| |
Longitudinal evaluation of cardiovascular disease-associated biomarkers in relation to abacavir therapy
|
| |
| |
[Research Letters]
AIDS:Volume 22(18)30 November 2008p 2540-2543
Hammond, Emma; McKinnon, Elizabeth; Mallal, Simon; Nolan, David
Centre for Clinical Immunology and Biomedical Statistics, Murdoch University, South Street, Murdoch, Western Australia 6150, Australia.
Correspondence to Emma Hammond, Centre for Clinical Immunology and Biomedical Statistics, Murdoch University, South Street, Murdoch, WA 6150, Australia.
"In conclusion, this study sought evidence from longitudinal biomarker analyses for an abacavir-specific signal that could provide a link between abacavir treatment and susceptibility to cardiovascular events. In general, inflammatory and metabolic indicators were not significantly worse on abacavir-based therapy, and we were unable to detect an effect of abacavir treatment that was consistently evident in both ART-naive and ART-experienced patients. Among participants who were ART-naive at baseline, the introduction of effective HIV therapy was associated with a reduction of proinflammatory cytokine and D-dimer levels, as previously suggested [4,16]. In this context, the nonsignificant increase in IL-6 levels is a notable exception and, given recent findings from the SMART study [2], this warrants further consideration. However, it is notable that IL-6 levels were not influenced by the introduction of abacavir with ART-experienced patients (P = 0.9), suggesting that this effect is not directly attributable to abacavir treatment. Otherwise, our findings are concordant with previous studies in demonstrating that previous thymidine NRTI treatment, and consequent lipoatrophy, has adverse effects on serum adipokine levels [15]. These results suggest that relationships between HIV infection, choice of NRTI therapy, and cardiovascular disease outcomes are likely to be complex, and that longitudinal study design will be an important element in unravelling the clinically important observations that have been revealed by the D:A:D and SMART studies."
Data from the D:A:D study indicate that patients at high risk of cardiovascular disease appeared to be more prone to developing myocardial infarction (MI) when on abacavir, but not on stavudine or zidovudine [1]. This association, which was neither related to cumulative abacavir exposure nor persistent after abacavir cessation, has since been supported by data from the self-monitoring, analysis, and reporting technology (SMART) trial, which also identified an approximately 20% increase in C-reactive protein (CRP) and interleukin (IL)-6 inflammatory markers in abacavir-treated patients [2]. In each case, the analysis of cross-sectional data has precluded any ability to investigate whether abacavir is directly responsible for these effects. Alternative explanations may include the preferred use of abacavir for patients with high cardiovascular risk, due to its perceived metabolic safety [3]. Incomplete suppression of HIV infection may also represent a risk factor, as in the SMART study, in which D-dimer levels were found to be highly predictive of cardiovascular events in untreated HIV disease [4].
This study provides a longitudinal evaluation of biomarkers of cardiovascular relevance during abacavir therapy in previously treatment-naive patients as well as treatment-experienced patients with HIV infection. We evaluated levels of D-dimer, as well as highly sensitive C reactive protein (hs-CRP), IL-6, IL-8, tumour necrosis factor (TNF)β, monocyte chemoattractant protein (MCP-1), hepatocyte growth factor (HGF) [5-8], as well as adiponectin and leptin [9,10], as inflammatory and metabolic markers potentially relevant to cardiovascular disease.
Western Australian HIV observational cohort participants were selected according to treatment choice, and samples were obtained from plasma aliquots collected in EDTA, processed by centrifugation within 6 h of collection, and stored at -80C during routine viral load measurements. There were no MIs in the study population. In order to investigate specific effect of abacavir treatment in both antiretroviral therapy (ART)-naive and ART-experienced individuals, study selection comprised: 15 ART-naive patients initiating abacavir, but not stavudine or zidovudine; 13 abacavir-naive individuals switching to abacavir from thymidine nucleoside reverse transcriptase inhibitor (NRTI) therapy; and 13 abacavir-experienced patients switching from thymidine NRTI therapy while continuing abacavir treatment. Demographic, clinical, and ART data are provided in Table 1. Each participant was sampled at least once, both at baseline and after initiating or switching therapy, with a median (range) of 3 (2-4) samples per person, and a time period of 25 (3-67) months since initiating or switching treatment.
D-dimer was measured by latex immunoassay (Diagnostica Stago, Asnieres, France) on the STA-R coagulation analyser (Diagnostica Stago). The coefficient of variation for intra-assay was less than 3%, and for interassay less than 5%. Hs-CRP was measured by latex-enhanced nephelometry (N High Sensitivity CRP assay) on a Behring II nephelometer (Dade Behring Inc., Deerfield, Illinois, USA). All other analytes were determined by immunoassay using the LINCOplex kit (Luminex xMAP Technology; Lumunex, Moorabbin, Australia) on a luminex 200 platform (Luminex). Human Serum Adipokine Panel B (HADK2-61K-B) was used for IL-6, IL-8, TNFβ, MCP-1, HGF, and leptin. Human Serum Adipokine Panel A (HADK1-61K-A) was used for adiponectin.
Statistical analysis of log-transformed analyte data was undertaken within a linear regression framework, utilizing mixed effect models to accommodate multiple measures per person. Univariate and multivariate analyses were performed, with adjustment (where appropriate) for demographic factors (sex, non-Caucasian race), HIV disease status (viral load, CD4 cells/μl), and severity of lipoatrophy [11] (%leg-fat/BMI, as determined by Dual energy X-ray absorptiometry scanning). Secondary analyses also assessed the influence of HIV protease inhibitor or nonnucleoside reverse transcriptase inhibitor (NNRTI) therapy, as well as previous zidovudine versus stavudine exposure.
Values for plasma analytes and the results for the primary analyses of biomarker changes are provided in Table 1, with median values for all analytes within the range reported in previous studies of diverse populations [12-14]. Amongst ART-naive individuals initiating abacavir, reduced on-treatment D-dimer levels were observed (P = 0.08), particularly after 1 year of therapy (P = 0.03), although the correlation with duration was not linear (P = 0.2). Significant decreases from baseline were also observed for IL-8 (P = 0.03), TNFβ (P = 0.009), and MCP-1 (P = 0.006) levels. IL-6 levels increased marginally (P = 0.08), whereas no significant changes were observed in levels of HGF (P = 0.8) or hs-CRP (P = 0.3). Leptin levels increased after initiating abacavir-based therapy (P = 0.004), whereas a marginal increase in adiponectin levels (P = 0.1) reflected some early increases rather than a trend over the duration of therapy. Adjusting for baseline demographics or HIV status (viral load or CD4 cells/μl) had negligible effect on any of the results, except it is noted that decreases in tumour necrosis factor were greater in those with higher baseline viral load.
Amongst patients switching from thymidine NRTI therapy, baseline values for all analytes were similar when comparing abacavir-naive and abacavir-experienced individuals (P > 0.2) as well as protease inhibitor versus NNRTI treatment use (P > 0.2). There was a nonsignificant trend towards decreasing D-Dimer levels following cessation of thymidine NRTI therapy, in both abacavir-initiating (P = 0.3) and abacavir-continuing groups (P = 0.2). Levels of hs-CRP, IL-6, IL-8, and MCP-1 were not significantly altered with switching for either group (P > 0.2). A small increase in TNFβ levels was observed in those initiating abacavir (P = 0.04), and a small decrease in HGF levels amongst those continuing abacavir on switching (P = 0.05), with no change from baseline markedly abrogated by covariate adjustment.
Among thymidine-NRTI-experienced patients, baseline adiponectin levels were decreased relative to the ART-naive group (P = 0.03), and correlated with the severity of lipoatrophy (measured by nadir %leg-fat/BMI, P = 0.03) as shown by previous studies [15]. Following NRTI switching, adiponectin levels remained stable (P = 0.5), irrespective of inclusion of abacavir in the baseline regimen or inclusion of a protease inhibitor or NNRTI, likely reflecting the loss of adipocyte cells in subcutaneous fat tissue. Conversely, leptin levels increased after switching from thymidine NRTI therapy regimen (P = 0.006), irrespective of previous abacavir exposure, choice of protease inhibitor versus NNRTI therapy, or severity of lipoatrophy.
In conclusion, this study sought evidence from longitudinal biomarker analyses for an abacavir-specific signal that could provide a link between abacavir treatment and susceptibility to cardiovascular events. In general, inflammatory and metabolic indicators were not significantly worse on abacavir-based therapy, and we were unable to detect an effect of abacavir treatment that was consistently evident in both ART-naive and ART-experienced patients. Among participants who were ART-naive at baseline, the introduction of effective HIV therapy was associated with a reduction of proinflammatory cytokine and D-dimer levels, as previously suggested [4,16]. In this context, the nonsignificant increase in IL-6 levels is a notable exception and, given recent findings from the SMART study [2], this warrants further consideration. However, it is notable that IL-6 levels were not influenced by the introduction of abacavir with ART-experienced patients (P = 0.9), suggesting that this effect is not directly attributable to abacavir treatment. Otherwise, our findings are concordant with previous studies in demonstrating that previous thymidine NRTI treatment, and consequent lipoatrophy, has adverse effects on serum adipokine levels [15]. These results suggest that relationships between HIV infection, choice of NRTI therapy, and cardiovascular disease outcomes are likely to be complex, and that longitudinal study design will be an important element in unravelling the clinically important observations that have been revealed by the D:A:D and SMART studies.
|
|
| |
| |
|
|
|
