|
Rate of AIDS diseases or death in HIV-infected antiretroviral therapy-naive individuals with high CD4 cell count and high viral load: >650 CD4s and viral load <10,000 better
|
|
|
AIDS:Volume 21(13)20 August 2007p 1717-1721
The UK Collaborative HIV Cohort (CHIC) Study Steering Committee
From the Department of Primary Care & Population Sciences, Royal Free & University College Medical School, London.
"....Our results tell us that the decreasing trend in rate of AIDS or death with higher CD4 cell count in ART-naive people is present even among those with CD4 cell count > 500 cells/μl; those with CD4 cell count 500-650 cells/μl carry a 55% raised risk of AIDS or death compared with those with CD4 cell count ≥ 650 cells/μl (P = 0.01)..... Our findings suggest that risk of AIDS and death might be reduced by using ART to raise CD4 cell counts even among patients with high CD4 cell counts. This information contributes to the rationale for evaluating the risk/benefits of initiation of ART in patients with high CD4 cell count (e.g., > 500 cells/μl) in a randomized trial, for example comparing with deferral to 350 cells/μl....Lower CD4 cell count was associated with a raised rate of AIDS/death, as were higher viral load, older age and injection drug use as transmission group, after adjustment for the CD4 cell count. Findings for the outcome of severe AIDS or death were similar, with a 0.79-fold lower rate of severe AIDS or death per 100 cells/_l higher CD4 cell count within the CD4 cell count groups ≥ 350 cells..."
Abstract
Objective: To assess the absolute rate of AIDS and death in antiretroviral therapy (ART)-naive patients with a high CD4 cell count. Such information would be helpful in the design of a trial investigating early initiation of ART.
Design: Analysis of data from an ongoing HIV cohort study.
Methods: The rate of (severe) AIDS or death and death alone was evaluated in ART-naive patients according to the current CD4 cell count, focusing on CD4 cell counts ≥ 350 cells/μl among patients in the UK CHIC Study.
Results: In a total of 30 313 person-years of follow-up, there were 1557 AIDS or death events. The rate of AIDS or death in persons with most recent CD4 cell count 350-499, 500-649 and > 650 cells/μl was 2.49, 1.54 and 0.96 per 100 person-years, respectively. The rate ratio for those with CD4 cell count 500-649 cells/μl compared with those with CD4 cell count ≥ 650 cells/μl was 1.55 [95% confidence interval (CI), 1.11-2.17; P = 0.01]. In a Poisson regression model based on person years with CD4 cell count ≥ 350 cells/μl, there was a strong effect of CD4 cell count on rate of AIDS or death (rate ratio, 0.84; 95% CI, 0.76-0.93; P = 0.001), independent of viral load and age.
Conclusions: The trend of decreasing rate of AIDS and death with higher CD4 cell count is present throughout the CD4 cell count ≥ 350 cells/μl range in ART-naive people.
Discussion
The risk of AIDS and death in persons with CD4 cell counts > 350 cells/_l is low relative to those with lower CD4 cell counts, particularly those with counts < 50 cells/_l (at least 30-fold lower, according to our results; Table 1). Yet the risk at higher counts is not negligible and, as ART efficacy has improved (both in terms of specific drugs and the number of non-cross-resistant regimens available) and associated toxicities have become lower in incidence and severity and/or better managed, it is appropriate to focus on risk at these higher levels. This is particularly the case for ART-naive patients with high CD4 cell counts, for whom consideration is being given to early ART initiation. Our results tell us that the decreasing trend in rate of AIDS or death with higher CD4 cell count in ART-naive people is present even among those with CD4 cell count > 500 cells/_l; those with CD4 cell count 500-650 cells/_l carry a 55% raised risk of AIDS or death compared with those with CD4 cell count ≥ 650 cells/_l (P = 0.01). Many studies have presented rates of AIDS or AIDS/death according to the current CD4 cell count (e.g., [6-13]), some focusing on risks in the high CD4 cell count range [8], but none to our knowledge has established a significant trend in rate in the CD4 cell count range ≥ 500 cells/_l among ART-naive subjects.
Our findings suggest that risk of AIDS and death might be reduced by using ART to raise CD4 cell counts even among patients with high CD4 cell counts. This information contributes to the rationale for evaluating the risk/benefits of initiation of ART in patients with high CD4 cell count (e.g., > 500 cells/_l) in a randomized trial, for example comparing with deferral to 350 cells/_l. Another key endpoint to consider in such a trial is that of serious nonfatal cardiovascular, renal or liver events. A limitation of our analysis is that the incidence of such events cannot be evaluated in our study as such information was not recorded. Other cohort studies have found that rate of death from these events increases with lower CD4 cell count [14-17]. The SMART trial of CD4 cell count-guided interruption of ART versus continuous ART found that the incidence of a composite endpoint consisting of such events was significantly higher in the CD4 cell count-guided interruption arm, when it had been hypothesized that it would be lower, owing to less ART use [18]. It is not clear from these overall results from SMART, however, whether this raised incidence in those using less ART was a consequence of the absence of ART per se or having stopped ART. Only a large randomized trial carried out for a significant duration will be able to reliably inform us of the net risks or benefits - in terms of AIDS diseases, cardiovascular, renal or liver events and deaths - of early versus deferred initiation of ART. Like SMART, such a trial is likely to require a substantial international collaborative effort.
There are further limitations to this analysis. We do not currently have sufficiently reliable information on cause of death to detail these causes. Further, the amount of person time of observation, while substantial, is not sufficient for precise estimates of rates of AIDS and death by viral load at very high CD4 cell counts, so collaborative joint analyses may be useful to improve estimates. Another limitation is that unlike in a randomized trial the AIDS events occurring in this study are not subject to review by an endpoint committee and so the certainty and standardization of diagnosis is not as reliable as would be the case in a trial.
In summary, the trend of decreasing rate of AIDS and death with higher CD4 cell count is present across the CD4 cell count range ≥ 350 cells/_l and may even persist across CD4 cell counts ≥ 500 cells/_l. These are important observations for considerations of a randomized trial of early ART.
Introduction
The increasing success of current antiretroviral therapy (ART) for HIV infection in suppressing HIV replication and preventing HIV-related disease [1-3], along with the promise of further new drugs becoming available (e.g., Cooper et al. [4]), have led to the hope that many people with HIV may potentially be able to live something approaching a normal lifespan. With this new outlook, a reevaluation of acceptable levels of clinical risk is appropriate. Levels of risk that were previously considered low and not justifying (additional) intervention should perhaps now prompt consideration of more aggressive management. For example, it could now be appropriate to immediate initiate ART in ART-naive patients with a high CD4 cell count. The net benefit of early versus delayed ART remains to be proven, however, and should be tested in a randomized trial [5]. This might involve, for example, randomizing ART-naive patients with high CD4 cell count (e.g., > 500 cells/μl) to immediate initiation of ART or to deferral until reaching a level more consistent with current guidelines (e.g., in the range 300-350 cells/μl) [5]. A fuller understanding of the absolute risks at different high CD4 cell count levels would provide useful background. This paper presents analyses from the large observational UK CHIC Study with the aim of providing such information.
Methods
Patients
The UK Collaborative HIV Cohort (CHIC) Study contains clinical information collected as part of routine care on all patients seen since 1 January 1996 in 10 clinics in the UK. This includes information on all CD4 cell counts and viral load measures, incidence of AIDS diseases and reports of death. In addition, the national death registration data on all UK deaths occurring in people aged < 55 years is checked against the database to maximize ascertainment.
Calculation of rates of AIDS or death
For ART-naive patients, person time was counted from the time of each qualifying CD4 cell count until the first of the following: start of ART, next CD4 cell count, date of first AIDS disease, death, 6 months after the date of the CD4 cell count. Qualifying CD4 cell counts were those made when there had been no prior treatment use and no previous AIDS disease.
The joint effects of CD4 cell count, viral load and age on the rate of AIDS and death was evaluated in a Poisson regression model. For this analysis, there was an additional requirement for a CD4 cell count to qualify, namely that a viral load value was available within the past 6 months. Follow-up from the CD4 cell count was until the first of the following: next CD4 cell count, date of first AIDS disease since January 1996, death, 6 months after the date of the CD4 cell count, 6 months after the date of the current viral load, the date of starting ART. Since patients could contribute person time related to several of their CD4 cell count measures and some patients contributed person time relating to many more CD4 cell counts than others, the population is most appropriately described by the distribution of demographic variables at the level of CD4 cell counts, rather than at the level of individual patients.
All P values are two sided and analyses were performed using SAS version 9.1 (Statistical Analysis Software, version 8.2, Cary, North Carolina, USA).
Results
So far, a total of 25 274 patients have contributed to the UK CHIC Study. Of these, 17 609 contributed a total of 30 313 person-years to the analysis of rates of AIDS or death in ART-naive patients. The median age across CD4 cell counts was 34 [interquartile range (IQR), 30-40], 16% of CD4 cell counts were in females, 72% were in men who have sex with men and 20% in heterosexuals.
Table 1 gives the rates of AIDS and death according to the current CD4 cell count. The median viral load (copies/ml) at the time of the CD4 cell count defining each period of follow-up was 7000 (3.9 log), 16 000 (4.2 log), 23 000 (4.4 log), 37 000 (4.6 log), 84 000 (4.9 log) and 195 000 (5.3 log) for CD4 cell count categories ≥ 650, 500-649, 350-499, 200-349, 50-349 and 0-49 cells/μl, respectively. A total of 1557 AIDS or death events occurred (overall rate 5.1/100 person-years). The decreasing risk of AIDS or death with higher CD4 cell count extended through the CD4 cell count range > 500 cells/μl: comparisons of the rates in 500-649 and ≥ 650 cells/μl categories gave a rate ratio of 1.55 [95% confidence interval (CI), 1.11-2.17; P = 0.01]. For the CD4 cell count categories ≥ 350 cells/μl, rates of AIDS or death were also examined according to the current viral load. The rates of AIDS or death during person time with viral load of < 10 000, 10 000-99 999 and ≥ 100 000 copies/ml, respectively, for the three categories of CD4 cell counts were as follows: 1.5/100 person-years (28 AIDS or death events/1878.6 person-years), 2.1/100 person-years (57 events/2740.8 person-years) and 3.1/100 person-years (26 events/826.7 person-years) for the 350-499 cells/μl group; 0.9 person-years (14 events/1511.5 person-years), 0.8/100 person-years (12 events/1594.0 person-years) and 2.7/100 person-years (10 events/376.8 person-years) for the 500-649 cells/μl category; and 0.5/100 person-years (10 events/1942.1 person-years), 1.1/100 person-years (13 events/1159.1 person-years) and 0.9/100 person-years (2 events/228.5 person-years) for the ≥ 650 cells/μl category.
![person-1.gif](../images/081407/person-1.gif)
The first AIDS events occurring at CD4 cell count > 350 cells/_l were also examined: 63 (20%) were Kaposi's sarcoma (compared with 16% overall), 62 (20%) oesophageal candidiasis (17% overall), 42 (14%) tuberculosis (13% overall), 35 (11%) herpes simplex (6% overall), 37 (12%) recurrent bacterial infections (6% overall), 20 (6%) Pneumocystis jiroveci pneumonia (19% overall), 17 (5%) cryptosporidiosis (3% overall) and 13 (4%) lymphoma (3% overall).
Also shown in Table 1 are rates of severe AIDS/death (severe AIDS is defined as an AIDS disease other than recurrent bacterial pneumonia, recurrent herpes simplex, tuberculosis and oesophageal candidiasis) and rates of death alone. CD4 cell count trends are similar, extending into the high CD4 cell count range.
A Poisson regression model was fitted to evaluate the independent contributions of CD4 cell count, viral load, age, transmission group and gender to the rate of AIDS (Table 2), focusing on the higher CD4 cell count range (CD4 cell count ≥ 350 cells/_l). Lower CD4 cell count was associated with a raised rate of AIDS/death, as were higher viral load, older age and injection drug use as transmission group, after adjustment for the CD4 cell count. Findings for the outcome of severe AIDS or death were similar, with a 0.79-fold lower rate of severe AIDS or death per 100 cells/_l higher CD4 cell count within the CD4 cell count groups ≥ 350 cells/_l (95% CI, 0.69-0.91; P < 0.001), while for death the equivalent figure was 0.82 (95% CI, 0.66-1.01; P = 0.06).
![CD4-2.gif](../images/081407/CD4-2.gif)
|
|
|
|
|
|
|