Rapid clearance of virus after acute HIV-1 infection: correlates of risk of AIDS

OBJECTIVE

Our objective was to define early virologic and immunologic determinants of human immunodeficiency virus (HIV) type 1 disease progression among 22 case subjects with acute infection from the Trinidad Seroconvertor Cohort.

METHODS

A linear segmented regression model was fitted to sequential quantitative virus load measurements. Parameters of virus kinetics during different phases of primary infection were correlated with clinical and immunologic end points, by use of Kaplan-Meier estimates and Cox regression.

RESULTS

Ten individuals developed acquired immunodeficiency syndrome (AIDS)-defining events. In univariate analysis, progression to AIDS was associated with rate of initial HIV clearance (P=.002), virus load during set point (P=.008), and CD4(+) cell count during steady state (P=.04). In the multivariate analysis, a rapid rate of initial clearance was the sole independent predictor of subsequent progression to AIDS and was associated with a 92% reduction in the risk of AIDS. The rate of initial clearance is inversely correlated with the number of early symptoms (r=-0.66; P=.0008). However, symptoms did not predict subsequent risk of AIDS.

CONCLUSION

Among a subset of patients, rapid clearance of plasma HIV-1 after peak viremia is associated with lower viral set point, prolonged virus suppression before loss of virologic control, and decreased risk of AIDS. These findings are consistent with the hypothesis that effective immune responses during the earliest phase of infection are important determinants of the subsequent natural history.

First report of human immunodeficiency virus transmission via an RNA-screened blood donation

BACKGROUND AND OBJECTIVES

Blood banks in the USA have recently introduced minipool nucleic acid amplification testing (MP-NAT) of blood products to reduce the transmission of human immunodeficiency virus (HIV) and hepatitis C virus (HCV) by transfusions. However, MP-NAT is limited in its ability to detect preseroconversion samples with very low viral RNA loads.

MATERIALS AND METHODS

To determine whether a red blood cell unit, from an MP-NAT-negative donation, transmitted HIV when transfused to a patient, we compared the viral sequences from the blood donor and recipient. The implicated donation was also tested by commercially available NAT assays at a range of dilution factors to determine whether the infectious unit could have been detected using individual-donation NAT (ID-NAT).

RESULTS

Phylogenetic linkage of HIV sequences in the blood donor and recipient confirmed the transmission of HIV by blood transfusion, the first such case identified since introduction of MP-NAT screening in 1999. Viral RNA was reliably detected by ID-NAT, but only inconsistently detected by MP-NAT.

CONCLUSIONS

Even following the introduction of MP-NAT, a preseroconversion donation with a viral load of <or= 150 copies of RNA/ml went undetected and resulted in an HIV transmission. Implementation of ID-NAT will further reduce such rare transmissions, but at a considerable cost per infectious unit interdicted.

Multicenter evaluation of a new, automated enzyme-linked immunoassay for detection of human immunodeficiency virus-specific antibodies and antigen

A collaborative multicenter study was conducted to evaluate the sensitivity, specificity, and precision of a three-step, fully automated, qualitative microparticle-based enzyme-linked immunoassay (AxSYM HIV Ag/Ab Combo; Abbott Laboratories), designed to simultaneously detect (i). antibodies against human immunodeficiency virus type 1 (HIV-1) and/or type 2 (HIV-2) and (ii). HIV p24 antigen. A significant reduction in the HIV seroconversion window was achieved by combining detection of HIV antibodies and antigen into a single assay format. For 22 selected, commercial HIV seroconversion panels, the mean time of detection with the combined-format HIV antigen-antibody assay was reduced by 6.15 days compared to that with a similar third-generation single-format HIV antibody assay. The quantitative sensitivity of the combination assay for the p24 antigen (17.5 pg/ml by use of the p24 quantitative panel VIH SFTS96') was nearly equivalent to that of single-format antigen tests. The combination assay demonstrated sensitive (100%) detection of anti-HIV immunoglobulin in specimens from individuals in CDC stages A, B, and C and from individuals infected with different HIV-1 group M subtypes, group O, or HIV-2. The apparent specificity for hospitalized patients (n = 1938) was 99.90%. In a random population of 7900 volunteer blood donors, the specificity (99.87%) was comparable to that of a third-generation single-format HIV antibody assay (99.92%) on the same donor specimens. In addition, the combination assay was robust to potential interfering specimens. The precision of the combination was high, with intra- and interrun variances of <or=9.3% for each precision panel specimen or assay control and <or=5.3% for the negative assay control.

Effector HIV-specific cytotoxic T-lymphocyte activity in long-term nonprogressors: associations with viral replication and progression

Ex vivo effector cytotoxic T-lymphocyte (CTL) activity was assessed in 27 members of the Australian Long-Term Nonprogressor cohort and correlated with genetic, virological, and immunological markers. The 27 individuals were antiretroviral naive with CD4(+) T-cell counts of >500 cells/ microl for more than 8 years after human immunodeficiency virus type 1 (HIV-1) infection. Effector CTL activity was determined using a standard ex vivo chromium release assay. Individuals with CTL activity (HIV-1 env(IIIB) or pol or gag) were then compared to those without CTL activity in relation to plasma HIV-1 RNA, ICD p24 antigen, beta(2)-microglobulin, CD4 and CD8 T-cell counts, CCR5 and CCR2b genotypes, and progression to CD4 <500 cells/microl or commencement of antiretroviral treatment. Of the 27 individuals examined, 19 had no detectable effector CTL activity. The eight individuals with detectable CTL activity had significantly higher plasma levels of HIV-1 RNA (P = 0.014), immune complex dissociated p24 antigen (P = 0.006), and beta(2)-microglobulin (P = 0.009). There was increased risk of progression within 4 years of study entry in individuals with detectable effector CTL activity, higher plasma levels of HIV-1 RNA, higher beta(2)-microglobulin levels, and higher immune complex dissociated p24 antigen levels at enrollment (P = 0.017, P = 0.004, P = 0.027, P = 0.008 respectively). Multivariate analysis demonstrated viral load remained the strongest predictor of disease progression within this group (P = 0.017). There were no significant associations between CTL response and chemokine receptor genotype. These findings demonstrate the importance of HIV replication in generating an effector CTL response and show that effector CTL activity may be an early predictor of progression in people with long-term asymptomatic HIV infection.

Performance of a quantitative human immunodeficiency virus type 1 p24 antigen assay on various HIV-1 subtypes for the follow-up of human immunodeficiency type 1 seropositive individuals

The heat-denatured signal-amplified p24 antigen assay is a low-cost test allowing the determination of plasma levels of HIV-1 p24 antigen in infected patients. This assay may be appropriate for monitoring disease progression in HIV seropositive patients in developing countries. Only a few data on the clinical validation of the test are available for HIV-1 non-subtypes B viruses that represent the vast majority of virus circulating in Africa. The present study was undertaken to evaluate and compare the performance of a heat-denatured signal-amplified p24 assay for the determination of p24 viral load in the plasma of individuals infected with different subtypes of HIV-1 and using the RT-PCR-based RNA viral load test as the gold standard. A total of 120 plasma samples from individuals infected with HIV-1 strains belonging to group M (subtypes A-->H) and group O, as well as recombinant strains, were tested in parallel with the heat-denatured signal-amplified p24 assay and the RNA viral load. Plasma p24 levels appeared to be correlated significantly with the plasma RNA viral loads (R=0.751, P<0.0001). The heat-denatured p24 antigen assay was capable of measuring the plasma level of p24 derived from all the HIV-1 subtypes and recombinants selected for this study, in contrast to the RNA viral load test which lacked sensitivity towards HIV-1 group O. The heat-denatured signal-amplified p24 assay is a reliable, sensitive and a more affordable tool that can be used for the follow-up of patients infected with B and non-B subtypes as well as recombinant forms of HIV-1 in developing countries.

The induction of Toll-like receptor tolerance enhances rather than suppresses HIV-1 gene expression in transgenic mice

Microbial-induced proinflammatory pathways are thought to play a key role in the activation of human immunodeficiency virus type 1 (HIV-1) gene expression. The induction of Toll-like receptor (TLR) tolerance leads to a complex reprogramming in the pattern of inflammatory gene expression and down-modulates tumor necrosis factor alpha (TNF-alpha), interleukin (IL)-1, and IL-6 production. Using transgenic (Tg) mice that incorporate the entire HIV-1 genome, including the long-terminal repeat, we have previously demonstrated that a number of different TLR ligands induce HIV-1 gene expression in cultured splenocytes as well as purified antigen-presenting cell populations. Here, we have used this model to determine the effect of TLR-mediated tolerance as an approach to inhibiting microbial-induced viral gene expression in vivo. Unexpectedly, Tg splenocytes and macrophages, rendered tolerant in vitro to TLR2, TLR4, and TLR9 ligands as assessed by proinflammatory cytokine secretion and nuclear factor-kappaB activation, showed enhanced HIV-1 p24 production. A similar enhancement was observed in splenocytes tolerized and then challenged with heterologous TLR ligands. Moreover, TLR2- and TLR4-homotolerized mice demonstrated significantly increased plasma p24 production in vivo despite lower levels of TNF-alpha. Together, these results demonstrate that HIV-1 expression is enhanced in TLR-reprogrammed host cells, possibly reflecting a mechanism used by the virus to escape the effects of microbial-induced tolerance during natural infection in vivo.

Changes in Immunoglobulin Isotypes and Immunoglobulin G (IgG) Subclasses During Highly Active Antiretroviral Therapy

SUMMARY

The effects of highly active antiretroviral therapy (HAART) on immunoglobulin isotypes and immunoglobulin G (IgG) subclasses were studied in 12 patients in early stages of HIV-1 infection. Blood samples were obtained at enrollment and 2, 4, 8, 12, 24, 48, and 120 weeks after initiation of HAART. Immunoglobulin concentrations were determined by nephelometry, and anti-p24-specific IgG and IgG1 levels were determined by an enzyme immunoassay. Overall time changes were analyzed in analysis of variance models. IgG and IgG1 levels showed a marked overall decline, whereas other immunoglobulin isotypes and IgG subclasses did not change significantly. Anti-p24-specific IgG1 levels decreased considerably and significantly more in virus isolation-negative patients than in virus isolation-positive patients, as defined according to the ability to isolate HIV-1 from their CD4+ T cells after initiation of therapy. Anti-p24 IgG levels showed a similar but overall weaker decline in the two groups. However, the anti-p24 IgG1 level followed the biphasic decline in plasma viremia more closely than the anti-p24 IgG level, with an initial sharp decline that leveled off with time. These findings suggest that the main reduction in immunoglobulin levels is caused by reduced HIV-1-specific antigen stimulation rather than a general reduction in immune activation. Using anti-p24 IgG1 as a parameter of response to the effect of HAART merits further investigation.

Viral RNA and p24 Antigen as Markers of HIV Disease and Antiretroviral Treatment Success

HIV-1 RNA has become the standard for monitoring antiretroviral therapies. Dogma predicts, however, that a viral protein like p24 should be at least as good a marker of HIV disease activity, provided that it is measured with sufficient sensitivity and accuracy. Simple modifications including use of a more efficient virus lysis buffer, heat-mediated destruction of antibodies interfering with antigen detection, and tyramide signal amplification for increased sensitivity have highly improved the HIV-1 p24 antigen assay. The p24 antigen assay is inferior to RT-PCR in detecting viral particles, but the presence of extraviral p24 antigen in most samples makes largely up for this. p24 antigen testing is similarly sensitive and specific in diagnosing pediatric HIV infection, in predicting CD4+ T cell decline and clinical progression at early and late stage of infection, and suitable for antiretroviral treatment monitoring in both adults and children. Notably, p24 antigen was measurable even in patients with stably suppressed viremia, and its concentrations were correlated negatively with the concentrations of CD4+ T cells and positively with the concentrations of activated CD8+ T cell subsets. p24 antigen is an excellent marker of HIV expression and disease activity and can be used in the same fields of application as HIV RNA is used. The test is validated for subtype B, but requires further studies for non-B subtypes.

Alternatives to HIV-1 RNA concentration and CD4 count to predict mortality in HIV-1-infected children in resource-poor settings

Cheaper, simpler alternatives to CD4 lymphocyte count and HIV-1 RNA detection for assessing the prognosis of HIV-1 infection are needed for resource-poor settings. However, little is known about the predictive value of alternative assays, in particular in children. We assessed the prognostic value of total lymphocyte count, immune complex-dissociated p24 antigen, white blood cell count, packed-cell volume (haematocrit), and serum albumin for mortality in 376 HIV-1-infected, mainly African-American or Hispanic children enrolled during March, 1988 to January, 1991. In a Cox proportional hazards model, including all assay-alternatives to CD4 and RNA, total lymphocyte count (p<0.0001) and serum albumin (p=0.0107) independently predicted mortality. Further assessment of these markers is warranted in resource-poor settings.

Dynamics of HIV viremia and antibody seroconversion in plasma donors: implications for diagnosis and staging of primary HIV infection

OBJECTIVES

The characterization of primary HIV infection by the analysis of serial plasma samples from newly infected persons using multiple standard viral assays.

DESIGN

A retrospective study involving two sets of archived samples from HIV-infected plasma donors. (A) 435 samples from 51 donors detected by anti-HIV enzyme immunoassays donated during 1984-1994; (B) 145 specimens from 44 donors detected by p24 antigen screening donated during 1996-1998.

SETTING

Two US plasma products companies.

MAIN OUTCOME MEASURES

The timepoints of appearance of HIV-1 markers and viral load concentrations during primary HIV infection.

RESULTS

The pattern of sequential emergence of viral markers in the 'A' panels was highly consistent, allowing the definition and estimation of the duration of six sequential stages. From the 'B' panels, the viral load at p24 antigen seroconversion was estimated by regression analysis at 10 000 copies/ml (95% CI 2000-93 000) and the HIV replication rate at 0.35 log copies/ml/day, corresponding to a doubling time in the preseroconversion phase of 20.5 h (95% CI 18.2-23.4 h). Consequently, an RNA test with 50 copies/ml sensitivity would detect HIV infection approximately 7 days before a p24 antigen test, and 12 days before a sensitive anti-HIV test.

CONCLUSION

The sequential emergence of assay reactivity allows the classification of primary HIV-1 infection into distinct laboratory stages, which may facilitate the diagnosis of recent infection and stratification of patients enrolled in clinical trials. Quantitative analysis of preseroconversion replication rates of HIV is useful for projecting the yield and predictive value of assays targeting primary HIV infection.

HIV-1 p24 antigen testing in blood banks: results from Saudi Arabia

HIV-1 p24 antigen testing was introduced to increase sensitivity in the early detection of HIV infection in blood donors. Since the introduction of HIV-1 p24 antigen testing in Saudi Arabia, we have failed to detect a single positive case. Over a three-year period, only four indeterminates were detected out of 24,654 blood donors. All four proved negative by confirmatory testing. Based on this experience, we believe that resources would be better directed to pooled nucleic acid testing and that p24 serological testing should be abandoned.

Detection of an early HIV-1 infection by HIV RNA testing in an Italian blood donor during the preseroconversion window period

BACKGROUND

The implementation of NAT technologies for HIV screening has further reduced the diagnostic window in recent HIV infection. There is still a debate regarding the cost effectiveness of genomic screening of blood donations for transfusion-transmitted viruses (HBV, HCV, HIV).

STUDY DESIGN AND METHODS

Since October 2001, at the Transfusion Service of Verona, single-donation NAT testing for HCV and HIV-1 (Procleix TMA HIV-1/HCV Assay) of all blood donations has been performed.

CASE REPORT

A case of acute HIV-1 infection detected by HIV NAT in a repeat blood donor who donated during the preseroconversion window period is reported. All blood components donated were discarded, and the donor started antiretroviral therapy 2 weeks after blood donation. HIV-1 p24 antigen was still negative 10 days after the HIV-1 RNA-positive blood donation. Seroconversion was documented by Day 41 after donation.

CONCLUSION

This case report testifies that HIV NAT screening of blood donation is effective in preventing the transmission of HIV infection through blood components.

HIV-1 p24 antigen is a significant inverse correlate of CD4 T-cell change in patients with suppressed viremia under long-term antiretroviral therapy

An HIV-1 p24 antigen test involving signal amplification-boosted ELISA of heat-denatured plasma was evaluated prospectively in 55 patients whose viral RNA in plasma had previously been suppressed for at least 6 months under antiretroviral combination therapy. During a median follow-up of 504 days, CD4 counts increased by a median of 62 cells per year. By univariate and multivariate linear regression analysis, the level of p24 antigen as expressed by the absorbance/cutoff ratio was a significant inverse correlate of both the CD4 count in a sample (p =.013) and its annual change in a patient (p <.0001). The p24 antigen retained significance even among 48 individuals whose HIV-1 RNA, apart from occasional blips, remained below 400 copies/mL. Batch-wise retesting of 70 samples from 5 such patients with a further improved procedure showed measurable p24 antigen in all but 1 sample and an inverse correlation with both the CD4 count (p =.0331) and percentage (p <.0001), thus confirming the prospectively generated data. Comparison of p24 antigen and HIV-1 RNA concentrations indicate that the p24 antigen detected in these samples is not associated with viral RNA-containing particles and may originate from other compartments of virus expression.

Transmission of HIV type 1 through blood transfusion from an antibody-negative/p24 antigen-negative donor

Transmission of HIV through transfusion occurred in a 78-year-old male after receiving red blood cells from a 27-year-old woman, who was negative for HIV antibodies and p24 antigen at the time of donation, but seroconverted thereafter. Plasma viral load at donation was 2538 HIV RNA copies/ml. Phylogenetic analyses demonstrated that viruses from both patients clustered tightly together, belonged to subtype B clade, and had no primary drug resistance mutations. This case highlights the residual risk of HIV transmission from donors in the window period, and supports the implementation of nucleic acid testing in blood banks.

HIV-specific immunity during structured antiviral drug treatment interruption

The immunologic correlates associated with control of viremia in HIV disease are poorly understood. We hypothesized that structured antiviral drug treatment interruptions could be utilized to better understand the relationship between HIV-specific immunity and viral replication. We thus examined the effects of two 8 weeks antiviral structured treatment interruptions (STIs) in a cohort of HIV-1 chronically infected individuals on highly active antiretroviral treatment (HAART) with (n = 13) and without (n = 12) therapeutic HIV immunizations. In this study, we observed that p24 gag antigen (np24) stimulated MIP-1beta levels and T helper immune responses prior to antiviral drug discontinuation were associated with control of viremia. Stronger and earlier production of gag peptide stimulated gamma interferon was observed in the immunized group during the structured antiviral drug interruptions. These results support the concept that HIV-specific immune responses are associated with control of viremia. Further study of immune-based therapies that enhance HIV-specific immunity is warranted.

False negative HIV-1 proviral DNA polymerase chain reaction in a patient with primary infection acquired in Thailand

BACKGROUND

Qualitative HIV-1 proviral DNA PCR tests have three main diagnostic applications. These include direct detection of viral sequences in the pre-seroconversion window period which may be positive up to 8 days prior to the development of HIV specific antibodies; resolution of indeterminate HIV serological tests and in the diagnosis of neonates born to seropositive mothers where maternal antibodies may be detectable for up to 15 months past partum.

METHODS

A total of three serial specimens from a single patient following symptomatic primary HIV-1 infection were assessed by commercially available serological assays for antibodies to HIV and HIV-1 antigen and nucleic acid amplification assays for proviral DNA (Amplicor HIV-1 test, Roche Molecular Systems, CA, USA) and RNA (HIV MONITOR ver1.5, Roche Molecular Systems) according to manufacturers recommendations. A subsequent modified PCR protocol for HIV proviral DNA was performed which included modified primers (SK145/SKCC1B) and altered thermal cycling parameters.

RESULTS

We describe a case of HIV infection acquired in Thailand by heterosexual transmission, where a commercially available HIV proviral DNA PCR assay remained negative despite a typical evolving serological profile consistent with seroconversion.

CONCLUSION

These data support the use of HIV DNA PCR tests only as a second line supplemental test to licensed standard HIV diagnostic testing strategies, and should be used with care in cases where non-B subtype infection is a possibility.

[Screening of HIV in blood banks. Evaluation of fourth generation kits]

Use of detection tests for p24 HIV antigen (p24Ag) in blood banks in Argentina is recommended by the Argentinean Society of Hemotherapy and Immunohematology. In the blood bank of the National University of Cordoba (Argentina), the recent implementation of the p24Ag screening test has considerably increased the cost of the battery of screening tests and its use in all blood donations has not produced the benefits expected. A 4th generation EIA was evaluated for the screening of HIV in comparison with the currently used assays in the blood bank of National University of Cordoba (3rd generation EIA + p24Ag assay). For this comparison, 11 serum samples from subjects with early HIV infection (early seroconversion period) were tested, as well as 27 serum samples from asymptomatic HIV-infected subjects and other 39 from non-HIV infected subjects. The 3rd generation EIA and the 4th generation EIA showed the same sensitivity value (100%) but the specificity of the 3rd generation EIA was higher (97.5%) comparing with 4th generation (95.1%). Besides, the p24Ag test failed to detect 2 samples from subjects with early HIV infection. These results indicate a good performance of both 3rd and 4th generation assays for screening of HIV. However, due to the lowest cost of 4th generation EIA kit, it could replace the currently used assays for HIV screening in regional blood banks. This screening assay will lead to gain in effectiveness and reduced costs until the detection of HIV RNA can be implemented in blood banks.

[Assay for simultaneous detection of HIV p24 antigen and anti-HIV antibody]

OBJECTIVE

To develop a rapid assay for simultaneous detection of HIV p24 antigen (Ag) and anti-HIV antibody (Ab).

METHODS

HIV-1 gp41 antigen and HIV-2 gp36 antigen were expressed by recombinant baculovirus insect system and purified by immunochromatography. p24 monoclonal antibody (mAb) was obtained from p24 hybridoma cell line. Purified antigen and mAb were dot blotted to nitrocellular membrane; 20 nm colloidal gold-anti-human IgG ab and p24 ab complex were used for this test. Previously detected 39 sera specimens were tested in this study to compare with the result of HIV test with commercial HIV test kit.

RESULTS

20 mg/L purified gp41 Ag and gp36 Ag were obtained from recombinant baculovirus-insect cell system; 1.5 mg/L p24 mAb was obtained from p24 mAb hybridoma cell line. Compared the test result of 39 sera with commercial HIV test kits, consistency rate was 100%.

CONCLUSIONS

The rapid assay for simultaneous detection of HIV p24 antigen and anti-HIV antibody provides a simple, sensitive and reliable test for HIV diagnosis.

HIV production from purified monocytes isolated from antiretroviral-naïve and protease inhibitor-treated HIV-1-infected patients

PURPOSE

We obtained peripheral blood monocytes from HIV-1-infected patients and differentiated them into macrophages in vitro to investigate the contribution of purified monocytes to plasma viremia in HIV-infected patients.

METHOD

We compared the production of HIV-1 from purified monocytes isolated from patients who were receiving antiviral therapy to those who were not. We obtained monocytes from 29 antiretroviral-naïve (ARVN) patients and 35 protease inhibitor (PI)-treated HIV-infected patients (20 with undetectable plasma HIV). Cells were cultured with phorbol myristate acetate to induce HIV replication.

RESULTS

HIV p24 core antigen was present in supernatants of 9/29 (31%) of ARVN patients (mean = 102 pg/mL). In contrast to ARVN patients, p24 was not found in supernatants of any PI-treated patient (p =.01). Thus, purified monocytes from some ARVN patients produce HIV, but we were unable to detect HIV production from monocytes from PI-treated patients, regardless of their virologic response.

CONCLUSION

Monocytes may contribute to plasma viremia in ARVN patients but may not do so in PI-treated patients.

Lack of antitumor activity and intolerance of interleukin-4 in patients with advanced HIV disease and Kaposi's sarcoma

Kaposi's sarcoma (KS), the most common malignancy associated with HIV infection, is caused by the Kaposi sarcoma herpesvirus (KSHV). Exacerbations of KSHV are associated with increased human interleukin-6 (HuIL-6), and elevated IL-6 could be related to the development of KS. IL-4, a cytokine with pleiotropic effects, suppresses IL-6 in vivo and modestly inhibits AIDS-KS-derived cells in vitro. Suppression of IL-6 by exogenous IL-4 could result in antitumor activity. We report the results of a clinical trial to test this hypothesis. A phase I/II dose escalation safety, tolerance, and efficacy trial was conducted in patients with biopsy-proven AIDS-related KS, at two university medical centers. Patients were scheduled to receive IL-4 (0.5, 1.5, 3.0, or 4.0 microg/kg/day) administered subcutaneously (s.c.) in sequential cohorts. Patients were continued on study as long as the drug was tolerated or the disease progressed. Patients were followed for antitumor activity, effects on viral replication, immune status, and clinical and laboratory toxicity. Seventeen patients were enrolled at two sites over a 21-month period. There were 15 males and 2 females, and 1 patient was Hispanic. All patients had a Karnofsky score >70. Patients enrolled only into the two lower dose cohorts (0.5 and 1.5 microg/kg/day). Both groups had similar baseline characteristics. The median time on treatment was only 7.4 and 8.4 weeks for the 0.5 and 1.5 microg/kg/day dose levels, respectively. There was significant neutropenia, with 6 patients having grade 3 or greater toxicity requiring granulocyte colony-stimulating factor (G-CSF). Three patients on a dose of 1.5 microg/kg/day stopped treatment due to protocol-defined toxicity. There were no appreciable effects on CD4/CD8 counts. HIV viral RNA did not significantly change over time. However, in several people, it appeared to decline with treatment and rebound with discontinuation of treatment. Corresponding changes were noted in the HIV immunocomplex dissociated (ICD) p24 antigen. One patient had a partial response, 11 patients had stable disease, and 5 patients had disease progression during the short period of treatment. The maximum tolerated dose for IL-4 in patients with advanced AIDS-related KS is 1.5 microg/kg/day. At this dose level, IL-4 is poorly tolerated and is not an effective KS treatment. Treatment of the majority of patients is discontinued because of drug-related toxicity or because of disease progression. Future studies of IL-4 should be confined to studies of cytokine manipulation of the underlying HIV infection, as there appears to be little antitumor activity.