| |
Viral Load Predicts CD4 Decline to <350, Takes 2.5 Yrs
|
| |
| |
HIV Diagnosis at CD4 Count Above 500 Cells/mm3 and Progression to Below 350 Cells/mm3 Without Antiretroviral Therapy
The UK Collaborative HIV Cohort (CHIC) Study Steering Committee
JAIDS Journal of Acquired Immune Deficiency Syndromes:Volume 46(3)1 November 2007pp 275-278
"For untreated patients with CD4 count in the 350 to 500 cells/mm3 range, the rate of AIDS or death is approximately 2.5 per 100 person-years,15 which translates into a predicted 6.1% cumulative risk over an average 2.5 years spent with CD4 count in this range while deferring ART. As expected,16 viral load at baseline was a strong predictor of the length of time taken for the CD4 count to reach <350 cells/mm3, with the median ranging from 0.7 years in those with viral load ≥500,000 copies/mL to 4.7 years in those with viral load <1000 copies/mL."
Abstract
Summary: A trial to evaluate the risks and benefits of initiation of antiretroviral therapy (ART) in patients with high CD4 count (eg, ≥500 cells/mm3), in comparison with deferral (eg, to <350 cells/mm3), merits consideration. Two issues for consideration in designing such a trial are the proportion of patients seen in clinics who present with high CD4 count and the time it will take for those randomized to deferring ART to reach a level where ART must be initiated. Among 13,572 patients in the UK CHIC Study presenting since 1996, 3631 (27%) had a count ≥500 cells/mm3. Among 4268 ART-naive patients with at least one CD4 count in the 500 to 650 cells/mm3 range, the median time to <350 cells/mm3 (or start of ART) was 2.5 years, with a range of 2.1 to 3.1 years depending on the analysis approach. Viral load at baseline was a strong predictor of the time taken for the CD4 count to reach <350 cells/mm3, with the median ranging from 0.7 years in those with viral load ≥500,000 copies/mL to 4.7 years in those with <1000 copies/mL. This provides timely background data on ART-naive patients seen in clinical practice to support design of a trial to compare immediate with deferred ART in people with high CD4 count.
There has been a growing appreciation that antiretroviral therapy (ART) drugs have sustained and highly durable efficacy in clinical practice,1,2 that accumulation of drug resistance to the 3 main drug classes is slow,3 and that new drugs are becoming available that allow at least 3 lines of non-cross-resistant regimens to be constructed.4 Furthermore, results from the Strategies for Management of Anti-Retroviral Therapy (SMART) trial have found that those randomized to using episodic CD4-count guided ART (and hence interrupting or staying off ART at baseline) experienced a greater incidence of liver, renal, and cardiovascular disease and non-AIDS mortality than those randomized to continuous ART, suggesting that the net effect of ART may be to reduce incidence of these conditions.5 Furthermore, cohort studies have found that the rate of death from many of these events increases with lower CD4 count.6-9 This has led to the view that ART should currently be initiated at a CD4 count of 350 cells/mm3 and that efforts should be made to assess the risks and benefits of starting ART in people with much higher CD4 count (eg, CD4 ≥500 cells/mm3), compared with deferring to a level around 300 to 350 cells/mm3 in a randomized trial.10 Two issues for such a trial are (i) the proportion of patients who are likely to be eligible for the trial, and subsequently for ART initiation at high CD4 count if it is found to be beneficial, and (ii) the time it will take for those randomized to deferring ART to reach <350 cells/mm3, which indicates the length of additional ART use with early initiation, and also allows estimates of the AIDS/death risk in those deferring ART during the period in which the CD4 count is declining to 350 cells/mm3. To help with feasibility assessment and design of such a potential trial, we have studied these issues in the UK Collaborative HIV Cohort (CHIC) Study, a large observational clinic-based cohort study.
PATIENTS AND METHODS
The UK CHIC Study has been described previously.11 In brief, the study currently contains clinical information collected as part of routine care on all patients seen since January 1, 1996, in 10 clinics in the UK. This includes information on all CD4 counts and viral load measures, incidence of AIDS diseases, and reports of death. So far, a total of 25,274 patients have contributed to the UK CHIC Study.
To understand what proportion of newly presenting ART-naive patients might have a CD4 count sufficiently high to be eligible for a trial of early ART, we assessed the distribution of CD4 count at presentation (ie, the first measured CD4 count for ART-naive patients entering the clinic) for all patients presenting within the period 1996 to 2005, inclusive. Generalized linear models with log link and Poisson error (using generalized estimating equations [GEE] with an unstructured covariance) were used to assess unadjusted (1 univariable model for each covariate) and adjusted (multivariable model) relative risks for the association between various covariates and the probability of the CD4 count at presentation being ≥500 cells/mm3.12
For the main analysis of the time to fall from above 500 cells/mm3 to below 350 cells/mm3 in ART-naive patients, we included ART-naive patients who experienced at least 1 CD4 count in the 500 to 650 cells/mm3 range after January 1, 1996 (only patients alive after this date are included), and followed from the time of the first such count until the first CD4 count <350 cells/mm3 or start of ART, whichever occurred first. The upper limit of 650 cells/mm3 was chosen because it was felt that patients entering a trial would be likely to be in this range rather than at a higher count. We also considered variations of the analytical approach. All analyses were performed using SAS version 9.1 (Statistical Analysis Software, Cary, NC).
RESULTS
CD4 Counts in Newly Presenting and Current ART-Naive Patients
Of a total of 13,572 patients presenting in the period 1996 to 2005, 7285 (54%) were men who acquired HIV through sex with another man (MSM), 422 (3%) acquired HIV through injection drug use (IDU), 5084 (37%) acquired HIV through heterosexual sex, and 781 (6%) were other or unknown; 3583 (26%) were female; 7111 (52%) were of white ethnicity, 3778 (28%) were of black African/black non-Caribbean ethnicity, and the remaining 2683 (20%) were of other or unknown ethnicity. The median age at presentation was 34 (interquartile range [IQR]: 29 to 39). At presentation, 3631 (27%) had a CD4 count ≥500 cells/mm3, 1852 (51%) of which fell in the 500 to 650 cells/mm3 range. The median viral load (measured on the same day or within 6 months before the CD4 count, n = 9639) was 10,400 copies/mL (IQR: 1700 to 46,000), with a breakdown as follows: <1000 copies/mL, n = 959 (10%); 1000 to 9999, n = 1830 (19%); 10,000 to 49,999, n = 2481 (26%); 50,000 to 99,999, n = 1271 (13%); and ≥100,000, n = 3098 (32%). Factors associated with presenting at CD4 ≥500 cells/mm3 were assessed in a multivariable model (Table 1). Older age, being a heterosexual male, and being of black African ethnic origin, and presentation before 2000 were all strongly independently associated with a lower chance of presenting at a CD4 count ≥500 cells/mm3.
In addition to the prevalence of CD4 count ≥500 cells/mm3 at presentation, we also considered the prevalence of CD4 count ≥500 cells/mm3 among naive patients under follow-up on July 1, 2005. Nine hundred fifty (38%) of all 2510 ART-naive patients under follow-up had a CD4 count ≥500 cells/mm3 (of which 512 [54%] were in the 500 to 650/mm3 range; median viral load 10,000 copies/mL, IQR 1600 to 36,000). A further 769 (31%) had a CD4 count in the 350 to 500 cells/mm3 range and 791 (32%) <350 cells/mm3.
To illustrate the sensitivity of this estimate to differences in definition, we considered the effect of the following changes in the approach to the analysis on the median time to CD4 count <350 cells/mm3 or start of ART: (i) restricting eligibility for analysis to those never previously measured below 500 cells/mm3 (median, 2.7 years; IQR: 1.1 to 5.3); (ii) changing the definition of the endpoint so that we required 2 consecutive values <350 cells/mm3 instead of 1 only (median, 3.1 years; IQR: 1.3 to 6.0); (iii) expanding eligibility to all those with a CD4 count ≥500 cells/mm3 while ART-naive, instead of <650 cells/mm3, and defining the baseline as the date of the first such count (median, 2.9 years; IQR: 1.2 to 6.0); (iv) basing the analysis on those with first CD4 count at presentation of 500 to 650 cells/mm3 (median, 2.1 years; IQR: 0.7 to 4.6); and (v) censoring at the start of ART, instead of considering this as part of the endpoint (median, 3.0 years; IQR: 1.5 to 5.8). Finally, by defining the baseline as January 1, 1998, and including all ART-naive patients with CD4 count 500 to 650 cells/mm3 (n = 292), the median time was 3.0 years (IQR: 1.6 to 5.7).
We also assessed factors associated with the time to <350 cells/mm3 or start of ART in Cox models. Lower CD4 count at baseline (adjusted relative hazard [RH] per 50 cells/mm3 higher: 0.90; 95% confidence interval [CI]: 0.84 to 0.96; P = 0.001), lower CD4 count nadir (RH per 100 cells/mm3 higher: 0.86; 95% CI: 0.80 to 0.92; P < 0.0001), and higher baseline viral load (RHs of 1.71 [95% CI: 1.39 to 2.10], 2.29 [95% CI: 1.87 to 2.81], 2.97 [95% CI: 2.36 to 3.73], 4.62 [95% CI: 3.70 to 5.77], and 7.98 [95% CI: 5.92 to 10.77] for viral load 1000 to 9999, 10,000 to 49,999, 50,000 to 99,000, 100,000 to 499,999, and >500,000 cps/mL, respectively, relative to <1000; P < 0.0001 in each case; Figure 1B) were associated with a raised rate of reaching the endpoint, as was older age (RH: 1.07 per 10 years older; 95% CI: 1.00 to 1.13; P = 0.04). Females were also at higher risk. However, this appeared to be attributable to the greater tendency to start ART early on account of pregnancy, because this gender effect was not present when we censored at early ART initiation rather than considering it as part of the endpoint. There was no significant effect of race/ethnicity or risk group.
DISCUSSION
Our results indicate that 27% of newly presenting patients in large UK clinics have a CD4 count ≥500 cells/mm3. The proportion was highest in MSM and was also higher in females infected heterosexually and all IDUs, compared with males infected heterosexually. It was also higher in whites and those of other ethnicity, compared with those of black African origin. Furthermore, of ART-naive patients under care in July 2005, 38% had a CD4 count ≥500 cells/mm3. These proportions suggest that a trial of immediate (vs. deferral until reaching a CD4 count of around 300 to 350 cells/mm3) therapy in patients with CD4 count ≥500 cells/mm3 should be relevant to a substantial proportion of individuals presenting with HIV. As testing for HIV becomes more mainstream and widespread, people should be diagnosed earlier in their infection, and these proportions would be expected to increase somewhat. Indeed, we identified a greater chance of the CD4 count at presentation being ≥500 cells/mm3 in those diagnosed after 2000 than before.
Our estimates suggest that the median time it takes for the CD4 count in a person with a level in the 500 to 650 cells/mm3 range to fall to below 350 cells/mm3 or to start ART is 2.5 years, with a range of 2.1 to 3.1 years depending on the analysis approach. There was, however, a wide variability among patients-our main analysis suggests 10% with times of >8 years. We are not aware of previous estimates directly comparable with this. The absolute rate of CD4 count decline has been found to be greater at higher counts13 and, consistent with this, has been generally modeled as linear on a square-root scale.14 Our results suggest that initiating ART at CD4 counts above 500 cells/mm3 would increase the time that patients would potentially spend on ART by approximately 2.5 to 3 years on average, compared with waiting to start ART until the count has fallen below 350 cells/mm3. For untreated patients with CD4 count in the 350 to 500 cells/mm3 range, the rate of AIDS or death is approximately 2.5 per 100 person-years,15 which translates into a predicted 6.1% cumulative risk over an average 2.5 years spent with CD4 count in this range while deferring ART. As expected,16 viral load at baseline was a strong predictor of the length of time taken for the CD4 count to reach <350 cells/mm3, with the median ranging from 0.7 years in those with viral load ≥500,000 copies/mL to 4.7 years in those with viral load <1000 copies/mL.
In summary, we have provided background data on ART-naive patients with CD4 count ≥500 cells/mm3 seen in clinical practice to support the design of a trial to compare immediate ART with deferral to a CD4 count around 350 cells/mm3.
|
|
| |
| |
|
|
|
