Atherosclerosis and heart disease

Atherosclerosis is the hardening and narrowing of the arteries. The process begins with damage to the endothelium, a single layer of cells that line the arteries. Once the endothelium becomes damaged (for example by smoking), platelets can get caught up in the tiny nicks on the normally non-stick surface of the artery wall. White blood cells can also be caught in these platelet accumulations, and migrate through the wall of the artery (the intima), where they become macrophage cells.

Macrophages actively accumulate low-density lipoprotein (LDL) cholesterol once inside the artery wall, and begin to accumulate in clumps that become visible in the arteries of many adults by the age of 30.

Fat-laden macrophages play an active role in attracting more macrophages and causing further damage, by stimulating inflammation or bursting out of the artery wall. Eventually fatty streaks begin to stimulate the migration of cells which generate connective tissue and smooth muscle cells, forming a lesion called a plaque which partially obstructs an artery. Over time the plaque will attract deposits of calcium, and calcification of the artery is an important marker of the development of atherosclerosis.

High lipid levels in the blood can lead to this narrowing, but other factors such as smoking, diet and stress also appear to affect the speed at which this process occurs. Ageing reduces the elasticity of the arteries, making any blockage or narrowing more difficult for the blood to flow past.

Coronary artery disease, one example of atherosclerosis, is a form of heart disease where the arteries that provide blood and oxygen to the heart become narrowed. If an artery which supplies the heart with blood is partially blocked, chest pain or angina may result. Angina may precede a heart attack. Other factors that increase the risk of heart disease are obesity, high blood sugar levels and lack of exercise. Drug treatment can reduce the strain on the heart if atherosclerosis occurs.

Thickening of the wall of the carotid artery (the artery in the neck which supplies the brain with blood) is another sign of atherosclerosis. The technical term for this condition is intima media thickness (IMT). Thickening of the carotid artery is a surrogate marker for coronary atherosclerosis, so some studies investigating the risk of heart attack in HIV-infected people have measured IMT.

High LDL cholesterol is associated with narrowing of the arteries when other risk factors such as hypertension, diabetes, obesity and smoking are present. Although high triglyceride levels are often seen in people with atherosclerosis, they are not associated with narrowing of the arteries in the absence of high cholesterol levels. In the presence of high triglyceride levels, LDL cholesterol particles become more atherogenic - that is, they cause more damage to the arteries, because they can pass more easily into the walls of arteries. These particles also persist for longer in the presence of high triglyceride levels.

HIV infection itself may also contribute to the development of atherosclerosis by causing gaps to open up in the endothelium of blood vessels (which promotes plaque formation), as well as inducing apoptosis of endothelial cells (Field 2004).

Cardiac abnormalities and HIV

HIV itself may increase a persons risk of heart disease (Klein 2002; Mary-Krause 2001). Cardiac abnormalities (including reduced heart muscle mass and reduced ability of the heart to relax between contractions allowing blood to flow into the heart) have been observed among people with HIV and may foreshadow later heart failure. Studes have also linked HIV infection and higher viral loads to impaired endothelial function, the ability of the blood vessels to dilate in response to increased blood flow (Solages 2006). This is the earliest detectable sign of future cardiovascular disease.

Children and infants who are infected with HIV from birth have worse cardiac function than other children (Lipshultz 2002). A five year study of approximately 300 HIV-infected infants and young children has found 18-39% developed heart dysfunction and that this was associated with an increased risk of death (Starc 2002).

The mechanisms which lead to cardiac abnormalities among adults and children with HIV have not been established but malnourishment is one factor associated with abnormalities (Martinez-Garcia 2000).

Cardiovascular risk factors among people with HIV

Risk factors for heart disease are widespread among people with HIV, regardless of any influence of therapy or HIV infection. Epidemiologists and doctors at the Royal Free Hospital in London reviewed risk factors among patients at the clinic. Forty-five per cent of the Royal Free cohort were smokers and 7% had an alcohol intake above the weekly recommended limit. Seventy-two per cent of current smokers had previously tried to stop.

• 20% had a body mass index (BMI) above 26kg/m², a level associated with a higher risk of cardiovascular disease.

• 34% were aged over 40.

• 29% reported a family history of heart disease.

• 14% had high blood pressure.

• 18% had total cholesterol above 6.3mmol/l.

• 36% had non-fasting triglycerides above 2.3mmol/l.

• 10% had high density lipoprotein (HDL) cholesterol below 0.9mmol/l.

Among patients on highly active antiretroviral therapy (HAART; 73% of the cohort), raised cholesterol and triglyceride levels were significantly more common, as was a family history of heart disease (38 vs. 30%, p = 0.03).

However, patients on HAART were significantly less likely to smoke (43 vs. 47%, p = 0.03). It appeared that those who were ex-smokers (24% of the cohort) were more likely to have given up after starting HAART (65% of ex-smokers) (Smith 2002).

Multiple risk factors have also been identified in other cohorts. In the SMART study, a comparison of HIV treatment strategies in both treated and untreated patients with CD4 counts above 350 cells/mm³, 18% of men and 25% of women already had a metabolic syndrome (see Metabolic changes on antiretroviral therapy for a definition of the metabolic syndrome).

Evidence of a link?

The first reports of premature heart attack and coronary artery disease among people taking protease inhibitors emerged in the late 1990s. A letter in the Annals of Internal Medicine reported four men aged between 35 and 44 years who had heart attacks after more than two years on protease inhibitor therapy. All four men were smokers and three had high lipids (Flynn 1999; Gazzard 1999). There have also been anecdotal reports of arterial blockages in the legs and the spleen believed to be caused by narrowing of the arteries. However, there is no clear evidence that these changes were caused by protease inhibitors - one of the patients already had a history of angina and a family history of early onset heart disease, so the development may have been a coincidence.

These early case reports led researchers to analyse large databases of HIV-infected patients to assess the rates of heart disease. To date, evidence is inconclusive although there is a growing body of evidence which supports the theory that people taking protease inhibitors who develop high lipids are at increased risk of heart attack.

The most important evidence to date comes from the DAD study. DAD (Data collection on Adverse events of anti-HIV Drugs) is a unique collaborative effort begun in 1999, in which data from eleven cohorts of HIV-positive people (eight European, one multinational, one Australian, and one North American) have been pooled in order to track the long-term toxicity of HIV therapies. The project is supported by the European Agency for the Evaluation of Medical Products (EMEA), and the corresponding United States agency, the Food and Drug Administration (FDA), along with eight pharmaceutical companies who market antiretrovirals.

The DAD study population included 23,468 people, three quarters of whom were male. The median age was 39. At entry, over half the cohort were smokers, approximately 20% had raised total cholesterol levels (above 6.3 mmol/l), over a third raised triglyceride levels (above 2.3 mmol/l), and a quarter had low levels of the cardio-protective HDL cholesterol (Friis-Moller 2003).

Nineteen per cent had taken no antiretroviral drugs, 67% had never taken a protease inhibitor, and 34% had never taken a non-nucleoside reverse transcriptase inhibitor (NNRTI). Combination antiretroviral therapy was defined here as any regimen containing a protease inhibitor or an NNRTI in combination with nucleoside reverse transcriptase inhibitors (NRTIs).

DAD set out to detect a two-fold increase in the rate of heart attacks with increasing exposure to combination antiretroviral therapy. When designing any research study, the investigators calculate how much data they will need to gather in order to address their objectives. In this case, DAD needed to amass at least 30,000 patient years of follow-up, and passed this target last year.

During 36,199 person years of follow-up, 126 people had a heart attack, 36 of whom died. The overall event rate was 3.5 heart attacks per 1,000 person years of follow-up, though the rate increased with the duration of exposure to antiretrovirals (some participants were already on therapy at the time they entered the study). The rate in people on treatment for less than one year was 2.2 heart attacks per 1,000 person years of follow-up, but rose to 5.5 in those who had been on treatment for more than four years. Whether this incremental risk will continue for longer periods is not yet clear, as there were relatively few people in the study who had been on treatment for more than six years.

After the data were adjusted for the effects of known cardiovascular risk factors (for example age, gender, smoking status), each year of exposure to antiretroviral therapy increased the risk of heart attack by 26%. The risk was also associated with older age, male gender, smoking status, presence of diabetes, higher cholesterol levels and previous experience of cardiovascular disease. Indeed, certain of these traditional risk factors were more significant predictors of heart attack risk than antiretroviral exposure, such as raised cholesterol at study entry, which increased risk by 60%.

Concerned that their findings should not be taken out of context, the DAD steering committee issued a Position Statement to coincide with the studys presentation at the Tenth Retroviruses Conference in Boston, in 2003. Cautioning that cohort studies cannot prove causality, only association, they point to the well-established benefits of HAART in extending survival in people with HIV, saying, although of concern, the absolute risk of myocardial infarction [heart attack] remains low and does not outweigh the marked effectiveness of combination antiretroviral therapy in preventing HIV-related complications.

Nevertheless the DAD study is particularly important because it is a prospective study in a closed cohort with an average follow-up of five years, because all myocardial infarctions in the study were independently validated within two months of the event, and because complete baseline risk factor data are available to allow properly controlled analyses.

The DAD study has also reported on the risk of cerebral infarction (stroke) and cardiac procedures (CCVEs). In this analysis, an additional 81 patients experienced at least once CCVE other than a myocardial infarction. Thirty-eight cases of stroke were reported, of which nine were fatal. Thirty-nine cases of invasive procedures intended to modify cardiac risk were reported, including angioplasty and bypass surgery. None of the individuals died.

The incidence of first CCVE was 5.7 cases per 1,000 person-years of follow-up (95% confidence interval [CI] 5.0 - 6.5), and increased with longer exposure to antiretroviral therapy (p < 0.001). Given the range of events reported, this is still a relatively low frequency, and is not enough for the investigators to pin the blame on a particular class of antiretroviral.

To put this into perspective with the other, classic risk factors for CCVE, the authors examined the relative risk (RR) through multivariate analysis. They were, in order of risk: previous history of CCVE (RR, 7.12 p < 0.001); male gender (RR, 1.82 p = 0.02); smoking (RR, 1.66 p = 0.008); family history of CCVE (RR, 1.62 p = 0.03); older age (RR per five years older, 1.42 p < 0.001); antiretroviral therapy (RR per year of exposure, 1.26 p < 0.001).

Additional analyses tested the association between CCVE and a number of metabolic and physiological causes. Factors independently associated with the risk of CCVE, were, in order of relative risk: diabetes (RR, 2.22 p < 0.001); high blood pressure (RR, 1.79 p = 0.001); high triglycerides (RR, 1.30 per log₁₀ copies higher p = 0.006); high cholesterol (RR, 1.11 per mM higher; p = 0.008; DAD Writing Committee 2004).

In a second prospective study, all HIV-positive patients admitted with an acute myocardial infarction to a Los Angeles hospital between 1998 and 2000 were monitored and matched with HIV-negative controls. Twenty-four HIV-positive patients experienced heart attacks during that time. Over 15 months of follow-up, the HIV-positive patients had a much higher incidence of second acute myocardial infarction (20 vs. 4%) and rehospitalisation for a coronary event (45 vs. 11%) than sex- and age-matched controls. The investigators observed that protease inhibitors did not seem to be solely responsible for abnormally high blood lipids seen in 24 patients who had had heart attacks. Consequently, the authors attributed the higher rate of secondary cardiac events among HIV-positive patients to HIV infection itself, rather than PI therapy (Matetzky 2003).

Key findings of retrospective studies include:

• A comparison of 951 people who took antiretroviral therapy without a protease inhibitor and 373 people who took a PI-based regimen found that incidence of heart attack increased fivefold among those who took a PI. However, the increase was not statistically significant (Jutte 1999).

• An analysis of the US Veterans Administration (VA) cohort, which included 36,666 patients who received HIV care between 1993 and 2001 did not find any increase in the incidence of cardiovascular events after the introduction of HAART (Bozzette 2003). The VA analysis has good follow-up because patients who move around the United States will continue to seek their care from the VA, but the average follow-up is shorter than that of the D:A:D study.

• Another database of individuals involved in the Kaiser Permanente health care plan, which individuals may opt out of (and thus be lost to follow-up), showed an increased rate of heart attack and coronary artery disease in people with HIV compared with adults without HIV in an age-matched analysis, but no relationship to type of treatment (Klein 2002, 2003). The analysis did not control for other cardiovascular risk factors. The database, which contains 15,000 patient-years of follow-up among HIV-positive individuals, may not be as accurate as the Veterans Administration cohort, due to the ability of patients to drop-out.

• An analysis of the HIV Outpatient Study, a US database covering 5676 patients, with 17,712 years of patient follow-up, found that individuals taking a protease inhibitor (PI) were at least three times more likely to experience a heart attack compared to those not taking a PI, but had no increased risk for angina or stroke (Holmberg 2002). An event rate of 1.42 cases per 1,000 patients years of follow-up in PI recipients versus an event rate of 0.46 per 1,000 patient years in non-PI recipients). The incidence of heart attacks rose significantly after 1996, the year that drugs from the PI class began to be licensed for treatment of HIV.

• A review of the Johns Hopkins University HIV cohort reported a higher incidence of angina, myocardial infarction and stroke in 2,671 patients followed for 7,330 person-years and compared with an age and race matched control population. The researchers said that the rate among HIV-positive individuals was two to three times higher than the level that would be expected in the HIV-negative population (Moore 2003).

• A review of the Insight database of 7542 US patients found that one year of protease inhibitor treatment increased the risk of subsequent cardiovascular events by 60% over a median of 3.5 years of follow-up (Iloeje 2005).

• A review of the French Hospital database found that people who had been taking PIs for more than 30 months had a 3-fold increase in risk of heart attack (Mary-Krause 2001).

• Another small study also reported no increased incidence of heart attack among people on PIs (Coplan 2001).

As a consequence of the possible increased risk associated with HAART, a recent editorial in the British Medical Journal written by Filip Moerman and colleagues has argued that cardiovascular risk should be taken into account when considering when a patient should start anti-HIV therapy. They advocate that patients with a risk of more than 20% over ten years should delay starting HAART until the CD4 cell count has fallen to near 200 cells/mm³, allowing them time to undertake lifestyle changes and possibly take lipid-lowering medication, rather than starting HAART when their CD4 cell count is closer to the upper recommended limit of 350 cells/mm³ (Moerman 2005).

What is the risk of heart disease?

Whilst the D:A:D study and a number of retrospective cohort studies provide some evidence to suggest that the risk of cardiovascular disease is higher in people receiving HAART, the follow-up is relatively short-term and we do not know whether all people with HIV will have an increased risk, or whether HAART or HIV disease are accelerating atherosclerosis in people who already have other, classic risk factors.

The main limitation of cohort evidence is the restricted follow-up of patients. Most studies of cardiovascular risk factors have followed patients for fifteen to twenty years.

In the absence of such lengthy follow-up, clinicians have been obliged to make guesses based on lipid, glucose and body composition changes, which are all good predictors of the long-term risk of cardiovascular disease in HIV-negative people.

Elevations in total and LDL cholesterol, low HDL cholesterol levels, increased waist measurement and high plasma glucose levels have been detected in a substantial proportion of people with HIV receiving HAART. See Metabolic changes on antiretroviral therapy in Anti-HIV therapy: Body fat and metabolic changes whilst on treatment.

A comprehensive risk assessment

In order to establish what the risk might be, it is necessary to look at a number of questions:

• What is the existing risk of heart disease in the population and what are the metabolic changes which predict heart disease in all adults?

• What patterns of metabolic changes occur in people taking antiretroviral therapy?

• Taking this pattern and the information about how these metabolic changes determine the risk of heart disease in the general population, what can we conclude about the effects of antiretroviral therapy on coronary heart disease risk?

• How do we decide when heart disease could pose a greater risk than HIV infection for someone considering whether to start or continue treatment?

The most comprehensive analysis of the risk of heart disease has been carried out by Matthias Egger of the University of Bristol. He used a number of different datasets to model the possible impact of metabolic changes on heart disease risk.

Baseline risks in the HIV-negative male population (Caerphilly Heart Disease study) predict the subsequent risks of fatal and non-fatal coronary heart disease over 15 to 20 years of follow-up:

• Ever smoked: 2.3-fold increase in risk compared to non-smokers.

• Diabetes or impaired glucose tolerance (fasting glucose 6 to 7mM): 2.3-fold increase in risk.

• Triglycerides above 2.0mmol/l (175mg/dl) - 1.8-fold increase in risk.

• High blood pressure (above 140 / 90) - 1.5-fold.

• Cholesterol above 5.2mmol/l (200mg/dl) - 1.5-fold.

• Reduced HDL cholesterol (below 1.0mmol/l or 38mg/dl) - 1.4-fold.

What happens when these figures are compared with metabolic changes on HAART? Matthias Egger used data gathered by Andrew Carr in Sydney to assess the severity of metabolic changes in HAART patients, and compared these with the relative risks for people with elevated cholesterol levels and other metabolic changes derived from the Caerphilly study (see previous paragraph). The purpose of this exercise was to demonstrate the degree to which HAART may add to an individual's risk of coronary heart disease, over and above the influence of age and lifestyle.

The average age in the Sydney lipodystrophy study was 40, and 98% of the participants were men.

A number of different patterns of metabolic changes were seen in the Sydney cohort, and these patterns hold true for all people taking HAART, although different combinations of drugs have been associated with very different patterns.

This assessment found that the increase in risk ranged from 1.3 to fourfold higher, with impaired glucose tolerance alone associated with the smallest increase in risk, and the presence of diabetes, elevated cholesterol and elevated triglycerides associated with the greatest increase in risk.

If serious lipid elevations occur, Matthias Egger calculates that the risk of coronary heart disease within five years are as follows:

• A male smoker aged 50 in the United States has a 14% risk of coronary heart disease within five years, compared to an 8% risk in France. United Kingdom levels of risk are broadly similar to those in the United States, but the level of risk may be even higher in Scotland and Ireland.

• A female non-smoker aged 30 has a 1% risk of coronary heart disease within five years, but a 50 year old male non-smoker has a 9% risk.

This absolute risk needs to be balanced against the risk of developing AIDS or dying if treatment is not started, and the risk of dying despite starting treatment, argues Matthias Egger.

For example, the MACS cohort (a study of the risk of developing AIDS among 1604 gay men in the US) shows that the risk of developing AIDS within three years does not climb above 10% until viral load rises above 41,000 copies/ml. Once it hits this level, a CD4 cell count decline below 750 cells/mm³ is associated with a 16% risk of developing AIDS within three years, climbing to a 40% risk in those with CD4 cell counts below 350 cells/mm³.

In contrast, an individual with a CD4 cell count between 201 and 350 cells/mm³ has a risk of 8% when their viral load lies below 41,000 copies/ml, and might decide that as an overweight 50 year old smoker, any deferral of anti-HIV therapy is worth the risk, given the higher five year risk of a cardiovascular event.

If these figures are translated into the numbers needed to treat in order to harm, only 18 non-smoking, 50 year-old men would need to be treated and develop lipodystrophy for one death to occur as a result of coronary heart disease within five years.

Amongst 50 year-old male smokers starting treatment, only ten would need to start treatment and develop lipodystrophy for one death to occur within five years due to coronary heart disease.

Another concern is that the protective effect of gender against atherosclerosis may be undermined by greater increases in lipid levels among women. A small prospective study of 27 men and 13 women with HIV and 35 healthy controls found that triglycerides, cholesterol and LDL cholesterol levels increased more significantly in women than in men after starting HAART, and that sex differences in the LDL:HDL cholesterol ratio disappeared as a consequence. Thus, calculations based on risk assessments in the HIV-negative population which suggest that female sex reduces the risk of heart disease may be flawed (Pernerstorfer 2001).

Other indicators of changes in risk of heart disease

As well as the identification of factors which increase the risk of heart disease, some researchers have begun to look at changes in arteries of the heart and neck. Several studies have reported damage to the arteries in people with HIV (Maggi 2000; Stein 2001; Bauer 2002) although studies often lack control groups with similar risk factors apart from anti-HIV therapy.

Two studies presented at the Tenth Retroviruses Conference in early 2003 showed conflicting findings. Both studies looked at thickening of the carotid artery (the neck artery which feeds blood to the brain), the phenomenon known as intima media thickening (IMT). One study found no evidence that protease inhibitor therapy was associated with IMT (Currier 2003). A late breaker study found that a low CD4 cell count was a risk factor for thickening of the carotid artery in 106 HIV-infected people on HAART, suggesting an association between risk of heart disease and HIV infection. Importantly, this study also found some preliminary evidence that risk of heart disease was linked to exposure to PIs. The average rate of IMT was 0.1mm/year in the HIV-infected patients on treatment, and progression was predicted by age and duration of PI use. In contrast, age-matched historical controls showed a thickening rate of 0.04mm/year. The progression rate among the HIV-infected patients is based on follow-up data from only 22 people, so these findings require confirmation (Hsue 2003).

Research presented at the International AIDS Conference in 2000, and the Second International Workshop on Adverse Drug Reactions and Lipodystrophy showed some evidence of arterial damage and IMT among people on HAART, and elevated risk factors for coronary artery disease in people with HIV-related fat redistribution. However, not all of these studies took into account the age, weight, cholesterol levels or other risk factors related to cardiovascular disease in participants, and so must be treated with caution (Seminari 2000; Mercie 2002; Hadigan 2001b; Maserati 2000; Goebel 2000).

An increased incidence of plaques in the carotid artery has also been identified. In a study of 239 HIV-positive individuals subjected to ultrasound scanning of the carotid artery, 55 (52%) of the protease inhibitor-treated patients had plaque lesions present in their carotid vessels and 25 had increased IMT indicative of pathology. In group II, 19 patients (15%) had plaques detected, and ten had increased IMT. In group three, nine patients (14%) had plaque lesions present, including three with increased IMT.

In further analysis, the investigators found that treatment with a protease inhibitor (p < 0.001) and elevated triglycerides (p = 0.01) were both significantly associated with an increased risk of vascular damage. The investigators also found that protease inhibitors were associated with elevated cholesterol (p < 0.001).

The investigators also found that the odds ratio (OR) of developing carotid plaques increased with age, with a 1.09 increase in OR per year. Smoking, a well-recognised risk factor for vascular damage, was also found to be significant (OR 2.22). The investigators also found that individuals with a CD4 cell count between 200 and 500 cells/mm³ were more likely to have vascular damage than patients with either a CD4 cell count below 200 cells/mm³ or above 500 cells/mm³ (OR 2.45, p = 0.001). However, the most significant predictor of carotid plaques was treatment with a protease inhibitor (p < 0.001; Maggi 2004).

Several further mechanisms have also been reported. Protease inhibitors may increase expression of CD36 on the surface of macrophages, leading to greater accumulation of cholesterol within these cells, and subsequent accretion of fat-laden macrophages in foam cells in the walls of blood vessels. This increase in CD36 expression occurred independently of plasma cholesterol levels, and mice lacking the CD36 receptor did not develop atherosclerotic lesions after treatment with protease inhibitors, whereas normal mice did (Dressman 2002).

A study published in mid-2002 has also reported that protease inhibitors are associated with a thickening of parts of the heart, and that this may impact on heart function (Meng 2002).

Endothelial function is also impaired in the presence of a combination of protease inhibitors and nucleoside analogues in the test tube, due to suppression of nitric oxide-mediated vasodilation as a result of elevated levels of TNF-alpha and insulin resistance. Nitric oxide production rises when insulin sensitivity is restored (Agrawal 2002).

Coronary artery calcification is a more direct measure of the development of atherosclerosis than increases in cholesterol or triglycerides. Calcification occurs when calcium is deposited in plaques, which are accumulations of cholesterol and other debris which narrow or 'fur up' arteries, reducing blood flow. The rupture of unstable plaques can cause blood clots or thromboses which are life-threatening. The degree of coronary calcification has been found to predict the risk of subsequent cardiovascular events in individuals with other risk factors. Significantly higher levels of coronary calcification have been found in individuals who receive protease inhibitor treatment compared with individuals who receive non-PI containing antiretroviral therapy. The study excluded individuals with diagnosed heart disease, high blood pressure, diabetes, and those who smoked more than one packet of cigarettes a day (Meng 2002).

While these studies suggest that HAART may predispose people to atherosclerosis, other studies have found that HAART did not impact on the arteries (Depairon 2000; Currier 2000) or that individuals on PIs had very mild arterial thickening (Cheminot 2000). A recent case-control study including HIV-positive patients on HAART regimens including protease inhibitors, those not taking protease inhibitors and HIV-negative controls has also questioned the role of protease inhibitors in IMT. This study found that IMT was no associated with protease inhibitor use or HIV status, but rather 'traditional' risk factors such as age and body mass index (Currier 2005). While by and large these studies provide encouraging evidence that HAART will not significantly damage the arteries, this type of research may not accurately predict heart attack and stroke.

Another potential indicator of cardiovascular disease, silent myocardial ischemia (SMI), has also been monitored. This occurs when the supply of oxygen to the muscles of the heart is restricted, but no pain is felt. It is an early warning sign of heart disease and increased risk of heart attack or stroke due to narrowing of the arteries. Exercise stress testing may reveal coronary artery disease in patients with no history or symptoms; silent myocardial ischemia is predictive of major coronary events in HIV-negative adults with no symptoms of heart disease.

A French study of 99 patients with no prior history of cardiovascular disease showed no association between presence of SMI and duration of antiretroviral therapy. SMI was associated with elevated plasma glucose and total cholesterol, and central fat accumulation (Duong 2002).

Antiretroviral therapy and risk of stroke

A persons risk of stroke is thought to be increased with AIDS. Also known as cerebral infarction, stroke occurs when not enough blood reaches the brain. According to a US study published in 1996, patients with AIDS had an annual incidence of stroke of 1.3% compared to a rate of 0.025% in the general population (Pinto 1996).

There is emerging evidence that PI-based treatments may add to this increased risk of stroke among people with HIV.

Protease inhibitors were strongly associated with development of lesions in the carotid vessels (the arteries which connect the heart and the brain) in HIV-infected people in one study carried out in Italy. These lesions may lead to stroke. Of the PI-treated patients 29/55 (57.7%) developed carotid lesions compared to 7/47 (15%) of HIV-infected people who hadnt taken PIs and 6.7% of a local control group. While PI therapy was most strongly associated with the lesions, smoking and degree of immune deficiency were also associated with increased risk of lesions (Maggi 2000).

While the authors found a correlation between high cholesterol and high triglycerides, they suggested that this may be due to the frequency of high triglycerides amongst those on PIs, rather than indicative of a causal connection. The authors suggest further research is needed to establish the mechanism involved in the development of carotid lesions amongst those on PIs. In the meantime, they recommend periodic ultrasonographic tests of the vascular wall for those on PI therapy.

Controversies over lipid changes on antiretroviral therapy

However, not all clinicians agree that the cholesterol elevations seen in HAART patients are predictive of heart disease.

Dr Stefan Mauss, one of Germanys leading HIV clinicians, looked at the make up of the cholesterol in the bloodstreams of 172 of his patients, and found that it may not be as atherogenic as first thought. Cholesterol is divided up into three sorts HDL (high density lipoprotein), LDL (low density lipoprotein) and VLDL (very low density lipoprotein). A total cholesterol test measures the sum of the three parts. VLDL cholesterol normally represents around 10-15% of circulating cholesterol in a healthy adult. Dr Mauss found that the ratio of VLDL cholesterol to total cholesterol was much higher in HIV-infected people than in people with elevated cholesterol in the general population. When VLDL cholesterol is elevated in this way, the risk of heart disease is less, because many of the VLDL particles are simply too big to pass through the wall of the artery to deposit their cholesterol. High VLDL levels are associated with elevated triglyceride levels because VLDL particles carry triglycerides to the tissues. Elevated triglyceride levels occur because of increased production and turnover of fatty acids in the liver on protease inhibitor therapy (Mauss 2001).

In fact, some species of VLDL cholesterol are very harmful. VLDL particles can be divided into 6 sub-groups, and groups V5 and V6 (the largest particles) carry the highest risk of coronary heart disease. A low level of HDL cholesterol coupled with high levels of V5 and V6 VLDL increases the risk of coronary heart disease substantially

A group at the University of Washington analysed VLDL sub-groups and other lipid profiles in treatment-naﶥ patients and in 15 people taking nelfinavir and 17 taking efavirenz after 16 weeks of therapy. Most individuals were taking AZT/3TC, although a few of the nelfinavir group were taking ddI/d4T.

The group found that whilst triglyceride levels rose significantly in both groups (26.5% in the NFV group and 38.3% in the EFV group, non-significant difference), the change in VLDL levels was modest, and largely occurred in the less atherogenic subset of medium-sized VLDL particles.

However, a clear difference did emerge in HDL cholesterol levels. Although HDL cholesterol is considered to be `good cholesterol, increases in levels of smaller HDL particles confer an increase in CHD risk. The rise in small HDL particle levels was significantly greater in the nelfinavir group when compared to the efavirenz group (44.8% vs 25.2%), whereas the efavirenz group experienced a significantly greater increase in large HDL cholesterol particles (8.2% vs 40.9%, p<0.05).

LDL cholesterol levels also increased more in the nelfinavir group than the efavirenz group (45% vs 11.7%, p<0.05) (Simon 2002).

Given that participants were taking different nucleoside backbones, these data are best treated as proof of principle rather than clinical gospel. What they suggest is the value of measuring cholesterol sub-set changes in licensing studies and large strategic trials like Initio and ACTG 384.

There is other clear evidence that different drugs tend to produce different patterns of cholesterol elevation. Nevirapine treatment, for example, has been shown to result in a 40% increase in HDL cholesterol levels after two years on treatment in the Atlantic study, compared to a non-significant change in LDL cholesterol levels over the same period. In the indinavir arm of the same study, LDL cholesterol rose significantly (by 14%) but HDL cholesterol levels did not change significantly. The total cholesterol levels rose by 22% in both the indinavir and the nevirapine arms in this study, but only one of these changes could be defined as atherogenic (Van der Valk 2001b). Efavirenz treatment has also been shown to result in cholesterol elevations that largely comprise HDL cholesterol.

Despite these attempts to downplay the risk of heart disease among people taking protease inhibitors, other clinicians are more cautious. They point to the fact that heart attacks and other cardiovascular events in people under the age of 45 have been associated with as few as one or two of the risk factors associated with heart disease. For example, a study of HIV-negative individuals who developed premature coronary heart disease (before the age of 50) in Switzerland found that compared to an age-matched sample of the general population, people with premature heart disease were significantly more likely to smoke (85 vs 30%), but were no more likely to have elevated lipid levels (40% vs 35%). Thirty-five percent of those who developed premature heart disease had only one risk factor, and 35% had only low risk factors (Mooser). Another study of HIV-negative individuals from the Framingham database found that very high lipid levels are not necessary for individuals to experience an increased risk of heart disease (Lloyd-Jones 2001). Quite marginal increases in the presence of other risk factors such as raised blood pressure and smoking resulted in an elevated risk.

The findings described above suggest that there may be value to intervening even when an HIV-infected individual does not have a large number of risk factors, if triglyceride levels are greatly elevated. Although other analyses have not shown triglycerides to be an independent risk factor in the development of heart disease, the Framingham analysis suggests that the LDL:HDL cholesterol ratio and triglycerides may be a more relevant marker of cardiovascular risk in HIV disease.

In a review presentation given at the 2002 International AIDS Conference, Dr Jens Lundgren of Copenhagen University said that there seems to be a lagtime between starting protease inhibitors and increased risk of cardiovascular events. This lagtime appears to be at least 20 months.

Increased risk of high blood pressure

High blood pressure is a known risk factor for heart disease. Abdominal obesity, elevated serum triglycerides, insulin resistance, hypercholesterolemia and hypertension have been associated with an increased risk of cardiovascular disease in HIV-negative adults - the so-called `Syndrome X'.

A review of the MACS cohort - HIV-positive and HIV-negative men with long-term follow-up data available - showed that HIV-positive men not on treatment were significantly less likely to have systolic hypertension than HIV-negative men, and had a similar risk of systolic hypertension to HIV-negative men for the first two years after initiating HAART. However, their risk increased in comparison to HIV-negative men after more than two years on HAART (OR 1.80). Dystolic hypertension risk did not increase with the duration of HAART (Seaberg 2003).

A case control study has shown that people with lipodystrophy had significantly higher blood pressure than HIV-positive people without lipodystrophy on HAART. Although there was a trend towards an association between elevated triglyceride levels and hypertension in this study, the only significant relationship was between an increased waist-hip ratio (central fat accumulation) and the development of hypertension. Increased systolic blood pressure was more likely to become elevated in the event of an increased waist-hip ratio. The authors of the study proposed that interventions to manage hypertension should be instituted in patients with three or more elevated systolic blood pressure readings (Sattler 2001).

One retrospective study compared development of high blood pressure among people on indinavir with those on nelfinavir. Twenty-two percent of the indinavir recipients developed high blood pressure compared with 8% of the nelfinavir group. Statistical analysis showed that only indinavir significantly increased a person's risk of hypertension (Hewitt 1999b).

A study of 2302 HIV-positive individuals in Hawaii observed an increase in average blood pressure over time independent of age, sex and body weight, and significantly higher blood pressure in those receiving HAART (Chow 2000).

Recently, elevated blood pressure has been linked with treatment with ritonavir-boosted lopinavir (Kaletra) in a cohort of 444 patients from the United States. These patients were 2.5 times more likely to have high blood pressure than patients taking efavirenz (Sustiva), although this seemed to be secondary to the drug's effect of increasing body mass index. A similar effect was seen with tenofovir (Viread) and 3TC (lamivudine, Epivir) treatment.  The same study also revealed a link between starting antiretroviral therapy at very low CD4 cell counts (below 50 cells/mm³) and high blood pressure (Crane 2006).

Insulin resistance and cardiovascular risk

Insulin resistance increases the risk of cardiovascular disease even in the absence of hyperglycemia. In the presence of insulin resistance, LDL cholesterol particles are smaller, denser and more atherogenic.

If unchecked, insulin resistance may develop into diabetes, which significantly increases the risk of cardiovascular disease. Indeed, the development of diabetes leads to risk of a cardiovascular event that is equivalent to the risk of further cardiovascular events experienced by a person who has already had one CVD event within the previous eight years.

The British HIV Association recommends that insulin resistance should be treated with metformin.

Infants and risk of heart abnormalities

There have been reports linking the use of antiretroviral drugs during pregnancy to heart problems in infants. However, the balance of current evidence indicates that NRTIs do not cause neurologic or cardiac problems in infants exposed in the womb (Mofenson 2000). One recent prospective study of over 600 infants found zidovudine had no impact on the left ventricular structure or heart function of infants as assessed during 14 months of follow-up. AZT exposure also did not affect fractional shortening at 10-14 months while the impact on left ventricular mass was inconsistent (Lipshultz 2000).

Reducing the risk

Actions the individual can take to reduce the risk of heart disease and arterial blockages are discussed fully in Treating body fat and metabolic changes in Anti-HIV therapy: Body fat and metabolic changes whilst on treatment. Recommendations include: regular exercise and reduced consumption of saturated fats (found in meat, processed cakes and biscuits, butter, full cream dairy products, fast foods) which can increase your blood cholesterol. Stopping smoking and reducing alcohol intake to less than three units a day is also desirable.

Heart surgery may be an option for people who are developing a substantial blockage in a coronary artery. A stent can be inserted to keep the artery open. This operation has been performed in HIV-infected people (Boccara 2001).

Thrombosis

Thrombosis is the formation of a thrombus or blood clot within a blood vessel. Among HIV-infected individuals, age over 45 years and the presence of opportunistic illnesses, especially CMV, are associated with a greater risk of thrombosis. Drugs associated with thrombosis among HIV-infected individuals are indinavir and megestrol acetate. Sex, race and mode of exposure have not been associated with thrombosis (Sullivan 2000).

A link between the protease inhibitor (PI) class of anti-HIV drugs and the risk of thrombosis has been found in a Swedish study of 363 HIV-positive individuals, of whom 266 were taking PI-containing regimens and 97 were not on treatment. The study found significantly higher levels of plasminogen activator inhibitor type 1 (PAI-1) in the treated patients. Plasminogen activator inhibitor type 1 (PAI-1) is a protein that inhibits the break down of blood clots, and high levels of PAI-1 will lead to increased clotting of the blood. A reduced tendency to break down blood clots, called hypercoagulability, leads to a higher risk of thrombosis. The only factor significantly associated with elevated PAI-1 levels was protease inhibitor therapy, regardless of prior medical history, age, duration of therapy, specific PI drug or AIDS diagnosis (Koppel 2002).

In this study, elevated PAI-1 levels were associated with other risk factors for heart disease - such as elevated triglyceride levels, elevated insulin levels and higher levels of visceral adipose tissue. This is in line with established knowledge which links elevated levels of PAI-1 in the blood with an increased risk of various cardiovascular problems, especially heart attack. When patients were re-evaluated after twelve months, two patients in the PI group had developed cardiovascular disease and one had developed high blood pressure, while one untreated patient also developed high blood pressure and another developed diabetes.

Research into heart attack and heart disease

Friis-Moller reported on the DAD study which looked at the frequency of heart attacks (myocardial infarctions or MI) in people exposed to HAART. During over 36,000 patient years of follow-up, 126 people developed an MI, 36 of which were fatal. The incidence of MI increased with additional years on combination antiretroviral therapy, resulting in a 26% increased risk of MI per year of drug exposure. Traditional risk factors such as gender, age, smoking, raised cholesterol and diabetes were observed to predict risk of MI. And overall, the frequency of reported MI remained low at 3.5 cases per 1,000 patient years of follow-up.

Bozzette (2003) conducted a retrospective study of the risk of cardiovascular and cerebrovascular disease among the 36,766 HIV-infected patients who attended Veterans Affairs facilities between January 1993 and June 2001. Overall, there were 1,207 admissions for cardiovascular disease, 1,764 for cardiovascular or cerebrovascular disease and 2,006 deaths from these causes. However, after the introduction of HAART, the rate of admission for cardiovascular and cerebrovascular disease fell from 1.7 events per 100 patient-years in 1995 to 0.9 events per 100 patient-years in 2001. Death rates from any cause fell from 21.3 per 100 patient-years in 1995 to 5.0 per 100 patient-years in 2001. Where admission or death due to cardiovascular and cerebrovascular disease did occur, it was more likely to be amongst older patients with more advanced HIV disease and pre-existing cardiovascular or vascular disease.

Matetzky (2003) compared 24 HIV-positive patients who had experienced heart attacks to 48 age- and sex-matched HIV-negative controls. This study found that the HIV-positive patients were more likely to suffer a second cardiac event within 15 months. None of the HIV-positive patients died in hospital as a result of their acute myocardial infarction (AMI). Although the HIV-positive patients matched the 48 HIV-negative controls for age and sex, the HIV-positive patients had a much higher incidence of second AMI (20% versus 4%) and rehospitalisation for a coronary event (45% versus 11%). 30% experience vascular disease elsewhere in the body. The investigators observed that protease inhibitors did not seem to be solely responsible for abnormally high blood lipids seen in their 24 heart-attack patients, and suggested that premature coronary artery disease was likely to be a result of HIV infection itself rather than treatment with HAART.

Holmberg (2002) investigated the rate of myocardial infarction (heart attack), angina or stroke among 5,672 HOPS participants who were seen more than once at HOPS clinics between January 1993 and January 2002, giving 17,712 person years of follow-up. Average age was 43, and 82% of the group were men. The number of heart attacks increased after the introduction of protease inhibitors (PIs). PIs were not associated with increased risk of angina or stroke. 19 heart attacks were reported in 3,247 people taking PIs, an event rate of 1.42 per 1,000 patients years of follow-up; versus two heart attacks in 2,425 people not taking PIs, event rate 0.46). Heart attacks were significantly associated with PI use, diabetes and raised lipids.

Klein (2003) reported on the Kaiser Permanente observational database. The study population included only those HIV-positive men who were between the ages of 35 and 64 at entry (n=4,480). A median 4.5 years of observation provided 18,792 person years of follow-up. 4,025 people experienced 8,106 person years of non-PI exposure time (median 1.1 years), and 2,860 people experienced 10,686 person years of PI exposure (median 3.9 years). HIV-positive cases were compared with a randomly recruited control group of HIV-negative men aged between 35 and 64. These 39,425 individuals contributed 211,221 person years of follow-up. Between 1996 and 2002, 100 HIV-positive men were admitted to hospital for coronary heart disease (CHD), including 65 myocardial infarctions. Overall CHD rates were higher in HIV-positive men than HIV-negative men, when these were controlled for age (6.6 CHD events per 1,000 person years of follow-up versus 3.3 respectively). MI rates were also higher in men with HIV (3.8 versus 2.6).

Starc (2002) conducted a prospective study of 205 HIV-infected children enrolled at the average age of 1.9 years and 93 HIV-infected infants enrolled prenatally or as newborns. In the toddler group, the 5-year incidence of left ventricular fractional shortening was 28.0%. Half of the children diagnosed with heart failure had died within one year. In the infant group, 10.7% in the HIV-infected compared with 3.1% in the HIV-uninfected children developed left ventricular shortening over five years.

Mary-Krause (2001) reported a rising incidence of heart attack (myocardial infarction) with increasing duration of PI therapy among people with HIV. The analysis included 19,795 men who took a PI, giving 36,907 person years of follow-up. During this time, there were 54 heart attacks. Subjects were divided into three groups according to the duration of their PI use; either less than 18 months, between 18 and 29 months, and more than 30 months. Heart attack incidence rates increased with increasing exposure to PIs, though the number of people exposed for more than 30 months was limited. In those with the shortest exposure time, the rate was 8.9 per 10,000 person years of follow-up, and rose to 19.2 and 34.7 in the remaining exposure groups respectively. For comparison, the incidence rate in a French historical control group, matched for gender and age, was 10.8 cases per 10,000 person years. Again, HIV itself was associated with an 2-3-fold increased risk of heart attack.

Moore conducted a case control study matching HIV-positive patients who experienced cardiovascular events in the Johns Hopkins HIV clinic observational database with five age, gender and race matched HIV-positive controls. 2,671 patients were followed for a total of 7,330 person years; 43 CHD and 37 CVD events were recorded. Cases were significantly more likely to have received PI treatment, d4T treatment, to be older (46 vs 41 years)

Research into cardiovascular risks

Cheseaux (2002) studied cholesterol levels and risk of heart disease in 49 HIV-infected children who had been taking protease inhibitors. Children with high cholesterol (over 15 mmol/L) may develop heart disease in childhood. 3/49 children had cholesterol above 10 but none were over 15. The authors concluded that children treated with PIs are unlikely to develop heart disease in childhood but that long-term monitoring is necessary.

Whetham (2002) studied cardiovascular risk factors among HIV-infected patients taking PIs and attending the Mortimer Market HIV Clinic in London. Of 204 patients, 7 had coronary heart disease (CHD). Among the remaining patients, risk factors for cardiovascular disease (e.g. smoking, family history, diabetes, elevated cholesterol) were common. 14% (24) had a 15-30% risk of CHD within 10 years and 2% a greater than 30% risk of CHD.

Hewitt (2001) retrospectively analysed the risk of high blood pressure by anti-HIV treatment. Indinavir but not nelfinavir was associated with high blood pressure.

Hadigan (2001) compared 71 HIV-infected individuals with fat redistribution with 213 controls. Fasting high insulin in the HIV group was not associated with abdominal fat. High lipids, low HDL and impaired glucose tolerance were common in the HIV group. Women with lipodystrophy had larger waist circumference than controls whereas men with lipodystrophy had larger hip circumference. The authors reported risk factors for coronary artery disease were markedly higher in the lipodystrophy group.

Hewitt (1999b) identified 340 people in Buffalo, USA, treated with PI-containing drugs with full cardiovascular and HIV results. At baseline, indinavir recipients had an average age of 39, median CD4 count of 154, median viral load of 73,000, and an 84% rate of normal blood pressure. The nelfinavir group was similar: average age 37, median CD4 228, median viral load 31,000 and 84% normal blood pressure. 22% and 37% of the indinavir and nelfinavir groups were women respectively. 37 (22%) cases of high blood pressure developed in the indinavir group and 15 cases in the nelfinavir group (8.5%). Average time to hypertension was 195 day (IDV) and 157 days (NFV).

Cormier (1999) from Agouron assessed the risk of heart attack and heart pain, stroke, pancreatitis and peripheral vascular disease among people receiving nelfinavir or control therapy. Data from 1229 people from 5 clinical trial with an average time of 12.4 months on nelfinavir found one unconfirmed heart attack in the nelfinavir group and no heart attacks occurred in the control group. The data did not find any evidence that nelfinavir increases the risk of this adverse events.

Sattler (2001) conducted a retrospective review of blood pressure in HIV-positive and HIV-negative adults. Hypertension was defined as three or more diastolic blood pressure measurements above 90mmHg or three systolic measurements above 140mmHg, or a combination of three elevated diastolic or systolic measurements. Blood pressure readings were normally taken after approximately 30 minutes waiting time in the clinic. The first measurement was defined as the value obtained six months prior to commencing HAART. Forty-two patients with lipodystrophy were age and sex matched with to HIV-positive individuals without lipodystrophy, and a control group of 13 HIV-negative individuals was recruited from clinic staff. The average age was 42 in the lipodystrophic group and 40 in the HIV-positive control group. The lipodystrophy group included a significantly smaller number of Hispanic patients compared to the HIV-positive control group, but was well matched in other respects. Of note, the waist-to-hip ratio did not differ significantly between the HIV-positive control group and the lipodystrophic patients, but the ratio was significantly lower in the HIV-negative control group (p=0.001). The HIV-negative group did not differ significantly in any other body composition or metabolic characteristics except for lower fasting triglyceride levels (92mg/dL vs 308 and 208mg/dL in the lipodystrophic and HIV-positive control groups, p=0.001). Lipodystrophy was characterised by lipoatrophy in 88% of patients, and abdominal fat accumulation was reported in 69%. Lipid or glucose abnormalities were present in 86% of the lipodystrophy group. Hypertension was significantly more common in the lipodystrophic group after an average of 21 months on HAART. 74% of the lipodystrophic patients, compared with 48% of the HIV-positive control group met the study definition of elevated blood pressure (p=0.01). The average for the three highest blood pressure measurements was also calculated; the mean diastolic value was 153mmHg for the lipodystrophic group, compared to 144 for the HIV-positive control group, while the mean systolic values were 92 and 87 respectively (p=0.01). After controlling for a family history of hypertension the difference diminished, but remained significant (P<0.05, OR 3.8). Increased waist-hip ratio after commencing HAART was associated with elevated blood pressure, and was significantly correlated with systolic blood pressure. Elevated systolic blood pressure is a independent risk factor for cardiovascular disease. There was a trend towards association between triglyceride elevations and systolic blood pressure elevation. The authors proposed that interventions to manage hypertension should be instituted in patients with three or more elevated systolic blood pressure readings.

Research evidence of heart and artery damage

Currier (2003) reported from ACTG 5078, a matched case control study assessing carotid IMT in HIV-positive people on protease inhibitors (PIs) for more than two years, compared to HIV-positive people who had not received PIs, and to HIV-negative people. This study excluded people with a family history of cardiovascular disease, or with diabetes, or who had a viral load above 10,000 copies, as these factors made matching difficult. This study measured the thickness of the walls of the carotid artery (which carries blood to the brain). The thickness of this artery is called the intima media thickness or IMT. Thickening of the IMT indicates vascular disease which may be a precursor to heart attack or stroke. Based on single measure of each carotid artery, Curriers group found no differences in the IMT of HIV-positive people on protease inhibitors (PIs) for more than two years, HIV-positive people who had not received PIs, and to HIV-negative people.

Hsue (2003) reported on carotid IMT in 106 HIV-positive adults, and compared these with data from historical controls. IMT was derived by calculating the mean of twelve intima medial segments measured. As with Curriers study, Hsue found that increased IMT was associated with traditional cardiovascular risk factors such as age, LDL cholesterol, hypertension, and smoking. In multivariate analysis, Hsue observed that having a nadir CD4 count below 200 was a further predictor of risk. Carotid IMT was increased at baseline compared to large studies performed in HIV-negative populations. Preliminary follow-up data on 22 patients found that the median rate of IMT progression was 0.1 mm/year, and progression was predicted by age and duration of PI use. This rate appears to compare poorly with reported progression rates in HIV-negative populations of less than 0.04mm/year.

Meng (2002) performed echocardiography in 98 HIV-infected adults aged 25-45 years. 56.1% had taken PIs for a man duration of 30 months. There was no significant difference in the left ventricular systolic (contraction) function and cardiac valve regurgitation between those who had and had not taken PIs. However, there was a significant difference in left ventricular wall thickness and a measure of blood velocity. These measures were significantly associated with duration of PI therapy. The authors conclude that there is an association between continued PI intake and left ventricular hypertrophy (increased mass due to increasing cell size rather than more cells).

Bauer (2002) reported that PIs had direct toxicities on the endothelium (the cells that line arteries) of rats and humans. Endothelial cell dysfunction is an initiating event in vascular lesion formation and atherosclerosis.

Duong (2002) conducted exercise stress tests (which can reveal silent myocardial ischemia [SMI] or restricted blood flow in the heart) with 99 HIV-infected people who had been taking HAART for at least 12 months. 11 had evidence of SMI. Of these 11, 2 had diabetes and 2 had high cholesterol before starting HAART.

One study of SMI among HIV-infected people has been conducted (Duong 2002). Central fat accumulation was more common in those with SMI ( 54% vs 17%, p=.004), as were elevated levels of cholesterol (5.46 vs 6.47 mmol/L, p=0.04) and glucose (6.47 vs 5.46 mmol/L, p=0.007 but there was no difference in duration of exposure to any antiretroviral agents, body mass index, smoking, hypertension or family history of cardiovascular disease.

Maggi (2000) reported that 29 of 55 HIV-positive patients on PI treatment had some arterial dysfunction measured by ultrasonography, versus 7 of 47 PI-naﶥ individuals and 7 of 104 HIV-negative adults. Vascular and carotid lesions were most strongly linked to protease inhibitor therapy (p=0.001); hypertriglyceridemia disappeared as an independent risk factor when patients were stratified by treatment group and the authors concluded that vascular damage in PI-treated individuals was not mediated by hyperlipidemia.

Depairon (2000) reported that whilst the incidence of plaques in the carotid artery was higher in HIV-positive patients (n= 168) than in HIV-negative patients (n=68), no association was found with protease inhibitor therapy. Similar results were reported by Currier (2000). Plaques were independently associated with age, male gender, LDL cholesterol levels and smoking.

Stein (2001) found impaired vasodilation (expansion of the arteries) and endothelial dysfunction (damage to cells that line the arteries and heart cavities) in 21 of 28 non-smoking HIV-infected individuals on HAART.

Talwani (2002) studied coronary artery calcium (CAC) in 60 HIV-infected men aged over 40 years who had either never taken antiretroviral therapy or who had been on treatment for at least 6 months. CAC is an established marker for detecting heart disease. Each HIV-infected man was matched with 3 HIV-negative men from an existing database. 18% of cases and 17% of controls had clinically significant CAC while 33% and 37% had detectable CAC. However, the small number of cases meant the study was not powerful to show statistical significance.

Koppel (2002) studied metabolic and haemostatic (clotting and bleeding) markers in 363 HIV-positive individuals. 266 were receiving protease inhibitor(PI)-containing regimens and 97 were treatment-naﶥ. Patients on treatment had taken PIs for an average of 27 months and had been infected with HIV for longer (106 months vs 68 months). Each group had similar CD4 cell counts around 480 cells/mm³. 16% of the PI group had metabolic disturbances, compared to 8% of the treatment-naﶥ group, a non-significant difference. The study found significantly higher levels of plasminogen activator inhibitor type 1 (PAI-1 - a protein that inhibits the break down of blood clots) in the treated patients. The only factor significantly associated with elevated PAI-1 levels was PI therapy, regardless of prior medical history, age, duration of therapy or AIDS diagnosis. Elevated PAI-1 levels were significantly associated with elevated triglyceride levels, elevated insulin levels and higher levels of visceral adipose tissue. However, the PAI-1 elevation was more severe in the PI-treated group. No association could be found between particular PIs and elevated levels of PAI-1, but duration of exposure to ddI was correlated with elevated PAI-1 levels. When patients were re-evaluated after twelve months, two patients in the PI group had developed cardiovascular disease and one had developed high blood pressure, while one untreated patient also developed high blood pressure and another developed diabetes. Among patients in the PI group who switched or discontinued therapy, no significant change in PAI-1 levels was detected.

Mercie (2002) analysed 129 people on HAART with fat redistribution. Bifurcation of carotid artery intima-media was significantly thicker in those with lipodystrophy by univariate analysis. On multivariate analysis, age and body mass index were associated with intima-media thickness.

Seminari (2002) investigated 11 PI-naive and 16 PI-experienced patients for risk of cardiovascular disease. Fibrinogen was significantly higher in the PI group. Both groups had intimal thickening, with a trend to greater thickening among those on PIs.

Cheminot (2000) reported that individuals on PIs had very mild arterial thickening.

Henry (1998b) reported three cases of coronary artery disease in three men 40 years or younger who had been on protease inhibitor therapy. He also reported four cases of pancreatitis among patients on protease inhibitors.

References

See References - body fat and metabolic changes for full references.