elevated white blood cell count and carotid plaque...

Elevated White Blood Cell Count and CarotidPlaque Thickness

The Northern Manhattan Stroke Study

Mitchell S. Elkind, MS, MD; Jianfeng Cheng, MS, MD; Bernadette Boden-Albala, MPH;Myunghee C. Paik, PhD; Ralph L. Sacco, MS, MD

Background and Purpose—Elevated leukocyte count has been associated with cardiovascular and cerebrovascular diseasein several epidemiological studies. We sought to determine whether white blood cell count (WBC) is associated withcarotid plaque thickness in a stroke-free, multiethnic cohort.

Methods—For this cross-sectional analysis, WBC was measured in stroke-free community subjects undergoing carotidduplex Doppler ultrasound. Maximal internal carotid plaque thickness (MICPT) was measured for each subject.Demographic and potential medical confounding factors were analyzed with linear and logistic regression to calculatethe effect of quartile of WBC on MICPT. Odds ratios (ORs) and 95% confidence intervals (CIs) for the effect of quartileof WBC on MICPT$75th percentile were calculated. All analyses were stratified by race-ethnicity.

Results—The mean age of the 1422 subjects was 68.6610.2 years; 40.0% were men; 24.4% were white, 46.9% Hispanic,and 26.7% black. Among Hispanics, compared with the lowest quartile of WBC, those in the highest quartile hadsignificantly increased MICPT (mean difference50.30 mm, P50.0086) after adjustment for age, sex, and otheratherosclerotic risk factors. There was no significant increase for blacks or whites. The OR for MICPT$75th percentile(1.9 mm) was significantly increased for Hispanics (OR, 2.8; 95% CI, 1.4 to 5.6), marginally elevated for blacknon-Hispanics (OR, 1.6; 95% CI, 0.8 to 3.2), and not increased for white non-Hispanics (OR, 0.5; 95% CI, 0.2 to 1.1).

Conclusions—Relative elevation in WBC is associated with carotid atherosclerosis, but this relationship differs byrace-ethnicity. The association is strongest in Hispanics, intermediate in black non-Hispanics, and not present in whitenon-Hispanics in this population. Chronic subclinical infection or inflammation may account for this association.(Stroke. 2001;32:842-849.)

Key Words: atherosclerosisn cerebrovascular disordersn epidemiologyn risk factors

K nown risk factors fail to account for all cases ofischemic stroke and cardiovascular disease. Recent ev-

idence suggests that atherosclerosis is an inflammatory dis-ease.1 Chronic infection withChlamydia pneumoniae,Heli-cobacter pylori, cytomegalovirus, and other organisms hasbeen postulated as a potential risk factor for atherosclerosis,heart disease, and stroke.2–6 Recent data have also suggestedthat carotid duplex Doppler ultrasound is a useful way tostudy atherosclerotic risk factors, because asymptomatic ca-rotid wall thickening and plaque formation may be a precur-sor to clinical vascular events. Several studies, includingearlier work from our own laboratory,7,8 have shown thatmaximal internal carotid plaque thickness (MICPT) is asso-ciated with vascular risk factors, such as diabetes mellitus,hypertension, hypercholesterolemia, and smoking. We soughtto determine whether white blood cell count (WBC) is

associated with MICPT in a cross-sectional analysis of astroke-free, elderly, multiethnic urban population.

Subjects and MethodsThe Northern Manhattan Stroke Study (NOMASS) includes anongoing population-based cohort study designed to determine strokeincidence, risk factors, and prognosis in a multiethnic, urban popu-lation. Northern Manhattan consists of the area north of 145th Streetand south of 218th Street that is bordered on the west by the HudsonRiver and on the east by the Harlem River. In 1990,'260 000people lived in the community, with 40% older than 39 years of ageand a race-ethnic mixture consisting of 20% black, 63% Hispanic,and 15% white residents.9

Selection of NOMASS CohortThe methods of subject recruitment and enrollment have beendescribed in previous publications.7,10 Briefly, random digit dialingof '29 000 households was performed by Audits and Surveys, Inc.

Received July 28, 2000; final revision received November 16, 2000; accepted December 6, 2000.From the Department of Neurology (M.S.E., R.L.S.) and Sergievsky Center (M.S.E., B.B.-A., R.L.S.), College of Physicians and Surgeons, Columbia

University, New York, NY; Columbia-Presbyterian Medical Center of New York Presbyterian Hospital, New York, NY (M.S.E., R.L.S.); and Divisionsof Biostatistics (J.C., M.C.P.), Sociomedical Sciences (B.B.-A.), and Epidemiology (R.L.S.), Joseph P. Mailman School of Public Health, ColumbiaUniversity, New York, NY.

Presented in abstract form at the American Heart Association Stroke Conference, New Orleans, La, February 10, 2000.Correspondence to Mitchell S. Elkind, MD, Neurological Institute, 710 W 168th St, New York, NY 10032. E-mail [email protected]© 2001 American Heart Association, Inc.

Stroke is available at http://www.strokeaha.org

842

by guest on May 17, 2018

http://stroke.ahajournals.org/D

ownloaded from

http://stroke.ahajournals.org/

Community participants were enrolled if they (1) had never beendiagnosed with stroke, (2) were.40 years of age, and (3) resided inNorthern Manhattan for$3 months in a household with a telephone.In-person evaluations were performed at the hospital; those subjectswho were not able to come to the hospital did not undergo Dopplerimaging and were not included in this analysis. The telephoneresponse rate was 94%, and 70% of those respondents participated inan in-person evaluation. The study was approved by the InstitutionalReview Board at Columbia-Presbyterian Medical Center. All partic-ipants gave consent directly or through a surrogate when appropriate.

Index Evaluation of SubjectsData were collected through interviews, in-person measurements,and collection of fasting blood specimens for lipid and glucosemeasurements by trained research assistants, and physical andneurological examinations were done by the study physicians, asdescribed elsewhere.10 When possible, data were obtained directlyfrom subjects with the standardized data collection instruments.When the subject was unable to provide answers, a proxy knowl-edgeable about the subject’s history was interviewed. Direct subjectdata were obtained from 99% of stroke-free subjects.

Assessments were conducted in English or Spanish, depending onthe primary language of the participant. Race-ethnicity was based onself-identification through a series of interview questions modeledafter the US Census and conforming to the standard definitionsoutlined by Directive 15.11 All participants responding affirmativelyto being of Spanish origin or identifying themselves as Hispanicwere classified as such. All participants classifying themselves aswhite without any Hispanic origin or as black without any Hispanicorigin were classified as white, non-Hispanic or black, non-Hispanic,respectively.

Standardized questions were adapted from the Behavioral RiskFactor Surveillance System12 by the Centers for Disease Control andPrevention regarding the following conditions: hypertension, diabe-tes, hypercholesterolemia, peripheral vascular disease, transient is-chemic attack, cigarette smoking, and cardiac conditions, such asmyocardial infarction, coronary artery disease, angina, congestiveheart failure, atrial fibrillation, other arrhythmias, and valvular heartdisease. Standard techniques were used to measure blood pressure,height, weight, and fasting glucose as described in prior publica-tions.13 Fasting lipid panels (including total cholesterol, low-densitylipoprotein [LDL], high-density lipoprotein, and triglycerides) weremeasured with a Hitachi 705 automated spectrometer (Boehringer).Hypertension was defined as in prior publications,14 and diabetesmellitus was defined as a fasting blood glucose level$127 mg/dL,the subject’s self-report of such a history, or insulin or oralhypoglycemic use. The definitions are noted in the table footnotes.

Assessment of WBCWBCs were measured with automated cell counters via standardtechniques (Coulter STK-R and Coulter STK-S, Coulter Electronics,and Sysmex SE-9500, TOA Medical Electronics). Whole blood wascollected in 5-cm3 EDTA-anticoagulated tubes by a trained phlebot-omist. The automated cell counter aspirated a sample from thecollection tube, and after lysis of red blood cells and platelets, WBCswere counted by use of a standard direct current detection method.Normal values for WBC in the hematology laboratory are 3.54 to9.063109/L. Quality control is maintained by the laboratory withstandard procedures. The coefficient of variation for repeated mea-surements on samples from individual hospitalized patients is main-tained at#2.5%.

Assessment of MICPTThe method for assessment of MICPT has been described in aprevious publication.7 Briefly, MICPT was assessed by trainedultrasonographers experienced in the use of duplex ultrasonographyfor research purposes and blinded to the participant’s risk factors.Ultrasonography was performed on a Siemens Quantum 2000 duplexultrasound system with 7.5-MHz scanning frequency in B mode and5.0-MHz frequency in pulsed Doppler mode. With the subject lying

in a supine position, the extracranial carotid arteries were imaged inthe longitudinal (anterior, lateral, and posterior views) and transverseplanes. Both internal carotid arteries were examined for the presenceof atherosclerotic plaque, defined as an area of focal hyperechoicwall thickening. If no atherosclerosis was identified, MICPT wasrecorded as zero. If plaque was imaged, the view showing thethickest plaque was frozen, and the intimal-medial wall thickness(including the plaque) was measured with an electronic cursor andrecorded as the MICPT for that artery. For this analysis, the greaterof the right and left MICPT was used.

Statistical AnalysesMeans were calculated for continuous variables, and proportionswere calculated for categorical variables. Bonferroni-correctedt testswere conducted for comparisons of means;x2 tests, for comparisonsof proportions. Regression analysis using linear and quadratic termswas first performed with WBC as a continuous independent variableand MICPT as the dependent variable, stratified by each of the 3race-ethnic subgroups. Residual plots were examined, and severalsubsequent analyses were performed, excluding influential points toestablish the robustness of the associations. Subjects were thendivided into 4 quartiles defined by WBC. Multivariate linear regres-sion using a nonautomated procedure, incorporating demographicand clinical variables, was then used to build models for theassociation of WBC and MICPT, stratified by race-ethnicity. Thedependent variable for these analyses, MICPT, was expressed as acontinuous variable, despite a skewed distribution, because of thelarge sample size and stable variance. Conventional atheroscleroticrisk factors were chosen for the final model on the basis of anassociation with MICPT in univariate analysis or findings of asignificant association in previous analyses from our laboratory.Analyses were then conducted to determine the effect of quartile ofWBC on MICPT as a dichotomous variable using multivariatelogistic regression, with an MICPT$75th percentile (1.9 mm) as thecutoff. Additional logistic regression analyses were then performedwith subjects stratified by sex to analyze differences in the associa-tions of MICPT and plaque thickness among the different popula-tions. Statistical significance was determined at thea50.05 levelwith 2-sided tests. Statistical analyses were conducted with SAScomputer software (SAS Institute).

ResultsThe mean age of the 1422 participants was 68.6610.2 years(median, 69 years; range, 40 to 99 years). Forty percent(n5569) were men; 24.4% (n5347) of the participants werewhite non-Hispanic, 26.7% (n5379) black non-Hispanic,46.9% (n5667) Hispanic, and 2.0% (n529) “other.” Whitenon-Hispanics (mean age, 72.769.6 years) were older thanthe black non-Hispanics (mean age, 70.3610.3 years) andHispanics (mean age, 65.669.4 years). The distribution ofsociodemographic factors, comorbid vascular diseases, andconventional atherosclerotic risk factors is shown in Table 1.

The mean WBC for the entire cohort was6.2962.013109/L (median, 6.03109/L; interquartile range,4.9 to 7.33109/L; range, 1.5 to 25.73109/L; Table 2). Inunivariate linear regression models stratified by race-ethnicity, WBC considered as a continuous variable wassignificantly associated with MICPT among Hispanics(b50.098, P,0.001) but not among blacks or whites(b50.03, P50.33 andb50.03, P50.35, respectively). At-tempts to fit quadratic models showed no improvement in themodels. Analysis after deletion of possible influential pointsdid not change the main results. Further analyses wereconducted with WBC quartiles as the independent variable,with those in the lowest quartile as the reference group (WBC,4.93109/L). There were differences between the different

Elkind et al Elevated WBC Count and Carotid Plaque Thickness 843



ownloaded from


race-ethnic groups in WBC (Table 2). Black non-Hispanicshad the lowest mean WBC (P,0.0001).

The mean MICPT for the entire cohort was 1.1461.23 mm(median, 1.0 mm; interquartile range, 0 to 1.9 mm). Therewere also significant differences in MICPT among the 3major race-ethnic groups (Table 2). MICPT among Hispanicswas significantly less than that among non-Hispanics, butthere was no significant difference between white and blacknon-Hispanics. Among Hispanics, the mean MICPT values ineach of the 4 quartiles of WBC were 0.72, 0.75, 0.74, and1.19 mm. The difference between mean MICPT in the highestand the other 3 quartiles was significant (P,0.05). There wasno difference between mean MICPT by quartile of WBCamong the other race-ethnic subgroups.

In linear regression models including quartile of WBC onlyand stratified by race-ethnicity, among Hispanics those in thehighest quartile of WBC had a significantly increased MICPTcompared with those in the lowest quartile (mean differ-ence50.47,P,0.0001; Table 3). For the second and thirdquartiles, there was no significant association (Table 3). Afteradjustment for demographic variables (age and sex), thereremained an association for Hispanics (b50.41, P50.004)but not for black or white non-Hispanics (Table 3). Aftermultiple linear regression analysis adjusting for potential

confounders, a statistically significant association for thehighest quartile of WBC and MICPT remained among His-panics, although inclusion of the covariates attenuated therelationship (b50.30, P50.0086). Thus, among Hispanics,on average those in the highest quartile for WBC had anMICPT that was 0.3 mm thicker than those in the lowestquartile. The final model also included the following covari-ates, which were significantly associated with MICPT amongHispanics: age from 65 to 80 years (b50.41,P,0.0001), age.80 years (b51.03, P,0.0001), male sex (b50.19,P50.02), LDL (b50.0038,P50.0003), coronary artery dis-ease (b50.37, P50.0005), and current cigarette smoking(b50.46, P50.0002). Current smoking was highly signifi-cantly associated with WBC level among Hispanics, withmean WBC among current smokers and current nonsmokersof 7.56 and 6.233109/L (P,0.0001), respectively. Inclusionof smoking in the model, however, did not eliminate theeffect of WBC on MICPT but did reduce the magnitude of theeffect (b50.40, P50.0004 in model without smoking;b50.30, P50.0086 in the model with smoking). Hyperten-sion and diabetes mellitus were also included in the finalmodel because of their recognized importance as risk factorsfor atherosclerosis and because they were associated withMICPT in the black or white populations but not convention-ally significant among Hispanics.

Among white non-Hispanics, further models revealed nosignificant association of WBC quartile with plaque thickness(Table 3). Age from 65 to 80 years (b50.39,P50.02), age.80 years (b51.00, P,0.0001), hypertension (b50.46,P50.0004), current cigarette smoking (b50.53, P50.007),and coronary artery disease (b50.37,P50.01) were signifi-cantly associated with plaque thickness, but sex, LDL, anddiabetes mellitus were not. Similarly, among black non-Hispanics, there was no association of WBC with plaquethickness (Table 3). Age from 65 to 80 years (b50.77,P,0.0001), age.80 years (b51.41, P,0.0001), diabetesmellitus (b50.51,P50.001), and current smoking (b50.75,P5,0.0001) were associated with plaque thickness amongblack non-Hispanics. Because blacks are known to have alower WBC than white populations15 (a condition known asbenign leukopenia), additional analyses were performed withthe use of within-group quartiles for blacks, but this producedno material change in the results.

TABLE 2. Distribution of WBC and MICPT by Race-Ethnicity

Hispanic(n5667)

White Non-Hispanic(n5347)

Black Non-Hispanic(n5379)

Overall*(n51422)

WBC, 3109/L

Mean6SD 6.4061.97 6.5362.01 5.8862.07 6.2561.92

Median 6.3 6.1 5.5 6.0

Interquartile range 5.1–7.6 5.1–7.3 4.5–6.9 4.9–7.3

MICPT, mm

Mean6SD 0.8761.09 1.4061.27 1.3861.32 1.1461.23

Median MICPT 0.0 1.3 1.2 1.0

Interquartile range 0.0–1.5 0.0–2.0 0.0–2.2 0.0–1.9

*Twenty-nine subjects were characterized as other race-ethnicity.

TABLE 1. Characteristics of Participants

nPrevalence, %or mean6SD

Completed high school 713 50.1

Hypertension 820 57.8

Diabetes mellitus 309 21.8

Cardiac disease 282 19.9

Current smoking 213 15.0

Ever smoked 775 54.6

Total cholesterol, mg/dL 1378 205640.9

HDL, mg/dL 1374 47.3622.6

LDL, mg/dL 1372 130.6637.3

HDL indicates high-density lipoprotein. Hypertension was defined as asystolic blood pressure recording of $160 mm Hg or a diastolic blood pressurerecording of $95 mm Hg or the patient’s self-report of a history of hyperten-sion or antihypertensive use. Diabetes mellitus was defined by a fasting bloodglucose level .126 mg/dL, the patient’s self-report of such a history, or insulinor hypoglycemic use.

844 Stroke April 2001



ownloaded from


Further analyses used a dichotomous outcome variable,MICPT $75th percentile (1.9 mm), to assess the effect ofWBC quartile on marked thickening of the vessel wall (Table4). Among Hispanics, compared with those in the lowestquartile of WBC, those in the highest quartile had a signifi-cantly elevated risk of MICPT$1.9 mm after adjustment fordemographic and other risk factors. There was no significantincrease for those in the intermediate quartiles. Among blacknon-Hispanics, a significant increase in risk was found in amodel adjusted for demographic factors for both the third andfourth quartiles, but this was attenuated after adjustment forthe other atherosclerotic risk factors. Among white non-Hispanics, there was no evidence of an association.

We further stratified our Hispanic population by sex todetermine whether the association of WBC with MICPT heldfor both men and women (Table 5). An increase in the oddsfor MICPT $1.9 mm was found for both men and women,

although the magnitude of this risk was greater for men thanwomen (odds ratio [OR], 3.29; 95% confidence interval [CI],1.24 to 9.58; and OR, 2.61; 95% CI, 1.02 to 7.41, respective-ly). A formal test for an interaction between sex and WBC didnot reveal any significant effect.

DiscussionThis cross-sectional study supports a previously describedassociation between WBC and subclinical carotid plaqueformation and suggests that the association may varyamong particular race-ethnic groups. We found an associ-ation of WBC and MICPT in an elderly, urban, mostlyHispanic population in whom the burden of stroke andother vascular diseases is high. There also may be variationby sex, because the relationship appears stronger amongmen than women.

TABLE 3. Predicted Mean Differences and 95% CIs in MICPT by Quartile of WBC Stratified by Race andAdjusted for Demographic and Other Risk Factors

Quartile of WBC

Hispanic Non-Hispanic White Non-Hispanic Black

b 95% CI b 95% CI b 95% CI

Adjusted for demographic factors* (n5667) (n5347) (n5379)

Quartile 1† z z z z z z z z z z z z z z z z z z

Quartile 2 0.05 20.18–0.29 20.28 20.66–0.10 0.06 20.28–0.40

Quartile 3 0.03 20.19–0.26 0.03 20.34–0.41 0.36 0.01–0.70

Quartile 4 0.42 0.20–0.65 20.08 20.46–0.29 0.24 20.10–0.58

Adjusted for above andconventional risk factors‡

(n5640) (n5335) (n5363)

Quartile 1† z z z z z z z z z z z z z z z z z z

Quartile 2 0.06 20.17–0.29 20.32 20.69–0.05 0.01 20.32–0.35

Quartile 3 20.05 20.27–0.17 20.06 20.43–0.30 0.24 20.09–0.58

Quartile 4 0.30 0.08–0.53 20.24 20.62–0.15 20.04 20.39–0.32

*Demographic factors are age and sex.†First quartile is the referent group.‡Conventional risk factors are LDL, hypertension, diabetes mellitus, coronary artery disease, and current cigarette smoking.

TABLE 4. ORs for Association of Quartile of WBC With MICPT >1.9 mm

Quartile of WBC

Hispanic Non-Hispanic White Non-Hispanic Black

OR 95% CI OR 95% CI OR 95% CI

Adjusted for demographicfactors*

(n5667) (n5347) (n5379)

Quartile 1† 1.00 z z z 1.00 z z z 1.00 z z z

Quartile 2 1.61 0.80–3.30 0.58 0.29–1.17 1.30 0.69–2.43

Quartile 3 1.20 0.59–2.46 0.84 0.43–1.64 1.95 1.05–3.64

Quartile 4 3.08 1.67–5.96 0.67 0.34–1.34 2.07 1.10–3.91

Adjusted for above andconventional risk factors‡

(n5640) (n5335) (n5363)

Quartile 1† 1.00 z z z 1.00 z z z 1.00 z z z

Quartile 2 1.80 0.87–3.81 0.55 0.26–1.17 1.22 0.62–2.40

Quartile 3 1.04 0.49–2.23 0.75 0.37–1.53 1.76 0.91–3.43

Quartile 4 2.77 1.44–5.59 0.51 0.24–1.10 1.55 0.77–3.15

*Demographic factors are age and sex.†First quartile is the referent group.‡Conventional risk factors are LDL, hypertension, diabetes mellitus, coronary artery disease, and current cigarette smoking.




ownloaded from


Previous studies of atherosclerotic risk factors using high-resolution carotid duplex Doppler ultrasound have examinedWBC among other hemostatic and infectious risk factors.One study16 found WBC to be independently predictive ofprogression of atherosclerosis over 2 years in a small sampleof Finnish men. Another study17 that investigated severalhematologic factors did not find an association, but that studyexamined patients with established arterial diseases. Studiesof clinical atherosclerotic outcome events have found anassociation between leukocyte count and risk of atheroscle-rotic heart disease18–23 and stroke.24,25 Our study providesadditional evidence that in certain populations elevations inWBC may be independently associated with a marker ofsubclinical atherosclerosis.

In the population examined in this study, the association ofleukocytes with atherosclerosis differed by race-ethnicity.Among Hispanics, there appeared to be a threshold for therelationship of WBC with MICPT, with the increase inMICPT occurring only in the highest quartile of WBC. It isimportant to recognize, however, that even the highest quar-tile, WBC $7.33109/L, still fell mostly within the normalrange of WBC. Among black non-Hispanics, there was atrend toward an increased MICPT in those with WBC in thethird and fourth quartiles, but this association was notstatistically significant after adjustment for other risk factors.The number of black non-Hispanics was small relative to thenumber of Hispanics, however. Among white non-Hispanics,our smallest group, there was no significant association ofWBC and MICPT. There was also evidence of some sex-related heterogeneity in the association of WBC and MICPT.Among Hispanics, the relationship was slightly strongeramong men than women. Formal tests for an interactionbetween WBC and sex were negative, however.

We7 and others26 have found other differences in theprevalence of carotid wall thickening between Hispanics andnon-Hispanics. Hispanics have less MICPT, even after ad-

justment for age and other conventional risk factors, than donon-Hispanic whites and blacks in our population. Thedifferent distribution of MICPT among Hispanics comparedwith non-Hispanic groups in our population does not explainthe difference in the association of WBC with MICPT,however. Our analyses were stratified by race-ethnicity, sothe findings within each group represent true findings.Whether an association among WBC and MICPT would befound among white or black populations with differentdistributions of MICPT is unclear from these data. Moreover,using group-specific thresholds for the quartile values ofWBC and the MICPT cutoff for the dichotomous analysis didnot change the results.

Other studies have found an association between WBC andcardiovascular disease among whites, but ours did not. Ourwhite population may differ from the white populationsexamined in other studies, however. Among the studies thatfound WBC to be a predictor of clinical cardiovascularevents, such as the Framingham, Multiple Risk Factor Inter-vention, Caerphilly, and Speedwell studies, the participantswere generally,60 years of age.21–23 The mean age of ourwhite population, however, was 72.7 years compared withmean ages among our Hispanics of 65.6 years and among ourblack population of 70.3 years. In the intima-media thicknessprogression study of men from Finland,16 participants rangedfrom 42 to 60 years of age. In that population, moreover,hypertension was not found to be associated with IMTprogression, whereas in our study, hypertension was associ-ated with MICPT among whites. In studies that looked at bothmiddle-aged and older participants,20 the effect of WBC wasstronger in those,65 years of age. In addition, the number ofparticipants in the white non-Hispanic group was relativelysmall (n5379) compared with the black non-Hispanic andHispanic subgroups, making the results less robust for thatgroup.

The effect of risk factors other than WBC on atheroscle-rosis may also differ between Hispanics and non-Hispanics.

TABLE 5. ORs for Association of Quartile of WBC with MICPT >1.9 mm AmongHispanic Men and Women

Quartile of WBC

Hispanics

Men Women

OR 95% CI OR 95% CI

Adjusted for age (n5275) (n5392)

Quartile 1* 1.00 z z z 1.00 z z z

Quartile 2 1.78 0.68–4.92 1.43 0.50–4.28

Quartile 3 0.91 0.33–2.58 1.48 0.54–4.32

Quartile 4 3.71 1.58–9.55 2.50 1.02–6.83

Adjusted for age andconventional risk factors†

(n5267) (n5373)

Quartile 1* 1.00 z z z 1.00 z z z

Quartile 2 1.97 0.70–5.92 1.59 0.55–4.87

Quartile 3 0.78 0.25–2.43 1.37 0.48–4.15

Quartile 4 3.29 1.24–9.58 2.61 1.02–7.41

*First quartile is the referent group.†Conventional risk factors are LDL, hypertension, diabetes mellitus, coronary artery disease, and

current cigarette smoking.




ownloaded from


LDL7 and the ratio of apolipoprotein (apo)B to apoA-I8 arealso associated with increasing MICPT in Hispanics but notin the other race-ethnic groups in our population. Otherstudies have found that the effect of cardiovascular riskfactors may differ between blacks and whites.27,28 The ma-jority (88%) of our Hispanic population is composed ofCaribbean Hispanics from the Dominican Republic, PuertoRico, or Cuba. Risk factors in this group could also differfrom those in other Hispanic groups in the United States, suchas Mexicans and those from other countries. However,although Caribbean Hispanics may differ from MexicanHispanics, it is also possible that Hispanics from thesedifferent Caribbean countries also differ from one another inimportant ways. It should be emphasized that AmericanHispanics are a heterogeneous group, that the classification ofindividuals as “Hispanic” can be problematic,29 and thatanalyses based on any assumptions of homogeneity should beviewed cautiously. Nonetheless, we used the method ofself-identification to determine race-ethnicity, which is sub-scribed to by the US Census and recommended for allresearch studies.11 Moreover, if the term “Hispanic” is insuf-ficiently accurate, nondifferential misclassification by race-ethnicity should only serve to underestimate the effects of anyfindings.

The findings that Hispanics in our population have lesscarotid plaque and that their atherosclerosis may be associ-ated with different risk factors (LDL, WBC) than theirnon-Hispanic neighbors may have important implications.The recognition that atherosclerotic risk factors may havespecificity for certain race-ethnic populations may be of helpin targeting particular risk factor reduction strategies to thosepopulations. The underlying cause of this race-ethnic vari-ability remains uncertain. Genetic, dietary, and other envi-ronmental factors may all play a role. One of the benefits ofthe NOMASS study design is that it allows the study of 3different race-ethnic groups living in the same relativelylimited geographic area within northern Manhattan andthereby minimizes some of the environmental heterogeneitythat comes from studying populations from different regionsof the country. There may still be important cultural differ-ences among the different race-ethnic groups, however, thatlead to different effective environments. We did not includedietary data, for example, in this analysis.

The magnitude of the effect of elevated WBC on MICPTamong Hispanics in our population is probably clinicallymeaningful. Those in the highest quartile of WBC had anadjusted increase in plaque thickness of 0.3 mm comparedwith those in the lowest quartile. In the Cardiovascular HealthStudy (CHS) in which measurements were made with aslightly different technique, for every 0.55-mm increase inintima-media thickness of the internal carotid artery, therewas a 30% increase in risk of stroke or myocardial infarc-tion.30 For those with an intima-media thickness of$1.81 mm in CHS, the relative risk of stroke or myocardialinfarction was 2.47 (95% CI, 1.59 to 3.85). The cutoff of1.81 mm in CHS represented the 80th percentile and is verysimilar to the 75th percentile threshold of MICPT$1.9 mmin our dichotomous model.

Elevated leukocytes may lead to clinical atheroscleroticevents either by an effect on chronic atherosclerosis or byinducing acute thrombotic events. Prospective observationalstudies have found elevations in WBC to be predictive ofischemic stroke and coronary artery disease, perhaps by aneffect on plaque rupture.18,24Data from experimental work onanimals and in vitro data show that leukocytes also play animportant role in atherogenesis.1 Macrophages and T lym-phocytes are prominent in human atheromas, even in theearliest stages of the disease process,31 suggesting that im-mune processes also may play an initiating or early role in thedevelopment of the lesion in human beings. Our data supportat least a partial role for leukocytes in the chronic process ofatherosclerosis.

It is not known whether this association between relativelyelevated leukocyte count and atherosclerosis reflects ongoingchronic subclinical infection. Several observational epidemi-ological studies,5,6,32 including one in our own population,33

have suggested an association between chronic infection withC pneumoniaeand stroke risk. The Atherosclerosis Risk inCommunities (ARIC) Study Investigators4 reported a similarmagnitude of association ofC pneumoniaeIgG antibodieswith asymptomatic carotid atherosclerosis among patients 45to 64 years of age (adjusted OR, 2.0; 95% CI, 1.2 to 3.4).Other ways in which leukocytes could affect atherosclerosisinclude inflammatory mechanisms independent of infection.

Smoking is likely to be a partial confounder in therelationship between WBC and atherosclerosis. We andothers have found that current cigarette smoking is associatedwith WBC34 and subclinical atherosclerosis.16,35Smoking canincrease the mortality from chronic bronchitis36 and possiblyother infections.37 Smoking itself may thus be the cause of theincreased MICPT, with elevated WBC also resulting from thesmoking. We found, however, that there is an increase inMICPT even after adjustment for current smoking amongHispanics. Residual confounding or differential reporting ofsmoking history dependent on WBC status could mask theeffect of smoking on MICPT even among the Hispanicpopulation. Other risk factors are also associated with ele-vated WBC,35 and could confound the relationship betweenMICPT and WBC.

Further prospective studies of the relationship betweenWBC, as well as other inflammatory and infectious markers,and ischemic stroke are needed. Although many studies haveinvestigated the relationship between infection and athero-sclerotic heart disease, these may not reflect the relationshipbetween infection and stroke. In northern Manhattan, large-artery atherosclerosis accounts for a minority (only 10% to20%) of ischemic stroke38; embolic and small-vessel causesof stroke are probably more common. Further studies need totake into account the several etiological subtypes of stroke.

Our study has several limitations. Because of our cross-sectional design, we are unable to claim that elevated WBCleads to an increase in plaque thickness. The converse, thatparticipants with greater carotid plaque thickness developelevated WBC, could just as well be true. The participants inthis study were all stroke-free subjects from the communityselected at random; thus, it is unlikely that there would be aselection bias for a high incidence of infections. Our study




ownloaded from


also assesses a measure of subclinical atherosclerosis,MICPT, rather than clinical end points such as myocardialinfarction or stroke. Several recent studies, however, haveshown that these measures are predictive of clinical ischemicevents.30,39,40

Our measurement of MICPT also differs from that used inother studies of carotid wall thickness. For these analyses, weused the maximum thickness of the internal carotid arteryplaque. This measurement differs from the measurement ofintima-media thickness made in several other stud-ies.4,16,30,40–42We have previously reported on the use of thismethod and found associations with atherosclerotic riskfactors, including age, smoking, hyperglycemia, hyperten-sion, LDL, apoA-I and apoB, consistent with findings instudies using methods which measure intima-media thick-ness.7,8 Additionally, some30,41,42 have suggested that thecommon carotid artery wall thickness is the most reliablemeasurement for studies of preclinical studies of atheroscle-rosis. Others have shown, however, that internal carotidartery measurements39 or combined measurements of internaland common carotid arteries43 are as reliable and useful inpredicting clinical events as are those of the common carotidartery alone. We are currently measuring intima-media thick-ness at several sites in the common and internal carotidarteries and in the future may have further data on theassociation of WBC and other risk factors and intima-mediathickness, as well as on the relative merits of intima-mediathickness and MICPT.

Our study did not provide serological data on infection ordata on other markers of inflammation, such as C-reactiveprotein or leukocyte adhesion molecules. Further studies areongoing to determine the association of specific cytokinemarkers with carotid atherosclerosis in this population. Wealso did not have data on clinical infection and therefore wereunable to make statements about the underlying causes of theelevated WBC. Although this information would be useful, itis difficult to know how to apply it to the development ofatherosclerosis. In asymptomatic populations, the role ofelevated WBC may be more chronic. Also, it would be usefulto know whether the measurements we made of WBC arestable over time. The effect of such measurement bias shouldbe attenuated, however, by the large number of participantsenrolled.

It would also be helpful to have data on the leukocytedifferentials of the WBC of the subjects. It is unclear whetherthe elevated risk of MICPT is associated primarily withneutrophils or lymphocytes. Recent evidence suggests thatcirculating levels of specific white cell types, such as theCD41CD28null subset of T lymphocytes, may be associatedwith unstable angina.44 It remains unknown whether circulat-ing levels of specific WBC types might be associated withsubclinical atherosclerosis.

In summary, our study supports an association betweenWBC and carotid atherosclerosis in at least some populations.Evidence from pilot clinical trials in patients with coronaryartery disease,45–47 as well as animal studies,48 alreadysuggests that the risk of atherosclerotic disease associatedwith certain infections, such asC pneumoniae, may bemodifiable. Corroboration from larger, prospective studies of

the role of inflammatory and infectious markers in athero-sclerosis and stroke might lead to clinical trials with novelanti-inflammatory or anti-infectious therapies to retard ath-erosclerosis or prevent incident and recurrent stroke.

AcknowledgmentsThis work was supported by grants from the Columbia UniversityOffice of Clinical Trials (pilot grant, Dr Elkind), National Institute ofNeurological Disorders and Stroke (R01 NS 27517 and R01 NS29993, Dr Sacco; T32 NS 07153, Dr Elkind), the General ClinicalResearch Center (2 M01 RR00645), an American Heart AssociationScientist Development Grant (Dr Elkind), and Centers for DiseaseControl cooperative agreement (Dr Elkind). We acknowledge thesupport of Dr J.P. Mohr, Director of Cerebrovascular Research, andthe technical support of Drs Sam Trocio, Oscar Ramos, Roque Sia,and Romel Ramas, ultrasonographers in the NeurovascularLaboratory.

References1. Ross R. Atherosclerosis: an inflammatory disease.N Engl J Med. 1999;

340:115–126.2. Saikku P, Leinonen M, Mattila K, Ekman M-R, Nieminen MS, Mäkelä

PH, Huttunen JK, Valtonen V. Serologic evidence of an association of anovel Chlamydia, TWAR, with chronic coronary heart disease and acutemyocardial infarction.Lancet. 1988;2:983–985.

3. Thom DH, Grayston JT, Siscovick DS, Wang SP, Weiss NS, Daling JR.Association of prior infection withChlamydia pneumoniaeand angio-graphically demonstrated coronary artery disease.JAMA. 1992;268:68–72.

4. Melnick SL, Shahar E, Folsom AR, Grayston JT, Sorlie PD, Wang SP,Szklo M. Past infection byChlamydia pneumoniaestrain TWAR andasymptomatic carotid atherosclerosis.Am J Med. 1993;95:499–504.

5. Wimmer MLJ, Sandmann-Strupp R, Saikku P, Haberl RL. Association ofchlamydial infection with cerebrovascular disease.Stroke. 1996;27:2207–2210.

6. Cook PJ, Honeybourne D, Lip GYH, Beevers DG, Wise R, Davies P.Chlamydia pneumoniaeantibody titers are significantly associated withacute stroke and transient cerebral ischemia.Stroke. 1998;29:404–410.

7. Sacco RL, Roberts JK, Boden-Albala B, Gu Q, Lin I-F, Kargman DE,Berglund L, Hauser WA, Shea S, Paik MC. Race-ethnicity and deter-minants of carotid atherosclerosis in a multiethnic population.Stroke.1997;28:929–935.

8. Prisant LM, Zemel PC, Nichols FT, Zemel MB, Sowers JF, Carr AA,Thompson WO, Bond MG. Carotid plaque associations among hyper-tensive patients.Arch Intern Med. 1993;153:501-506.

9. United States Census of Population and Housing. Public use microdatasample. Available at www.census.gov. 1990.

10. Sacco RL, Gan R, Boden-Albala B, Lin I-F, Kargman DE, Hauser WA,Shea S, Paik MC. Leisure-time physical activity and ischemic stroke risk:The Northern Manhattan Stroke Study.Stroke. 1998;29:380–387.

11. Office of Management and Budget. Race and ethnic standards for federalstatistics and administrative reporting.Federal Register. May 4, 1978;43:19269. Directive No. 15.

12. Gentry EM, Kalsbeek WD, Hegelin GC, Jones JT, Gaines KL, FormanMR, Marks JS, Trowbridge FL. The Behavioral Risk Factor Surveys, II:design, methods, and estimates from combined state data.Am J Prev Med.1985;1:9–14.

13. Sacco RL, Roberts JK, Boden-Albala B, Gu Q, Lin IF, Kargman DE,Berglund L, Hauser WA, Shea S, Paik MC. Race-ethnicity and deter-minants of carotid atherosclerosis in a multi-ethnic population: theNorthern Manhattan Stroke Study.Stroke. 1997;28:929–935.

14. Sacco RL, Elkind M, Boden-Albala B, Lin I-F, Kargman DE, HauserWA, Shea S, Paik MC. The protective effect of moderate alcohol con-sumption on ischemic stroke.JAMA. 1999;281:53–60.

15. Reed WW, Diehl LF. Leukopenia, neutropenia, and reduced hemoglobinlevels in healthy American blacks.Arch Intern Med. 1991;151:501–505.

16. Salonen R, Salonen JT. Progression of carotid atherosclerosis and itsdeterminants: a population-based ultrasonography study.Atherosclerosis.1990;81:33–40.

17. Cortellaro M, Baldessare D, Cofrancesco E, Tremoli E, Colombo A,Boschetti C, Paoletti R. Relation between hemostatic variables and




ownloaded from


increase of common carotid intima-media thickness in patients withperipheral arterial disease.Stroke. 1996;27:450–454.

18. Friedman GD, Klatsky AL, Siegelaub AB. The leukocyte count as apredictor of myocardial infarction.N Engl J Med. 1974;290:1275–1278.

19. Zalokar JB, Richard JL, Claude JR. Leukocyte count, smoking, andmyocardial infarction.N Engl J Med. 1981;304:465–468.

20. Prentice RL, Szatrowski T, Fujikura T, Kato H, Mason MW, HamiltonHH. Leukocyte counts and coronary heart disease in a Japanese cohort.Am J Epidemiol. 1982;116:496–509.

21. Grimm RH, Neaton JD, Ludwig W. Prognostic importance of the whiteblood cell count for coronary, cancer, and all-cause mortality.JAMA.1985;254:1932–1937.

22. Yarnell JWG, Baker IA, Sweetnam PM, Bainton D, O’Brien JR,Whitehead PJ, Elwood PC. Fibrinogen, viscosity, and white blood cellcount are major risk factors for ischemic heart disease.Circulation.1991;83:836–844.

23. Kannel WB, Anderson K, Wilson PW. White blood cell count andcardiovascular disease: insights from the Framingham study.JAMA.1992;267:1253–1256.

24. Grau AJ, Buggle F, Becher H, Werle E, Hacke W. The association ofleukocyte count, fibrinogen and C-reactive protein with vascular riskfactors and ischemic vascular diseases.Thromb Res. 1996;82:245–255.

25. Prentice R, Szatrowski TP, Kato H, Mason MW. Leukocyte counts andcerebrovascular disease.J Chronic Dis. 1982;35:703–714.

26. D’Agostino RB, Burke G, O’Leary D, Rewers M, Selby J, Savage PJ,Saad MF, Bergman RN, Howard G, Wagenknecht L, Haffner S. Ethnicdifferences in carotid wall thickness: the Insulin Resistance Atheroscle-rosis Study.Stroke. 1996;27:1744–1749.

27. Arnett DK, Tyroler HA, Burke G, Hutchinson R, Howard G, Heiss G.Hypertension and subclinical carotid artery atherosclerosis in blacks andwhites: the Atherosclerosis Risk in Communities Study.Arch Intern Med.1996;156:1983–1989.

28. Novick AC, Zaki S, Goldfarb D, Hodge EE. Epidemiologic and clinicalcomparison of renal artery stenosis in black patients and white patients.JVasc Surg. 1994;20:1–5.

29. Swallen KC, West DW, Stewart SL, Glaser SL, Horn-Ross PL. Predictorsof misclassification of Hispanic ethnicity in a population-based cancerregistry.Ann Epidemiol. 1997;7:200–206.

30. O’Leary DH, Polak JF, Kronmal RA, Manolio TA, Burke GL, WolfsonSK. Carotid-artery intima and media thickness as a risk factor for myo-cardial infarction and stroke in older adults.N Engl J Med. 1999;340:14–22.

31. Munro JM, van der Walt JD, Munro CS, Chalmers JA, Cox EL. Animmunohistochemical analysis of human aortic fatty streaks.Hum Pathol.1987;18:375–380.

32. Fagerberg B, Gnarpe J, Gnarpe H, Agewall S, Wikstrand J.Chlamydiapneumoniaebut not cytomegalovirus antibodies are associated withfuture risk of stroke and cardiovascular disease.Stroke. 1999;30:299–305.

33. Elkind MS, Lin IF, Grayston TJ, Sacco RL.Chlamydia pneumoniaeandthe risk of first ischemic stroke: the Northern Manhattan Stroke Study.Stroke. 2000;31:1521–1525.

34. Yarnell JW, Sweetnam PM, Rogers S, Elwood PC, Bainton D, Baker IA,Eastham R, O’Brien JR, Etherington MD. Some long term effects ofsmoking on the haemostatic system: a report from the Caerphilly andSpeedwell Collaborative Surveys.J Clin Pathol. 1987;40:909–913.

35. Nieto FJ, Szklo M, Rock R, Mercuri M. Leukocyte count correlates inmiddle-aged adults: the Atherosclerosis Risk in Communities (ARIC)Study.Am J Epidemiol. 1992;136:525–537.

36. Jacobs DR, Adachi H, Mulder I, Kromhout D, Menotti A, Nissinen A,Blackburn H. Cigarette smoking and mortality risk: twenty-five-yearfollow-up of the Seven Countries Study.Arch Intern Med. 1999;159:733–740.

37. Wimmer C, Gluch H, Franzreb M, Ogon M. Predisposing factors forinfection in spine surgery: a survey of 850 spinal procedures.J SpinalDisord. 1998;11:124–128.

38. Sacco RL, Kargman DE Gu Q, Zamanillo MC. Race-ethnicity and deter-minants of intracranial atherosclerotic cerebral infarction: the NorthernManhattan Stroke Study.Stroke. 1995;26:14–20.

39. O’Leary DH, Polak JF, Kronmal RA, Manolio TA, Burke GL, WolfsonSK. Carotid-artery intima and media thickness as a risk factor for myo-cardial infarction and stroke in older adults.N Engl J Med. 1999;340:14–22.

40. Bots ML, Hoes AW, Koudstaal PJ, Hofman A, Grobbee DE. Commoncarotid intima-media thickness and risk of stroke and myocardialinfarction: the Rotterdam Study.Circulation. 1997;96:1432–1437.

41. Blankenhorn DH, Selzer RH, Crawford DW, Barth JD, Liu CR, Liu CH,Mack WJ, Alaupovic P. Beneficial effects of colestipol-niacin therapy onthe common carotid artery: two- and four-year reduction of intima-mediathickness measured by ultrasound.Circulation. 1993;88:20–28.

42. Hodis HN, Mack WJ, LaBree L, Selzer RH, Liu CR, Liu CH, AlaupovicP, Kwong-Fu H, Azen SP. Reduction in carotid arterial wall thicknessusing lovostatin and dietary therapy.Ann Intern Med. 1996;124:548–556.

43. Chambless LE, Heiss G, Folsom AR, Rosamond W, Szklo M, SharrettAR, Clegg LX. Association of coronary heart disease incidence withcarotid arterial wall thickness and major risk factors: the AtherosclerosisRisk in Communities (ARIC) Study, 1987–1993.Am J Epidemiol. 1997;146:483–494.

44. Liuzzo G, Kopecky SL, Frye RL, O’ Fallon WM, Maseri A, Goronzy JJ,Weyand CM. Perturbation of the T-cell repertoire in patients withunstable angina.Circulation. 1999;100:2135–2139.

45. Gupta S, Leatham EW, Carrington D, Mendall MA, Kaski JC, Camm AJ.ElevatedChlamydia pneumoniaeantibodies, cardiovascular events, andazithromycin in male survivors of myocardial infarction.Circulation.1997;96:404–407.

46. Gurfinkel E, Bozovich G, Daroca A, Beck E, Mautner B. Randomisedtrial of roxithromycin in non-Q-wave coronary syndromes: ROXIS pilotstudy.Lancet. 1997;350:404–407.

47. Gurfinkel E, Bozovich G, Beck E, Testa E, Livellara B, Mautner B.Treatment with the antibiotic roxithromycin in patients with acute non-Q-wave coronary syndromes: the final report of the ROXIS Study.EurHeart J. 1999;20:121–127.

48. Muhlestein JB, Anderson JL, Hammond EH, Zhao L, Trehan S, SchwobeEP, Carlquist JF. Infection withChlamydia pneumoniaeaccelerates thedevelopment of atherosclerosis and treatment with azithromycin preventsit in a rabbit model.Circulation. 1998;97:633–636.




ownloaded from


SaccoMitchell S. Elkind, Jianfeng Cheng, Bernadette Boden-Albala, Myunghee C. Paik and Ralph L.

Manhattan Stroke StudyElevated White Blood Cell Count and Carotid Plaque Thickness: The Northern

Print ISSN: 0039-2499. Online ISSN: 1524-4628 Copyright © 2001 American Heart Association, Inc. All rights reserved.

is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Stroke doi: 10.1161/01.STR.32.4.842

2001;32:842-849Stroke.

http://stroke.ahajournals.org/content/32/4/842World Wide Web at:

The online version of this article, along with updated information and services, is located on the

http://stroke.ahajournals.org//subscriptions/

is online at: Stroke Information about subscribing to Subscriptions:

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

document. Permissions and Rights Question and Answer process is available in the

Request Permissions in the middle column of the Web page under Services. Further information about thisOnce the online version of the published article for which permission is being requested is located, click

can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.Strokein Requests for permissions to reproduce figures, tables, or portions of articles originally publishedPermissions:



ownloaded from

http://stroke.ahajournals.org/content/32/4/842

http://www.ahajournals.org/site/rights/

http://www.lww.com/reprints

http://stroke.ahajournals.org//subscriptions/


elevated white blood cell count and carotid plaque...

Documents