INTRODUCTION

With more than 28 million people living with HIV (PLHIV) on antiretroviral therapy (ART) worldwide, HIV is now considered a chronic communicable disease associated with declining HIV-related mortality.1 However, a rise in traditional cardiovascular disease (CVD) risk factors, coupled with the effects of ART and HIV infection, drives the excess risk of developing metabolic syndrome (MetS) and subsequent atherosclerotic CVD in PLHIV compared with their HIV-uninfected counterparts.2,3

Globally, CVD is the leading cause of mortality4 with low and middle income countries (LMICs) bearing the brunt of the disease primarily due to the effects of urbanization and behavioral change.5 The relative risk of myocardial infarction among women with living with HIV (WLHIV) versus without HIV is higher than that among men with versus without HIV.6 Heightened immune activation stemming from unique responses to the virus and the presence of comorbid obesity compared with men, in addition to accelerated reproductive aging, poses WLHIV at a greater predisposition to CVDs.7 These factors compound the preexisting risk associated with the use of exogenous hormones and pregnancy-related complications8 and intimate partner violence in women, which are associated with poorer outcomes in CVD.9,10 Sub-Saharan Africa has the highest prevalence of uncontrolled hypertension in young adults11 and more prominent in women.12 Globally, 19.1 million women and girls are living with HIV, with one-fifth residing in South Africa.1 In addition, the epidemics of HIV and hypertension are exacerbated by the third epidemic of obesity in women, and the introduction of dolutegravir (DTG) has heightened the risk of obesity in this population.13,14

The rise in CVD risk in PLHIV has led to the recommendation for the integration of CVD risk assessments into HIV care by the World Health Organization since 2016.15 South Africa included baseline CVD risk assessments in the 2019 HIV treatment guidelines,16 but implementation has been slow. Clinical trials focusing on lifestyle modification, such as physical activity and dietary programs, have demonstrated the benefit of lifestyle modification interventions on traditional risk factors for CVD in both men and WLHIV;17 however, studies are minimal in LMICs. Until REPRIEVE,18 there were no evidence-based medical therapies to prevent CVD in PLHIV, although time to scale-up statin use in PLHIV older than 40 years in LMICs could take a while. Upscaling tailored noncommunicable disease management interventions is essential for reaching the Sustainable Development Goal target to reduce noncommunicable disease–related premature mortality by one-third by 2030 relative to 2015.19

The Integration of cardiovascular disease SCreening and prevention in the HIV MAnagement plan for women of reproductive age (ISCHeMiA) study set out to determine the effectiveness of the WHO Package of Essential noncommunicable disease interventions for primary health care in low-resource settings20 in identifying and modifying risk factors for CVD in WLHIV receiving HIV care at a research clinic compared with WLHIV who receive routine care at a primary health care (PHC) clinic.

METHODS Design

The study was conducted at a research clinic (intervention arm) and a PHC (control arm) in Umlazi, a periurban suburb in South Africa. The detailed methodology of the ISCHeMiA study has been described elsewhere.21 In summary, it is a prospective, 2-arm quasiexperimental study comparing a primary health care CVD risk-reduction intervention plan (intervention arm [I-arm]) with routine care (control arm [c-arm]). Women aged 18 years to younger than 50 years receiving ART for a minimum of 1 year were enrolled from November 2018 until May 2019 and were followed up until June 2021 to February 2022. Women in the I-arm were recruited from a cohort of participants coenrolled in the PEPFAR PROMise Ongoing Treatment Evaluation (PROMOTE) observational study at the CAPRISA Umlazi clinical research site based at the Prince Mshiyeni Memorial Hospital. The PROMOTE study was implemented to provide long-term adherence and safety outcomes of using combination ART from the standard of care health care providers.22 With the exception of routine weight and blood pressure assessment, there were no screening and intervention procedures for CVD provided through the PROMOTE study before ISCHeMiA entry. Women in the C-arm were enrolled and followed up at the Prince Mshiyeni Hospital Gateway PHC.

The Intervention Intervention

Following informed consent, sociodemographic, conventional and HIV-related CVD risk factors were assessed through annual questionnaires and standardized clinical and laboratory procedures performed by trained research staff in the I-arm. Procedures included regular screening for obesity, hypertension, diabetes mellitus (DM), and high cholesterol and prompt initiation of treatment when indicated. Every woman in the I-arm received a lifestyle modification advice handout on diet, physical activity, alcohol use, and smoking cessation (see Appendix 1, Supplemental Digital Content, https://links.lww.com/QAI/C268) at the study entry, which was discussed for up to 15 minutes with the research participant and emphasized briefly at each study visit. Participants with a body mass index (BMI) of ≥30 kg/m2, hypertension, DM, and high cholesterol were referred to a dietitian to schedule an appointment for a formal educational class on recommended diet and exercise.

Control

Participants in the C-arm continued their management per standard of care at the PHC with no study intervention. Following informed consent, information collected at study entry visits 1 and 2 were obtained through the local clinic medical records. Because of the standard of care at PHC, only age, BMI, blood pressure (BP), and HIV factors were available at baseline. At the end of the study, a complete set of sociodemographic, conventional, and HIV-related CVD risk factors were collected in both assigned groups, as described in the intervention arm above.

Ethical Considerations and Approvals

The University of KwaZulu-Natal Biomedical Research Ethics Committee approved the ISCHeMiA study (Ref BFC 220/18), and the study maintained ethical standards aligned with the Helsinki Declaration (1964, amended in 2008). Approvals were similarly obtained from the regional Department of Health, Prince Mshiyeni Memorial Hospital, and PROMOTE publication committee. The trial was registered with the Pan African Clinical Trial Registry database, PACTR201808524461224.

Statistical Analysis

With a 20%–30% estimated prevalence of MetS in PLHIV without any specific intervention in the risk factor modification of CVD and a proposed 10%–15% estimated incidence of MetS in the I-arm, 132 participants per arm would allow for 80% power with an additional 10% added to allow for loss to follow-up. Metabolic syndrome was the primary outcome of interest in this analysis and was defined as having 3 or more of the following subcomponents at a given visit: waist circumference (WC) ≥ 80 cm, elevated triglycerides ≥ 1.7 mmol/L, reduced high-density lipoprotein (HDL) levels < 1.3 mmol/L, elevated systolic blood pressure (SBP) ≥ 130 and/or diastolic blood pressure (DBP) ≥ 85 mm Hg or on any antihypertensive treatment, and an elevated fasting glucose ≥ 5.6 mmol/L or receiving treatment for type 2 DM (T2DM).23 Secondary outcome measure cutoff points were elevated high sensitivity C-reactive protein (hsCRP) ≥ 3 mg/L, BMI ≥ 30 kg/m2, urine albumin-to-creatinine ratio (U-ACR) ≥ 3 mg/mmol, fasting total cholesterol (TC) ≥ 5 mmol/L, fasting low-density lipoprotein (LDL) ≥ 3 mmol/L, newly diagnosed hypertension (BP ≥140/90 mm Hg), and newly diagnosed T2DM (fasting blood glucose [BG] ≥7 mmol/L). Diet was categorized as unhealthy based on participant's response to a low fruit/vegetable intake, high-fat, and high-salt diet. Exercise was defined as physical activity of >30 min/d and week.

Data were collected using standardized questionnaires, clinical templates, and laboratory-generated reports, captured on Microsoft Excel, and analyzed using IBM SPSS statistics software, version 25.0. End-of-study (EOS) sociodemographic, lifestyle practice, and HIV-related risk factors (ART regimen and duration, CD4 count, and viral load) were represented by frequencies, means (SD) at the end of the study. Baseline variables and related methodology were previously summarized and published.21

End-of-study prevalence of MetS, subcomponents of MetS, and other CVD risk factors were compared between arms using relative risk ratios (RR). The risk ratios were estimated using multivariate logistic regression, unadjusted and adjusted for age, baseline BMI, SBP, hypertension and diabetes diagnosis, tobacco smoking, ART duration, and CD4 count; 95% confidence intervals for the difference between arms were based on the Wald statistics. Risk ratio P values were used to summarize the strength of evidence against the null hypothesis of no difference, but a strict cutoff was not used. Instead, the 95% confidence interval served as the primary focus of statistical inference, where the inclusion of value of 1 would be interpreted as nonsignificant.

Independent samples t tests between means across arms were used to test mean differences between the continuous CVD risk factors variables and between baseline and EOS components in the intervention arm. The McNemar test was used to determine change in categorical variables from baseline to end of study. Comparison in variables from baseline to EOS were limited to participants who completed EOS visits.

The proportion of women in each arm who transitioned to DTG-based ART were described, and trends in BMI were observed in I-arm participants who did and did not switch to DTG. Pearson χ2 test was used to assess for significant differences in the status of MetS and other CVD risk factors from baseline to EOS between participants in the I-arm who transitioned to DTG and those who did not.

RESULTS

Of the 372 women enrolled from November 2018 to May 2019 (I-arm = 186, C-arm = 186), a total of 269 WLHIV (I-arm = 149 and C-arm = 120) were included in the EOS analyses at a median age of 34 years (31.0–38.0 years) in the I-arm and 38 years (32.0–41.0 years) in the C-arm (Fig. 1 Consort diagram). All women were Black South African.

FIGURE 1.:

CONSORT diagram. Source: Author created using Microsoft word Consort 2010 template.

Table 1 and 2 summarizes participants' sociodemographic, CVD risk factors and HIV-related characteristics by the assigned group at baseline and EOS. Sociodemographic data such as employment status and HIV characteristics were comparable between both arms. Table 3 displays the comparisons of MetS and other CVD risk factors by assigned group at EOS. Risk ratios were adjusted for baseline age, tobacco smoking status, BMI, diagnosis of hypertension and T2DM, SBP, ART duration, and CD4 count. The EOS prevalence of MetS was higher in the C-arm versus I-arm (estimated difference 3.2%); however, confidence interval included the null value (RR = 1), and the difference between arms was not statistically significant.

TABLE 1. - Baseline HIV-Related Characteristics and CVD Risk Factors by Assigned Group Variable Study Intervention
N = 149 Control Arm
N = 120 P Sociodemographic characteristics  Age, yr, mean (SD) 29.7 (6.8) 27.8 (6.6) 0.018*  Known with hypertension on treatment, n (%)   Yes 8 (5.4) 2 (1.7) 0.73†  Known with T2DM on treatment, n (%)   Yes 1 (0.7) 1 (0.8) 1.000†  BMI mean (SD) 29.7 (6.8) 27.8 (6.6) 0.018*  Systolic BP mm Hg, mean (SD) 117.1 (13.1) 114.2 (18.8) 0.141  Systolic BP ≥130 mm Hg, n (%)   Yes 23 (15.5) 15 (12.7) 0.512†  Tobacco smoking, n (%)   Yes 6 (4.0) 3 (2.5) 0.88† HIV-related characteristics  ART duration, yr, mean (SD) 4.1 (0.8) 4.7 (2.9) <0.0.097*  CD4, cells/uL, mean (SD), n (%) 947 (272) 657 (279) <0.001*  CD4 >500 144 (96.6) 79 (66.4) <0.001†

Data are n (%), mean (SD), P value.

*Independent samples t test

†Pearson χ2 test.

Mean HDL and hsCRP were abnormal in both arms. There was a significant difference in mean ± SD of WC between arms (95% CI: −10.57 to −3.69; P < 0.001). Waist circumference at a higher threshold of ≥88 cm was 81/149 (54.4%) and 58/120 (48.3%) in I and C-arms, respectively. When excluding measurements from the 12 women currently pregnant at EOS, the mean ± SD WC in the I-arm remained high at 94.5 ± 13.8 cm. There was no difference between the mean ± SD BMI, excluding currently pregnant women of 31.5 ± 6.97 kg/m2 and including pregnancy.

After adjusting for baseline risk factors, participants in the I-arm were 0.2 times less likely to have elevated fasting BG levels (95% CI: 0.063 to 0.784; P = 0.02). The 95% confidence interval did not include the null value (RR = 1) and was statistically significant. Although there were by-arm differences in LDL levels, these were not significant (95% CI: 0.387–1.095; P = 0.115).

The reports of a perceived healthy diet and duration of physical activity were higher in the I-arm compared with C-arm. However, reports of tobacco smoking, but not alcohol use, were also higher in the I-arm (Tables 1 and 2). Within the I-arm, the percentage of participants who reported having an unhealthy diet improved significantly from baseline (P < 0.001) (Table 4).

TABLE 2. - Sociodemographic and HIV-Related Characteristics by Assigned Group at End-Of-Study Variable Study Intervention
N = 149 Control Arm
N = 120 P Sociodemographic characteristics  Age, yr, mean (SD) 35.0 (5.42) 37.1 (6.09) 0.0031*  Employment, n (%)   Yes 56 (37.6) 49 (40.8) 0.587†  Parity, mean (SD) 3 (1.414) 2.17 (1.001) 0.257*  Pregnant/recent birth, n (%)   Yes 12 (8.1) 17 (14.2) 0.108†  Known family history of CVD, n (%)   Yes 23 (15.4) 26 (21.7) 0.177† HIV-related characteristics  ART regimen, n (%)   EFV/FTC/TDF 91 (61.1) 73 (60.8) 0.886†   DTG/3 TC/TDF 56 (37.6) 44 (36.7)   AZT/3 TC + lopinavir/ritonavir 2 (1.3) 2 (1.7)   ABC/3 TC + EFV 0 (0) 1 (0.8)  ART duration, yr, mean (SD) 5.76 (0.886) 6.83 (2.882) <0.001*  HIV-1 RNA, copies/mL, n (%) N = 119   <50 132 (88.6) 102 (85.7) 0.176†   50–200 6 (4) 11 (9.2)   >200 11 (7.4) 6 (5)  CD4, cells/uL, mean (SD), n (%) 909.30 (265.1) 680.45 (286.5) <0.001†  CD4 >500 142 (95.3) 83 (69.2)

Data are n (%), mean (SD), P value.

*Independent samples t test

†Pearson's χ2 test.

EFV/FTC/TDF, Efavirenz/Emtricitabine/Tenofovir disoproxil fumarate; DTG/3 TC/TDF, Dolutegravir/Lamivudine/Tenofovir disoproxil fumarate; AZT/3 TC, Zidovudine/Lamivudine; ABC/3 TC, Abacavir/Lamivudine.

TABLE 3. - Comparisons of MetS and Other CVD Risk Factors by Assigned Group at End Of Study Variable Study Intervention Control Arm Unadjusted RR (95% CI) Adjusted RR (95% CI) or Mean Difference (95% CI) P N = 149* N = 120* MetS at EOS, n (%)  Yes 25 (16.8) 24 (20.0) 0.84 (0.489 to 1.364) 1.14 (0.212 to 3.328) 0.861†  No 124 (83.2) 96 (80.0) Ref Ref WC mean (SD), n (%) 95.4 (13.97) 88.3 (14.55) −7.13 (−10.57 to −3.69) <0.001‡  WC ≥80 130 (87.8)§ 90 (75.0) 1.17 (1.055 to 1.243) 1.04 (0.765 to 1.203) 0.745†  WC <80 18 (12.2) 30 (25.0) Ref Ref BMI, kg/m2 mean (SD), n (%) 31.5 (6.98) 29.9 (7.29) −1.57 (−3.308 to 0.175) 0.078‡  BMI ≥30 78 (52.3)§ 49 (40.8) 1.31 (1.019 to 1.596) 1.27 (0.53 to 3.035) 0.593†  BMI <30 65 (43.6) 69 (57.5) Ref Ref Trig, mmol/L mean (SD), n (%) 0.84 (0.53) 0.79 (0.49) −0.048 (−0.173 to 0.764) 0.448‡  Trig ≥1.7 12 (8.1) 3 (2.5) 3.22 (0.942 to 9.648) 3.21 (0.81 to 10.779) 0.095†  Trig <1.7 137 (91.9) 117 (97.5) Ref Ref HDL, mmol/L mean (SD), n (%) 1.41 (0.41) 1.41 (0.34) −0.087 (−0.087 to 0.095) 0.929‡  HDL <1.3 64 (43.0) 51 (42.5) 1 (0.745 to 1.275) 1.01 (0.701 to 1.347) 0.930†  HDL ≥1.3 85 (57.0) 69 (57.5) Ref Ref TC, mmol/L mean (SD), n (%) 3.97 (0.87) 4.05 (0.85) 0.072 (−0.136 to 0.279) 0.497‡  TC ≥5.0 16 (10.7) 15 (12.5) 0.86 (0.43 to 1.623) 0.88 (0.396 to 1.796) 0.731†  TC <5.0 133 (89.3) 105 (87.5) Ref Ref LDL, mmol/L mean (SD), n (%) 2.40 (0.73) 2.56 (0.71) 0.152 (−0.022 to 0.325) 0.086‡  LDL ≥3.0 28 (18.8)§ 36 (30.0)‖ 0.6 (0.371 to 0.938) 0.67 (0.378 to 1.095) 0.115†  LDL <3.0 121 (81.2) 84 (70.0) Ref Ref BG, mmol/L mean (SD), n (%) 4.64 (0.79) 5.18 (2.43) 0.543 (−0.126 to 0.961) 0.011‡  BG ≥5.6 4 (2.7) 16 (13.3) 0.2 (0.067 to 0.591) 0.23 (0.063 to 0.784) 0.018†  BG <5.6 145 (97.3) 104 (86.7) Ref Ref Known with T2DM, n (%) 1 (0.7) 1 (0.8) SBP, mmHG mean (SD), n (%) 118.71 (12.10) 121.07 (18.80) 2.35 (−1.554 to 6.265) 0.236‡  SBP ≥130 34 (22.8) 28 (23.3) 0.98 (0.613 to 1.472) 1.08 (0.661 to 1.647) 0.396†  SBP <130 115 (77.2) 92 (76.7) Ref Ref DBP, mmHG mean (SD), n (%) 74.87 (10.09) 77.08 (13.84) 2.20 (−0.774 to 5.177) 0.146‡  DBP ≥85 28 (18.8) 27 (22.5) 0.84 (0.503 to 1.312) 0.52 (0.244 to 1.018) 0.057†  DBP <85 121 (81.2) 93 (77.5) Ref Ref Known with hypertension, n (%) 17 (11.4) 3 (2.5) U-ACR, mg/mmol mean (SD), n (%) 7.13 (45.91) 3.16 (6.73) −3.97 (−13.915 to 5.975) 0.432‡  U-ACR ≥3 12 (15.4) 21 (24.7) 0.62 (0.309 to 1.156) 0.49 (0.202 to 1.057) 0.071†  U-ACR <3 66 (84.6) 64 (75.3) Ref Ref hsCRP, mg/mL mean (SD), n (%) 11.3 (33.80) 7.51 (10.75) −3.77 (−9.572 to 2.027) 0.201‡  hsCRP ≥3 58 (38.9) 43 (35.8) 0.93 (0.647 to 1.253) 0.73 (0.449 to 1.094) 0.135†  hsCRP <3 91 (61.1) 76 (63.3) Ref Ref Unhealthy diet, n (%)  Yes 38 (25.5) 52 (43.3) 0.68 (0.584 to 0.858) 0.67 (0.56 to 0.86) 0.004†  No 111 (74.5) 68 (56.7) Ref Ref Exercise, min/wk, mean (SD), n (%) 68.72 (155.6) 25.49 (68.8) −43.23 (−71.64 to −14.83) 0.003‡  <30 min/wk 54 (37) 18 (15.8) 0.36 (0.2 to 0.626) 0.29 (0.15 to 0.564) <0.001†  >30 min/wk 92 (63) 96 (84.2) Ref Ref Smoking, n (%)  Current 5 (3.4) 1 (0.8) 0.23 (0.026 to 1.975) Not possible 0.812†  Past 11 (7.4) 3 (2.5) 0.32 (0.087 to 1.144) Not possible 0.08†  Never smoked 133 (89.3) 116 (96.7) Ref Ref Alcohol, n (%)  Yes 46 (30.9) 42 (35) 1.12 (0.801 to 1.485) 1.35 (0.932 to 1.776) 0.106†  No 103 (69.1) 78 (65) Ref Ref

Data are n (%).

*Some variables have a different N (WC N in study intervention = 148; U-ACR N study intervention = 78, N control arm = 85).

†RR, adjusted RR and 95% CI, and P value.

‡Mean (SD), mean difference (95% CI), and P value.

§5 current pregnancy.

‖0 current pregnancy.

Trig, triglycerides.

TABLE 4. - Difference Between Baseline and EOS CVD Risk Factors in the Intervention Arm Variable Baseline
Mean (SD) End-Of-Study
Mean (SD) Mean Difference (95% CI) P WC, cm
N = 144 90.2 (14.619) 95.5 (13.924) −5.32 (−6.738 to −3.912) <0.001 BMI, kg/m2
N = 144 29.6 (6.806) 31.5 (6.977) −1.91 (−2.346 to −1.474) <0.001 Trig, mmol/L
N = 148 1.004 (1.649) 0.84 (0.532) 0.16 (−0.092 to 0.422) 0.208 HDL, mmol/L
N = 148 1.31 (0.331) 1.40 (0.405) −0.09 (−0.157 to −0.034) 0.002 TC, mmol/L
N = 148 4.04 (0.782) 3.96 (0.856) 0.08 (−0.036 to 0.196) 0.176 LDL, mmol/L
N = 149 2.29 (0.759) 2.40 (0.725) −0.11 (−0.195 to −0.024) 0.013 BG, mmol/L
N = 147 4.54 (0.436) 4.62 (0.758) −0.08 (−0.188 to 0.031) 0.156 SBP, mmHG
N = 148 117 (13.114) 118 (12.129) −1.68 (−3.576 to 0.198) 0.079 DBP, mmHG
N = 148 74 (11.829) 75 (10.113) −1.25 (−2.926 to 0.426) 0.143 hsCRP, mg/mL
N = 148 7.17 (9.809) 11.34 (33.908) −4.17 (−9.794 to 1.439) 0.144 n/N (%) n/N (%) P * MetS  Yes 28/149 (18.8) 25/149 (16.8) 0.700  No 121/149 (81.2) 124/149 (83.2) Unhealthy diet  Yes 75/149 (50.3) 38/149 (25.5) <0.001  No 74/149 (49.7)