status and plasma fibrinogen; the level of plasma fibrinogen is higher among lower socioeconomic groups
[7,14,15]. However, it is uncertain whether socioeco- nomic status contributes to the increase of plasma
fibrinogen after adjustment for other determinants of plasma fibrinogen [16,17]. Among these other determi-
nants of high plasma fibrinogen level are winter season, old age, menopause, adverse early-life circumstances,
genetic factors and physical conditions such as inflam- mation, obesity, diabetes and hypertension. Further-
more,
life-style factors
such as
low alcohol
consumption, lack of physical exercise, poor diet and especially smoking, as well as adverse psychosocial
environment can
contribute to
increased plasma
fibrinogen level [18,19]. The purpose of this study is to investigate if a relationship between socioeconomic
status and plasma fibrinogen exists in Japanese male employees, and to assess whether this possible relation-
ship is independent of other determinants of fibrinogen.
2. Method
2
.
1
. Study population The cross-sectional study was conducted in a metal-
products factory in Japan. We analyzed full-time white and blue collar male employees aged 20 – 58 registered
in the factory as of 1st May in 1996, who underwent a medical checkup during the period between April 1996
and March 1997. The study excluded junior employees who had worked for less than 1 year, short-term em-
ployees from recruitment agencies and a few directors, as well as 43 subjects who underwent a medical
checkup at another medical facility and four subjects whose plasma fibrinogen data were not available. Alto-
gether 4375 males took part in the study and the participation rate was 92.9. We were unable to
analyse females in this study because of very little variation in their employment grade.
2
.
2
. Anthropometric measurements, blood pressure and blood analysis
Weight, height, and waist and hip circumferences were measured with subjects dressed in the factory
uniform made of thin cloth. The waist circumstance at umbilicus level and the hip circumference at the widest
level were measured by experienced nurses. Height was used as a possible marker of early life influences [20].
Blood pressure was measured after the subject had rested on a chair for 5 min or longer. Non-fasting
venous blood samples were collected into siliconized tubes containing 3.8 sodium citrate, and were
analysed for plasma fibrinogen by the Clauss method. We also measured white blood cell count WBC by an
automated cell counter SYSMEX 9421, Japan and hemoglobin A1c HbA1c by high-velocity liquid chro-
matography. Blood samples were analyzed in two labo- ratories; 96 and 4 of the samples, respectively. There
were no significant differences in the mean values of the blood analyses between these laboratories.
2
.
3
. Questionnaire A self-administered questionnaire was used to collect
information about life-style, including smoking habit, alcohol consumption and physical activity at leisure.
Smoking habit was classified as ‘non-smokers’, ‘ex- smokers’, ‘smokers of less than 21 cigarettes a day’ and
‘smokers of 21 or more cigarettes a day’. Alcohol consumption was measured in terms of grams of
ethanol consumed per week, and was based on quantity of specified drink consumed weekly. Physical activity at
leisure was classified as ‘almost no exercise’, ‘light exer- cise per week’, ‘brisk and sweating exercise once or
twice per week’ and ‘brisk and sweating exercise more than two times per week’.
2
.
4
. Employment grade and educational background Employment grade and educational background were
based on data obtained from the factory records. Nine employment grades that depended on salary existed.
The highest three grades in this population consisted of managers, and the third grade included those who were
at supervisory positions. As the number of subjects in the highest two grades was only 27, the subjects of the
highest three employment grades were put together into one grade. Other employees were classified into six
hierarchical grades. The basic annual salary was 9 450 000 12 350 000 yen in the combined highest
grade and 4 450 000 yen in the lowest grade. Educa- tional background was determined in terms of the total
years of education, and was classified as ‘more than 15 years’, ‘13 – 15 years’, ‘11 – 12 years’ and ‘less than 11
years’ of education.
2
.
5
. Statistical analysis Regression analyses were used to adjust for age and
other explanatory factors. Linear regression analyses were performed with the logarithm of plasma fibrino-
gen as the dependent variable. Logistic regression analyses were used to calculate odds ratios of having
plasma fibrinogen above the fifth quintile, 2.85 gl in these subjects. These odds ratios show the relative
change in odds odds = P1 − P, where P is the prob- ability of having fibrinogen above the fifth quintile for
a unit change in the explanatory variable. All calcula- tions were carried out using the SAS software 6.12.
Additionally, as the mean fibrinogen level of subjects in
winter December – February was not significantly dif- ferent from those in other seasons, we did not take
season into account as a covariate.
3. Results