Reported alcohol consumption and cognitive decline: The northern Manhattan study

BACKGROUND

Moderate alcohol intake may slow cognitive decline, and both vascular and neurodegenerative mechanisms have been implicated.

METHODS

We examined reported alcohol intake and cognitive decline in a community-based cohort of Hispanic, black and white individuals (n = 1,428). The role of the apolipoprotein E epsilon4 (APOE4) allele as a modifier was also studied.

RESULTS AND CONCLUSIONS

Reported drinking was as follows: 300 participants (21%) were 'never' drinkers, 622 (44%) 'past' drinkers, 145 (10%) reported taking less than 1 drink weekly, 330 (23%) 1 drink weekly up to 2 daily and 31 (2%) more than 2 drinks daily. A positive relationship was seen between reported alcohol intake and cognition. Drinking less than 1 drink a week (p = 0.09), between 1 drink weekly up to 2 drinks daily (p = 0.001) and more than 2 drinks daily (p = 0.003) were associated with less cognitive decline on the modified Telephone Interview for Cognitive Status compared to never drinkers. This dose-response relationship was not modified by the presence of an APOE4 allele in a subsample.

Levels of acute phase proteins remain stable after ischemic stroke

Background

Inflammation and inflammatory biomarkers play an important role in atherosclerosis and cardiovascular disease. Little information is available, however, on time course of serum markers of inflammation after stroke.

Methods

First ischemic stroke patients ≥40 years old had levels of high-sensitivity C-reactive protein (hsCRP), serum amyloid A (SAA), and fibrinogen measured in plasma samples drawn at 1, 2, 3, 7, 14, 21 and 28 days after stroke. Levels were log-transformed as needed, and parametric and non-parametric statistical tests were used to test for evidence of a trend in levels over time. Levels of hsCRP and SAA were also compared with levels in a comparable population of stroke-free participants.

Results

Mean age of participants with repeated measures (n = 21) was 65.6 ± 11.6 years, and 13 (61.9%) were men, and 15 (71.4%) were Hispanic. Approximately 75% of patients (n = 15) had mild strokes (NIH Stroke Scale score 0–5). There was no evidence of a time trend in levels of hsCRP, SAA, or fibrinogen for any of the markers during the 28 days of follow-up. Mean log(hsCRP) was 1.67 ± 1.07 mg/L (median hsCRP 6.48 mg/L) among stroke participants and 1.00 ± 1.18 mg/L (median 2.82 mg/L) in a group of 1176 randomly selected stroke-free participants from the same community (p = 0.0252).

Conclusion

Levels of hsCRP are higher in stroke patients than in stroke-free subjects. Levels of inflammatory biomarkers associated with atherosclerosis, including hsCRP, appear to be stable for at least 28 days after first ischemic stroke.

Alcohol intake, carotid plaque, and cognition: the Northern Manhattan Study

BACKGROUND AND PURPOSE

Moderate alcohol intake has been associated with better cognitive performance, implicating vascular and neurodegenerative processes. Few studies to clarify the importance of vascular disease have included direct measures of atherosclerosis or minority populations at higher risk of vascular disease and dementia.

METHODS

The Northern Manhattan Study includes stroke-free community based Hispanic (54%), black (25%), and white (22%) participants. We performed a cross-sectional study of alcohol intake and performance on the Mini-Mental State Examination (MMSE) in subjects with sonographic measurement of maximal carotid plaque thickness and adjusted for sociodemographic and vascular risk factors.

RESULTS

The median MMSE score was 27 (interquartile range 24 to 29; n=2215). Reported alcohol intake was divided into 5 groups: never (n=509), past (n=494), <1 drink/week (n=300), 1/week to < or =2 drinks/day (n=796), and >2 drinks/day (n=116). Drinking 1 drink weekly up to 2 daily was associated with better performance on the MMSE (odds ratio=1.19; 95% CI, 1.10 to 1.26) compared with nondrinkers in women (P< or =0.0001) but not in men, adjusting for sociodemographic and vascular risk factors. Maximal carotid plaque thickness (mean 1.1 mm; SD 1.2 mm) was not associated with alcohol intake and did not mediate the relationship between alcohol and cognition.

CONCLUSIONS

Moderate alcohol consumption was independently associated with better cognitive performance in women from this multiethnic sample. Carotid plaque was not a mediator of this association suggesting alcohol may impact cognition through a separate vascular or degenerative pathway.

Relative elevation in baseline leukocyte count predicts first cerebral infarction

BACKGROUND

Atherosclerosis is an inflammatory disease, and leukocyte levels are associated with future risk of ischemic cardiac disease.

OBJECTIVE

To investigate the hypothesis that relative elevations in leukocyte count in a stroke-free population predict future ischemic stroke (IS).

METHODS

A population-based prospective cohort study was performed in a multiethnic urban population. Stroke-free community participants were identified by random-digit dialing. Leukocyte levels were measured at enrollment, and participants were followed annually for IS, myocardial infarction (MI), and cause-specific mortality. Cox proportional hazards regression models were used to calculate hazard ratios (HRs) and 95% CIs for IS, MI, and vascular death after adjustment for medical, behavioral, and socioeconomic factors.

RESULTS

Among 3,103 stroke-free community participants (mean age 69.2 +/- 10.3 years) with baseline leukocyte levels measured, median follow-up was 5.2 years. After adjusting for stroke risk factors, each SD in leukocyte count (1.8 x 10(9) cells/L) was associated with an increased risk of IS (HR 1.22, 95% CI 1.05 to 1.42), and IS, MI, or vascular death (HR 1.13, 95% CI 1.02 to 1.26). Compared with those in the lowest quartile of leukocyte count, those in the highest had an increased risk of IS (adjusted HR 1.75, 95% CI 1.08 to 2.82). The effect on atherosclerotic and cardioembolic stroke was greater than in other stroke subtypes.

CONCLUSION

Relative elevations in leukocyte count are independently associated with an increased risk of future ischemic stroke and other cardiovascular events.

Moderate Alcohol Consumption Reduces Risk of Ischemic Stroke

Background and Purpose— Moderate alcohol consumption is protective against coronary disease, but its relationship to ischemic stroke (IS) is controversial.

Methods— Stroke-free participants ≥40 years of age identified by random-digit dialing were enrolled in a prospective cohort study between 1993 and 2001. Alcohol consumption was assessed through in-person interview and categorized as none in the past year, ≥1 drink in past month to ≤2 per day (moderate drinkers), and >2 drinks daily. Lifetime drinking was also assessed. Cox proportional hazard regression modeling was used to assess hazard ratios and their 95% CIs for the association of drinking with risk of stroke and vascular events.

Results— Mean age among participants (n=3176) was 69.1±10.3 years; 62.8% were women, 20.8% were non-Hispanic white, 24.5% non-Hispanic black, and 52.4% were Hispanic. No alcohol in the previous year was present in 62.3%, and 32.5% drank moderately. After adjusting for other risk factors compared with those who did not drink in the past year, moderate drinkers had a reduced risk of IS (0.67; 95% CI, 0.46 to 0.99) and IS, myocardial infarction, or vascular death (0.74; 95% CI, 0.59 to 0.94). Results were similar when never-drinkers were used as referent group. Reduction in risk was seen for nonatherosclerotic IS subtypes, and results stratified by age, sex, and race–ethnicity were similar.

Conclusion— Moderate alcohol consumption is associated with decreased risk of IS in a multiethnic population. This effect is independent of other risk factors and holds for nonatherosclerotic stroke subtypes.

Total homocysteine is associated with white matter hyperintensity volume: the Northern Manhattan Study

BACKGROUND

Total homocysteine (tHcy) has been implicated as a risk factor for stroke and dementia, but the mechanism is unclear. White matter hyperintensities may be a risk factor for both, but studies of the relationship between tHcy and quantitative measures of white matter hyperintensity volume (WMHV) are lacking, especially in minority populations.

METHODS

A community-based sample of 259 subjects with baseline tHcy levels underwent pixel-based quantitative measurement of WMHV. We examined the relationship between tHcy and WMHV adjusting for age, sociodemographics, vascular risk factors, and B12 deficiency.

RESULTS

Higher levels of tHcy were associated with WMHV adjusting for sociodemographics and vascular risk factors.

CONCLUSIONS

These cross-sectional data provide evidence that tHcy is a risk factor for white matter damage.

Total homocysteine and cognition in a tri-ethnic cohort: the Northern Manhattan Study

OBJECTIVE

Several studies implicate elevated homocysteine as a risk factor for dementia and cognitive decline, but most studies have involved subjects older than 55 years from homogeneous populations. The authors examined homocysteine and cognition in a tri-ethnic community sample 40 years and older.

METHOD

The Northern Manhattan Study includes 3,298 stroke-free subjects. Of these 2,871 had baseline fasting total homocysteine (tHcy) levels and Mini-Mental State Examination (MMSE) scores available. The authors used multiple linear regression to examine the cross-sectional association between baseline tHcy levels and mean MMSE scores adjusting for sociodemographic and vascular risk factors.

RESULTS

Homocysteine levels were related to age, renal function, and B12 deficiency. Those with B12 deficiency had tHcy levels five points higher (9.4 vs 14.4 nmol/L). Mean MMSE scores differed by age, sex, and race-ethnic group. Those with hypertension, diabetes, cardiac disease, and B12 deficiency had lower MMSE scores. In multivariate analyses, elevated tHcy was associated with lower mean MMSE scores for those older than 65 but not for those 40 to 64. Adjusting for B12 deficiency and sociodemographic factors the mean MMSE was 2.2 points lower for each unit increase in the log tHcy level (95% CI -3.6, -0.9). Adding vascular risk factors to the model did not attenuate this effect (mean MMSE -2.2 points; 95% CI -3.5, -0.9).

CONCLUSIONS

Elevated homocysteine was independently associated with decreased cognition in subjects older than 65 in this tri-ethnic cohort, adjusting for sociodemographic and vascular risk factors.

Homocysteine and the risk of ischemic stroke in a triethnic cohort: the NOrthern MAnhattan Study

BACKGROUND AND PURPOSE

The level of total homocysteine (tHcy) that confers a risk of ischemic stroke is unsettled, and no prospective cohort studies have included sufficient elderly minority subjects. We investigated the association between mild to moderate fasting tHcy level and the incidence of ischemic stroke, myocardial infarction, and vascular death in a multiethnic prospective study.

METHODS

A population-based cohort was followed for vascular events (stroke, myocardial infarction, and vascular death). Baseline values of tHcy and methylmalonic acid were measured among 2939 subjects (mean age, 69+/-10; 61% women, 53% Hispanics, 24% blacks, and 20% whites). Cox proportional models were used to calculate hazard ratios (HRs) and 95% CIs in tHcy categories after adjusting for age, race, education, renal insufficiency, B12 deficiency, and other risk factors.

RESULTS

The adjusted HR for a tHcy level > or =15 micromol/L compared with <10 micromol/L was greatest for vascular death (HR=6.04; 95% CI, 3.44 to 10.60), followed by combined vascular events (HR=2.27; 95% CI, 1.51 to 3.43), ischemic stroke (HR=2.01; 95% CI, 1.00 to 4.05), and nonvascular death (HR=2.02; 95% CI, 1.31 to 3.14). Mild to moderate elevations of tHcy of 10 to 15 micromol/L were not significantly predictive of ischemic stroke, but increased the risk of vascular death (2.27; 95% CI, 1.44 to 3.60) and combined vascular events (1.42; 95% CI, 1.06 to 1.88). The effect of tHcy was stronger among whites and Hispanics, but not a significant risk factor for blacks.

CONCLUSIONS

Total Hcy elevations above 15 micromol/L are an independent risk factor for ischemic stroke, whereas mild elevations of tHcy of 10 to 15 micromol/L are less predictive. The vascular effects of tHcy are greatest among whites and Hispanics, and less among blacks.

Abdominal obesity and risk of ischemic stroke: the Northern Manhattan Stroke Study

BACKGROUND AND PURPOSE

Obesity is well recognized as a risk factor for coronary heart disease and mortality. The relationship between abdominal obesity and ischemic stroke remains less clear. Our aim was to evaluate abdominal obesity as an independent risk factor for ischemic stroke in a multiethnic community.

METHODS

A population-based, incident case-control study was conducted July 1993 through June 1997 in northern Manhattan, New York, NY. Cases (n=576) of first ischemic stroke (66% >or=BORDER="0">65 years of age; 56% women; 17% whites; 26% blacks; 55% Hispanics) were enrolled and matched by age, sex, and race-ethnicity to stroke-free community controls (n=1142). All subjects were interviewed and examined and had measurements of waist-to-hip ratio (WHR). Odds ratios (ORs) of ischemic stroke were calculated with gender-specific quartiles (GQs) and gender-specific medians of WHR adjusted for stroke risk factors and body mass index (BMI).

RESULTS

Compared with the first quartile, the third and fourth quartiles of WHR had an increased risk of stroke (GQ3: OR, 2.4; 95% CI, 1.5 to 3.9; GQ4: OR, 3.0; 95% CI, 1.8 to 4.8) adjusted for other risk factors and BMI. Those with WHR equal to or greater than the median had an overall OR of 3.0 (95% CI, 2.1 to 4.2) for ischemic stroke even after adjustment for other risk factors and BMI. Increased WHR was associated with a greater risk of stroke in men and women and in all race-ethnic groups. The effect of WHR was stronger among younger persons (test for heterogeneity, P<0.0002) (<65 years of age: OR, 4.4; 95%CI, 2.2 to 9.0; >or=65 years of age: OR, 2.2; 95% CI, 1.4 to 3.2). WHR was associated with an increased risk among those with and without large-artery atherosclerotic stroke.

CONCLUSIONS

Abdominal obesity is an independent, potent risk factor for ischemic stroke in all race-ethnic groups. It is a stronger risk factor than BMI and has a greater effect among younger persons. Prevention of obesity and weight reduction need greater emphasis in stroke prevention programs.

Apolipoproteins and carotid artery atherosclerosis in an elderly multiethnic population: the Northern Manhattan stroke study

The association of apolipoproteins A-I and B (apo A-I and apo B) with cardiovascular disease has been studied in younger populations, but there is sparse information in the elderly. We determined whether apo A-I and apo B were associated with carotid artery atherosclerosis (CAA) in 507 stroke-free elderly community residents (mean age 70.1+/-11.7 years, 60% women, 41% Hispanics, 30% African American, 28% Caucasian). CAA severity was normal (no plaque or carotid stenosis) in 39%, mild (maximum plaque thickness < or =1.8 mm or carotid stenosis <40%) in 25%, and moderate/severe (maximum plaque thickness >1.8 mm or carotid stenosis > or =40%) in 36%. CAA severity increased with age in all race/ethnic groups (P<0.01). CAA was similar among African Americans and Caucasians, but less in Hispanics (age adjusted OR: 0.5, CI: 0.4-0.8). apo A-I <1.2 g/l (OR: 2.0, CI: 1.0-3.3) and apo B > or =1.4 g/l (OR: 2.0, CI: 1.1-3.6) were associated with moderate-severe CAA. An apo B/apo A-I ratio > or =1 was associated with moderate-severe CAA (OR: 2.4, CI: 1.3-4.4), and the association varied by race (Hispanics OR: 4.3, CI: 1.8-10; non-Hispanics, OR: 1.4, CI: 0.6-3.2). Total cholesterol, triglycerides and low density lipoprotein cholesterol were not associated with moderate-severe CAA, while high density lipoprotein cholesterol was protective (OR: 0.4, CI: 0.2-0.8). Thus, in an elderly population, apo A-I and B were determinants of moderate-severe CAA, and the degree of association varied by race/ethnicity

Relation of Apo(a) size to carotid atherosclerosis in an elderly multiethnic population

Lipoprotein(a) [Lp(a)] is a novel risk factor for atherosclerosis, whose role in multiracial populations has been debated. We recently demonstrated a significant association of elevated levels of Lp(a) carried in particles containing small apolipoprotein(a) [apo(a)] isoforms with coronary artery disease in African American and white men. To extend these findings, we investigated the associations between Lp(a) levels, apo(a) size, and maximum internal carotid artery plaque thickness (MPT) in a randomly selected elderly multiethnic population (173 men and 253 women, consisting of 135 African Americans, 146 Hispanics, and 145 whites; mean age 70.5+/-11.4 years). Lp(a) levels were not associated with MPT. Among white men, MPT was associated with a small apo(a) isoform size (P=0.03) as well as with the amount of Lp(a) carrying the small apo(a) size (P=0.04), and the latter showed a borderline association in African American men (P=0.07). Among white women, but not in Hispanic or African American women, MPT was associated with the amount of Lp(a) carrying a small apo(a) isoform size (P<0.01). For all patients, the amount of Lp(a) carrying the small apo(a) size was associated with carotid atherosclerosis when there was control for age, sex, ethnicity, high density lipoprotein cholesterol, low density lipoprotein cholesterol, triglycerides, diabetes mellitus, hypertension, waist-to-hip ratio, and current smoking status (P=0.03). This association was significant for all men (P=0.03) and for white women (P=0.007). The results suggest that molecular properties of apo(a) are important in determining the atherogenicity of Lp(a).

Mortality and causes of death after first ischemic stroke: the Northern Manhattan Stroke Study

OBJECTIVE

To analyze the early and long-term causes of death after first ischemic stroke in the multiethnic northern Manhattan community.

METHODS

In the prospective, population-based Northern Manhattan Stroke Study, 980 patients with first ischemic stroke (mean age 70 years; 56% women; 49% Caribbean Hispanic, 31% black, 20% white) were followed for a mean of 3 years. Causes of death were classified as vascular (incident stroke, recurrent stroke, cardiac) or nonvascular. Life table analyses were used to assess mortality risks among different race-ethnic groups. Early (< or =1 month) vs long-term (> 1 month to 5 years) causes of death were compared.

RESULTS

Among the 980 patients followed, 278 (28%) died; 47 (5%) died during the first month. Cumulative mortality risk was 5% at 1 month, 16% after 1 year, 29% after 3 years, and 41% after 5 years. The proportion of vascular deaths among all deaths was 75% at 1 month and 43% thereafter (p = 0.001). Stroke, either incident (53%) or recurrent (4%), caused early deaths in 57% and long-term deaths in 14% (p = 0.001). Overall mortality risks did not differ significantly among race-ethnic groups. However, the proportion of incident stroke-related early deaths was 85% in Caribbean Hispanic patients, 33% in white patients, and 25% in black patients (p = 0.002).

CONCLUSIONS

Among patients with first ischemic stroke, incident stroke is the leading cause of early deaths. A large proportion of long-term deaths are nonvascular in origin. Despite similar overall mortality rates in race-ethnic groups, our data suggest a higher incident stroke-related early mortality among Caribbean Hispanics.

Matched case-control data analysis with selection bias

Case-control studies offer a rapid and efficient way to evaluate hypotheses. On the other hand, proper selection of the controls is challenging, and the potential for selection bias is a major weakness. Valid inferences about parameters of interest cannot be drawn if selection bias exists. Furthermore, the selection bias is difficult to evaluate. Even in situations where selection bias can be estimated, few methods are available. In the matched case-control Northern Manhattan Stroke Study (NOMASS), stroke-free controls are sampled in two stages. First, a telephone survey ascertains demographic and exposure status from a large random sample. Then, in an in-person interview, detailed information is collected for the selected controls to be used in a matched case-control study. The telephone survey data provides information about the selection probability and the potential selection bias. In this article, we propose bias-corrected estimators in a case-control study using a joint estimating equation approach. The proposed bias-corrected estimate and its standard error can be easily obtained by standard statistical software.

Race-ethnic disparities in the impact of stroke risk factors: the northern Manhattan stroke study

BACKGROUND AND PURPOSE

Stroke risk factors have been determined in large part through epidemiological studies in white cohorts; as a result, race-ethnic disparities in stroke incidence and mortality rates remained unexplained. The aim in the present study was to compare the prevalence, OR, and etiological fraction (EF) of stroke risk factors among white, blacks, and Caribbean Hispanics living in the same urban community of northern Manhattan.

METHODS

In this population-based incident case-control study, cases (n=688) of first ischemic stroke were prospectively matched 1:2 by age, sex, and race-ethnicity with community controls (n=1156). Risk factors were determined through in-person assessment. Conditional logistic regression was used to calculate adjusted ORs in each race-ethnic group. Prevalence and multivariate EFs were determined in each race-ethnic group.

RESULTS

Hypertension was an independent risk factor for whites (OR 1.8, EF 25%), blacks (OR 2.0, EF 37%), and Caribbean Hispanics (OR 2.1, EF 32%), but greater prevalence led to elevated EFs among blacks and Caribbean Hispanics. Greater prevalence rates of diabetes increased stroke risk in blacks (OR 1.8, EF 14%) and Caribbean Hispanics (OR 2.1 P<0.05, EF 10%) compared with whites (OR 1.0, EF 0%), whereas atrial fibrillation had a greater prevalence and EF for whites (OR 4.4, EF 20%) compared with blacks (OR 1.7, EF 3%) and Caribbean Hispanics (OR 3.0, EF 2%). Coronary artery disease was most important for whites (OR 1.3, EF 16%), followed by Caribbean Hispanics (OR 1.5, EF 6%) and then blacks (OR 1.1, EF 2%). Prevalence of physical inactivity was greater in Caribbean Hispanics, but an elevated EF was found in all groups.

CONCLUSIONS

The prevalence, OR, and EF for stroke risk factors vary by race-ethnicity. These differences are crucial to the etiology of stroke, as well as to the design and implementation of stroke prevention programs.

High-density lipoprotein cholesterol and ischemic stroke in the elderly: the Northern Manhattan Stroke Study

CONTEXT

Elevated high-density lipoprotein cholesterol (HDL-C) levels have been shown to be protective against cardiovascular disease. However, the association of specific lipoprotein classes and ischemic stroke has not been well defined, particularly in higher-risk minority populations.

OBJECTIVE

To evaluate the association between HDL-C and ischemic stroke in an elderly, racially or ethnically diverse population.

DESIGN

Population-based, incident case-control study conducted July 1993 through June 1997.

SETTING

A multiethnic community in northern Manhattan, New York, NY.

PARTICIPANTS

Cases (n = 539) of first ischemic stroke (67% aged >/=65 years; 55% women; 53% Hispanic, 28% black, and 19% white) were enrolled and matched by age, sex, and race or ethnicity to stroke-free community residents (controls; n = 905).

MAIN OUTCOME MEASURE

Independent association of fasting HDL-C levels, determined at enrollment, with ischemic stroke, including atherosclerotic and nonatherosclerotic ischemic stroke subtypes.

RESULTS

After risk factor adjustment, a protective effect was observed for HDL-C levels of at least 35 mg/dL (0.91 mmol/L) (odds ratio [OR], 0.53; 95% confidence interval [CI], 0.39-0.72). A dose-response relationship was observed (OR, 0.65; 95% CI, 0.47-0.90 and OR, 0.31; 95% CI, 0.21-0.46) for HDL-C levels of 35 to 49 mg/dL (0.91-1.28 mmol/L) and at least 50 mg/dL (1.29 mmol/L), respectively. The protective effect of a higher HDL-C level was significant among participants aged 75 years or older (OR, 0.51; 95% CI, 0.27-0.94), was more potent for the atherosclerotic stroke subtype (OR, 0.20; 95% CI, 0.08-0.50), and was present in all 3 racial or ethnic groups studied.

CONCLUSIONS

Increased HDL-C levels are associated with reduced risk of ischemic stroke in the elderly and among different racial or ethnic groups. These data add to the evidence relating lipids to stroke and support HDL-C as an important modifiable stroke risk factor.

Elevated white blood cell count and carotid plaque thickness : the northern manhattan stroke study

BACKGROUND AND PURPOSE

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

METHODS

For this cross-sectional analysis, WBC was measured in stroke-free community subjects undergoing carotid duplex 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 calculate the effect of quartile of WBC on MICPT. Odds ratios (ORs) and 95% confidence intervals (CIs) for the effect of quartile of WBC on MICPT >/=75th percentile were calculated. All analyses were stratified by race-ethnicity.

RESULTS

The mean age of the 1422 subjects was 68.6+/-10.2 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 had significantly increased MICPT (mean difference=0.30 mm, P:=0.0086) after adjustment for age, sex, and other atherosclerotic 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 black non-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 by race-ethnicity. The association is strongest in Hispanics, intermediate in black non-Hispanics, and not present in white non-Hispanics in this population. Chronic subclinical infection or inflammation may account for this association.

Bivariate binary data analysis with nonignorably missing outcomes

One of the objectives in the Northern Manhattan Stroke Study is to investigate the impact of stroke subtype on the functional status 2 years after the first ischemic stroke. A challenge in this analysis is that the functional status at 2 years after stroke is not completely observed. In this paper, we propose a method to handle nonignorably missing binary functional status when the baseline value and the covariates are completely observed. The proposed method consists of fitting four separate binary regression models: for the baseline outcome, the outcome 2 years after the stroke, the product of the previous two, and finally, the missingness indicator. We then conduct a sensitivity analysis by varying the assumptions about the third and the fourth binary regression models. Our method belongs to an imputation paradigm and can be an alternative to the weighting method of Rotnitzky and Robins (1997, Statistics in Medicine 16, 81-102). A jackknife variance estimate is proposed for the variance of the resulting estimate. The proposed analysis can be implemented using statistical software such as SAS.

Predictors of resource use after acute hospitalization: the Northern Manhattan Stroke Study

OBJECTIVE

To determine demographic and clinical predictors of discharge destinations following acute care hospitalization for stroke in the community of northern Manhattan.

METHODS

A group of 893 patients (mean age, 70 +/- 12 years; 56% women; 51% Hispanic, 30% African-American, 19% white) who survived acute care hospitalization for a first ischemic stroke were followed prospectively. Stroke severity was assessed by the NIH Stroke Scale and categorized as mild (< or = 5), moderate (6 to 13), and severe (> or = 14). Polytomous logistic regression was used to determine predictors for rehabilitation and nursing home placement versus returning home.

RESULTS

Among the survivors of acute stroke care hospitalization, 611 (68%) patients were discharged to their homes, 168 (19%) to rehabilitation, and 114 (13%) to nursing homes. Patients with moderate and severe neurologic deficits had more than a threefold increased risk of being sent to a nursing home and more than an eightfold increased risk of being sent to rehabilitation. Age over 65 and cognitive impairment were associated with placement to a nursing home (age over 65: OR, 2.4; 95% CI, 1.0 to 5.6; cognitive impairment: OR, 2.9; 95%, CI 1.4 to 5.7), and rehabilitation (age over 65: OR, 1.8; 95% CI, 1.1 to 2.9; cognitive impairment: OR, 2.9; 95% CI, 1.4 to 5.7).

CONCLUSION

Our results demonstrated that one-third of patients with acute stroke from the community of northern Manhattan required placement in a temporary or a long-term disability care institution following acute care hospitalization. Severity of stroke is an important factor that influences discharge planning following acute care hospitalization and its reduction can improve health care resource usage.

Frequency and clinical determinants of dementia after ischemic stroke

OBJECTIVE

To investigate the frequency and clinical determinants of dementia after ischemic stroke.

METHODS

The authors administered neurologic, neuropsychological, and functional assessments to 453 patients (age 72.0 +/- 8.3 years) 3 months after ischemic stroke. They diagnosed dementia using modified Diagnostic and Statistical Manual of Mental Disorders, 3rd ed., revised criteria requiring deficits in memory and two or more additional cognitive domains as well as functional impairment.

RESULTS

The authors diagnosed dementia in 119 of the 453 patients (26.3%). Regarding dementia subtypes, 68 of the 119 patients (57.1%) were diagnosed with vascular dementia, 46 patients (38.7%) were diagnosed with AD with concomitant stroke, and 5 patients (4.2%) had dementia for other reasons. Logistic regression suggested that dementia was associated with a major hemispheral stroke syndrome (OR 3.0), left hemisphere (OR 2.1) and right hemisphere (OR 1.8) infarct locations versus brainstem/cerebellar locations, infarcts in the pooled anterior and posterior cerebral artery territories versus infarcts in other vascular territories (OR 1.7), diabetes mellitus (OR 1.8), prior stroke (OR 1.7), age 80 years or older (OR 12.7) and 70 to 79 years (OR 3.9) versus 60 to 69 years, 8 or fewer years of education (OR 4.1) and 9 to 12 years of education (OR 3.0) versus 13 or more years of education, black race (OR 2.6) and Hispanic ethnicity (OR 3.1) versus white race, and northern Manhattan residence (OR 1.6).

CONCLUSIONS

Dementia is frequent after ischemic stroke, occurring in one-fourth of the elderly patients in the authors' cohort. The clinical determinants of dementia include the location and severity of the presenting stroke, vascular risk factors such as diabetes mellitus and prior stroke, and host characteristics such as older age, fewer years of education, and nonwhite race/ethnicity. The results also suggest that concomitant AD plays an etiologic role in approximately one-third of cases of dementia after stroke.

Effect of hypervolemic therapy on cerebral blood flow after subarachnoid hemorrhage : a randomized controlled trial

BACKGROUND AND PURPOSE

Cerebral blood flow (CBF) is reduced after subarachnoid hemorrhage (SAH), and symptomatic vasospasm is a major cause of morbidity and mortality. Volume expansion has been reported to increase CBF after SAH, but CBF values in hypervolemic (HV) and normovolemic (NV) subjects have never been directly compared.

METHODS

On the day after aneurysm clipping, we randomly assigned 82 patients to receive HV or NV fluid management until SAH day 14. In addition to 80 mL/h of isotonic crystalloid, 250 mL of 5% albumin solution was given every 2 hours to maintain normal (NV group, n=41) or elevated (HV group, n=41) cardiac filling pressures. CBF ((133)xenon clearance) was measured before randomization and approximately every 3 days thereafter (mean, 4.5 studies per patient).

RESULTS

HV patients received significantly more fluid and had higher pulmonary artery diastolic and central venous pressures than NV patients, but there was no effect on net fluid balance or on blood volume measured on the third postoperative day. There was no difference in mean global CBF during the treatment period between HV and NV patients (P=0.55, random-effects model). Symptomatic vasospasm occurred in 20% of patients in each group and was associated with reduced minimum regional CBF values (P=0.04). However, there was also no difference in minimum regional CBF between the 2 treatment groups.

CONCLUSIONS

HV therapy resulted in increased cardiac filling pressures and fluid intake but did not increase CBF or blood volume compared with NV therapy. Although careful fluid management to avoid hypovolemia may reduce the risk of delayed cerebral ischemia after SAH, prophylactic HV therapy is unlikely to confer an additional benefit.

Using missing data techniques to explore the lack of survival effect: illustration with the CABG patch trial

The Multicenter Automatic Defibrillator Implantation Trial (MADIT) showed a conclusive 54 per cent reduction in mortality in patients with inducible sustained monomorphic ventricular tachycardia (VT) and impaired left ventricular function who received an implantable defibrillator compared with those who did not. The Coronary Artery Bypass Graft (CABG) Patch Trial, which studied a patient population with a similar extent of left ventricular dysfunction and overall cardiovascular risk, demonstrated no mortality benefit from placement of an implantable defibrillator. All patients in the MADIT trial were 'VT inducible', while this criterion was neither required nor evaluated for entry into the CABG Patch Trial. A statistical approach to estimating with good accuracy the fraction of CABG Patch patients who were inducible at the time of their randomization from the prevalence of VT inducibility in the surviving CABG Patch Trial control population during follow-up is presented. This more generally applicable approach estimates the mixing percentage using missing data techniques. We present the mathematical and physiological basis of the assumptions underpinning the mixture model and its estimation procedure. The mixture model forms the basis for the electrophysiological substudy to the CABG Patch Trial, which directly tests the hypothesis that the difference in the frequency of inducible VT between the MADIT and CABG Patch patients populations is sufficient to account for the difference in effect on mortality.

Treatment for the secondary prevention of stroke in older patients: the influence of dementia status

OBJECTIVE

To investigate the influence of dementia status on treatment for the secondary prevention of stroke in older patients.

DESIGN

Based on patient examinations and medical record review, we investigated the frequency of aspirin and/or warfarin use at hospital discharge for the prevention of recurrent stroke in older patients hospitalized with acute ischemic stroke.

SETTING

A large academic medical center.

PARTICIPANTS

A cohort of 272 patients, mean age 72.1 +/- 8.5 years.

MEASUREMENTS

We performed neurologic examinations and reviewed medical records to investigate the effects of a clinical diagnosis of dementia and other potentially relevant factors on treatment with aspirin or warfarin at hospital discharge.

RESULTS

Thirty-one patients (11.4%) were not prescribed aspirin or warfarin at hospital discharge. Logistic regression determined that dementia (odds ratio (OR) = 2.57, 95% confidence interval (CI), 1.04-6.30) was a significant independent determinant of nontreatment with aspirin or warfarin, adjusting for abnormal gait (OR = 2.01, CI, .88-4.59); discharge to a nursing home or other institutional residence (OR = 2.55, CI, .83-7.81); cardiac disease (OR = .39, CI, .16-.95); cortical infarct location (OR = .45, CI, .18-1.10); male sex (OR = .47, CI, .20-1.15); age 80+ (OR = 1.14, CI, .46-2.82) and age 70-79 (OR = .96, CI, .32-2.88) versus age 60-69.

CONCLUSIONS

Our results suggest that dementia is a significant independent determinant of nontreatment with aspirin or warfarin when otherwise indicated for the prevention of recurrent stroke. The underutilization of aspirin and warfarin in older stroke patients with dementia may be a modifiable basis for their increased risk of recurrence and death.

Validity of telephone interview data for vascular disease risk factors in a racially mixed urban community: the Northern Manhattan Stroke Study

The aims of our study were to assess the validity and reliability of a telephone survey instrument designed to measure vascular disease risk factors and to assess whether these measurements were influenced by age, gender, race/ethnicity, or other sociodemographic variables. Subjects were sampled and interviewed using random digit dialing methodology from the multiethnic community of northern Manhattan. For the validity study, 261 consecutive subjects were clinically assessed in-person within 60 days of the telephone interview. A retest reliability study of the telephone interview was conducted in 92 randomly selected subjects within 30 days of the initial interview. The telephone interview instrument had a sensitivity of more than 55% and a specificity of 74% or greater for various vascular disease risk factors. Sensitivity, specificity, and positive predictive value did not vary significantly or systematically among whites, blacks, and Hispanics, but subjects with access to health care were more likely to provide valid data. The reliability substudy indicated a good reliability for the telephone interview. These results support the validity of telephone interviews for estimating the prevalence of vascular risk factors in urban populations.

Assessment of cognition in advanced AD: The test for severe impairment

Article abstract The Test for Severe Impairment (TSI) was compared with the Mini-Mental State Examination (MMSE) and a modified MMSE (mMMSE) in a multisite, longitudinal study of AD. The TSI correlated highly with the MMSE (r = 0.83) and the mMMSE (r = 0.82), but was not redundant. There was a wide range of scores on the TSI among those scoring in the severely impaired range on the MMSE and mMMSE. The slope of cognitive change over time detected by the TSI was greater than that revealed by the MMSE or the mMMSE. Performance on the TSI was a significant predictor of survival. The TSI is a valid measure that is sensitive to cognitive change over time in severely demented patients with AD.

The protective effect of moderate alcohol consumption on ischemic stroke

CONTEXT

Moderate alcohol consumption has been shown to be protective for coronary heart disease, but the relationship between moderate alcohol consumption and ischemic stroke is more controversial.

OBJECTIVE

To determine the association between alcohol consumption and risk of ischemic stroke.

DESIGN

Population-based case-control study conducted between July 1993 and June 1997.

SETTING

Multiethnic population in northern Manhattan, New York, NY, aged 40 years or older.

PATIENTS AND OTHER PARTICIPANTS

Cases (n = 677) had first ischemic stroke and were matched to community controls (n = 1139) derived through random digit dialing by age, sex, and race/ethnicity. Mean +/- SD age of cases was 70.0+/-12.7 years; 55.8% were women; 19.5% were white, 28.4% black, and 50.7% Hispanic.

MAIN OUTCOME MEASURE

First ischemic stroke (fatal or nonfatal).

RESULTS

Moderate alcohol consumption, up to 2 drinks per day, was significantly protective for ischemic stroke after adjustment for cardiac disease, hypertension, diabetes, current smoking, body mass index, and education (odds ratio [OR], 0.51; 95% confidence interval [CI], 0.39-0.67). This protective effect of alcohol consumption was detected in both younger and older groups, in men and women, and in whites, blacks, and Hispanics. In a quadratic model of stroke risk, increased risk of ischemic stroke was statistically significant among those consuming 7 or more drinks per day (OR, 2.96; 95% CI, 1.05-8.29).

CONCLUSIONS

Moderate alcohol consumption was independently associated with a decreased risk of ischemic stroke in our elderly, multiethnic, urban subjects, while heavy alcohol consumption had deleterious effects. Our data support the National Stroke Association Stroke Prevention Guidelines regarding the beneficial effects of moderate alcohol consumption.

Risk factors for early recurrence after ischemic stroke: the role of stroke syndrome and subtype

BACKGROUND AND PURPOSE

Information regarding risk factors for early recurrence is limited. Our aim was to identify the clinical predictors of early recurrence after ischemic stroke.

METHODS

We prospectively examined 297 patients (mean age, 72.0+/-8.4 years) hospitalized with ischemic stroke to identify recurrent strokes occurring within 90 days of the index stroke. Survival free of recurrence was estimated using Kaplan-Meier analysis stratified by demographic variables; vascular risk factors; stroke syndrome, subtype, vascular territory, and severity; scores on the Barthel Index and Mini-Mental State Examination during hospitalization; blood pressure on admission; and selected laboratory data. We estimated the relative risk (RR) of early recurrence associated with those variables using proportional hazards analysis.

RESULTS

We identified 22 recurrent events in the first 90 days after the index stroke, resulting in an early stroke recurrence rate of 7.4%, and death occurred immediately after recurrence in 6 of the 22 patients. A major hemispheric stroke syndrome (RR=2.9; 95% confidence interval [CI]=1.2 to 7.1), atherothrombotic stroke mechanism (RR=3.3; CI=1.3 to 8.3), and atrial fibrillation (RR=2.2; CI=0.8 to 6.1) were independent predictors of early recurrence, after adjustment for demographic variables. Conclusions-Early recurrence was frequent and resulted in increased mortality. Attention to the clinical features of the index stroke, including the presenting syndrome and the ischemic mechanism, and the recognition of atrial fibrillation may help in the selection of patients for the initiation of targeted interventions to prevent early recurrence and subsequent mortality.

Lipid and lipoprotein levels remain stable in acute ischemic stroke: the Northern Manhattan Stroke Study

Serum lipoproteins including lipoprotein(a), Lp(a), are emerging as possible biological markers for cerebrovascular disease. Existing data on Lp(a) and serum lipids levels following acute ischemic stroke (AIS) are however equivocal. To determine whether serum Lp(a) and other lipid levels obtained within 24 h of acute ischemic stroke onset changed over the ensuing 4 weeks and whether these levels are related to an acute phase response, acquired nutritional deficiency, and neurovascular data, we conducted repeated measurement analyses among 19 subjects (mean age 65.0 +/- 12.1 years; 32% women) presenting with AIS (evaluated within 9.7 +/- 12.7 h). Eleven of the subjects had a moderate-to-severe stroke, defined by NIH stroke severity scale, and seven patients had a large cerebral infarction. Seven serial measurements of Lp(a), total cholesterol, high density lipoprotein cholesterol, low density lipoprotein cholesterol, and other lipoproteins, major acute phase reactants and albumin levels were collected for each subject over 4 weeks. The mean initial levels, (mg/dl), of total cholesterol, LDL cholesterol, HDL cholesterol, triglycerides, Lp(a), apolipoproteins A-I and B were: 225 +/- 57.6, 154 +/- 56.0, 40 +/- 10.4, 181 +/- 93.7, 52 +/- 28.6, 130 +/- 24.6, and 141 +/- 46.1, respectively. There were no significant changes in mean serum lipid, apolipoprotein or Lp(a) levels over the 4-week study period, analyzed by a random effects model to test for time trend. In addition, there were no significant changes in established acute phase or nutritional markers (C-reactive protein, alpha 1-glycoprotein, haptoglobin or serum albumin). Our findings suggest that serum lipid, apolipoprotein and Lp(a) levels remain stable following AIS, consistent with the absence of acute phase response or nutritional deficiency.

Stroke incidence among white, black, and Hispanic residents of an urban community: the Northern Manhattan Stroke Study

Stroke mortality is reported to be greater in blacks than in whites, but stroke incidence data for blacks and Hispanics are sparse. The aim of this study was to determine and compare stroke incidence rates among whites, blacks, and Hispanics living in the same urban community. A population-based incidence study was conducted to identify all cases of first stroke occurring in northern Manhattan, New York City, between July 1, 1993, and June 30, 1996. The population of this area was approximately 210,000 at that time, based on 1990 US Census data. Surveillance for hospitalized and nonhospitalized stroke consisted of daily screening of all admissions, discharges, and computed tomography logs at Columbia-Presbyterian Medical Center, the only hospital in the region, and review of discharge lists from outside hospitals, telephone surveys of random households, and contacts with community physicians, Visiting Nurses' Services, and community agencies. Stroke incidence increased with age and was greater in men than in women. The average annual age-adjusted stroke incidence rate at age > or =20 years, per 100,000 population, was 223 for blacks, 196 for Hispanics, and 93 for whites. Blacks had a 2.4-fold and Hispanics a twofold increase in stroke incidence compared with whites. Cerebral infarct accounted for 77 percent of all strokes, intracerebral hemorrhage for 17 percent, and subarachnoid hemorrhage for 6 percent. These data from the Northern Manhattan Stroke Study suggest that part of the reported excess stroke mortality among blacks in the United States may be a reflection of racial/ethnic differences in stroke incidence.

Leisure-time physical activity and ischemic stroke risk: the Northern Manhattan Stroke Study

BACKGROUND AND PURPOSE

Physical activity reduces the risk of premature death and cardiovascular disease, but the relationship to stroke is less well studied. The objective of this study was to investigate the association between leisure-time physical activity and ischemic stroke in an urban, elderly, multiethnic population.

METHODS

The Northern Manhattan Stroke Study is a population-based incidence and case-control study. Case subjects had first ischemic stroke, and control subjects were derived through random-digit dialing with 1:2 matching for age, sex, and race/ethnicity. Physical activity was recorded through a standardized in-person interview regarding the frequency and duration of 14 activities over the 2 prior weeks. Conditional logistic regression was used to calculate odds ratios (OR) and 95% confidence intervals after adjustment for medical and socioeconomic confounders.

RESULTS

Over 30 months, 369 case subjects and 678 control subjects were enrolled. Mean age was 69.9 +/- 12 years; 57% were women, 18% whites, 30% blacks, and 52% Hispanics. Leisure-time physical activity was significantly protective for stroke after adjustment for cardiac disease, peripheral vascular disease, hypertension, diabetes, smoking, alcohol use, obesity, medical reasons for limited activity, education, and season of enrollment (OR = 0.37; 95% confidence interval=0.25 to 0.55). The protective effect of physical activity was detected in both younger and older groups, in men and women, and in whites, blacks, and Hispanics. A dose-response relationship was shown for both intensity (light-moderate activity OR = 0.39; heavy OR = 0.23) and duration (<2 h/wk OR = 0.42; 2 to <5 h/wk OR = 0.35; > or =5 h/wk OR = 0.31) of physical activity.

CONCLUSIONS

Leisure-time physical activity was related to a decreased occurrence of ischemic stroke in our elderly, multiethnic, urban subjects. More emphasis on physical activity in stroke prevention campaigns is needed among the elderly.

Multiple imputation methods for the missing covariates in generalized estimating equation

This paper discusses the missing covariates problem in the generalized estimating equation (GEE) model. Estimates by various multiple imputation techniques (MI) are examined and compared to the sample average imputation method (SA) through simulations and an example. The simulation results show that, under the correct model specification, the MI estimators have negligible bias and have fairly similar efficiencies as the SA estimator. A practical advantage of the MI estimates is that the standard errors can be more easily computed than the SA estimates.

Generalized estimating equation model for binary outcomes with missing covariates

This paper presents an approach to handling missing covariates in the generalized estimating equation (GEE) model for binary outcomes when the probability of missingness depends on the observed outcomes and covariates. The proposed method is to replace the missing quantities in the estimating function with consistent estimates. In special cases, the proposed model reduces to a weighted GEE model for the completely observed units, where the weight is the inverse of the probability of missingness. Our method can be viewed as an extension of the mean score method by Reilly and Pepe (1995, Biometrika 82, 299-314) to the GEE context. Under certain regularity conditions, the estimates of the regression coefficients obtained by the proposed method are consistent and asymptotically normally distributed. The finite sample properties of the estimates are illustrated via computer simulations. An application to the study of dementia among stroke patients is presented.

Meta-analysis of the Hachinski Ischemic Score in pathologically verified dementias

Our objectives were to investigate the utility of the Hachinski Ischemic Score (HIS) in differentiating patients with pathologically verified Alzheimer's disease (AD), multi-infarct dementia (MID), and "mixed" (AD plus cerebrovascular disease) dementia, and to identify the specific items of the HIS that best discriminate those dementia subtypes. Investigators from six sites participated in a meta-analysis by contributing original clinical data, HIS, and pathologic diagnoses on 312 patients with dementia (AD, 191; MID, 80; and mixed, 41). Sensitivity and specificity of the HIS were calculated based on varied cutoffs using receiver-operator characteristic curves. Logistic regression analyses were performed to compare each pair of diagnostic groups to obtain the odds ratio (OR) for each HIS item. The mean HIS (+/- SD) was 5.4 +/- 4.5 and differed significantly among the groups (AD, 3.1 +/- 2.5; MID, 10.5 +/- 4.1; mixed, 7.7 +/- 4.3). Receiver-operator characteristic curves showed that the best cutoff was < or = 4 for AD and > or = 7 for MID, as originally proposed, with a sensitivity of 89.0% and a specificity of 89.3%. For the comparison of MID versus mixed the sensitivity was 93.1% and the specificity was 17.2%, whereas for AD versus mixed the sensitivity was 83.8% and the specificity was 29.4%. HIS items distinguishing MID from AD were stepwise deterioration (OR, 6.06), fluctuating course (OR, 7.60), hypertension (OR, 4.30), history of stroke (OR, 4.30), and focal neurologic symptoms (OR, 4.40). Only stepwise deterioration (OR, 3.97) and emotional incontinence (OR, 3.39) distinguished MID from mixed, and only fluctuating course (OR, 0.20) and history of stroke (OR, 0.08) distinguished AD from mixed. Our findings suggest that the HIS performed well in the differentiation between AD and MID, the purpose for which it was originally designed, but that the clinical diagnosis of mixed dementia remains difficult. Further prospective studies of the HIS should include additional clinical and neuroimaging variables to permit objective refinement of the scale and improve its ability to identify patients with mixed dementia.

Race-ethnicity and determinants of carotid atherosclerosis in a multiethnic population. The Northern Manhattan Stroke Study

BACKGROUND AND PURPOSE

Risk factors for carotid atherosclerosis have been studied in white populations but infrequently in multiethnic cohorts. The aim of this study was to determine the importance of race-ethnicity and other factors associated with carotid atherosclerosis in a mixed population of Hispanics, blacks, and whites.

METHODS

As part of the Northern Manhattan Stroke Study, 526 stroke-free community residents (aged > or = 40 years; 41% men, 59% women; 46% Hispanic, 31% black, 23% white) were recruited through random-digit dialing and had vascular risk factor evaluations. Maximum internal carotid artery plaque thickness (MICPT) was measured with B-mode ultrasound. The frequency distribution of MICPT was examined in the three race-ethnic groups, and multivariate regression was performed to identify factors that were independently associated with MICPT.

RESULTS

Mean MICPT in the entire sample was 1.5 +/- 1.4 mm, increased directly with age, and was greater in whites and blacks than Hispanics. Other independent determinants of MICPT included smoking, glucose, LDL cholesterol, and hypertension. After we controlled for these covariates, Hispanic (versus non-Hispanic) race-ethnicity was still an independent determinant of less carotid plaque. There was a significant interaction between race-ethnicity and LDL cholesterol, with a greater effect of increasing LDL cholesterol among Hispanics.

CONCLUSIONS

Atherosclerotic risk factors were predictive of MICPT in this mixed-ethnic cohort. Hispanics had significantly less carotid plaque after adjustment for other known risk factors, but they also had a greater impact of increasing LDL cholesterol.

Dementia after stroke increases the risk of long-term stroke recurrence

BACKGROUND

Although risk factors for first stroke have been identified, the predictors of long-term stroke recurrence are less well understood. We performed the present study to determine whether dementia diagnosed three months after stroke onset is an independent risk factor for long-term stroke recurrence.

METHODS

We examined 242 patients (age = 72.0 +/- 8.7 years) hospitalized with acute ischemic stroke who had survived the first three months without recurrence and followed them to identify predictors of long-term stroke recurrence. We diagnosed dementia three months after stroke using modified DSM-III-R criteria based on neuropsychological and functional assessments. The effects of conventional stroke risk factors and dementia status on survival free of recurrence were estimated using Kaplan-Meier analyses, and the relative risks (RR) of recurrence were calculated using Cox proportional hazards models.

RESULTS

Dementia (RR = 2.71, 95% CI = 1.36 to 5.42); cardiac disease (RR = 2.18, CI = 1.15 to 4.12); and sex, with women at higher risk (RR = 2.03, CI = 1.01 to 4.10), were significant independent predictors of recurrence, while education (RR = 1.90, CI = 0.77 to 4.68), admission systolic blood pressure >160 mm Hg (RR = 1.80, CI = 0.94 to 3.44) and alcohol intake exceeding 160 grams per week (RR = 1.86, CI = 0.79 to 4.38) were weakly related.

CONCLUSIONS

Our results suggest that dementia significantly increases the risk of long-term stroke recurrence, with additional independent contributions by cardiac disease and sex. Cognitive impairment may be a surrogate marker for multiple vascular risk factors and larger infarct volume that may serve to increase the risk of recurrence. Alternatively, less aggressive medical management of stroke patients with cognitive impairment or noncompliance of such patients with medical therapy may be bases for an increased rate of stroke recurrence.

Multiple imputation for the Cox proportional hazards model with missing covariates

We present three multiple imputation estimates for the Cox model with missing covariates. Two of the suggested estimates are asymptotically equivalent to estimates in the literature when the number of multiple imputations approaches infinity. The third estimate can be implemented using standard software that could handle time-varying covariates.

Risk factors for incident dementia after stroke. Role of hypoxic and ischemic disorders

BACKGROUND AND PURPOSE

Stroke significantly increases the risk of dementia in the elderly, yet the risk factors for incident dementia after ischemic stroke are not well understood. We attempted to determine whether hypoxic-ischemic (HI) disorders, which may result from comorbid medical conditions (eg. seizures, cardiac arrhythmias, pneumonia), would be an independent risk factor for the development of new dementia after stroke.

METHODS

We prospectively followed 185 initially nondemented patients with ischemic stroke (age, 70.3 +/- 7.7 years) for a maximum of 52.8 months. We diagnosed the presence of dementia at annual examinations based on neuropsychological testing and modified DSM-III-R criteria. HI disorders were identified by record review or examination during hospitalization. We used Kaplan-Meier analysis to determine the cumulative proportion of patients with and without HI disorders who survived free of dementia and used Cox models to estimate the relative risk of dementia associated with HI disorders.

RESULTS

The cumulative proportion (+/- SE) surviving without dementia was 51.7 +/- 10.9% in the HI group versus 78.2 +/- 4.3% in the non-HI group after 52.8 months of observation. The relative risk of incident dementia associated with HI events was 4.3 (95% confidence interval = 1.9 to 9.6) after we adjusted for demographic factors, recurrent stroke, and baseline cognitive function.

CONCLUSIONS

We conclude that HI disorders may be a significant independent risk factor for incident dementia after stroke, even after adjustment for other recognized predictors of cognitive decline. Recognition of HI cerebral damage as a possible pathogenic mechanism for dementia after stroke may allow targeted therapeutic interventions to prevent subsequent cognitive deterioration.

Difference in clinical implications of CD4 counts among HIV-infected homosexual men and injection drug using men and women

While the relationship between CD4 counts and clinical symptoms is well established among homosexual men, the same is not true for injection drug using men and women (IDUM and IDUW). In this paper we investigate whether CD4 counts have the same clinical implications for IDUM and IDUW as for homosexual men. We estimated the CD4 counts at which 50 per cent of the HIV-infected but AIDS-free population has AIDS related complex (ARC) based on three biannually measured CD4 counts. The analyses involve interval, right and left censored threshold data. We took the parametric approach, assuming that the threshold values for ARC arise from a family of distributions that includes symmetric, left or right skewed distributions, in which the logistic and extreme value distributions are embedded as special cases. The resulting estimates of median thresholds of CD4 counts for ARC were 249, 424 and 755 for homosexual men, IDUM, and IDUW, respectively. The results were robust with respect to the assumptions on the underlying distribution.

Multivariate survival analysis using piecewise gamma frailty

In this note we propose a frailty model called piecewise gamma frailty for correlated survival data with random effects having a nested structure. In frailty models, a dependence function defined as a hazard ratio of one member given the failure time of another member in a unit is determined by the distributional assumptions on frailty. In the piecewise gamma frailty model, the nested structure of random effects or frailty allows the dependence function to vary over the time periods. This model includes existing models such as the piecewise exponential model (Breslow, 1974, Biometrics 30, 89-100) and the gamma frailty model (Clayton, 1978, Biometrika 65, 141-151; Oakes, 1982, Journal of the Royal Statistical Society, Series B 44, 414-428) as special cases. A study of familial aggregation of epilepsy is used to illustrate the proposed method.

Parametric variance function estimation for nonnormal repeated measurement data

Zeger and Liang (1986, Biometrics 42, 121-130) proposed a procedure for analyzing nonnormal longitudinal data in the context of the generalized linear model. This procedure is extended to model variance heterogeneity, allowing the observations to come from distributions with different scale parameters. Loss of efficiency is evaluated when heterogeneity of scale factor is ignored.