A cross-sectional assessment of metabolic syndrome in HIV-infected people of low socio-economic status receiving antiretroviral therapy

Our cross-sectional analysis is part of a longitudinal program involving PLWH of predominantly low SES. Using ATPIII criteria, 33% of PLWH taking ART were estimated to have MetS at the time of their enrollment into “Healthy Living for Better Days,” an exercise training and nutrition education community program. The presence of MetS in almost one in three individuals was driven by a high rate of abdominal obesity and hypertension. Higher use of PIs, elevated BP and HbA1c, greater self-reported daily caloric intake and consumption of carbohydrates, sugar, and added sugar, and higher GL were recorded among the individuals with MetS, compared to those without it. Elevated HbA1c and high total sugar consumption were the strongest predictors and accounted for 30% of the occurrence of MetS.

Different definitions of MetS have been utilized to examine its incidence, and most studies have reported a rate of 11% to 26% among PLWH. The same rates have been comparable or slightly less in HIV-negative individuals. For example, the incidence of MetS in the NHANES dataset among adults 20 years of age and older was estimated to be 23% [3]. Overall, our PLWH cohort of 20 years of age and older had higher rates of MetS than previously reported in the general population. However, most of our participants were 40 to 59 years old, and their age-specific incidence of MetS was similar to the NHANES cohort (37% vs. 41% for males and 37% vs. 37% for females) [16]. It is still unclear if HIV infection per se or ART use has a dominant mechanistic role in the development of MetS. Similarly to one cohort from the northeastern US, we found no association between duration of HIV infection and MetS [7], while the information on duration of ART use among our participants was not available at baseline. However, it is worth noting that all of our subjects were receiving ART, with the overall prevalence of MetS in our cohort being higher than in other PLWH cohorts. The same rate is comparable (from 31% to 45%) to three younger PLWH Italian cohorts, where all study participants also had ART exposure [10-12].

The characteristics of MetS have been reported to differ between PLWH and the general population, and these differences could be attributed to HIV and/or its associated therapies. The most commonly achieved metabolic criteria for MetS in PLWH were hypertriglyceridemia and low HDL-C, while the least common criterion was increased WC [13]. The opposite was true for our cohort. The rates of hypertriglyceridemia and low HDL-C were among the lowest in our study. In regard to increased WC, HIV-associated lipoatrophy and loss of abdominal subcutaneous fat can actually lead to lower rates of central obesity [13]. Nonetheless, overall three out of four individuals and almost all individuals with MetS had high WC, making this trait the most common metabolic abnormality among our participants. The second most prevalent component of MetS in our study was hypertension. No specific ART has been associated with elevated BP, and underlying factors associated with hypertension, similar to the general population, are BMI and age [13]. The higher rate of obesity in our cohort could explain the increased rate of hypertension, both compared to the general population. Our findings are similar to the NHANES dataset [3], where the rates of abdominal obesity and hypertension, in addition to low HDL-C, were the most common features contributing to MetS. Therefore, we conclude that the overall incidence of MetS and the rates of its individual traits in our PLWH cohort resemble more the general population, rather than other PLWH cohorts.

Specific ART medications are known to affect individual components of MetS, and PI use is independently associated with metabolic abnormalities [17]. Protease inhibitors, a class of ART, are hypothesized to cause peripheral lipodystrophy, a syndrome that can lead to central fat accumulation, hyperlipidemia, and insulin resistance [18]. In our analysis, PI exposure was not an independent risk factor for MetS, but we did observe a higher prevalence of MetS and less favorable lipid and glucose profiles among the participants taking PIs, compared to the ones on non-PI based ART.

Of all lipid and non-lipid markers of CVD, the greatest difference, with a moderate trend toward significance, was observed for HDL-C. The individuals without MetS had on average 7 mg/dl higher HDL-C, compared to those with MetS. Knowing that for every 1 mg/dl increase in HDL-C, CVD risk is reduced by 2-3% [19], this near statistical significance should not be ignored as potentially clinically relevant. It is also worth noting that none of our participants received nevirapine, a non-nucleoside reverse transcriptase inhibitor drug known to increase HDL-C levels [20].

The American Diabetes Association (ADA) has recently recommended the use of HbA1c as one of the diagnostic tests for type 2 diabetes and also suggests a range of 5.7% to 6.4% as a category of increased risk for diabetes (i.e., pre-diabetes) [21]. Additionally, most epidemiological studies have shown that HbA1c, unlike FG, is an independent risk factor for CVD [21,22]. According to the ADA standards, the overall average HbA1c levels for both of our groups (5.7% and 6.7%) can be identified as pre-diabetes risk, while the MetS group, compared to those without MetS, may even possess a greater risk for cardiometabolic abnormalities. Hemoglobin A1c has rarely been examined in PLWH in relation to MetS, while in our cohort and among all traditional risk factors, including FG and homeostasis model assessment, it represents the strongest predictor for MetS. Thus, our findings suggest that HbA1c might be a useful diagnostic criterion for MetS in this patient population.

Many factors, such as a sedentary lifestyle and excessive caloric intake, contribute to the incidence of MetS [23]. We did not observe any differences between the groups on anthropometric measurements and levels of physical fitness, while caloric intake and four variables related to carbohydrate consumption (total carbohydrate, total sugar, added sugar, and GL) significantly differed between the groups. Total carbohydrate intake and GL have been associated with MetS in the general population [24,25], while (added) sugars have been linked with decreased intake of essential micronutrients [26] and are hypothesized to contribute to weight gain [27]. For a similar body size, individuals with MetS consumed on average more than 600 calories per day, than those without MetS.

A significant portion of excessive caloric intake in the MetS group came from carbohydrates and more specifically from the combination of naturally occurring sugar and added artificial sugar to food and beverages. A World Health Organization report from 2002 recommended that sugars be limited to less than 10% of total energy per day. Their new draft guidelines are proposing a reduction to below 5% of total daily energy intake for additional benefits [28]. The average intake of sugar for both of our groups is substantially greater (>20% of total daily energy intake) than the recommendations, contributing together with high levels of HbA1c, to the presence of MetS. Additionally, the Dietary Guidelines for Americans recommend limiting total intake of discretionary calories, including both added sugar and solid fats, to 5%-15% per day [29]. In our study, both groups consumed added sugar alone more than the recommended range for added sugar and solid fats combined. More specifically, the individuals with MetS consumed approximately 19% (equivalent to approximately 27 teaspoons) of daily calories from added sugar, compared to 15% (16 teaspoons) among the ones without MetS. In summary, our study indicates that unhealthy dietary habits consisting of higher intake of carbohydrates in the form of sugars and resulting in chronic hyperglycemia may contribute to a higher rate of MetS in our cohort.

Lastly, our program participants were of predominantly low SES, and two-thirds reported having only completed high school or less, earning less than $15,000 per year, and even more were unemployed. Low SES has been independently associated with chronic conditions and diseases including MetS [30], but the same finding was not observed in our cohort. However, due to a small number of our participants with an average or above average income, these results should be interpreted with caution. Future studies should explore the potential influence of (low) SES on MetS among PLWH.

Strengths of the current study included a program population with extensive information obtained on metabolic risk factors, including physical fitness levels and dietary habits as lifestyle characteristics highly associated with MetS. Additionally, selecting a population of PLWH who are predominantly minority and low SES provided an opportunity for them to participate in health-promotion activities not typically available as part of their conventional treatment plan. Whereas several other protocols utilized a 6–8 hour fast before blood sampling, which could overestimate TGs and MetS rates, we utilized a 12-hour fast, thus avoiding exaggerated TGs and ensuring that serum chylomicrons would be eliminated. If participants consumed any solid or liquid calories on the morning of blood draw, they were re-scheduled for the following day and told to adhere to the 12-hour fast.

Our results are hindered by several limitations. Our study design is observational and cross-sectional, which precludes causality between changes in risk factors and the rate of MetS. All of our investigators were equally trained in performing assessments. Regardless, it was not expected that all measurements (e.g., WC and BP) would always be conducted in a uniform manner. We utilized the Omron automatic oscillometric device and not a mercury sphygmomanometer, the gold standard in obtaining BP. Nonetheless, a recent study reported a high correlation between Omron and mercury measurements (SBP: r = 0.92 and DBP: r = 0.79) and low mean between-device differences (SBP: −1.6 mmHg and DBP: −0.6 mmHg, p < 0.05 for both) [31]. The same study reported that Omron underestimates high BP, which could only result in lower overall estimates of MetS in our analysis. In our research, dietary assessment was based on a short instrument and as such more prone to error in its estimate of true dietary intake. Lastly, we did not obtain HIV RNA levels, CD4 cell counts, and the duration of exposure to ART to further examine the association between HIV specific risk factors and MetS in our cohort.