Serum P24 Antigen Levels Research Articles

HIV-1 Tat is present in the serum of people living with HIV-1 despite viral suppression

ObjectivesDespite successful human immunodeficiency virus (HIV) control with combination antiretroviral therapy (cART), individuals with HIV still face health risks, including cancers, cardiovascular and neurocognitive diseases. An HIV protein, Tat, is potentially involved in these HIV-related diseases. Previous studies demonstrated circulating Tat in the blood of untreated people with HIV. Here, we measured Tat levels in the serum of cART-treated people with HIV to examine the effect of cART on Tat production. MethodsSerum samples from 63 HIV-positive and 20 HIV-seronegative individuals were analyzed using an ELISA assay that detected Tat concentrations above 2.5 ng/mL. ResultsAmong HIV-positive individuals, the Tat level ranged from 0 to 14 ng/mL. 25.4% (16 out of 63) exceeded the 2.5 ng/mL cut-off, with a median HIV Tat level of 4.518 [3.329-8.120] ng/mL. No correlation was revealed between Tat levels and CD4+ T cell counts, serum HIV RNA, p24 antigen, or anti-Tat levels. ConclusionsDespite cART, circulating HIV Tat persists and may contribute to HIV-related diseases. This emphasizes the need for further research on the mechanisms of Tat action in non-infected cells where it can penetrate upon circulation in the blood.

T activation marker evaluation in ARC patients treated with AZT. Comparison with CD4+ lymphocyte count in non-progressors and progressors towards AIDS

Reductions in the percentage and absolute number of CD4+ lymphocytes, as well as abnormally high levels of activated peripheral T lymphocytes (CD3+ HLA-DR+ phenotype) and an increased proportion of CD8+ cells coexpressing the CD57 surface antigen (involved in natural killer activity) have been reported in HIV infection and associated with disease progression. We prospectively measured these subsets of lymphocytes in 34 patients with advanced AIDS-related complex (ARC) treated with azidothymidine (AZT). Peripheral blood lymphocyte phenotyping was performed before treatment, then at weeks 12 and 24. A striking fall in the proportion of activated T lymphocytes from baseline was observed (P less than 0.001) at week 24. In contrast, the percentage of CD4+ cells showed a slight and transient rise at week 12 (P less than 0.05). No modification in levels of CD8+ or CD8+ CD57+ cells was detected during the study. Of the 34 patients, 11 developed AIDS, and 23 remained AIDS-free during 51 weeks of follow-up. Similar patterns of change in CD4+ and HLA-DR+ CD3+ lymphocytes were found in the AIDS progressors and nonprogressors. Likewise, HIV p24 antigenaemia showed parallel decreases in both groups of patients. Although changes in CD4+ cells, p24 antigenaemia and HLA-DR-reactive T lymphocytes were not predictive of clinical outcome, large differences existed between the two groups prior to the initiation of therapy. The short-term onset of AIDS was associated with lower CD4+ cell numbers, higher levels of serum p24 antigen and a greater proportion of activated T lymphocytes. Our results suggest that the possible interest of T lymphocyte activation markers, in conjunction with conventional phenotyping, should be investigated further.

HIV viral load monitoring.

Markers of HIV infection are used in the clinical setting to predict prognosis in HIV-infected persons, measure disease progression and monitor responses to therapy. In the past, the most commonly used tests were CD4 cell counts and serum p24 antigen levels, although serum beta-2 microglobulin, serum neopterin, serum IgA levels and CD8 cell numbers were used as adjunctive measures by some clinicians. Since the early to mid 1990s new technologies for the detection of HIV RNA in human plasma have allowed relatively accurate and precise measurement of plasma HIV RNA concentrations, which directly relate to the production of HIV-1 in tissues, particularly the lymphoid tissues. The available assays to quantify plasma HIV RNA, in combination with CD4 counts, enable clinicians to monitor responses to antiretroviral therapy and to predict the risk for disease progression.1,2,3,4

Comparison of hydroxychloroquine with zidovudine in asymptomatic patients infected with human immunodeficiency virus type 1

Hydroxychloroquine (HCQ), an antimalarial agent used to treat patients with autoimmune diseases, has been shown to suppress human immunodeficiency virus type 1 (HIV-1) replication in T cells and monocytes in vitro by inhibiting posttranscriptional modification of the virus. An initial randomized, placebo-controlled clinical trial conducted in 38 asymptomatic HIV-1—infected patients who had CD4+ counts between 200 and 500 cells/mm 3 demonstrated that the amount of recoverable virus declined significantly in the HCQ group compared with the placebo group over the 8-week study period. These preliminary observations were expanded into a second 16-week clinical trial comparing the efficacy of HCQ with that of zidovudine (ZDV) in 72 asymptomatic HIV-1—infected patients with CD4+ counts between 200 and 500 cells/mm 3. Patients were randomly assigned to receive either HCQ 800 mg/d (n = 35) or ZDV 500 mg/d (n = 37) for 16 weeks. No adverse reactions to the study medications were observed in either the HCQ or ZDV group. Patients in both groups had reduced levels of recoverable HIV-1 RNA in the plasma, reduced levels of cultured virus, and reduced levels of serum p24 antigen after the 16-week study period. However, no difference was noted in absolute CD4+ counts between the two groups. Interleukin-6 and serum immunoglobulin G levels were significantly reduced in the HCQ group but not in the ZDV group. These findings support the results of the previous clinical trial. Thus HCQ may be potentially useful in the treatment of patients with HIV-1 infection

Comparison of virus burden in blood and sequential lymph node biopsy specimens from children infected with human immunodeficiency virus

BACKGROUND: Lymph nodes serve as reservoirs for the replication of human immunodeficiency virus (HIV) type 1. Comparison of serial measurements of virus burden in lymph nodes and peripheral blood after a change in antiretroviral therapy may provide insights into pathogenic mechanisms and permit a more accurate assessment of a therapeutic response. STUDY DESIGN: Nevirapine was added to the drug regiment of eight children with HIV infection treated with the combination of zidovudine and didanosine who had increasing levels of serum p24 antigen. Lymph node biopsies were performed at entry and after 12 weeks of therapy. RESULTS: Neither CD4 counts nor p24 antigen level correlated with the degree of viremia as measured by ribonucleic acid copy numbers in plasma. Correlations were found between HIV DNA copy number in peripheral blood mononuclear cells and HIV DNA copy number in lymph nodes ( p = 0.02), as well as between peripheral blood CD4 counts and lymph node architecture. The HIV signals in the lymph nodes conformed to the anatomic organization of apical light zones in the germinal centers; however, in more advanced disease stages, organized germinal centers disappeared as evidenced by a decline in the extent of the follicular dendritic network. CONCLUSIONS: Lymph node biopsies in this small number of HIV-infected children revealed a progressive loss of an organized architecture, especially of the follicular dendritic network. This correlated with a progressive loss of CD4 + cells but not with other measures of disease stage, including viral load, as measured by ribonucleic acid copy numbers. (J P EDIATR 1996;129:410-8)

Relation between stage of disease and neurobehavioral measures in children with symptomatic HIV disease.

To study the relationships between stage of HIV disease, reflected by CD4+ lymphocyte percentages and p24 antigen levels, and HIV-associated central nervous system (CNS) abnormalities, measured by computed tomography (CT) brain-scan ratings and neurobehavioral tests. Consecutive case series. Government medical research center. Eighty-six previously untreated children with symptomatic HIV-1 disease. CD4% measures correlated significantly with overall CT brain-scan severity ratings (r = -0.45; P < 0.001) as well as with its component parts (cortical atrophy, white matter abnormalities, and intracerebral calcifications); they were of comparable magnitude for vertically and transfusion-infected children. CD4% measures were also associated with the general level of cognitive function (r = 0.32; P < 0.005). Furthermore, patients with detectable serum p24 antigen levels (n = 39) had CT brain scans that were more abnormal than patients with undetectable p24 levels (n = 20; CT abnormality ratings of 21.3 versus 35.9; P < 0.02); similar differences were found for the cortical atrophy and calcification ratings. p24 levels also correlated with the overall CT brain-scan severity rating (r = 0.34; P < 0.01). Degree of CT brain-scan abnormality and level of cognitive dysfunction were significantly associated with the stage of HIV-1 disease, as reflected by either CD4 leukocyte measures or elevations of p24 antigen. The relation between the CT brain-scan lesions and markers of HIV disease (both CD4 and p24) suggest that these CNS abnormalities are most likely associated with HIV-1 infection, and further support the hypothesis that the interaction between systemic disease progression and CNS manifestations is continuous rather than discrete.

Antiretroviral activity of stavudine (2′,3′-didehydro-3′-deoxythymidine, D4T)

Stavudine, 2′,3′-didehydro-3′-deoxythymidine (D4T), is a potent inhibitor of HIV-1 reverse transcriptase in vitro. In clinical studies, stavudine has excellent oral bioavailability in excess of 80%. The dose-limiting toxicity is peripheral neuropathy, which occurred in 15% of stavudine versus 6% of zidovudine-treated patients for 80 weeks in a randomized, blinded, phase III trial. Stavudine-treated groups have experienced significant increases in mean CD4 cell counts and decreases in both mean serum p24 antigen levels and infectious HIV titers in peripheral blood mononuclear cells. In subjects with prior zidovudine treatment, the duration of these responses is limited; CD4 counts and serum p24 antigen levels return to baseline after approximately 6 months. The effect of stavudine on clinical outcome and survival has not yet been established in comparative trials. Stavudine offers an additional therapeutic option to those individuals who are refractory to or intolerant of other available antiretrovirals.

Distinct Changes in HIV Type 1 RNA versus p24 Antigen Levels in Serum during Short-Term Zidovudine Therapy in Asymptomatic Individuals with and without Progression to AIDS

Serum HIV-1 RNA and p24 antigen levels were examined in 28 seropositive asymptomatic individuals participating in a trial on the efficacy of zidovudine. Sixteen individuals remained asymptomatic until 4 years after the onset of the trial, whereas 12 individuals were diagnosed with an AIDS-defining event. The serum HIV-1 RNA load and p24 antigen levels were determined before the onset of therapy and during the first 8 weeks of therapy to establish whether the patterns of change were predictive of clinical outcome. Among the 28 participants 43% had measurable pretreatment concentrations of p24 antigen. Initiation of zidovudine therapy was followed by a similar decline of p24 antigen levels in nonprogressors as well as progressors and, therefore, these groups could not be distinguished on the basis of this parameter. HIV-1 RNA was detected in the pretreatment samples of 82% of the individuals and could be detected in p24 antigen-positive as well as p24 antigen-negative individuals. Similar changes in HIV-1 RNA load during zidovudine therapy were observed in p24 antigen-positive and -negative individuals. Analysis of the HIV-1 RNA response according to clinical outcome demonstrated that HIV-1 RNA copy numbers had declined significantly after 4 weeks of therapy in both nonprogressors and progressors, but the decline in RNA load was much stronger in the nonprogressors. Our data show that the HIV-1 RNA load in serum can be used to monitor the response to antiviral therapy in p24 antigen-positive as well as -negative individuals. Posttreatment changes in p24 antigen levels are not indicative for clinical outcome, whereas RNA copy numbers are.(ABSTRACT TRUNCATED AT 250 WORDS)

Sustained Increases In Cd4 Cell Counts In Asymptomatic Human

Forty-four asymptomatic patients infected with human immunodeficiency virus type 1 (HIV-1), who had 200-799 CD4 cells/microL, received oral prednisolone (0.5 mg/kg for 6 months; 0.3 mg/kg thereafter). After 1 year of treatment, no major side effect or AIDS events had occurred. The percentage of DR+ and CD25+ phenotypes in CD4 T cells decreased significantly as did levels of serum IgG, IgA, and beta 2-microglobulin. Serum p24 antigen and HIV RNA levels remained stable. CD4 cell counts increased significantly at all time points (median increase at 1 year, 119 cells/microL). Peripheral blood mononuclear cell apoptosis after overnight stimulation with anti-CD3 monoclonal antibodies was strongly inhibited at all times. In asymptomatic seropositive patients, immunotherapy for 1 year with glucocorticoids was safe and led to sustained increases in CD4 cell counts and to improvement or stabilization of other biologic markers of disease activity.

Stavudine In Patients With Aids And Aids-Related Complex: Aids Clinical Trials Group 089

In a phase I trial of stavudine in AIDS or AIDS-related complex (ARC), antiviral effects and safety were assessed in 41 patients treated with dosages of 0.5-12.0 mg/kg/day. Among evaluable patients, 10% increases in CD4 lymphocyte counts were sustained in 24 (60%) of 40 during treatment; an NAUC response (normalized area under the CD4 cell count-versus-time curve > 1.0) was observed in 31 (91%) of 34 at 10 weeks and in 20 (80%) of 25 at 24 weeks; 15 (83%) of 18 had decreases in p24 antigenemia; and 24 (60%) of 40 gained > or = 2.5 kg body weight. Median CD4 lymphocyte levels remained above baseline for 6 months in patients receiving > 0.5 mg/kg/day. Median serum p24 antigen levels remained below baseline for > or = 1 year in patients with p24 antigen responses. The principal toxicity was peripheral neuropathy, which generally resolved after drug discontinuation but limited the dosage to < or = 2.0 mg/kg/day. Additional trials assessing the effect of stavudine on overall morbidity and mortality are ongoing.

Alternating nevirapine and zidovudine treatment of human immunodeficiency virus type 1-infected persons does not prolong nevirapine activity.

The potential use of an alternating treatment strategy with nevirapine and zidovudine in prolonging the antiretroviral effects of nevirapine was evaluated. Ten human immunodeficiency virus type 1 (HIV-1)-infected p24 antigen-positive persons who had not received prior antiretroviral therapy were treated for 9-13 weeks with an alternating regimen of 1 week of nevirapine (200 mg/day) and 3 weeks of zidovudine (600 mg/day). Serum p24 antigen levels declined during the first week of nevirapine treatment (median, 59%); however, subsequent courses of nevirapine were characterized by rising p24 antigen levels, while antigen levels remained stable or declined during zidovudine treatment. Serum beta 2-microglobulin levels and CD4+ cell counts exhibited similar responses. HIV-1 isolates obtained from 2 patients revealed 40- and 1000-fold reductions in nevirapine sensitivity after 8 weeks. These findings demonstrate that alternating treatment with zidovudine and nevirapine does not prolong the effectiveness of nevirapine and does not prevent the development of nevirapine resistance.

Auditory P300 abnormalities and leukocyte activation in HIV infection.

To evaluate whether P300 testing might serve as a screening modality for the early detection of HIV-related neuropathology, we tested 26 HIV-infected men (23 without neurologic symptoms, 2 with peripheral neuropathy, 1 with AIDS-associated dementia) and 15 controls. Although they had no overt neurologic symptoms, the P300 latency was delayed or undetectable in 30% of patients without clinically evident neurologic disease. P300 latencies did not correlate with peripheral blood CD4 T-cell count, serum quinolinic acid or p24 antigen levels, or the numbers of activated peripheral blood monocytes. Three individuals with abnormal P300 latencies had been HIV-seropositive for < or = 1 year, suggesting that delayed evoked responses detect early neurologic dysfunction. P300 responses do not predict imminent dementia. In only one previously asymptomatic individual with abnormal P300 waveforms have overt neurologic symptoms developed during a 2-year followup. Extended longitudinal studies will be necessary to define the predictive value of P300 latencies in the development of AIDS-related dementia. However, the sensitivity, quantitative nature, and speed of administration of this test suggest that it may be useful for identification of early neurologic involvement in HIV infection.

An Open-Label Dose-Escalation Trial of Oral Dehydroepiandrosterone Tolerance and Pharmacokinetics in Patients with HIV Disease

Dehydroepiandrosterone (DHEA) is a naturally occurring adrenal steroid reported to have immunomodulatory and antiviral activity in cellular and animal models as well as modest in vitro antiretroviral activity against human immunodeficiency virus (HIV). A phase I dose-escalation study was performed to evaluate the safety and pharmacokinetics of DHEA in subjects with symptomatic HIV disease and an absolute CD4 lymphocyte count between 250 and 600 cells/microliters. Thirty-one subjects were evaluated and monitored for safety and tolerance. The oral drug was administered three times daily in doses ranging from 750 mg/day to 2,250 mg/day for 16 weeks. Some immunological and virological parameters were monitored as well. The drug was well tolerated and no dose-limiting side effects were noted. Dose proportionality was evidenced neither by the serum DHEA nor by DHEA-S time-concentration curves for the three dosing groups. However, the study cohort appeared to consist of two subpopulations with markedly different bioavailability for a given DHEA dose. No sustained improvements in CD4 counts nor decreases in serum p24 antigen or beta-2 microglobulin levels were observed. However, serum neopterin levels decreased transiently by 23-40% at week 8 compared with baseline in all dosing groups. DHEA was well tolerated by patients with mild symptomatic HIV disease; evaluation of this agent for efficacy in HIV disease would require randomized, controlled trials.

Serum levels of soluble CD8, neopterin, β2-microglobulin and p24 antigen as indicators of disease progression in children with AIDS on zidovudine therapy

To test the hypothesis that serum levels of soluble markers in children change after initiation of zidovudine therapy and that the extent and pattern of these longitudinal changes correlates with clinical outcome. We measured serum levels of soluble CD8, neopterin, beta 2-microglobulin (beta 2M), and p24 antigen, and CD4 cell counts, before the initiation of zidovudine therapy and at 12, 24 and 48 weeks of treatment in 24 HIV-1-infected children (Centers for Disease Control classification P2) and 15 controls. Soluble CD8 levels were elevated before therapy in 70% of the infected children; subsequent decreases were associated with lower risk of disease progression. The mean serum neopterin level before treatment was elevated in infected children; decreases in neopterin levels marginally reflected improved or stable clinical status. Serum beta 2M levels and CD4+ cell counts were not associated with clinical outcome. Only 10 out of the 24 patients had detectable levels of serum p24 antigen before treatment; again, the amount of decline after initiation of therapy did not predict clinical outcome. Decreasing levels of soluble CD8 and neopterin in HIV-1-infected children receiving zidovudine therapy might reflect a good response to treatment and a slowing of disease progression.

Alternating and Intermittent Regimens of Zidovudine and Dideoxycytidine in Patients with AIDS or AIDS-Related Complex

To determine whether alternating regimens consisting of zidovudine and 2',3'-dideoxycytidine (ddC) reduce the toxicity and maintain or increase the antiretroviral effect associated with each drug alone. An unblinded, randomized (phase II) clinical trial in which seven treatment regimens were compared. Outpatient clinics of 12 AIDS Clinical Trials Units. One hundred thirty-one patients with the acquired immunodeficiency syndrome (AIDS) or AIDS-related complex and serum p24 antigenemia (> or = 70 pg/mL). Treatments included weekly or monthly alternating zidovudine (200 mg every 4 hours) and ddC (0.01 or 0.03 mg/kg body weight every 4 hours); weekly intermittent zidovudine, 200 mg every 4 hours, or ddC, 0.03 mg/kg every 4 hours; and continuous zidovudine. Toxicity, CD4 cell counts, serum p24 antigen levels, and clinical end points. Data were analyzed for the first 48 weeks of therapy (median follow-up, 40 weeks). Hematologic toxicity was significantly less frequent in patients who received zidovudine therapy every other week (11% to 15%) or every other month (11% to 14%) than in those who received continuous zidovudine therapy (33%) (P < 0.02). Weekly alternating therapy with zidovudine and ddC, 0.03 mg/kg, or intermittent therapy with ddC, 0.03 mg/kg, produced high rates of peripheral neuropathy (41% and 50%, respectively). Neuropathy occurred in 10% to 21% of patients in the other three alternating-therapy limbs and in 17% of patients receiving zidovudine alone (intermittently or continuously). Initial increases in CD4 cell counts were sustained in three alternating-therapy limbs, but counts returned to baseline by week 28 in the remaining limbs. The median weight gain at week 48 was significantly greater in patients treated with alternating regimens (0.9 to 3.8 kg) compared with those treated with continuous zidovudine therapy (-0.7 kg) (P = 0.008). Patients treated with alternating regimens and those treated with continuous zidovudine had similarly sustained decreases in p24 antigen levels. These findings suggest that alternating therapy with zidovudine and ddC reduces the toxicity associated with each drug alone while maintaining strong antiretroviral activity.

Immunodominant epitopes of HIV-1 p17 and p24.

Immunodominant antibody-binding sites were mapped using overlapping synthetic peptides of the structural proteins p17 and p24 of human immunodeficiency virus type 1 (HIV-1). Using sera from HIV-1-infected individuals at a variety of disease states, three major epitopes were identified within p17 and one within p24. Antibodies which recognized these epitopes were present in all risk groups throughout all stages of HIV infection, regardless of the presence of high levels of serum p24 antigen.

Phase I study of continuous-infusion soluble CD4 as a single agent and in combination with oral dideoxyinosine therapy in children with symptomatic human immunodeficiency virus infection

To determine the safety and pharmacokinetics of recombinant soluble CD4 (sCD4) administered by continuous intravenous infusion to children with symptomatic human immunodeficiency virus type 1 infection, we conducted a phase I study at the National Cancer Institute. Three dose levels of sCD4 were evaluated: 100, 300, and 1000 micrograms/kg per day. After an initial 12 weeks of treatment with sCD4 alone, dideoxyinosine at a dose of 90 mg/m2 every 8 hours was added and subjects were observed for an additional 12 weeks. Combination therapy was continued in patients in whom it was well tolerated. In addition to toxicity and pharmacokinetic monitoring, surrogate markers of antiviral activity were evaluated. Eleven children were enrolled in the study. During the 12 weeks of treatment with sCD4 alone, and during subsequent sCD4 plus dideoxyinosine combination therapy, no significant toxic reaction attributable to sCD4 or dideoxyinosine was encountered. Low-level anti-CD4 antibodies developed in two patients. Steady-state sCD4 levels increased proportionately at higher doses. The CD4 cell counts and serum p24 antigen levels did not provide evidence of antiviral activity. We conclude that sCD4 was well tolerated at doses up to 1000 micrograms/kg per day when administered by continuous intravenous infusion; however, evidence of in vivo antiviral activity was not observed in this study.

Isolation of HIV-1 from Plasma of Infected Individuals

Experimental conditions affecting the successful propagation of HIV-1 from the plasma of seropositive individuals were examined. It was determined that whole blood samples collected with lithium heparin as the anticoagulant, immediate plasma separation, and immediate culturing were best suited for obtaining viable virus from plasma. Virus was isolated by infecting fresh phytohemagglutinin-stimulated normal donor peripheral blood mononuclear cells (PBMCs) with plasma followed by weekly cocultivation with new target cells. The plasma virus isolation rate was the greatest and HIV-1 titers were the highest for those individuals with less than 200 CD4+ cells/mm3 and decreased as the level of CD4+ cells approached normal values. We were able to obtain positive cultures from 29.5% of those patients with CD4+ counts greater than 500/mm3. HIV-1 titers in plasma also correlated with high serum p24 antigen levels when serum was treated with glycine to dissociate antigen-antibody complexes.

P24 antibody production in p24 antibody-negative HIV-infected subjects.

To determine if HIV-infected patients with no detectable serum antibodies to p24 are producing antibodies to p24 (anti-p24), blood was obtained from 49 HIV-infected patients at various stages of infection. Serum p24 antigen levels were measured and peripheral blood mononuclear cells were cultured for 1 week without mitogenic stimulation. The presence of anti-p24 in culture supernatants and sera was determined by radioimmunoprecipitation assays. Cells from 89% of the patients who had anti-p24 in their sera spontaneously synthesized anti-p24 in vitro. Similarly, cells from 83% of the HIV-infected patients who had no detectable anti-p24 in their sera spontaneously produced anti-p24 in vitro. Thus the absence of anti-p24 in serum did not reflect suppression in the ability of patients' cells to synthesize and secrete antibodies to p24. However, cells from patients whose sera contained anti-p24 spontaneously synthesized more anti-p24 than did cells from patients whose sera lacked anti-p24, suggesting that these two groups of patients may represent individuals with inherently high or low responses to p24 epitopes, respectively.

HIV-1 sensitivity to zidovudine and clinical outcome in children

In adults with the acquired immunodeficiency syndrome, long-term monotherapy with zidovudine selects for human immunodeficiency virus type 1 (HIV-1) strains with substantially reduced in-vitro susceptibility to the drug. We have assessed the relation between in-vitro resistance to zidovudine and clinical outcome in children, in whom disease progression is more rapid than in adults. We studied 23 children with symptoms of HIV-1 disease during extended monotherapy with zidovudine. An in-vitro assay was used to determine the concentration of zidovudine required to inhibit by 50% the replication of viral isolates (IC 50) obtained after 9 to 39 months of treatment. Viral stocks of high enough titre to yield reproducible results were obtained from 19 of the children. During the following 6 months of therapy, 9 children were stable, 7 deteriorated, and 3 died. There was a highly significant relation between decreased zidovudine susceptibility and poor clinical outcome (p<0·001) but no relation between IC 50 and age at start of therapy or length of time on treatment. Age-adjusted CD4 lymphocyte counts were lower at the start of treatment (p=0·02) and at the time of sampling (p=0·01) in children whose viral isolates had an increased zidovudine IC 50. Initial serum p24 antigen levels were not predictive of subsequent emergence of resistant virus, but at the time of sampling for viral sensitivity higher p24 antigen levels were associated with raised IC 50 (p=0·004). The findings suggest that most children who become unresponsive to monotherapy with zidovudine, as judged by clinical criteria, will have changes in in-vitro sensitivity to the drug. In these children, an alternative antiretroviral therapy should be considered.