Search

As the HIV-infected population ages and the burden of chronic comorbidities increases, adherence to medications for HIV and diabetes and hypertension is crucial to improve outcomes. We pilot-tested a pictorial aid intervention to improve medication adherence for both HIV and common chronic conditions. Adult patients with HIV and diabetes (DM) and/or hypertension (HTN) attending a clinic for underserved patients and at risk for poor health outcomes were enrolled. Patients were randomized to receive either a pictorial aid intervention (a photographic representation of their medications, the indications, and the dosing schedule) or a standard clinic visit discharge medication list. Adherence to antiretroviral therapy (ART) for HIV and therapy for DM or HTN was compared. Predictors of ART adherence at baseline were determined using logistic regression. Medication adherence was assessed using medication possession ratio (MPR) for the 6-month interval before and after the intervention. Change in adherence by treatment group was compared by ANOVA. Among the 46 participants, there was a trend towards higher adherence to medications for HIV compared with medications for hypertension/diabetes (baseline median MPR for ART 0.92; baseline median MPR for the medication for the comorbid condition 0.79, p = 0.07). The intervention was feasible to implement and satisfaction with the intervention was high. With a small sample size, the intervention did not demonstrate significant improvement in adherence to medications for HIV or comorbid conditions. Patients with HIV are often medically complex and may have multiple barriers to medication adherence. Medication adherence is a multifaceted process and adherence promotion interventions require an approach that targets patient-specific barriers.

Objective: To compare the effectiveness of initial highly active antiretroviral therapy with either: a single protease inhibitor (PI); ritonavir (RTV)/saquinavir (SQV); or efavirenz (EFV) plus nucleoside reverse transcriptase inhibitors., Design: Cohort study., Setting: Urban HIV clinic., Patients: Five-hundred and forty-five HIV-1-infected individuals with minimal antiretroviral exposure who started combination therapy with > or = 3 antiretroviral drugs and > or = 1 NRTI to which they had not previously been exposed (single PI, 416; RTV/SQV, 68; EFV, 61)., Main Outcome Measures: HIV-1 RNA < 400 copies/ml within 8 months of starting therapy; time to HIV-1 RNA rebound to > 1000 copies/ml in the subset of patients achieving initial viral suppression; change in CD4 cell count from baseline within 12 months of starting therapy., Results: By intent-to-treat analysis, initial viral suppression was achieved by 72% of patients in the EFV group, compared to 49% in the single PI group (P = 0.001) and 51% in the RTV/SQV group (P = 0.019). Among patients who achieved initial viral suppression, time to viral rebound was similar in the three groups. Durable viral suppression (> or = 3 consecutive HIV-1 RNA levels < 400 copies/ml for > 6 months) was achieved by 53% of patients in the EFV group, 26% in the single PI group, and 29% in the RTV/SQV group (P < 0.05 for both comparisons with EFV). The median CD4 cell count increase was 139 x 10(6) cells/l, and was similar in the three groups., Conclusions: In agreement with a recent clinical trial, use of initial EFV-based combination antiretroviral therapy was associated with higher rates of viral suppression than PI-based therapy in a clinical cohort.

Background: Pancreatitis is a known adverse effect of the nucleoside reverse transcriptase inhibitors, particularly didanosine. Hydroxyurea has been used to potentiate the antiviral efficacy of didanosine, but recently there has been concern that severe and even fatal pancreatitis may be more likely to occur when hydroxyurea is used in combination with didanosine. We investigated the incidence of pancreatitis in patients using nucleoside analogues with or without hydroxyurea., Methods: Data were obtained from patients followed longitudinally on the Johns Hopkins HIV Clinic. Incidence rates of pancreatitis were calculated for each antiretroviral regimen that included zidovudine, stavudine, didanosine (+ hydroxyurea), and didanosine + stavudine (+ hydroxyurea). Poisson regression was used to compare the relative rate of pancreatitis for each regimen adjusting for other covariates., Results: A total of 2613 patients received at least one of the nucleoside reverse transcriptase inhibitor-containing regimens. There were 33 cases of pancreatitis. The crude incidence rate of pancreatitis ranged from 0.18 cases per 100 person-years on therapy for zidovudine to 6.25 cases per 100 person-years for didanosine + hydroxyurea. Compared to didanosine alone, and adjusting for CD4 cell count and other variables, the relative risk (RR) of pancreatitis was 8.56 [95% confidence interval) CI, 1.85-35.59] for didanosine + hydroxyurea, and 2.35 (95% CI, 0.46-11.89) for didanosine + stavudine + hydroxyurea. For any use of hydroxyurea, the RR = 4.01 (95% CI, 1.02-15.89). Other risk factors for pancreatitis included a CD4 cell count < 200 x 106 cells/l, female sex, and a history of pancreatitis., Conclusions: Our data show that the risk of pancreatitis is four-fold higher when hydroxyurea is used. The use of hydroxyurea with didanosine should probably be discouraged if other treatment options are available.

Data derived from studies conducted before 1996 consistently showed that anemia was a common occurrence in patients with human immunodeficiency (HIV) Infection and an Independent risk factor for early death. Correction of anemia was associated with reversal of this increased risk. Highly active antiretroviral therapy (HAART) has been shown to reduce HIV disease progression and mortality. In the HAART era, HIV-related anemia is still common and independently associated with decreased survival, with a decreased risk of mortality associated with recovery from anemia. Epoetin alfa treatment has been shown to correct anemia and significantly improve quality of life (QOL) in patients with HIV disease. Additional data on the effect of correction of anemia with epoetin alfa treatment on survival in patients with HIV infection are needed.

To determine the relationship of combination antiretroviral therapy and bacterial pneumonia, we assessed incidence of and risk factors for bacterial pneumonia in 1,898 human immunodeficiency virus (HIV)-infected patients with CD4 cell counts < 200/mm(3) followed in the Johns Hopkins HIV clinic between 1993 and 1998. A total of 352 episodes of bacterial pneumonia occurred during 2,310 patient-years of follow-up. Incidence of bacterial pneumonia decreased from 22.7 episodes/100 person-years (py) in the first half of 1993 to 12.3 episodes/100 py in the first half of 1996, reaching a nadir of 9.1 episodes/100 py in the second half of 1997 (p < 0.05). The use of protease inhibitor-containing regimens was associated with a decreased risk of bacterial pneumonia (risk ratio [RR] 0.55, 95% CI 0.31 to 0.94). Lower CD4 cell counts (RR 2.22, 95% CI 1.54 to 3.18), injection drug use as HIV transmission category (RR2.0, 95% CI 1.43 to 2.76), and prior Pneumocystis carinii pneumonia (RR 3.88, 95% CI 1.65 to 9.16) were also significantly associated with bacterial pneumonia. Trimethoprim-sulfamethoxazole and macrolide use were not significantly associated with risk of bacterial pneumonia. There has been a dramatic decline in the incidence of bacterial pneumonia resulting from the use of combination antiretroviral therapy containing protease inhibitors.

Background: Sensory neuropathy is a common adverse effect of the nucleoside analogue anti-retroviral drugs didanosine (ddl) and stauvudine (d4T). These drugs are increasingly being used in combination, and it is not currently known whether the incidence of neuropathy is higher with combination compared to individual drug use. It is also not known if hydroxyurea, used to potentiate the antiviral efficacy of these drugs, may also increase the risk of neuropathy. The purpose of this analysis is to investigate if the combination of ddl and d4T, with or without hydroxyurea, has a higher incidence of neuropathy than a single drug regimen., Methods: Data were obtained from patients followed longitudinally by the Johns Hopkins AIDS Services. Incidence rates of development of neuropathy were calculated for each of five regimens: ddl (+/- hydroxyurea), ddl + d4T (+/- hydroxyurea), and d4T. Cox proportional hazard regression was used to compare the relative risk of neuropathy for each regimen adjusting for CD4 cell count, other drugs received, and time on therapy., Results: A total of 1116 patients received at least one of the five regimens. There were 117 cases of neuropathy. The crude incidence rate of neuropathy ranged from 6.8 cases per 100 person-years for ddl to 28.6 cases per 100 person-years for ddl + d4T + hydroxyurea. Compared with ddl alone, and adjusting for CD4 cell counts and other variables, the relative risk of neuropathy was 1.39 [95% confidence interval (CI): 0.84-2.32] for d4T alone, 2.35 (95% CI: 0.69-8.07) for ddl + hydroxyurea, 3.50 (95% CI: 1.81-6.77) for ddl + d4T, and 7.80 (95% CI: 3.92-15.5) for ddl + d4T + hydroxyurea., Conclusions: Based on the data, the risk of neuropathy is additive or even synergistic for ddl + d4T + hydroxyurea compared with ddl or d4T alone. The combination of ddl + d4T also increases the risk of neuropathy but less than when hydroxyurea is included.

Context: Use of antiretroviral drugs, including protease inhibitors, for treatment of human immunodeficiency virus (HIV) infection has been anecdotally associated with hepatotoxicity, particularly in persons coinfected with hepatitis C or B virus., Objectives: To ascertain if incidence of severe hepatotoxicity during antiretroviral therapy is similar for all antiretroviral drug combinations, and to define the role of chronic viral hepatitis in its development., Design: Prospective cohort study., Setting: University-based urban HIV clinic., Patients: A total of 298 patients who were prescribed new antiretroviral therapies between January 1996 and January 1998, 211 (71%) of whom received protease inhibitors as part of combination therapy (median follow-up, 182 days) and 87 (29%) of whom received dual nucleoside analog regimens (median follow-up, 167 days). Chronic hepatitis C and B virus infection was present in 154 (52%) and 8 (2.7%) patients, respectively., Main Outcome Measure: Severe hepatotoxicity, defined as a grade 3 or 4 change in levels of serum alanine aminotransferase and aspartate aminotransferase, evaluated before and during therapy., Results: Severe hepatotoxicity was observed in 31 (10.4%) of 298 patients (95% confidence interval [CI], 7.2%-14.4%). Ritonavir use was associated with a higher incidence of toxicity (30%; 95% CI, 17.9% -44.6%). However, no significant difference was detected in hepatotoxicity incidence in other treatment groups, ie, nucleoside analogs (5.7%; 95% CI, 1.2%-12.9%), nelfinavir (5.9%; 95% CI, 1.2%-16.2%), saquinavir (5.9%; 95% CI, 0.15%-28.7%), and indinavir(6.8%; 95% CI, 3.0%-13.1 %). Although chronicviral hepatitis was associated with an increased risk of severe hepatotoxicity among patients prescribed nonritonavir regimens (relative risk, 3.7; 95% CI, 1.0-11.8), most patients with chronic hepatitis C or B virus infection (88%) did not experience significant toxic effects. Rate of severe toxicity with use of any protease inhibitor in patients with hepatitis C infection was 12.2% (13/107; 95% CI, 6.6%-19.9%). In multivariate logistic regression, only ritonavir (adjusted odds ratio [AOR], 8.6; 95% CI, 3.0-24.6) and a CD4 cell count increase of more than 0.05 x 10(9)/L (AOR, 3.6; 95% CI, 1.0-12.9) were associated with severe hepatotoxicity. No irreversible outcomes were seen in patients with severe hepatotoxicity., Conclusions: Our data indicate that use of ritonavir may increase risk of severe hepatotoxicity. Although hepatotoxicity may be more common in persons with chronic viral hepatitis, these data do not support withholding protease inhibitor therapy from persons coinfected with hepatitis B or C virus.

Background: The use of protease inhibitor-containing (PI) combination antiretroviral therapy has led to a reduction in the incidence of opportunistic illness and mortality (events) in HIV infection. We wished to quantify the changing incidence of these events in our clinical practice and delineate the relationship between CD4, HIV-1 RNA, and development of events in patients receiving PI combination therapy., Methods: We assessed HIV-infected patients with CD4 counts < or =500 cells x10(6)/l. We calculated the incidence of events from 1994 through 1998 and analyzed the association of temporal changes in event incidence and use of antiretroviral therapy. In patients on PI combination therapy, we determined the probability of achieving and maintaining an undetectable HIV-1 RNA response and determined the association of CD4, HIV-1 RNA, and developing an event., Results: The incidence of opportunistic illness declined from 23.7 events/100 person-years in 1994 to 14.0 events/100 person-years in 1998 (P<0.001). Mortality declined from 20.2 deaths/100 person-years in 1994 to 8.4 deaths/ 100 person-years in 1998 (P<0.001). Use of PI combination therapy was associated with a relative rate of opportunistic illness or death of 0.66 [95% confidence interval (CI), 0.51-0.85; P<0.001]. The relative incidence of each of 16 opportunistic illnesses was approximately the same in 1998 as in 1994 except for lymphoma, cervical cancer and wasting syndrome which do not appeared to have declined in incidence. Approximately 60% of patients who received PI therapy achieved an undetectable HIV-1 RNA, and 65% of these patients maintained durable suppression of HIV-1 RNA. Achieving an undetectable HIV-1 RNA was associated with a decreased risk of an event, and was the variable most strongly associated with an increase in CD4 level. By multivariate analysis, the concurrent CD4 level was most strongly associated with developing an event., Conclusions: We observed a significant decline in the incidence of opportunistic illness and death from 1994 through 1998 associated with combination antiretroviral therapy. Patients who develop events while being treated with PI combination therapy were not likely to have achieved an undetectable HIV-1 RNA and are likely to have a low concurrent CD4 level.

Background: In clinical trials, highly active antiretroviral therapy (HAART) reduces plasma HIV-1 RNA levels to less than 500 copies/mL in 60% to 90% of patients with HIV-1 infection. The performance of such therapy outside of the clinical trial setting is unclear., Objective: To determine factors associated with failure to suppress HIV-1 RNA levels and adverse drug reactions in a cohort of patients in whom protease inhibitor-containing therapy was begun in a large urban clinic., Design: Retrospective cohort study., Setting: Johns Hopkins HIV Clinic in Baltimore, Maryland., Patients: 273 protease inhibitor-naive patients began taking a protease inhibitor regimen containing at least one other antiretroviral drug to which the patients had not previously been exposed., Measurements: Demographic variables, plasma HIV-1 RNA levels, CD4+ lymphocyte counts, and adverse drug reactions., Results: Levels of HIV-1 RNA were undetectable in 42% of the cohort at 1 to 90 days, in 44% at 3 to 7 months, and in 37% at 7 to 14 months. Factors associated with failure to suppress viral load at two or more time points included higher rates of missed clinic appointments, nonwhite ethnicity, age 40 years or younger, injection drug use, lower baseline CD4+ lymphocyte count, and higher baseline viral load. In a multivariate model, only higher rates of missed clinic appointments were independently associated with viral suppression at 1 year. Ritonavir was associated with adverse drug reactions about twice as frequently as indinavir or nelfinavir, and women experienced significantly more adverse effects than men., Conclusions: Unselected patients in whom HAART is started in a clinic setting achieve viral suppression substantially less frequently than do patients in controlled clinical trials. Missed clinic visits were the most important risk factor for failure to suppress HIV-1 RNA levels. Studies are needed to identify interventions that maximize the performance of HAART in inner-city clinics.

Objectives: In the USA, Medicaid is the principal payer of the health care costs of patients with HIV infection. We wished to determine how the costs to Medicaid of patients in Maryland infected with HIV have changed in the setting of highly active antiretroviral treatment., Design: Observational cohort study., Methods: Analysis of combined economic and clinical data of patients from the Johns Hopkins HIV Service, the provider of primary and sub-specialty care for a majority of HIV-infected patients in the Baltimore metropolitan region. All patients were enrolled in Medicaid and received care longitudinally in Maryland from 1 January 1995 through 31 December 1997. Monthly Medicaid payments were calculated for all inpatient and outpatient services by fiscal year, CD4 cell count, and use of protease inhibitors., Results: For inpatients with a CD4 cell count < or = 50 x 10(6) cells/l, the total health care average monthly payments remained unchanged ($2629 in 1995, $2585 in 1997). Total mean monthly payments increased for those with a CD4 cell count > 50 x 10(6) cells/l (CD4 cell count 50-200 x 10(6) cells/l, $1172 in 1995 and $1615 in 1997, P < 0.05; CD4 cell count 201-500 x 10(6) cells/l, $1078 in 1995 and $1305 in 1997, P < 0.05). However, when data were stratified according to use of a protease inhibitor-containing regimen (used during approximately 50% of follow-up time in 1996-1997) it was found that hospital inpatient payments decreased significantly in all CD4 strata for patients on a protease inhibitor-containing regimen whereas pharmacy payments increased significantly. Inpatient payments associated with treating opportunistic illness were lower in 1996-1997 for patients receiving protease inhibitor therapy compared with those not receiving protease inhibitors. On balance, total health care payments tended to be slightly lower for patients receiving a protease inhibitor regimen., Conclusion: Although protease inhibitor-containing antiretroviral regimens are being used by only about half of our Medicaid-insured patients, when they are used, there are significantly lower hospital inpatient and community care costs, as well as lower costs associated with the treatment of opportunistic illness. Even with the concurrent increase in their pharmacy costs, total health care costs were stable or slightly lower for these patients. We believe this is a favorable result suggesting a good clinical value being achieved without an increase in costs.

Objective: To assess the effect of prior zidovudine (ZDV) use on subsequent response to stavudine (D4T)-containing regimens., Design: Analysis of data from prospective observational database., Methods: Patients were ZDV-experienced if they had previously received more than 90 days of ZDV and ZDV-naive if they had never received ZDV. HIV-1 RNA and CD4 cell counts were compared at 3, 6, and 12 months after initiation of D4T. Univariate and multivariate analyses were performed, adjusting for baseline HIV-1 RNA and CD4 cell count, age, sex, race, HIV transmission category, time since enrollment, and protease inhibitor use., Results: No difference was found between ZDV-experienced (n = 130) and naive (n = 98) patients in age, sex, race, transmission category, use of a concurrent protease inhibitor, or baseline CD4 cell count and HIV-1 RNA. There was no difference in the median decline in HIV-1 RNA (-1.29 log10 copies/ml for experienced patients versus -1.19 log10 copies/ml for naive patients; P = 0.39), in achieving HIV-1 RNA < 400 copies/ml at 3 months (51% versus 49%; P = 0.79) or 6 months (48% versus 56%; P = 0.33). There was no difference in CD4 cell response (+73 x 10(6)/l versus + 87 x 10(6)/l; P = 0.51). By multivariate adjustment in a repeated measures analysis, there was no significant difference in achieving undetectable HIV-1 RNA or in CD4 cell response between experienced and naive patients., Conclusion: No difference in response to a D4T-containing regimen between ZDV-experienced and naive patients was found over a 1-year period. In contrast to previous trials, most patients in this study also received a protease inhibitor. These findings may be more relevant in the current era of highly active antiretroviral therapy.

Context: The US Public Health Service and the International AIDS Society-USA recently published recommendations for antiretroviral therapy (ART) for persons infected with human immunodeficiency virus (HIV); however, anecdotal evidence suggests that HIV-infected injection drug users (IDUs) may not be receiving optimal care as defined by the recommendations., Objective: To assess ART use in HIV-infected IDUs., Design: A cross-sectional survey of self-reported ART use between July 1996 and June 1997 in IDUs., Setting: A community-based clinic affiliated with Johns Hopkins University, Baltimore, Md., Participants: A total of 404 HIV-infected IDUs with CD4+ cell counts less than 0.50 x 10(9)/L recruited into a longitudinal study in 1988 and 1989., Main Outcome Measure: Self-reported ART use was assessed: no current therapy, monotherapy, or combination therapy with or without a protease inhibitor., Results: One half (199/404 [49%]) of patients reported no recent ART. A total of 14% (58/404) had monotherapy, 23% (90/404) were receiving combination therapy without a protease inhibitor, and 14% (57/404) had triple-combination therapy with a protease inhibitor. A multivariate analysis of factors associated with ART showed that care continuity and recent HIV-related outpatient visit (odds ratio [OR], 4.30; 95% confidence interval [CI], 2.36-7.81 and OR, 2.84; 95% CI, 1.66-4.88, respectively), CD4+ cell count of less than 0.20 x 10(9) (OR, 2.41; 95% CI, 1.51-3.84), no current drug use and being in drug treatment (OR, 2.16; 95% CI, 1.34-3.47; OR, 2.12; 95% CI, 1.23-3.66, respectively), and unemployment (OR, 2.31; 95% CI, 1.21-4.40) were associated with reporting ART use. In other analysis, less likely to receive protease inhibitors were current drug injectors (OR, 0.5; 95% CI, 0.3-1.0) and those recently incarcerated (OR, 0.2; 95% CI, 0.03-0.9), but patients with acquired immunodeficiency syndrome were more likely to receive protease inhibitors (OR, 2.0; 95% CI, 0.9-4.6). Protease inhibitor use doubled (P<.01) from July and December 1996 to January and June 1997 (7.7% and 14.8%, respectively)., Conclusions: Those IDUs infected with HIV who were not receiving ART tended to be active drug users without clinical disease who have less contact with health care providers. Although we do not have information on clinical judgment regarding treatment decisions or whether persons were prescribed therapy not taken, the proportion of subjects reporting receiving ART suggests that strategies for improving treatment in this population are indicated. Expanding simultaneous treatment services for HIV infection and substance abuse would enhance the response to these related epidemics.

Background: Medication nonadherence in the treatment of chronic diseases compromises the effectiveness of therapy. Little information is available about the extent of medication adherence or determinants of medication adherence in HIV disease, an issue of increasing importance in this new therapeutic era of combination antiretroviral therapy., Methods: We studied 244 HIV-infected Medicaid-insured patients attending an HIV hospital-based clinic regarding the extent of and predictors of adherence to antiretroviral therapy and Pneumocystis carinii pneumonia (PCP) prophylaxis. Patients were asked to report medications being taken, patterns of use, and knowledge and attitudes about HIV therapies. Medical record report of type, dose, and frequency of medication was compared with self-report using the kappa statistic. Urine sulfamethoxazole assay was obtained from patients prescribed sulfamethoxazole-trimethoprim., Results: Among patients prescribed antiretroviral therapy, 60% reported > or = 80% adherence in the previous 7 days; 49% reported > or = 80% adherence with PCP prophylaxis in the previous seven days. Seventy-nine percent of patients who reported taking daily sulfamethoxazole-trimethoprim had detectable urinary sulfamethoxazole. In multivariate analysis, > or = 80% adherence to antiretroviral therapy was associated with taking medication < or = twice a day (odds ratio [OR]=1.44; 95% confidence interval [CI], 1.01, 1.96), being likely to take medication when not at home, (OR=1.41; 95% CI, 1.04, 2.00) and patients' belief in their ability to adhere to therapy (OR=1.57; 95% CI, 1.13, 2.17). For PCP prophylaxis, > or = 80% adherence was associated with presence of family (OR=2.39; 95% CI, 1.01, 5.63) and patients' belief in their ability to adhere to therapy (OR=2.87; 95% CI, 1.44-1.78). Sociodemographic characteristics and belief in the efficacy of medications were not associated with adherence., Conclusions: A relatively low level of adherence to antiretroviral therapy and PCP prophylactic regimens was found. Although our results are principally from patients receiving antiretroviral monotherapy, these findings may have important implications for patients receiving highly active antiretroviral therapy (HAART). Decreasing the complexity of antiretroviral regimens, and working with patients to modify identified barriers to adherence may improve effectiveness of medications and prolong survival.

The aim of this study was to evaluate the association between acyclovir use and survival in HIV-infected patients. To achieve this, we used survival analysis in an observational cohort of HIV-infected patients enrolled in primary care at an urban HIV clinic. We measured survival in a cohort of HIV-infected patients who had CD4 cell counts < or = 500/mm3 and who enrolled for care at a single urban HIV clinic between December 1988 and April 1995. We compared survival in users of acyclovir alone, zidovudine alone, and acyclovir and zidovudine in combination with the survival of those using neither drug. Factors associated with improved survival were identified using Cox proportional hazards analysis. Among the 1408 patients enrolled, there were no significant differences in overall survival between acyclovir users and non-users. After adjustment for CD4 cell count, the use of other antiretroviral agents, race, transmission risk and a history of herpesvirus infection, acyclovir use alone was independently associated with a relative hazard (RH) of death of 1.008 (P = 0.969); zidovudine use alone with a RH of 0.559 (P < 0.001); and combination use of acyclovir and zidovudine associated with a RH of 1.062 (P = 0.788). Therefore we conclude that the use of acyclovir is not associated with prolonged survival in this cohort of HIV-infected patients.

Unlabelled: OBJECTIVES--To determine if racial-ethnic differences exist in survival, disease progression, and development of myelosuppression in zidovudine-treated patients with advanced human immunodeficiency virus (HIV) disease. DESIGN--Prospective observational study.Setting: -Hospital and private clinics in 12 metropolitan centers.Patients: -The study included 754 non-Hispanic white, 165 black, and 106 Hispanic patients with the acquired immunodeficiency syndrome (AIDS) or advanced AIDS-related complex (ARC) who received up to 2 years of zidovudine therapy. OUTCOME MEASURES--Survival, development of Pneumocystis carinii pneumonia (PCP), other opportunistic infections, and myelosuppression. RESULTS--At initiation of zidovudine therapy, Hispanic and particularly black patients had more advanced HIV disease than white patients, as indicated by lower baseline CD4+ counts, hematocrits, and AIDS-defining diagnoses. Black patients with AIDS also had a worse prognosis compared with white and Hispanic patients with AIDS. The product-limit survival rates at 2 years for white, black, and Hispanic patients with AIDS were 40%, 27%, and 39%, respectively (black vs white, P = .01; Hispanic vs white, P = .32, by the log-rank test). The respective proportions of patients who developed PCP at 2 years were 46%, 66%, and 44% (black vs white, P = .0001; Hispanic vs white, P = .86) and for other opportunistic infections the proportions were 56%, 63%, and 63%, respectively (black vs white, P = .03; Hispanic vs white, P = .09). There were no significant racial-ethnic differences in survival or in the development of opportunistic infections for patients with ARC, and there were no differences in the incidence of myelosuppression or dose reduction or suspension for patients with either ARC or AIDS. After adjusting for more advanced HIV disease (mainly low CD4+ counts and hematocrits), black race was no longer a significant independent predictor of survival. Adjustment for racial differences in the use of PCP prophylaxis accounted for most of the excess risk for the development of PCP in black patients compared with white patients with AIDS. CONCLUSIONS--Racial differences in survival and the development of opportunistic infections are mainly due to the more advanced HIV disease in black patients when zidovudine therapy is started and to their less frequent use of PCP prophylaxis. Innovative approaches are needed to ensure more widespread use of and earlier access to zidovudine therapy and PCP prophylaxis. [ABSTRACT FROM AUTHOR]