to 0.7 log10 copies/ml increase) (P < 0.0001). Twenty-four events occurred in patients of group 1, according to the criteria of progression of the CD4 cell count and viral load defined above. Their mean CD4 cell count (± SD) was 120 ± 73 × 106 cells/l, the mean increase (± SD) was 95 ± 58 × 106 cells/l, the mean viral load (± SD) was 3.1 ± 0.5 log10 copies/ml and the mean reduction (± SD) was -2.2 ± 0.8 log10 copies/ml. Twenty-four events occurred in patients of group 2, with a mean CD4 cell count (± SD) of 31 ± 30 × 106 cells/l, a mean increase of 15 ± 21, a mean plasma viral load (± SD) of 5.5 ± 0.4 log10 copies/ml, and a mean reduction of 0.4 ± 1.0 log10 copies/ml. Two patients could not be classified due to the absence of sufficient laboratory data. The CD4 cell count (25 ± 25 × 106 cells/l in group 1 and 16 ± 25 × 106 cells/l in group 2) and viral load (5.34 ± 0.67 in group 1, and 5.50 ± 0.42 log10 copies/ml in group 2) were not different between the two groups at baseline (P > 0.05). At the onset of the AIDS-defining events, there was a strong difference between the two groups with regard to the CD4 cell count (P < 0.0001) and viral load (P < 0.0001).
Cytomegalovirus infection
Eighteen CMV infections were observed in 18 patients, which corresponded to a mean incidence of 7.3 cases per 100 person-years. The overall incidence of CMV infections in patients with a baseline CD4 cell count less than 50 × 106 cells/l was 14.8 cases per 100 person-year. The mean time to onset after introduction of indinavir (10 patients), or ritonavir (eight patients) was 61 ± 69 days. A primary visceral CMV infection was observed in nine patients, and a relapse was observed in another nine patients, but in a different site for two of them. The lesion was retinal in 15 cases, always unilateral, neurological in two cases, intestinal in one case.
Fourteen of the 18 infections occurred less than 2 months after starting protease inhibitor treatment (for an incidence of 13.3 cases per 100 person-years): 11 retinal lesions, two neurological diseases, and one intestinal disease. This event constituted the first manifestation of CMV in seven cases (without primary prophylaxis) that occurred for six who were patients in group 1, and a relapse in seven cases (all with maintenance therapy) occurred for five that were patients of group 1 (Table 1). In every case, the initial ocular fundus examination was normal or cicatricial. Blood cultures for CMV were performed in 10 of the 14 patients at baseline: three were positive, seven were negative. The PCR for plasma CMV DNA was positive in one of the seven patients with negative cultures. At the start of treatment, the mean CD4 cell count (± SD) was 19 ± 31 ↔ 106 cells/l, and the mean viral load was 5.27 log10 copies/ml (4.4-6.4). At the time of diagnosis the 11 group 1 patients had a mean CD4 cell count of 111 ± 73 × 106 cells/l (11 ± 6%), a mean gain in CD4 of 82 ± 50 × 106 cells/l. Four patients had a CD4 cell count greater than 100 × 106 cells/l. Ten out of the 11 patients had a viral load below the limit of detection, which varied according to the technique used. In the three others cases, CMV infection occurred in two patients in group 2 (Table 2), and in one patient whose response to treatment could not be determined, due to the absence of follow-up viral load. The three earliest cases of retinal CMV infection were symptomatic: one on the tenth day of treatment in the form of acute retinal necrosis, the other two on the fourteenth and twenty-first days, corresponding to a relapse of previously cured CMV retinitis. They were associated with acute uveitis and sudden loss of vision in one eye and occurred despite a marked effect of treatment in terms of laboratory parameters, by the first month of treatment. The same applies to the case of VZV acute retinal necrosis, occurring on the seventh day of treatment.
The CMV diseases were resolved in every case in response to high-dose IV ganciclovir (12 patients) or foscavir (two patients), but symptomatic patients had complete visual loss or very marked reduction of visual acuity. One patient died 2 months later from undocumented interstitial pneumonia.
Late CMV retinitis (first episode = 2, relapse = 2) were observed in four patients with an incidence of 2.4 cases per 100 person-year. Only one case occurred in patients of group 1: the patient, treated for 9 months with IV ganciclovir for CMV encephalitis, stopped treatment following a recent undetectable viral load (< 2.3 copies/ml) and a CD4 cell count of 250 × 106 cells/l. At that time ocular fundus examination was normal and CMV cultures were negatives. A primary episode of unilateral retinitis was observed 21 days later on routine ocular fundus examination. The three other cases of retinitis were observed in group 2 patients. These patients did not receive any prophylaxis or had stopped maintenance therapy a few month before the occurrence of retinitis. All four patients were treated with intravenous ganciclovir and were cured without any ocular sequelae.
Mycobacterial infections
Thirteen mycobacterial infections were observed in 13 patients which corresponded to an incidence of 5.3 cases per 100 person-year. Eleven patients were treated with indinavir and two with ritonavir, all in combination with two nucleoside analogues. At least one nucleoside analogue was modified at the time of introduction of the protease inhibitor in seven patients. The infection appeared or was detected an average of 64 ± 50 days after starting protease inhibitor treatment, eight infections occurred early, five later.
The eight early infections were six first episodes (none of the four patients with M. avium complex infection had received prophylaxis), and two relapses despite maintenance therapy (Table 1). Five out of these eight infections, occurred in group 1 patients: M. avium infection in three cases, Mycobacterium tuberculosis infection in one case, Mycobacterium malmoense infection in one case. These five patients had a mean CD4 cell count of 103 ± 106 × 106 cells/l, with an increase of 84 ± 87 × 106 cells/l and a mean viral load of 3 ± 0.5 log10 copies/ml, with a mean reduction of 2.5 ± 1.8 log10 copies/ml. One M. avium complex infection was observed with CD4 cell count of 70 × 106 cells/l (9%) and four out of the five patients had a viral load below the limit of detection. In three patients, the infection occurred in patients of group 2: two M. avium complex infection and one M. Szulgai and Mycobacterium Xenopi infection (first episode) (Table 2).
The five late infections were observed 110 ± 28 days after starting protease inhibitor: two primary episodes of tuberculosis infection, and three relapses (only one patient was on maintenance therapy). Three out of the five patients were in group 1: relapse of M. avium infection, relapse of M. Szulgai bone infection and a first case of M. tuberculosis infection each in one patient. Mycobacterium avium complex and M. Szulgai infections occurred several weeks after the patients had stopped maintenance therapy, 105 and 138 days after starting treatment, with CD4 counts of 125 × 106 cells/l (7%) and 70 × 106 cells/l (9%), and a gain of 121 and 64 × 106 cells/l, respectively. The two other infections which were observed after 2 months of treatment, one relapse of M. avium complex infection and one episode of pulmonary tuberculosis occurred in patients of group 2 and their biological data are shown Table 2.
All patients were cured without sequel by specific treatment, except for one patient with tuberculosis who died 1 month later from undocumented meningoencephalitis.
Oesophageal candidiasis
Five cases of oesophageal candidiasis were observed: two primary sites and three recurrences without maintenance therapy. They occurred 74 ± 41 days after introduction of indinavir (three patients) or ritonavir (two patients). Two episodes of oesophageal candidiasis occurred during the first 2 months of treatment (one patient of group 1, and one patient of group 2).
Three episodes of oesophageal candidiasis occurred later: in one patient with a CD4 cell count of 97 × 106 cells/l (14%), a gain of 61 × 106 cells/l, and a viral load of 3.3 log10 copies/ml (group 1), but this patient was simultaneously taking antibiotic treatment for M. avium complex and in two patients of group 2 (Table 2).
Cryptococcus neoformans infections
Three cryptococcal infections were observed, namely, the recurrence of meningitis in two patients and a primary episode of pulmonary and cutaneous cryptococcosis in a third patient. One patient was treated with ritonavir and the other two were treated with indinavir. In one case, the infection occurred during the first 2 months of treatment, despite a marked effect of treatment on laboratory parameters (patient of group 1). In two cases, the infection occurred later, 120 and 150 days after starting treatment, corresponding to a recurrence in both cases, occurring during fluconazole prophylaxis in one case (this patient could not be classified due to missing data) and in the other case after fluconazole had been stopped for several weeks, when the CD4 cell count was 180 × 106 cells/l, with a gain of 160 × 106 cells/l, and a viral load of 3.2 log10 copies/ml with a reduction of 1.9 log10 copies/ml. A favourable course in response to treatment was observed in all cases.
Pneumocystis pneumonia
Three cases of Pneumocystis carinii infections were observed as a first episode in three cases. Two episodes occurred during the first 2 months of treatment with indinavir and the third case occurred 240 days after starting ritonavir. The three patients were in group 2; two of the three patients had been receiving trimetho-prim-sulfamethoxazole prophylaxis.
Cerebral toxoplasmosis
Two Toxoplasma gondii brain infections occurred during the first 2 months of treatment, in two patients who were in group 2. Trimethoprim-sulfamethoxazole prophylaxis was ongoing in one case. One patient was cured without neurological sequelfl by pyrimethamine and sulfadiazine, and the other died 2 months later.
Other events
The other clinical events observed were: one case each of cryptosporidiosis, progressive multifocal leukoencephalopathy, visceral leishmaniasis, Kaposi sarcoma, cerebral lymphoma, encephalitis and VZV acute retinitis. Five of the six events occurred during the first 2 months of treatment, one of them despite effective HAART regimen treatment (group 1: VZV infection), and the four others events in patients whose CD4 cell count did not change. Progressive multifocal leukoencephalopathy occurred later, in a context of persistently low CD4 cell count (13 × 106 cells/l), and high viral load (5.6 log10 copies/ml)
Discussion
This study emphasises the development of AIDS-defining events despite highly active antiretroviral treatment including protease inhibitor, especially during the first 2 months of treatment. The incidence of primary visceral infections was 6.1% during the 6.1 months of follow-up, which is similar to the incidence of 6% of AIDS-defining events observed with the zidovudine, lamivudine, indinavir combination [2], despite the absence of a change of nucleoside analogue when protease inhibitors had started for some patients.
The profile of infections during HAART was modified: 24 (48%) of the 50 clinical events occurred despite an increase in the absolute CD4 cell count of more than 100% and a reduction of the HIV-1 viral load by more than 1.5 log10 copies/ml and 19 patients had an undetectable plasma viral load at the time of the clinical event. CMV infection with a CD4 count over 100 × 106 cells/l used to be rare before the age of protease inhibitors [13,14] and CMV retinitis generally occurred in patients with CD4 cell counts less than 50 × 106 cells/l [15,16]. With HAART, the development of CMV retinitis in patients with CD4 cell counts higher than 150 × 106 cells/l is now encountered [9,10] and must be taken into account when advising prophylaxis and surveillance of patients.
Eighteen of the 24 events in group 1 occurred during the first 2 months of treatment which is in favour for a delayed restoration of immune function and raising the problem of the functional value of the CD4 and CD8 cells rapidly acquired after starting triple combination therapy including protease inhibitor. The incidence of CMV infections during the first 2 months of follow-up, equal to 17.3 cases per 100 person-year, was at least identical or higher than the incidence observed in the studies conducted in patients who did not receive protease inhibitors, reported by Pertel [16] and by Gallant [15]. The development of a CD8+ lymphocyte specific cytotoxic response is essential to control cytomegalovirus in mice [17] and the recovery of a CMV-specific cytotoxic T-lymphocyte reaction of CD8 lymphocytes ensures the prevention of CMV disease in transplant recipients [18]. Loss of the cytotoxic response of CD4 lymphocytes and interleukin 2 production in response to memory antigens can occur early in the course of HIV infection. Immune reconstitution in response to HAART is partial and deferred, which may partly explain the development of these early infections. The increase of cytotoxic T-lymphocyte proliferation reactions of CD4+ cells in the presence of CMV or tuberculin antigen is obtained in vitro after 1 to 6 months of HAART [2]. These in vitro immunological data agree with our clinical observations of early CMV and M. avium complex infections, despite marked efficacy of antiretroviral treatment and rapid reduction of their incidence after the second month of treatment. This study does not exclude the hypothesis raised by other authors [9] that the introduction of HAART could promote, over a brief period, the development of certain infections, such as cytomegalovirus infections, in which the infectious agent, already present, may replicate asymptomatically during introduction of treatment. The explosive clinical features of the three cases of early CMV infections and the VZV infection are in favour of this hypothesis. The possibility of a specific effect of the protease inhibitor on the development of these early opportunistic infections as well as drug interactions must be studied [19]. As some patients with a CD4 cell count less than 100 × 106 cells/l remain at risk of developing a first episode or relapse of cytomegalovirus infection during the first 2 months of HAART, it appears important to assess the value of pre-emptive treatment in the case of signs of active CMV replication when starting such antiretroviral treatment. The predictive value of CMV blood and urine cultures, serum antigen pp65, plasma and leukocyte CMV DNA assays have been demonstrated [20,21,22]. The modalities, beneficial effect, duration, and acceptability of this treatment in this situation need to be determined.
Six clinical events occurred later despite an effect of treatment on laboratory parameters (patients in group 1). Five of the six opportunistic infections were relapses, occurring several weeks after stopping maintenance therapy, although the infection had been previously well controlled and considered to be cured by the clinician. They were one case of CMV retinitis (CD4 = 133 × 106 cells/l, 19%), one episode of febrile M. avium complex bacteremia (CD4 = 70/ × 106 cells/l, 9%), a M. Szulgai bone infection (CD4 = 125 × 106 cells/l, 7%), oesophageal candidiasis (CD4 = 97 × 106 cells/l, 14%), and neuromeningeal cryptococcosis (180 × 106 cells/l, 7%) which were observed an average of 146 days after starting protease inhibitor. These late events emphasise that the course of immune restoration can be heterogeneous and incomplete, and that the absolute number of CD4 cells, even associated with an undetectable viral load, is not always sufficient to allow safe discontinuation of long-term maintenance therapy.
We conclude that in patients at high risk of developing opportunistic infections prior to the institution of HAART, prophylaxis for cytomegalovirus diseases and M. avium infections should not be discontinued during the first 2 months of treatment unless significant increase in the CD4 cell count. This attitude can be extended to P. carinii and T. gondii prophylaxis. In the same way, CMV disease and M. avium complex infection prophylaxis must be considered before the introduction of protease inhibitors in high risk patients. Following 2 months of successful HAART therapy, primary prophylactic regimen may not be necessary, but maintenance therapy for previous diagnosed opportunistic infections, especially for CMV, M. avium complex and invasive fungal infections should be carried on up to the right time of sufficient immune restoration, which have to be determined by further studies.
Acknowledgements
We thank all the members of GERICCO (Groupe d'Epidémiologie et de Recherche en Infectiologie du Centre-Ouest) who participated to the collection of the data, especially P. Choutet, B. Becq-Giraudon and P. Weinbreck.
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. Autran B, Garcelain G, Li TS, et al.: Positive effects of combined antiretroviral therapy on CD4+ T cell homeostasis and function in advanced HIV disease. Science 1997, 277: 112-116.
3. Hammer SM, Squires KE, Hugues MD, et al.: A controlled trial of two nucleoside analogues plus indinavir in persons with human immunodeficiency virus infection and CD4 cell counts of 200 per cubic millimeter or less. N Engl J Med 1997, 337:725-733.
4. Centers for Diseases Control and prevention: Update trends in AIDS incidence 1996. MMWR 1997, 46:861-867.
5. Mouton Y, Alfanderi S, Valette M, et al.: Impact of protease inhibitors on AIDS-defining events and hospitalizations in 10 French AIDS reference centres. AIDS 1997, 11:F101-F105.
6. Torres RA, Barr M: Impact of combination therapy for HIV infection on inpatient census. N Engl J Med 1997, 336:1531-1532.
7. Cameron B, Heath-Chiozzi M, Danner S et al.: Prolongation of life and prevention of AIDS in advanced HIV immunodeficiency with Ritonavir. Lancet 1998, 351:543-549.
8. USPHS/IDSA Prevention of opportunistic infections Working Group: Guidelines for the prevention of opportunistic infections in persons infected with human immunodeficiency virus. MMWR 1997, 46: RR-12.
9. Jacobsen MA, Zegans M, Pavan PR, et al.: Cytomegalovirus retinitis after initiation of highly active antiretroviral therapy. Lancet 1997, 349:1443-1445.
10. Gilquin J, Piketty C, Thomas V, et al.: Acute CMV Infection in AIDS patients with CD4 count above 100 x 106 cells following receiving combination therapy including protease inhibitors. AIDS 1997, 11:1659-1660.
11. Centers for Disease Control: Revision of the CDC surveillance case definition for acquired immunodeficience syndrome. MMWR 1987, 36 (supp 1S):3S-15S.
12. Moore R, Chaison R: Natural history of opportunistic disease in HIV infected urban clinical cohort. Ann Intern Med 1996, 124:633-642.
13. Kuppermann BD, Petty JG, Richman DD, et al.: Correlation between CD4+ counts and prevalence of cytomegalovirus retinitis and human deficiency virus related non infectious retinal vasculopathy in patients with acquired immunodeficiency syndrome. Am J Ophthal 1993, 115:575-582.
14. Fekrat S, Dunn JP, Lee D, Miller T, Jabs DA: Cytomegalovirus retinitis in HIV-infected patients with elevated CD4+ counts. Arch. Ophthalmol 1995, 113:18.
15. Gallant JE, Moore RD, Richman DD, Keruly J, Chaisson RR, and the Zidovudine Epidemiology Study Group: Incidence and natural history of cytomegalovirus disease in patients with advanced human immunodeficiency virus disease treated with zidovudine. J Infect Dis 1992.166:1223-1227.
16. Pertel P, Hirchtick R, Phair J, Chmiel J, Poggensee L, Murphy R: Risk of developing cytomegalovirus retinitis in persons infected with the human immunodeficiency virus. J Acquir Immune Defic Syndr 1992, 5:1069-1074.
17. Reddehase M, Mütter W, Munch K, Buhring HJ, Koszinowski U: CD8-positive T lymphocytes specific for murine cytomegalovirus immediate early antigens mediate protective immunity. J Virol 1987, 61:3102-3108.
18. Reusser P, Ridell S, Meyers J, Greenberg P: Cytotoxic T lymphocyte response to cytomegalovirus following allogeneic bone marrow transplantation: pattern of recovery and correlation with cytomegalovirus infection and disease. Blood 1991, 78:1373-1380.
19. Aussel C, Breittmayer JP, Ticchioni M, Pelassy C, Bernard A: Regulation of T cell activation by cytochrome P450 inhibitors. Cell Immunol 1994, 155:436-445.
20. Gerard L, Leport C, Flandre P, et al.: Cytomegalovirus viremia and the CD4+ lymphocytes count as predictors of CMV disease in patients infected with human immunodeficiency virus. Clin Infect Dis 1997, 24:836-840.
21. Shinkai M, Bozette SA, Powderly W, Frame P, Spector SA: Utility of urine and leukocyte cultures and plasma DNA polymerase chain reaction for identification of AIDS patients at risk for developing human cytomegalovirus disease. J Infect Dis 1997, 175:302-308.
22. Bowen EF, Sabin CA, Wilson P, et al.: Cytomegalovirus viraemia detected by polymerase chain reaction identifies a group of HIV positive patients at high risk of CMV disease. AIDS 1997, 11:889-893.
© 1998 Lippincott Williams & Wilkins, Inc.