Comparing the prevalence of hypertension between HIV-positive and HIV-negative adults: a global systematic review and meta-analysis of cross-sectional studies
K Davis1, P Perez-Guzman1, A Hoyer2, R Brinks3, E Gregg4, KN Althoff5, AC Justice6,7, P Reiss8,9,10, S Gregson1,11, M Smit1,12
1Imperial College London, London, United Kingdom; 2Ludwig-Maximilians-University Munich, Munich, Germany; 3University Hospital Duesseldorf, Duesseldorf, Germany; 4Department of Epidemiology and Biostatistics, Imperial College London, London, UK; 5Department of Epidemiology, Johns Hopkins University, Baltimore, MD, USA; 6Yale University, New Haven, CT, USA; 7The Veterans Affairs Connecticut Healthcare System, New Haven, CT, USA; 8Amsterdam University Medical Centers, University of Amsterdam, Amsterdam, the Netherlands; 9Amsterdam Institute for Global Health and Development, Amsterdam, the Netherlands; 10HIV Monitoring Foundation, Amsterdam, the Netherlands; 11 Biomedical Research and Training Institute, Harare, Zimbabwe; 12Infectious Disease Unit, University Hospitals Geneva, Switzerland
Objectives/aim: As life expectancy among people living with HIV (PLHIV) has increased dramatically since effective antiretroviral therapy was introduced, ageing PLHIV are increasingly affected by non-communicable diseases, including hypertension. Hypertension, though often asymptomatic, is a key risk factor for other conditions like chronic kidney disease and cardiovascular disease. In recent years, several observational studies have found a differing prevalence of hypertension among PLHIV compared with HIV-negative individuals, with some finding an elevated burden and others finding a reduced burden. However, this evidence has not been systematically assessed, undermining efforts to develop hypertension prevention and care strategies. We aimed to perform the first global systematic review and meta-analysis of hypertension prevalence by HIV status.
Methods: We searched MEDLINE, EMBASE, Global Health and Cochrane Central Register of Controlled Trials to 23 August 2019, and reference lists of included articles, to identify published cross-sectional studies presenting hypertension prevalence by HIV status among adults aged over 15. Our risk of bias assessments addressed adequacy of sample sizes, participant selection and HIV and hypertension status measurement. We pooled risk ratios (RRs) for prevalent hypertension with random effects models. We also explored whether any difference in hypertension prevalence was associated with study-level factors through sub-group analyses and meta-regression. Our sensitivity analyses examined the impact of removing studies at high risk of bias and applying the Hartung-Knapp modification to account for small sample sizes. The study followed PRISMA guidelines and is registered with PROSPERO (CRD42019151359).
Results: Of 20,305 studies identified, 51 were eligible (1,943,047 participants). Overall, global hypertension prevalence was lower among PLHIV than HIV-negative individuals (RR: 0.90, 95% confidence interval [CI]: 0.84, 0.97, I2=97%). The relationship varied by continent (Figure 1; HPN: hypertension), with prevalence higher among PLHIV in North America (1.13, 1.00, 1.27) and lower among PLHIV in Africa (0.75, 0.66, 0.84) and Asia (0.77, 0.63, 0.95). Meta-regression con-firmed that the risk ratio was higher for North America than Africa (coefficient: 1.46, 95% CI: 1.16, 1.83). Removing two studies at high risk of bias indicated that the risk ratio for Europe was also higher than the risk ratio for Africa (1.28, 1.04, 1.57). Our results were robust to use of the Hartung-Knapp modification.
(Abstract O17)
Conclusion(s)/discussion: Our findings demonstrate that the relative vulnerability of PLHIV to hypertension differs by region. This suggests that policy for hypertension prevention and care will require tailoring to local needs. Our results also highlight the need for further studies of PLHIV and matched negative controls, in order to better understand the dynamics and mechanisms of observed context-specific trends and shape care accordingly. Efforts to safeguard the long-term quality of life of PLHIV will rely on multidiscipli-nary work to respond to the varying challenges posed by comorbidities.
Abstract O18
Antiviral Therapy 2020; 25 Suppl 1:A47
Carotid atherosclerosis in suppressed HIV patients and in a healthy sample: prediction by cardiovascular risk equations
M Saumoy1,2, S Di Yacovo1,2, I Subirana3,4, JM Valdivielso5, A Imaz1,2, B García1,2, J Tiraboschi1,2, D Podzamczer1,2, M Grau3,6,7
1HIV and STD Unit, Infectious Disease Service, Bellvitge University Hospital; 2Bellvitge Biomedical Research Institute, Hospitalet de Llobregat, Spain; 3IMIM – Hospital del Mar Medical Research Institute, Barcelona, Spain; 4Consortium for Biomedical Research in Cardiovascular Diseases (CIBERCV), Barcelona, Spain; 5Vascular and Renal Translational Research Group. UDETMA. Biomedical Research Institute of Lledia, Lérida, Spain; 6Consortium for Biomedical Research in Epidemiology and Public Health (CIBERESP), Barcelona, Spain; 7Department of Medicine, University of Barcelona, Barcelona, Spain
Background: Carotid plaque assessment may improve cardiovascular (CV) risk prediction and could reclassify patients to a higher risk, which would benefit of intensive preventive measures. The aim of the study was to assess the relation between carotid plaque and several CV risk equations in an HIV-infected suppressed sample and to compare with a healthy sample.
Methodology: Cross-sectional study of two samples matched by age and sex: (1) HIV- patients randomly selected from an HIV Unit in 2014–16; (2) community-based sample randomly selected from Girona (Spain) in 2016. All individuals were aged 30 to 80 years and had no history of CV disease. HIV patients had undetect-able viral load on the last 6 months. The presence of plaque in the common, bulb and internal carotid was assessed by B-mode ultrasound. 10-year CV risk was assessed by 5 equations: three designed for the general population (Framingham, REGICOR [Framingham calibrated and validated for the Spanish population] and Systematic Coronary Risk Evaluation [SCORE]) and calibrated to HIV-infected population (COMVIH [Framingham calibrated to Spanish HIV population] and the reduced D:A:D).
Results: 758 participants were included (379 per group), 78% men, age 50 (10). HIV-sample: CD4 718 (338) cell/ul, duration of antiretroviral therapy 14.7 (6.6) years. There were no differences between samples in the presence of carotid plaque (HIV 33.2%, control 31.3%; adjusted odds ratio [95% CI] 1.15 [0.79,1.68]).
All three CV risk equations for the general population showed that HIV-infected patients have a slight but significantly higher risk of future severe CV outcomes than controls (Framingham [7.52 versus 6.48], REGICOR [2.67 versus 2.16], SCORE [0.40 versus 0.26]; P≤0.001 for all comparisons). However, there were no significant differences between groups in the percentage of HIV-infected patients or controls in different risk categories for Framingham (P=0.843), REGICOR (P=0.375) or SCORE (P=1.000). The COMVIH equation in the HIV-infected group classified a higher number of patients in the moderate and high CV risk categories. The presence of carotid plaque increased alongside the increase in risk category, with no significant differences according to HIV serostatus. When CV risk equations for general population were applied, among 18% to 27% of participants classified in the low CV risk category had carotid plaque; the percentage of carotid plaque in those classified in the moderate risk category was between 47% to 76%. These percentages were similar when COMVIH or reduced D:A:D were used. See Figure 1.
(Abstract O18)
Conclusions: An important number of participants clas-sified in the low and moderate CV risk categories by all the equations and in both groups had carotid plaque. Assessment of carotid plaque by ultrasound may be useful for identifying HIV-infected patients who could benefit from more aggressive management of preventive measures.
Abstract O19
Antiviral Therapy 2020; 25 Suppl 1:A49
Selective drop-out of HIV-positive AGEhIV Cohort participants may bias estimates of long-term adverse health effects of ageing with HIV
E Verheij1,2, A Boyd3, FW Wit1,3, SO Verboeket1,2, MF Schim van der Loeff4, P Reiss1,2,3
1Amsterdam University Medical Centers, University of Amsterdam, Department of Global Health and Division of Infectious Diseases, Amsterdam Infection and Immunity Institute and Amsterdam Public Health Research Institute, Amsterdam, the Netherlands; 2Amsterdam Institute for Global Health and Development, Amsterdam, the Netherlands; 3HIV Monitoring Foundation, Amsterdam, the Netherlands; 4Public Health Service of Amsterdam, Department of Infectious Diseases, Amsterdam, the Netherlands
Aims: Attrition is an important challenge for longitudinal studies, especially when attrition is associated with the studied exposure or outcome variables (that is, selective drop-out). This problem is likely to occur in cohort studies aimed at evaluating clinical outcomes in ageing people with HIV (PWH). A higher burden of co- and multimorbidity in ageing PWH could lead to earlier study discontinuation and thereby to biased estimates of incident comorbidities in PWH. This analysis aims to (1) quantify attrition and its reasons and (2) determine factors possibly associated with attrition in the AGEhIV Cohort Study.
Methods: From October 2010 through September 2012 589 HIV-positive and 550 lifestyle comparable HIV-negative participants, aged ≥45 years, were included into the cohort. Participants underwent extensive health evaluation (±1.5 h) during biennial study visits. In this analysis follow-up data through October 2019 were included. Attrition was defined as not having attended a study-visit without advance notice within 6 months of its scheduled date, or withdrawal of consent. Follow-up data were analysed in intervals of 2 years. Observation time was censored at time of attrition, death or the last scheduled study-visit, whichever occurred first. Complementary log-log regression was used to (1) estimate hazard ratios (HRs) for attrition during each interval, comparing HIV-positive and HIV-negative participants, and (2) assess factors possibly associated with attrition, including HIV-status, socio-demographics, having ≥1 comorbidities and depressive symptoms (categorized as ≤8, 9-15 or ≥16 points on the Center for Epidemiologic Studies Depression [CES-D] questionnaire).
Results: During the 9-year observation period, attrition occurred in 178 (29.8%) HIV-positive and 90 (16.4%) HIV-negative participants, of whom 87 (48.9%) and 41 (45.6%), respectively, withdrew consent. Reasons for withdrawal were more commonly health-related (for example, study-visit was too strenuous or poor health condition prevented study participation) in HIV-positive (n=33, 37.9%) versus HIV-negative (n=8, 19.5%; P=0.01) participants. Thirty HIV-positive and 8 HIV-negative participants died. HIV-status was associated with attrition (univariable HRHIV1.9, 95% CI=1.5, 2.5; P<0.001; Figure 1), this effect was attenuated after adjustment for gender, age at enrolment, non-Dutch origin, level of education, and having ≥1 comorbidity or depressive symptoms at a study-visit (multivariable HRHIV1.5, 95% CI=1.1, 2.0; P=0.003). Non-Dutch origin, lower educational level, having ≥1 comorbidities or severe depressive symptoms (≥16 points on the CES-D) each were also independently associated with attrition, without significant interactions with HIV-status for any of these factors.
(Abstract O19)
Conclusions: HIV-positive status was independently associated with greater risk of attrition, with poor health often reported as the reason. This was only partly explained by having greater numbers of comorbidities or depressive symptoms, each of which were equally associated with attrition in HIV-positive and negative participants. Indicators of poor health not captured by our study may contribute to the increased attrition rate among PWH. These findings suggest that selective drop-out may lead to underestimates of long-term adverse health effects of ageing in PWH.
Abstract O20
Antiviral Therapy 2020; 25 Suppl 1:A50
Choline-containing lipid species are associated with hepatic steatosis in people with HIV
CL Gabriel1,2, F Ye3, R Fan3, CN Wanjalla4, M Mashayekhi5, S Bailin4, JR Koethe2,4,6
1Division of Gastroenterology, Hepatology and Nutrition, Vanderbilt University Medical Center, Nashville, TN, USA; 2Tennessee Center for AIDS Research, Vanderbilt University Medical Center, Nashville, TN, USA; 3Department of Biostatistics, Vanderbilt University Medical Center, Nashville, Tennessee, USA; 4Division of Infectious Diseases, Vanderbilt University Medical Center, Nashville, TN, USA; 5Division of Diabetes, Endocrinology and Metabolism, Vanderbilt University Medical Center, Nashville, Tennessee, USA; 6Veterans Affairs Tennessee Valley Healthcare System, Nashville, TN, USA
Objectives/aims: People with HIV (PWH) have a higher prevalence of non-alcoholic fatty liver disease (NAFLD) than HIV-negative persons. The mechanisms that give rise to this disparity are not known but may be due to virus-specific effects, antiretroviral therapy or changes in the intestinal microbiome. Collectively, these factors contribute to alterations of the plasma lipidome that promote lipid accumulation, inflammation and fibrosis in the liver. A more complete understanding of the plasma lipidome will provide targets for specific therapies and diagnostics for NAFLD in PWH. In this study, we define the plasma lipidome associated with NAFLD as compared with other ectopic lipid depots in PWH.
Methods: Adult PWH on antiretroviral therapy with sustained viral suppression were recruited at a large academic medical centre. Participants underwent non-contrasted abdominal CT scans and the size and/or density of skeletal muscle, liver and adipose tissue were measured. Fasted plasma was collected from all participants for untargeted liquid chromatography-mass spectrometry lipidomics profiling. The abundance of lipid species relative to the size and/or density of tissues as measured on CT scan was assessed using linear regression modelling.
Results: The mean age of the cohort was 46 years old and BMI was 33.3. 78% of participants were male and 53% were White. The average CD4+ T-cell count was 498 cells/mm3 and the average duration of antiretrovi-ral therapy was 8.6 years. A higher abundance of choline-containing species, including phosphatidylcholine and lysophosphatidylcholine was associated with lower levels of hepatic steatosis (as measured by higher CT attenuation; Figure 1). Conversely, higher abundance of triglycerides and oxidized triglyceride species was associated with greater hepatic steatosis. Visceral fat attenuation showed similar relationships with triglycerides and choline-containing lipid species, whereas visceral fat volume and pericardial fat volume had an inverse relationship with these species.
(Abstract O20)
Discussion: A decreased abundance of plasma choline-containing species and increased levels of triglyceride species is associated with hepatic steatosis in PWH. Plasma choline deficiency has previously been shown to be associated with hepatic steatosis and cardiovascular disease in HIV and could be secondary to intestinal dysbiosis. Ongoing studies will determine whether changes in the intestinal microbiome are associated with a ‘pro-NAFLD’ plasma lipidome in PWH.
