Highly active combination antiretroviral therapy with protease inhibitors, nucleoside-analogue and non-nucleoside reverse transcriptase (RT) inhibitors has become the standard of care of HIV infection in the U.S.[1,2]. Prior to the introduction of combination antiretroviral therapy, the natural history of the development of opportunistic complications of HIV infection was relatively well-documented[3-7]. Advancing immunosuppression, specifically the CD4 level, could be used as a marker of the risk of developing individual opportunistic illness and death. Since the introduction of highly active combination antiretroviral therapy, the incidence of opportunistic infection and death has declined[8-12].
A landmark study published in 1996 revealed the importance of the serum HIV level as a predictor of developing an AIDS-defining illness and survival, independent of the CD4 level. Clinical trials have demonstrated that PI-containing combination therapy suppresses plasma HIV RNA levels more completely and durably than dual nucleoside regimens and that inhibition of HIV replication is associated with more sustained increases in CD4 cell levels and reductions in opportunistic diseases[14-15]. The relative contribution of HIV viral load suppression and CD4 cell increases on disease outcomes in patients receiving combination therapy, however, has not been quantified. Whether CD4 level maintains the same association with development of opportunistic infection and death as in earlier years is not well-delineated, and several questions regarding this relationship lack clear answers. For example, is the nadir CD4 a better predictor of opportunistic infection and death than a subsequent higher level resulting from therapy? Is the serum HIV-1 level now more important than the CD4 level as a predictor of opportunistic infection and death, or can a relationship be defined using both parameters that is a better predictor of opportunistic infection and death than either parameter alone?
To address these issues, we assessed the development of opportunistic infection and death in our cohort of HIV-infected patients followed longitudinally on the Johns Hopkins AIDS Service in Baltimore, Maryland. Our goal was to determine first how the patterns of opportunistic infection and death had changed with the use of combination antiretroviral therapy, looking both at the incidence and the relative risk among various infections over time. We then focused on those patients who developed an opportunistic infection or died to better delineate the relationship between CD4, HIV-1 RNA, and development of these outcomes.
The Johns Hopkins AIDS Service provides care for a large proportion of HIV-infected patients in the Baltimore metropolitan area. Longitudinal primary and subspecialty care are integrated in one hospital-based HIV clinic. An observational clinical database has been a part of the AIDS Service since 1990. Information from the clinical records is reviewed and abstracted by trained medical record technicians onto structured data collection forms, and then entered into an automated database. The Clinic medical record, the main hospital medical record and various institutional automated databases are abstracted. Medical records from other institutions in which the patients received care are also routinely abstracted. Comprehensive demographic, clinical, laboratory, pharmaceutical, and psychosocial data are collected at times corresponding to enrollment into the HIV Clinic and at 6-month intervals thereafter. Information on death is obtained from a death registry maintained by the Clinic that receives reports from families, funeral homes, other medical institutions, and local coroners. In addition, death records of the Maryland Bureau of Vital records and the national Social Security death index are regularly searched.
Definitions and criteria used
Diagnoses of opportunistic Illness
We focused on several major opportunistic diseases that are included in the Category C AIDS surveillance criteria of the Centers for Disease Control and Prevention  and are associated historically with advanced immunosuppression and elevated mortality rates.
CD4 cell levels were measured by flow cytometry (Becton-Dickinson, Baltimore, MD, USA). HIV-1 RNA levels were measured using the Roche Amplicor assay (Roche Molecular Systems, Branchburg, New Jersey, USA) which has a lower detectability threshold of approximately 400 copies/ml. The frequency of monitoring was determined by individual clinicians _caring for the patients, and was measured every _3 months, on average.
Antiretroviral regimens were categorized as monotherapy with a nucleoside-analog reverse transcriptase inhibitor (RTI), two RTIs, three or more drugs that included a protease inhibitor (PI). These combinations could be any of the following: one PI and two RTIs, or two PIs and one or two RTIs, one or two PIs, and one or two RTIs. We did not include patients whose only PI was saquinavir [hard gel capsule (Hgc)] because of low bioavailability which excludes it as a first-line recommended therapy. Pneumocystis carinii pneumonia (PCP) prophylaxis included either trimethoprim-sulfamethoxazole or dapsone. Mycobacterium avium complex (MAC) prophylaxis included either rifabutin, clarithromycin or azithromycin. All therapies were counted once the patient had received at least 30 days of that therapy.
Incidence of opportunistic illness and death from 1994-1998
The sample for this analysis included patients _who contributed follow-up time to our cohort after _January 1, 1994. Our analysis focused on the time after a patient was first reported to have a CD4 ≤_500 cells__∞_106/l, even if the CD4 subsequently increased to _> 500 cells_∞_106/l. Five complete years were analyzed, beginning January 1, 1994 and extending to December 31, 1998. For this first set of analyses, we determined the incidence of developing an opportunistic illness or of dying during each year from 1994 through 1998. This analysis was done to determine whether combination antiretroviral therapy had been associated with a decrease in development of opportunistic and death in our cohort as it had been in other cohorts. To calculate annual incidence rates, we computed the amount of time each patient contributed to that year. Within this patient denominator, we computed the number of events of interest (e.g., death or opportunistic illness) as the numerator for the determination of incidence. If more than one event occurred in a year, each was counted. Events were reported per 100 person-years with Poisson 95% confidence intervals. We then determined the incidence of each individual opportunistic illness in 1994 and compared it with the incidence in 1996 and 1998. We compared the rank order in incidence of each opportunistic illness in 1994 with 1998. For this calculation, only the person-time until the individual event was included.
Multivariate Poisson regression was used to analyze the association of demographic and clinical patient characteristics and the relative rate of developing an opportunistic illness or dying. For this analysis, all events were taken into account. The Poisson regression analysis was conducted using PROC GENMOD (SAS, Inc., Cary, NC, USA) using a general estimating equations methodology with an exchangeable correlation structure to adjust the variance for the contribution of multiple intervals by each individual. We first divided the sample time into 1-month intervals. Each interval was categorized by CD4 count (stratified by whether a CD4 of <_50, 51-100, 101-200, or >_200 cells_∞_106/l had previously been achieved), prior AIDS-defining opportunistic illness, use of PCP and MAC prophylaxis, age (stratified as >_40 or ≤?40 years), sex, (male, female), race (non-Hispanic white, African-American), and HIV transmission risk group [injecting drug use (IDU), men who have sex with men (MSM), heterosexual contact], and calendar year. Two analyses were done. The first of these did not include the antiretroviral regimen, and the second included the antiretroviral regimen (zero, one, two, three or more drugs) as a covariate. The interval was classified by the regimen if it had preceded that interval. However, a regimen of greater potency would supercede a regimen of lesser potency. For example, a patient receiving a two-drug regimen would have all subsequent intervals classified two drugs until a three-or-more drugs regimen was used. At that point, the patient would reclassify as three or more drugs. Drug classification was continued even if the patient discontinued therapy as a method to approximate an intention-to-treat analysis. The results are reported as relative rates with 95% confidence intervals.
Opportunistic illness and death in patients receiving a three-or-more drug PI-containing therapy
For the analyses in this section, we focused on those patients who received three-or-more antiretroviral drug combination regimens and who had baseline CD4 and HIV-1 RNA levels. We first calculated the Kaplan-Meier estimated probability over time of achieving an undetectable HIV-1 RNA (e.g. ≤_400 copies/ml) after beginning of therapy. In those patients who achieved an undetectable HIV-1 RNA, we _calculated the Kaplan-Meier estimated probability of subsequently having an HIV-1 RNA ≥ 1000 copies/ml on at least two measurements.
We computed Kaplan-Meier estimates of the probability over time of developing an opportunistic illness or dying. These estimates were stratified by whether or not the patient achieved an undetectable HIV-1 RNA, and, in those who achieved an undetectable HIV-1 RNA, by whether or not it subsequently increased to _≥ 1000 copies/ml. A second Kaplan-Meier analysis was carried out stratifying the estimates by CD4 level and whether or not an undetectable HIV-1 RNA was achieved. For both of these analyses, the log-rank test was used to statistically compare strata.
Our final step was an Anderson-Gill multiplicative _hazards regression of the development of opportunistic illness or death as a function of baseline and current CD4 levels, nadir HIV-1 RNA, as well as demographic (age, sex, race, HIV transmission risk group) and clinical (prior AIDS-defining opportunistic illness, PCP and MAC prophylaxis, first protease inhibitor used) variables. This method allows for the analysis of multiple events per patient. This analysis was carried out using PROC PHREG (SAS, Inc). CD4 and HIV-1 RNA levels were analyzed as time-dependent covariates in the model. Unlike a fixed covariate which maintains a constant value, a time-dependent covariate can change in value over time with each risk set (e.g. event). Fair comparisons are made at each risk set (e.g. time of event) of the values of variables at that point in follow-up time for each individual. This allows for a better assessment of the effect of a measure that can change over time such as the CD4 and HIV-1 RNA. It also minimizes length of follow-up bias where a person who is followed for a longer period of time has a greater opportunity to have a change in the value of a variable. The results of this analysis are reported as estimates of the relative hazard of an event occurring with associated 95% confidence limits.
Incidence of opportunistic infection and death from 1994-1998
A total of 3211 patients who had a CD4 level ≤ 500 cells_∞_106/l were studied. Of these patients, 2281 (71%) were men, 2481 (77%) were African-American, and 688 (21%) were non-Hispanic white. Risk factors for HIV transmission included 1614 (50%) IDU, 816 (25%) MSM, and 781 (25%) heterosexual contact with an infected or high-risk partner or an unknown risk factor. Median age was 36 years (range: 17-76 years). There were no significant differences annually from 1994 to 1998 in the distribution of any of these demographic characteristics. The median CD4 level at time of first measured CD4 level below 500 cells_∞_106/l showed a non-significant increase from a median of 252 cells_∞_106/l in 1994 to 271 cells_∞_106/l in 1998. PCP prophylaxis was used in 64% of patients and MAC prophylaxis in 35% of patients.
The incidence of opportunistic illness and death by year is shown in Table 1. The incidence of all AIDS-defining opportunistic illnesses combined declined from 23.7 events/100 person-years in 1994 to 14.0 events/100 person-years in 1998 (P_<_0.001). Similarly, mortality decreased from 20.2 death/100 person-years in 1994 to 8.4 deaths/100 person-years in 1998 _(P_<_0.001). There was an increase in use of three or more drugs over time, a decrease in one-drug therapy, and an increase in two-drug regimens that peaked in 1996, but declined thereafter. The annual incidence of each opportunistic illness in 1994, 1996, and 1998 is shown in Table 2. Most of the illnesses declined in incidence from 1994 to 1998, with the possible exception of lymphoma, wasting syndrome, and cervical cancer.
Our assessment of the association between antiretroviral therapy and annual incidence is shown in Table 3. In the first analysis, there was a decrease in the relative rate of opportunistic illness or death in 1995 [relative risk (RR) = 0.88], 1996 (RR = 0.72), 1997 (RR = 0.58), and 1998 (RR = 0.64) compared with 1994. Lower CD4 level was also associated with a higher risk of an event, however, age, sex, race, HIV transmission risk group, and prophylaxis were not associated. In the second model, antiretroviral drug regimen was added to the analysis. Treatment with combination antiretroviral therapy is associated with a decreased relative rate of opportunistic infection or death with a decreased relative risk of a magnitude similar to that seen in model 1 for calendar years 1997 and 1998. Adjusting for treatment, year was no longer associated with risk of progression, indicating that treatment rather than some other temporal trend is associated with the decline in incidence.
Opportunistic illness and death in patients receiving three-or-more drug PI-containing therapy
We focused our remaining analyses on the development of these outcomes in patients on three-or-more drug antiretroviral therapy (n_=_748). Patients who received this therapy were similar demographically to the overall population, however CD4 level was lower than for the overall population [median = 212 cells__∞_106/l (Q1_=_71, Q3_=_359)].
The median duration of time on three-or-more drug therapy was 332 days (range: 30-1072 days; total of 759 person-years on therapy). Dual PI therapy was used initially in 127 (17%) patients who started three-or-more drug therapy. This was ritonavir-saquinavir in 117 patients and other combinations in 10 patients. A single PI [indinavir_=_286, nelfinavir_=_232, ritonavir_=_84, saquinavir (soft gel capsule; sgc) = 19] was used in the remaining 621 patients. A total of 229 of 748 (31%) patients changed from their initial PI regimen to a regimen with an alternative PI at a later date.
Kaplan-Meier estimates of achieving an undetectable HIV-1 RNA showed that by 6 months after starting therapy, 50% (95% CI; 46-54%) of patients had achieved an undetectable HIV-1 RNA. This increased to 65% (95% CI: 61-69%) by 12 months and was relatively stable at 67% (95% CI: 62-72%) by 18 months. Kaplan-Meier estimates of the HIV-1 RNA subsequently increasing to >_1000 copies/ml in those who achieved an undetectable HIV-1 RNA showed that 32% (95% CI: 27-37%) of patients who achieved an undetectable HIV-1 RNA had >_1000 copies/ml after 1 year increased to 60% (95% CI: 52-68%) by _18 months.
The association of CD4 level, HIV-1 RNA and time to event is shown in Table 4 for each opportunistic _illness and death in patients receiving three or more drugs. The median time to the event after start of _therapy ranged from 72 days for toxoplasmosis to 260 days for PCP. Only 29% of all events (23 opportunistic illnesses and five deaths) occurred within the first 90 days of starting therapy. Baseline and concurrent CD4 levels tended to be low in patients who had events, although CD4 levels were highest in patients who developed lymphoma, Kaposi‚s sarcoma, and bacterial pneumonia. The median change in the CD4 after start of therapy was no greater than +53 cells for any of the events. Although the median decline in HIV-1 RNA level was 1 log or more for several of the outcomes, the concurrent HIV-1 RNA was elevated above undetectable in almost all of the patients who developed an event.
Of the patients on three-or-more drug therapy, 532 patients received PCP prophylaxis and 290 patients received MAC prophylaxis. A total of 64 (12%) patients of the PCP prophylaxis-receiving (85 person-years) had their primary PCP prophylaxis stopped when the CD4 level had increased to > 200 cells__∞_106/l, and 63 (22%) patients (81 person-years) had their primary MAC prophylaxis stopped when the CD4 level had increased to > 75 cells_∞_106/l. No events (e.g. PCP, MAC, or toxoplasmosis) have occurred in any of these patients. No patient in our cohort received primary cytomegalovirus prophylaxis.
The Kaplan-Meier plot of an opportunistic illness or death stratified by achieving an undetectable HIV-1 RNA is shown in Fig. 1a. By 1 year after the start of therapy, the probability of an event was 18% in patients who did not achieve an undetectable HIV-1 RNA versus 9% in patients who did (P_<_0.001, log rank test). There is no significant difference between achieving and maintaining an undetectable HIV-1 RNA and achieving and not maintaining an undetectable HIV-1 RNA (P_= 0.72, log rank test). Figure 1b shows the Kaplan-Meier plot of an event stratified both by achieving an undetectable HIV-1 RNA and by whether or not the CD4 count fell below 100 cells__∞_106/l. Undetectable HIV-1 RNA was not further stratified by subsequent increase to >1000 copies/ml because there was no significant difference between the strata. The largest differences are between those who did or did not have a CD4 level < 100 cells_∞_106/l (log rank P_<_0.0001). Differences between achieving an undetectable HIV-1 RNA or not were not as great as for the CD4 stratification (log rank P_<_0.01).
This association between the HIV-1 RNA response and CD4 response is further illustrated in Fig. 2 which plots the median change from baseline in CD4 level and associated HIV-1 RNA levels over time. Patients who did not achieve an undetectable HIV-1 RNA had no more than a 30 cells_∞_106/l median change in CD4 level at 6 months, declining to 20 cells_∞_106/l by 12 months. Patients who achieved an undetectable HIV-1 RNA had a median increase of 60 cells_∞_106/l by 6 months of therapy and 100 cells_∞_106/l by 1 year of therapy. The increase in CD4 is similar for patients who did or did not maintain an undetectable HIV-1 RNA.
Finally, we analyzed by multiplicative proportional hazards regression the associations of CD4 and HIV-1 RNA levels as time-dependent variables with the development of opportunistic infection or death in patients receiving three-or-more drug antiretroviral therapy (Table 5). In the first analysis excluding _concurrent CD4 level, the nadir HIV-1 RNA was associated with development of an event. Patients who had undetectable HIV-1 RNA had a significantly lower relative hazard (RH)_=_0.33 of an event. A baseline CD4 level of < 50 cells_∞_106/l was associated with a significantly increased hazard (RH_=_2.69) of an event. However, when concurrent CD4 level was included in the second model, the HIV-1 RNA levels and baseline CD4 level were no longer significant. Age, race, sex, HIV risk group, and initially used protease inhibitor were not associated with outcome adjusting for the above variables.
Our analysis of an HIV-infected clinical cohort yielded several results regarding the development of opportunistic infection and death in the era of combination antiretroviral therapy. First, there has been a decline in the incidence of almost every opportunistic infection and in mortality from 1994 to 1998. This decline appears to be due to use of combination antiretroviral therapy, particularly with a three-or-more drug _regimen that contains a PI. However two-drug RT therapy has contributed to the decline. The decline in the annual incidence of opportunistic infection and death is similar to the decline found by others in the US, Canada, and Europe[8-10]. These cohorts were principally composed of white men. Our population was predominantly African-American men, but age, sex, race, and HIV risk group were not associated _significantly with the event rate suggesting that the effectiveness of combination antiretroviral therapy does not differ by these demographic patient characteristics.
Our analysis shows that along with the mortality rate, each of the opportunistic infections declined, although the relative relationship among them was similar to that found in the era prior to introduction of combination antiretroviral therapy. This suggests that there is probably no differential recovery of antigen-specific memory CD4 lymphocytes for these infections. The recovery of CD4 T lymphocytes with antiretroviral therapy has been an area of very active research in recent years. The initial increase in the CD4 level appears to be due to redistribution of memory CD4 T lymphocytes from lymph nodes. This is followed by a reduction in T-cells activation, coupled with improved T-cell _reactivity to recall antigens, with production of naive T cells more slowly over 4-6 months[21-25].
Neither lymphoma, wasting syndrome, or cervical cancer appeared to decline in incidence from 1994 to 1998. Concerns have been raised regarding a possible increase in the incidence of lymphoma in HIV-infected patients being treated over the long-term with combination antiretroviral therapy. Although our study does not show an increase, neither was there a decrease, and subtle changes could occur that are not detectable with our sample size. Wasting syndrome can be more difficult to accurately diagnose in patients on protease inhibitors since the lipodystrophy syndrome that occurs with these drugs could cause weight loss in some patients. We doubt that the incidence of other illnesses were affected by temporal or therapeutic changes in the application or interpretation of the criteria for diagnosis. Most of the same providers were caring for these patients in the same HIV practice from 1994 to 1998, and our methods for defining the diagnoses and collecting the data did not change over this time period.
Opportunistic infection and death continues to occur in patients who are receiving potent combination antiretroviral therapy, although at a lower rate than prior to the use of these drugs. Most of the opportunistic illnesses that occurred in patients on three-or-more drug antiretroviral therapy occurred when the concurrent CD4 level was low, most commonly with a median of <_50 cells_∞_106/l. In particular, no episodes of PCP occurred when the CD4 was > 200 cells_∞_106/l, and no episode of MAC occurred when the CD4 was greater than 75 cells_∞_106/l, consistent with these CD4 thresholds for using prophylaxis. Similarly, cytomegalovirus and toxoplasmosis occurred at very low CD4 levels. In contrast, patients developing wasting syndrome, recurrent bacterial pneumonia, non-Hodgkin lymphoma, Kaposi‚s sarcoma, AIDS dementia complex and candida esophagitis had a higher median concurrent CD4 levels and some had CD4 levels above 200 cells_∞_106/l.
The serum HIV-1 RNA level is used principally to assess the efficacy of antiretroviral therapy in clinical trials and in clinical practice. Reduction of the HIV-1 RNA to undetectable levels within the first 3 to _6 months of therapy is correlated with a sustained durable suppression of HIV-1 RNA, increase in CD4 level, and significant improvement in clinical progression of disease[14,15]. Our analyses indicate that suppression of HIV-1 RNA to undetectable levels is associated with an increase in the CD4 level and with a decrease in development of opportunistic illness or death. In our multivariate analysis, the significant association between an undetectable HIV-1 RNA and development of an event was no longer significant when concurrent CD4 was included. This is consistent with the association of HIV-1 RNA response and CD4 response but suggests that is the CD4 level that is most strongly associated with whether or not opportunistic infection or death will occur.
Unlike others, we did not find that most events occurred in the first 2 to 3 months after starting potent combination antiretroviral therapy. Instead, we found a distribution of events from as little as 30 days to as great as almost 2 years, with a median time of event of generally between 6 months and 1 year after starting therapy. This finding and our laboratory results suggest lack of durable suppression of HIV-1 RNA and poorer CD4 response over time. Most patients who developed an opportunistic infection or died failed to achieve an undetectable HIV-1 RNA. We were able to provide some data regarding the cessation of prophylactic drug therapy for PCP and MAC. Across our entire patient sample, PCP prophylaxis was discontinued in 64 patients and MAC prophylaxis was discontinued in 63 patients after the CD4 increased above prophylaxis threshold levels. None of these patients developed PCP or MAC. Other studies have also shown that PCP and MAC are unlikely to occur when prophylaxis is _discontinued in patients who have had elevation of their CD4 level above the threshold where prophylaxis is recommended[28-32]. Our data provide further support for the safety of discontinuing prophylactic therapy for PCP and MAC at appropriate CD4 thresholds. None of the opportunistic illnesses that _developed in patients on potent combination antiretroviral therapy presented in an unusual manner, such as mycobacterial lymphadenitis[33,34].
Similar to the results from the Swiss HIV Cohort, we found that achieving an undetectable HIV-1 RNA was associated with a decreased risk of opportunistic illness or death irrespective of whether or not the HIV-1 RNA level subsequently increased to detectable levels. These results reflect only relatively short-term effects, and longer follow-up will be needed to make a more definitive inference. It is of interest that sustained _suppression of HIV-1 RNA in patients treated with PIs is not as strongly linked with disease progression as in those treated with dual nucleoside therapy. These findings suggest that viral rebound in patients on PI therapy may be due to resistant viruses with attenuated pathogenicity, as evidenced by sustained CD4 cell recovery in the face of viremia.
In summary, our analysis has described a decline in opportunistic illness and death in a clinical cohort characterized by a high percentage of patients of minority race and a history of injecting drug use. This decline appears to be a result of the use of potent combination antiretroviral therapy and has affected most, but not all, illnesses. In patients receiving three-or-more drugs combination therapy, the CD4 level is probably the best immediate predictor of the risk of developing an opportunistic illness or of dying, with the HIV-1 RNA response a predictor of outcome in as much as it correlates with subsequent CD4 level. The concurrent CD4 level is probably the best correlate of development of opportunistic illness or death. As the duration of time on potent combination antiretroviral therapy increases further, it will continue to be important to assess the changing natural history of HIV disease and the _effectiveness and durability of current and new HIV therapies in clinical practice.
1. Carpenter CCJ, Fischl MA, Hammer SM, et al
. Antiretroviral therapy for HIV infection in 1998.JAMA
2. Panel on Clinical Practices for Treatment of HIV Infection. Guidelines for the use of antiretroviral agents in HIV-infected adults and adolescents.
Department of Health and Human Services. MMWR
1998, 47 RR-1
3. Phair J, Munoz A, Detels R, et al
. The risk of Pneumocystis carinii pneumonia among men infected with human immunodeficiency virus type I. N Engl J Med
4. Moore RD, Chaisson RE. Natural history of opportunistic disease in an HIV-infected urban clinical cohort.Ann Intern Med
5. U.S. Department of Health and Human Services, Centers for Disease Control and Prevention. 1997 USPHS/IDSA guidelines for the prevention of opportunistic infection in persons infected with human immunodeficiency virus.MMWR
6. Yarchoan R, Venson DJ, Pluda JM et al
. CD4 count and the risk for death in patients infected with HIV receiving antiretroviral therapy.Ann Intern Med
7. Moore RD, Keruly JC, Richman DD, et al
. Natural history of advanced HIV disease in patients treated with zidovudine.AIDS
8. Palella FJ, Delaney KM, Moorman AC, et al
. Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. N Engl J Med
9. Forrest DM, Seminari E, Hogg RS, et al
. The incidence and spectrum of AIDS-defining illnesses in persons treated with antiretroviral drugs.Clin Infect Dis
10. Mocroft A, Vella S, Benfield TL, et al
. Changing patterns of mortality across Europe in patients infected with HIV-1. Lancet
11. Centers for Disease Control and Prevention. Update: Trends in AIDS Incidence - United States, 1996.MMWR
12. Hogg RS, Heath KV, Yip B, et al
. Improved survival among HIV-infected individuals following initiation of antiretroviral therapy.JAMA
13. Mellors JW, Rinaldo CR Jr, Gupta P, et al
. Prognosis in HIV-1 infection predicted by the quantity of virus in plasma.Science
14. Gulick RM, Mellors J, 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
15. Hammer SM, Squires KE, Fischl MA, et al
. A controlled trial of two nucleoside analogs plus indinavir in persons with human immunodeficiency virus infection and CD4 cell counts of 200 per cubic millimeter or less.N Engl J Med
16. Moore RD. Understanding the clinical and economic outcomes of HIV therapy: The Johns Hopkins HIV clinical practice cohort.J Acquir Immune Defic Syndr
17. Centers for Disease Control and Prevention. 1993 Revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults.MMWR
18. Diggle PJ, Lian KY, Zeger SL. Analysis of Longitudinal Data.
New York, NY: Oxford University Press; 1994.
19. Clayton D, Hills M. Statistical Models in Epidemiology
. New York, NY: Oxford University Press; 1993.
20. Autran B, Carcelain G, Li TS, et al
. Positive effects of combined antiretroviral therapy on CD4+ T-cell homeostasis and function in advanced HIV disease.Science
21. Connors M, Kovacs S, Gea-Banacloche JC, et al
. HIV infection induces changes in CD4+ T-cell phenotype and depletions within the CD4+ T-cell repertoire that are not immediately restored by antiviral or immune-based therapies. Nat Med
22. Pakker NG, Notermans DW, Boer RDD, Roos MTL, Wolf FD, Hill A. Biphasic kinetics of peripheral blood T cells after triple combination therapy in HIV-1 infection: a composite of redistribution and proliferation.Nat Med
23. Gorochov G, Neumann AU, Kereveur A, et al
. Perturbation of CD4 and CD8 T-cell repertoires during progression to AIDS and regulation of the CD4 repertoires during antiviral therapy.Nat Med
24. Li TS, Tubiana R, Katlama C, Calvez V, Ait Mohand H, Autran B. Long-lasting recovery in CD4 T-cell function and viral load reduction after highly antiretroviral therapy in advanced HIV-1 disease.Lancet
25. Lederman MM, et al
. Immunologic response associated with 12 weeks of combination antiretroviral therapy consisting of zidovudine, lamivudine, and ritonavir: results of AIDS Clinical Trial Group protocol 315.J Infect Dis
26. Carr A, Samaras K, Burton S, et al
. A syndrome of peripheral lipodystrophy, hyperlipidaemia and insulin resistance in patients receiving HIV protease inhibitors.AIDS
27. Michelet C, Arvieus C, Francois C, et al
. Opportunistic infections occurring during highly active antiretroviral treatment.AIDS
28. Vasquez EG, Frances JM, Delgado RG, Mainez C, Gorgolas M, Fernandez Guerero ML. Withdrawal of Pneumocystis carinii pneumonia (PCP) prophylaxis after HAART in patients with AIDS.12th World AIDS Conference,
Geneva, Switzerland, June 1998 [abstract 22186].
29. Schneider MME, Borleffs JCC, Stolk RP, Jaspers CAJJ, Hoepelman AIM. Discontinuation of prophylaxis for Pneumocystis carinii pneumonia in HIV-1 infected patients treated with highly active antiretroviral therapy.Lancet
30. Rodriguez-Guardado A, Maradona JA, Carton L, Casado L, Asensi V. Pneumocystis carinii prophylaxis can be discontinued after CD4 + cell recovery over 200 ∞ 106/1 [letter]. AIDS
31. Ravaux I,
Chadapavid S, Quinson AN, Dahan V, Gallais H. Discontinue primary prophylaxis regimens in selected HIV-infected patients treated with HAART.12th World AIDS Conference,
Geneva, Switzerland, June 1998 [abstract 31225].
32. Reiss P, Weverling GJ, Mocroft A, et al
. Can prophylaxis for PCP and cerebral toxoplasmosis be safely discontinued in HIV-infected patients successfully treated with HAART? The EuroSIDA study.12th World AIDS Conference
, Geneva, Switzerland, June, 1998 [abstract 12268].
33. Race EM, Adelson-Mitty J, Kriegel GR, et al
. Focal mycobacterial lymphadenitis following initiation of protease inhibitor therapy in patients with advanced HIV-1 disease.Lancet
34. Jacobson MA, French M. Altered natural history of AIDS-related opportunistic infection in the era of potent combination antiretroviral therapy.AIDS
35. Ledergerber B, Egger M, Opravil M, et al
. Clinical progression and virological failure on highly active antiretroviral therapy in HIV-1 patients: a prospective cohort study.Lancet
36. Katzenstein DA, Hammer SM, Hughes MD, et al
. The relation of virologic and immunologic markers to clinical outcomes after nucleoside therapy in HIV-infected adults with 200 to 500 CD4 cells per cubic millimeter.N Engl J Med
Keywords:© 1999 Lippincott Williams & Wilkins, Inc.
Opportunistic illness; combination therapy; natural history; protease inhibitors; CD4; HIV-1 RNA