From the Institute of Atherosclerosis Research (K.B., H.S., G.A.), the
Institute of Epidemiology and Social Medicine (K.B.), and the Department of
Neurology (K.B., F.S.), University of Muenster, Muenster, Germany.
Correspondence to Klaus Berger, MD, MPH, MSc, Institute of Epidemiology and Social Medicine, Muenster University, Domagkstr 3, 48129 Muenster, Germany. E-mail bergerk{at}uni-muenster.de
Methods—We studied a prospective cohort of 12 866 male
employees, aged 30 to 65 years, in 52 companies in northwestern
Germany, with an average follow-up of 7.2 years. Participants were free
of self-reported stroke, transient ischemic attack, and
myocardial infarction at baseline. Physical examination, blood tests,
and a face-to-face interview to assess presence of various risk factors
were performed at the workplace. Follow-up was done by standardized
mailed questionnaire. Main outcome measure was first stroke
occurrence.
Results—Overall stroke incidence was 42.4 per 100 000
person-years, increasing from 10.1 per 100 000 person-years in the age
category 30 to 39 years to 33.6, 80.6, and 159.2 per 100 000
person-years in the age categories 40 to 49, 50 to 59, and 60 years and
older, respectively. After adjustment for potential confounders, the
relative risks of total stroke associated with systolic blood
pressure
Conclusions—This occupational cohort had a 2-fold lower stroke
incidence than that observed in cohorts of the general population. In
the absence of a strong healthy-worker effect, moderate differences in
behavioral risk factors and a higher treatment rate for hypertension
contribute to the explanation of this favorable stroke incidence.
Stroke is a leading cause of disabling morbidity and death in many
countries of the world.3 4 5 Stroke events
in actively employed individuals often cause death or premature
retirement because of disease-related disability. Because of a strong
age dependency, stroke events are not as frequent in working as in
general populations but still have considerable socioeconomic impact.
Incidence data on stroke are available from only 2
registers6 7 in Germany. The Prospective
Cardiovascular Muenster Study (PROCAM) offered the
first opportunity to evaluate stroke occurrence and magnitude of stroke
risk factors in a German prospective cohort study.
Blood was drawn from each participant in a sitting position. Levels of
total cholesterol, triglycerides, and HDL
cholesterol were measured using enzymatic
assays11 ; HDL cholesterol was
measured with a precipitation method from Boehringer
Mannheim12 using a Hitachi 737
autoanalyzer.13 LDL
cholesterol was calculated by the Friedewald formula if
triglycerides were <400 mg/dL.14 A
subject was considered to be glucose intolerant if a diagnosis of
diabetes mellitus was known or if glucose in the specimen of whole
fasting blood was
During follow-up, participants were mailed a brief questionnaire every
2 years asking about the occurrence of new
cardiovascular and cerebrovascular events, including
myocardial infarction and stroke. Deaths were usually reported by
families, and death certificates were reviewed in all cases. For all
events, hospital records or records from the attending
physician were requested, as well as an eyewitness account in case of
death. The total follow-up was 96% after a maximum of 2 reminders by
mail and phone.
A definite stroke was defined as a focal neurological deficit that
lasted longer than 24 hours and was attributable to a vascular event.
Strokes were classified into ischemic and hemorrhagic subtypes
on the basis of mode of onset, clinical findings, and test results. The
necessary information was extracted from all available medical
records using a standardized stroke assessment
form.15 The rate of MRI and/or CT scan
performance in stroke cases was 83%. Because of the small
number (n=6), hemorrhagic strokes are not presented separately
in this analysis. Strokes were categorized as undetermined if
clinical data, although consistent with stroke, did not allow a
distinction between ischemic and hemorrhagic subtype. Severity
of stroke at hospital discharge or at time of stabilization for
patients who were not hospitalized was determined using the following
6-grade scale: no residual impairment (grade 1); minor nonfunctionally
impairing deficit (grade 2); mild functional deficit with some
restriction of lifestyle (grade 3); moderate deficit significantly
interfering with activities of daily life (grade 4); dependent state
requiring chronic care (grade 5); and fatality (grade 6).
All reported events were first independently classified by 2 study
neurologists (K.B., F.S.) after review of medical records and all
other available information. Interrater reliability (kappa statistic)
was calculated to compare the 2 classifications. The interrater
reliability was good (
For the analysis, only first stroke events were included.
Because of the very small number of strokes among women and younger
men, the analysis was limited to male participants aged 30 to
65 years without a prior history of myocardial infarction and stroke.
Because of their ongoing working status, 29 pensioners aged 66 years
and older with part-time work contracts were also included. The total
number of participants was 12 866, contributing 92 093 person-years
over an average follow-up of 7.2 years.
Statistical Analysis
To better demonstrate a potential healthy-worker effect, we compared
participants of the PROCAM Study with those of a general
population–based study (the Augsburg MONICA Survey, 1984/1985). The
latter, conducted at about midpoint of the course of the PROCAM study,
did not assess stroke and had a participation rate of 80.0%. Marked
differences (Table 2
During an average of 7.2 years of follow-up, 39 total strokes
were observed. Thirty-two were of ischemic, 6 of hemorrhagic,
and 1 of undetermined subtype. The overall incidence of total stroke
was 42.4 per 100 000 person-years in this cohort of 12 866 working
men. Stroke occurrence showed an expected strong age dependency. The
incidence increased from 10.1 per 100 000 person-years in the age
category 30 to 39 years to 33.6, 80.6, and 159.2 per 100 000
person-years in the age categories 40 to 49, 50 to 59, and 60 years and
older, respectively. Of the observed 39 strokes, 11 were fatal, 10
caused a major deficit, and 17 caused a minor functional deficit. No
information on severity was available for 1 stroke. Differences in
baseline characteristics were reflected in different rates of stroke in
categories of major risk factors (Table 3
In addition to age, 4 baseline risk factors were statistically
significantly associated with incident stroke in both
univariate and multivariate regression
analyses (Table 4
Compared with studies of the general population, the stroke incidence
observed in this occupational cohort was about 2-fold lower. The 2
German population-based stroke registers also reported higher rates.
Eisenblätter et al6 found an annual
incidence of 88 per 100 000 men aged 25 to 64 years, which is twice as
high. However, they report results from the MONICA stroke register in
the former German Democratic Republic. Differences in methods and
healthcare systems do not allow a true comparison. Kolominsky-Rabas et
al7 describe an even higher overall annual
incidence of 146 per 100 000 men in the community-based stroke
register in Erlangen, Germany. This overall incidence rate is caused
mainly by a high mean age (73.4 years) of stroke case subjects in this
register. Also, age-specific incidence rates were higher in all age
categories compared with those in the PROCAM
Study.22 Again, differences in methods (community
register versus prospective cohort study) do not allow a true
comparison.
Other general population studies have found incidences among males of
30 per 100 000 person-years in the third decade of life to 320 per
100 000 person-years in the age group of 55 to 64
years.23 24 25 26 One study reported incidences among
healthy men only in their fourth decade of age
separately.26 Their rate of 48 per 100 000
person-years is only half as high as the rate among all men of this age
in this study and comes close to the rate in our study.
The lower overall incidence in our study may be explained by several
factors. An initially expected healthy-worker effect is not among them,
since the observed differences in known stroke risk factor levels
compared with a German general population study are small. However, job
status is an important factor of social class. Differences in
incidences of cardiovascular and cerebrovascular
diseases according to social class are well
known.1 27 Participation at baseline was
optional. This might have attracted health-conscious employees, since
the chronically ill are less likely to volunteer for enrollment in any
study.28 Health consciousness includes motivation
for initiation and adherence to treatment if diagnosed with a crucially
elevated risk factor. The assessed treatment rate in the PROCAM Study
among employees with hypertensive blood pressure values or aware of
hypertension was almost twice as high as in the general population
survey (44.3% versus 25.0%). In the absence of a marked difference in
risk profiles, substantially higher treatment rates in addition to
small differences in risk factor levels might help explain a low
overall stroke incidence. Other behavioral factors that might
contribute to an explanation were either not comparable because of
different assessment methodology (physical activity) or remained
insignificant (alcohol consumption). Nevertheless, having 1 or more
classic risk factors for stroke is of special importance for those
employees with high (
This analysis is subject to a number of limitations. First,
events were self-reported. Underreporting of stroke from case subjects,
especially those left with a functional deficit, would result in an
underestimation of stroke incidence in this cohort. However, since the
number of reported events was considerably larger than the number of
validated strokes, underreporting seems unlikely. Also, we observed the
same strong age dependency and increases in event rates across
categories of known stroke risk factors as in studies of general
populations. This supports the internal validity of our data. The
observed low stroke incidence is in accordance with a 30% lower
overall mortality in this cohort30 based on
comparisons with German vital statistics (Federal Statistics Office).
Self-reported stroke was validated by medical records. The
validation procedure has been successfully applied to other studies
before and was recently the subject of an interobserver agreement
study15 and a review.31
Risk factor status was only measured once at baseline. Participants who
changed smoking status, for example, during follow-up might have
contributed to misclassification of exposure. However, since all
information about risk factors status was assessed at baseline and
before a stroke event, this misclassification would have been
nondifferential.
In conclusion, we found that participants in our male working cohort
had a very low incidence of stroke compared with those in general
population studies. This finding cannot be explained by a
healthy-worker effect. Carrying one of the classic stroke risk factors
is of special importance in this group of employees because these
individuals face high RR of stroke. Low incidence does not mean low
severity. When a stroke occurred, only 43% of case subjects were left
with a minor functional deficit or a better outcome. Because the risk
magnitudes associated with hypertension and smoking are
high,21 32 early and effective treatment of blood
pressure, even at levels currently not strictly recommended for
treatment, and avoidance of smoking should be encouraged by companies.
This would contribute to a reduction of the burden of disease for
employees, their families, and the companies themselves.
Received February 18, 1998;
revision received April 24, 1998;
accepted May 14, 1998.
© 1998 American Heart Association, Inc.
Original Contributions
Incidence and Risk Factors for Stroke in an Occupational Cohort
The PROCAM Study
Abstract
Top
Abstract
Introduction
Subjects and Methods
Results
Discussion
References
Background and Purpose—The purpose
of this study was to assess the incidence of stroke and the magnitude
of classic stroke risk factors in an occupational cohort of
white-collar and blue-collar workers. 
120, 121 to 140, and 
141 mm Hg were 1.00
(reference), 2.99 (95% confidence interval, 0.85 to 10.49), and 5.56
(1.56 to 19.88). The risks associated with smoking status of
never/past, 20 cigarettes per day, and >20 cigarettes per day) were
1.00 (reference), 1.65 (0.62 to 4.42), and 3.56 (1.78 to 7.15),
respectively. A history of hypertension at baseline (yes versus no) was
independently associated with a relative risk of 2.37 (1.20 to 4.71)
for total stroke and a history of diabetes mellitus (yes versus no)
with a risk of 2.21 (1.00 to 4.87). A comparison of risk factor levels
with a general population study revealed only small differences.
Key Words: cerebrovascular disorders • epidemiology • incidence • occupation • risk factors • stroke
Introduction
Top
Abstract
Introduction
Subjects and Methods
Results
Discussion
References
Occupational cohorts
differ in many aspects of morbidity from cohorts based on the general
population. Disease incidences are in general lower and risk factor
profiles more favorable among working
populations,1 2 unless work-related factors
represent risks for specific diseases. The aim of most studies
conducted in occupational cohorts is to evaluate potential associations
between these work-related factors and certain diseases of interest.
The rare chance to prospectively assess cardiovascular
and cerebrovascular diseases in an occupational cohort offers an
opportunity to contrast incidence rates and risk factor profiles with
those from general populations.
Subjects and Methods
Top
Abstract
Introduction
Subjects and Methods
Results
Discussion
References
A detailed description of participants and methods of the PROCAM
Study has previously been published.8 9 Briefly,
25 000 male and female blue- and white-collar employees, aged 16 to 65
years, in 52 companies and public authorities in the district of
Muenster in the northwestern part of Germany were recruited for study
participation on a voluntary basis from 1979 through 1990. Baseline
examinations took place during paid working hours at the workplace and
included face-to-face assessment of medical history using standardized
questionnaires, standardized measurement of blood pressure and
anthropometric data, a resting ECG, and collection of a blood sample
after a 12-hour fast for the determination of various laboratory
parameters. Participation rates in each company varied
between 40% and 80%. Participation was free of charge, both to
employees and to their employers. All findings were reported to the
participants themselves and their general practitioners.
The assessment of medical history included questions about current
medical treatment; prior history of cardiovascular and
cerebrovascular disease, hypertension, and diabetes mellitus; family
history of cardiovascular disease; and information
about smoking status, alcohol consumption, and physical activity. A
participant was considered a current smoker if he or she smoked
cigarettes daily within the last 12 months. The recommendations of the
American and British Heart Associations were followed for blood
pressure measurements.10 Systolic and
diastolic readings were taken in the left arm with the
subject seated and the arm at heart level. One measurement was taken at
the start of the interview by the examining physician and one at
its end. Only the second measurement was used in the
analysis. 120 mg/dL. Laboratory analyses were carried
out at the Institute of Atherosclerosis Research at the
University of Muenster. The methods of laboratory analyses were
validated by regular analyses of reference sera supplied by the
national German INSTAND proficiency testing program and the
international quality assurance program of the US Centers for Disease
Control and Prevention. 
=0.74) for total stroke, good (=0.74) for
ischemic subtype, and excellent (=1.0) for hemorrhagic
subtype. In a second step, the 2 neurologists made an end-point
decision by commonly confirming or not confirming each reported
event.
Analysis was performed using the STATA Statistical
Package. Comparisons between groups were based on Student's
t test for continuous variables and the
2 test for discrete variables. For the
latter, Fisher's exact test was used for groups of <10
participants. Age-standardized rates of stroke occurrence were
calculated using the age distribution of those participants who did not
develop a stroke (noncases). Participants were classified in categories
of the studied variables to determine incidence rates across risk
factors. The Cox proportional hazards model16 was
used to estimate the relative risks (RRs) of major stroke risk factors
for total stroke, controlling for important confounders. The assumption
of proportional hazards was evaluated for categories of
systolic blood pressure. No significant change in RRs over time
was found. Thus, stroke events that occurred during the complete
follow-up period were included in the analysis. Tests of linear
trend in RRs were calculated using the respective categories as an
ordinal variable in the proportional hazards model. The RRs and
95% significance levels were calculated using the lowest category as
reference. Separate models were calculated for systolic and
diastolic blood pressures, since both risk factors were
highly correlated.
Results
Top
Abstract
Introduction
Subjects and Methods
Results
Discussion
References
Baseline characteristics of cohort members are given in Table 1
. Stroke case subjects were
significantly older and had higher systolic and
diastolic blood pressure at baseline than noncases. They
more often reported a history of hypertension but were less likely to
be receiving current blood pressure treatment (47.0% versus 73.0%) if
presence of hypertension was known. Of all study participants, 10.5%
were newly diagnosed with hypertension, following the World Health
Organization definition of hypertension (15.4% in cases, 10.5% in
noncases) at baseline. Additionally, borderline isolated
systolic hypertension (defined as systolic blood
pressure between 140 to 159 mm Hg and diastolic blood
pressure <90 mm Hg) was diagnosed in 10.3% of cases and 6.6%
of noncases. Differences in smoking status, history of diabetes
mellitus, and family history of stroke were of borderline significance.
Considerable differences were also observed in lipid profiles of study
participants. Those who subsequently developed a stroke had, in
general, more unfavorable profiles than those who did not, most
markedly with respect to total cholesterol.
View this table:
[in a new window]
Table 1. Baseline Characteristics of Participants (Men Aged
30–65 Years) in the PROCAM Study According to Stroke Case Status ) were observed in
several variables such as obesity, alcohol consumption, and
hypertension treatment rate. Additionally, small differences were found
in systolic blood pressure, body mass index (BMI), and total
cholesterol.
View this table:
[in a new window]
Table 2. Risk Factor Comparison1
Between Participants (Men
Aged 30–65 Years) in the PROCAM Study (Occupational Cohort) and the
Augsburg MONICA Survey 1984/1985 (General Population) ). Prominent increases were especially
observed for elevated systolic and diastolic blood
pressure.
View this table:
[in a new window]
Table 3. Age-Adjusted Rates per 100 000 Person-Years for
Total Stroke During 7.2 Years of Follow-Up in Different Categories of
Major Stroke Risk Factors ). In 2
different models, risks for systolic and diastolic
blood pressure were calculated separately. Risk magnitudes were lower
for diastolic blood pressure (RR, 2.20 [95% confidence
interval, 0.89 to 5.48] for 81 to 90 mm Hg; RR, 3.59 [95%
confidence interval, 1.39 to 9.26] for 91 mm Hg), whereas the
risk magnitudes of the other variables in both models (smoking,
history of hypertension, diabetes mellitus) did not change. Known
history of hypertension at baseline was found to be an independent risk
factor for total stroke of considerable magnitude. Significant trends
of risk increase were observed across categories of systolic
and diastolic blood pressure as well as smoking status.
View this table:
[in a new window]
Table 4. Major Risk Factors for Total Stroke During 7.2 Years
of Follow-Up in the PROCAM Study
Discussion
Top
Abstract
Introduction
Subjects and Methods
Results
Discussion
References
In this cohort of 12 866 working men from different industry
branches and with different job training, a low overall incidence of
stroke was observed. Strong increases in event rates across categories
of various potential risk factors were found. Regression
analysis revealed that systolic blood pressure and
current smoking represented the strongest risks for total
stroke, underscoring their known status as most important risk
factors.17 18 19 20 21 140 mm Hg) systolic blood
pressure. They had a 5-fold risk of stroke compared with those with
pressure <120 mm Hg. This is remarkable because the former
category included a considerable number (14%) of individuals with
so-called borderline isolated systolic hypertension, for which
only moderate risk magnitudes have been
described.29
Acknowledgments
The PROCAM Study is funded by the Ministry of Science and
Research of Northrhine-Westfalia and the Public Health Insurance
Societies of Westphalia and Rhineland.
References
Top
Abstract
Introduction
Subjects and Methods
Results
Discussion
References
This article has been cited by other articles:
 |
|
 |
|
  E. Touze and P. M. Rothwell Sex Differences in Heritability of Ischemic Stroke: A Systematic Review and Meta-Analysis Stroke, January 1, 2008; 39(1): 16 - 23. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. K.J. Oksala, M. Heikkinen, J. Mikkelsson, T. Pohjasvaara, M. Kaste, T. Erkinjuntti, and P. J. Karhunen Smoking and the Platelet Fibrinogen Receptor Glycoprotein IIb/IIIA PlA1/A2 Polymorphism Interact in the Risk of Lacunar Stroke and Midterm Survival Stroke, January 1, 2007; 38(1): 50 - 55. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. I. Paraskevas, S. S. Daskalopoulou, M. E. Daskalopoulos, and C. D. Liapis Secondary Prevention of Ischemic Cerebrovascular Disease. What Is the Evidence? Angiology, September 1, 2005; 56(5): 539 - 552. [Abstract] [PDF]
|
|
|
|
 |
|
 G. Assmann Calculating global risk: the key to intervention Eur. Heart J. Suppl., July 1, 2005; 7(suppl_F): F9 - F14. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Konrad, C. Langer, G.A. Muller, K. Berger, R. Dziewas, F. Stogbauer, D.G. Nabavi, R. Junker, E.B. Ringelstein, and G. Kuhlenbaumer Protease Inhibitors in Spontaneous Cervical Artery Dissections Stroke, January 1, 2005; 36(1): 9 - 13. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Flossmann, U. G.R. Schulz, and P. M. Rothwell Systematic Review of Methods and Results of Studies of the Genetic Epidemiology of Ischemic Stroke Stroke, January 1, 2004; 35(1): 212 - 227. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Asia Pacific Cohort Studies Collaboration Cholesterol, coronary heart disease, and stroke in the Asia Pacific region Int. J. Epidemiol., August 1, 2003; 32(4): 563 - 572. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. M Dean, C. L. Richards, and F. Malouin Walking speed over 10 metres overestimates locomotor capacity after stroke Clinical Rehabilitation, April 1, 2001; 15(4): 415 - 421. [Abstract] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||