Relationship of ethnicity and CD4 Count with glucose metabolism among HIV patients on Highly-Active Antiretroviral Therapy (HAART)

HIV/AIDS affects U.S. minorities disproportionately, reflecting a change in the distribution of the disease by racial/ethnic groups since the beginning of the epidemic[1, 2]. The widespread use of highly-active antiretroviral therapy (HAART) has slowed the progression of HIV infection to AIDS and has reduced associated morbidity and mortality[3]. However, HIV patients on HAART are prone to metabolic abnormalities, including insulin resistance, lipodystrophy and diabetes (1). Lipodystrophy - characterized by peripheral fat loss, visceral fat increase, hypertriglyceridemia, and low HDL-C - is often associated with insulin resistance[4, 5], a risk factor for developing diabetes. Although the frequencies of impaired glucose tolerance (IGT) and diabetes have not been systematically evaluated in HIV patients of different ethnicities and geographic areas, the rates are likely to be higher than in the general population. An early study of HIV patients with lipodystrophy and abdominal fat accumulation reported a 35% prevalence of IGT and 8% prevalence of diabetes, compared to 5% and 0.5% respectively for matched non-HIV controls[4]. The high frequency of obesity (an independent risk factor for metabolic abnormalities and CVD) among HIV/HAART patients has an inimical effect on their immune reconstitution and body composition[6].

Diabetes and cardiometabolic syndrome are highly prevalent among ethnic minorities in non-HIV infected populations[7]. The relationship between HbA1c and fasting serum glucose is not homogenous across racial/ethnic groups: e.g., African-Americans and Hispanics have higher HbA1c than non-Hispanic Whites (NHWs) after adjusting for fasting glucose concentration, glucose area under the curve after oral glucose, insulin response and insulin resistance[8, 9]. Possible important interactions between ethnicity and specific aspects of glucose metabolism may be obscured in epidemiologic studies that utilize limited parameters (e.g., only fasting serum glucose, or HbA1c) to measure dysglycemia: e.g., significant differences have been noted in HbA1c levels between African-Americans and NHWs at elevated serum glucose levels[8]. Among HIV patients, elevated HbA1c has been associated with higher CD4 cell count, older age, and ethnicity[10]. Similar data from relatively small studies and heterogeneous HIV populations suggest that metabolic (including glycemic) abnormalities among HIV patients may also be affected by ethnicity[11, 12]. However, in the context of diabetes risk, the relationship of ethnicity and its interactions with metabolic parameters and the degree of immune reconstitution following HAART have not been investigated systematically in HIV patients.

The purpose of this study was to assess the relationship of ethnicity with several simultaneously measured glycemic parameters - HbA1c, fasting serum glucose and insulin, and post-challenge serum glucose and insulin levels - among hypertriglyceridemic HIV patients on stable HAART who participated in the Heart Positive study. We hypothesized that ethnicity and CD4 level on stable HAART would exert an influence on glycemic regulation in these hypertriglyceridemic but otherwise healthy HIV patients.

Heart Positive (Clinicaltrials.gov ID NCT00246376) was a large multiethnic study of healthy patients with HIV infection on stable HAART[13]. The primary outcomes have been reported[13]. The protocol was approved by the Institutional Review Boards of Baylor College of Medicine and the two recruitment centers. Written informed consent was obtained (in Spanish or English) from all subjects prior to their participation.

Hispanics were significantly younger, had shorter HIV infection duration, fewer years of HAART and lower systolic blood pressure, and reported a lower rate of smoking, drug use and family history of diabetes, than African-Americans or NHWs (based on Bonferroni Post-Hoc analysis). Furthermore, they had significantly lower weight, height and body cell mass than African-Americans or NHWs. African-Americans had significantly more fat mass than Hispanics or NHWs (P < 0.01). Both African-Americans and Hispanics had higher BMI and percent body fat than NHW’s. NHWs had higher resting VCO₂, VO₂ and REE than Hispanics or African-Americans (P < 0.001). There were no significant ethnic diffenences in frequency of PI use or in fasting plasma levels of lipids, adiponectin, hsCRP, leptin, and free fatty acids. or.

No significant differences were noted for frequency of use of first generation PI drugs (indinavir, nelfinavir, saquinavir) compared to more recent PI drugs (atazanavir, darunavir), or for use of thymidine analogue drugs, with respect to ethnic group or CD4 stratum. No differences were noted between the low and moderate-high CD4 groups in demographic / clinical / anthropometric or endocrine / hormonal / energy expenditure parameters except for a higher frequency of PI use and lower body cell mass among subjects with moderate-high CD4 count, higher testosterone level in the low CD4 group, and higher VO₂ with a trend towards higher VCO₂ in the low CD4 group.

Insulin sensitivity and glycemic parameters by ethnicity and CD4 cell countare presented in Table 1. The 2-hour glucose value in the OGTT differed significantly by ethnicity but not by CD4 strata. IGT (defined by 2-h glucose level >140 but <200 mg/dl) was observed in 32% of the subjects - these were significantly older than those with normal glucose tolerance (P < 0.001). Compared to either Hispanics or NHWs, African-Americans manifested greater impairment of glucose tolerance, as indicated by higher 2-hour glucose (P < 0.05) and HbA1c (P < 0.001) levels. Hispanics had higher insulin levels at 90 min and 120 min of the OGTT than NHW’s or African-Americans. No ethnic difference was noted in calculated measures of insulin resistance such as area under the curve (AUC) for insulin or the insulinogenic index.

Multivariate analysis (repeated measures ANOVA) assessed the effect of ethnicity on the OGTT glucose response curves after controlling for confounders (age, sex, race, obesity, fasting glucose, HOMA-IR, family history of diabetes, blood pressure, duration of HIV, and PI drug use) and also measured interactions of ethnicity with age, sex and CD4 strata. Ethnicity alone did not distinguish the glucose response curves in a statistically significant manner. Interaction effects of ethnicity*sex, and ethnicity*age were also not significant, but an interaction between ethnicity and CD4 strata was significant for the glucose response curves (P = 0.023). Interaction effects of ethnicity with CD4 count imply that the effect of ethnicity is modified by the CD4 count of the subjects. To elucidate how ethnicity affects the relationship between glycemic excursions and CD4 levels, we plotted post-challenge glucose data against ethnic groups for each CD4 stratum (Figure 1). In the moderate-high CD4 group, Hispanic subjects had the highest mean post-challenge glycemic excursion (Figure 1a), while in the low CD4 group, African-Americans had the highest mean post-challenge glycemic excursion (Figure 1b).

This is the first study to assess ethnic differences in glycemic parameters among HIV patients on a stable HAART regimen, specifically those with hypertriglyceridemia. The selection criteria for the Heart Positive study permitted us to investigate metabolic factors that differentiate glycemic control between ethnic groups of HIV patients without a history of diabetes, and to analyze a large data set of both fasting and post-prandial glucose responses. In bivariate analyses, African-Americans, Hispanics and NHW’s had similar fasting glucose levels, but African-Americans had higher HbA1c and 2-hour OGTT glucose values than Hispanics and NHWs. Significant ethnic differences noted in the multivariate models for 2-hour glucose and HbA1c suggested that ethnicity affects postprandial glycemic control in HIV patients on HAART. These effects persisted after adjusting for group differences in age, sex, HIV years, HAART years, smoking, obesity, fasting glucose, fasting insulin, and lipid levels. Overall, African-Americans and Hispanics had higher glucose values after oral glucose challenge as well as higher HbA1c levels than NHW’s, and an interaction between ethnicity and CD4 count modulated the glycemic response.

A previous study demonstrated that HIV-infected African American women have higher HbA1c levels than NHWs[10]. While controlling for ethnicity, age and CD4 count provided a reasonably accurate reflection of glycemic control in that study, the present results indicate that the greater impairment in glycemic control among HIV infected African-Americans is not due to unmeasured or suboptimally measured risk factors. A robust model that included all possible risk factors (demographic, anthropometric, clinical, and metabolic parameters) improved our confidence in the existence of ethnic disparities in specific glycemic parameters in this cohort who are at high risk for metabolic complications. It is possible that the effect of ethnicity in part “pre-dated” HIV infection or HAART therapy, since African American and Hispanics the general US population is known to have increased risk for impaired glucose tolerance and diabetes compared to NHW’s. The cross-sectional design of this study and absence of a HIV-negative control group did not permit us to estimate the specific effects of HIV / HAART on the risk of dysglycemia. However the interaction between ethnicity and CD4 level suggests that specific effects related to the immune status resulting from chronic HIV infection and its treatment magn ify the risk of dysglycemia in minority ethnic groups.

Previous studies have shown inconsistent associations of CD4 cell count with abnormal glucose levels in HIV patients. Low CD4 cell count was associated with IGT and diabetes in patients co-infected with HIV, hepatitis C and hepatitis B virus[16]. Estrada et al. reported that patients with HIV-associated lipodystrophy and IGT have lower CD4 counts than normoglycemic patients[17]. El-Sadr et al. found an inverse relationship between CD4 counts and insulin resistance in HIV patients[18], but no relationship between CD4 counts and fasting glucose concentrations. This finding is consistent with our data which showed CD4-related differences relate to post-challenge glucose levels, not to fasting glucose levels. Shikuma et al. reported no association between fasting glucose levels and CD4 levels but did not measure post-challenge glucose[19]. In contrast to our results demonstrating no differences in either fasting or post-challenge insulin levels, these investigators also noted that their small sample (7–8 subjects in the respective groups) of HIV patients with CD4 <200/cc had higher fasting insulin levels than those with CD4 >500/cc.

The design of this study makes it difficult to specify the reasons for the relationships between ethnicity, CD4 counts and glucose tolerance in these patients. However, the literature suggests potential mechanisms that should be investigated in future studies. Strictly speaking these results may be generalizable to only to HIV patients with hypertriglyceridemia, however such patients comprise a large proportion of those with HIV infection on various forms of antiretroviral therapy. Additional limitations include the fact that not all ART drug classes or single drugs that could possibly affect insulin sensitivity were included in the multivariate analyses. However, drugs of the protease inhibitor class (which are the most strongly associated with both insulin resistance and defects in glucose uptake) were taken into consideration (i.e., use of PI was controlled for in the analyses).

Abnormal glycemic responses following an oral glucose challenge could be due to insulin resistance or impaired insulin secretion. HOMA-IR, a marker of insulin resistance calculated from fasting values of glucose and insulin, did not differ between patients of the different CD4 strata, suggesting that an aspect of the glycemic defect might involve defective insulin response to an oral challenge. Consistent with this notion, AUC insulin following the oral challenge and the insulinogenic index (a measure of insulin response relative to the glucose surge) did not differ between the two CD4 strata, indicating that persons with both low and high CD4 cell counts may have inadequate insulin secretory response in the face of higher post-challenge glucose levels. It is possible that either low or high CD4 levels might reflect underlying immunologic mechanisms that could affect beta cell mass or function. These possibilities are supported by recent studies showing that, compared to healthy non-HIV persons, HIV patients have higher rates of IGT[20] without necessarily having greater degrees of insulin resistance[21].

While the CD4+ T cell level (low or moderate-high) achieved on a stable HAART regimen was not by itself associated with dysglycemia, ethnicity appeared to interact with low or moderate-high CD4 levels to affect glycemic response to an oral glucose challenge. Activated CD4 cells play a significant role in non-insulin mediated glucose disposal[22]. The association of dysglycemia with low CD4 levels could be due to impaired non-insulin mediated glucose disposal caused by a diminished “sink” of activated CD4 cells (normally equal in total mass to perhaps a third of the liver) or to a greater degree of chronic systemic inflammation. Association of defective glucose disposal with high CD4 counts might be due to post-HAART recovery of CD4 T cells that are unable to utilize glucose effectively because of low GLUT-1 or IL7-R expression or to the influence of genetic factors such as CCL3L1 or CCR5 polymorphisms known to be associated with the development of type 1 diabetes[23]. These findings will be important in formulating hypotheses for an immunologic or molecular basis of impaired glucose metabolism in HIV patients.

These findings have prominent implications for clinical practice. Ethnicity should be considered to affect risk for developing dysglycemia and Metabolic Syndrome among HIV patients. In the U.S., the disproportionately higher burden of HIV/AIDS coupled with the projected population growth among African Americans and Hispanics underscore the need for early screening and intervention as a cost-effective public health strategy[24, 25]. Consistent with non-HIV population data[26], we found that African-American ethnicity is associated with IGT, both independently and through an interaction with low CD4 count. Earlier studies investigating the association of ethnicity with IGT or insulin resistance in HIV patients have shown inconsistent results. NHW ethnicity, Hispanic ethnicity, and non-Caucasian race each have been shown to increase risk for insulin resistance or IGT[11]. However, El-Sadr et al.[18] actually found an inverse relationship between African-American ethnicity and insulin resistance in a large HIV cohort, while the Nutrition For Healthy Living study found no relation between ethnicity and insulin resistance in HIV patients[21]. These studies involved predominantly NHW’s[18], whereas our cohort had a mixed ethnic distribution resembling that of the overall US population.

Development of ethnically focused and culturally appropriate treatment approaches is highly relevant to clinical management of the growing HIV epidemic among minorities. The rate of African Americans living with HIV infection (1,202.2) is 8 times higher than that of NHWs (156.7) and 3 times higher than that of Hispanics (487.3) per 100,000 population[24]. HIV’s economic burden is predicted to grow markedly, based on the projected dramatic increase in the minority population as well as in HIV infections and complications associated with HAART[27]. In this context, the present findings in a large, multiethnic cohort of hypertriglyceridemic but otherwise apparently healthy HIV patients on stable HAART have important clinical implications for targeted therapy of HIV/HAART-associated glycemic complications.

Among hypertriglyceridemic HIV patients on HAART, African-Americans and Hispanics are at increased risk of developing diabetes. Ethnicity also interacts with CD4+ T cell count attained on stable HAART to affect post-challenge glycemic response. Clinicians should understand ethnic-specific factors in developing treatment strategies for dysglycemia, and should be aware of the need for vigilant screening and possibly earlier treatment for HIV-infected African-Americans with low CD4 count and Hispanics with high CD4 count following HAART.