Introduction

In the era of effective antiretroviral therapy (ART), HIV-infected adults can live near normal lifespans but face high rates of metabolic disease stemming from both traditional (Western diet, sedentary lifestyle) and HIV-/ART-related contributors (chronic inflammation and immune activation, gut microbiome disturbances, drug toxicities). Obesity and visceral adiposity are common in treated HIV infection, and have both traditional and HIV-/ART-associated contributors. As in the general population, excess adiposity is associated with numerous metabolic and inflammatory consequences. Although the mechanisms by which HIV and ART contribute to changes in fat quality and quantity are incompletely elucidated, an understanding of the importance of fat health is emerging. Here we review recent advances in our understanding of the prevalence, incidence, pathophysiology, and consequences of excess adiposity in treated HIV infection.

Burden of Obesity and Visceral Adiposity

Obesity is traditionally defined as a body mass index (BMI) > 30 kg/m2. However, in persons with low muscle mass, excess adiposity may exist within the “normal” BMI range of 18.5–24.9 kg/m2. For this reason, some authorities have advocated using body fat > 25% for men or > 33% for women to define obesity [1], although these cutoffs have not been validated in the setting of HIV infection. Additionally, HIV-infected persons may have increased BMI-to-visceral adipose tissue (VAT) ratios [2,3,4], creating another scenario where traditional BMI guidelines may underestimate adiposity in this population. Despite this, up to two thirds of HIV-infected adults are classified as overweight or obese by standard BMI criteria in recent cohorts [5,6,7,8,9], reflecting global HIV obesity rates.

Modern ART initiation is often associated with weight gain. In an AIDS Clinical Trial Group (ACTG) study of ART initiation in resource-diverse settings (A5175), more than 25% of participants were classified as overweight or obese at entry, and approximately 40% of participants were overweight or obese by week 144 [8]. Some weight gain following ART initiation may be attributable to a “return to health” phenomenon; however, excessive weight gain can occur, with persons with the highest pre-ART HIV-1 RNA or lowest CD4+ T lymphocyte counts at risk for greater weight gain [10, 11•]. Further exemplifying the fact that weight gain can represent differential effects depending on the host, weight gain among underweight persons has been associated with a decline in circulating high-sensitivity C-reactive protein levels [12] and improved survival [13], whereas weight gain among overweight or obese individuals has been associated with significant increases in circulating levels of the monocyte activation marker soluble CD14 [12], no mortality benefit [13] and a ≥ 67% prevalence of multimorbidity [14].

Importantly, many studies have described weight gain following ART initiation without additional details regarding the type of tissue involved, and lean mass and fat mass may both increase with ART initiation [11•, 15•]. However, other studies have specifically demonstrated increases in fat following ART initiation. For example, ACTG study A5260, which randomized participants to raltegravir, ritonavir-boosted darunavir, or ritonavir-boosted atazanavir, each with a backbone of tenofovir disoproxil fumarate/emtricitabine, demonstrated a mean computed tomography-quantified VAT gain of 26% after 96 weeks of ART that did not vary significantly between randomization arms [11•]. However, in subset analyses using waist circumference as a surrogate for VAT (Pearson correlation between waist circumference and VAT = 0.52, p < 0.0001 [16]), women appeared to have greater VAT gain on raltegravir vs. protease inhibitors than men [17], suggesting sex differences in drug effects may exist.

Additionally, Grant et al. examined longer-term changes in body composition in HIV-infected adults on ART, and demonstrated that after an initial 96 weeks of ART, dual x-ray absorptiometry-quantified trunk fat gains continued at a slower rate than in the first 96 weeks of therapy, but at faster rates than HIV-uninfected controls [15•]. As VAT gains of as little as 5% have been associated with increased risk for the metabolic syndrome [18] and central fat accumulation has been associated with short-term mortality risk in HIV infection [19], these size of these VAT changes are highly clinically significant.

Quality and Quantity

Given associations between obesity and immunometabolic disturbances in both the general population and HIV-infected persons, significant emphasis has been placed on the impact of fat quantity to cardiometabolic disease and the inflammatory milieu in treated HIV infection. Fat quality is less well understood, but is likely at least as important as quantity. An example of this is metabolically healthy obesity, in which persons have BMI ≥ 30 kg/m2 without overt cardiometabolic disease. Indeed, in the general population, metabolically healthy obesity has been associated with less VAT and systemic inflammation, more favorable immune cell profiles, and higher fat utilization than the metabolically unhealthy obese [20,21,22,23]. However, other (but not all) data suggests that this state is not completely benign, and can be associated with increased risk of progression to the metabolic syndrome and/or progression of cardiovascular disease (CVD) [24, 25]. Conflicting data is further complicated by the use of varying populations and definitions to define the metabolically healthy but overweight or obese phenotype. Of note, metabolically healthy obesity has recently been documented in HIV-infected men at a similar prevalence to that observed in HIV-uninfected men [26], and studies of whether metabolically healthy obesity differs in HIV-infected vs. HIV-uninfected persons are underway.

Why some persons remain metabolically healthy despite obesity remains unknown, but fat function likely plays a major role: Normal, healthy adipocytes are small, well differentiated, and contain a modest lipid droplet. During fat gain, adipocytes either increase in number (hyperplasia) or volume (hypertrophy). While hyperplastic adipocytes maintain normal functioning, large, hypertrophied adipocytes become hypoxic, expand their lipid droplet, and recruit pro-inflammatory immune cells, primarily activated macrophages [27,28,29].

Activated macrophages stimulate a local, pro-inflammatory type 1 immunologic response and lose the ability to store iron, which leads to iron deposition in the adipose tissue and subsequent reactive oxygen species production and mitochondrial dysfunction [30]. Adipocyte hypertrophy also suppresses adiponectin production, exacerbating the pro-inflammatory environment. Ultimately, increased transforming growth factor-β production [31] triggers pro-fibrotic processes in an attempt to limit further adipocyte hypertrophy [32]. However, when caloric excess persists and fibrosis limits adipocyte expansion, ectopic fat deposition occurs in sites such as the liver and skeletal muscle. Ectopic fat deposition is associated with additional inflammation and metabolic dysregulation [33, 34], including the development of insulin resistance and CVD [35, 36]. Thus, maintaining adipose tissue function, or quality, is critical to preventing immunometabolic consequences of fat dysfunction independent of fat quantity.

Much of the data on adipose tissue function in the general population is derived from the setting of obesity. Given that both HIV and ART are associated with adipose tissue disturbances (Fig. 1 [37]), extrapolation of these data to HIV-infected persons must proceed with caution. Much of the available data on adipose tissue dysfunction in HIV-infected persons is in the setting of lipodystrophy. While lipodystrophy in treated HIV infection still occurs, obesity prevalence is increasing, and some persons may experience an overlap syndrome with components of both obesity and lipodystrophy. While central lipohypertrophy is associated with increased adipose tissue inflammation and apoptosis [38], further research is needed to understand the intersections of HIV-, ART-, and obesity-induced adipose tissue disturbances, and to define whether these intersections may lead to differential clinical consequences in treated HIV infection.

Fig. 1
figure 1

Effects of HIV infection, ART, and obesity on adipose tissue biology. Reproduced with permission from J Infect Dis. 2013;208(8):1194–1201. IL-6, interleukin-6; MCP-1, macrophage chemotactic protein-1; MIP-1α, macrophage inflammatory protein-1α; mtDNA, mitochondrial DNA; PPAR-γ, peroxisome proliferator-activated receptor-γ; TNF-α, tumor necrosis factor-α

Notably, fibrosis of the subcutaneous adipose tissue in nonobese, HIV-infected adults has recently been documented, accompanied by the finding that adipose tissue fibrosis improves with continued suppressive ART [39]. While additional study is needed, this finding suggests that a pro-fibrotic, pro-inflammatory stimulus in the abdominal adipose tissue of HIV-infected persons exists that is not attributable to ART alone. Whether this stimulus is HIV itself [40••, 41•, 42••], microbial products resulting from gut barrier disruption in HIV infection [43, 44] or other currently unidentified triggers remains to be seen. However, these provocative data suggest that adipose tissue dysfunction and its associated metabolic disturbances may be related to a chronic inflammatory stimulus within the adipose tissue, and that continued ART allows for suppression of the inflammatory stimulus such that pro-fibrotic pathways can be suppressed and wound healing can occur. Additional research is needed to determine whether reductions in adipose tissue fibrosis in HIV-infected persons on suppressive ART will translate into improved adipocyte function and clinical benefit.

Consequences of Obesity and Visceral Adiposity

Obesity and visceral adiposity are integral players in the development of multiple non-AIDS comorbid disease states, including CVD and liver disease, which are now leading causes of death among HIV-infected persons [45, 46]. The development of comorbid disease in HIV-infected persons is believed to have both traditional and HIV-/ART-specific contributors: Obesity, and particularly VAT accumulation, are associated with systemic [12, 47] and adipose tissue inflammation [48], insulin resistance, dyslipidemia [49], and increased oxidative stress [50]. At the same time, adipose tissue is a potential HIV reservoir, leading to recruitment of T lymphocytes and macrophages into adipose tissue [40••]. HIV also alters adipocyte differentiation [41•, 51, 52]. One proposed mechanisms for this is inhibition of peroxisome proliferator-activated receptor-γ (a regulator of adipogenesis, lipogenesis, insulin sensitivity, and normal cytokine/adipokine expression) target gene expression and activation of glucocorticoid target gene expression by the HIV protein Vpr, leading to accelerated lipolysis, increased macrophage infiltration into adipose tissue, diminished white adipose tissue quantity and hepatic steatosis [41•]. The HIV Tat protein may also impair adipogenesis and promote adipose tissue inflammation [52].

HIV-associated chronic inflammation and immune activation may play a role in the development of visceral adiposity, with circulating CD8+ T lymphocyte activation linked to VAT accumulation [53]. Activation of the renin-angiotensin system has been independently associated with treated HIV infection and VAT accumulation [54, 55], and appears to independently predict insulin resistance [56•]. Similar to obesity, HIV-associated hormonal imbalances (e.g., hypogonadism, growth hormone deficiency) and alteration of the gut microbiome may play adjunctive roles [57, 58]. Finally, aging is associated with physiologic central fat redistribution, adipocyte senescence, and chronic inflammation [59], which may be enhanced in HIV-infected individuals. Whether the combination of aging, HIV infection, and VAT accumulation/obesity has synergistic or additive effects in this population is incompletely understood, however.

Cardiovascular Disease

Obesity and visceral adiposity are documented risk factors for CVD and diabetes mellitus (a CVD risk equivalent) in HIV-infected adults [60, 61], and VAT, intrahepatic fat, and epicardial fat are all associated with CVD independent of traditional CVD risk factors in this population [36, 61]. Additionally, HIV infection has been associated with increased type 2 diabetes mellitus [62] and CVD [63, 64] risk. Whether obesity and HIV infection are additive or synergistic to CVD risk, however, is not fully understood. In a recent study of HIV-infected participants, obesity was associated with reduced insulin resistance and greater systemic inflammation but not greater carotid intima media thickness or greater impairment of arterial flow-mediated dilatation compared to normal weight, HIV-infected persons on identical ART regimens [65]. In another study of older HIV-infected adults with traditional cardiovascular risk factors on suppressive ART, high rates of undetectable circulating endothelial progenitor cell levels were observed (suggesting markedly reduced vascular reparative capacity) that did not vary by BMI [66].

Fatty Liver Disease

Approximately 30–40% of HIV-infected adults are estimated to have nonalcoholic fatty liver disease (NAFLD, defined as ≥ 5% hepatic steatosis without other demonstrable causes) [67,68,69], and this prevalence may increase substantially among patients with elevated transaminase levels [70]. While NAFLD may be associated with progressive liver disease in the forms of nonalcoholic steatohepatitis (NASH), hepatic fibrosis, and, ultimately, cirrhosis, liver failure, and hepatocellular carcinoma [71], CVD accounts for most of the excess morbidity and mortality associated with NAFLD [72, 73]. This association with morbidity and mortality is independent of traditional CVD risk factors [74], but is tightly linked to excess adiposity and its consequences. In fact, 80–90% of adults with NAFLD have generalized obesity, visceral adiposity, metabolic syndrome or type 2 diabetes [67]. Interestingly, intrahepatic triglyceride accumulation, the root cause of NAFLD, is more closely linked to metabolic complications than VAT quantity [75, 76].

NAFLD may have unique origins in HIV infection. Although the exact mechanisms are not well understood, there may be several contributing factors not common among HIV-uninfected persons. First, HIV infection is characterized by persistent inflammation and immune activation [77], which (1) could help explain higher rates of NASH and liver disease severity in HIV infection (63 vs. 37% in HIV-uninfected) [78] and (2) contributes to greater insulin resistance, furthering metabolic dysregulation in both adipose tissue and the liver [79, 80]. In addition to traditional risk factors (older age, sedentary lifestyle), HIV-/ART-specific factors (dyslipidemia, microbial translocation, mitochondrial dysfunction) likely contribute. As such, both traditional and HIV-/ART-specific NAFLD contributors should be considered when developing therapeutics.

Given the high prevalence of NAFLD in HIV infection, NAFLD’s associations with CVD, the independent association of HIV with CVD [81] and increasingly high rates of traditional CVD risk factors in HIV-infected persons on ART, HIV-infected adults with NAFLD are primed for adverse cardiovascular outcomes and aggressive measures should likely be taken to prevent and treat NAFLD in this population. However, no standard of care for NAFLD exists in HIV infection or in the general population beyond diet and exercise recommendations. The thiazolidinedione pioglitazone [82] and growth hormone-releasing factor analog tesamorelin [83] have shown promise for the treatment of NAFLD in HIV/hepatitis C virus coinfected and HIV monoinfected adults, respectively, but have not yet been recommended for use. While a large number of agents are being developed for NAFLD and NASH treatment [84], future research efforts to define how the pathophysiology of NAFLD may differ in HIV infection are needed to allow for the development of targeted NAFLD therapies for this population.

Cognitive Decline

In HIV-infected persons, poorer neurocognitive function has been associated with both increased waist circumference (a marker of visceral adiposity [85]) [86] and obesity [87]. Among a subset of middle-aged, HIV-infected, and at risk HIV-uninfected women in the Womens Interagency HIV Study, higher leptin levels (indicative of higher adiposity) correlated strongly with poorer neurocognitive testing performance [88]. Similarly, in the Multicenter AIDS Cohort Study, VAT was strongly associated with regional brain atrophy (which precedes neurocognitive decline), irrespective of HIV serostatus [89•]. As such, mounting evidence suggests that, similar to middle-aged persons in the general population, obesity and visceral adiposity may have detrimental effects on cognition among HIV-infected persons. Whether these relationships persist into older age and/or weight loss can improve cognitive function in HIV-infected persons requires further study.

Functional Decline

Obesity and the metabolic syndrome are established risk factors for the development of physical function impairment or frailty among middle-aged or older HIV-infected adults [90,91,92,93]. Obesity has also been associated with fall risk in HIV-infected women [94]. The exact mechanisms underlying these relationships are incompletely understood, but may be related to adipocytokine imbalances [95] and/or chronic inflammation and immune activation [96, 97]. Complicating this fact is the observation of faster rates of functional decline and high rates of sarcopenia among HIV-infected vs. HIV-uninfected persons [15•, 98,99,100,101], although sex differences may exist. As the obesity and aging epidemics in HIV-infected persons collide, understanding the mechanisms by which obesity and visceral adiposity contribute to functional decline will play an important role in maximizing physical function and preventing further decline in this population.

Interventions

Given the potential consequences of obesity and visceral adiposity in HIV infection, prevention and treatment of these states, including HIV- and ART-specific contributors, are important but incompletely understood components of HIV care. A complete review of therapeutic options for obesity and visceral adiposity in treated HIV infection is beyond the scope of this article, although a consensus guideline has recent been published [102••]. To date, data on the efficacy of weight loss and dietary interventions in HIV are mixed, although structured exercise with or without dietary intervention reduces abdominal obesity in most studies [103,104,105,106,107], and disproportionate SAT loss has not been observed, eliminating fear of worsening of lipoatrophy in persons with mixed lipodystrophy. Based on available data, ≥ 30 min of moderate-intensity physical activity most days of the week plus reduction of caloric intake at least 500 kcal/day below usual intake is generally recommended to attain and sustain significant (≥ 5%) weight loss [102••, 108]. While weight loss may be effective, recent data suggests that HIV-related stigma may prevent participation in traditional weight management programs [109], suggesting the possible need for programs specifically targeting HIV-infected persons.

The role of ART in weight management is evolving, as initiation of most preferred ART regimens (whether non-nucleoside reverse transcriptase inhibitor-, protease inhibitor- or integrase strand transfer inhibitor-based) is associated with gain of both subcutaneous adipose tissue and VAT, and data supporting ART switches to improve regional or generalized adiposity are lacking [102••].

Tesamorelin is the only FDA-approved intervention to reduce visceral fat in HIV infection, and no pharmacologic weight loss interventions for generalized obesity [110] are specifically approved in HIV infection. As such, additional studies in HIV-infected populations are needed before specific pharmacologic therapies for fat loss can be recommended. However, as in the general population, persons with BMI ≥ 27 kg/m2 with comorbidity or BMI > 30 kg/m2 without comorbidity are likely to receive benefits from weight loss [102••], whether it be pharmacologic, diet and exercise-based, and/or through behavioral modification.

Conclusions

Obesity and visceral adiposity in HIV-infected persons are frequent and have important physiologic consequences that contribute to morbidity and mortality. The pathophysiology of these states, while overlapping with that of the general population, likely includes HIV- and ART-specific risk factors. Further research is needed to define the pathophysiology of and develop interventions for obesity and lipohypertrophy in HIV-infected persons, with the ultimate goal of preventing morbidity and mortality in this vulnerable population.