High prevalence of the metabolic syndrome in HIV-infected patients: impact of different definitions of the metabolic syndrome
Worm, S.W., Friis-Møller, N., Bruyand, M., Monforte, A.D., Rickenbach, M., Reiss, P., El-Sadr, W., Phillips, A., Lundgren, J., Sabin, C. and Mallal, S. (2010) High prevalence of the metabolic syndrome in HIV-infected patients: impact of different definitions of the metabolic syndrome. AIDS, 24 (3). pp. 427-435.
*Subscription may be required
Introduction: This study describes the characteristics of the metabolic syndrome in HIV-positive patients in the Data Collection on Adverse Events of Anti-HIV Drugs study and discusses the impact of different methodological approaches on estimates of the prevalence of metabolic syndrome over time.
Methods: We described the prevalence of the metabolic syndrome in patients under follow-up at the end of six calendar periods from 2000 to 2007. The definition that was used for the metabolic syndrome was modified to take account of the use of lipid-lowering and antihypertensive medication, measurement variability and missing values, and assessed the impact of these modifications on the estimated prevalence.
Results: For all definitions considered, there was an increasing prevalence of the metabolic syndrome over time, although the prevalence estimates themselves varied widely. Using our primary definition, we found an increase in prevalence from 19.4% in 2000/2001 to 41.6% in 2006/2007. Modification of the definition to incorporate antihypertensive and lipid-lowering medication had relatively little impact on the prevalence estimates, as did modification to allow for missing data. In contrast, modification to allow the metabolic syndrome to be reversible and to allow for measurement variability lowered prevalence estimates substantially.
Discussion: The prevalence of the metabolic syndrome in cohort studies is largely based on the use of nonstandardized measurements as they are captured in daily clinical care. As a result, bias is easily introduced, particularly when measurements are both highly variable and may be missing. We suggest that the prevalence of the metabolic syndrome in cohort studies should be based on two consecutive measurements of the laboratory components in the syndrome definition.
|Publication Type:||Journal Article|
|Murdoch Affiliation:||Centre for Clinical Immunology and Biomedical Statistics|
|Publisher:||Lippincott Williams & Wilkins Ltd.|
|Copyright:||© 2010 Lippincott Williams & Wilkins, Inc.|
|Notes:||Simon Mallal appears as part of the D:A:D Study Group|
|Item Control Page|