Table of Contents
Cohort studies show that cardiovascular disease affects HIV-positive people more often than HIV-negative comparison groups. People with HIV carry a heavy burden of classic and HIV-specific cardiovascular risk factors. HIV itself and combination antiretroviral therapy (cART) appear to inflate cardiovascular risk about 50% in adults and children. At the same time, cART eases cardiovascular disease risk in various ways. Hypertension is highly prevalent in HIV populations and has a profound impact on cardiovascular and overall mortality. A Mediterranean diet rapidly cut rates of MI, stroke, and cardiovascular death in a randomized trial of high-risk people in the general population of Spain. But studies of this diet in small HIV-positive groups have yielded mixed results. Smoking prevalence stands twice higher in HIV-positive than HIV-negative US residents, and smoking may account for one quarter of all deaths and new diagnoses of cardiovascular disease, non-AIDS cancer, and bacterial pneumonia in people with HIV. US research shows that many HIV providers do not even know if their patients smoke. Some evidence suggests heavy alcohol drinking may boost cardiovascular risk in HIV-positive men more than in HIV-negative men. More than 7% of HIV-positive people in the United States have stage 3 or worse chronic kidney disease, which has a great impact on risk of cardiovascular events and heart failure.
When someone fashioned the handy acronym HAART in 1996, it seemed an apt and mnemonically friendly way to name the triple-drug strategy that wondrously reversed the course of AIDS. But it took only 2 years of HAART use to discover what a ruefully sardonic moniker HAART would be. In May 2008, fewer than 24 months after the watershed Vancouver AIDS Conference, Keith Henry and colleagues in Minnesota reported "severe premature coronary artery disease" in two men taking protease inhibitors -- one 26 years old and the other 37.1 Looking at 124 people taking PIs, Henry discovered that a third of them had high lipids. He urged colleagues "to be aware that patients receiving protease inhibitors have the potential for accelerated atherosclerosis." (This watershed report appeared as a 1-page letter in The Lancet, after the journal rejected a full-length article with angiograms.)
Everyone knows now that Henry's two young men were no anomalies, but instead harbingers of a heart disease surge that dogs HIV clinicians and disables or kills their patients to this day. In March 2013 a PubMed search for "cardiovascular disease" and "HIV" returned 5225 entries in a remorseless crescendo more than doubling from 163 citations in 1998, the year of Henry's study, to 389 in 2012 (Figure 1).
Published reports on cardiovascular disease (CVD) and HIV swelled from 163 in 1998, only 2 years after the introduction of HAART, to 389 in 2012.
Today only the most forbearing experts mapping the interlacings of HIV and cardiovascular disease can resist the heart/HAART homophone. But whatever their rhetorical indiscretions, after 15 years these experts -- and allied pathophysiologists, epidemiologists, and statisticians -- have offered some solid answers to questions posed by Keith Henry's 1998 case reports:1 Did HIV infection cause heart disease in these young men? Or was it the protease inhibitors? Or both? Or something else? Unfortunately for those preferring clean causal pathways, the answer to all four questions seems to be yes.
At the same time, HIV heart experts hasten to caution, a causal relationship between HIV or combination antiretroviral therapy (cART) and cardiovascular disease can be established only in a randomized controlled trial, and such a trial -- were it even feasible -- would have to be large and long.2 This particular HIV/heart brain trust propose that HIV and antiretrovirals can flick the risk of cardiovascular disease (CVD) in three ways (Figure 2)2:
An international panel of experts on heart disease in HIV-positive people outlined three routes by which HIV, its treatment, or both may contribute to development of cardiovascular disease.2
"Importantly," these authorities add, "there is substantial evidence to suggest that all 3 mechanisms are in operation and affect the risk of cardiovascular disease in patients infected with HIV."2 In other words, it's hard to figure out what's going on. But steadfast research since those first case reports has afforded HIV clinicians a firm footing from which to evaluate, test, counsel, and treat people with HIV who may be headed for heart trouble. The thousands of studies addressing these issues can hardly be boiled down to tidy take-home ABCs, but several points -- discussed in detail throughout this issue -- are clear:
Cardiovascular disease kills more people in the United States than any other malady, felling 600,000 people a year,3 a number higher than the population of Luxembourg. For every 3 people in the United States who die of something else, 1 dies of heart disease. So maybe we shouldn't be surprised that so many Americans with HIV get heart disease and die from it. But when one considers the grisly array of mortal threats people with HIV face -- still led by AIDS in most analyses -- heart disease exacts a stunning toll in the United States and Western Europe.
A EuroSIDA study with follow-up starting in January 2001 tracked death rates and causes in 12,844 HIV-positive people.4 During follow-up AIDS arose in 1025 people and 339 (33%) of them died. In the same period 1058 people had a serious non-AIDS diagnosis and 462 (44%) died. Of the 1058 non-AIDS diagnoses, heart disease accounted for 384 (36%), more than attributed to non-AIDS cancer (380), liver-related disease (183), pancreatitis (81), or end-stage renal disease (35).
The Data Collection on Adverse Events of Anti-HIV Drugs (DAD) Study gathers and deciphers stats on more than 49,000 HIV-positive people in Europe, the United States, and Australia. Parsing input from 33,308 cohort members with data culled up to February 2008, the DAD team counted 2482 deaths, with AIDS causing the highest proportion (29.9%).5 Cardiovascular deaths accounted for 11.6% of the tally, just behind liver-related deaths (13.7%) and ahead of deaths from non-AIDS cancer, invasive bacterial infection, kidney disease, and pancreatitis.
The HIV Outpatient Study (HOPS) has tracked HIV-positive people in 10 US cities since 1993. An analysis of 6945 patients seen from 1996 through 2004 logged 702 deaths.6 In 2004 cardiovascular disease, liver disease, and non-AIDS cancers each accounted for 23.5% of non-AIDS deaths. Rates of cardiovascular death did not rise significantly over the study period. And in one analysis figuring non-AIDS diseases as primary or secondary causes of death with or without an AIDS disease, the cardiovascular death rate fell significantly from 1996 through 2004 (P = 0.01).
In contrast, French researchers canvassing a larger, national HIV cohort charted a burgeoning trend in heart deaths.7 Surveys in 2000, 2005, and 2010 saw the proportion of deaths caused by heart trouble balloon from 8% to 10% to 14%, a highly significant jump (P < 0.0001). In 2010 AIDS retained its top mortality ranking, accounting for 25% of deaths, followed by non-AIDS non-liver cancers (22%) and cardiovascular disease (14%). Heart disease vaulted from fourth place in 2005 to third in 2010, swapping spots with liver disease.
On its website the CDC lists nine heart risk factors grouped into four categories (Table 1).8 Besides leaving out a personal history of heart disease, the list omits at least 11 risk factors closely studied in people with HIV, many of which apply to the general population as well (Table 1). The following review considers most of these risk factors, one by one, with special emphasis on variables probed in studies with clinical endpoints.
|Table 1. Cardiovascular Risk Factors in the General Population and in People With HIV|
|CDC List of Risk Factors8||Risk Factors Studied in People With HIV|
High blood pressureF
High blood pressureF
Personal history of heart disease*
Poor kidney function
Vitamin D deficiency or insufficiency
Excessive alcohol use
Excessive alcohol use
Hormonal contraceptive use
Lower CD4 count†
Higher viral load†
* Not considered in this review article.
† Considered in a separate review article in this issue.
F The Framingham tool for estimating 10-year risk of myocardial infarction considers age, gender, total cholesterol, high-density lipoprotein cholesterol, smoking, systolic blood pressure, and anti-hypertensive therapy in people who do not already have heart disease or diabetes.
This review leaves out male gender and older age, virtually certain predictors of higher heart disease risk in any study including men and women across a range of ages. This analysis also sets aside the knotty question of whether heart disease arises at a younger age in HIV populations. (A recent Veterans Aging Cohort Study audit found it does not.9) This review also excludes analysis of many individual biomarkers because clinicians are unlikely to measure things like D-dimer and sCD14 when reckoning a patient's cardiovascular risk.
A spate of DAD Study analyses piqued interest and stirred controversy with their findings that certain protease inhibitors, abacavir, and didanosine upped the risk of myocardial infarction in people with HIV. But these studies, and countless others, consistently show that classic risk factors weigh heavily in the risk equations of people with HIV. A 2007 DAD inquest found, for example, that age, male gender, a previous cardiovascular event, smoking, diabetes, and dyslipidemia each strongly and independently inflated the risk of myocardial infarction.10
Yet classic cardiovascular risk factors do not hold true across studies of all HIV populations -- they vary with the make-up of the study group and the methods applied. For example, a comparison of 1525 HIV-positive veterans and 843 HIV-negative veterans found that the HIV group shouldered a heavier burden of prevalent cardiovascular disease, hypertension, diabetes, obesity, hazardous drinking, and renal disease.11 But veterans without HIV were more likely to smoke and to have HCV infection and out-of-line lipids.
Statisticians working with the French Hospital Database on HIV figured in 2010 that HIV-positive men have a 40% higher risk of myocardial infarction (MI) than men in the general population, and HIV-positive women have almost a tripled MI risk compared with other women.12 In the United States a comparison of 3851 HIV-positive people in a Boston healthcare system and over 1 million people without HIV also found that HIV-positive men had a 40% higher MI risk than the comparison group, while women again had a 3-fold higher risk.13
A study comparing all 3953 HIV patients in Denmark from 1995 through 2004 and a general-population group of 373,856 people yielded similar results.14 Compared with the general population, HIV-positive people who had not begun combination antiretroviral therapy had a 39% higher risk of getting admitted to the hospital for ischemic heart disease, though this difference stopped short of statistical significance (adjusted relative risk, 1.39, 95% confidence interval [CI] 0.81 to 2.33). After people with HIV started antiretroviral therapy, they had a doubled (and significant) risk of hospital admission for heart disease (adjusted relative risk 2.12, 95% CI 1.62 to 2.76). But this risk did not grow through the first 8 years of antiretroviral treatment.
Although analyses like these12-14 try to account for the impact of critical risk factors, heart disease adepts who scrutinize such studies urge caution in parsing the results. The University of Wisconsin's James Stein, among the top authorities in this field, counsels that most data suggesting heightened heart disease risk with HIV come from observational studies "with important methodological limitations, including short durations of follow-up, low CVD event rates, incomplete ascertainment of risk factors and events, and a lack of HIV-negative controls."15 The Massachusetts study, for example, had an HIV-negative comparison group and adjusted calculations for age, gender, race, hypertension, diabetes, and lipids but couldn't account for one cardinal cardio factor -- smoking.13 Danish investigators were also unable to adjust for smoking in their hospital admission study.14
Meta-analysis of studies assessing heart disease risk in HIV-positive people compared with HIV-negative groups determined that antiretroviral-naive people with HIV ran a 61% higher risk (relative risk [RR] 1.61, 95% CI 1.43 to 1.81).16 Antiretroviral-treated people had a doubled risk of cardiovascular disease compared with the general population (RR 2.00, 95% CI 1.70 to 2.37) and a 52% higher risk than treatment-naive HIV-positives (RR 1.52, 95% CI 1.35 to 1.70).
After considering this study, US antiretroviral guideline writers echoed James Stein in noting that such comparisons of HIV-positive groups and the general population must typically omit a few critical factors like smoking and often cannot tame a statistical bugbear called competing risks.17 Competing risks can skew statistical analyses when people in a study group succumb to some illness or outcome other than the one being analyzed. For example, an HIV group being assessed for incident myocardial infarction may die first from kidney failure (the competing risk), whereas if they had lived they may have ended up in the group with a new MI.
Researchers working with the Veterans Aging Cohort Study (VACS) tried to conjure a relatively unbiased estimate of myocardial infarction risk with HIV by comparing HIV-positive veterans with an age- and race-matched HIV-negative group behaviorally similar to the HIV group.18 Rates of drinking and cocaine use, for example, were similar in vets with and without HIV, and the Framingham risk score was 6 (low risk) in both groups. None of these veterans had heart disease when they joined the study group, and all were in care at some point between April 2003 and December 2009.
Through a median follow-up of 5.9 years, 82,459 veterans had 871 acute MIs, and MI incidence proved consistently higher in vets with HIV in three age brackets (P < 0.05 for all comparisons):
MI incidence (per 1000 person-years) in veterans with and without HIV:
After statistical adjustment for Framingham risk factors (see Table 1 footnote), comorbidities, and substance use, veterans with HIV had almost a 50% higher MI risk (adjusted hazard ratio [aHR] 1.48, 95% CI 1.27 to 1.72). When the researchers focused only on veterans with a viral load below 500 copies/mL, this HIV-positive subgroup still had almost a 40% higher MI risk than veterans without HIV (aHR 1.39, 95% CI 1.17 to 1.66).
Despite the clever plan of this VACS study,18 like all efforts to reckon whether HIV-positive people run a higher risk of heart disease, it falls short in several ways underlined by the authors. From their list of five possible limitations, the most important is that 97% of study participants were men, so the results do not apply to women. Another limitation they do not mention, perhaps because it is so obvious, is that these veterans have ready access to free care for life. So the results may not hold for the many HIV-positive US men who fall in and out of care and have trouble paying for it.
Myocardial infarction and other heart maladies are hardly the only cardiovascular diseases that seem to affect HIV-positive people more than coevals without HIV. A comparison of HIV-positive and negative people in a Boston healthcare system figured that those with HIV had about a 20% higher risk of ischemic stroke (resulting from clots rather than ruptured vessels).19 The study focused on HIV-positive and matched HIV-negative people seen between 1996 and 2009. Over that period stroke incidence measured 5.27 per 1000 person-years in the HIV group and 3.75 in the non-HIV group. After statistical adjustment for demographics and stroke risk factors, people with HIV had about a 20% higher ischemic stroke risk (aHR 1.21, 95% CI 1.01 to 1.46, P = 0.043). The study linked a higher viral load to a heightened stroke risk.
Some evidence suggests swifter vascular disease progression in people with than without HIV. Using noninvasive ultrasonography to measure carotid intima-media thickness (cIMT), a sturdy marker of subclinical atherosclerosis, offers a safe way to track such changes (Figure 3). A cIMT at or above 0.9 mm is abnormal. cIMT is especially useful in HIV populations, James Stein notes, because most HIV groups studied are relatively young and have a low short-term risk of cardiovascular disease.15
Carotid intima-media thickness (cIMT) measures the two inner layers of the carotid artery, the intima and the media, and offers a noninvasive look at subclinical atherosclerosis and progression. (Illustration from Servier Medical Art.)
Meta-analysis of 19 cross-sectional studies confirmed significantly higher cIMT in people with HIV than in HIV-negative comparison groups.20 A longitudinal comparison of cIMT in people with and without HIV found that 148 HIV-positive people had an abnormal average baseline cIMT (0.91+/-0.33 mm), significantly higher than the average 0.74+/-0.17 mm in 63 age- and sex-matched controls.21 Comparing 121 HIV-positive and 27 HIV-negative people with a second cIMT a year later showed significantly greater progression in the HIV group (0.074+/-0.13 mm versus -0.006+/-0.05 mm). A nadir CD4 count at or below 200 cells/mm3 tended to predict cIMT progression (P = 0.082).
Age averaged a relatively young 45 years in this longitudinal comparison.21 Another longitudinal study tracked cIMT over time in HIV-positive children and young adults from 2 to 21 years old.22 Even at these tender ages, the 35-person HIV group had a significantly higher (and already abnormal) baseline cIMT than did 37 matched controls in both the internal carotid artery (0.90 versus 0.78 mm, P = 0.01) and the common carotid artery (1.00 versus 0.95 mm, P = 0.05). After 48 weeks of follow-up, though, cIMT of both arteries decreased significantly in the HIV group (-0.23 mm and -0.15 mm, P = 0.01 for both). Over those 48 weeks, CD4 percent rose and low-density lipoprotein (LDL) cholesterol fell in people with HIV, findings leading the authors to suggest that "lipid control, immune restoration, and viral suppression with continuous antiretroviral therapy" may prevent cIMT worsening in children and young adults.
Still, the prospect of nearly lifelong HIV infection and cART poses special concerns for children infected at birth or early in life. A 3035-child US study found that cardiomyopathy developed in 99 of them through a median 5.5 years of follow-up.23 Cardiomyopathy incidence stood at 5.6 per 1000 child-years, a rate 40 times higher than in the general population. Triple antiretroviral therapy halved the risk of cardiomyopathy, but taking zidovudine boosted cardiomyopathy risk 90%. (US pediatric antiretroviral guidelines list zidovudine as a preferred first-line antiretroviral for infants, children, and adolescents through puberty.24) Ongoing research on cardiovascular disease in HIV-positive children deserves special attention not only from pediatricians, but also from clinicians who will start caring for these youngsters when they reach their late teens and 20s.
Given the added heart risk burden HIV groups tote, can clinicians rely on risk formulas devised for the general population, like the Framingham Risk Score? The short answer seems to be no -- because the Framingham index does not account for important HIV-specific variables. The DAD Study group fashioned three HIV-specific risk tools -- one for myocardial infarction, one for coronary heart disease, and one for a composite endpoint.25 All three models proved more accurate than Framingham in a 22,625-person analysis. The final article in this issue of RITA! details differences between the Framingham and DAD tools. And in an interview in this issue, James Stein advises HIV clinicians on what they can learn from a Framingham score.
Like heart disease itself, aberrant lipids trouble a high fraction of all US residents, not just people with HIV. CDC head counters figure that one third of American adults have high LDL ("bad") cholesterol.26 That rate tops the 27% prevalence of high non-HDL cholesterol charted by HIV Outpatient Study (HOPS) investigators among US men in a survey of 3166 cART-treated men and women in care in 2006-2010.27 But 81% of these men had some sort of dyslipidemia, 41% had low HDL cholesterol, 32% had high triglycerides, and their median age was only 47. Among women in this study group, 67% had dyslipidemia of some sort, including 27% with low HDL cholesterol. More than half of these women had hypertension, 32% were obese, and their median age was only 45.
The HOPS study group had taken cART for a median of 6.8 years, and treatment almost certainly contributed to their bad lipid numbers. Current US antiretroviral guidelines list all ritonavir-boosted PIs, efavirenz, and abacavir (but not integrase inhibitors or the CCR5 antagonist maraviroc) as lipid mischief-makers (Table 2).17 But HIV itself, in antiretroviral-naive people, can send lipids off on wayward paths, boosting triglycerides and cutting HDL cholesterol.28 And studies comparing HIV-positive and negative groups consistently find worse lipid scores in people with HIV.2
|Table 2. Antiretrovirals Linked to Abnormal Lipids|
|Nucleoside Analogs||Nonnucleosides||Protease Inhibitors|
|Stavudine > zidovudine > abacavir||Efavirenz||All ritonavir-boosted PIs||LPV/r > DRV/r > ATV/r|
|↑ LDL and TG||↑ TG, LDL, and HDL||↑ LDL, HDL, and TG||↑ TG|
Source: Panel on Antiretroviral Guidelines for Adults and Adolescents.17
ATV, atazanavir; DRV, darunavir; HDL, high-density lipoprotein cholesterol; LDL, low-density lipoprotein cholesterol;
LPV, lopinavir; r, ritonavir; TG, triglycerides
Lofty triglycerides do heighten myocardial infarction risk in people with HIV, according to a 33,308-person DAD Study analysis.29 The DAD team figured that the overall impact of high triglycerides on MI risk is small -- though still independent of other factors -- when the analysis included those other factors.
This DAD analysis followed people enrolled in this European-American-Australian cohort at some point from 1999 through 2008. During 178,835 person-years of follow-up, the investigators recorded 580 MIs. Every triglyceride doubling upped the MI risk 67% in an unadjusted analysis. Layering on one statistical adjustment after another, the DAD team found that relative risk fell with each adjustment but remained independent of other risk factors:
MI risk per triglyceride doubling with HIV:
Because the relative MI risk dwindled when the analysis considered other risk factors, the DAD team questions whether drugs that cut triglyceride levels would make a dent in MI incidence among people with HIV.29 These investigators note that European AIDS Clinical Society Guidelines do not recommend niacin or fibrates to treat high triglycerides in people with HIV.30 In a review of cardiovascular risk and capricious lipids in people with HIV, US cardiologist James Stein suggests high triglycerides should become a target of lipid-lowering therapy only if levels exceed 500 mg/dL, when pancreatitis poses a threat.31 (Stein's review, accessible online, is loaded with advice on managing dyslipidemia in people with HIV. He also addresses lipid control in the interview in this issue.)
Of course lipid values other than triglycerides sway MI risk. In the DAD analysis considering triglycerides, total cholesterol, and HDL cholesterol at the same time (second bullet above), every mmol/L (39 mg/dL) higher total cholesterol boosted MI risk 26% (RR 1.26, 95% CI 1.20 to 1.32, P < 0.001) and HDL cholesterol below 0.9 mmol/L (35 mg/dL) doubled the risk (RR 2.02, 95% CI 1.39 to 2.95, P < 0.001).29
US HIV/heart guidelines from the HIV Medicine Association and allied groups are a decade old, but little has changed to affect bedrock management principles: measure fasting lipids before people start cART and within 3 to 6 months after starting a new combo.28 If changing diet, exercise, and smoking habits doesn't control lipids, start statins (that don't interact with prescribed antiretrovirals) for high LDL or non-HDL cholesterol and fibrates for lofty triglycerides. These guidelines are linked at reference 28 below.
In his 2012 lipid review, James Stein opines, "if there is a single take-home message about treating dyslipidemia to reduce [coronary heart disease] risk [in people with HIV], it is to put patients on statin therapy."31 Simply stated, Stein observes, statins saves lives: A meta-analysis of statins versus no statins in 160,000 people in the general population found that every 39-mg/dL (1 mmol/L) drop in LDL cholesterol with statins over 5 years trimmed all-cause mortality 10%, coronary heart disease (CHD) mortality 20%, MI and CHD mortality 26%, and major cardiovascular events 21%.32
Recent research links statin therapy to a higher risk of diabetes in the general population33,34 and in people with HIV.35 In all these studies the statin-related diabetes risk was small and apparently outweighed by the cardiovascular benefits of these drugs. Another recent study tied statin use to lower all-cause mortality in 25,884 people with cART-induced virologic suppression.36
How well do US clinicians follow lipid therapy guidelines in people with HIV? The report card features some high marks and some low marks, at least for clinicians seeing people in the HIV Outpatient Study cohort from 2002 through 2009.37 Among more than 1300 cohort members who had their 10-year cardiovascular risk figured, 28% had less than a 10% 10-year risk, 18% had a 10% to 20% risk, and 20% had a 10-year risk above 20%. Using National Cholesterol Education Program Adult Treatment Panel III (NCEP) guidelines as the standard, the HOPS team found that 81% to 87% of eligible patients got treated for high LDL/non-HDL cholesterol and 56% to 91% got prescriptions for high triglycerides. But only 2% to 11% took lipid drugs for low HDL cholesterol, and only 46% to 69% who needed anti-hypertensives got them. The investigators concluded that "a large percentage of at-risk patients who were eligible for pharmacologic treatment did not receive recommended interventions and did not reach recommended treatment goals."37
Abnormal lipids -- and what to do about them -- have preoccupied HIV clinicians and researchers since the first report of coronary artery disease in cART-treated people.1 But early on it became clear that flaring lipids are hardly the only heart worry in people with HIV -- and hardly the only trigger for vascular "events." Among classic cardiovascular risk factors, hypertension and diabetes represent two of the most treatable conditions.
The CDC estimates that one third of Americans have high blood pressure.38 An HIV Outpatient Study analysis logged even higher rates in US men and women with HIV.27 The 3166 people studied had a median age of 47 years and had taken cART for a median of 6.8 years; 21% were women and more than half smoked or used to smoke. Similar high proportions of women (57.4%) and men (54.4%) had hypertension. Almost one third of these women were obese.
Untreated or inadequately treated hypertension has a profound impact on morbidity and mortality (and not only cardiovascular mortality) in people with HIV. A DAD Study analysis of 33,308 HIV-positive cohort members figured that current hypertension doubled the risk of cardiovascular death (adjusted relative rate [aRR] 2.04, 95% CI 1.57 to 2.66) and more than doubled the risk of liver death (aRR 2.34, 95% CI 1.83 to 2.99).5 Hypertension also independently hoisted chances of all-cause mortality and AIDS mortality.
Meta-analysis of 2242 HIV-positive people in 11 studies determined that hypertension independently magnified the odds of left ventricular dysfunction almost as much as 10 years of age.39 The adjusted odds ratio for hypertension stood at 2.3 (95% CI 1.2 to 4.5), compared with 2.5 (95% CI 1.70 to 3.6) for every decade of age.
A Swiss HIV Cohort Study analysis of 2595 people with HIV and confirmed hypertension calculated that every 10 mm Hg higher systolic blood pressure boosted the risk of cardiovascular disease 18% (hazard ratio 1.18, 95% CI 1.06 to 1.32).40 "Insufficient control of hypertension was associated with increased risk for cardiovascular events," the Swiss team noted, "indicating the need for improved management of hypertension in HIV-infected individuals."
Diabetes affects 11.3% of US residents 20 years old or order, according to a 2011 CDC estimate.41 In contrast, big HIV cohort studies in Europe record much lower diabetes prevalence: 2.5% of 17,852 DAD Study members (from Europe, Australia, and Israel),42 2.7% of 8033 Swiss HIV Cohort Study participants,43 and 3% of 394 HIV-positive people at a London hospital.44
Compared with these European cohorts, US studies tabulate much higher diabetes prevalence in HIV-positive people -- perhaps reflecting the older age in these US groups than the European groups (Figure 4) and the high diabetes rate in the US population at large.41 A Multicenter AIDS Cohort Study (MACS) comparison of 534 men with HIV and 322 at-risk men without HIV charted an 11.4% diabetes prevalence in the HIV group and an 8.0% rate in the HIV-negative group, a nonsignificant difference (P = 0.16).45 But average age was significantly younger in the HIV group (48.9 versus 52.6, P < 0.0001).
Diabetes prevalence was much lower in three European HIV cohorts42-44 than in three US HIV cohorts45-47 or the US population at large.41 Younger age in the European groups than the US groups could partly explain the difference between cohorts, but US national data41 indicate that diabetes is highly prevalent throughout the US, at a rate of 11.3%. In comparison, Diabetes UK estimates that 2.5 million people in England had diabetes in 2012, or 4.7% of the 53 million people in England.48 (DAD data from Europe, Israel, Australia; SHCS, Swiss HIV Cohort Study; London, Chelsea and Westminster Hospital; MACS, Multicenter AIDS Cohort Study; VACS, Veterans Aging Cohort Study; WIHS, Women's Interagency HIV Study. Ages are medians, mean, or range [for USA]. USA national estimate from CDC.41)
A Veterans Aging Cohort Study of 3227 vets with HIV and 3240 without HIV found a significantly lower diabetes prevalence in the HIV group (14.9% versus 21.4%, P < 0.0001),46 though prevalence in this largely male HIV-positive contingent was higher than among HIV-positive men in the MACS analysis.45 Ages averaged 49.6 in the veterans HIV group and 50.8 in the HIV-negative group (P < 0.001). A recent Women's Interagency HIV Study (WIHS) survey logged a diabetes prevalence of 12.3% in 1797 women with HIV and 14.0% in 679 without HIV.47 These HIV-positive and negative women had median ages of only 39 and 35 and a collective body mass index in the overweight range.
Whether HIV and cART confer a higher diabetes risk remains open to question -- at least for men. MACS and WIHS studies from the mid-2000s differed in determining the impact of HIV on diabetes risk -- MACS findings a higher diabetes risk with HIV in men49 and WIHS finding no higher diabetes risk with HIV in women.50 The Veterans Aging Cohort Study discerned a lower diabetes prevalence with HIV than without HIV in a mostly male population.46
At least two factors contribute to these seemingly contradictory results -- what antiretrovirals people are taking and how the researchers define diabetes. For example, a nationwide French study of 1046 HIV-positive people charted a diabetes incidence of 14.1 per 1000 person-years.51 Incidence peaked in 1999-2000 at 23.2 and fell afterwards, a turnaround at least partly reflecting abandonment of indinavir, stavudine, and didanosine, all of which heightened diabetes risk in this analysis. CD4 count, CD4/CD8 ratio, and viral load did not affect diabetes risk, but traditional risk factors did (older age, overweight, and waist-to-hip ratio).
The French team defined diabetes by a confirmed high blood glucose and/or starting anti-diabetic medication.51 The two US studies that found a higher diabetes risk with than without HIV in men49 but not in women50 relied on a single blood glucose level (or anti-diabetic medication or a clinical diagnosis). The 97.5% male veterans study that discerned a lower diabetes risk with HIV relied on a confirmed high blood glucose (or other dual clinical criteria).46 And a Swiss HIV Cohort Study analysis, which used confirmed fasting glucose to define diabetes, observed similar age- and gender-specific diabetes incidence in HIV-positive cohort members and in a population-based cohort of HIV-negative people.52
When WIHS researchers updated their diabetes incidence analysis using multiple confirmed diagnostic criteria, they found an independently higher incident diabetes risk in HIV-positive women when diagnosis depended on a confirmed high blood glucose.47 HIV doubled the diabetes risk in these women after statistical adjustment for age, body mass index, and other variables. The WIHS investigators cautioned that relying on an unconfirmed blood glucose can result in an overestimate of diabetes incidence.
The Swiss study buttressed earlier work linking incident diabetes to nucleosides with or without protease inhibitors -- but not to nucleosides plus nonnucleosides.52 Among protease inhibitors the association held true for the first-generation protease inhibitor indinavir, but not for atazanavir or lopinavir. Three nucleoside combinations -- none used routinely today -- upped the risk of incident diabetes: didanosine/stavudine, stavudine/lamivudine, and didanosine/tenofovir. Reviewing all recent antiretroviral data, US guidelines list diabetes or insulin resistance as a side effect of three nucleosides (zidovudine, stavudine, and didanosine) and two protease inhibitors (indinavir and lopinavir/ritonavir).17
Regardless of whether HIV makes diabetes more likely in women, men, or both, no one doubts the potentially deadly impact of this chronic and often poorly controlled disease. The 33,308-person DAD study analysis that linked current hypertension to higher death rates from cardiovascular disease, liver disease, AIDS, and all causes also found that current diabetes independently raised the risk of death in those four categories.5
One reason HIV-positive people in the United States have high rates of hypertension and diabetes (see preceding section) is the growing girth of the populace at large. Besides causing or contributing to hypertension and diabetes, obesity heightens the risk of wanton lipids, coronary heart disease, and stroke.53 The CDC figures more than one third of US adults and 17% of children are obese. In 2000, the CDC reports, no state had an obesity prevalence topping 30%; in 2010, 12 states had crossed that line.54
In fact, a recent CDC analysis found a higher obesity prevalence in the general US population than in a nationally representative sample of people with HIV.55 This study focused on 4040 HIV-positive adults in 23 health departments across the United States, comparing them with people in the 2009-2010 National Health and Nutrition Examination Survey (NHANES). Obesity (body mass index above 30 kg/m2) affected 35.7% of NHANES participants and 22.8% of people with HIV. Age-adjusted obesity prevalence in HIV-positive women exceeded the general population rate (40% versus 36%), but HIV-positive men had an obesity rate less than half that of general-population men (17% versus 36%). Nearly half of HIV-positive women under 40 years old (45%) were obese.
Obesity prevalence fell with age in women with HIV and rose with age in the general population (Figure 5).55 Women with HIV ran a twice higher risk of obesity than HIV-positive men (adjusted prevalence ratio 2.12, 95% CI 1.87 to 2.41). Less education and less advanced HIV infection also made obesity more likely.
CDC analysis of a nationally representative sample of people with HIV found that obesity prevalence dwindled with age in HIV-positive women but rose with age in women in the general population (NHANES).55 Linear trend for age P < 0.01 for both groups.
Cohort studies verify the savage impact of high weight and visceral fat on cardiovascular risk in people with HIV. FRAM study investigators compared 586 HIV-positive men and women with 280 HIV-negative controls, calculating regional body fat by whole-body magnetic resonance imaging and figuring heart risk with the Framingham Risk Score.56 Splitting visceral adipose tissue (VAT) levels into four quartiles, they found that the median Framingham score rose with higher VAT quartiles in people with and without HIV. But in each VAT quartile the Framingham score was significantly or nearly significantly higher in the HIV group than in the HIV-negative group. "Increased VAT is associated with cardiovascular disease risk" with or without HIV infection, the FRAM team concluded, "but the risk is higher in HIV-infected individuals relative to controls at every level of VAT."56
Overweight and obese people with HIV also heft a heavier burden of other morbidities familiar to HIV clinicians, according to a 1833-person study at the University of Alabama at Birmingham.57 Earlier work by this team uncovered a 45% prevalence of overweight and obesity (>25 kg/m2) among HIV-positive men and women before they began cART in this clinic.58 The newer study classified participants as underweight, normal weight, overweight, or obese and grouped 15 common non-HIV conditions into three clusters -- metabolic, behavioral, and substance use. While 35% of participants were underweight or normal weight, 36% were overweight and 29% obese. Obesity independently predicted having one or more conditions in at least two of the disease clusters (adjusted odds ratio 1.52, 95% CI 1.15 to 2.00). The University of Alabama team urged colleagues to "embrace HIV care as complex chronic disease management of multiple overlapping conditions within the context of primary care."57
Italian and Canadian researchers proposed one step toward that daunting goal. They devised a simple tool combining triglycerides (TG) and waist circumference (WC) that predicted a higher Framingham Risk Score -- as well as higher VAT and rates of metabolic syndrome and type 2 diabetes -- in 1481 men and 841 women with HIV in an Italian study group (Figure 6).59 Researchers divided people into four groups: low WC/low TG, low WC/high TG, high WC/low TG, and high WC/high TG using cutoffs of ≥90 cm and ≥2.0 mmol/L (177 mg/dL) for men and ≥85 cm and ≥1.5 mmol/L (133 mg/dL) for women. Men in the high TG/high WC group had the most VAT (208 cm2), the highest Framingham score (10.3), and the highest prevalence of metabolic syndrome and type 2 diabetes, when compared with other groups of men. Women in the high TG/high WC box also had elevated VAT (average 150 cm2) as well as the highest Framingham score (2.9) and the highest rates of metabolic syndrome, hypertension, and type 2 diabetes, when compared with other groups of women.
A foursquare tool dividing HIV-positive men (M) and women (W) into four groups according to high or low triglycerides (TG) plus high or low waist circumference (WC) predicted Framingham Risk Score, type 2 diabetes, hypertension, metabolic syndrome, and visceral adipose tissue in a 2322-person study.59 (Waist circumference in cm; triglycerides in mmol/L; 2 mmol/L = 177 mg/dL; 1.5 mmol/L = 133 mg/dL).
A North Carolina comparison of 92 HIV-positive adults and 92 age-matched HIV-negative people found that overweight/obesity prevalence in the HIV group climbed from 52% to 66% during the first 12 months of cART, a relative increase of 27% (P = 0.002).60 HIV-positive women gained significantly more weight than men, and people starting a protease inhibitor regimen gained significantly more than those starting other regimens. People who began cART with fewer than 200 CD4 cells/mm3 added significantly more pounds than those starting at higher CD4 counts. Nearly everyone in the HIV-negative group, 93%, was overweight or obese at the start of follow-up, and that rate did not change during the study.
Diet and exercise -- or at least supplanting a sedentary lifestyle with some vigorous pursuits -- offers the surest path to weight reduction while often tempering cardiovascular risk. Research shows that structured exercise programs can cut fat and build lean body mass. But because most exercise studies in people with HIV are small and completion rates often modest, this article focuses on diet and its impact on heart disease.
Whether obese, overweight, or normal weight, many American have bad diets, a failing glaringly reflected in a study of 265 men and 56 women with HIV in Boston and Providence.61 About 3 in 10 women and 1 in 10 men were obese, while one third of women and 40% of men were overweight. Figuring dietary intake by 3-day food records, the researchers found that total fat and saturated fat intakes exceeded US recommendations for both men and women in all body mass index categories.
Heavier people did not eat more than normal-weight people, but they ate worse, wresting less energy from every kilocalorie (kcal) gulped: average energy intake per kilogram waned significantly from normal weight to overweight to obese in women (33 to 25 to 19 kcal) and in men (40 to 33 to 28 kcal) (Figure 7).61 Diets of overweight and obese people contained significantly less fiber than diets of normal-weight people among both women (11.3 to 9.3 to 6.9 g for normal, overweight, and obese women) and men (13.2 to 12.8 to 11.7 g) (Figure 7). A low-fiber diet bespeaks a lack of whole grains, fruits, vegetables, nuts, and seeds.
As body mass index (BMI) category rose from normal, to overweight, to obese in a study of 265 men (M) and 56 women (W) with HIV, (1) average energy intake per kilogram fell significantly, and (2) median fiber content dipped significantly.61 (See text for exact values.)
Worse diets in heavier people in this study probably contributed to three factors intimately linked to cardiovascular risk -- significantly worse insulin resistance in both men and women (Figure 8), and significantly higher triglycerides and total cholesterol in men.61
HIV-positive women and men in each higher body mass index (BMI) category had a higher prevalence of insulin resistance (HOMA IR >3.5).61
A study comparing 356 HIV-positive adults with 162 HIV-negative people in the same community determined that, despite consuming similar shares of calories, the HIV group ate significantly more total fat, saturated fat, and cholesterol.62 Using 4-day food records or 24-hour recall, this Boston study of 197 men and 159 women with HIV also found that the HIV group derived a significantly higher percentage of calories from saturated fat and trans fat than the 73 men and 89 women in the HIV-negative group. Triglycerides rose 8.7 mg/dL for each gram of fat an HIV-positive person swallowed (P = 0.005).
People who improve their diets reap health benefits, plentiful research attests. Anyone who had occasion to browse the internet or scan a newspaper in the past few months will know results of the randomized clinical-endpoint PREDIMED trial pitting a Mediterranean diet against advice to eat a low-fat diet: People in the two Mediterranean diet groups (supplemented by extra-virgin olive oil or additional nuts) had a 30% lower risk of myocardial infarction, stroke, or death from cardiovascular disease than the low-fat group after only 4.8 years of follow-up.63
Adherence to the Mediterranean diet prescribed in this trial (Table 3) was good. The PREDIMED researchers believe their striking results "are particularly relevant given the challenges of achieving and maintaining weight loss."63 Earlier, a systematic review rated a Mediterranean diet the type of diet most likely to ward off coronary heart disease in the general population.64
|Table 3. Mediterranean Diet Prescribed in the Spanish PREDIMED Trial63|
|Olive oil||Soda drinks|
|Tree nuts and peanuts||Commercial bakery goods, sweets, pastries|
|Fresh fruits||Spread fats|
|Fish (especially fatty fish), seafood||Red and processed meats|
|Sofrito (tomato and onion sauce)|
|Wine with meals (only for habitual drinkers)|
This widely lauded study may have special pertinence for people with HIV because study participants had a high risk of heart disease but a clean cardiovascular slate when they entered the trial. A comparison of risk factors shows, though, that PREDIMED participants were a whole lot closer to a heart attack than 33,308 antiretroviral-treated DAD Study participants in 2010:5 The PREDIMED contingent ran a higher heart risk by age (67 versus 39 in DAD), body mass index (29 versus 23 kg/m2), hypertension prevalence (82% versus 14%), and diabetes prevalence (48% versus 3%). The DAD cohort had a twice higher proportion of current smokers (35% versus 14%). More than half of PREDIMED study participants, 57%, were women, and 97% were white Europeans. In the DAD study 26% were women and 54% white. So whether the profound cardiovascular benefit seen with a Mediterranean diet in PREDIMED would hold true in contemporary HIV populations -- at least over the short term -- remains an open question.
A few studies have appraised Mediterranean fare in people with HIV. A pilot randomized trial in HIV-positive Hong Kong patients found pluses and minuses with a 12-month Mediterranean diet versus a low-fat, low-cholesterol diet.65 Of the 48 people randomized to one diet or the other, 36 (75%) completed 12 months of follow-up, which included regular dietary consultation. Dietary adherence was good, and use of specific nucleosides and protease inhibitors was similar between study arms. People in the low-fat/cholesterol crew had unfavorable body fat changes in triceps skinfold, hip circumference, and waist-to-hip ratio. Triglycerides rose in the low-fat/cholesterol group while remaining unchanged in the Mediterranean group. The Mediterranean arm had significant jumps in total cholesterol at 9 months (P = 0.03) and 12 months (P = 0.01), whereas the low-fat group did not. Because of missing data, the researchers did not analyze HDL and LDL cholesterol, so the total cholesterol findings are hard to interpret.
A larger cross-sectional US study linked Mediterranean eating habits to improvements in three heart disease indicators -- insulin resistance, triglycerides, and HDL cholesterol.66 This study involved 247 HIV-positive people with abnormal fat distribution seen at a Boston center, all of whom had complete metabolic profiles available. Researchers figured how closely their diet fit a Mediterranean plan by calculating a Mediterranean Diet Score (MedDietScore).67 A higher MedDietScore meant (1) a lower rate of insulin resistance (standardized beta -0.15, P = 0.03), (2) marginally lower triglycerides (standardized beta -0.16, P = 0.13), and (3) higher "good" HDL cholesterol (standardized beta 0.15, P = 0.01).
Two studies in Croatia gauged the impact of a Mediterranean diet and other variables on lipids and body fat in the first year of cART.68,69 Both studies relied on a 150-item questionnaire to rank people in a low Mediterranean adherence group (below 4 points on a 0-to-9 scale) or a moderate to high adherence group (4 to 9 points). Analysis of 117 people interviewed between May 2004 and June 2005 discerned no link between Mediterranean diet and serum lipids.68 Notably, people in this study were still taking lipid-malefic antiretrovirals such as indinavir and stavudine, both of which were associated with higher total cholesterol.
A similar questionnaire-based dietary analysis of 136 Croatians in the first year of cART during the same period focused on lipoatrophy and lipohypertrophy assessed by self-report and physical exam.69 Compared with nonsmoking participants with a moderate to high Mediterranean diet score, nonsmokers with a low diet score had nonsignificantly higher odds of lipoatrophy (adjusted odds ratio 4.53, 95% CI 0.86 to 23.92, P = 0.076), while smokers with a low diet score had significantly higher lipoatrophy odds (adjusted OR 3.42, 95% CI 1.21 to 9.67, P = 0.014), as did smokers with a moderate to high diet score (adjusted OR 4.39, 95% CI 1.35 to 14.26, P = 0.021). People with a moderate to high Mediterranean diet score had 70% lower odds of lipohypertrophy (adjusted OR 0.3, 95% CI 0.1 to 0.7, P = 0.012).
Although these small studies65,6668,69 hint that Mediterranean meal planning can score cardiovascular pluses for people with HIV, hints are all they provide (Table 4). Results of the randomized PREDIMED trial63 strongly suggest that HIV-positive people with a heart risk as high as people in this trial can ward off ischemic heart disease by eating more olive oil, nuts, fish, and fresh produce.
|Table 4. Impact of Mediterranean Diet on Heart Risk Factors in HIV Studies|
|Body Fat||Triglycerides||Cholesterol||Insulin Resistance|
But will other healthy diets, leavened with a little exercise, do as well? A small randomized US trial found heart marker benefits with a 6-month low-fat/high-fiber diet plus 3 hours of physical activity weekly.70 This trial randomized 34 HIV-positive adults with metabolic syndrome to physical activity plus counseling that emphasized a diet low in saturated, polyunsaturated, and trans fat and high in omega 3 fatty acids and fiber or to a control group whose members got monthly counseling sessions on healthy eating. After 6 months the intervention group did significantly better on measures of waist circumference, systolic blood pressure, hemoglobin A1C, lipodystrophy score, and activity measured by the Modifiable Activity Questionnaire. Lipids did not improve significantly in the intervention group compared with the control group. And as in many studies of diet and/or exercise, getting people to stick to the program was not easy. Four people dropped out of the intervention arm and 2 left the control arm for an overall 6-month dropout rate of 18%.
"Aside from having a history of cardiovascular disease," write HIV heart expert James Stein and colleagues, "current cigarette smoking is the most powerful predictor of CVD events among patients with HIV."71 Everyone knows that lots of people with HIV smoke more than the most leather-lunged film noir antihero, and everyone knows tobacco sears a deeply corrosive path through many a major organ. The data are so uniform and unequivocal they hardly bear repeating. But this article will detail some of these dreary numbers -- and offer a few suggestions on getting people to quit -- in hopes that some clinician readers will pluck out a fact or two that will scare patients into stopping, or that some clinician readers will recommit themselves to the difficult chore of helping patients quit.
Why so many HIV-positive people smoke remains unclear. An unadorned hunch is simply that the groups most likely to get infected with HIV include a high fraction of smokers. There seems to be no evidence that people start smoking because they learn they have HIV infection.
The first nationally representative estimate of smoking prevalence in US residents with HIV, unveiled in 2013 by the CDC, found that 42% of HIV-positive people in care smoke, compared with 21% of the US general population.72 While 37% of people with HIV never smoked, 58% of the general population never lit up. Men made up the biggest proportion of current HIV-positive smokers (72%), while women accounted for 27% of that group and transgenders for 1%. Blacks accounted for 43% of current HIV-positive smokers, followed by whites (36%), Hispanics (16%), and others (5%). Statistical analysis adjusted for age determined that HIV-positive people in care have a twice higher smoking prevalence than the general population (standardized prevalence ratio 1.9). The prevalence ratio stayed near that mark after individual adjustment for gender (1.9), race/ethnicity (2.1), education level (2.0), and poverty level (1.7).
SMART trial investigators recorded a current smoking rate of 40.5% in 5472 HIV-positive participants from 33 countries73 -- nearly the same as the CDC's US estimate.72 SMART researchers tallied a 24.8% former-smoker rate in this group, compared with 20% in the CDC study. The SMART analysis went on to address a bigger question: what does smoking do to people with HIV? To find an answer they (1) figured hazard ratios for major clinical endpoints by comparing SMART participants who smoked at study entry with participants who never smoked and (2) calculated population-attributable risk percentages (see note 7474) for the same endpoints.
Compared with SMART enrollees who never got a nicotine high, current smokers ran more than a doubled risk of dying from any cause (adjusted hazard ratio [aHR] 2.4, P < 0.001), a doubled risk of incident major cardiovascular disease (aHR 2.0, P = 0.002), more than a doubled risk of bacterial pneumonia (aHR 2.3, P < 0.001), and almost a doubled risk of non-AIDS cancer (aHR 1.8, P = 0.008).73Quitting made a difference. Comparing former smokers with current smokers, the SMART team found significantly higher adjusted hazard ratios for all-cause mortality (1.5, P = 0.04), major cardiovascular disease (1.6, P = 0.02), AIDS-related disease (1.6, P = 0.03), non-AIDS cancer (2.3, P < 0.001), and bacterial pneumonia (1.5, P = 0.01) in current smokers.
Comparing current smokers with former and never smokers indicated that 24.3% of all deaths could be attributed to smoking, as could 25.3% of major cardiovascular diagnoses, 30.6% of non-AIDS cancer diagnoses, and 25.4% of bacterial pneumonia diagnoses (Figure 9).74 The SMART team warns that "significant reductions in morbidity and mortality among HIV-infected patients achieved by advances in HIV therapy may be undercut by increases in adverse clinical outcomes attributable to smoking."73
Among START trial participants who died or got diagnosed with a major disease, smoking explained about one quarter of deaths or new diagnoses.73
In a group 6 times bigger than the SMART population, DAD Study investigators linked current smoking to a 2.2 times higher rate of death from non-AIDS cancers, a 90% higher rate of cardiovascular death, and a 44% higher rate of death from any cause.5 Unlike the SMART analysis, the DAD team found that former smokers matched current smokers in adjusted death rates for non-AIDS cancer, cardiovascular disease, or any cause. "Studies in the general population suggest that although smoking cessation leads to decreased risk" of death, the DAD team noted, "ex-smokers remain at an excess risk for a number of years after cessation, with the risk of malignancy in particular remaining raised for up to 10 years."5
How many years do risks of heart disease and death stay elevated after a person stops smoking? Another DAD analysis suggested not too many.75 This study involved 27,136 HIV-positive people with a reported smoking status, divided into never, previous, and current smokers. None had a history of heart disease. The researchers counted how many people had (1) a myocardial infarction, (2) coronary heart disease (MI plus invasive coronary artery procedure or death from other coronary heart disease), (3) cardiovascular disease (coronary heart disease plus carotid artery endarterectomy or stroke), and (4) death from any cause. Poisson regression analysis to determine how quitting affected these endpoints factored in cohort, calendar year, age, sex, family heart disease history, time-updated diabetes and lipids, cumulative cART, cumulative indinavir, cumulative lopinavir, and current abacavir. Every additional year without smoking trimmed the risk of the three cardiovascular endpoints significantly (or nearly significantly, P < 0.06) (Figure 10). This analysis found a nonsignificant trend to lower all-cause mortality with time since quitting.
With each year since DAD Study participants quit smoking, incidence rate ratios for myocardial infarction (MI), coronary heart disease (CHD), and cardiovascular disease (CVD) (see text for definitions) declined significantly or nearly significantly.75
A nationwide study of everyone in Denmark with diagnosed HIV infection and a matched group of HIV-negative people figured that a 35-year-old smoker with HIV could expect to live to the age of 62.6, whereas a 35-year-old HIV-positive nonsmoker would probably celebrate his 78th birthday.76 These researchers calculated a much higher death risk due to smoking than figured in the SMART study (61.5% versus 24.3%). Among people without HIV in the Danish study, smoking could explain 34.2% of all deaths. HIV-positive people who smoked had more than a 4.4 times higher risk of death than HIV-positive people who never lit up. The death risk was only 1.7 times higher in HIV-positive former smokers than in people who never smoked.
Smoking appears to abuse arteries more in people with HIV than in HIV-negative people, according to results of a study comparing carotid intima-media thickness (cIMT, Figure 3) in 166 men and women with HIV and 152 healthy HIV-negative people.77 Multivariate regression modeling that considered gender, race, and classic heart risk factors identified a significant three-way interaction between age, smoking burden, and HIV status with respect to cIMT (P < 0.01). This interaction indicated that more smoking and older age had a bigger negative impact on cIMT in people with than without HIV.
Many a clinician will attest familiarity with the foregoing dizzying data -- or similar findings that crowd the medical literature. And most HIV-positive smokers say they're already thinking of quitting.78 Yet HIV providers in a US veterans study were less likely than non-HIV providers even to know if their patients smoked.79 This 143-provider analysis determined that infectious disease specialists were almost 3 times more likely than generalists not to know whether a patient smoked. If findings like these hold true for HIV providers outside the Veterans Affairs system, educating providers seems a good place to start a smoke-ending campaign in people with HIV.
HIV heart maven James Stein and coauthors advise colleagues to take the 5A approach to encouraging patients to break the smoking habit (Figure 11).71 Because effective medications have become available, Stein and colleagues believe "pharmacotherapy is the preferred approach to smoking cessation." Nicotine replacement products now come in gum, lozenges, transdermal patches, inhalers, and sprays.73 In the interview in this issue, Stein notes that dual pharmacologic therapy -- with a nicotine patch and lozenge -- proved most effective in one randomized trial. Bupropion or varenicline may work for people who have no success with nicotine replacement. SMART investigators caution, though, that these drugs may interact with antiretrovirals, so consulting an HIV pharmacologist or a reliable drug interaction website is prudent.73
HIV heart sages recommend the 5As for getting patients to conquer their nicotine habit.71
Clinicians may not realize that the Department of Health and Human Services rounded up 24 physicians and scientists who assembled 276 pages of guidelines on treating tobacco use and dependence.80 Providers who prefer not to scour every page of that report would do well to scan the "ten key guideline recommendations" on pages 6 through 8, accessible at the link in the reference list.80 The National Heart Lung and Blood Institute offers a straightforward online patient-directed guide, "Strategies to quit smoking."81
SMART investigators stress that smokers often make many attempts to stop before succeeding.73 Smokers with HIV should know this so they can muster the resolve to try again after one or a few futile tries. And clinicians should know this so they don't quit supporting would-be quitters. Stein and colleagues observe that the best time to campaign against smoking is before a patient starts.71 Avoiding nicotine is easier than subduing the addiction, and years of cumulative damage can be sidestepped. Clinicians caring for adolescents with HIV, take note.
Besides tobacco use, two other substance problems figure mightily in cardiovascular risk. The CDC lists one of them, alcohol,8 but not the other, cocaine.
Getting drunk boosts chances of transmitting or acquiring sexually transmitted infections, including HIV. The CDC offers a sobering list of long-term health risks from drinking, including dementia, stroke, and neuropathy; depression, anxiety, and suicide; liver, colon, mouth, throat, and esophageal cancer; alcoholic hepatitis and cirrhosis; pancreatitis and gastritis; miscarriage and stillbirth; and cardiovascular disease -- including myocardial infarction, cardiomyopathy, atrial fibrillation, and hypertension.82
Research in the general population shows that, compared with abstinence, low to moderate drinking (20 g daily) eases the risk of coronary heart disease, while heavy drinking (70 g daily) hikes the risk.83 This J-shaped curve describing the relation between alcohol volume and heart risk starts its upward swing into the danger zone between 25 and 50 g daily. Five ounces of wine, 12 ounces of beer, and 1.5 ounces of 80-proof liquor contain about 14 g (1.2 tablespoons) of alcohol,82 so routinely downing four drinks puts one on the road to perdition.
The most illuminating study of alcohol and heart disease in people with HIV found that heavy drinking boosts cardiovascular risk in HIV-positive men -- perhaps more than in HIV-negative men.84 This Veterans Aging Cohort Study (VACS) analysis involved 4743 HIV-positive and negative (and demographically similar) veterans. The analysis excluded women and lifetime abstainers. The VACS team defined infrequent or moderate drinking as 14 or fewer drinks weekly and no binge drinking; hazardous drinking meant more than 14 drinks weekly or binge drinking. The VACS team defined cardiovascular disease by self-report survey and ICD-9 codes. The 2422 HIV-positive and 2321 HIV-negative men both averaged about 50 years in age; about two thirds were black and one quarter white.
Compared with HIV-negative vets, the HIV-positive group had higher proportions of infrequent or moderate drinkers (45.9% versus 42.9%) and current hazardous drinkers (33.2% versus 30.9%), but a lower proportion who ever had an alcohol dependence diagnosis (20.9% versus 26.2%) (P < 0.001). The HIV-positive group included significantly lower proportions with key heart risk factors -- hypercholesterolemia, diabetes, hypertension, current smoking, and high body mass index.
Among HIV-positive vets, hazardous drinking and alcohol abuse and dependence each independently boosted the odds of cardiovascular disease prevalence about 50%, compared with infrequent or moderate drinking, after adjustment for age, race/ethnicity, traditional heart risk factors, HCV and liver disease, kidney disease, exercise, education, CD4 count, and adherence to cART:
Notably, the links between heavy drinking and heart disease did not hold true in HIV-negative vets, a finding "suggesting the effect of alcohol may be more pronounced among those infected with HIV."84
In HIV-positive vets, a familiar list of classic risk factors also hoisted odds of heart disease in this analysis, including older age, high cholesterol, diabetes, hypertension, and current smoking. Kidney disease (defined as glomerular filtration rate below 30 mL/min) more than doubled chances of cardiovascular disease (aOR 2.39, 95% CI 1.24 to 4.61). Regular exercise lowered odds of heart disease almost 20%, though that association fell short of statistical significance (aOR 0.81, 95% CI 0.62 to 1.05).
More than a few gay and bisexual men, injection drug users, and other substance abusers use cocaine. Some clinicians may not realize that many HIV-positive US women also use cocaine and crack cocaine. A study of 1686 HIV-positive women enrolled in the Women's Interagency HIV Study in 1996-2004 found that 29% used crack during the study period.85 An analysis controlling for other risk factors figured that persistent crack users ran more than a 3 times higher risk of AIDS death than nonusers. Persistent crack users also lost more CD4 cells and had higher viral loads than women who did not use crack.
A longitudinal study of 736 gay and bisexual men in San Francisco found that cocaine, methamphetamine, and popper use declined over 48 months in older men but rose during the same period in younger men.86 Compared with men who did not use these drugs, those you used them less than weekly or at least weekly were more likely to have condom-free anal sex with an HIV-positive or status-unknown partner.
HIV clinicians who care for youngsters should realize cocaine habits can start in high school. The national Youth Risk Behavior Survey found that more than 1 in 20 youngsters in the 9th to 12th grades used cocaine at least once in 1993.87 That rate rose to almost 1 in 10 in 1999, then drifted down to about 7% in 2011.
Cocaine is not kind to the heart. A general-population study of 479 people 50 and younger admitted to the coronary care unit at a Barcelona hospital found that cocaine use prevalence vaulted from 6.8% in 2001 to 21.7% in 2008 (P = 0.035).88 People younger than 30 had more than a 4 times higher cocaine use rate by urine testing than people 45 to 50 (18.2% versus 4.1%, P = 0.035). Cocaine users had bigger MIs (by troponin I level) than did nonusers, and more cocaine users died in the hospital (8.3% versus 0.8%, P = 0.030).
Two studies of HIV-positive cocaine users at Johns Hopkins University uncovered evidence linking both HIV and cocaine use to coronary artery calcification, an early stage in plaque development that can culminate in coronary heart disease. A cross-sectional study of 192 African Americans in the Baltimore area found a higher prevalence of coronary calcification (measured by computed tomography) in HIV-positive cocaine users (37.6% of 85) than in HIV-negative people who used cocaine (29.8% of 47), HIV-positive people who did not use cocaine (28.6% of 28), or HIV-negative people who did not use cocaine (18.8% of 32).89 These people averaged about 38 years in age and had no symptoms of cardiovascular disease.
Statistical analysis to reckon the impact of HIV and cocaine use (alone and together) on coronary calcification factored in age, body mass index, LDL cholesterol, triglycerides, mean corpuscular volume, and systolic blood pressure. Compared with total calcification volume in HIV-negative nonusers, the total value was higher in HIV-negative cocaine users (regression estimate [RE] 2.59), higher still in HIV-positive nonusers (RE 2.92), and highest in HIV-positive cocaine users (RE 3.49). Compared with HIV-negative nonusers, the other three groups all had a significantly higher number of lesions and a significantly higher total calcium score.
Later work by this Johns Hopkins team involved 165 HIV-positive African Americans from 25 to 54 years old and recruited from August 2003 through June 2007.90 Median age stood at 44 years, and 36% were women; nobody had cardiovascular symptoms. Computed tomography detected significant (50% or greater) coronary artery stenosis in 24 people (15%). Among people who used cocaine at least 15 years and took antiretrovirals for at least 6 months, that rate reached 42%. Regression analysis determined that using cocaine at least 15 years hoisted the odds of significant stenosis almost 8 times (aOR 7.75, 95% CI 2.26 to 31.2), while cART for at least 6 months more than quadrupled the odds (aOR 4.35, 95% CI 1.30 to 16.4).
Duration of stavudine or Combivir drove the association between longer cART and significant stenosis. At least 6 months of stavudine boosted the odds 18 times, while at least 6 months of Combivir raised the odds almost 6 times. Antiretrovirals not associated with significant stenosis in this 6-month duration analysis were zidovudine or lamivudine alone (that is, not as part of Combivir), didanosine, efavirenz, nevirapine, nelfinavir, indinavir, lopinavir, and atazanavir. Odds for abacavir could not be calculated because not many people had used it. Long-term cocaine use, these researchers concluded, "imposes an alarming risk of coronary artery disease."90
Kidneys, the fist-sized filters facing each other across the spine, do much more than remove waste (and drugs) from the body: they also balance bodily fluids, release hormones that regulate blood pressure, produce vitamin D, and help make red blood cells.91 Every day 200 quarts of fluid sluice through the kidneys, percolating through a million nephrons (Figure 12). The kidneys extract 2 quarts of fluid daily and return 198. The National Kidney Foundation estimates that 26 million adults in the United States have chronic kidney disease, which poses a high threat of heart disease.92
Analysis of 17,264 veterans with HIV -- more than 95% of them men -- found that two indicators of poor kidney function independently raised the risk of a new atherosclerotic cardiovascular event or heart failure.94 When a person had both indicators of poor kidney function, risks for these two outcomes were considerably higher. Kidney nephron shown at upper left, from Servier Medical Art. (See text for variables in model.)
In a nationally representative sample, the CDC recently estimated that 7.6% of HIV-positive adults in care in the United States have stage 3 or worse chronic kidney disease, defined as estimated glomerular filtration rate (eGFR) below 60 mL/min/1.73m2.93 Among 20- to 39-year-olds, HIV-positive people had more than a 4 times higher rate of chronic kidney disease than the general population (prevalence ratio [PR] 4.6); among 40- to 59-year-olds, people with HIV had an 80% higher rate (PR 1.8). Among people 60 and older, the general population had a higher rate, probably partly because more HIV-negative than HIV-positive people with chronic kidney disease survive past 60. In HIV-positive adults in care, the CDC identified five factors associated with chronic kidney disease -- older age, female sex (adjusted PR 1.4), HIV duration longer than 10 years (adjusted PR 1.4), an AIDS diagnosis, and a CD4 count under 350 cells/mm3 (adjusted PR 1.6).
Recent cohort studies link poor kidney function -- measured as eGFR, albuminuria, or proteinuria -- with cardiovascular disease in people with HIV. Persistent albuminuria indicates that a damaged kidney is spilling albumin into urine. Researchers at the San Francisco Veterans Affairs Medical Center parsed records of 17,264 HIV-positive people in the Veterans Health Administration to catalog newly diagnosed cardiovascular disease (defined as coronary, cerebrovascular, or peripheral arterial disease) and new cases of heart failure.94
In this national sample the 1194 cohort members with eGFR below 60 mL/min averaged 52 years in age, compared with 46 in the 16,070 members with an eGFR at or above 60 mL/min. About 45% of study participants were black, about 35% white, and only 3% women. Through a median 7 years of follow-up, the investigators counted 370 heart failures and 833 atherosclerotic cardiovascular events.
People with an eGFR below 30 mL/min and albuminuria at or above 100 mg/dL had about a 6-fold higher rate of new cardiovascular disease than people with an eGFR at or above 60 mL/min and no albuminuria. Incidence of atherosclerotic cardiovascular events and albuminuria grew progressively as albumin levels rose (from 0 to 30 to 100 or more mg/dL) and as eGFR waned (from 60 or higher to 30 to 59 to under 30).
A full multivariate model adjusted for age, sex, race, and time-updated hypertension, diabetes mellitus, chronic obstructive lung disease, dyslipidemia, smoking, CD4 count, viral load, and antiretroviral therapy. In this analysis albuminuria above 30 mg/dL and eGFR below 60 mL/min each independently raised the risk of an atherosclerotic cardiovascular event and the risk of heart failure (Figure 12). When a person had both albuminuria and a sub-60 eGFR, the risk of an atherosclerotic cardiovascular event and the risk of heart failure were even higher.
These investigators believe their results "are clinically relevant because they may help providers to identify HIV-infected persons at high risk for CVD events."94 The National Kidney Foundation recommends screening for albuminuria in people with chronic kidney disease risk factors, including diabetes, hypertension, systemic illnesses, age over 60, and family history of chronic kidney disease.95 The Foundation advises confirming a positive test with a second urine test.
A case-control study at the Johns Hopkins HIV Clinic confirmed the graded impact of worsening kidney function on cardiovascular risk.96 This study involved 315 HIV-positive adults, 63 who had a myocardial infarction or a cerebrovascular accident and 252 who did not. The 252 randomly selected control patients had no history of heart disease and matched the 63 case patients by age, race, and sex. Age averaged 49.5 in both groups, 63.5% were men, and 84% were black.
Multivariate analysis (adjusted for diabetes, hypertension, previous cardiac events, dyslipidemia, viral load, and CD4 count) linked every 10 mL/min lower eGFR to 20% higher odds of a cardiovascular event (aOR 1.2, 95% CI 1.1 to 1.4, P = 0.009). In the same analysis proteinuria, defined as a urine dipstick reading at least above 1+, nearly tripled the odds of a cardiovascular event, though that association stopped short of statistical significance (aOR 2.9, 95% CI 0.9 to 9.0, P = 0.070). In a separate analysis, proteinuria compounded the impact of low eGFR on cardiovascular risk.
The Johns Hopkins investigators noted that their findings reflect results in the general population but assume greater importance in people with HIV, who have a 3- to 5-fold higher kidney disease prevalence than people without HIV.93,97,98 They proposed that their findings "suggest the potential value of early screening and treatment of chronic kidney disease in HIV-1-infected patients, particularly those with other cardiovascular risk factors."96
In its report on chronic kidney disease prevalence with HIV, the CDC recommends (1) routine screening for chronic kidney disease, (2) aggressive management of related conditions including diabetes, hypertension, and obesity, and -- among people who do have chronic kidney disease -- (3) avoidance of nephrotoxic drugs and referral to a nephrologist.93
HIV infection is an inflammatory disease marked by ongoing immune activation. Even when cART corrals HIV replication, low-level inflammation and immune activation may persist and tweak up the risk of cardiovascular disease. How? HIV heart guru James Stein explains that relentless inflammation, immune activation, and viremia hamper a blood vessel's ability to dilate and generate an anticoagulant surface.31 And clumpy cells are a big enemy of cardiovascular health.
The literature on inflammation, immune activation, and heart health in people with HIV has ballooned to Brobdingnagian proportions. Searching for cardiovascular disease + HIV + inflammation on pubmed returned 371 articles in April 2013. The same search on Google Scholar gives you "about 52,000" returns. Most of these studies point in the same direction: inflammation is bad for your heart. But James Stein cautions that pinning down the precise inflammation-linked risk, and figuring which markers predict best, "likely will require several thousand subjects, more than a decade of follow-up, reliable biomarker/imaging tests, and strict endpoint adjudication."31
One place to start through the thicket of research on inflammation and HIV-related heart disease is with HCV infection, an overtly inflammatory illness that often coexists with HIV infection. Meta-analysis of 12 studies linked HCV infection to a higher risk of coronary artery disease in the general population.99 Of the 6 best studies analyzed, three found a significant association between HCV and coronary artery disease, two found a nonsignificant association, and one figured HCV protects against coronary artery disease.
Six recent studies on HCV, HIV, and cardiovascular risk yielded divergent results reflecting the different study populations, methods, and endpoints.
Women's Interagency HIV Study (WIHS) investigators measured cIMT and carotid plaque in 1865 HIV-positive women in 2004 and 2005.100 Median cIMT was similar in HCV-infected and HCV/HIV-coinfected women, and higher than in HIV-monoinfected women. But after statistical adjustment for other cardiovascular risk factors, HCV infection was not associated with cIMT or with carotid plaque.
A DAD Study analysis involved 33,347 HIV-positive men and women in Europe, the US, and Australia who had 517 myocardial infarctions during followup for an incidence of 3.3 per 1000 person-years.101 Incidence was marginally lower in HCV-seropositive people than HCV-negative people (2.7 versus 3.3 per 1000 person-years). After statistical adjustment for relevant variables, HCV seropositivity was not associated with incident myocardial infarction (rate ratio 0.86, 95% CI 0.61 to 1.19). There were 295 strokes during the study period (1.47 per 1000 person-years with HCV and 1.91 without HCV), and HCV positivity did not affect stroke risk after statistical adjustment. Active HBV infection did not affect rates of MI or stroke.
But a large Veterans Affairs study did link HCV/HIV infection to a higher risk of cerebrovascular disease (stroke and transient ischemic attack) and to a trend toward a higher MI rate.102 This analysis involved 19,424 HIV-positive veterans, 32% of them coinfected with HCV and HIV and 97% of them men. The investigators identified HCV infection by diagnostic codes and HCV-antibody positivity. There was no HCV/HIV-negative control group. After statistical adjustment for potentially confounding factors, HCV/HIV coinfection was linked to a 20% higher risk of cerebrovascular disease compared with HIV alone (adjusted hazard ratio 1.20, 95% CI 1.04 to 1.38, P = 0.013), while coinfection was nonsignificantly associated with a 25% higher risk of acute MI (adjusted hazard ratio 1.25, 95% CI 0.98 to 1.61, P= 0.072).
Another US veterans study did find an independent (though small) association between HCV infection and acute myocardial infarction in a comparison of HIV-positive and negative veterans.18 This Veterans Aging Cohort Study involved 27,350 HIV-positive and 55,109 age-, race-, and site-matched HIV-negative veterans, 97% of them men. No one had a history of cardiovascular disease. The researchers defined HCV infection by ICD-9 code or positive HCV antibody. During 5.9 years of follow-up, 871 veterans had an acute MI. Statistical analysis adjusted for multiple risk factors associated HCV infection with about a 20% higher MI risk when compared with HCV-negative vets (adjusted hazard ratio 1.19, 95% CI 1.01 to 1.40).
A medical record review at the University of Rochester compared 239 people with HIV, 167 with HCV, and 182 with both HIV and HCV with gender-, race-, and age-matched uninfected people in the NHANES database.103 After statistical adjustment for confounders, HCV/HIV-coinfected people had a 2% higher Framingham Risk Score than the general population (P = 0.03) and a 4.1-year older vascular age (P = 0.01). People infected with HCV but not HIV had a 2.4% higher Framingham Risk Score than the general population (P < 0.001) and a 4.4-year older vascular age (P < 0.001). But HIV infection alone did not confer a higher Framingham score or an older vascular age.
Comparing 18 HCV/HIV-coinfected people with 22 HIV-monoinfected people, French investigators recorded a significantly higher prevalence of subclinical carotid plaque in the coinfected group, even though LDL cholesterol and blood pressure were lower in coinfected people.104 Chronic HCV infection was associated with 10-fold higher odds of plaque (OR 10, 95% CI 1.5 to 72, P = 0.02).
The French team suggested that "HCV infection might be considered as not only a liver infection but also as a metabolic disease in HIV patients, justifying regular cardiovascular surveillance."104 The VA team observed, though, that HCV may inflate rates of some cardiogenic conditions, including metabolic syndrome and diabetes, while it appears to lower levels of total cholesterol, LDL cholesterol, and triglycerides.102 Of the three studies that assessed HCV impact on clinical endpoints, the two predominantly male VA studies uncovered evidence implicating HCV in MI and stroke,18,102 but the DAD study (with 74% of participants men) saw no HCV tie to MIs (Table 5).101
|Table 5. Cardiovascular Disease Risk With HCV in Clinical Endpoint Studies|
|Study||n (% Men)||Comparison Groups||Study Years (Follow-Up)||Endpoints||HCV Association|
|VACS18||27,350 (97%)||55,109 matched HIV-negative veterans||2003-2009 (median 5.9 y)||Acute MI||19% higher MI risk|
|Veterans102||19,424 (97%)||6136 HIV/HCV+ vs 13,288 HIV+ only||1996-2004 (mean 3.9 y)||Acute MI, cerebrovascular disease (CVD)||25% higher MI risk;* 20% higher CVD risk|
|DAD101||33,347 (74%)||5084 HCV/HIV+ vs 16,731 HIV only†||1999-2007||MI, stroke||No association|
VACS, Veterans Aging Cohort Study.
* Not significant.
† 11,532 had unknown HCV status.
The DAD team tabulated results of 18 previous general-population studies assessing cardiovascular disease risk with HCV: eight found an association and 10 did not. Nailing down whether HCV substantially inflates an already high cardiovascular risk in people with HIV requires further study.
Two recent studies saw links between markers of inflammation and all-cause mortality in people with HIV, and one of them extended that association to atherosclerotic cardiovascular disease and heart failure. Albumin levels fall in the face of inflammation, providing an inverse marker of the inflammatory process.105 To gauge the impact of serum albumin on mortality and heart disease in people with HIV, University of California San Francisco researchers turned to a national veterans database, the HIV Clinical Case Registry.106 This analysis included 25,522 HIV-positive veterans enrolled between 1986 and 2007 who had serum albumin, serum creatinine, and urine dipstick measures between 1986 and 2007. The three primary outcomes were time to death, atherosclerotic cardiovascular disease, and hospital admission for heart failure. The investigators broke serum albumin levels into five brackets: at or above 4.0, 3.5 to 3.9, 3.0 to 3.4, 2.5 to 2.9, and under 2.5 g/dL.
Compared with the highest albumin bracket, each lower bracket independently boosted chances of mortality -- both with baseline albumin and (even more so) with time-updated albumin. Lower baseline albumin did not affect chances of atherosclerotic cardiovascular disease, but lower updated albumin did. Both lower baseline and lower updated albumin hoisted hazard ratios for heart failure:
Hazard ratios (and 95% CIs) for mortality:
Hazard ratios (and 95% CIs) for atherosclerotic cardiovascular events:
Hazard ratios (and 95% CIs) for heart failure:
The associations between low albumin and mortality were strongest in the first year of follow-up, though still usually significant after 2 or 3 years.
The investigators106 acknowledged the difficulty in explaining why low serum albumin (versus high urine albumin in the just-described study94) predicts dire clinical outcomes. Serum albumin levels can fall because of poor nutrition, liver disease, kidney disease, and chronic inflammation. But a sensitivity analysis that excluded people with liver and kidney dysfunction found nearly identical associations between low serum albumin and the three endpoints. That result suggested to the investigators "that a more transient process such as inflammation is responsible for the lower levels of albumin."106 They proposed that "serum albumin captures a dynamic process of inflammation in HIV infection that has clinical importance in the short-term."
A prospective study of 327 HIV-positive people at Boston's Tufts University found significantly higher risks of all-cause mortality with higher high-sensitivity C-reactive protein (hsCRP), a classic marker of inflammation, and with cIMT, a verified signal of subclinical atherosclerosis.107 The study involved 242 men and 85 women, 52% of them white, with an average age of 44 years. None of them had overt cardiovascular disease. Through a median followup of 3.1 years, 38 people (12%) died. Five of these deaths (13%) had primary or secondary cardiovascular causes, and one was sudden and unexplained.
Statistical analysis adjusted for age, gender, race, body mass index, cigarette smoking, CD4 count, viral load, LDL cholesterol, HDL cholesterol, and hsCRP determined that cIMT above versus below 0.655 mm almost tripled the risk of death (adjusted hazard ratio 2.74, 95% CI 1.26 to 5.97, P = 0.01). In a similar analysis adjusted for cIMT, hsCRP at or above versus below 3 mg/L more than doubled the risk of death (adjusted hazard ratio 2.38, 95% CI 1.15 to 4.90, P = 0.02). hsCRP was almost 3 times higher in people who died than in those who did not (3.2 versus 1.3 mg/L, P < 0.001), and a significantly higher proportion of people who died had hsCRP above 3 mg/L (51% versus 25%, P < 0.001). An earlier study of 209 HIV-positive US women identified baseline CRP as an independent predictor of mortality through a median 45 months of follow-up.108
In Boston's Partners HealthCare System, elevated CRP and HIV each independently doubled chances of acute myocardial infarction.109 This analysis involved 487 HIV-positive patients and 69,870 HIV-negative people in care between January 1997 and December 2006. Everyone had CRP measured in the past 3 years and more than 1 week before an acute MI. A statistical model adjusted for age, sex, race, hypertension, diabetes, dyslipidemia, elevated CRP, and HIV status determined that high CRP and HIV each independently doubled the odds of acute MI (Figure 13). People with HIV and elevated CRP had quadrupled odds of acute MI compared with HIV-negative people with normal CRP levels. At the end of the third review article in this issue of RITA!, this study's principal investigator, Steven Grinspoon, offers his insights on when measuring CRP in people with HIV may pay off in practice.
In 487 HIV-positive and 68,870 HIV-negative people in care in Boston, elevated CRP and HIV infection each independently doubled the odds of acute MI.109 People with HIV and elevated CRP had quadrupled odds of acute MI compared with HIV-negative people with normal CRP.
Research links low vitamin D to cardiovascular disease in the general population, but so far studies of vitamin D and cardiovascular risk in people with HIV are small, rely on surrogate markers, and yield mixed results. All of these studies weigh vitamin D in relation to cIMT, a marker of atherosclerosis, and some examine other cardiovascular proxies.
The largest heart-related vitamin D study involved 139 HIV-positive adults in a San Francisco group, all of whom had vitamin D measured as 25-hydroxyvitamin D [25(OH)D], the standard way to measure this vitamin.110 The study group averaged 45 years in age, 84% were men, 54% white, and 32% black. Half of these people (52%) had vitamin D insufficiency, defined as a level at or below 30 ng/mL. A statistical model adjusted for classic heart risk factors and HIV variables determined that average cIMT increased (worsened) significantly from about 0.8 mm in people with normal vitamin D (above 30 ng/mL) to about 1.0 mm in those with deficient vitamin D (15 to 30 ng/mL) and to about 1.1 mm in those with vitamin D below 15 ng/mL (P = 0.021). cIMT was an average 0.13 mm greater in people with 25(OH)D below 30 ng/mL than in people with normal 25(OH)D. The authors observe that research in the general population links every 0.10-mm greater cIMT to a 15% higher MI risk and an 18% higher stroke risk. The study is limited by its cross-sectional nature and the inability to account for the possible impact of individual antiretrovirals. Other work, for example, links efavirenz to low vitamin D.
In three smaller cross-sectional analyses of vitamin D and cIMT, one study tied lower 25(OH)D to greater cIMT but two studies did not. The study that found a link involved 56 adults with HIV.111 Median age stood at 49, and most participants were men (85%) and white (52%). Although this analysis did not tie 25(OH)D to inflammatory or endothelial markers, lower 25(OH)D conferred a 10 times higher risk of common carotid IMT above the median for the study group (P < 0.01). The association was not significant for internal carotid IMT.
An analysis in the Hawaii Aging With HIV-Cardiovascular Cohort Study involved 100 people with a median age of 52, most of them male (86%) and white (60%).112 Analysis of 50 cIMT measurements found a significant correlation between 25(OH)D and brachial artery flow-mediated dilation but not cIMT (r = -0.05, P = 0.76). A third cross-sectional study involved 30 HIV-positive children and young adults with a median age of 11, three quarters of them black and 37% male.113 These researchers found no significant correlation between 25(OH)D and cIMT, inflammatory markers, or lipids. But 25(OH)D correlated inversely with insulin resistance -- the lower the 25(OH)D, the greater the insulin resistance.
A recent review of randomized trials, meta-analyses, and other evidence in the general population concluded that adequate vitamin D may protect against cardiovascular disease -- as well as musculoskeletal maladies, infectious diseases, autoimmune diseases, type 1 and type 2 diabetes, several cancers, neurocognitive dysfunction, and mental illness.114 This mega-analysis also tied low vitamin D to all-cause mortality. With such a catalog of benefits-in-waiting, checking HIV-positive people for vitamin D and supplementing those deficient may seem a sensible hedge. But randomized trials in the general population -- and in people with HIV -- show that swallowing high doses of vitamin D3 does not ipso facto translate into sounder health. A double-blind, placebo-controlled trial of 45 HIV-positive adults who took 4000 IU of vitamin D daily or placebo for 12 weeks found that supplementation modestly improved vitamin D status and non-HDL cholesterol but did not change endothelial function and worsened insulin resistance.115
A Women's Interagency HIV Study of 885 HIV-positive and 408 HIV-negative women linked progestin-only hormonal contraceptives to lower HDL cholesterol (-3 mg/dL, 95% CI -5 to -1 with HIV, -6 mg/dL, -9 to -3 without HIV) and greater insulin resistance (HOMA-IR +0.86, 95% CI 0.51 to 1.22 with HIV, and +0.56, 95% CI 0.12 to 1.01 without HIV).116 Estrogen/progestin hormonal contraceptives were associated with higher HDL. The WIHS investigators suggested that combined hormonal contraceptives may be preferred for women with HIV, but they cautioned clinicians to check for interactions with antiretrovirals.