The present cross-sectional study investigated whether MetS, a cluster of metabolic abnormalities, had a stronger impact on neurocognitive deficits among PLHIV than among HIV− persons. We found that (in a sample with average age of 50 years) MetS had an independent significant effect on global neurocognitive deficits among PLHIV but not among their HIV− counterparts. Among PLHIV, MetS was most strongly associated with the neurocognitive domains of learning, fine motor skills, and executive function. Also among PLHIV, diabetes and elevated triglycerides were the MetS components most strongly associated with increased global neurocognitive deficits.
The presence of a significant association between MetS and global neurocognitive deficits among PLHIV but not among HIV− persons indicates that HIV might worsen the impact of MetS on neurocognitive functioning. Although the mechanisms underlying the link between MetS and NCI are likely varied, at least some of these mechanisms, such as systemic inflammation and lower blood–brain barrier integrity, may predispose PLHIV to a higher risk of neurocognitive deficits. The rate of MetS was much lower among HIV− persons than among PLHIV in the present study, which might have affected our ability to find a significant effect in the HIV− group. Yet, HIV serostatus groups were of comparable age, underscoring the importance of MetS in HIV-associated NCI and the potential for even larger effects in older samples.
Interestingly, among PLHIV, the association between MetS and GDS remained significant after considering the impact of significant covariates. This suggests that the link between MetS and worse neurocognition in this group is likely unexplained by other factors known to affect NCI in HIV, including HIV disease burden. It also highlights the importance of treating MetS in PLHIV to maintain or improve neurocognition.
Although the effect of MetS on individual neurocognitive domains was generally small to medium in this sample of middle- to older-aged PLHIV, MetS was most notably associated with worse learning, executive function, and fine motor skills (Figure 2). We did not assess for peripheral neuropathy in the present study, which might have confounded the association between MetS and fine motor skills. Of note, our findings linking MetS to executive functioning and learning in PLHIV are consistent with previous work in HIV− individuals.[15–18] However, MetS has also been linked with other neurocognitive domains in the HIV− population (ie, recall, processing speed), which had very small effects in our sample of PLHIV. This suggests that MetS may differentially affect specific neurocognitive domains in the context of HIV infection.
Among PLHIV, diabetes and increased triglycerides were the MetS components most strongly associated with increased global neurocognitive deficits, with medium effect sizes. Although only the association between a diagnosis of diabetes mellitus and GDS remained statistically significant when accounting for relevant covariates, the separate effect of both diabetes and triglycerides on GDS was comparable with that of the overall MetS as a cluster of metabolic abnormalities. Other individual components of MetS (ie, elevated waist circumference, reduced HDL-C, and elevated blood pressure) were not significantly associated with GDS. Our findings are consistent with previous studies that demonstrated significant associations between diabetes mellitus and worse cognitive performance in HIV[25,26] and indicate that this might be an important risk factor for HIV-associated NCI. Our results are also in-line with previous findings showing that other components of MetS, such as hypertension and HDL-C, might not be as important for neurocognition among PLHIV. Contrary to our findings, a previous cross-sectional study showed that greater waist circumference increased the risk for NCI among PLHIV in the context of considering the impact of BMI and other variables. Why these studies have disparate findings is unclear. Measurements of waist circumference and neurocognition followed similar procedures in both studies and thus are less likely to have played a notable role. Small demographic differences between participants in the 2 studies in mean age (approximately 4 years) and education (approximately 1 year) also seem unlikely to explain the difference in results. Future studies with larger and more diverse samples in terms of sex and race/ethnicity might help determine whether these factors modify the association of waist circumference and neurocognitive deficits in HIV.
Whereas investigating rates of MetS in HIV infection was not the focus of the present study, PLHIV in the current sample had nearly twice the rate of MetS than the HIV− comparison group. This increased prevalence of MetS in HIV is consistent with previous studies and with findings linking worse HIV disease to MetS.[12,13] Higher HIV viral load has been linked to MetS, possibly due to the increased inflammation and immune activation resulting from poorly controlled HIV infection. Alterations in lipid and glucose metabolism that result from undergoing various ART modalities, particularly protease inhibitors,[43–45] might also play a role in increasing the prevalence of MetS in HIV. Yet, findings linking ART use to HIV have been mixed.[4,12–14,44] Furthermore, a population-based study in the United States did not find an increased risk for MetS among PLHIV. Importantly, this previous study included PLHIV who were evaluated between 2000 and 2003 and had an average age of almost 10 years younger than participants in the present study. HIV disease has been transformed considerably since these earlier days of the US HIV epidemic, which might explain some of the disparate findings.
The present study has a number of limitations. First, the cross-sectional design prevents us from ascribing directionality to our findings. Longitudinal observational studies and intervention research will be needed to test potential causal relationships. Second, the relatively small sample sizes and the demographic characteristics of our cohort might have limited our power to detect associations. For example, despite previous findings showing a robust association between MetS and neurocognitive functioning,[5–7] we found that among HIV− participants, MetS was unrelated to global neurocognitive deficits. This null finding may have been a consequence of the young age of participants relative to those from previous studies, who tended to be at least 60 years old. As MetS is typically considered a condition of older age, too few of our HIV− participants may have developed MetS (only 21% compared with 39% of PLHIV). Furthermore, as expected, neurocognitive deficits were significantly higher among PLHIV than among HIV− individuals. These factors likely limited our ability to detect significant associations between MetS and neurocognitive performance in the HIV−group. Due to a small number of women living with HIV and relatively low numbers of specific racial/ethnic minority participants in our study, we were unable to examine how these other potentially important demographic factors might affect the relationship between MetS and NCI. This seems particularly important in light of previously established differences on MetS by sex and race/ethnicity[11,46,47] and higher rates of NCI among Hispanics living with HIV. Future studies should use a larger, more diverse sample to examine the effects of age, sex, and race/ethnicity.
Overall, the present findings underscore the importance of considering cardiometabolic disease as a potential risk factor for poor neurocognitive outcomes among PLHIV. Additional longitudinal studies will best determine whether there is a causal role of MetS on neurocognitive deficits. Intervention research, including investigations of whether treatment for components of the MetS might affect NCI, would help identify treatment approaches that might be best suited to lessen the development of MetS and its potential impact on NCI among PLHIV.
J Acquir Immune Defic Syndr. 2019;81(1):95-101. © 2019 Lippincott Williams & Wilkins