aBrighton Health Care NHS Trust, Royal Sussex County Hospital, Brighton, East Sussex BN2 5BE, UK; and bDepartment of Primary Care and Population Sciences, Royal Free and University College Medical School, London NW3 2PF, UK.
Received: 14 December 2001; revised: 14 June 2002; accepted: 3 July 2002.
This study highlights risk factors for CD4 cell losses and clinical consequences of antiretroviral treatment interruptions. The largest decreases in CD4 cell counts occur in patients with lower nadir CD4 cell counts or more advanced disease. Falls are greater in those with higher rates of CD4 cell loss before highly active antiretroviral therapy, and are associated with a significant increase in clinical/AIDS events during treatment interruption. Caution is therefore advised when considering treatment interruption in those with advanced disease.
Although the possible beneficial effects of treatment interruptions remain unproved, many patients stop therapy for a variety of reasons [1,2]. However, the information available to guide patients and clinicians with regard to the risks of CD4 cell loss or disease progression when taking a treatment interruption is limited [3–6]. In particular, few studies have assessed the clinical sequelae of treatment interruptions, and to date there are limited data suggesting clinical progression when therapy is stopped [2,7].
To study the possible risk factors for CD4 cell loss during treatment interruption and to assess the clinical sequelae, we identified patients attending the HIV clinic in Brighton who had taken a treatment interruption of at least 2 months. All patients had to have received highly active antiretroviral therapy (HAART) with three or more drugs for 3 or more months before taking a treatment interruption. In addition, patients had to have at least two CD4 cell counts off therapy during treatment interruption. Data were collected on demographic, treatment, CD4 cell count and viral load history, and clinical events, both on and off treatment.
Linear regression methods were used to assess the predictors of CD4 cell change. Univariate regression models considered the effects of the following variables on outcome: CD4 cell count; percentage and viral load at the start of treatment interruption; nadir CD4 cell count before treatment interruption; rate of CD4 cell loss before HAART (available for 28 patients); age; number of drugs and drug classes exposed to at the start of treatment interruption; time since first starting antiretroviral treatment; clinical stage at baseline; and whether the patient took a treatment interruption because of adverse events. Those factors significant in univariate models were included in a multivariable regression analysis. All analyses were performed using PROC REG in the statistical software package, SAS. To assess the risk of clinical progression during the treatment interruption, the event rate was calculated as the number of events divided by the person-years of follow-up, after starting HAART but before treatment interruption, and during treatment interruption itself. Differences in these rates were tested for significance using Poisson regression.
Forty patients met the criteria for inclusion. The median age of patients was 38 years (range 24–62). Thirty-nine were men (predominantly homosexual) and 33 were born in the UK (82.5%). The median nadir CD4 cell count was 182 cells/mm3 (range 6–523), with a median previous yearly rate of CD4 cell change of −59 cells/mm3 (range −294, 100). At the start of treatment interruption patients had been on treatment for a median of 2.6 years (range 0.5–10.5) and on HAART for a median of 1.5 years (range 0.5–4.0). Patients had been exposed to a median of six different antiretroviral agents (3–11), with 17 (42.5%) having been exposed to all three drug classes. At the time of treatment interruption, the majority of patients were receiving three drugs (36; 90%). Eighteen patients (45%) had a viral load of less than 500 copies/ml and 13 (32.5%) had a history of an AIDS-defining illness at baseline. The median length of treatment interruption was 214 days (61–1036).
There was a drop in the median CD4 cell count from 348 to 216 cells/mm3 during the treatment interruption, with a median individual patient drop of 88 cells/mm3 (drop of 443 to increase of 133) (P = 0.0001). CD4 cell percentages decreased from 17.5 to 12.5%, with a median patient drop of 5% (drop of 17% to increase of 4%, P = 0.0001). The viral load increased from 2.81 to 5.17 log10 copies/ml, with individual changes ranging from a drop of 1.22 log10 copies/ml to an increase of 4.08 log10 copies/ml (median increase of 2.38 log10 copies/ml, P = 0.0001). The individual plots of CD4 cell changes (Fig. 1) did not suggest any changes in the rate of CD4 cell loss over treatment interruption, but these results should be interpreted cautiously because of the possibility of information censoring as those with more rapid CD4 cell losses restart therapy. In a multivariable analysis, those with higher baseline CD4 cell counts (P = 0.02), but lower CD4 cell nadir (P = 0.02) lost more CD4 cells. In addition, the CD4 cell loss was greater in those with a higher rate of CD4 cell loss before starting HAART (P = 0.02), even after adjusting for the baseline and nadir CD4 cell count.
Patients were followed for a total of 41.7 person-years over treatment interruption. During this time, 31 patients (77.5%) experienced at least one clinical event, with 65 events occurring in total, of which six were AIDS-defining. The median time to the first clinical event (Kaplan–Meier) was 121 days, with 15, 30, 43 and 64% of individuals developing a clinical event by 1, 2, 3 and 6 months after the start of the treatment interruption, respectively. This gave incidence rates of 1.56 per year for clinical events and 0.14 per year for AIDS events. Between starting antiretroviral therapy and the start of the treatment interruption, patients had experienced 106 clinical events, of which five were AIDS-defining, over a total follow-up period of 135.29 years. Incidence rates before the treatment interruption were 0.78 per year for clinical events and 0.04 per year for AIDS events. These rates were significantly lower (P = 0.0001 for all clinical events, P = 0.03 for AIDS events) than those seen during the treatment interruption.
Event rates in the different CD4 cell strata are shown in Table 1. At CD4 cell counts of less than 50 cells/mm3, the event rate is actually higher while on treatment than during the treatment interruption, although this is not statistically significant. This probably reflects a selection bias, whereby of those with very low CD4 cell counts, only those without clinical events may risk stopping treatment. At higher CD4 cell counts, the event rate is higher in those taking a treatment interruption, and this is statistically significant for CD4 cell counts greater than 200 cells/mm3.
In line with findings from other studies [4,5], our results suggest that those at most risk of a large CD4 cell loss are those who currently have high CD4 cell counts but previously had much lower nadir values before starting treatment. These results may suggest some irreversible loss of immune function in those starting treatment at very low CD4 cell counts, and may thus have important implications for patients who choose to delay therapy to a late stage, or those who present late for treatment.
We have found an increase in the rates of clinical and AIDS events during treatment interruption. As clinical/AIDS events are far more pertinent to patients than CD4 cell loss, these results are important. Our results suggest that treatment interruptions are independently associated with an increased risk of clinical events regardless of the CD4 cell loss, as the difference in risk is apparent in most CD4 cell strata. A finer stratification of CD4 cell counts is desirable, but with the sample size in this study, we were unable to do this. Larger studies that consider this issue are needed.
We have demonstrated that immunological changes seen en en in individuals during treatment interruption do indeed translate into clinical deterioration. The CD4 cell loss is greatest in those individuals with a history of advanced clinical disease or low nadir CD4 cell counts. We would urge caution in such individuals if treatment interruption is considered, and close monitoring during this time is essential to ensure that the risk of disease progression is minimized.
1.Mocroft A, Youle M, Moore A, Sabin CA, Madge S, Cozzi Lepri A, et al
. Reasons for modification and discontinuation of antiretrovirals: results from a single treatment centre. AIDS
2.Taffe P, Richenbach M, Hirschel B, Opravil M, Furrer H, Janin P, et al
. Impact of occasional short interruptions of HAART on the progression of HIV infection: results from a cohort study. AIDS
3.Garcia F, Plana M, Ortiz GM, Bonhoeffer S, Soriano A, Vidal C, et al
. The virological and immunological consequences of structured treatment interruptions in chronic HIV-1 infection. AIDS
4.Sabin C, Phillips A, Fusco J, Youle M, Gill J, Barbour J, et al
. The effect of treatment interruption in patients with virologic failure: results from a multi-cohort collaborative study.
In: 8th Conference on Retroviruses and Opportunistic Infections
. Chicago, 4–8 February 2001 [Poster no. 365].
5.Youle M, Janossy G, Turnbull W, Tilling R, Loveday C, Mocroft A, et al
. Changes in CD4 lymphocyte counts after interruption of therapy in patients with viral failure on protease inhibitor-containing regimens. AIDS
6.Miller V, Sabin C, Hertogs K, Bloor S, Martinez-Picado J, D'Aquila R, et al
. Virological and immunological effects of treatment interruptions in HIV-1 infected patients with treatment failure. AIDS
7.Deeks SG, Wrin T, Liegler T, Hoh R, Hayden M, Barbour JD, et al
. Virological and immunological consequences of discontinuing combination antiretroviral-drug therapy in HIV-infected patients with detectable viraemia. N Engl J Med