|Home | About | Journals | Submit | Contact Us | Français|
An estimated 34 million men, women, and children are infected with human immunodeficiency virus type 1 (HIV-1), the virus that causes acquired immunodeficiency syndrome (AIDS). Current technology cannot eradicate HIV-1, and most patients with HIV-1-infection (HIV+) will require lifelong treatment with combined antiretroviral therapy (cART). Stroke was recognized as a complication of HIV-1 infection since the early days of the epidemic. Potential causes of stroke in HIV-1 include opportunistic infections, tumors, atherosclerosis, diabetes, hypertension, autoimmunity, coagulopathies, cardiovascular disease, and direct HIV-1 infection of the arterial wall. Ischemic stroke has emerged as a particularly significant neurological complication of HIV-1 and its treatment due to the aging of the HIV+ population, chronic HIV-1 infection, inflammation, and prolonged exposure to cART. New prevention and treatment strategies tailored to the needs of the HIV+ population are needed to address this issue.
Stroke has received increased attention with the maturation of the human immunodeficiency virus type 1 (HIV-1) epidemic. This topic is difficult to study, because HIV-1 infected (HIV+) populations differ in respect to risk factors such as age, diabetes, dyslipidemia, and access to combined antiretroviral therapy (cART). cART is so important that many authors distinguish between groups according to the availability of treatment. We define the cART era as beginning in 1996 with the availability of protease inhibitor (PI) drugs. However, many countries have no treatment access and are still considered as ‘pre-cART’.
Stroke was first reported in HIV+ patients in the 1980s [Anders et al. 1986a]. An unexpectedly high incidence of stroke was noted in HIV+ children and young adults without classic risk factors [Park et al. 1990; Qureshi et al. 1997]. Autopsy series found stroke in 4–34% of patients with acquired immunodeficiency syndrome (AIDS) [Pinto, 1996]. These strokes were primarily ischemic in nature and often went undiagnosed during life.
In the pre-cART era, stroke typically occurred in the setting of advanced AIDS [Anders et al. 1986b; McGuire and So, 1994; Mizusawa et al. 1988; Park et al. 1990; Pinto, 1996; Price et al. 1992], such as fungal meningitis, toxoplasmosis encephalitis, tuberculous meningitis, neurosyphilis, lymphoma, coagulopathies, vasculitis, and cardiac disorders [Berger et al. 1990; Kieburtz et al. 1993; Mizusawa et al. 1988; Perriens et al. 1992; Pinto, 1996].
Stroke was also reported in HIV+ patients without secondary complications. This coincided with the identification of new entities such as HIV-associated vasculopathy, a unique and poorly understood condition characterized by occlusive disease of medium and large sized arteries [Bhagavati and Choi, 2008; Tipping et al. 2007], as well as HIV-associated aneurysmal arteriopathy, a condition that is predominantly observed in HIV+ children and young adults and is characterized by fusiform or saccular dilatation of the cerebral arteries [Patsalides et al. 2002; Tipping et al. 2006].
The direct effect of HIV-1 on stroke pathogenesis has been difficult to establish. An early case control study of autopsy-diagnosed stroke compared patients with AIDS aged 20–55 years with age- and gender-matched HIV-seronegative (HIV–) controls who died of other diseases [Berger et al. 1990]. Eight percent of the AIDS group had stroke, which was not statistically different than the controls. An African retrospective case control study compared HIV+ patients with age-matched HIV– controls and found no increased incidence of stroke in the HIV+ group [Hoffmann et al. 2000]. Pinto reviewed 6 pre-cART clinical and 11 autopsy series of HIV+ patients (total n = 1885) and found that only 1.3% had stroke [Pinto, 1996]. Pinto concluded that the evidence for an association of HIV-1 and stroke was lacking.
In contrast, other studies reported a positive association between HIV-1 and stroke, particularly in young adults without conventional cardiovascular risk factors. Engstrom and colleagues reported an annual risk of stroke of 0.75% in patients with AIDS versus 0.025% for a reference population aged 35–45 years [Engstrom et al. 1989]. A cohort of HIV+ children was found to have an annual incidence of stroke of 1.3%; in the subset that came to autopsy, 24% had a stroke [Park et al. 1990]. Qureshi and colleagues performed a retrospective, case control study of stroke and compared 113 HIV+ patients aged 19–44 years with matched HIV– patients from the same area who were hospitalized for asthma [Qureshi et al. 1997]. The authors chose this control group because of the lack of relationship between asthma and stroke. HIV-1 was associated with all strokes with an odds ratio of 2.3, and with ischemic stroke with an odds ratio of 3.2. In a radiological study of 426 pediatric HIV+ patients, 2.6% had cerebrovascular lesions [Patsalides et al. 2002]. Cole and colleagues conducted a pre-cART population-based study of adults aged 15–44 and compared patients with AIDS with an age-matched group composed of both HIV+ patients without AIDS and HIV– controls [Cole et al. 2004]. AIDS conferred 9 times the risk for ischemic stroke and 12 times the risk for hemorrhagic stroke.
In summary, these studies indicate that HIV+ patients have a higher rate of stroke, particularly ischemic stroke, than comparable HIV– controls. This effect is most striking in younger people.
The introduction of cART in 1996 resulted in a decline in neurologic infections and tumors [Sacktor et al. 2001] and increased lifespan for HIV+ patients. However, within 2 years, premature atherosclerosis was reported in cART-exposed HIV+ patients [Henry et al. 1998] and data collected from longitudinal studies of cART-treated patients reported an increase in myocardial infarction and ischemic stroke.
The Data collection on Adverse events of anti-HIV Drugs (DAD) Study found a significantly elevated risk of myocardial infarction and stroke, beyond what was anticipated for the increasing age of their cohort [d’Arminio et al. 2004]. Subsai and colleagues studied neurological disease in an HIV+ Thai population and discovered that, post-cART, the incidence of brain opportunistic infections dropped, and the rate of ischemic and hemorrhagic stroke increased [Subsai et al. 2006]. Longo-Mbenza and colleagues studied stroke in HIV+ Africans with access to cART [Longo-Mbenza et al. 2011]. Stroke was the first manifestation of HIV-1 in 12 of 116 participants, and was associated with age under 45 years, autoimmune disease, and the metabolic syndrome. Many of these African patients had risk factors for stroke similar to those seen in industrialized countries, such as smoking, insulin resistance, cardiac disease, and obesity. The investigators suggested that such factors be considered in future international HIV-1 stroke studies. Ovbiagele and Nath reviewed data from the United States Nationwide Inpatient Sample, and found that while stroke hospitalizations declined from 1997 to 2006 by 7% in the general population, there was a concurrent increase of 60% in stroke hospitalizations among HIV+ patients [Ovbiagele and Nath, 2011]. Chow and colleagues analyzed data from a US healthcare system and found that the incidence of ischemic stroke was 5.27 per 1000 person years in HIV+ versus 3.75 in HIV– patients [Chow et al. 2012]. HIV-1 was an independent predictor of stroke after controlling for demographics and traditional stroke risk factors. The Swiss HIV Cohort Study [Hasse et al. 2011] reported that the multivariable hazard ratio for stroke (17.7; confidence interval 7.06–44.5) was elevated in participants aged 65 years and older, in contrast to previous studies that found an association between stroke and HIV+ status in younger groups.
HIV-1 has been linked to the induction of anticardiolipin antibodies (ACAs), antiphospholipid antibodies, and antiprothrombin antibodies [Saif and Greenberg, 2001]. Studies in the HIV– population suggest that ACAs are a risk factor for thrombosis and stroke, and act by impairing the activity of the polypeptide tissue factor pathway inhibitor [Forastiero et al. 2003]. However, ACAs are commonly seen in HIV+ patients without stroke, therefore this mechanism requires further exploration [Galrao et al. 2007].
Protein S is a vitamin K-dependent anticoagulant. Decreased levels or impaired function of protein S leads to a propensity for venous thrombosis. In an HIV+ African cohort, protein S deficiency was the most common coagulopathy associated with stroke [Mochan et al. 2003]. Subsequent studies indicated that protein S deficiency is related to HIV-1 serostatus and is not causally related to stroke [Mochan et al. 2005]. Protein C deficiency is associated with an increased incidence of venous thromboembolism whereas little association with arterial thrombosis has been found in any group [Morris et al. 2010].
Infectious vasculitis is a known cause of stroke. Based primarily on immunohistochemical studies, Nagel has argued that there is insufficient evidence that HIV-1 causes cerebral vasculitis or infects the arterial wall [Nagel et al. 2010; Tipping et al. 2006; Bissel and Wiley, 2004]. He opined that most reported cases of HIV-1 vasculopathy did not adequately exclude VZV, a virus known to infect cerebral arteries. In contrast, Eugenin and colleagues, using fluorescence confocal microscopy, reported that HIV-1 could directly infect human arterial smooth muscle [Eugenin et al. 2008]. HIV-1 sequences have also been detected by polymerase chain reaction in an intracerebral blood vessel of one pediatric case [Dubrovsky et al. 1998].
Most cases of aneurysmal dilation of the cerebral arteries occur in HIV+ children and young adults [Park et al. 1990]. Affected patients develop fusiform or saccular aneurysms, often involving the large blood vessels of the Circle of Willis, and can present with ischemic or hemorrhagic stroke. Pathology is notable for medial fibrosis, damage or loss of the muscularis and internal elastic lamina, and intimal hyperplasia [Dubrovsky et al. 1998] with little or no inflammation [Goldstein et al. 2010]. Later studies indicate that HIV-1 is associated with thinning of the arterial medial layer in autopsied subjects without stroke [Gutierrez et al. 2012]. The authors suggested that this medial thinning might be a precursor to aneurysmal arteriopathy [Gutierrez et al. 2012]. Radiologically documented cases of HIV-1 arteriopathy have responded to cART, implicating HIV-1 as a pathogen [Bhagavati and Choi, 2008; Cutfield et al. 2009].
The ability of HIV-1 to infect vascular cells remains controversial. However, it is agreed that HIV-1 infects other elements in the cerebrovascular milieu, such as perivascular monocytes, macrophages, and astrocytes [Bell, 1998]. Infected cells can release neurotoxic viral proteins such as gp120, Tat, and Nef [Duffy et al. 2009; Kanmogne et al. 2007; Maniar et al. 2012; Toborek et al. 2005] that damage vascular endothelium and potentially increase the probability of stroke [Annunziata, 2003; Strazza et al. 2011]. In vitro, these viral proteins produce changes in vessel morphology, endothelial apoptosis, loss of collagen, loss of membrane glycoproteins, and disruption of the blood–brain barrier [Fischer-Smith and Rappaport, 2005; Huang et al. 2001; Park et al. 2001a]. HIV-1 and its viral proteins can also upregulate the expression of chemoattractants, adhesion molecules and proinflammatory cytokines that damage the blood–brain barrier and increase leukocyte migration into the brain [Fischer-Smith and Rappaport, 2005; Mu et al. 2007; Park et al. 2001b]. Similarly, both infected and uninfected but activated cells can release pathogenic factors. By altering the vasoactive properties of blood vessels and the production of endothelial vasodilators, these proinflammatory cytokines can predispose vessels to vasospasm, thrombosis, and atherosclerosis [Vila and Salaices, 2005].
In the absence of cART, HIV-1 is associated with the development of abnormalities such as increased carotid intimal thickness [Lorenz et al. 2008], carotid arterial wall stiffness [Seaberg et al. 2010], and abnormalities in vascular compliance and distensibility [Oliviero et al. 2009], findings which are associated with atherosclerotic disease. Suppressive cART improves markers of immune activation, inflammation and coagulation, but elevated levels persist despite effective treatment [Neuhaus et al. 2010]. HIV-1 may drive atherogenesis through activating endothelial and immune cells, increasing the numbers of circulating atherogenic immune cells, and altering lipid levels and function [Zanni and Grinspoon, 2012]. The mechanisms by which HIV-1 predisposes the vascular system to atherosclerosis, myocardial infarction, and chronic inflammation [Hunt, 2012; Maniar et al. 2012; Zanni and Grinspoon, 2012] have been reviewed and are beyond the scope of this article. Other factors which may contribute to stroke by increasing inflammation and atherosclerosis include coinfection with cytomegalovirus or hepatitis C virus [Maniar et al. 2012].
cART may increase the risk for stroke and heart disease [d’Arminio et al. 2004; Worm et al. 2010; Rasmussen et al. 2011]. It is likely that the risk of stroke is also influenced by the stage of disease at which cART is initiated [Subsai et al. 2006], treatment duration [d’Arminio et al. 2004; Worm et al. 2012] and other factors. Cardiovascular disease in HIV+ has been associated with a low nadir (lowest ever) CD4+ count [Ho et al. 2012], but no studies have examined the relationship of nadir CD4+ to stroke. Hypothetically, a successful cART regimen that lowers viral load and increases CD4+ should reduce the likelihood of stroke associated with opportunistic infections. Conversely, a cART regimen that significantly increases lifespan may increase the risk for stroke because of prolonged exposure to HIV-1, cART-related, and age-related risk factors. Bozzette and colleagues reported that cART reduced overall mortality from all causes, including cardiovascular and cerebrovascular diseases [Bozzette et al. 2003]. However, his cohort was treated for a mean duration of less than 2 years [Bozzette et al. 2003]. Corral and colleagues reported that a shorter duration of cART was associated with more vascular complications [Corral et al. 2009]. In contrast, the DAD: study reported that prolonged use (6 years) of cART was associated with more cardiovascular and cerebrovascular disease than would be expected by aging alone [d’Arminio et al. 2004].
Elevated cholesterol and triglycerides are common in patients receiving cART and are associated primarily with the use of PIs [Calza et al. 2004a]. The PI atazanavir (Reyataz, Bristol-Meyers Squibb, Princeton, New Jersey USA) is associated with less dyslipidemia [Murphy et al. 2010], whereas boosted tipranavir (Aptivus, Boehringer Ingleheim Pharmaceuticals, Inc., Ridgefield CT, USA)/ritonavir (Norvir, Abbott Laboratories, North Chicago, IL, USA) and lopinavir/ritonavir (Kaletra, Abbott Laboratories, North Chicago, IL, USA) are associated with more dyslipidemia. Nucleoside reverse transcriptase inhibitors (NRTIs), particularly stavudine (Zerit, Bristol-Meyers Squibb Company, Bristol, NJ, USA), have also been associated with dyslipidemia [Malvestutto and Aberg, 2011]. Use of the NRTI abacavir (Ziagen, GlaxoSmithKline for ViiV Healthcare, Research Triangle Park, NC, USA) was reported to be associated with myocardial infarction [Sabin et al. 2008]; however, later studies found no association with myocardial infarction or stroke [Bedimo et al. 2011; Cruciani et al. 2011]. Non-nucleoside reverse transcriptase inhibitors (NNRTIs) such as nevirapine (Viramune, Boehringer Ingleheim Pharmaceuticals Inc., Ridgefield, CT, USA), etravirine (Intelence, Janssen Therapeutics,Titusville, NJ, USA), and efavirenz (Sustiva, Bristol-Meyers Squibb, Princeton, NJ, USA) have less negative effects on lipids [Malvestutto and Aberg, 2011]. Fusion inhibitors such as enfuvirtide (Fuzeon, Genentech, South San Francisco, CA, USA), CCR5 receptor inhibitors such as maraviroc (Selzentry, ViiV Healthcare, Research Triangle Park, NC, USA) and integrase inhibitors such as raltegrivir (Isentress, Merck & Co., Inc., Whitehouse Station, NJ, USA) have neutral effects on lipids [Malvestutto and Aberg, 2011].
Diabetes can predispose people to stroke via accelerated atherogenesis and thrombosis, impairment of cerebral blood flow and impaired cerebral autoregulation. HIV-1 itself is not associated with an increased risk of diabetes. However, patients treated with cART regimens, especially those that contain the NRTI stavudine and some PIs, have an increased prevalence of type II diabetes and glucose intolerance [Calza et al. 2004b; Paik and Kotler, 2011]. NRTIs may cause diabetes by their toxic effect on muscle mitochondria, which can lead to insulin resistance, whereas PI drugs may inhibit glucose uptake by cells [Paik and Kotler, 2011]. The fusion inhibitors, integrase inhibitors, and entry inhibitors have not been reported to significantly affect glucose metabolism. Diabetes is also associated with other adverse effects of cART, known as lipodystrophy and lipoatrophy. Body fat is redistributed away from the face and extremities and accumulates centripetally in the trunk and viscera, leading to an increased waist-to-hip ratio. Visceral fat accumulation is a risk factor for myocardial infarction, and may be related to stroke.
Hypertension is a risk factor for cerebrovascular disease. HIV-1 has inflammatory effects on the vascular endothelium which could result in hypertension [Sen et al. 2012]. One study found that hypertension was associated with increased duration of HIV-1 disease [Manner et al. 2012]. However, it is unknown whether HIV-1 causes hypertension in the absence of other conditions such as increased body mass index, which is a common consequence of PI use [Bergersen et al. 2003; Chow et al. 2003; Crane et al. 2006].
The treatment of stroke in HIV+ patients has not garnered much attention. However, stroke may be more common than many better known complications of AIDS. An HIV-centric approach to stroke may be appropriate for several reasons. Many HIV+ patients remain inadequately treated and are at risk for stroke due to brain infection. When these stroke suspects arrive at a hospital, they are triaged to determine whether they have a hemorrhage, or an ischemic stroke that requires immediate (within 4.5 h) thrombolysis [Hacke et al. 2008]. HIV-1 introduces a third variable, infection. There are few guidelines indicating when a lumbar puncture should supersede emergency thrombolysis. Most emergency rooms are poorly equipped to perform on-site HIV-1 testing. No studies have reported whether HIV+ patients with stroke in the setting of a brain infection are at increased risk of bleeding if they receive emergency thrombolysis.
In addition, stroke prevention may also be complicated by drug interactions with a cART regimen. All PIs are metabolized by and inhibit the cytochrome P450 system (CYP) 3A4 [Fichtenbaum and Gerber, 2002]. The NNRTIs nevirapine (Viramune) and efavirenz (Sustiva) are inducers of CYP3A4. Interactions with the P450 enzyme (CP450) system can result in drug toxicity and the loss of virologic control. Many of the drugs used in stroke prevention are either metabolized by these pathways or have additive toxicity.
There is a growing body of literature on drug interactions, which may be applicable to the treatment of stroke in the HIV+ population. For this publication, the web sites http://hivinsite.ucsf.edu/InSite and www.hiv-druginteractions.org were used as references, as well as specific articles. However, the literature on drug interactions is rapidly evolving, and we suggest that physicians consult multiple sources of information on a regular basis.
Recombinant tissue plasminogen activator or rtPA (Activase [Alteplase] Genentech, South San Francisco, CA, USA) is a thrombolytic drug used for the treatment of acute ischemic stroke [Wardlaw et al. 2012]. rtPA has not been reported to have serious interactions with cART; however, it is theoretically possible that bleeding could occur if rtPA is given with tipranavir due to increased inhibition of platelet aggregation by tipranivir [Ingelheim, 2011].
Dabigatrin (Pradaxa, Boehringer Ingleheim Pharmaceuticals, Ridgefield, CT, USA) is a direct thrombin inhibitor used to prevent thrombus formation [Dahl, 2012]. There are no reported interactions with cART. Warfarin (Coumadin, Bristol-Meyers Squibb, Princeton, New Jersey, USA) is an anticoagulant for long-term use that is metabolized by the CYP450 system. There are many interactions reported with PIs and NNRTIs [Liedtke and Rathbun, 2010]. Patients receiving cART and warfarin require aggressive monitoring of their international normalized ratio.
Aspirin is the most commonly used drug for stroke prevention. No specific drug interactions have been reported with cART, but there is a theoretical risk of renal damage when it is administered with tenofovir, as well as of increased bleeding when given with tipranavir. Tornero and colleagues observed that aspirin is under utilized in the HIV+ population for the primary prevention of cardiovascular events [Tornero et al. 2010]. No significant drug interactions have been reported among HIV+ patients treated with aspirin-extended release dipyridamole combinations (Aggrenox, Boehringer Ingleheim Pharmaceuticals, Ridgefield, CT, USA) and cART, or with clopidogrel (Plavix, Bristol-Myers Squibb/Sanofi Pharmaceuticals, Bridgewater NJ, USA), another antiplatelet agent.
Statins are the first-line therapy for hypercholesterolemia. Simvastatin (Zocor, Merck & Co., Inc., Whitehouse Station, NJ, USA), lovastatin (Mevacor, Merck & Co., Inc., Whitehouse Station, NJ, USA), and atorvastatin (Lipitor, Pfizer Inc., NY, NY,USA) are metabolized by the CYP450 pathway and have significant interactions with both PIs and NNRTIs [Fichtenbaum et al. 2002; Penzak and Chuck, 2002], resulting in elevated statin levels and clinical toxicity. Pravastatin (Pravachol, Bristol-Myers Squibb, Princeton, NJ USA) is not significantly metabolized by the CYP450 enzyme system, so the risk of toxicity is lower [Penzak and Chuck, 2002]. Drugs such as rosuvastain (Crestor, AstraZeneca Pharmaceuticals LP, Wilmington, DE, USA) can cause myositis and rhabdomyolysis; this can also be caused by some cART drugs such as raltegrivir [Kostapanos et al. 2010], raising concerns about additive toxicity. Fibrates such as gemfibrozil (Lopid, Pfizer, NY, NY, USA) and fenofibrate (Tricor, Abbott Laboratories, North Chicago, IL, USA) are used to treat hypertriglyceridaemia. They are metabolized via the CYP450 system and may interact with components of cART, resulting in lower gemfibrozil levels [Busse et al. 2009]. The use of marine ω3 fatty acids, that is, fish oil, to manage hypertriglyceridemia in cART-treated HIV+ patients is a promising area, and there are no significant interactions with cART [Oliveira and Rondo, 2011].
Diet and exercise are the first steps to manage diabetes in obese HIV+ cART-exposed patients. Obesity is a greater problem in cART-exposed patients than is wasting [Crum-Cianflone et al. 2008]. Insulin has not been reported to interact with cART. Metformin (Glucophage, Bristol-Meyers Squibb, Princeton, NJ, USA) is used to treat type II diabetics with normal renal function and helps manage dyslipidemia. However, metformin carries a small risk of lactic acidosis, which can also occur with the use of NRTIs such as stavudine and didanosine ([Videx] Bristol-Meyers Squibb, Princeton, NJ, USA) [Dragovic and Jevtovic, 2012]. Thiazolidinedione medications are used to control type II diabetes and have been studied as a possible treatment for lipodystrophy.
Varenicline (Chantix, Pfizer Inc., NY, NY, USA) is used to treat nicotine dependence. There are no reported interactions with cART. Tornero and Mafe reported an uncontrolled study in which varenicline was offered to cART-treated HIV+ patients for smoking cessation [Tornero and Mafe, 2009]. The tolerance for varencicline was poor, primarily because the patients were not motivated to quit, but no unusual drug interactions were noted. Varenicline was also studied in the Lung HIV Study and adverse effects were similar to those seen in the general population [Ferketich et al. 2012].
The treatment of hypertension in HIV+ patients parallels that in the general population. There is a potential for drug interactions between calcium channel blockers and both NNRTIs and PIs. This topic has been reviewed in depth by Peyriere and colleagues [Peyriere et al. 2012].
Stroke is a common and serious complication of HIV-1 infection and of cART. There are no HIV-specific protocols for the prevention, evaluation, or treatment of acute ischemic stroke despite the observation of an increased probability of an infectious cause of stroke in HIV+ patients, as well as specific adverse drug interactions between cART and medications used to treat stroke. The intersection of HIV-1 and stroke provides a window into neurovascular injury caused by a unique set of mechanisms that may be useful to gain insight into other diseases.
Funding: This research was supported by the National Institutes of Health (NIMH U01MH083500, NCI 5U01CA066529, NIMH 1R01MH083553, NIDA R01 DA030913, and NINDS P50 NS044378).
Conflict of interest statement: ES has been a consultant to Pfizer and Neurogesx.
Elyse J. Singer, National Neurological AIDS Bank, Department of Neurology, David Geffen School of Medicine at UCLA, 11645 Wilshire Blvd, Suite 770, Los Angeles, CA 90025, USA.
Miguel Valdes-Sueiras, Department of Neurology, David Geffen School of Medicine at University of California Los Angeles, Los Angeles, CA, USA.
Deborah L. Commins, Department of Pathology, University of Southern California Keck School of Medicine, Los Angeles, CA, USA.
William Yong, Department of Pathology and Laboratory Medicine, David Geffen School of Medicine at University of California Los Angeles, Los Angeles, CA, USA.
Margrit Carlson, Department of Medicine, Division of Infectious Diseases, David Geffen School of Medicine at University of California Los Angeles, Los Angeles, CA, USA.