Lipoatrophy

LA, characterised by loss of subcutaneous fat⁽Reference Saint-Marc, Partisani and Poizot-Martin¹⁰⁰⁾, is distinctly different from the traditional HIV wasting syndrome, characterised by a disproportionate decrease in lean body mass⁽Reference Grinspoon, Corcoran and Miller¹⁰¹⁾. LA, as a side effect of ART, is seen mainly in the face (facial LA) and the extremities (peripheral LA)⁽Reference Saint-Marc, Partisani and Poizot-Martin⁶³⁾. Fat wasting of the face usually presents as malar or temporal wasting⁽Reference Omolayo and Sealy²²⁾. Peripheral fat wasting typically occurs in the arms, shoulders, buttocks and legs⁽Reference Engelson, Kotler and Tan¹⁰²⁾. The latter type of fat wasting is often accompanied by prominent superficial veins, which contribute to the emaciated appearance observed in these individuals⁽Reference Bergersen, Sandvik and Ellingsen¹⁰³⁾.

Initial reports attributed the development of LA to PI⁽Reference Carr, Samaras and Burton²⁰⁾; however, it is now known that the use of NRTI such as d4T is more strongly linked with its development⁽Reference Seminari, Tinelli and Minoli^61, Reference Nolan, Hammond and James¹⁰⁴⁾. Some NRTI combinations, such as d4T and didanosine (ddI), are contraindicated as a result of their severe lipoatrophic side effects⁽Reference Grinspoon and Carr¹⁰⁵⁾. In addition to type of ART, a number of other risk factors for LA have been identified including older age⁽Reference Grinspoon and Carr¹⁰⁵⁾, a decrease in BMI before ART⁽Reference Lichtenstein, Delaney and Armon⁶⁹⁾, white race⁽Reference Lichtenstein, Delaney and Armon⁶⁹⁾, use of PI for greater than 2 years⁽Reference Carr, Samaras and Thorisdottir⁹⁸⁾, and factors relating to disease progression including lower CD4 cell count⁽Reference Lichtenstein, Delaney and Armon⁶⁹⁾, duration and severity of HIV infection⁽Reference Lichtenstein, Ward and Moorman^65, Reference Jacobson, Knox and Spiegelman¹⁰⁶⁾ and prior diagnosis of AIDS⁽Reference Lichtenstein, Ward and Moorman⁶⁵⁾.

Lipohypertrophy

LH is characterised by adipose tissue accumulation mainly in the intra-abdominal (‘Crix belly’)⁽Reference Viraben and Aquilina^19, Reference Carr, Samaras and Burton^20, Reference Engelson, Kotler and Tan^102, Reference Miller, Jones and Yanovski^107, Reference Dinges, Chen and Snell¹⁰⁸⁾ and dorsocervical (‘buffalo hump’) regions⁽Reference Hengel, Watts and Lennox^17, Reference Roth, Kravcik and Angel^18, Reference Lo, Mulligan and Tai¹⁰⁹⁾. Other characteristic features of LH include breast enlargement, observed in both males and females⁽Reference Mutimura, Stewart and Rheeder^59, Reference Paton, Earnest and Ng^68, Reference Goujard, Boufassa and Deveau^77, Reference Bernasconi, Boubaker and Junghans^80, Reference Savès, Raffi and Capeau⁸¹⁾, accumulation of adipose tissue on the anterior region of the neck⁽Reference Palella, Chmiel and Riddler¹¹⁰⁾, side of the neck⁽Reference Mutimura, Stewart and Rheeder^59, Reference Paton, Earnest and Ng⁶⁸⁾, under the axillae⁽Reference Palella, Chmiel and Riddler¹¹⁰⁾ and in the suprapubic region⁽Reference Guaraldi, Orlando and Squillace¹¹¹⁾, and localised or generalised lipomas⁽Reference Miller, Carr and Emery⁵⁶⁾. LH is distinct from simple visceral fat accumulation, as it is associated with a decrease, rather than an increase, in subcutaneous fat⁽Reference Gkrania-Klotsas and Klotsas^90, Reference Dinges, Chen and Snell¹⁰⁸⁾. It is worth noting that abdominal LH is the most commonly identified lipohypertrophic change in HALS patients⁽Reference Boufassa, Lascaux and Meyer^52, Reference Miller, Carr and Emery^56, Reference Puttawong, Prasithsirikul and Vadcharavivad^57, Reference Mutimura, Stewart and Rheeder^59, Reference Paton, Earnest and Ng^68, Reference Goujard, Boufassa and Deveau^77–Reference Savès, Raffi and Capeau^81, Reference Engelson, Kotler and Tan¹⁰²⁾.

Risk factors associated with the development of LH in the context of HIV and ART include age, female sex, having a BMI of greater than 25 kg/m²⁽Reference Manfredi, Calza and Chiodo²⁶⁰⁾, and having a low CD4 cell count⁽Reference Lichtenstein, Ward and Moorman^65, Reference Lichtenstein, Delaney and Armon⁶⁹⁾. The type of ART has also been shown to play a role in the pathogenesis of LH. Jacobson et al. ⁽Reference Jacobson, Tang and Spiegelman⁴⁶⁾ demonstrated that LH was observed in both patients who have and have not been exposed to PI, indicating that PI are not the only cause of LH. Thymidine analogues in particular have been shown to increase the risk of developing LH⁽Reference van Griensven, De Naeyer and Mushi⁷¹⁾. Novel drugs, such as the peptidic HIV-1 fusion inhibitor enfuvirtide, have also recently been implicated in the development of LH⁽Reference Cooper, Cordery and Reiss¹¹²⁾. In addition to the type of ART, a longer duration of treatment has been associated with an increased risk of developing LH⁽Reference Martínez, Mocroft and García-Viejo⁷⁹⁾.

Dyslipidaemia

Before the advent of ART, evidence suggested that HIV infection itself caused abnormalities of blood lipids⁽Reference Grunfeld, Kotler and Hamadeh^113, Reference Grunfeld, Kotler and Shigenaga¹¹⁴⁾. One study investigating lipid abnormalities associated with seroconversion in men found that HIV infection was associated with a reduction in total cholesterol, LDL and HDL⁽Reference Riddler, Smit and Cole¹¹⁵⁾. Subsequent initiation of ART in the same subjects led to a significant increase in total cholesterol and LDL concentrations from baseline to follow-up, confirming the role of ART in the pathogenesis of dyslipidaemia in HIV⁽Reference Riddler, Smit and Cole¹¹⁵⁾.

The prevalence of lipid disorders in HIV-infected individuals treated with ART has been shown to vary from 24 to 72 %⁽Reference Chêne, Angelini and Cotte^53, Reference Thiébaut, Daucourt and Mercié^64, Reference Pujari, Dravid and Naik^70, Reference Zannou, Denoeud and Lacombe^75, Reference Elgalib, Aboud and Kulasegaram^88, Reference Tomažič, Silič and Karner^116, Reference Lesi, Soyebi and Eboh¹¹⁷⁾. Characteristic lipid abnormalities associated with HALS include elevated total cholesterol and LDL, elevated TAG⁽Reference Carr, Samaras and Burton²⁰⁾ and reduced HDL⁽Reference Friis-Møller, Weber and Reiss¹¹⁸⁾. Early studies attributed the development of dyslipidaemia to PI therapy⁽Reference Carr, Samaras and Chisholm^119, Reference Periard, Telenti and Sudre¹²⁰⁾. Subsequent studies have, however, shown that both NRTI and NNRTI are involved in the development of lipid abnormalities in HIV⁽Reference van Leth, Phanuphak and Stroes^121, Reference Jones, Sawleshwarkar and Michailidis¹²²⁾. Furthermore, both in vitro and in vivo studies have demonstrated an association between cART and the development of more pronounced lipid abnormalities⁽Reference Friis-Møller, Weber and Reiss^118, Reference Kosmiski, Miller and Klemm¹²³⁾. A recent UK study found that impaired postprandial TAG clearance in HIV patients receiving ART was exacerbated by a combination of NRTI and PI⁽Reference Ware, Jackson and Wootton¹²⁴⁾. A recent retrospective cohort study from Brazil found that PI increased serum TAG but not total cholesterol concentrations in 102 HIV-infected patients⁽Reference Monnerat, Cerutti Junior and Caniçali¹²⁵⁾. In the same study NNRTI were associated with an increase in total cholesterol with no significant effect on TAG levels. Similarly, Walmsley et al. ⁽Reference Walmsley, Cheung and Fantus⁷²⁾ in their prospective cohort study of HIV patients found that after 12 months of treatment with NNRTI, only total cholesterol concentrations increased significantly. Results pertaining to the duration of ART and risk of lipodystrophy are inconsistent, with some showing that increased duration increases risk of dyslipidaemia⁽Reference Savès, Raffi and Capeau^81, Reference Friis-Møller, Weber and Reiss¹¹⁸⁾, while other studies have shown no effect of duration on the risk of dyslipidaemia⁽Reference Mondy, Overton and Grubb³²⁾.

A number of factors have been identified which are associated with an increased risk of dyslipidaemia in patients receiving ART. Similar to the general population, dyslipidaemia in HIV has been shown to occur to a greater extent in female patients⁽Reference Pernerstorfer-Schoen, Jilma and Perschler¹²⁶⁾. Although African-Americans in the general population have been shown to have a lower prevalence of hypertriacylglycerolaemia⁽Reference Gaillard, Schuster and Osei¹²⁷⁾, Foulkes et al. ⁽Reference Foulkes, Wohl and Frank¹²⁸⁾ found that exposure to PI induced the greatest increase in TAG concentrations in black compared with white and Hispanic populations. This may indicate a role for race/ethnicity in increasing the risk of dyslipidaemia in HIV. It is important to note, however, that this study had, according to the authors, limited power, making it difficult to detect small interaction effects within these racial/ethnic groups. A number of polymorphisms of genes including APOA5, APOC3, APOE, sterol-regulatory element-binding protein-1c (SREBP1c) and TNF have also been associated with an increased risk of dyslipidaemia in HIV-infected individuals⁽Reference Guardiola, Ferré and Salazar^129–Reference Tarr Philip, Taffé and Bleiber¹³³⁾.

It has been suggested that the diagnosis of dyslipidaemia in HIV-infected individuals should be made using recommendations for non-HIV-infected individuals⁽Reference Gkrania-Klotsas and Klotsas⁹⁰⁾. For the general population, dyslipidaemia is diagnosed using a fasting lipid profile and defined using the NCEP ATP III criteria. Ideally, fasting lipid profiles should be offered to patients before initiation of ART in order to gain an insight into the exact changes caused thereafter by ART⁽Reference Gkrania-Klotsas and Klotsas⁹⁰⁾. LDL levels are the primary target of the NCEP ATP III guidelines, which recommend that lifestyle modifications be trialled first, followed by statins, to lower LDL⁽⁸⁶⁾.

Glucose abnormalities

Before the ART era, the development of T2DM in HIV-infected individuals was attributed to the anti-microbial medication pentamidine⁽Reference Abourizk, Lyons and Madden¹³⁴⁾ and was relatively uncommon⁽Reference Bradbury and Samaras¹³⁵⁾. Following the introduction of PI, however, a greater number of reported glucose disorders began to emerge in HIV-infected individuals⁽Reference Dubé, Johnson and Currier^15, Reference Eastone and Decker^136, Reference Visnergarwala, Krause and Musher¹³⁷⁾.

Abnormalities in fasting blood glucose concentration have been found in up to 20 % of patients⁽Reference Thiébaut, Daucourt and Mercié^64, Reference Walmsley, Cheung and Fantus^72, Reference Mercier, Gueye and Cournil^74, Reference Elgalib, Aboud and Kulasegaram⁸⁸⁾, while prevalence figures for impaired fasting glucose⁽Reference Mutimura, Stewart and Rheeder^59, Reference Mercier, Gueye and Cournil^74, Reference Zannou, Denoeud and Lacombe^75, Reference Savès, Raffi and Capeau^81, Reference Elgalib, Aboud and Kulasegaram⁸⁸⁾ and impaired glucose tolerance⁽Reference Puttawong, Prasithsirikul and Vadcharavivad^57, Reference Walmsley, Cheung and Fantus^72, Reference Jevtovic, Dragovic and Salemovic^73, Reference Savès, Raffi and Capeau^81, Reference Carr, Samaras and Thorisdottir^98, Reference Tomažič, Silič and Karner¹¹⁶⁾ have been shown to vary from 3·8 to 18 % and from 7 to 37 %, respectively. In comparison, the prevalence of impaired glucose tolerance and impaired fasting glucose in the general population is 8·4 and 6·3 %, respectively⁽Reference van Dieren, Beulens and van der Schouw¹³⁸⁾.

Puttawong et al. ⁽Reference Puttawong, Prasithsirikul and Vadcharavivad⁵⁷⁾ and Tomažič et al. ⁽Reference Tomažič, Silič and Karner¹¹⁶⁾ identified the prevalence of insulin resistance in 30 and 38 % of their HIV subjects, respectively. The prevalence of diabetes in the general population has been shown to be 9·8 % for men and 9·2 % for women⁽Reference Danaei, Finucane and Lu¹³⁹⁾, while in patients receiving ART, prevalence has been shown to range from 7 to 27 %⁽Reference Jevtovic, Dragovic and Salemovic^73, Reference Savès, Raffi and Capeau^81, Reference Carr, Samaras and Thorisdottir^98, Reference Tomažič, Silič and Karner¹¹⁶⁾. Although the aforementioned studies have shown a relationship between ART and glucose abnormalities in HALS, a number of studies have failed to show a relationship with either glycaemic parameters⁽Reference Sobieszczyk, Hoover and Anastos^47, Reference Chêne, Angelini and Cotte^53, Reference Carr, Hudson and Chuah¹⁴⁰⁾ or insulin resistance⁽Reference Dubé, Parker and Tebas¹⁴¹⁾, highlighting the inconsistencies that currently exist in the literature.

Risk factors for the development of glucose abnormalities in the context of HIV and ART have been recently reviewed and were found to include older age, existing LA, non-white race, family history of T2DM, and disease factors, such as co-infection with hepatitis C⁽Reference Gkrania-Klotsas and Klotsas⁹⁰⁾. Furthermore, a recent study from Bangkok found that the risk of pre-diabetes in HIV-infected patients receiving ART increased with each 5 kg increase in body weight⁽Reference Srivanich, Ngarmukos and Sungkanuparph¹⁴²⁾. In the same study, the NNRTI nevirapine was found to be protective for pre-diabetes. Both in vitro and in vivo studies have demonstrated the negative effect of PI on glucose homeostasis in HIV⁽Reference Behrens, Dejam and Schmidt^27, Reference Lee, Seneviratne and Noor^143, Reference Woerle, Mariuz and Meyer¹⁴⁴⁾. Results from the Women's Interagency HIV Study showed that longer-term exposure to NRTI increased the incidence of T2DM, indicating their role in increasing the risk of glucose abnormalities in HALS⁽Reference Tien, Schneider and Cole¹⁴⁵⁾.

Diagnosis of glucose disorders in HALS is similar to the general population and has been made on the basis of guidelines from the International Diabetes Federation⁽⁸⁹⁾ and the American Diabetes Association⁽¹⁴⁶⁾. According to these guidelines, fasting plasma glucose greater than 5·6 mmol/l is defined as impaired glucose tolerance and a value greater than 7 mmol/l is indicative of frank diabetes. The American Diabetes Association criteria for diagnosing abnormalities of glucose metabolism state that patients must present with symptoms (polyuria, polydipsia, weight loss) and a random glucose of greater than 11·1 mmol/l for a diagnosis of diabetes to be made. Furthermore, Wohl et al. ⁽Reference Wohl, McComsey and Tebas¹⁴⁷⁾ recommend follow-up with fasting blood glucose every 3–6 months for at-risk patients and those undergoing changes in their ART regimen.

Hypertension

Both LA and LH have been shown to be independently associated with hypertension in HIV-infected individuals receiving ART⁽Reference Crane, Grunfeld and Harrington¹⁴⁸⁾. As for the general population, hypertension in HIV patients is associated with an increased risk of CVD⁽Reference de Arruda Junior, Lacerda and Moura¹⁴⁹⁾. A recent UK study of HIV patients with the metabolic syndrome found that raised systolic blood pressure was associated with risk factors such as being male, higher BMI and higher CD4 cell count and viral load⁽Reference Elgalib, Aboud and Kulasegaram^88, Reference Sattler, Qian and Louie¹⁵⁰⁾. Crane et al. ⁽Reference Crane, Grunfeld and Harrington¹⁴⁸⁾ suggest that increased BMI may be involved in mediating hypertension associated with LH in HALS. When the authors adjusted for BMI, patients with LA had an increased risk of hypertension compared with those without anthropometric abnormalities⁽Reference Crane, Grunfeld and Harrington¹⁴⁸⁾. The role of ART in mediating hypertension is somewhat unclear. A study by Thiébaut et al. ⁽Reference Thiébaut, El-Sadr and Friis-Møller¹⁵¹⁾ showed that ART was not independently associated with any negative effects on blood pressure; in fact, use of NNRTI was associated with a lower risk of hypertension in this group.

Carotid artery intima thickness

Arterial stiffness has been shown to be an independent predictor of cardiovascular morbidity and mortality in the general population⁽Reference Franklin¹⁵²⁾. Exposure to ART in HIV-infected individuals is associated with thickening of the carotid artery intima and arterial stiffness⁽Reference van Vonderen, Smulders and Stehouwer^153, Reference Vigano, Bedogni and Cerini¹⁵⁴⁾. Recent findings from the Women's Interagency HIV Study and the Multicenter AIDS Cohort found a significant association between HIV-related immunosuppression and increased carotid artery stiffness, independent from the impact of ART or other traditional atherosclerotic risk factors⁽Reference Seaberg, Benning and Sharrett¹⁵⁵⁾. These results suggest that disease factors may predict the development of arterial stiffness and subsequent atherosclerosis in HALS.

Endothelial dysfunction

Endothelial dysfunction is a critical initial step in the progression of atherosclerosis in HIV-infected individuals⁽Reference Shankar and Dubé¹⁵⁶⁾. A recent prospective study showed that the presence of lipodystrophy predicted endothelial dysfunction in fifty-five HIV-infected patients, independent of other CVD risk factors⁽Reference Masiá, Padilla and García²⁹⁾. Contrary to initial findings, different classes of ART have been implicated in the pathogenesis of endothelial dysfunction in HALS⁽Reference Masiá, Padilla and García²⁹⁾. Currently, results appear conflicting, some showing that use of ART contributes to endothelial dysfunction⁽Reference Stein, Klein and Bellehumeur¹⁵⁷⁾, some showing no association between ART and endothelial function⁽Reference Dubé, Shen and Mather¹⁵⁸⁾, and others showing improved endothelial function following treatment in previously ART-naive subjects⁽Reference Torriani, Komarow and Parker¹⁵⁹⁾. Interestingly, recent in vitro work has shown increased oxidative stress and cellular senescence in human coronary artery endothelial cells following long-term exposure to ritonavir and lopinavir–ritonavir⁽Reference Lefèvre, Auclair and Boccara¹⁶⁰⁾, highlighting a potential mechanism for PI-associated endothelial dysfunction. Larger long-term prospective studies are, however, required to determine the effect of ART on endothelial dysfunction in vivo.

Atherosclerosis

Patients with lipodystrophy have been shown to be at a higher risk of atherosclerosis⁽Reference Coll, Parra and Alonso-Villaverde¹⁶¹⁾. Calza et al. ⁽Reference Calza, Manfredi and Pocaterra¹⁶²⁾ recently reviewed the link between HIV infection, ART and the development of premature atherosclerosis. Similar to the general population, the most commonly identified risk factors associated with atherosclerosis were age, smoking, increased BMI, hypertension and dyslipidaemia. Of nine studies, four found an association between the use of PI and premature atherosclerosis. Furthermore, three of five studies showed that HIV infection itself was associated with atherosclerosis. This, coupled with the association between risk of atherosclerosis and CD4 cell count⁽Reference Kaplan, Kingsley and Gange¹⁶³⁾, indicates that disease factors play an important role in the pathogenesis of atherosclerosis.

CVD

It has been well established that ART contributes to a ‘metabolic syndrome’ encompassing abdominal obesity, atherogenic dyslipidaemia, insulin resistance, endothelial dysfunction and inflammation, known as HALS. In recent years, therefore, research has begun to focus on the deleterious effects of ART on risk of CVD⁽Reference Bradbury and Samaras¹³⁵⁾.

Early reports of CVD appeared in peer-reviewed literature shortly after the introduction of PI⁽Reference Henry, Melroe and Huebsch^164, Reference Vittecoq, Escaut and Monsuez¹⁶⁵⁾. Evidence for the association between ART and increased risk of CVD is, at present, inconsistent. Some studies show no association between the use of ART and risk of CVD or cerebrovascular disease⁽Reference Bozzette, Ake and Tam¹⁶⁶⁾, while others show a positive association for PI⁽Reference Friis-Møller, Sabin and Weber^30, Reference Friis-Møller, Weber and Reiss^118, Reference Sabin, Worm and Weber^167, Reference Barbaro, Di Lorenzo and Cirelli¹⁶⁸⁾. Research has shown that between 5 and 31 % of patients with HIV/AIDS are at risk for cardiovascular events⁽Reference Jevtovic, Dragovic and Salemovic^73, Reference Mallewa, Higgins and Garbett^169, Reference Alvarez, Salazar and Galindez¹⁷⁰⁾, and, similar to the general population, patients with the metabolic syndrome have a greater risk than those without⁽Reference Alvarez, Salazar and Galindez¹⁷⁰⁾.

Variations in observed risk could be explained by differences in the risk factors of the study population. Commonly identified risk factors for MI or cardiovascular events include AIDS before ART initiation, age over 40 years, cigarette smoking⁽Reference Jevtovic, Dragovic and Salemovic⁷³⁾, family history of CVD, diagnosis of dyslipidaemia, hypertension, lipodystrophy or T2DM⁽Reference Sabin, Worm and Weber¹⁶⁷⁾ or pre-existing vascular disease⁽Reference Bozzette, Ake and Tam³⁵⁾. Unlike the general population, Bozzette et al. ⁽Reference Bozzette, Ake and Tam³⁵⁾ showed that risk of serious cardiovascular events was lower for African-American subjects, indicating that race/ethnicity may also be a risk factor. It has also been found that the prevalence of CVD is higher for patients receiving a combination of PI and NNRTI⁽Reference Friis-Møller, Weber and Reiss¹¹⁸⁾. In a recent review, Schafer et al. ⁽Reference Schafer, Short and Squires¹⁷¹⁾ referred to studies which show an increased risk of CVD associated with recent, but not cumulative, use of abacavir, a NRTI. However, a recent 96-week randomised controlled trial did not find an association between the NRTI combination abacavir–lamivudine and cardiovascular morbidity and mortality in HIV-infected individuals⁽Reference Martin, Amin and Cooper¹⁷²⁾. These researchers suggest that differences in results may be attributed to variations in pre-study viral load among patients. The increased longevity observed in the HIV population as a result of advanced drug therapy has also been associated with an increase in the incidence of CVD⁽Reference Schillaci, Pucci and De Socio¹⁷³⁾. Evidence indicates that disease progression and associated immune deficiency in HIV patients are associated with an increased CVD risk⁽Reference Seaberg, Benning and Sharrett¹⁵⁵⁾. Recent evidence that a low CD4 cell count was associated with an increased prevalence of carotid artery lesions in HIV patients further supports this finding⁽Reference Kaplan, Kingsley and Gange¹⁶³⁾. Paradoxically, interruption of ART has been shown to increase CVD risk⁽Reference Phillips, Carr and Neuhaus¹⁷⁴⁾, suggesting that HIV infection itself may play a role in increasing the risk of CVD. A recent treatment interruption trial in Thai HIV-infected patients demonstrated an association between markers of CVD, including increased vascular cell adhesion molecule-1, decreased adiponectin, and increased HIV RNA replication⁽Reference Calmy, Gayet-Ageron and Montecucco¹⁷⁵⁾, which further supports this finding.

Currently, risk-prediction models such as the Framingham score are recommended for use in estimating CVD risk⁽Reference Calza, Manfredi and Pocaterra¹⁶²⁾. The Framingham equations, developed over a decade ago for use in non-HIV-infected individuals⁽Reference Anderson, Odell and Wilson¹⁷⁶⁾, have been used to estimate CVD risk in HIV-infected subjects⁽Reference Jevtovic, Dragovic and Salemovic^73, Reference Martin, Amin and Cooper^172, Reference Calmy, Gayet-Ageron and Montecucco¹⁷⁵⁾; however, studies assessing the accuracy of this model in HIV-infected patients are limited⁽Reference Friis-Møller and Worm¹⁷⁷⁾. Friis-Møller et al. ⁽Reference Friis-Møller, Weber and Reiss¹¹⁸⁾, in a large prospective cohort, used CVD risk-scoring estimates for the general population to determine cut-offs to define HIV patients at ‘high risk’ of CVD. More recently, May et al. ⁽Reference May, Sterne and Shipley¹⁷⁸⁾ have developed another risk model for predicting the risk of MI or death from CHD in HIV-infected men. These researchers use data from five cardiovascular cohorts of HIV-uninfected men and adapt the model for the known risk factors observed in HIV patients following initiation of ART. However, the authors state that only a modest change in CHD risk factors may be detected using the risk model. In addition, the model does not take into account changes in CHD risk attributable to lifestyle changes. To the best of the current authors' knowledge, this model has also not yet been evaluated.

Altered adipocyte inflammatory status

Studies of human adipose tissue from HIV-infected patients receiving ART have demonstrated an increase in the expression of genes relating to inflammation. In particular, HALS has been associated with an increase in pro-inflammatory cytokine expression⁽Reference Hammond, McKinnon and Nolan¹⁷⁹⁾, in addition to increased systemic pro-inflammatory cytokine activity⁽Reference Johnson, Albu and Engelson¹⁸⁰⁾. Increased circulating levels of TNF-α, IL-6 and IL-1β have been shown in both in vitro ⁽Reference Lagathu, Eustace and Prot^181, Reference Kim, Wilson and Wabitsch¹⁸²⁾ and ex vivo studies⁽Reference Johnson, Albu and Engelson^180, Reference Jan, Cervera and Maachi^183, Reference Sievers, Walker and Sevastianova¹⁸⁴⁾. IL-6 has been shown to mediate insulin resistance and may modulate insulin signalling in adipose tissue⁽Reference Bastard, Maachi and van Nhieu¹⁸⁵⁾. A large body of research has focused on the hypersecretion of TNF-α, which has a number of pathophysiological effects including mediating insulin resistance via reduction of insulin receptor kinase activity, inducing apoptosis and lipolysis⁽Reference Domingo, Vidal and Domingo^186, Reference Kovsan, Ben-Romano and Souza¹⁸⁷⁾, and down-regulating insulin receptor kinase substrate (IRS)-1 and GLUT-4⁽Reference Domingo, Vidal and Domingo^186, Reference Mallewa, Wilkins and Vilar¹⁸⁸⁾. These effects may occur via a number of mechanisms including: a reduction in insulin signalling, attenuating the anti-lipolytic action of insulin⁽Reference Rudich, Ben-Romano and Etzion¹⁸⁹⁾; down-regulation of inhibitory G-protein-coupled receptors, leading to enhanced cyclic AMP levels⁽Reference Zhang, Halbleib and Ahmad¹⁹⁰⁾; down-regulation of lipoprotein lipase⁽Reference Domingo, Vidal and Domingo¹⁸⁶⁾; and down-regulation of the function and expression of perilipin, a lipid droplet-associated protein, which protects the adipocyte from the hydrolytic action of cellular lipases⁽Reference Rydén, Arvidsson and Blomqvist¹⁹¹⁾. This increased cytokine production in HALS has also been suggested to induce a stress response in adipocytes, which may lead to physical damage of the cell⁽Reference Lagathu, Eustace and Prot^181, Reference Adler-Wailes, Guiney and Koo¹⁹²⁾.

In addition to an increase in inflammatory cytokine production, HALS has been associated with a reduced expression of adiponectin in both plasma and adipose tissue⁽Reference Giralt, Domingo and Guallar¹⁹³⁾. Adiponectin is a potent insulin sensitiser and, hence, its down-regulation contributes to insulin resistance⁽Reference Kim, Wilson and Wabitsch¹⁸²⁾. In vitro and ex vivo studies have shown reduced expression, secretion and release of adiponectin from adipose tissue⁽Reference Lagathu, Eustace and Prot^181, Reference Jan, Cervera and Maachi¹⁸³⁾, while in vivo studies in HALS patients have identified the presence of hypoadiponectinaemia, which is a risk factor for cardiovascular impairment⁽Reference Bezante, Briatore and Rollando¹⁹⁴⁾. Inhibition of adipocyte differentiation, such as that caused by PI, has been shown to down-regulate adiponectin expression⁽Reference Körner, Wabitsch and Seidel¹⁹⁵⁾. Furthermore, down-regulation of adiponectin expression by NRTI has been suggested to occur as a result of the reduction in fat mass associated with NRTI use⁽Reference Pacenti, Barzon and Favaretto^196, Reference Lindegaard, Keller and Bruunsgaard¹⁹⁷⁾. Mallewa et al. ⁽Reference Mallewa, Wilkins and Vilar¹⁸⁸⁾ also refer to the negative feedback loop that exists between cytokines, whereby high levels of TNF-α and IL-6 may inhibit the expression of adiponectin, which may also account for the observed reduction in adiponectin in HALS.

Adipose tissue macrophage infiltration, resulting in chronic low-grade inflammation, has also been suggested to contribute to the development of HALS⁽Reference Sevastianova, Sutinen and Kannisto¹⁹⁸⁾. Macrophage infiltration of adipose tissue has been shown to be greater in HALS patients compared with healthy controls⁽Reference Jan, Cervera and Maachi¹⁸³⁾. Recently, Hammond et al. ⁽Reference Hammond, McKinnon and Nolan¹⁷⁹⁾ demonstrated an increase in adipose tissue macrophage count associated with thymidine NRTI treatment.

Altered adipocyte functionality

Microarray analysis of gene expression during adipogenesis has revealed numerous effects of ART on genes involved in adipocyte lipid and glucose metabolism⁽Reference Pacenti, Barzon and Favaretto¹⁹⁶⁾. In a recent study, Sievers et al. ⁽Reference Sievers, Walker and Sevastianova¹⁸⁴⁾ showed that NRTI caused a general decrease in the expression of genes involved in adipocyte differentiation and lipid and glucose metabolism within the cell (CCAAT/enhancer-binding protein-α (C/EBPA), C/EBPB, cyclo-oxygenase-3 (COX3), GLUT4, hexokinase-1 (HEXOK1), perilipin (PLIN), SREBP1c), and an increase in markers of cell proliferation and genes involved in mitochondrial transcription (COX4, lamin-B (LAMINB), lamin A/C (LAMINA), proliferating cell nuclear antigen (PCNA), PPAR-γ co-activator-1b (PGC1B)).

Similarly, a number of in vitro studies have demonstrated changes in gene expression following exposure of adipocytes to antiretroviral drugs. Both PI and NRTI have been shown to down-regulate the expression of adipocyte differentiation genes such as Pparg, Cebpa, adiponectin (Adipoq), leptin (Lep), the scavenger receptor CD36 (Cd36), adipocyte lipid-binding protein-2 (Ap2), fatty acid synthase (fasn) and acetyl-coenzyme A carboxylase (Acc)⁽Reference Pacenti, Barzon and Favaretto^196, Reference Caron, Auclair and Lagathu¹⁹⁹⁾. In particular, the NRTI d4T and ZDV have been found to cause a reduction in mRNA expression of adipogenic markers involved in lipid accumulation including fatty acid synthase, acetyl-coenzyme A carboxylase and adipocyte lipid-binding protein-2⁽Reference Lagathu, Eustace and Prot^181, Reference Pacenti, Barzon and Favaretto^196, Reference Caron, Auclair and Lagathu^199, Reference Grigem, Fischer-Posovszky and Debatin²⁰⁰⁾. Pacenti et al. ⁽Reference Pacenti, Barzon and Favaretto¹⁹⁶⁾ demonstrated that NRTI modulate the expression of various transcription factors, such as Aebp1, Pou5f1 and Phf6, which may play a role in determination of the adipocyte phenotype.

Adiponectin plays a role in glucose and lipid metabolism within the adipocyte⁽Reference Kim, Wilson and Wabitsch¹⁸²⁾ and a number of in vitro studies have shown a reduction in adiponectin expression following exposure of 3T3-L1 murine and Simpson–Golabi–Behmel syndrome (SGBS) human adipocytes to PI⁽Reference Kim, Wilson and Wabitsch^182, Reference Pacenti, Barzon and Favaretto^196, Reference Grigem, Fischer-Posovszky and Debatin²⁰⁰⁾. These alterations in gene expression correspond with findings of altered adipocyte function including reduced capacity of insulin to activate lipogenesis⁽Reference Caron, Auclair and Lagathu¹⁹⁹⁾, decreased lipid accumulation⁽Reference Caron, Auclair and Lagathu¹⁹⁹⁾ and reduced adipocyte lipid content⁽Reference Lagathu, Eustace and Prot^181, Reference Kim, Wilson and Wabitsch¹⁸²⁾.

Two ex vivo studies have investigated gene expression in subcutaneous adipose tissue samples from HALS patients and found reduced nuclear mRNA expression of mitochondrial proteins (PGC-1α), transcription factors (PPAR-γ) and adipocyte metabolic markers (GLUT-4, lipoprotein lipase)⁽Reference Giralt, Domingo and Guallar^193, Reference Kim, Leclercq and Lanoy²⁰¹⁾. Further support for these findings comes from results by Kim et al. ⁽Reference Kim, Leclercq and Lanoy²⁰¹⁾, which showed that the expression of PPAR-γ increased after PI withdrawal. Moreover, mRNA expression of uncoupling protein-3 and preadipocyte factor-1, both inhibitors of adipocyte differentiation and metabolism, has been shown to be increased in HALS⁽Reference Giralt, Domingo and Guallar¹⁹³⁾. As with the work of Sievers et al. ⁽Reference Sievers, Walker and Sevastianova¹⁸⁴⁾, these findings suggest that ART impair mitochondrial biogenesis, adipocyte differentiation and metabolism, and are involved in the down-regulation of adipogenic transcription factors.

Mitochondrial toxicity

ART-mediated inhibition of mitochondrial DNA-polymerase-γ, leading to mitochondrial toxicity, has been suggested to not only be involved in cell death and loss of fat mass, but in the aetiology of alterations in adipose tissue function⁽Reference Hammond, McKinnon and Nolan¹⁷⁹⁾. As a result of these defects in adipose tissue function, the liver and skeletal muscles are exposed to increased concentrations of fatty acids, which has been associated with the development of the metabolic alterations seen in HALS⁽Reference Villarroya, Domingo and Giralt²⁰²⁾. Studies examining the effect of ART on mitochondrial toxicity are somewhat conflicting, with some showing limited or no effect of certain ART regimens on mitochondrial toxicity⁽Reference Venhoff, Setzer and Melkaoui^203, Reference Birkus, Hitchcock and Cihlar²⁰⁴⁾, while others found effects for both single ART and cART⁽Reference Hammond, McKinnon and Nolan^179, Reference Caron, Auclair and Lagathu^199, Reference Viengchareun, Caron and Auclair²⁰⁵⁾. According to Walker et al. ⁽Reference Walker, Setzer and Venhoff²⁰⁶⁾, mitochondrial toxicity is sometimes more pronounced with use of cART. A recent study examining the effect of switching from d4T to tenofovir found improvements in mitochondrial toxicity after just 1 month⁽Reference Ribera, Paradiñeiro and Curran²⁰⁷⁾. Mallewa et al. ⁽Reference Mallewa, Wilkins and Vilar¹⁸⁸⁾ suggest that these observed differences may be due to differing levels of affinity of the active metabolites of the drugs for mitochondrial DNA-polymerase-γ. Furthermore, PI and NRTI have been associated with increased oxidative stress, which has been shown to induce mitochondrial dysfunction in 3T3-F442A adipocytes⁽Reference Caron, Auclairt and Vissian^208, Reference Walker, Auclair and Lebrecht²⁰⁹⁾. This PI- and NRTI-associated mitochondrial dysfunction and oxidative stress have also been shown to trigger premature senescence in a number of cell models, including primary human fibroblasts⁽Reference Caron, Auclairt and Vissian²⁰⁸⁾, human coronary artery endothelial cells and peripheral blood mononuclear cells⁽Reference Lefèvre, Auclair and Boccara¹⁶⁰⁾. In the context of HIV and ART, it has been suggested that premature senescence may contribute to accelerated cellular ageing, which might increase the risk of premature CVD as observed in HALS⁽Reference Lefèvre, Auclair and Boccara¹⁶⁰⁾.

Pharmacological and surgical management

A number of pharmacological and surgical interventions have been used in the management of HALS. Pharmacological interventions include switching to more ‘lipid-friendly’ antiretrovirals⁽Reference Sension, de Andrade Neto and Grinsztejn²¹⁰⁾, use of synthetic growth hormone analogues to reduce excess visceral adipose tissue⁽Reference Sivakumar, Mechanic and Fehmie²¹¹⁾, statins to improve dyslipidaemia⁽Reference Tungsiripat and Aberg^212–Reference Fichtenbaum, Gerber and Rosenkranz²¹⁴⁾ and anti-diabetic drugs⁽Reference Feldt, Oette and Kroidl^215–Reference van Wijk, Hoepelman and de Koning²¹⁸⁾ to improve glucose abnormalities. A number of adverse events are associated with these pharmacological interventions, which range from drug–drug interactions⁽Reference Jiménez-Nácher, Alvarez and Morello²¹⁹⁾ to more serious side effects such as a higher virologic failure⁽Reference Martínez^220, Reference van Vonderen, Gras and Wit²²¹⁾ and increased risk of MI⁽Reference Psaty and Furberg²²²⁾ (Table 2).

Table 2 Adverse events associated with the pharmacological and surgical management of the HIV-associated lipodystrophy syndrome

PI, protease inhibitor; NRTI, nucleoside RT inhibitor.

To correct the morphological abnormalities associated with HALS, patients often undergo surgical procedures. These include liposuction⁽Reference Hultman, McPhail and Donaldson²²³⁾ and excisional lipectomy⁽Reference Warren and Borud²²⁴⁾ for LH and silicone gluteal prostheses⁽Reference Moreno, Miralles and Negredo²²⁵⁾, facial fillers⁽Reference Skeie, Bugge and Negaard^226–Reference Narciso, Bucciardini and Tozzi²²⁸⁾, facial grafting⁽Reference Fontdevila, Serra-Renom and Raigosa⁹⁶⁾ and fat transplantation⁽Reference Guaraldi, Squillace and De Fazio²²⁹⁾ for LA. Surgical interventions such as these are radical interventions and are associated with numerous adverse events, which often offset their success (Table 2).

Lifestyle interventions

Although pharmacological and surgical interventions have a role to play in the management of HALS, lifestyle interventions are increasingly being trialled as first-line strategies in the management in HALS, due to their greater safety and tolerability.

Exercise

A number of studies have investigated the role of exercise in improving the systemic parameters in HALS and have shown mixed results. One study failed to show an effect of exercise and resistance training in improving lipid parameters in HALS⁽Reference Terry, Sprinz and Stein²³⁰⁾, while four have shown a beneficial effect, particularly in reducing central fat accumulation and in increasing body weight and limb girth⁽Reference Roubenoff, Weiss and McDermott^231–Reference Fillipas, Cherry and Cicuttini²³⁴⁾. A recent cross-sectional study investigated the effect of leisure time physical activity on central fat accumulation in adults receiving ART and showed a significant negative correlation between leisure time physical activity and central fat⁽Reference Florindo and Jaime²³⁵⁾. As for the general population, exercise in HALS patients has proven effective in improving lipid parameters and insulin resistance. Yarasheski et al. ⁽Reference Yarasheski, Tebas and Stanerson²³⁶⁾ investigated the effect of exercise on dyslipidaemia and showed that progressive weight-lifting reduced serum TAG levels in eighteen men receiving ART. Furthermore, a recent study of twenty men receiving supervised strength and endurance training demonstrated increases in insulin-mediated glucose uptake and hence improved insulin sensitivity after 16 weeks of training⁽Reference Lindegaard, Hansen and Hvid²³⁷⁾. Overall, it appears that exercise has a beneficial effect in improving lipid parameters and central adiposity in HALS.

Nutrition

Relatively little is known about the influence of diet on the metabolic complications of HIV and associated lipodystrophy⁽Reference Hadigan²³⁸⁾. There are a number of studies that have generally investigated the area by cross-sectional analysis of diet and systemic parameters of HIV-positive adults with and without lipodystrophy. Dietary fibre intake has been shown to be positively associated with metabolic health in HIV-positive adults⁽Reference Hendricks, Dong and Tang^40, Reference Shah, Tierney and Adams-Huet^41, Reference Hadigan, Jeste and Anderson⁴³⁾. In another study, fibre had no association⁽Reference Gavrila, Tsiodras and Doweiko²³⁹⁾. A recent Brazilian cross-sectional study found that individuals with HIV who consumed more than two servings of dairy food per d had a lower BMI, waist circumference and blood pressure than those who consumed less than this amount⁽Reference Leite and Sampaio⁴²⁾. The authors of this study suggest that Ca intake may be involved in mediating these changes. In most cross-sectional studies no association was found between saturated fat⁽Reference Shah, Tierney and Adams-Huet^41, Reference Gavrila, Tsiodras and Doweiko^239, Reference Batterham, Garsia and Greenop²⁴⁰⁾, total fat⁽Reference Shah, Tierney and Adams-Huet^41, Reference Gavrila, Tsiodras and Doweiko^239, Reference Batterham, Garsia and Greenop²⁴⁰⁾ or other fat subclasses⁽Reference Gavrila, Tsiodras and Doweiko²³⁹⁾, with the exception of trans-fatty acids⁽Reference Shah, Tierney and Adams-Huet⁴¹⁾, and metabolic health in HIV-positive adults. Samaras et al. ⁽Reference Samaras, Wand and Law²⁴¹⁾ in their study of men with HALS showed that saturated fat intake was significantly positively associated with percentage body fat. Weak evidence suggests that polyunsaturated fat intake is positively associated with insulin sensitivity in HIV-infected individuals⁽Reference Hadigan, Jeste and Anderson⁴³⁾. Contrary to these findings, Samaras et al. ⁽Reference Samaras, Wand and Law²⁴¹⁾ demonstrated that fat subtype did not relate to fasting insulin, insulin resistance, total cholesterol, HDL, TAG, glucose or adiponectin concentrations in HALS.

Turčinov et al. ⁽Reference Turčinov, Stanley and Rutherford⁴⁵⁾ cross-sectionally investigated the diets of 136 HIV-positive Croatian adults on ART. Adherence to a Mediterranean diet was assessed by a 150-item questionnaire and a point scale that stratified subjects as having low or moderate to high adherence. Although HALS was not an inclusion factor in the study, it was determined that Croatians who did not smoke and moderately or highly adhered to the Mediterranean diet were least likely to have LA and LH. In another cross-sectional study, adherence to a Mediterranean-style diet was positively correlated with HDL and marginally negatively correlated with TAG levels⁽Reference Tsiodras, Poulia and Yannakoulia⁴⁴⁾.

Interestingly, a negative association between total and supplemental vitamin E intake and diastolic blood pressure has been shown among HIV-positive adults⁽Reference Gavrila, Tsiodras and Doweiko²³⁹⁾. Two association studies have shown that dietary energy intake is not associated with metabolic dysregulation among HIV-positive adults⁽Reference Shah, Tierney and Adams-Huet^41, Reference Gavrila, Tsiodras and Doweiko²³⁹⁾, and one has shown significant positive associations⁽Reference Batterham, Garsia and Greenop²⁴⁰⁾.

A number of intervention studies have investigated the effects of diet in mitigating the metabolic and morphological abnormalities of HALS (Table 3). Barrios et al. ⁽Reference Barrios, Blanco and Garcia-Benayas²⁴²⁾ showed that adherence to a low-fat diet for 6 months reduced total cholesterol by 10 % and TAG by 23 % among HIV-positive adults with hyperlipidaemia. Contrary to these findings, Ng et al. ⁽Reference Ng, Chan and Li²⁴³⁾ in a recent pilot randomised controlled trial found that HIV-infected individuals who adhered to a low-fat diet did not have reduced cholesterol levels and in fact had increased TAG levels after 1 year. The same authors found that HIV-infected individuals adopting a modified Mediterranean diet had significantly increased cholesterol levels after 9 and 12 months, while serum TAG levels in the same individuals remained unchanged over the same period⁽Reference Ng, Chan and Li²⁴³⁾. In a case report, Roubenoff et al. ⁽Reference Roubenoff, Schmitz and Bairos²⁴⁴⁾ found that a moderate-fat, low-GI, high-fibre diet, in combination with exercise, reduced total and trunk fat, LDL, fasting glucose and insulin resistance in one male HALS patient. Similarly, another study found that a low-fat diet and aerobic exercise significantly reduced body weight, body fat and waist:hip ratio in HALS patients⁽Reference Terry, Sprinz and Stein²³⁰⁾. One study investigated the effect of altering the fatty acid composition of the diet from medium- to long-chain fatty acids in HALS, and showed improvements in lipid profile after 3 months⁽Reference Vázquez, Reyes and Alcaraz²⁴⁵⁾. Owing to conflicting results, further randomised controlled trials are necessary before dietary recommendations can be made in this area.

Table 3 Intervention trials investigating the effect of nutrition in the HIV-associated lipodystrophy syndrome (HALS)

PI, prospective intervention; HIV+, HIV-positive; ART, antiretroviral therapy; TC, total cholesterol; ↓ , decrease; CT, control trial; M, male; HALS+, HIV patients with HALS; HALS − , HIV patients without HALS; REE, resting energy expenditure; ↑ , increase; RCT, randomised controlled trial; CR, case report; GI, glycaemic index; WHR, waist:hip ratio; MCT, medium-chain TAG; OLRT, open-label randomised trial; RIT, randomised intervention trial; NRTI, nucleoside RT inhibitor; RDBPCT, randomised double-blind placebo-controlled trial; PR, prospective randomised trial; t.d.s., ter die sumendum (three times per d); POL, prospective open label study.

* 10 d per month for 24 weeks.

† 16-week intervention followed by 16-week washout.

A number of studies have examined the effect of supplements, such as l-acetylcarnitine, uridine and niacin, on the metabolic and morphological abnormalities in HALS. l-Acetylcarnitine has been suggested to be involved in regulating fatty acid oxidation⁽Reference Famularo, Moretti and Marcellini²⁴⁶⁾ and in one study of HALS subjects supplementation with 4 g/d resulted in increased lipid oxidation, decreased intramyocellular TAG content, decreased plasma NEFA and lower insulin sensitivity compared with controls after 8 months⁽Reference Benedini, Perseghin and Terruzzi²⁴⁷⁾. Three interventions have trialled dietary uridine supplementation, which has been shown in vitro to prevent and treat mitochondrial toxicity⁽Reference McComsey, O'Riordan and Setzer²⁴⁸⁾. One study showed no effect on changes in fat or blood mitochondrial DNA levels⁽Reference McComsey, O'Riordan and Setzer²⁴⁸⁾, while the other two studies showed conflicting results – one finding no significant increase in limb fat mass following 24 weeks of supplementation⁽Reference Calmy, Bloch and Wand²⁴⁹⁾, while the other showed a significant increase in subcutaneous fat mass following 3 months of supplementation in lipoatrophic patients⁽Reference Sutinen, Walker and Sevastianova²⁵⁰⁾. Both studies used the same level of supplementation. Niacin, which has been shown to modulate lipoprotein metabolism and inhibit TAG synthesis⁽Reference Kamanna and Kashyap²⁵¹⁾, was used in one study examining the effect of combination therapy with diet, exercise and niacin in patients with highly active ART-associated dyslipidaemia. Treatment with a low-saturated fat diet, exercise and niacin significantly increased HDL concentrations, and total cholesterol:HDL ratio compared with controls after 24 weeks⁽Reference Balasubramanyam, Coraza and Smith²⁵²⁾.

An interesting set of studies by Kosmiski et al. ⁽Reference Kosmiski, Bessesen and Stotz^253–Reference Kosmiski, Bessesen and Stotz²⁵⁵⁾ has shown that lipodystrophy in HIV is associated with an increase in resting energy expenditure (REE) per kg lean body mass. Furthermore, 3 d of eu-energetic feeding, which normally would not induce a change in REE, resulted in a significant increase in REE among HIV-positive adults with lipodystrophy compared with HIV-positive adults without lipodystrophy and healthy controls⁽Reference Kosmiski, Bessesen and Stotz²⁵³⁾. The same researchers found that 3 d of hypo-energetic feeding induced a significant drop in REE and 3 d of hyper-energetic feeding induced a significant increase in REE in HIV-positive adults with lipodystrophy compared with HIV-positive adults and healthy controls⁽Reference Kosmiski, Bessesen and Stotz^253, Reference Kosmiski, Bessesen and Stotz²⁵⁵⁾. The group concluded that lipodystrophic subjects have higher REE per kg lean body mass than non-lipodystrophic subjects, that short-term over-feeding increases REE among lipodystrophic subjects and that short-term energy restriction reduces REE among lipodystrophic subjects. The authors suggest that hypermetabolism associated with lipodystrophy, and a form of adaptive thermogenesis invoked to dissipate energy that cannot be stored in a normal manner underlie these observations.

Despite weak support from observational studies, a number of intervention trials focusing on the role of n-3 long-chain PUFA (n-3 LC-PUFA) in mitigating the metabolic abnormalities in HALS patients have been pursued. In the pre-ART era, intervention trials investigating the immunomodulatory effects of EPA and DHA as an adjunct therapy in HIV patients were pursued⁽Reference Razzini and Baronzio²⁵⁶⁾. Their hypothesis was based on the immunomodulatory effects of EPA and DHA previously documented. Evidence strongly supports a role for n-3 LC-PUFA in HIV therapy, but in lipid lowering rather than immune regulation.

In a study of 120 HIV-positive adults on ART, 8 weeks of supplementation with 6 g n-3 LC-PUFA per d induced a 25·5 and 38·7 % reduction in plasma TAG concentrations among moderate and severe hypertriacylglycerolaemics, respectively⁽Reference De Truchis, Kirstetter and Perier²⁵⁷⁾. Similarly, plasma TAG concentrations decreased by 25 % following 4 weeks of supplementation with 1750 mg EPA and 1150 mg DHA per d among fifty-two HIV-positive adults with moderately raised TAG⁽Reference Wohl, Tien and Busby²⁵⁸⁾. In a study of 100 HIV-positive adults with hypertriacylglycerolaemia, fish oil supplements taken at 6 g/d for 8 weeks reduced TAG concentrations by 46 %, fenofibrates reduced TAG concentrations by 58 %, and the combination of fish oil and fenofibrates by 65·5 %⁽Reference Gerber, Kitch and Fichtenbaum²⁵⁹⁾. Manfredi et al. ⁽Reference Manfredi, Calza and Chiodo²⁶⁰⁾ showed that rates of TAG normalisation were non-significantly different, at 25·9 and 34 %, between HIV-positive subjects with raised TAG supplemented with ethyl esters of n-3 LC-PUFA or treated with pharmaceutical lipid-lowering therapy, respectively. Salmon oil, administered at 3 g/d, significantly reduced TAG concentrations after 12 to 24 weeks of supplementation in fifty-eight HIV-positive adults on ART⁽Reference Baril, Kovacs and Trottier²⁶¹⁾. The TAG-lowering effects of the n-3 LC-PUFA among HIV-positive adults are supported by three smaller prospective studies⁽Reference Ranieri^262–Reference Carter, Woolley and Jolley²⁶⁴⁾, although Virgili et al. ⁽Reference Virgili, Farriol and Castellanos²⁶⁵⁾ showed no significant effect among nine HIV-positive subjects receiving 1120 mg EPA and 720 mg DHA daily for 6 weeks. A review of 237 hospital charts from HIV-positive adults with hypertriacylglycerolaemia showed that the use of n-3 LC-PUFA supplements was associated with a 32 % reduction in TAG concentrations⁽Reference Normén, Yip and Montaner²⁶⁶⁾. Furthermore, at baseline 11 % of subjects used these dietary supplements, whereas at 6 months 25 % of subjects used the supplements⁽Reference Normén, Yip and Montaner²⁶⁶⁾. This demonstrates an enthusiasm and acceptance of these dietary supplements by HIV-positive adults with hypertriacylglycerolaemia. The effects of n-3 LC-PUFA on lipoprotein concentrations in HIV-positive adults are unclear, with no effect⁽Reference Ranieri²⁶²⁾, 11 % raised HDL⁽Reference Normén, Yip and Montaner²⁶⁶⁾ and 22·4 % raised LDL⁽Reference Wohl, Tien and Busby²⁵⁸⁾ reported.

EPA and DHA have been shown to have anti-inflammatory effects in vitro via their role as PPAR-γ ligands⁽Reference Calder²⁶⁷⁾ and modulation of the NF-κB signalling system⁽Reference Weldon, Mullen and Loscher^268, Reference Mullen, Loscher and Roche²⁶⁹⁾. Despite the strength of evidence to support anti-inflammatory effects of EPA and DHA in vitro, studies investigating the effects of n-3 LC-PUFA supplementation on cytokine production in HIV-positive adults are limited. One study found no effects on the concentration of the soluble TNF-α receptor following 6 months of dietary supplementation with a product containing 1·7 g n-3 LC-PUFA and 7·4 g arginine⁽Reference Pichard, Sudre and Karsegard²⁷⁰⁾. Another study demonstrated that among ten subjects consuming a bar containing 1·96 g n-3 LC-PUFA, PGF-1α secretion was decreased, and IL-1β and IL-6 secretion increased, from peripheral blood mononuclear cells⁽Reference Bell, Chavali and Bistrian²⁷¹⁾. Overall, n-3 LC-PUFA appear to have beneficial TAG-lowering effects; however, their role in modulating inflammation in HALS remains to be elucidated.