In 1996, for the first time during the AIDS epidemic, AIDS incidence and AIDS-related deaths decreased in the USA[1,2]. The estimated incidence of AIDS opportunistic illnesses declined 6% during 1996 compared with 1995, and estimated deaths decreased 23% during the same period[1,2]. These decreases in AIDS incidence and death have been attributed largely to successful prophylactic regimens to prevent the development of opportunistic infections and the use of combination antiretroviral therapies (ART)[1,2].
Clinical trials and cohort studies have been conducted to assess the effectiveness of new ART and determine which therapies will be most successful in routine clinical practice. The current most effective antiretroviral treatment seems to be the combination of two nucleoside analogue drugs with a protease inhibitor (PI)[3,4]. The goal of triple combination therapy is to suppress plasma HIV RNA to levels that are undetectable. This combination has decreased plasma HIV RNA levels and increased CD4 T-lymphocyte counts in persons infected with HIV[5,6]. In addition, Hammer et al.  found that triple ART which included a PI was associated with a decreased risk of death. These results were limited to patients with CD4 T-lymphocyte counts of ≤200×106 cells/l who had been previously treated with zidovudine (ZDV). Although only limited information is available to determine the survival benefits of triple-combination therapies, preliminary data allow us to determine some early effects of these regimens on survival.
The benefits of Pneumocystis carinii pneumonia (PCP) prophylaxis on survival have been associated with increased survival time[8-11]. However, those results are limited by the difficulty of examining the effects of PCP prophylaxis independent of ART because most patients who were receiving PCP prophylaxis in those studies were also receiving ART. Data on the effect of Mycobacterium avium complex (MAC) prophylaxis on survival are limited[12,13]. However, for patients diagnosed with disseminated MAC[14,15], MAC bacteremia, or M.kansasii , the use of antimycobacterial agents is associated with increased survival. Data on the effect of pneumococcal vaccine on survival are also limited.
The purpose of the analysis was to assess the effects of monotherapy, dual- and triple-combination ART, PCP and MAC primary prophylaxis, pneumococcal vaccination, and other potentially prognostic risk factors on the survival of persons whose AIDS diagnosis was made through 1997. We used observational data representing a large number of HIV-infected patients in a variety of clinical settings across the USA. In addition to examining the time from AIDS diagnosis to death and factors contributing to survival, we assessed temporal trends in survival by examining survival time among cohorts of patients whose diagnosis was made between 1990 and 1992 compared with those whose diagnosis was made between 1993 and 1995. These 2-year periods were selected to determine if the survival rate for this population was increasing over time.
HIV-infected persons were observed through the Adult/Adolescent Spectrum of Disease (ASD) project. The methods used in ASD have been published. Persons observed in the ASD project and included in this analysis are aged 13 years or older and are enrolled in 93 study hospitals and clinics in Atlanta, Dallas, Denver, Detroit, Houston, Los Angeles, New Orleans, San Antonio, and Seattle. Medical records of these patients are reviewed retrospectively for the 12 months before their study entry date and at 6 month intervals until the last patient contact or death. In addition to basic demographic information and mode of exposure to HIV, information abstracted from patient medical records included the prescribed use of antiretroviral, chemoprophylactic, and chemotherapeutic medications, CD4 T-lymphocyte counts, complete history of AIDS-defining diseases, and other infections. We included persons whose AIDS diagnosis (1993 case definition) was made during 1990 through December 1997, and for whom follow-up data were collected through January 1998.
The effect of prescribed ART on survival relative to the lack of prescribed ART was evaluated by multivariate analysis. Patients receiving ART were grouped into monotherapy, or into dual- or triple-combination therapies. Effects of individual drugs and individual drug combinations were assessed in a separate model. Previously and currently recommended ART for persons with HIV were examined[3,4]. Monotherapies included in the model were defined as the single or sequential use of the nucleoside analogue DNA polymerase reverse transcriptase inhibitors ZDV, didanosine (ddI), or stavudine (d4T). Dual- and triple-combination regimens were defined, respectively, as prescribed use of two or three specified therapies during one follow-up interval. Dual ART included ZDV combined with ddI, zalcitabine (ddC), or lamivudine (3TC), and d4T combined with ddI or 3TC. Triple ART was defined as ZDV combined with ddI, ddC, or 3TC and with either a PI (indinavir, nelfinavir, ritonavir, or saquinavir) or a non-nucleoside analogue reverse transcriptase inhibitor (NNRTI) (nevirapine or delavirdine), and as d4T combined with ddI or 3TC and with a PI or an NNRTI. The monotherapy and combination ART accounted for 98.3% of the total observation time for ART. A covariate for other ART (not described) was also included in the multivariate model.
Primary chemoprophylaxis against PCP was defined as any prescribed use of trimethoprim-sulfamethoxazole (TMP-SMX), dapsone, or aerosolized pentamidine, alone or in combination, during a follow-up interval prior to or in the absence of PCP diagnosis. Patients who received TMP-SMX or dapsone during the same interval in which they received a PCP diagnosis were considered to be receiving treatment, not prophylaxis, during that interval. Primary prophylaxis against disseminated MAC was determined similarly. Patients receiving clarithromycin, rifabutin, or azithromycin alone or in combination with no prior presumptive or definitiveMAC or M.kansasii diagnosis were coded as receiving primary MAC chemoprophylaxis. We assessed the effects of PCP and MAC primary prophylaxis on survival time relative to lack of prescribed use of these medications by controlling separately for the effects of treatment and secondary prophylaxis in the multivariate model. Because of the potential colinearity between opportunistic infection prophylaxis and ART, the effects of primary prophylaxis were determined in a subset analysis restricted to observation time of patients who were receiving ART.
The effect of pneumococcal vaccine on survival after AIDS diagnosis was assessed for patients who had any indication of vaccination in their medical records during the study period. Other potential risk factors in the survival analysis were minimum CD4 T-lymphocyte count during each 6 month interval, age at the end of each 6 month period, history of an AIDS opportunistic illness before the end of each 6 month period, year of AIDS diagnosis, sex, mode of HIV exposure, and race/ethnicity.
Trends in prescribed use of monotherapy, and dual- and triple-combination ART among persons with AIDS were examined by quarter-year of observation from 1990 through 1997. Similarly, trends in the prescribed use of PCP and MAC primary chemoprophylaxis and the receipt of pneumococcal vaccination during the study period were examined for persons with AIDS during the same period.
SAS statistical software was used for all statistical analyses. Andersen-Gill multiplicative hazards regression with a counting process framework was used to estimate survival risks. Using the counting process formulation, the data for each patient were represented by multiple observations, each observation for a specific interval, the values of the explanatory variables during that period, and the patient‚s vital status at the end of each interval. This method allowed the delayed entry of a patient at risk for death (left censoring); i.e., persons whose AIDS diagnosis is retrospective can contribute risk information at the start of prospective follow-up. Although deaths due to suicide, homicide, or drug overdose may be HIV-related, these cases were right-censored at the end of the interval preceding death.
We investigated assumptions of proportional hazards by using plots of weighted Schoenfeld residuals versus time. We used model stratification to study the interaction of the effects of primary prophylaxis, ART, and pneumococcal vaccination on risk of death with HIV exposure mode and varying levels of CD4 T- lymphocyte counts.
Survival probabilities by month since AIDS diagnosis were estimated by the Kaplan-Meier method. Persons whose AIDS diagnosis preceded the prospective record review were excluded from this computation of survival time. Cumulative survival probability curves were constructed of persons whose diagnosis was made during 1990-1992 and 1993-1995. The proportion of AIDS cases diagnosed as an immunologic case-CD4 T-lymphocyte count of <200×106 cells/l or CD4 cells <14% relative to all AIDS cases, immunologic or clinical (diagnosis of an AIDS-defining illness) -was determined from this same subset of patients whose diagnosis was made after entry into the ASD study.
The proportional hazards regression analysis included 19 565persons with AIDS. Of the patients, 15% were female, 41% were white, 44% black, and 13% Hispanic. Most patients were aged 25-44 years (78%) at the time of observation (4% were aged <25 years and 2% aged ≥60 years). Mode of exposure to HIV was recorded for 89% of the patients, most of whom were men who had sex with men (60%), followed by injecting drug users (17%), men who both have had sex with men and were injecting drug users (12%), and those infected by heterosexual contact (9%).
There were 9280 deaths during 29555 person-years of observation. Right censoring was performed for 188 patients whose causes of death were homicide, suicide, or drug overdose. However, estimated risks did not change with the inclusion of these deaths in the multivariate model. Prescription of all regimens of ART significantly reduced the risk of mortality (Table 1). Triple ART had the most profound effect on the risk of death, followed by dual ART, and the least effect was exhibited by monotherapy. The independent effects of these interventions on survival are displayed as a risk ratio and 95% confidence limits (CL). The effects of medications were not significantly different according to CD4 T-lymphocyte count or by HIV exposure category.
Among monotherapy regimens, persons prescribed d4T had a decreased risk of death compared with persons prescribed ddI. Among persons prescribed dual ART, d4T+ddI had a greater effect on the risk of death than both ZDV+ddC and ZDV+ddI. No significant differences were found among triple-combination ART regimens containing ZDV compared to those containing d4T.
The risk of death was decreased among persons receiving primary PCP prophylaxis, and persons receiving primary MAC prophylaxis. The receipt of pneumococcal vaccination was associated with slightly improved survival compared with persons who had not received pneumococcal vaccination.
Other factors affecting survival after the development of AIDS included age <25 years, age 45-59 years and age ≥60 years, compared with the reference group, aged 25-44 years. Black race, compared with white race, was associated with lower risk of death. The risk of death increased as CD4 T-lymphocyte count declined through the following categories: 200-499×106 cells/l, 50-199×106 cells/l, and 0-49×106 cells/l, compared with patients who had ≥500×106 cells/l (Table 2).
The use of ART changed dramatically between 1990 and 1997 (Fig. 1). ART was prescribed for 82% of the patients with AIDS in our study population. Antiretroviral monotherapy was prescribed for approximately 75% of patients observed during 1990 and declined to less than 5% in 1997. Dual ART use increased to almost 40% of patients by the end of 1996 and declined in 1997 as the proportion of patients prescribed triple-combination ART steadily increased-from its introduction in the fourth quarter of 1995-to almost 60% of all patients with AIDS by the fourth quarter of 1997.
The use of antimicrobial prophylaxis also changed during this time (Fig. 2). More than 63% of the total study population with AIDS and CD4 T cell counts of <200×106 cells/l were prescribed primary PCP prophylaxis. Approximately 25% of our study population with AIDS and CD4 counts of <75×106 cells/l were prescribed primary MAC prophylaxis, and 42% received pneumococcal vaccine. Primary PCP prophylaxis remained relatively stable throughout the study period; primary MAC prophylaxis steadily increased. By 1995, approximately 15% of patients were prescribed MAC prophylaxis; by the end of 1997, the proportion increased to more than 30%. Of persons who had not yet received pneumococcal vaccination, the percentage vaccinated declined from nearly 20% during the early 1990s to less than 15% during the mid-1990s.
Of the patients included in the Kaplan-Meier analysis, 67% were diagnosed as immunological cases, and 33% were diagnosed as clinical AIDS cases under the Centers for Disease Control (CDC) AIDS case definition. Median survival for all patients in this analysis was 32 months (95% CI, 32-33 months), and 62.7% survived 2 years or longer (95% CI, 61.6-63.8 months). Persons whose cases were diagnosed as immunological between 1990 and 1997 had a median survival of 39 months (95% CI, 39-41 months), which was significantly longer than the median survival of 19 months (95% CI, 19-20 months) for persons whose cases were diagnosed as clinical AIDS during this same period. The estimated probability of 2 year survival for persons with immunological AIDS was 72.9% (95% CI, 71.7-74.1%), which was significantly more than the 42.1% (95% CI, 40.2-44.0%) estimated probability of survival for persons diagnosed with clinical AIDS.
The estimated 4 year survival rate was greater for persons whose AIDS diagnosis was made between 1993 and 1995 (41.3; 95% CI, 38.4-44.1) compared with persons whose diagnosis was made between 1990 and 1992 (31.2; 95% CI, 29.6-32.8) (Fig. 3). Median survival increased from 31 months (95% CI, 30-32 months) among persons whose AIDS diagnosis was made between 1990 and 1992 to 35 months (95% CI, 35-38 months) among persons whose diagnosis was made between 1993 and 1995.
In our study, the use of triple ART was associated with increased survival of persons with AIDS. This is consistent with the findings of Hammer et al.  who compared ZDV-experienced patients who were taking either ZDV+3TC+indinavir or dual combinations of ZDV+3TC or ZDV+indinavir, and found decreased risk of death and delayed disease progression among patients who were receiving triple ART. Recent studies have demonstrated that PI decrease viral load and increase CD4 T-lymphocyte counts in patients infected with HIV[5,6], which will probably contribute to prolonged survival. When comparing the same combinations, Gulick et al.  found that the triple ART combination resulted in a greater decrease in HIV RNA and a greater increase in CD4 T-lymphocyte counts. The comparison of ZDV-experienced patients prescribed ZDV+ddC+saquinavir or dual combinations of ZDV+ddC or ZDV+saquinavir, showed a greater decrease in plasma HIV RNA and increased CD4 T-lymphocyte counts among patients who received triple ART. Palella et al. also found that combination therapy, particularly those regimens containing a PI, benefited survival.
Primary PCP and MAC prophylaxis were associated with increased survival time. Research has demonstrated the difficulty of analyzing the effects of prophylaxis on survival independent of ART. Most patients who receive prophylaxis for PCP and MAC are concurrently receiving ART; therefore, the effect of prophylactic therapies on survival independent of ART is difficult to determine. However, we were able to examine the effects of these primary prophylactic therapies in the presence of ART, allowing us to determine that both PCP and MAC primary chemoprophylaxis reduced the risk of death significantly.
For the purpose of this analysis, we assumed that patients receiving triple ART were prescribed these medications in accordance with published guidelines for the use of ART. ZDV monotherapy decreased the risk for an opportunistic illness [25,26] and improved the survival of persons infected with HIV[8,25-32]. However, the therapeutic effects of ZDV monotherapy are limited and decline over time[8,9,26,28,29]. When antiretroviral combination therapy was introduced, it was discovered that the use of ZDV+ddI or ZDV+ddC improved survival compared with ZDV monotherapy[32-34], and antiretroviral combination therapy became the standard of care. Triple ART is currently recommended for HIV-infected persons with symptomatic HIV infection and should be offered to patients with asymptomatic infections if CD4 T-lymphocyte counts are <500×106 cells/l or HIV RNA >20000 (reverse transcription- PCR) or >10000 (branched DNA).
We assumed in this analysis that medications were prescribed in accordance with the 1995 USPHS/IDSA Guidelines for the Prevention of Opportunistic Infections in Persons Infected with Human Immunodeficiency Virus . The 1995 USPHS/ IDSA guidelines recommend PCP prophylaxis with TMP-SMX for patients with CD4 T-lymphocyte counts of <200×106 cells/l, unexplained fever of >100°F for >2 weeks, or history of oropharyngeal candidiasis. For patients who cannot tolerate TMP-SMX, the recommended regimens are dapsone alone, dapsone combined with pyrimethamine and leucovorin, or aerosolized pentamidine. To prevent disseminated MAC, the 1995 USPHS/IDSA guidelines recommend MAC chemoprophylaxis with rifabutin for HIV-infected persons with CD4 T-lymphocyte counts <75×106 cells/l. Clarithromycin, azithromycin, and combinations of clarithromycin or azithromycin with rifabutin were being evaluated at the time these guidelines were published. The 1997 USPHS/IDSA guidelines changed the recommendation for MAC prophylaxis to include HIV-infected persons with CD4 T-lymphocyte counts of <50×106 cells/l. The drugs currently recommended for the prevention of MAC infection are clarithromycin or azithromycin. Rifabutin is still recommended if patients cannot tolerate clarithromycin or azithromycin. Under these guidelines, patients receiving PCP prophylaxis would have been prescribed ART, and patients receiving MAC prophylaxis would be simultaneously receiving both PCP prophylaxis and ART, unless contraindicated.
Additional positive effects on survival for patients receiving PCP chemoprophylaxis in this study may be attributed to the protective effects of TMP-SMX [36-38] and dapsone [36,39] against toxoplasmosis and serious bacterial infections[36,40,41]. Similarly, both clarithromycin and azithromycin may prevent bacterial infections of the respiratory tract, thus possibly contributing to increased survival in our patient population.
The CDC‚s Advisory Committee on Immunization Practices recommends that persons infected with HIV receive pneumococcal vaccination. The antibody response to pneumococcal vaccine is sustained for approximately 5 years in persons with normal immune function, but reimmunization after 5 years may not enhance immunity to the pneumococcus in persons with HIV. Some patients may have received a pneumococcal vaccination prior to the study period, resulting in only a slight difference between the vaccinated and non-vaccinated groups. Other factors influencing survival after AIDS diagnosis included age, race, and CD4 T-lymphocyte count. Our finding that younger age at the time of AIDS diagnosis is associated with decreased risk of death agrees with that of other studies[9,42-44]. Increased risk of death as CD4 T-lymphocyte counts decrease, as found in this analysis, has been documented by several other investigators[28,45,46]. Our finding that black race was associated with increased survival is similar to that of Friedland et al., who found the same results among black men with PCP. Lemp et al.  found that race did not influence survival, and Chaisson, Keruly and Moore  reported that race was not associated with disease progression; other studies have shown black race to be associated with decreased survival[27,47,48]. Further investigation of the association between race and survival is needed.
The rate of PCP-associated deaths began to decline after 1989, and the annual incidence of AIDS opportunistic illnesses and deaths declined between 1995 and 1996. Although these findings indicate successes with the use of ART and opportunistic illness chemoprophylaxis, severe opportunistic illnesses continue to be the most common cause of HIV-related mortality, and HIV-related mortality continues to be a leading cause of death among persons aged 25-44 years in the USA.
The increased probability of survival for persons whose AIDS diagnosis was made between 1993 and 1995 compared with those whose diagnosis was made between 1990 and 1992 could be due to the widespread use of ART and opportunistic illness chemoprophylaxis. This study showed that the prescription of antiretroviral monotherapy declined from approximately 80% of patients in 1990 to approximately 30% in 1995, and showed an increasing trend in the prescription of more effective dual ART from 1993 to 1995. The prescription of primary prophylaxis against PCP has remained relatively steady. However, the prescription of MAC primary prophylaxis has increased from a small proportion of patients in 1992 to almost 20% of patients by the end of 1995. These advances in ART regimens and the prescription of recommended chemoprophylaxis may contribute to longer survival of persons whose diagnosis was made in later years, although persons whose diagnosis was made earlier may also benefit from triple ART.
As this study was a medical record review, we may have incomplete death reporting of persons whose diagnosis was made in more recent years. This may have resulted in an underestimation of the risk of death for patients with AIDS.
Patients whose diagnosis was made between 1990 and 1992 had 6 years of potential follow-up, compared with 3 years for persons whose diagnosis was made between 1993 and 1995. This decreased follow-up time may have resulted in an overestimation of the probability of death in the more recent period, providing a conservative estimate of the differences between the two cohorts. Temporal trends in survival time could not be assessed for patients whose AIDS diagnosis was made after 1995 due to insufficient observation time.
Prospective studies have established the efficacy of ART and antimicrobial prophylaxis in prolonging the lives of HIV-infected persons[25,50,51]. Estimates of treatment effect in an observational study may be confounded by preferential use of therapy, poor compliance with, or lack of adherence to, prescribed treatments, and discontinuation of medications, potentially resulting in conservative estimates of the survival benefits.
We used a counting process method to estimate survival risks. A disadvantage of using this method is the inherent assumption that a covariate is independent of death. Patients who discontinue medication because of the perception that death is near will probably contribute observation time in which death occurs in the absence of medications, leading to an overestimation of benefit in association with these drugs. This disadvantage is balanced by the methodological advantages of incorporating time-varying data as it occurs and the use of all AIDS cases despite left-censored data.
Patients in the ASD study are recorded as receiving treatment or prophylaxis if prescription of the drug was documented during the 12-month initial period or the 6-month follow-up interval. Initiation and termination dates, reasons for prescription, and dosage of drugs are not documented in ASD. It is assumed that treatment remains unchanged during each discrete time interval. Concurrent therapy is assumed for ART if more than one drug was recorded during one interval and combinations met current standards for the treatment of HIV/AIDS, although these data may actually represent a switch to new medications. This potential misclassification of therapy would presumably result in an underestimate of the beneficial effect of combination therapies.
The median survival time from AIDS diagnosis to death was 32 months. However, we have demonstrated that survival time is lengthening. There has been a decreasing trend in the use of antiretroviral monotherapies between 1990 and 1997, accompanied by an increase in the use of dual therapies and, more recently, an increase in the use of triple-combination ART. We have also seen steady use of primary PCP prophylaxis, and increased use of primary MAC prophylaxis during this time. The proportion of previously unvaccinated persons who are receiving pneumococcal vaccine has declined.
Our results may represent the achievable benefits of therapy regimens used to treat HIV-infected patients in the USA. Dual-combination ART was associated with decreased risk of death when compared with monotherapy, and triple ART involving a PI or NNRTI and two nucleoside analogue drugs was associated with the lowest risk of death. The risk of death is 2.5 times lower for persons receiving triple ART than persons receiving monotherapy and 1.6 times lower than the risk for persons receiving dual therapy.
1. Centers for Disease Control. Update: trends in AIDS incidence - United States, 1996.MMWR
2. Centers for Disease Control. Update: trends in AIDS incidence, deaths, and prevalence - United States, 1996.MMWR
3. Carpenter CJ, Fischl MA, Hammer SM, et al.Antiretroviral therapy for HIV infection in 1996: recommendations of an international panel.JAMA
4. Centers for Disease Control and Prevention. Report of the NIH Panel to define principles of therapy of HIV infection and guidelines for the use of antiretroviral agents in HIV-infected adults and adolescents.MMWR
5. 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
6. Collier AC, Coombs RW, Schoenfeld DA, et al.Treatment of human immunodeficiency virus infection with saquinavir, zidovudine, and zalcitabine. AIDS Clinical Trials Group. N Engl J Med
7. Hammer SM, Squires KE, Hughes 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
8. Graham NMH, Zeger SL, Park LP, et al.The effects on survival of early treatment of human immunodeficiency virus infection. N Engl J Med
9. Saah AJ, Hoover DR, He Y, the Multicenter AIDS Cohort Study. Factors influencing survival after AIDS: report from the multicenter AIDS cohort study (MACS). J Aquir Immune Defic Syndr
10. Chaisson RE, Keruly JC, Moore RD. Race, sex, drug use, and progression of human immunodeficiency virus disease. N Engl J Med
11. Osmond D, Charlebois E, Lang W, Shiboski S, Moss A. Changes in AIDS survival time in two San Francisco cohorts of homosexual men, 1983 to 1993. JAMA
12. Pierce M, Crampton S, Henry D, et al.A randomized trial of clarithromycin as prophylaxis against disseminated Mycobacterium avium complex infection in patients with advanced acquired immunodeficiency syndrome. N Engl J Med
13. Moore, RD, Chaisson RE. Survival analysis of two controlled trials of rifabutin prophylaxis against Mycobacterium avium complex in AIDS. AIDS
14. Horsburgh Jr CR, Metchock B, Gordon SM, Havlik Jr JA, McGowan JE, Thompson III SE. Predictors of survival in patients with AIDS and disseminated Mycobacterium avium complex disease. J Infect Dis
15. Ives DV, Davis RB, Currier JS. Impact of clarithromycin and azithromycin on patterns of treatment and survival among AIDS patients with disseminated Mycobacterium avium complex. AIDS
16. Chin DP, Reingold AL, Stone EN, et al.The impact of Mycobacterium avium complex bacteremia and its treatment on survival of AIDS patients - a prospective study. J Infect Dis
17. Campo RE, Campo CE. Mycobacterium kansasii disease in patients infected with human immunodeficiency virus. Clin Infect Dis
18. Farizo KM, Buehler JW, Chamberland ME, et al.Spectrum of disease in persons with human immunodeficiency virus infection in the United States. JAMA
19. Centers for Disease Control. 1993 Revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. MMWR
20. SAS Institute, Inc. SAS/STAT User‚s Guide, Version 6.12
. Cary, NC: SAS Institute Inc.; 1989.
21. Andersen PK, Gill RD. Cox‚s regression model counting process: a large sample study. Ann Stat
22. Grambsch PM, Therneau TM. Proportional hazards tests and diagnostics based on weighted residuals. Biometrics
23. Kaplan EL, Meier P. Non-parametric estimation from incomplete observations. J Am Stat Assoc
24. Palella Jr FJ, Delaney KM, Moorman AC, et al.Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. N Engl J Med
25. Fischl MA, Richman DD, Greico MH, et al.The efficacy of azidothymidine (AZT) in the treatment of patients with AIDS and AIDS-related complex: a double-blind, placebo-controlled trial. N Engl J Med
26. Volberding PA, Lagakos SW, Grimes JM, et al.The duration of zidovudine benefit in persons with asymptomatic HIV infection: prolonged evaluation of protocol 019 of the AIDS clinical trials group. JAMA
27. Moore RD, Hidalgo J, Sugland BW, Chaisson RE. Zidovudine and the natural history of acquired immunodeficiency syndrome. N Engl J Med
28. Hanson DL, Horsburgh Jr CR, Fann SA, Havlik JA, Thompson III SE. Survival prognosis of HIV-infected patients.J Aquir Immune Defic Syndr
29. Lundgren JD, Phillips AN, Pedersen C, et al.Comparison of long-term prognosis of patients with AIDS treated and not treated with zidovudine. JAMA
30. Lemp GF, Payne SF, Neal D, Temelso T, Rutherford GW. Survival trends for patients with AIDS. JAMA
31. De Grottola V, Wulfsohn M, Fischl MA, Tsiatis A. Modeling the relationship between survival and CD4 lymphocytes in patients with AIDS and AIDS-related complex.J Aquir Immune Defic Syndr
32. Graham NMH, Hoover DR, Park LP, et al.Survival in HIV infected patients who have received zidovudine: comparison of combination therapy with sequential monotherapy and continued zidovudine monotherapy.Ann Intern Med
33. Delta Coordinating Committee. Delta: a randomised double-blind controlled trial comparing combinations of zidovudine plus didanosine or zalcitabine with zidovudine alone in HIV-infected individuals.Lancet
34. Hammer SM, Katzenstein DA, Hughes MD, et al.A trial comparing nucleoside monotherapy with combination therapy in HIV-infected adults with CD4 cell counts from 200 to 500 per cubic millimeter. N Engl J Med
35. Centers for Disease Control. USPHS/IDSA guidelines for the prevention of opportunistic infections in persons with human immunodeficiency virus: a summary. MMWR
36. Centers for Disease Control and Prevention. USPHS/IDSA guidelines for the prevention of opportunistic infections in persons with human immunodeficiency virus.MMWR
37. Bozette SA, Finkelstein DM, Spector SA, et al
. A randomized trial of three antipneumocystis agents in patients with advanced human immunodeficiency virus infection. N Engl J Med
38. Jones JL, Hanson DL, Chu SY, et al.
Toxoplasmic encephalitis in HIV-infected persons: risk factors and trends. AIDS
39. Torres RA, Barr M, Thorn M, et al.Randomized trial of dapsone and aerosolized pentamidine for the prophylaxis of Pneumocystis carinii pneumonia and toxoplasmic encephalitis. Am J Med
40. Hardy WD, Feinberg J, Finkelstein DM, et al.A controlled trial of trimethoprim-sulfamethoxazole or aerosolized pentamidine for secondary prophylaxis of Pneumocystis carinii pneumonia in patients with the acquired immunodeficiency syndrome: AIDS clinical trials group protocol 021. N Engl J Med
41. Mayer HB, Rose DN, Cohen S, Gurtman AC, Cheung TW, Szabo S. The effect of Pneumocystis carinii pneumonia prophylaxis regimens on the incidence of bacterial infections in HIV-infected patients.AIDS
42. Centers for Disease Control and Prevention. Prevention of pneumococcal disease: recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR
43. Janoff EN, Breiman RF, Daley CL, Hopewell PC. Pneumococcal disease during HIV infection: epidemiologic, clinical, and immunologic perspectives. Ann Intern Med
44. Friedland GH, Saltzman B, Vileno J, Freeman K, Schrager LK, Klein RS. Survival differences in patients with AIDS. J Aquir Immune Defic Syndr
45. Easterbrook PJ, Emami J, Gazzard B. Rate of CD4 cell decline and prediction of survival in zidovudine-treated patients. AIDS
46. Mocroft A, Youle M, Morcinek J, et al.Survival after diagnosis of AIDS: a prospective observational study of 2625 patients. Br Med J
47. Rothenberg R, Woelfel M, Stoneburner R, Milberg J, Parker R, Truman B. Survival with the acquired immunodeficiency syndrome: experience with 5833 cases in New York City. N Engl J Med
48. Harris JE. Improved short-term survival of AIDS patients initially diagnosed with Pneumocystis carinii pneumonia, 1984 through 1987. JAMA
49. Selik RM, Karon JR, Ward JW. Effect of the human immunodeficiency virus epidemic on mortality from opportunistic infections in the United States in 1993.J Infect Dis
50. Torres R, Thorn M, Newlands J, et al.Randomized trial of intermittent dapsone versus aerosolized pentamidine for primary and secondary prophylaxis of Pneumocystis carinii pneumonia (PCP). VI International Conference on AIDS.
San Francisco, June 1990 [abstract ThB407].
51. Chaisson RE, Keruly J, Richman DD, Moore RD. Pneumocystis prophylaxis and survival in patients with advanced human immunodeficiency virus infection treated with zidovudine: the zidovudine epidemiology group. Arch Intern Med
The Adult/Adolescent Spectrum of Disease Group: M. Thompson, AIDS Research Consortium of Atlanta, Atlanta, Georgia; F. Sorvillo, Los Angeles County Department of Health Services, Los Angeles, California; S. Buskin, Seattle-King County Department of Public Health, Seattle, Washington; G. Thompson, Texas Department of Health, Austin, Texas; L. Wotring, Michigan Department of Community Health, Detroit, Michigan; A. Davidson, Denver Department of Health and Hospitals, Denver, Colorado; S. Troxler, Louisiana Department of Health, New Orleans, Louisiana; W. McNeely, Department of Health and Human Services, Houston, Texas, USA.