- Email: [email protected]
Association Between Histoplasma Exposure and Stroke
Association Between Histoplasma Exposure and Stroke John D. Cleary, PharmD,*† Margaret Pearson, PharmD, MS,* Jake Oliver, PhD,‡ and Stanley W. Chapman, MD*†
Background and Purpose: The rate of cerebrovascular events within the stroke belt, a geographic area defined in the Southeastern United States, exceeds that of the rest of the nation. Despite evaluation of multiple risk factors for this disparity, specific causes for the stroke belt have not been elucidated. More than 45 years ago, the US Public Health Service and US Navy cooperative skin-testing program (1958-1965) documented adolescent exposure to Histoplasma capsulatum. Our purpose was to evaluate the association between exposure to Histoplasma capsulatum and subsequent development of stroke. Methods: A cross-sectional study of stroke in a cohort of US Navy veterans was designed to assess our hypothesis. Medical records from 23,795 men who participated in the cooperative skin-testing program and who received medical care at Veterans Administration Hospitals and Clinics were reviewed. A logistic regression model was used to estimate the odds ratio of stroke while controlling for multiple covariates. Because of the large number of possible risk factors for stroke, propensity scores were used to reduce bias. Results: The adjusted odds ratio for stroke in veterans exposed to Histoplasma capsulatum during adolescence was 1.34 (95% confidence interval: 1.1-1.6; P 5.0033). The increased risk was independent of traditional cerebrovascular event risk factors. Less frequent risk factors (atrial fibrillation, coronary heart disease, rheumatic heart disease, and prosthetic cardiac valvular replacement) were not controlled in this model. Conclusion: Exposure to Histoplasma capsulatum during adolescence was associated with an increased risk of stroke. Key Words: Cerebrovascular event—Histoplasma capsulatum—yeast—infection. Ó 2008 by National Stroke Association
Stroke is the third leading cause of death in the United States and is a source of significant morbidity for survivors. Geographic mapping of stroke incidence from as early as the 1930s has identified a stroke belt within the Southeastern United States wherein the rate of disease exceeds that of the rest of the nation.1,2 This disparity within
From the *School of Pharmacy, †Division of Infectious Diseases, and ‡Department of Preventive Medicine, University of Mississippi Medical Center, Jackson, Mississippi. Received May 2, 2007; revision received December 20, 2007; accepted January 8, 2008. Address correspondence to John D. Cleary, PharmD, School of Pharmacy, University of Mississippi Medical Center, 2500 N State St, Jackson, MS 39216. E-mail: [email protected]. 1052-3057/$—see front matter Ó 2008 by National Stroke Association doi:10.1016/j.jstrokecerebrovasdis.2008.01.015
312
the stroke belt continues today, with astounding effects on public health.3 Despite evaluation of multiple potential factors for a cerebrovascular event risk, including differences in socioeconomic status, environmental factors, and death certificate data,4-6 specific causes for the geographic distributions, especially the Southeastern US stroke belt, have not been elucidated (Fig 1, A). Atherosclerosis is one of the known risk factors for development of cerebrovascular events including transient ischemic attack (TIA) and stroke. The concept of an infectious cause contributing to the development of atherosclerosis is not new.7,8 Infectious agents implicated include Chlamydia pneumoniae, Helicobacter pylori, herpes virus, Cytomegalovirus (CMV), and the organisms causing periodontal disease. If indeed these organisms are involved in atherosclerosis and, therefore, pose subsequent risk for TIA or stroke, the ubiquitous epidemiology of each would render them unlikely candidates to shed light on
Journal of Stroke and Cerebrovascular Diseases, Vol. 17, No. 5 (September-October), 2008: pp 312-319
HISTOPLASMOSIS AND STROKE
313
Figure 1. (A) Data from Centers for Disease Control and Prevention/National Center for Chronic Disease Prevention and Health Promotion (http://apps.nccd.cdc.gov/giscvh/map. aspx) for all stroke cases from 1991 through 1998 graphed as stroke death rates that are age adjusted for average annual deaths/ 100,000 population. (B) US Public Health Service data on histoplasmin H-42 skin-test reactivity in US naval recruits as a percentage of individuals tested.
the excess stroke risk experienced within the Southeastern United States. One infectious organism that is endemic to the Southeast United States is Histoplasma capsulatum (Fig 1, B). In fact, results of histoplasmin skin reactivity from a 1958 to 1965 survey9 produced an epidemiologic map that appears to overlie much of the same boundary as the Southeastern US stroke belt. Our purpose was to assess the risk of nonhemorrhagic stroke, including the diagnosis of TIA, after Histoplasma capsulatum infection. Seroepidemiologic correlation is a key initial step in providing evidence that exposure or infection is associated with increased risk for vascular disease. Currently a database is available documenting exposure to Histoplasma more than 45 years ago in a population of US Navy recruits. These individuals are now, epidemiologically, at increased risk of experiencing a stroke-related event based on age alone. The US Public Health Service and US Navy cooperative skin-testing program enabled us to design a noninvasive, retrospective, cross-sectional study to investigate the association of Histoplasma capsulatum exposure with a cerebrovascular event.
Methods This study was designed as a retrospective, cross-sectional evaluation of the association between ischemic or thrombotic stroke, including TIAs, and adolescent or young adult exposure to Histoplasma capsulatum. The cohort was identified from a pool of male US Navy recruits tested in the US Public Health Service and US Navy cooperative skin-testing program (1958-1965) who had been receiving medical care within a Department of Veterans Affairs (VA) medical center since 1987. The cohort consisted of 1,228,038 US Navy recruits, of whom 23,795 were receiving care at a VA medical center. This research was conducted according to institutional and ethical standards as required by our institutional review board and office of compliance; VA approval was also obtained. Personal identifiers of patients were not released.
Assessment of Exposure During the cooperative skin-testing program, US Navy recruits received standard purified protein derivative
J.D. CLEARY ET AL.
314
(PPD)–Seibert (PPD-S) (tuberculin), PPD–Battey strain (PPD-B), variable antigen (VANT), and histoplasmin skin tests during the first few weeks of training. All tests were given as 0.1 mL of antigen, injected intradermally into the volar surface of the forearm. Tests were read usually at 48 hours by a small team of intensively trained US Public Health Service nurses. Measurements were recorded in millimeters at the widest transverse diameter of palpable induration. Data from these tests were used in reporting nationwide mapping of the prevalence of exposure to these organisms (Fig 1, B). Measurement of an induration greater than 5 mm was defined as a positive reaction. Each recruit provided, on arrival at the training center, demographic data and chronologic information of all places in which he had resided for 6 months or more since birth. The recruits were also asked to specify whether any place of residence had been a farm. In the current study, a case was identified through a computer search of the University of Mississippi Medical Center’s Mycotic Research Center naval recruit database. A case was defined as a recruit who received any medical care at a VA health care facility for any reason and had a diagnosis of an ischemic or thrombotic stroke, including TIAs, and coded using International Classification of Diseases (ICD) coding, sixth through ninth revisions. Diagnoses included by the investigators were cerebrovascular disease codes (433.0-437.9) including TIAs (436.9), but excluding hemorrhagic and heat stroke–related codes (746.9, 992, or 997.02) unless a cerebrovascular event code also occurred. A control was defined as a recruit, admitted for care at a VA health care facility, who did not have a diagnosis of stroke. Tuberculin skin-test results were also assessed as a separate control for infections associated with intracellular pathogens that lead to chronic inflammatory responses. VA medical center medical record data were reviewed for diagnostic evidence of a stroke, including imaging studies or Doppler ultrasound examinations of the carotid arteries. The World Health Organization definition of stroke is described as rapidly developing signs of focal or global disturbance of cerebral function, with no apparent cause other than stroke. This definition was used for the current study evaluations. Stroke was classified when clinical neurologic deficiency persisted for more than 24 hours or led to death. TIA was classified when resolution of clinical neurologic deficiency occurred within 24 hours and was included in our case cohort. Planned data collection included demographic information (age, race, sex, height, weight, alcohol abuse, tobacco abuse), stroke/TIA risk factors (history of stroke, migraine headaches, diabetes mellitus [insulin-dependent, noninsulin-dependent, not diabetic]), cardiovascular diseases (hypertension), morbid obesity, ischemic heart disease, myocardial infarction, congestive heart failure, cardiac arrhythmias, peripheral vascular disease, dyslipidemia (hypercholesterolemia, hypertriglyceridemia), serologic
evidence of infection (Chlamydia pneumoniae, CMV, Helicobacter pylori, Histoplasma capsulatum), steroid use (inhaled or systemic), and granulomatous disease. These stroke/ TIA risk factors were also identified by ICD codes. Codes selected for identification of risk-associated diseases to be used in logistic regression analysis includes ICD-associated terms from 1995 until 2005. The ICD codes included: obesity (278.*), history of diabetes (250.*), hypertension (401.*-405.*) (controlled, uncontrolled), hypercholesterolemia or hypertriglyceridemia (272.*), periodontal disease (522.* and 523.*), history of infection by an associated microbial pathogen (Chlamydia) (079.* and 078.*), Helicobacter pylori (041.86), CMV (07.85), migraine (346.*), ischemic heart disease (413.* and 414.*) or myocardial infarction (410.* and 411.*), arrhythmias (426.* and 427.*), congestive heart failure (428.*), peripheral vascular disease (443.*), and smoking. Smoking and ethanol use could not be accurately quantified from the records and, therefore, was collected as a dichotomous variable (yes/no) if present in the history. Atherosclerosis was not used as a stroke risk factor secondary to its primary role in the hypothesis.
Statistical Methodology Logistic regression analysis was used to evaluate the association of stroke and exposure to Histoplasma capsulatum (as evidenced by a positive histoplasmin skin reaction). Logistic regression allows for the simultaneous estimation of the effects of multiple covariates (i.e., vascular disease-associated risk factors including histoplasmin skin-reactivity data) on the odds of the response. However, the large number of known risk factors for stroke makes this prohibitive. Propensity scores were used to balance the effect of known risk factors with one covariate. The known risk factors used to estimate the propensity score included diabetes, hypertension, obesity, ischemic heart disease, myocardial infarction, congestive heart failure, arrhythmia, peripheral vascular disease, hypercholesterolemia, hypertriglyceridemia, periodontal disease, granulomatous disease, Chlamydia pneumoniae, CMV, Helicobacter pylori, and migraine headaches. The response variable within our model is occurrence of stroke. Matching of cohorts was not performed. The demographics of the recruits limit assessment beyond a male, white cohort. Statistical software package (SPSS Corp, Chicago, IL) was used for all described analysis. Literature was reviewed a priori10-13 to determine an expected adjusted odds ratio (OR) for risk of TIA or stroke and organism skin-test positivity. Required sample sizes to detect an OR of 1.5, with a 2-sided alpha level of 0.05 using software (nQuery Advisor, 3.0, STATCON, Boston, MA) revealed whether 5% of naval recruits had experienced TIA or stroke; sample size required to achieve 80% and 90% power to detect an OR of 1.5 were 1233 and 1637, respectively. These sample sizes would be
HISTOPLASMOSIS AND STROKE
315
Table 1. Patient demographics
Characteristic General recruit data Age (y, mean 6 SD) Race (Caucasian) (African American) Height (cm, mean 6 SD) Weight (kg, mean 6 SD) Hospital admission data Age in 2006 (years, mean 6 SD)
Stroke (n 5 672)
No stroke (n 5 23,123)
19.1 6 2.1 19.1 6 1.9 69.4% 71.8% 13.1% 9.5% 175 6 7.1 175.5 6 7.6 69.9 6 12.1 69.5 6 11.8 60.7 6 3.6
60.1 6 3.4
achieved from our nearly 23,000 naval recruits. A sample size adequate to maintain power at the 80% to 90% level with an OR of 1.5 would not exceed the population sample unless the proportion of naval recruits having experienced TIA or stroke decreased to less than 1%.
Results The actual cohort obtained from the US Navy consisted of 671,007 recruits participating in the US Public Health Service and US Navy cooperative skin-testing program (1958-1965). Of the recruits tested, all were male and 69% were white, 17 to 21 years old, and residents of the
continental United States. Approximately 250,000 had lived all their lives in a single county and received the Histoplasma skin test.9 Our sample consists of 23,795 of the 250,000 naval recruits skin tested, who were receiving medical care at a VA hospital. Patient demographics between those with stroke and control subjects were not significantly different (Table 1). Skin-test results were identified from the skin-testing program. As previously stated, nearly 250,000 recruits received the Histoplasma skin test during the program. More recruits received the PPD-S (135,890) and PPD-B (429,285) combined or the VANT (303,033). A substantial number of veterans received care at the VA medical centers and had positive Histoplasma reactions greater than or equal to 5 mm (N 5 2313) or greater than 10 mm (N 5 1580). Reactions to PPD-S greater than or equal to 5 mm were observed in 1960 recruits and 10 mm were measured in 1134. PPD-B resulted in more induration than the PPD-S measured at 5 and 10 mm for 7143 and 3527 veterans, respectively. VANT was measurable at greater than or equal to 5 mm in 3853 and greater than or equal to 10 mm in 1759. In recruits who had stroke or TIA, 3.2%, 2.2%, 2.1%, and 2.4% had positive skin-test results defined as an induration greater than 5 mm for Histoplasma, PPD-S, PPD-B, or VANT, respectively. There was no difference in PPD-S, PPD-B, or VANT reactivity between patients with and without stroke. The level of risk for known risk factors was first assessed by ORs (Table 2). Diabetes and hypertension
Table 2. Comparison of known stroke risk factors before and after propensity score stratification
Confounding disease Metabolic disorders Diabetes Obesity Hypercholesterolemia Hypertriglyceridemia Cardiovascular disease Hypertension Ischemic heart disease Myocardial infarction Congestive heart failure Arrhythmia Peripheral vascular disease Migraine Infectious diseases Periodontal disease Granulomatous disease Previous infectious diseases Chlamydia infection Cytomegalovirus infection Helicobacter pylori Histoplasmosis
Stroke (n 5 672)
No stroke (n 5 23,123)
OR before stratification (univariant)
OR after stratification (multivariant)
181 (26.9%) 57 (8.5%) 185 (27.5%) 10 (1.5%)
2969 (12.8%) 1461 (6.3%) 3378 (14.6%) 223 (1.0%)
2.50 1.37 2.22 1.55
0.91 0.94 1.18 0.88
419 (62.4%) 197 (29.3%) 63 (9.4%) 79 (11.8%) 142 (21.1%) 0 (0.0%) 28 (4.2%)
6840 (29.6%) 2834 (12.3%) 702 (3.0%) 861 (3.7%) 1959 (8.5%) 0 (0.0%) 265 (1.2%)
3.94 2.97 3.30 3.44 2.89 * 3.75
1.39 1.22 1.31 1.46 1.21 * 1.36
* *
* *
2.03 3.37 1.15 1.42
1.65 1.23 0.85 1.34
0 (0.0%) 0 (0.0%) 2 (0.3%) 4 (0.6%) 1 (0.2%) 135 (20.1%)
0 (0.0%) 6 (0.03%) 34 (0.2%) 41 (0.2%) 30 (0.1%) 3481 (15.1%)
Abbreviation: OR, odds ratio. *No recorded cases of peripheral vascular disease or periodontal disease.
J.D. CLEARY ET AL.
316
were associated with cases of stroke when compared with nonevent cases. Morbid obesity, or at least obesity coded by health care providers, was not independently associated with a stroke. Within this population, 3150 (13%) recruits had a diagnosis in 2003 of diabetes, 1518 (6.4%) had obesity, and 7259 (31%) had hypertension. The unadjusted level of risk for histoplasmosis was significant (OR 1.42; 95% confidence interval 1.2-1.7; P 5.0003). Tobacco use and ethanol abuse were unreliable by history leaving only qualitative data. These two variables did not impact our model and were dropped from the analysis. The goal of propensity scores is to stratify subjects into homogenous subgroups based on known risk factors.14 Through multiple logistic regression, propensity scores were computed by incorporating main effects of diabetes, hypertension, obesity, ischemic heart disease, arrhythmia, myocardial infarction, CMV, and migraine headaches. Two-way interactions include diabetes*migraines, hypertension*migraines, hypertension*myocardial infarction, and obesity*arrhythmia. The effect of each risk factor was assessed by OR before and after stratification (Table 2). For histoplasmosis, the association with a stroke remained significant after stratification (OR 1.34; 95% confidence interval 1.1-1.6; P 5 .0033). Assessment of region of habitation on enrollment into the US Navy revealed more patients in the at-risk region. Recruits from the Northeast (Connecticut; Delaware; Iowa; Maine; Maryland; Massachusetts; Michigan; Minnesota; New Hampshire; New Jersey; New York; Pennsylvania; Rhode Island; Vermont; Washington; Washington, DC; and Wisconsin) and Western (Alaska, Arizona, California, Colorado, Hawaii, Idaho, Kansas, Montana, Nebraska, Nevada, New Mexico, North Dakota, Oregon, South Dakota, Utah, Wyoming) continental United States and residents from foreign countries were categorized arbitrarily before analysis. The at-risk region includes the states in the Mississippi, Trinity, Red, and Ohio River regions. Of patients with stroke, 52% lived in this region as an adolescent or young adult and acquired histoplasmosis. Of the 24 patients who had an event from Louisiana, 14 (58.3%) of them had positive Histoplasma skin-test findings whereas only one (4.2%) had positive standard tuberculin skin-test result. Arkansas and Texas were the second and third most frequent states for patients. Mississippi was the only state in which the rate of histoplasmosis exposure (42.8%) was close to rates of tuberculin conversion (42.8%) in patients with stroke. Previous residence on a farm did not appear significant. Although tuberculin conversion served as a control for infections by intracellular organisms, there did not appear to be a similar geographic distribution or relation to a stroke when compared with Histoplasma skin testing. Global positioning analysis also allowed identification of current habitation for patients and control subjects. Crude mapping suggested that many
histoplasmosis skin test positive patients still live in proximity to the high-risk river basins.
Discussion Studies linking infection to vascular disease include: (1) seroepidemiologic associations, (2) pathologic demonstration of organisms in atherosclerotic plaques, (3) in vitro systems demonstrating activation of proinflammatory and procoagulant mediators, and (4) animal models of atherosclerosis induction. Currently there are no reports in the literature describing Histoplasma capsulatum as an infectious cause of atherosclerosis, nor are there seroepidemiologic reports relating an association of Histoplasma capsulatum with coronary artery disease or stroke. However, it is estimated that 10% of disseminated cases before 1953 experienced central nervous system (CNS) involvement.15 In a review of disseminated and nondisseminated cases between 1952 and 1960, CNS involvement was estimated to be 7.6%. It has been hypothesized that the higher rates in earlier studies were associated with increased severity of disease.16 Unfortunately, no study has tested the hypothesis that stroke or other cerebrovascular events could be an outcome associated with the spread of this yeast within the CNS. Increased dissemination in the era of human immunodeficiency virus raises additional concerns. It is known that Histoplasma capsulatum causes vascular pathology. Neovascular disease associated with ocular histoplasmosis syndrome involved the vascular endothelium in 12 of 18 specimens, with 9 and 10 specimens having fibrocyte and macrophage infiltration, respectively. Eight of 13 lesions were well demarcated with fibrovascular cores surrounded by outer faintly fluorescent rims; none were poorly demarcated.17 In the review of Karalakulasingam et al,18 Histoplasma capsulatum lesions within the CNS are classified into several histopathologic types, including perivenous granulomatosis and fibrinoid necrosis of arterial walls. Reports by other authors19,20 describe multiple CNS vessel narrowing and fibrinoid necrosis in meningeal arteries at the base of the brain. Necroscopy in case series and case reports of histoplasmosis from the 1950s to 1960s are enlightening. Vascular changes are evident in patients with obvious CNS involvement and secondary stroke. In a cerebrovascular event observed as a complication of histoplasmosis meningoencephalitis, small leptomeningeal vessels had mononuclear cell infiltration of outer vessel walls and early reactive intimal proliferation on hematoxylin and eosin staining on cross sections.21 Epithelioid cells, and histiocytic and rare giant cell infiltration, have also been observed below the meningeal vein intima.22 Several studies have assessed serologic markers of Chlamydia pneumoniae to support a link to stroke. Logistic regression analysis suggested a significant association of stroke or TIA with both acute and prior Chlamydia
HISTOPLASMOSIS AND STROKE
pneumoniae infection as evidenced by elevated antichlamydial IgG, IgM, and/or IgA titers.10 Logistic regression, with adjustment for traditional stroke risk factors, revealed an increased odds of first time stroke significantly associated with elevated antichlamydial IgA titers. A trend toward increased odds of association for first time stroke was seen with elevated antichlamydial IgG titers.11 Geometric means of antichlamydial IgA titers were significantly higher in patients with cerebral infarction or TIA than in matched control subjects. In addition, specific IgG titers in circulating immune complexes were also significantly higher in patients than control subjects.12,23 Pathologic features of plaques have been described as consistent with those of a chronic chlamydial granuloma of the arteries.24 All tissues positive for Chlamydia pneumoniae had lesions that were either definite atheromatous plaques or at least more than a diffuse intimal thickening. No coronary samples without lesions had evidence of Chlamydia pneumoniae.23,25 In specimens of carotid atheromatous plaque, a variety of cells demonstrated AY-6 (a Chlamydia pneumoniae-specific monoclonal antibody) immunoreactivity: endothelial cells in the intima near the atheromatous plaque and in the media; macrophages; and smooth-muscle cells that had migrated from the media into the plaque.26 Chlamydia pneumoniae DNA was identified by polymerase chain reaction in plaques from carotid endarterectomy samples,27 and by nested polymerase chain reaction in atherosclerotic middle cerebral arteries.28 Chronic or recurrent exposure to Histoplasma capsulatum may be the source of an additive risk of stroke. This specific hypothesis would be extremely difficult to test. However, evaluation of stroke incidence in naval recruits who remained in an endemic area compared with those who relocated after discharge may address this question. Associations between Helicobacter pylori or CMV infection and atherosclerosis have been explored. Serum samples from subjects from a large population-based study on the epidemiology and origin of atherosclerosis13 were used to determine prevalence of antibodies against Helicobacter pylori and CMV. When the analysis was restricted to low social status, anti-Helicobacter pylori antibodies were significantly correlated with both carotid atherosclerosis and intima-media thickness. No correlation between CMV IgG antibodies and either atherosclerosis or vessel wall thickness was found. But CMV, with female gender as a cofactor, was associated with statistically higher serum total cholesterol levels.29 Helicobacter pylori DNA was found in atherosclerotic plaques from carotid endarterectomy specimens in patients with severe vessel stenosis.30 In a case-control study of patients with acute ischemic stroke or TIA and matched control subjects, patients with atherothrombotic or cardioembolic stroke had higher serologic titers than control subjects by both univariate and multivariate analysis. For patients with atherosclerotic vessel disease, increased OR for
317
stroke associated with Helicobacter pylori seropositivity was noted in univariate but not in multivariate analysis.14 No evidence of any positive association for CMV seropositivity and ensuing stroke was found in a prospective, case-control study adjusting for cardiovascular disease. Of note, baseline prevalence of CMV seropositivity was similar between the case and control groups.31 The hypothesis has been raised that chronic infection, regardless of the particular infectious agent, may result in development of atherosclerosis. Indeed, in patients deemed to have chronic infections, serum levels of inflammatory markers were substantially enhanced. These patients experienced multiple unfavorable changes in vascular risk factors, and progressed to have atherosclerosis at a greater incidence than individuals without chronic infections.32 One chronic infection, periodontal disease, has been scrutinized as a cause of atherosclerosis. The secondary inflammation from lipopolysaccharideinduced cytokine release may promote atherosclerosis. A case-controlled study of 40 patients with stroke (age-/ sex-matched control subjects without stroke) revealed poor oral hygiene as a risk for stroke.33,34 A crosssectional study of 401 VA patients found poor oral hygiene associated with cerebrovascular accidents.34,35 We did not find periodontal disease coded in our patients. Ischemic brain injury (arterial obstruction) associated with local inflammation can be secondary to infection and is characterized by increased cytokines, including liver expressed C-reactive protein (CRP) and atrial natriuretic peptide. A poststroke hospital admission or discharge CRP that was elevated was predictive of increased mortality at 1 year. Increased mortality was observed in those with elevated CRP (57 of 128 patients) of whom 54.4% had a vascular event (nonfatal or fatal).36 In animal models of stroke, inflammatory proteins are increased in serum and urine for prolonged periods before the cerebrovascular event.37 It has been hypothesized that the increased inflammatory proteins either lead to increased inflammation within the arterial vasculature with secondary atherosclerosis or increased red blood cell aggregation, which leads to cerebrovascular accidents.38,39 In patients with elevated CRP concentrations in the blood, higher concentrations of IgG directed at oxidized low-density lipoproteins and malondialdehyde-low-density lipoproteins can also be observed.38 These events are suggestive that initiators of inflammatory responses may mediate stroke.
Summary Enhanced understanding of Histoplasma capsulatum as a potential inflammatory/infectious cause of stroke could have profound effects on public health. Reduced risk of a stroke through prevention (vaccine development and primary prevention) and altering the clinical course of a TIA (secondary prevention) could be achieved.
318
Seroepidemiologic correlation is a key initial step in providing evidence that this yeast infection may be associated with increased risk for TIA or stroke. Further observations conducted in an in vitro examination of intracellular growth and replication of Histoplasma capsulatum in human umbilical vein endothelial cells (Clonetics Inc, San Diego, Calif) and aortic smooth-muscle cells (ATCC, Manassas, Va) should enlighten the vascular pathogenesis. The Mycotic Research Center is also conducting a large endarterectomy study, prospectively evaluating patients with stroke for seroepidemiologic evidence and in vitro evaluation of vascular endothelium genomic for Histoplasma capsulatum antibodies and antigen, respectively. Acknowledgments: We would like to thank the University of Mississippi Medical Center’s Global Instruments Services for their assistance in mapping the geographic location of each patient based on current demographic data. The project would not have moved forward but for the tremendous insight and support on behalf of the Veterans Administration Benefits Office, The Office of the Navy, Centers for Disease Control and Prevention, and the US Public Health Service.
References 1. Wing S, Casper M, Davis WB, et al. Stroke mortality maps United States aged 35-74 years, 1962-1982. Stroke 1988; 19:1507-1513. 2. Lanska D. Geographic distribution of stroke mortality in the United States: 1939-41 to 1979-81. Neurology 1993; 43:1839-1851. 3. Howard G, Evans GW, Evans MS, et al. Is the stroke belt disappearing? An analysis of racial, temporal, and age effects. Stroke 1995;16:1153-1158. 4. Gillum RF, Ingram DD. Relation between residence in the Southeast region of the United States and stroke incidence: The NHANES I epidemiologic follow-up study. Am J Epidemiol 1996;144:665-673. 5. Howard G. Why do we have a stroke belt in the Southeastern United States? A review of unlikely and uninvestigated potential causes. Am J Med Sci 1999;317:160-167. 6. Perry HM, Roccella EJ. Conference report on stroke mortality in the Southeastern United States. Hypertension 1998;31:1206-1215. 7. Neito FJ. Infections and atherosclerosis: New clues form an old hypothesis? Am J Epidemiol 1998;148:937-948. 8. Shah PK. Link between infection and atherosclerosis. Who are the culprits: Viruses, bacteria, both, or neither [editorial]? Circulation 2001;103:5-7. 9. Edwards LB, Acquaviva FA, Livesay VT, et al. An atlas of sensitivity to tuberculin, PPD-B, and histoplasmin in the United States. Am Rev Resp Dis 1969;99(4 part 2):1-132. 10. Cook PJ, Honeybourne D, Lip GYH, et al. Chlamydia pneumoniae antibody titers are significantly associated with acute stroke and transient cerebral ischemia: The West Birmingham stroke project. Stroke 1998;29:404-410. 11. Elkind MSV, Lin I, Grayston JT, et al. Chlamydia pneumoniae and the risk of first ischemic stroke: The Northern Manhattan stroke study. Stroke 2000;31:1521-1525.
J.D. CLEARY ET AL. 12. Wimmer MLJ, Sandmann-Strupp R, Saikku P, et al. Association of chlamydial infection with cerebrovascular disease. Stroke 1996;27:2207-2210. 13. Mayr M, Keichl S, Willeit J, et al. Infections, immunity, and atherosclerosis: Associations of antibodies to Chlamydia pneumoniae, Helicobacter pylori, and Cytomegalovirus with immune reactions to heat-shock protein 60 and carotid or femoral atherosclerosis. Circulation 2000; 102:833-839. 14. D’Agostino RB Jr. Propensity score methods for bias reduction in the comparison of a treatment to a nonrandomized control group. Stat Med 1998;17:2265-2281. 15. Schulz DM. Histoplasmosis of the central nervous system. JAMA 1953;151:549. 16. Cooper RA, Goldstein E. Histoplasmosis of the central nervous system. Am J Med 1963;35:45-57. 17. Grossniklaus HE, Hutchinson AK, Capone A, et al. Clinicopathologic features of surgically excised choroidal neovascular membranes. Ophthalmology 1994;101(6): 1099-1111. 18. Karalakulasingam R, Arora KK, Adams G, et al. Meningoencephalitis caused by Histoplasma capsulatum. Arch Intern Med 1976;136:217-220. 19. Stone JH, Pomper MG, Hellmann DB. Histoplasmosis mimicking vasculitis of the central nervous system. J Rheumatol 1998;25:1644-1648. 20. Shapiro JL, Lux JJ, Sprofkin BE. Histoplasmosis of the central nervous system. Am J Pathol 1955;31:319-335. 21. McGregor JA, Kleinschmidt-DeMasters BK, Ogle J. Meningoencephalitis caused by Histoplasma capsulatum complicating pregnancy. Am J Obstet Gynecol 1985; 154:925-931. 22. Parsons RJ, Zarafonetis CJD. Histoplasmosis in man: Report of seven cases and a review of seventy-one cases. Arch Intern Med 1945;75:1-23. 23. Nieto FJ, Folsom AR, Sorlie PD, et al. Chlamydia pneumoniae infection and incident coronary heart disease: The atherosclerosis risk in communities study. Am J Epidemiol 1999;150:149-156. 24. Shor A. A pathologist’s view of organisms and human atherosclerosis. J Infect Dis 2001;183:1428-1429. 25. Kuo CC, Grayston JT, Campbell LA, et al. Chlamydia pneumoniae (TWAR) in coronary arteries of young adults (1534 years old). Proc Natl Acad Sci U S A 1995;92:6911-6914. 26. Yamashita K, Ouchi K, Shirai M, et al. Distribution of Chlamydia pneumoniae infection in the atherosclerotic carotid artery. Stroke 1998;29:773-778. 27. LaBiche R, Koziol D, Quinn T, et al. Presence of Chlamydia pneumoniae in human symptomatic and asymptomatic carotid atherosclerotic plaque. Stroke 2001;32:855-860. 28. Virok D, Kis Z, Karai L, et al. Chlamydia pneumoniae in atherosclerotic middle cerebral artery. Stroke 2001; 32:1973-1978. 29. Froberg MK, Seacotte N, Dahlberg E. Cytomegalovirus seropositivity and serum total cholesterol levels in young patients. Ann Clin Lab Sci 2001;31:157-161. 30. Ameriso SF, Fridman EA, Leiguarda RC, et al. Detection of Helicobacter pylori in human carotid atherosclerotic plaques. Stroke 2001;32:385-391. 31. Ridker PM, Hennekens CH, Stampfer MJ, et al. Prospective study of herpes simplex virus, Cytomegalovirus, and the risk of future myocardial infarction and stroke. Circulation 1998;98:2796-2799. 32. Kiechl S, Egger G, Mayr M, et al. Chronic infections and the risk of carotid atherosclerosis: Prospective results form a large population study. Circulation 2001; 103:1064-1070.
HISTOPLASMOSIS AND STROKE 33. Syrjanen J, Peltola J, Valtonen M, et al. Dental infections in association with cerebral infarction in young and middleaged men. J Intern Med 1989;225:179-184. 34. Xiaojung LI, Kolltveit KM, Tronstad L, et al. Systemic diseases caused by oral infection. Clin Microbiol Rev 2000; 13:547-558. 35. Loesche WJ, Lopatin DE. Interactions between periodontal disease, medical diseases and immunity in the older individual. Periodontal 2000;16:80-105. 36. Napoli MD, Papa F, Bocola V. C-reactive protein in ischemic stroke: An independent prognostic factor. Stroke 2001;32:917-924.
319 37. Sironi L, Tremoli E, Miller I, et al. Acute-phase proteins before cerebral ischemia in stroke-prone rats. Stroke 2001;32:753-760. 38. Hulthe J, Wikstrand J, Fagerberg B. Relationship between C-reactive protein and intima-media thickness in the carotid and femoral arteries and to antibodies against oxidized low-density lipoprotein in healthy me: The atherosclerosis and insulin resistance (AIR) study. Clin Sci (Colch) 2001;100:371-378. 39. Ami RB, Barshtein G, Zeltser D, et al. Parameters of red blood cell aggregation as correlates of the inflammatory state. Am J Physiol Heart Circ Physiol 2001;280: H1982-H1988.
Background and Purpose: The rate of cerebrovascular events within the stroke belt, a geographic area defined in the Southeastern United States, exceeds that of the rest of the nation. Despite evaluation of multiple risk factors for this disparity, specific causes for the stroke belt have not been elucidated. More than 45 years ago, the US Public Health Service and US Navy cooperative skin-testing program (1958-1965) documented adolescent exposure to Histoplasma capsulatum. Our purpose was to evaluate the association between exposure to Histoplasma capsulatum and subsequent development of stroke. Methods: A cross-sectional study of stroke in a cohort of US Navy veterans was designed to assess our hypothesis. Medical records from 23,795 men who participated in the cooperative skin-testing program and who received medical care at Veterans Administration Hospitals and Clinics were reviewed. A logistic regression model was used to estimate the odds ratio of stroke while controlling for multiple covariates. Because of the large number of possible risk factors for stroke, propensity scores were used to reduce bias. Results: The adjusted odds ratio for stroke in veterans exposed to Histoplasma capsulatum during adolescence was 1.34 (95% confidence interval: 1.1-1.6; P 5.0033). The increased risk was independent of traditional cerebrovascular event risk factors. Less frequent risk factors (atrial fibrillation, coronary heart disease, rheumatic heart disease, and prosthetic cardiac valvular replacement) were not controlled in this model. Conclusion: Exposure to Histoplasma capsulatum during adolescence was associated with an increased risk of stroke. Key Words: Cerebrovascular event—Histoplasma capsulatum—yeast—infection. Ó 2008 by National Stroke Association
Stroke is the third leading cause of death in the United States and is a source of significant morbidity for survivors. Geographic mapping of stroke incidence from as early as the 1930s has identified a stroke belt within the Southeastern United States wherein the rate of disease exceeds that of the rest of the nation.1,2 This disparity within
From the *School of Pharmacy, †Division of Infectious Diseases, and ‡Department of Preventive Medicine, University of Mississippi Medical Center, Jackson, Mississippi. Received May 2, 2007; revision received December 20, 2007; accepted January 8, 2008. Address correspondence to John D. Cleary, PharmD, School of Pharmacy, University of Mississippi Medical Center, 2500 N State St, Jackson, MS 39216. E-mail: [email protected]. 1052-3057/$—see front matter Ó 2008 by National Stroke Association doi:10.1016/j.jstrokecerebrovasdis.2008.01.015
312
the stroke belt continues today, with astounding effects on public health.3 Despite evaluation of multiple potential factors for a cerebrovascular event risk, including differences in socioeconomic status, environmental factors, and death certificate data,4-6 specific causes for the geographic distributions, especially the Southeastern US stroke belt, have not been elucidated (Fig 1, A). Atherosclerosis is one of the known risk factors for development of cerebrovascular events including transient ischemic attack (TIA) and stroke. The concept of an infectious cause contributing to the development of atherosclerosis is not new.7,8 Infectious agents implicated include Chlamydia pneumoniae, Helicobacter pylori, herpes virus, Cytomegalovirus (CMV), and the organisms causing periodontal disease. If indeed these organisms are involved in atherosclerosis and, therefore, pose subsequent risk for TIA or stroke, the ubiquitous epidemiology of each would render them unlikely candidates to shed light on
Journal of Stroke and Cerebrovascular Diseases, Vol. 17, No. 5 (September-October), 2008: pp 312-319
HISTOPLASMOSIS AND STROKE
313
Figure 1. (A) Data from Centers for Disease Control and Prevention/National Center for Chronic Disease Prevention and Health Promotion (http://apps.nccd.cdc.gov/giscvh/map. aspx) for all stroke cases from 1991 through 1998 graphed as stroke death rates that are age adjusted for average annual deaths/ 100,000 population. (B) US Public Health Service data on histoplasmin H-42 skin-test reactivity in US naval recruits as a percentage of individuals tested.
the excess stroke risk experienced within the Southeastern United States. One infectious organism that is endemic to the Southeast United States is Histoplasma capsulatum (Fig 1, B). In fact, results of histoplasmin skin reactivity from a 1958 to 1965 survey9 produced an epidemiologic map that appears to overlie much of the same boundary as the Southeastern US stroke belt. Our purpose was to assess the risk of nonhemorrhagic stroke, including the diagnosis of TIA, after Histoplasma capsulatum infection. Seroepidemiologic correlation is a key initial step in providing evidence that exposure or infection is associated with increased risk for vascular disease. Currently a database is available documenting exposure to Histoplasma more than 45 years ago in a population of US Navy recruits. These individuals are now, epidemiologically, at increased risk of experiencing a stroke-related event based on age alone. The US Public Health Service and US Navy cooperative skin-testing program enabled us to design a noninvasive, retrospective, cross-sectional study to investigate the association of Histoplasma capsulatum exposure with a cerebrovascular event.
Methods This study was designed as a retrospective, cross-sectional evaluation of the association between ischemic or thrombotic stroke, including TIAs, and adolescent or young adult exposure to Histoplasma capsulatum. The cohort was identified from a pool of male US Navy recruits tested in the US Public Health Service and US Navy cooperative skin-testing program (1958-1965) who had been receiving medical care within a Department of Veterans Affairs (VA) medical center since 1987. The cohort consisted of 1,228,038 US Navy recruits, of whom 23,795 were receiving care at a VA medical center. This research was conducted according to institutional and ethical standards as required by our institutional review board and office of compliance; VA approval was also obtained. Personal identifiers of patients were not released.
Assessment of Exposure During the cooperative skin-testing program, US Navy recruits received standard purified protein derivative
J.D. CLEARY ET AL.
314
(PPD)–Seibert (PPD-S) (tuberculin), PPD–Battey strain (PPD-B), variable antigen (VANT), and histoplasmin skin tests during the first few weeks of training. All tests were given as 0.1 mL of antigen, injected intradermally into the volar surface of the forearm. Tests were read usually at 48 hours by a small team of intensively trained US Public Health Service nurses. Measurements were recorded in millimeters at the widest transverse diameter of palpable induration. Data from these tests were used in reporting nationwide mapping of the prevalence of exposure to these organisms (Fig 1, B). Measurement of an induration greater than 5 mm was defined as a positive reaction. Each recruit provided, on arrival at the training center, demographic data and chronologic information of all places in which he had resided for 6 months or more since birth. The recruits were also asked to specify whether any place of residence had been a farm. In the current study, a case was identified through a computer search of the University of Mississippi Medical Center’s Mycotic Research Center naval recruit database. A case was defined as a recruit who received any medical care at a VA health care facility for any reason and had a diagnosis of an ischemic or thrombotic stroke, including TIAs, and coded using International Classification of Diseases (ICD) coding, sixth through ninth revisions. Diagnoses included by the investigators were cerebrovascular disease codes (433.0-437.9) including TIAs (436.9), but excluding hemorrhagic and heat stroke–related codes (746.9, 992, or 997.02) unless a cerebrovascular event code also occurred. A control was defined as a recruit, admitted for care at a VA health care facility, who did not have a diagnosis of stroke. Tuberculin skin-test results were also assessed as a separate control for infections associated with intracellular pathogens that lead to chronic inflammatory responses. VA medical center medical record data were reviewed for diagnostic evidence of a stroke, including imaging studies or Doppler ultrasound examinations of the carotid arteries. The World Health Organization definition of stroke is described as rapidly developing signs of focal or global disturbance of cerebral function, with no apparent cause other than stroke. This definition was used for the current study evaluations. Stroke was classified when clinical neurologic deficiency persisted for more than 24 hours or led to death. TIA was classified when resolution of clinical neurologic deficiency occurred within 24 hours and was included in our case cohort. Planned data collection included demographic information (age, race, sex, height, weight, alcohol abuse, tobacco abuse), stroke/TIA risk factors (history of stroke, migraine headaches, diabetes mellitus [insulin-dependent, noninsulin-dependent, not diabetic]), cardiovascular diseases (hypertension), morbid obesity, ischemic heart disease, myocardial infarction, congestive heart failure, cardiac arrhythmias, peripheral vascular disease, dyslipidemia (hypercholesterolemia, hypertriglyceridemia), serologic
evidence of infection (Chlamydia pneumoniae, CMV, Helicobacter pylori, Histoplasma capsulatum), steroid use (inhaled or systemic), and granulomatous disease. These stroke/ TIA risk factors were also identified by ICD codes. Codes selected for identification of risk-associated diseases to be used in logistic regression analysis includes ICD-associated terms from 1995 until 2005. The ICD codes included: obesity (278.*), history of diabetes (250.*), hypertension (401.*-405.*) (controlled, uncontrolled), hypercholesterolemia or hypertriglyceridemia (272.*), periodontal disease (522.* and 523.*), history of infection by an associated microbial pathogen (Chlamydia) (079.* and 078.*), Helicobacter pylori (041.86), CMV (07.85), migraine (346.*), ischemic heart disease (413.* and 414.*) or myocardial infarction (410.* and 411.*), arrhythmias (426.* and 427.*), congestive heart failure (428.*), peripheral vascular disease (443.*), and smoking. Smoking and ethanol use could not be accurately quantified from the records and, therefore, was collected as a dichotomous variable (yes/no) if present in the history. Atherosclerosis was not used as a stroke risk factor secondary to its primary role in the hypothesis.
Statistical Methodology Logistic regression analysis was used to evaluate the association of stroke and exposure to Histoplasma capsulatum (as evidenced by a positive histoplasmin skin reaction). Logistic regression allows for the simultaneous estimation of the effects of multiple covariates (i.e., vascular disease-associated risk factors including histoplasmin skin-reactivity data) on the odds of the response. However, the large number of known risk factors for stroke makes this prohibitive. Propensity scores were used to balance the effect of known risk factors with one covariate. The known risk factors used to estimate the propensity score included diabetes, hypertension, obesity, ischemic heart disease, myocardial infarction, congestive heart failure, arrhythmia, peripheral vascular disease, hypercholesterolemia, hypertriglyceridemia, periodontal disease, granulomatous disease, Chlamydia pneumoniae, CMV, Helicobacter pylori, and migraine headaches. The response variable within our model is occurrence of stroke. Matching of cohorts was not performed. The demographics of the recruits limit assessment beyond a male, white cohort. Statistical software package (SPSS Corp, Chicago, IL) was used for all described analysis. Literature was reviewed a priori10-13 to determine an expected adjusted odds ratio (OR) for risk of TIA or stroke and organism skin-test positivity. Required sample sizes to detect an OR of 1.5, with a 2-sided alpha level of 0.05 using software (nQuery Advisor, 3.0, STATCON, Boston, MA) revealed whether 5% of naval recruits had experienced TIA or stroke; sample size required to achieve 80% and 90% power to detect an OR of 1.5 were 1233 and 1637, respectively. These sample sizes would be
HISTOPLASMOSIS AND STROKE
315
Table 1. Patient demographics
Characteristic General recruit data Age (y, mean 6 SD) Race (Caucasian) (African American) Height (cm, mean 6 SD) Weight (kg, mean 6 SD) Hospital admission data Age in 2006 (years, mean 6 SD)
Stroke (n 5 672)
No stroke (n 5 23,123)
19.1 6 2.1 19.1 6 1.9 69.4% 71.8% 13.1% 9.5% 175 6 7.1 175.5 6 7.6 69.9 6 12.1 69.5 6 11.8 60.7 6 3.6
60.1 6 3.4
achieved from our nearly 23,000 naval recruits. A sample size adequate to maintain power at the 80% to 90% level with an OR of 1.5 would not exceed the population sample unless the proportion of naval recruits having experienced TIA or stroke decreased to less than 1%.
Results The actual cohort obtained from the US Navy consisted of 671,007 recruits participating in the US Public Health Service and US Navy cooperative skin-testing program (1958-1965). Of the recruits tested, all were male and 69% were white, 17 to 21 years old, and residents of the
continental United States. Approximately 250,000 had lived all their lives in a single county and received the Histoplasma skin test.9 Our sample consists of 23,795 of the 250,000 naval recruits skin tested, who were receiving medical care at a VA hospital. Patient demographics between those with stroke and control subjects were not significantly different (Table 1). Skin-test results were identified from the skin-testing program. As previously stated, nearly 250,000 recruits received the Histoplasma skin test during the program. More recruits received the PPD-S (135,890) and PPD-B (429,285) combined or the VANT (303,033). A substantial number of veterans received care at the VA medical centers and had positive Histoplasma reactions greater than or equal to 5 mm (N 5 2313) or greater than 10 mm (N 5 1580). Reactions to PPD-S greater than or equal to 5 mm were observed in 1960 recruits and 10 mm were measured in 1134. PPD-B resulted in more induration than the PPD-S measured at 5 and 10 mm for 7143 and 3527 veterans, respectively. VANT was measurable at greater than or equal to 5 mm in 3853 and greater than or equal to 10 mm in 1759. In recruits who had stroke or TIA, 3.2%, 2.2%, 2.1%, and 2.4% had positive skin-test results defined as an induration greater than 5 mm for Histoplasma, PPD-S, PPD-B, or VANT, respectively. There was no difference in PPD-S, PPD-B, or VANT reactivity between patients with and without stroke. The level of risk for known risk factors was first assessed by ORs (Table 2). Diabetes and hypertension
Table 2. Comparison of known stroke risk factors before and after propensity score stratification
Confounding disease Metabolic disorders Diabetes Obesity Hypercholesterolemia Hypertriglyceridemia Cardiovascular disease Hypertension Ischemic heart disease Myocardial infarction Congestive heart failure Arrhythmia Peripheral vascular disease Migraine Infectious diseases Periodontal disease Granulomatous disease Previous infectious diseases Chlamydia infection Cytomegalovirus infection Helicobacter pylori Histoplasmosis
Stroke (n 5 672)
No stroke (n 5 23,123)
OR before stratification (univariant)
OR after stratification (multivariant)
181 (26.9%) 57 (8.5%) 185 (27.5%) 10 (1.5%)
2969 (12.8%) 1461 (6.3%) 3378 (14.6%) 223 (1.0%)
2.50 1.37 2.22 1.55
0.91 0.94 1.18 0.88
419 (62.4%) 197 (29.3%) 63 (9.4%) 79 (11.8%) 142 (21.1%) 0 (0.0%) 28 (4.2%)
6840 (29.6%) 2834 (12.3%) 702 (3.0%) 861 (3.7%) 1959 (8.5%) 0 (0.0%) 265 (1.2%)
3.94 2.97 3.30 3.44 2.89 * 3.75
1.39 1.22 1.31 1.46 1.21 * 1.36
* *
* *
2.03 3.37 1.15 1.42
1.65 1.23 0.85 1.34
0 (0.0%) 0 (0.0%) 2 (0.3%) 4 (0.6%) 1 (0.2%) 135 (20.1%)
0 (0.0%) 6 (0.03%) 34 (0.2%) 41 (0.2%) 30 (0.1%) 3481 (15.1%)
Abbreviation: OR, odds ratio. *No recorded cases of peripheral vascular disease or periodontal disease.
J.D. CLEARY ET AL.
316
were associated with cases of stroke when compared with nonevent cases. Morbid obesity, or at least obesity coded by health care providers, was not independently associated with a stroke. Within this population, 3150 (13%) recruits had a diagnosis in 2003 of diabetes, 1518 (6.4%) had obesity, and 7259 (31%) had hypertension. The unadjusted level of risk for histoplasmosis was significant (OR 1.42; 95% confidence interval 1.2-1.7; P 5.0003). Tobacco use and ethanol abuse were unreliable by history leaving only qualitative data. These two variables did not impact our model and were dropped from the analysis. The goal of propensity scores is to stratify subjects into homogenous subgroups based on known risk factors.14 Through multiple logistic regression, propensity scores were computed by incorporating main effects of diabetes, hypertension, obesity, ischemic heart disease, arrhythmia, myocardial infarction, CMV, and migraine headaches. Two-way interactions include diabetes*migraines, hypertension*migraines, hypertension*myocardial infarction, and obesity*arrhythmia. The effect of each risk factor was assessed by OR before and after stratification (Table 2). For histoplasmosis, the association with a stroke remained significant after stratification (OR 1.34; 95% confidence interval 1.1-1.6; P 5 .0033). Assessment of region of habitation on enrollment into the US Navy revealed more patients in the at-risk region. Recruits from the Northeast (Connecticut; Delaware; Iowa; Maine; Maryland; Massachusetts; Michigan; Minnesota; New Hampshire; New Jersey; New York; Pennsylvania; Rhode Island; Vermont; Washington; Washington, DC; and Wisconsin) and Western (Alaska, Arizona, California, Colorado, Hawaii, Idaho, Kansas, Montana, Nebraska, Nevada, New Mexico, North Dakota, Oregon, South Dakota, Utah, Wyoming) continental United States and residents from foreign countries were categorized arbitrarily before analysis. The at-risk region includes the states in the Mississippi, Trinity, Red, and Ohio River regions. Of patients with stroke, 52% lived in this region as an adolescent or young adult and acquired histoplasmosis. Of the 24 patients who had an event from Louisiana, 14 (58.3%) of them had positive Histoplasma skin-test findings whereas only one (4.2%) had positive standard tuberculin skin-test result. Arkansas and Texas were the second and third most frequent states for patients. Mississippi was the only state in which the rate of histoplasmosis exposure (42.8%) was close to rates of tuberculin conversion (42.8%) in patients with stroke. Previous residence on a farm did not appear significant. Although tuberculin conversion served as a control for infections by intracellular organisms, there did not appear to be a similar geographic distribution or relation to a stroke when compared with Histoplasma skin testing. Global positioning analysis also allowed identification of current habitation for patients and control subjects. Crude mapping suggested that many
histoplasmosis skin test positive patients still live in proximity to the high-risk river basins.
Discussion Studies linking infection to vascular disease include: (1) seroepidemiologic associations, (2) pathologic demonstration of organisms in atherosclerotic plaques, (3) in vitro systems demonstrating activation of proinflammatory and procoagulant mediators, and (4) animal models of atherosclerosis induction. Currently there are no reports in the literature describing Histoplasma capsulatum as an infectious cause of atherosclerosis, nor are there seroepidemiologic reports relating an association of Histoplasma capsulatum with coronary artery disease or stroke. However, it is estimated that 10% of disseminated cases before 1953 experienced central nervous system (CNS) involvement.15 In a review of disseminated and nondisseminated cases between 1952 and 1960, CNS involvement was estimated to be 7.6%. It has been hypothesized that the higher rates in earlier studies were associated with increased severity of disease.16 Unfortunately, no study has tested the hypothesis that stroke or other cerebrovascular events could be an outcome associated with the spread of this yeast within the CNS. Increased dissemination in the era of human immunodeficiency virus raises additional concerns. It is known that Histoplasma capsulatum causes vascular pathology. Neovascular disease associated with ocular histoplasmosis syndrome involved the vascular endothelium in 12 of 18 specimens, with 9 and 10 specimens having fibrocyte and macrophage infiltration, respectively. Eight of 13 lesions were well demarcated with fibrovascular cores surrounded by outer faintly fluorescent rims; none were poorly demarcated.17 In the review of Karalakulasingam et al,18 Histoplasma capsulatum lesions within the CNS are classified into several histopathologic types, including perivenous granulomatosis and fibrinoid necrosis of arterial walls. Reports by other authors19,20 describe multiple CNS vessel narrowing and fibrinoid necrosis in meningeal arteries at the base of the brain. Necroscopy in case series and case reports of histoplasmosis from the 1950s to 1960s are enlightening. Vascular changes are evident in patients with obvious CNS involvement and secondary stroke. In a cerebrovascular event observed as a complication of histoplasmosis meningoencephalitis, small leptomeningeal vessels had mononuclear cell infiltration of outer vessel walls and early reactive intimal proliferation on hematoxylin and eosin staining on cross sections.21 Epithelioid cells, and histiocytic and rare giant cell infiltration, have also been observed below the meningeal vein intima.22 Several studies have assessed serologic markers of Chlamydia pneumoniae to support a link to stroke. Logistic regression analysis suggested a significant association of stroke or TIA with both acute and prior Chlamydia
HISTOPLASMOSIS AND STROKE
pneumoniae infection as evidenced by elevated antichlamydial IgG, IgM, and/or IgA titers.10 Logistic regression, with adjustment for traditional stroke risk factors, revealed an increased odds of first time stroke significantly associated with elevated antichlamydial IgA titers. A trend toward increased odds of association for first time stroke was seen with elevated antichlamydial IgG titers.11 Geometric means of antichlamydial IgA titers were significantly higher in patients with cerebral infarction or TIA than in matched control subjects. In addition, specific IgG titers in circulating immune complexes were also significantly higher in patients than control subjects.12,23 Pathologic features of plaques have been described as consistent with those of a chronic chlamydial granuloma of the arteries.24 All tissues positive for Chlamydia pneumoniae had lesions that were either definite atheromatous plaques or at least more than a diffuse intimal thickening. No coronary samples without lesions had evidence of Chlamydia pneumoniae.23,25 In specimens of carotid atheromatous plaque, a variety of cells demonstrated AY-6 (a Chlamydia pneumoniae-specific monoclonal antibody) immunoreactivity: endothelial cells in the intima near the atheromatous plaque and in the media; macrophages; and smooth-muscle cells that had migrated from the media into the plaque.26 Chlamydia pneumoniae DNA was identified by polymerase chain reaction in plaques from carotid endarterectomy samples,27 and by nested polymerase chain reaction in atherosclerotic middle cerebral arteries.28 Chronic or recurrent exposure to Histoplasma capsulatum may be the source of an additive risk of stroke. This specific hypothesis would be extremely difficult to test. However, evaluation of stroke incidence in naval recruits who remained in an endemic area compared with those who relocated after discharge may address this question. Associations between Helicobacter pylori or CMV infection and atherosclerosis have been explored. Serum samples from subjects from a large population-based study on the epidemiology and origin of atherosclerosis13 were used to determine prevalence of antibodies against Helicobacter pylori and CMV. When the analysis was restricted to low social status, anti-Helicobacter pylori antibodies were significantly correlated with both carotid atherosclerosis and intima-media thickness. No correlation between CMV IgG antibodies and either atherosclerosis or vessel wall thickness was found. But CMV, with female gender as a cofactor, was associated with statistically higher serum total cholesterol levels.29 Helicobacter pylori DNA was found in atherosclerotic plaques from carotid endarterectomy specimens in patients with severe vessel stenosis.30 In a case-control study of patients with acute ischemic stroke or TIA and matched control subjects, patients with atherothrombotic or cardioembolic stroke had higher serologic titers than control subjects by both univariate and multivariate analysis. For patients with atherosclerotic vessel disease, increased OR for
317
stroke associated with Helicobacter pylori seropositivity was noted in univariate but not in multivariate analysis.14 No evidence of any positive association for CMV seropositivity and ensuing stroke was found in a prospective, case-control study adjusting for cardiovascular disease. Of note, baseline prevalence of CMV seropositivity was similar between the case and control groups.31 The hypothesis has been raised that chronic infection, regardless of the particular infectious agent, may result in development of atherosclerosis. Indeed, in patients deemed to have chronic infections, serum levels of inflammatory markers were substantially enhanced. These patients experienced multiple unfavorable changes in vascular risk factors, and progressed to have atherosclerosis at a greater incidence than individuals without chronic infections.32 One chronic infection, periodontal disease, has been scrutinized as a cause of atherosclerosis. The secondary inflammation from lipopolysaccharideinduced cytokine release may promote atherosclerosis. A case-controlled study of 40 patients with stroke (age-/ sex-matched control subjects without stroke) revealed poor oral hygiene as a risk for stroke.33,34 A crosssectional study of 401 VA patients found poor oral hygiene associated with cerebrovascular accidents.34,35 We did not find periodontal disease coded in our patients. Ischemic brain injury (arterial obstruction) associated with local inflammation can be secondary to infection and is characterized by increased cytokines, including liver expressed C-reactive protein (CRP) and atrial natriuretic peptide. A poststroke hospital admission or discharge CRP that was elevated was predictive of increased mortality at 1 year. Increased mortality was observed in those with elevated CRP (57 of 128 patients) of whom 54.4% had a vascular event (nonfatal or fatal).36 In animal models of stroke, inflammatory proteins are increased in serum and urine for prolonged periods before the cerebrovascular event.37 It has been hypothesized that the increased inflammatory proteins either lead to increased inflammation within the arterial vasculature with secondary atherosclerosis or increased red blood cell aggregation, which leads to cerebrovascular accidents.38,39 In patients with elevated CRP concentrations in the blood, higher concentrations of IgG directed at oxidized low-density lipoproteins and malondialdehyde-low-density lipoproteins can also be observed.38 These events are suggestive that initiators of inflammatory responses may mediate stroke.
Summary Enhanced understanding of Histoplasma capsulatum as a potential inflammatory/infectious cause of stroke could have profound effects on public health. Reduced risk of a stroke through prevention (vaccine development and primary prevention) and altering the clinical course of a TIA (secondary prevention) could be achieved.
318
Seroepidemiologic correlation is a key initial step in providing evidence that this yeast infection may be associated with increased risk for TIA or stroke. Further observations conducted in an in vitro examination of intracellular growth and replication of Histoplasma capsulatum in human umbilical vein endothelial cells (Clonetics Inc, San Diego, Calif) and aortic smooth-muscle cells (ATCC, Manassas, Va) should enlighten the vascular pathogenesis. The Mycotic Research Center is also conducting a large endarterectomy study, prospectively evaluating patients with stroke for seroepidemiologic evidence and in vitro evaluation of vascular endothelium genomic for Histoplasma capsulatum antibodies and antigen, respectively. Acknowledgments: We would like to thank the University of Mississippi Medical Center’s Global Instruments Services for their assistance in mapping the geographic location of each patient based on current demographic data. The project would not have moved forward but for the tremendous insight and support on behalf of the Veterans Administration Benefits Office, The Office of the Navy, Centers for Disease Control and Prevention, and the US Public Health Service.
References 1. Wing S, Casper M, Davis WB, et al. Stroke mortality maps United States aged 35-74 years, 1962-1982. Stroke 1988; 19:1507-1513. 2. Lanska D. Geographic distribution of stroke mortality in the United States: 1939-41 to 1979-81. Neurology 1993; 43:1839-1851. 3. Howard G, Evans GW, Evans MS, et al. Is the stroke belt disappearing? An analysis of racial, temporal, and age effects. Stroke 1995;16:1153-1158. 4. Gillum RF, Ingram DD. Relation between residence in the Southeast region of the United States and stroke incidence: The NHANES I epidemiologic follow-up study. Am J Epidemiol 1996;144:665-673. 5. Howard G. Why do we have a stroke belt in the Southeastern United States? A review of unlikely and uninvestigated potential causes. Am J Med Sci 1999;317:160-167. 6. Perry HM, Roccella EJ. Conference report on stroke mortality in the Southeastern United States. Hypertension 1998;31:1206-1215. 7. Neito FJ. Infections and atherosclerosis: New clues form an old hypothesis? Am J Epidemiol 1998;148:937-948. 8. Shah PK. Link between infection and atherosclerosis. Who are the culprits: Viruses, bacteria, both, or neither [editorial]? Circulation 2001;103:5-7. 9. Edwards LB, Acquaviva FA, Livesay VT, et al. An atlas of sensitivity to tuberculin, PPD-B, and histoplasmin in the United States. Am Rev Resp Dis 1969;99(4 part 2):1-132. 10. Cook PJ, Honeybourne D, Lip GYH, et al. Chlamydia pneumoniae antibody titers are significantly associated with acute stroke and transient cerebral ischemia: The West Birmingham stroke project. Stroke 1998;29:404-410. 11. Elkind MSV, Lin I, Grayston JT, et al. Chlamydia pneumoniae and the risk of first ischemic stroke: The Northern Manhattan stroke study. Stroke 2000;31:1521-1525.
J.D. CLEARY ET AL. 12. Wimmer MLJ, Sandmann-Strupp R, Saikku P, et al. Association of chlamydial infection with cerebrovascular disease. Stroke 1996;27:2207-2210. 13. Mayr M, Keichl S, Willeit J, et al. Infections, immunity, and atherosclerosis: Associations of antibodies to Chlamydia pneumoniae, Helicobacter pylori, and Cytomegalovirus with immune reactions to heat-shock protein 60 and carotid or femoral atherosclerosis. Circulation 2000; 102:833-839. 14. D’Agostino RB Jr. Propensity score methods for bias reduction in the comparison of a treatment to a nonrandomized control group. Stat Med 1998;17:2265-2281. 15. Schulz DM. Histoplasmosis of the central nervous system. JAMA 1953;151:549. 16. Cooper RA, Goldstein E. Histoplasmosis of the central nervous system. Am J Med 1963;35:45-57. 17. Grossniklaus HE, Hutchinson AK, Capone A, et al. Clinicopathologic features of surgically excised choroidal neovascular membranes. Ophthalmology 1994;101(6): 1099-1111. 18. Karalakulasingam R, Arora KK, Adams G, et al. Meningoencephalitis caused by Histoplasma capsulatum. Arch Intern Med 1976;136:217-220. 19. Stone JH, Pomper MG, Hellmann DB. Histoplasmosis mimicking vasculitis of the central nervous system. J Rheumatol 1998;25:1644-1648. 20. Shapiro JL, Lux JJ, Sprofkin BE. Histoplasmosis of the central nervous system. Am J Pathol 1955;31:319-335. 21. McGregor JA, Kleinschmidt-DeMasters BK, Ogle J. Meningoencephalitis caused by Histoplasma capsulatum complicating pregnancy. Am J Obstet Gynecol 1985; 154:925-931. 22. Parsons RJ, Zarafonetis CJD. Histoplasmosis in man: Report of seven cases and a review of seventy-one cases. Arch Intern Med 1945;75:1-23. 23. Nieto FJ, Folsom AR, Sorlie PD, et al. Chlamydia pneumoniae infection and incident coronary heart disease: The atherosclerosis risk in communities study. Am J Epidemiol 1999;150:149-156. 24. Shor A. A pathologist’s view of organisms and human atherosclerosis. J Infect Dis 2001;183:1428-1429. 25. Kuo CC, Grayston JT, Campbell LA, et al. Chlamydia pneumoniae (TWAR) in coronary arteries of young adults (1534 years old). Proc Natl Acad Sci U S A 1995;92:6911-6914. 26. Yamashita K, Ouchi K, Shirai M, et al. Distribution of Chlamydia pneumoniae infection in the atherosclerotic carotid artery. Stroke 1998;29:773-778. 27. LaBiche R, Koziol D, Quinn T, et al. Presence of Chlamydia pneumoniae in human symptomatic and asymptomatic carotid atherosclerotic plaque. Stroke 2001;32:855-860. 28. Virok D, Kis Z, Karai L, et al. Chlamydia pneumoniae in atherosclerotic middle cerebral artery. Stroke 2001; 32:1973-1978. 29. Froberg MK, Seacotte N, Dahlberg E. Cytomegalovirus seropositivity and serum total cholesterol levels in young patients. Ann Clin Lab Sci 2001;31:157-161. 30. Ameriso SF, Fridman EA, Leiguarda RC, et al. Detection of Helicobacter pylori in human carotid atherosclerotic plaques. Stroke 2001;32:385-391. 31. Ridker PM, Hennekens CH, Stampfer MJ, et al. Prospective study of herpes simplex virus, Cytomegalovirus, and the risk of future myocardial infarction and stroke. Circulation 1998;98:2796-2799. 32. Kiechl S, Egger G, Mayr M, et al. Chronic infections and the risk of carotid atherosclerosis: Prospective results form a large population study. Circulation 2001; 103:1064-1070.
HISTOPLASMOSIS AND STROKE 33. Syrjanen J, Peltola J, Valtonen M, et al. Dental infections in association with cerebral infarction in young and middleaged men. J Intern Med 1989;225:179-184. 34. Xiaojung LI, Kolltveit KM, Tronstad L, et al. Systemic diseases caused by oral infection. Clin Microbiol Rev 2000; 13:547-558. 35. Loesche WJ, Lopatin DE. Interactions between periodontal disease, medical diseases and immunity in the older individual. Periodontal 2000;16:80-105. 36. Napoli MD, Papa F, Bocola V. C-reactive protein in ischemic stroke: An independent prognostic factor. Stroke 2001;32:917-924.
319 37. Sironi L, Tremoli E, Miller I, et al. Acute-phase proteins before cerebral ischemia in stroke-prone rats. Stroke 2001;32:753-760. 38. Hulthe J, Wikstrand J, Fagerberg B. Relationship between C-reactive protein and intima-media thickness in the carotid and femoral arteries and to antibodies against oxidized low-density lipoprotein in healthy me: The atherosclerosis and insulin resistance (AIR) study. Clin Sci (Colch) 2001;100:371-378. 39. Ami RB, Barshtein G, Zeltser D, et al. Parameters of red blood cell aggregation as correlates of the inflammatory state. Am J Physiol Heart Circ Physiol 2001;280: H1982-H1988.