| |
Cardio- and cerebrovascular events in HIV-infected persons
|
| |
| |
AIDS: Volume 18(13) 3 September 2004
The Writing Committee From the D:A:D: Study Group
"...the results of this study further support the hypothesis that CART is associated with increased risk of atherosclerosis. Further follow-up of the D:A:D cohort is necessary to assess whether the risk of CCVE will continue to increase with the duration of antiretroviral therapy and to evaluate further the magnitude of this increase... the D:A:D study does not currently have sufficient power to address the issue of whether protease inhibitors and non-nucleoside reverse transcriptase inhibitors are associated with the same risk of CCVE. These analyses are planned for the future.
207 patients experienced at least one CCVE (cerebrovascular event), with 268 events occurring in total. There were 134 MI among 126 patients (first event in 126 patients), 87 invasive cardiovascular procedures in 82 patients (first event in 39 patients), 41 strokes in 39 patients (first event in 38 patients), and six patients died from CCVE (first event in four patients). The fatality rate following the first event was 23.7% (49 of 207 cases). Of the 126 MI that were first events, 69 (54.8%) were categorized as definite coronary events, 37 (29.4%) as possible and 20 (15.9%, 17 of which were fatal) were unclassifiable. Of the 38 strokes that were first events, seven (18.4%) were hemorrhagic, 29 (76.3%) infarctions, and the underlying etiology was unknown in two cases (5.3%).
The incidence of first CCVE was 5.7 per 1000 person-years and increased with longer exposure to CART. At a conference I attended Sam Bozette expressed his doubt of the original findings of the DAD Study.
From the multivariable Poisson regression analyses,
-older age (RR per 5 years older, 1.42; P = 0.0001),
-smoking (RR, 1.66; P = 0.008),
-previous history of CCVE (RR, 7.12; P = 0.0001),
-family history of CCVE (RR, 1.62; P = 0.03) and
-male gender (RR, 1.82; P = 0.02) were independently associated with CCVE.
After controlling for these variables, as well as cohort, body mass index, HIV transmission group and ethnicity (none of which were significantly associated with CCVE in multivariable analyses), the incidence rate of CCVE increased by 26% per additional year of exposure to CART (RR per year of exposure, 1.26; 95% CI, 1.14-1.38; P < 0.0001). The results were comparable when including all 268 end-points in the analysis and controlling for multiple endpoints in the same individual.
Additional analyses tested the association between CCVE and a number of possible metabolic and physiological causative factors. Factors independently associated with the risk of CCVE from these analyses were (time-updated):
-cholesterol (RR, 1.11 per mmol higher; P = 0.008);
-triglycerides (RR, 1.30 per log2 mmol higher; P = 0.0006);
-diabetes mellitus (RR, 2.22; P = 0.0002); and
-hypertension (RR, 1.79; P = 0.001).
Adjustment for these factors in multivariable models that also included the risk factors for CCVE, led to a small reduction in the RR associated with exposure to CART, although this variable remained significantly associated with CCVE in all analyses (relative rates ranging from 1.21 to 1.25 in the individual analyses).
You should read the author discussion below on the limitations of interpreting the study findings.
Introduction
Recent results from the D:A:D (Data collection on Adverse events of anti-HIV Drugs) Study, a prospective observational cohort of 23 468 HIV-positive patients, indicated that the incidence of myocardial infarction (MI) increased by an average of 26% per year of exposure to CART, over the first 6 years of exposure.
In the primary analyses of the D:A:D study, a well-defined single endpoint, MI, was selected as the primary endpoint to facilitate reliable and reproducible event validation by means of established diagnostic criteria. The choice of this endpoint also allows comparisons with other studies that assess the risk of cardiovascular disease in HIV-infected or uninfected persons, and is particularly useful when evaluating possible mechanisms of disease. However, atherosclerotic disease is not confined to the coronary arteries and it is of interest to see whether the reported relationship between increased exposure to CART and MI is similar when considering additional cardio- and cerebro- vascular events (CCVE). As the D:A:D study also collected data on other CCVE, we developed a composite CCVE endpoint that included MI, stroke, invasive cardiovascular procedures and deaths from other cardiovascular causes.
The components of this endpoint were chosen as they can be objectively assessed and classified. In the background population, the risk factors for CCVE are largely the same as those for coronary artery disease. Thus, an additional aim was to consider the impact of known risk factors for MI on this composite endpoint in our cohort of HIV-positive individuals.
SUMMARY/Abstract
Recent results from the D:A:D Study indicated that the incidence of myocardial infarction (MI) increased by 26% per year of exposure to combination antiretroviral treatment (CART). The present study was performed to investigate whether this risk was similar when including other cardio- and cerebro-vascular disease events (CCVE).
D:A:D is an international collaboration of 11 cohorts, following 23 468 HIV-infected patients prospectively at 188 clinics in 21 countries situated in Europe, USA and Australia.
The end-point was the occurrence of a first CCVE during prospective follow-up, defined as the first of: acute MI, invasive cardiovascular procedures, stroke, or death from other cardiovascular disease. Relative rates (RR) for CCVE from Poisson regression models and 95% confidence intervals (CI) are reported. All models are adjusted for other risk factors for CCVE, including age, gender, ethnicity, family history, body mass index, and smoking status as well as cohort and HIV transmission group.
Over 36,145 person-years of follow-up, 207 patients experienced at least one CCVE (23.7% fatal). The first event was MI in 126 patients, invasive cardiovascular procedure in 39 patients, stroke in 38 patients, and death from other cardiovascular disease in four patients.
The incidence of first CCVE was 5.7 per 1000 person-years [95% confidence interval (CI) 5.0-6.5] and increased with longer exposure to CART (RR per year of exposure, 1.26; 95% CI, 1.14-1.38; P < 0.0001).
Authors conclude that CART increases the risk of CCVD, and this increase is comparable with how CART affects the risk of MI. This finding is consistent with the hypothesis that atherosclerosis is a side-effect of CART.
AUTHOR DISCUSSION
The results of this analysis of the D:A:D study indicate that, as with the primary MI endpoint, there is an increasing risk of CCVE with longer exposure to CART over and above that which can be explained by increasing age. The magnitude of this risk is comparable with that found when exclusively focusing on MI as endpoint. Other risk factors traditionally associated with CCVE in the HIV-uninfected population are confirmed in this cohort of HIV-infected individuals: age, male gender, smoking, previous CCVE and family history of CCVE. Moreover, other risk factors that have been associated with use of CART, such as increased serum triglycerides and cholesterol, diabetes mellitus and hypertension are also associated with increased risk of CCVE.
In order for patients and clinicians to make the most informed choices regarding antiretroviral therapy, it is important to be aware of the risks of all cardio- and cerebro-vascular events that may be associated with CART. A composite endpoint, such as the one used in this study, is arguably the most useful when developing prediction models for gauging an HIV-infected individual's risk of cardiovascular disease. Such information may also allow patients to modify aspects of their lifestyle (e.g. smoking and dietary habits) that are shown to be associated with increased risk of CCVE. However, caution should be taken when interpreting the results from studies of composite endpoints as these may be limited by biases related to differential associations between exposures and the individual outcomes. In our analyses of the composite endpoint (of which MI constituted 62%), the risk factors were very similar to those observed for MI alone. However, in the populations without HIV infection these risk factors have been reported to have a differential impact on specific single endpoints, with a more pronounced association between hypertension and the risk of stroke, and a stronger association between cholesterol levels and the risk of MI than with stroke. Given the low incidence of end-points other than MI at the current time in the D:A:D study, we are unable to confirm these differential findings in our study population.
With regards to stroke, our findings illustrate a number of problems of ascertainment, which may occur when considering this end-point in HIV-infected individuals. Unfortunately, the coding system applied for diagnostic classification of strokes in the population without HIV infection cannot be applied directly to cohorts of HIV-infected persons. Despite the decrease in incidence of clinical AIDS-defining central nervous system events since the introduction of CART, the prevalence of conditions such as AIDS-dementia complex/HIV encephalitis remains relatively high, and these conditions are associated with an increased risk of stroke-like episodes. Using MONICA clinical criteria for stroke and supporting imaging examinations and paraclinical information, only episodes that were considered to be of atherogenic/vascular origin - and not secondary to HIV central nervous system morbidity - were included in our analyses. One possible limitation of our study is that we do not systematically collect information on cocaine use, and thus cannot rule out that a fraction of the strokes may partially be attributable to cocaine usage. However, cocaine use was only recorded in a small number of the detailed stroke event reports. Furthermore it seems unlikely that cocaine use should be more prevalent in patients with longer CART exposure, and therefore we do not think that this is likely to introduce substantial bias.
In conclusion, the results of this study further support the hypothesis that CART is associated with increased risk of atherosclerosis. Further follow-up of the D:A:D cohort is necessary to assess whether the risk of CCVE will continue to increase with the duration of antiretroviral therapy and to evaluate further the magnitude of this increase. Moreover, as the number of less frequent CCVE events increases with longer follow-up, this will allow for more detailed assessment of risk of the individual events. Finally, although several studies have reported a differential effect of the drug classes on the development of metabolic complications, including our own, the D:A:D study does not currently have sufficient power to address the issue of whether protease inhibitors and non-nucleoside reverse transcriptase inhibitors are associated with the same risk of CCVE. These analyses are planned for the future, however, once enough events have occurred to ensure that this comparison is sufficiently powered .
Study Methods
D:A:D is an international collaboration of 11 cohorts, following patients at 188 clinics in 21 countries situated in Europe, USA and Australia. Details of the methods have previously been reported. Briefly, patients were included in the study irrespective of whether or for how long they had been receiving antiretroviral therapy. Demographic and HIV disease characteristics were collected on all participants and prospective data were collected at regular intervals. All collected information was transformed into a standardized format and merged into a central data set. The analyses reported in this paper are based on data collected up to July 2002.
All incident cases of MI, stroke, invasive cardiovascular procedures and deaths were reported to the study co-ordinating office for central validation and coding as detailed previously. For reported strokes, detailed information was collected including symptoms, duration of symptoms, central nervous system imaging findings by computerized tomography or magnetic resonance scanning, spinal fluid examination where available, and information of underlying disease possibly related to the episode (e.g. hypertension, atrial fibrillation, cocaine use, concomitant central nervous system disease). Reported strokes were classified as: (i) definite, (ii) stroke-like event, or (iii) not stroke, and were further classified by the type of lesion as infarction, hemorrhage, or unknown. MI and stroke events were also distinguished as non-fatal (survival to 28 days from onset) or fatal (all other events). Only definite events were included in the composite endpoint. The classification of MI and stroke was conducted according to criteria applied in the WHO MONICA study, and was performed with expert consultation independently of knowledge of a patient's antiretroviral treatment history.
Cerebrovascular lesions that were secondary to other concomitant central nervous system disease, or which did not fulfil the criteria for the stroke definition, were excluded. These included stroke-like events in patients with cerebral opportunistic infection, HIV dementia, brain lymphoma or non-HIV central nervous system infection (ii); and transient ischemic attacks, cerebral venous thrombosis, or hemorrhages secondary to trauma (iii).
Deaths from other CCVE included cases of death from end-stage ischemic heart disease other than MI.
The end-point of this analysis was the occurrence of a first CCVE during prospective follow-up, defined as the first of: acute MI, invasive cardiovascular procedures (specifically coronary artery angioplasty, bypass, or carotid artery endarterectomy), stroke, or death from other CCVE.
The statistical analyses have been described in detail elsewhere. Briefly, all analyses were based on an incidence rate approach. Only events that occurred prospectively following enrolment in D:A:D were included in the analyses. Follow-up was censored on the earliest date of the following: a first episode of CCVE, death from non-CCVE causes (if applicable), 1 February 2002, or 6 months after the individual's last contact for HIV care (defined on the basis of the most recent CD4 count/HIV-RNA level, initiation of new antiretroviral regimen or new AIDS-defining diagnosis). Relative rates (RR) for CCVE from Poisson regression models and 95% confidence intervals (CI) are reported. All models are adjusted for other risk factors for CCVE, including age, gender, ethnicity, family history, body mass index, and smoking status as well as cohort and HIV transmission group.
Results
Over 36,145 person-years of follow-up, 207 patients experienced at least one CCVE, with 268 events occurring in total. There were 134 MI among 126 patients (first event in 126 patients), 87 invasive cardiovascular procedures in 82 patients (first event in 39 patients), 41 strokes in 39 patients (first event in 38 patients), and six patients died from CCVE (first event in four patients). The fatality rate following the first event was 23.7% (49 of 207 cases; 95% CI, 17.9- 29.5%). Of the 126 MI that were first events, 69 (54.8%) were categorized as definite coronary events, 37 (29.4%) as possible and 20 (15.9%, 17 of which were fatal) were unclassifiable. Of the 38 strokes that were first events, seven (18.4%) were hemorrhagic, 29 (76.3%) infarctions, and the underlying etiology was unknown in two cases (5.3%).
Twenty additional reported cases of stroke were excluded from the endpoint, eight of which were subsequently classified as stroke-like events, and 12 as 'not stroke'.
The incidence of first CCVE was 5.7 per 1000 person-years (95% CI, 5.0-6.5) and increased with longer exposure to CART.
From the multivariable Poisson regression analyses, older age (RR per 5 years older, 1.42; 95% CI, 1.32-1.52; P = 0.0001), smoking (RR, 1.66; 95% CI, 1.14-2.42; P = 0.008), previous history of CCVE (RR, 7.12; 95% CI, 4.91-10.33; P = 0.0001), family history of CCVE (RR, 1.62; 95% CI, 1.05-2.50; P = 0.03) and male gender (RR, 1.82; 95% CI, 1.10-3.00; P = 0.02) were independently associated with CCVE. After controlling for these variables, as well as cohort, body mass index, HIV transmission group and ethnicity (none of which were significantly associated with CCVE in multivariable analyses), the incidence rate of CCVE increased by 26% per additional year of exposure to CART (RR per year of exposure, 1.26; 95% CI, 1.14-1.38; P < 0.0001). The results were comparable when including all 268 end-points in the analysis and controlling for multiple endpoints in the same individual.
Additional analyses tested the association between CCVE and a number of possible metabolic and physiological causative factors. Factors independently associated with the risk of CCVE from these analyses were (time-updated): cholesterol (RR, 1.11; 95% CI, 1.03-1.19 per mmol higher; P = 0.008); triglycerides (RR, 1.30; 95% CI, 1.12-1.51 per log2 mmol higher; P = 0.0006); diabetes mellitus (RR, 2.22; 95% CI, 1.46-3.37; P = 0.0002); and hypertension (RR, 1.79; 95% CI, 1.25-2.56; P = 0.001). Adjustment for these factors in multivariable models that also included the risk factors for CCVE, led to a small reduction in the RR associated with exposure to CART, although this variable remained significantly associated with CCVE in all analyses (relative rates ranging from 1.21 to 1.25 in the individual analyses).
|
|
| |
| |
|
|
|
