American Journal of Eptdemiotogy Copyright C 1992 by The Johns Hopkins University School of Hygiene and Pubic Health All rights reserved
Vol. 136, No. 5 Printed in U S A.
Leukocyte Count Correlates in Middle-aged Adults: The Atherosclerosis Risk in Communities (ARIC) Study
F. Javier Nieto,1 Moyses Szklo,1 Aaron R. Folsom,2 Robert Rock,3 and Michele Mercuri4 for the ARIC Investigators
Cross-sectional associations between leukocyte count and sociodemographic and cardiovascular risk factors were investigated in 14,679 participants aged 45-64 years in the Atherosclerosis Risk in Communities Study carried out in four US communities in 1986-1989. Leukocyte count was strongly associated with present or past history of cigarette smoking and was higher in males than in females and in white subjects than in black subjects. Among never smokers, no sex differences were evident after adjustment for other risk factors. Race-associated differences were substantially reduced after other factors were taken into account in multivariate analyses. In never smokers, leukocyte count was higher in those who reported poor health, and it was inversely associated with high density lipoprotein cholesterol, forced expiratory volume at 1 second, physical activity, and, among whites, height and socioeconomic indicators. It was directly associated with indices of body weight and body fat, heart rate, Wood pressure, hemoglobin, platelet count, uric acid, fasting insulin and glycemia, triglycerides, fibrinogen, antithrombin III, protein C, factors VII and VIII, and von Willebrand factor. The associations of leukocyte count with cardiovascular risk factors may either represent manifestation of subclinical disease or suggest that leukocyte count is part of the causal chain leading to atherosclerosis. Alternatively, the relation of leukocyte count to cardiovascular disease may be confounded by risk factors and thus be noncausal. Am J Epidemiol 1992;136:525-37. atherosclerosis; cardiovascular diseases; coronary disease; ethnic groups; hemostasis; leukocyte count; risk factors
Since Friedman et al. (I) reported in 1974 that leukocyte count was a significant predictor of myocardial infarction, several studies have shown direct associations between
leukocyte count and death from all causes (2), coronary heart disease incidence and mortality (3-9), cerebral infarction (10), cancer mortality (3, 11, 12), and incidence
Received for publication December 23, 1991, and in final form April 3, 1992. Abbreviations: ARIC, Atherosclerosis Risk in Communities, FEV,, forced expiratory volume in 1 second; HDL cholesterol, high density lipoprotein cholesterol ' Department of Epidemiology, School of Hygiene and Public Health, The Johns Hopkins University, Baltimore, MD. 2 Division of Eptdemiotogy, School of Public Health, University of Minnesota, Minneapolis, MN 3 Department of Laboratory Medicine, The Johns Hopkins Hospital, Baltimore, MD. 4 Division of Vascular Ultrasound Research, Bowman Gray School of Medicine, Winston-Salem, NC. Reprint requests to Dr F. Javier Nieto, Department of Epidemiology, School of Hygiene and Public Health, The Johns Hopkins University, 615 North Wolfe Street, Baltimore, MD 21205 The Atherosclerosis Risk in Communities Study is carried out as a collaborative study supported by National
Heart, Lung, and Blood Institute contracts N01-HC-55015, M01-HC-55016, N01-HC-55018, N01-HC-55019, N01-HC55020, N01-HC-55021, and N01-HC-55022. The authors acknowledge the contributions of the staff at the following Atherosclerosis Risk in Communities centers- Project Office, National Heart, Lung, and Blood Institute, Bethesda, MD—Donna J. Neal and Patricia A. Smith; Coordinating Center, University of North Carolina, Chapel Hill, NC—Cynthia Stahoviak, Myra Carpenter, Deborah Rubin-Williams, and Yi-Hsin Yang; Jackson Field Center, Jackson, MS-Bobbte Alhston, Jane Johnson, Dr. Qinton Smith, and Dr. Robert Smith; Forsyth County Field Center, University of North Carolina, Chapel Hill, NC—Regina DeLacy, Barbara Owens, Carmen Woody, and Billie Young; Minneapolis Field Center, University of Minnesota, Minneapolis, MN—Jennifer Hill, Ellie Justiniano, Byrna Lester, and Gail Murton; Washington County Reid Center, The Johns Hopkins University, Hagerstown, MD—Joyce Chabot, Carol Chnstman, Joel Hill, and Joan Netting.
526
Nieto et al.
of clinical hypertension (13). The biological significance of these associations is controversial. Mechanisms whereby leukocyte count may be a factor in the causal pathway of atherosclerosis and hypertension have been proposed (14). Leukocyte count may also be a marker for another risk factor, cigarette smoking, which is a strong determinant of most of the health outcomes listed above and is also strongly correlated to leukocyte count (15-17). However, most of the current evidence shows an independent value of leukocyte count as a predictor of cardiovascular disease, regardless of smoking status and other risk factors (3-5, 14). Other risk factors for atherosclerosis found to correlate with leukocyte count include high density lipoprotein cholesterol (HDL cholesterol) (inverse association), triglycerides (3, 17, 18), total cholesterol, and uric acid (6). Data from the first National Health and Nutrition Examination Survey (19) show higher leukocyte counts in whites than in blacks, regardless of sex and age, and higher leukocyte counts in young women and in older men, with the crossover occurring around age 40 years. Oral contraceptive use, short stature, body mass index, heart rate, low physical fitness, and systolic blood pressure have also been found to be associated with a higher leukocyte count (17, 18). The objectives of this report were twofold: first, to show the distribution of leukocyte count in adults aged 45-64 years in population samples obtained in four US communities, as reference for comparison with other populations, and second, to further study leukocyte count correlates, which may assist in elucidating the current controversy regarding the pathogenetic mechanisms linking leukocyte count to atherosclerosisrelated diseases. Some of the correlates of leukocyte count reported herein, including hemostatic factors and others, have never been reported before to our knowledge. MATERIALS AND METHODS Study population
The main goals of the Atherosclerosis Risk in Communities (ARIC) Study are to inves-
tigate the etiology and natural history of subclinical and clinical atherosclerosis. The study population consists of 15,800 subjects aged 45-64 years of both sexes, randomly selected from four communities in the US (approximately one fourth from each): selected Minneapolis, Minnesota, suburbs; Washington County, Maryland; Forsyth County, North Carolina; and Jackson, Mississippi. The first two communities are predominantly white. Blacks were exclusively sampled in Jackson and comprised 12 percent of the study sample from Forsyth County. The study objectives, design, sampling procedures, and examination procedures have been described in detail elsewhere (20). Briefly, a baseline interview and clinical examination were carried out between the fall of 1986 and the winter of 1989; participants have been reinterviewed annually by telephone and are currently being reexamined (3 years after the baseline visit). For the present cross-sectional analyses, only data from the baseline examination were used. Subjects of races other than white or black, subjects taking medication which induces hemoconcentration, subjects from whom blood samples were not drawn, and those with a leukocyte count of lower than 2,000 per mm3 or higher than 15,000 per mm3 were excluded. A total of 14,679 subjects were included in the present analyses. Examination procedures
The baseline examination took place in the study clinic at each community after a home interview by trained interviewers. Information on education (highest grade completed), occupation, income, cigarette smoking, ethanol intake, physical activity, the presence of diagnosed cardiovascular disease or diabetes, menopausal status, and estrogen replacement therapy was obtained from the interview. "Passive smoking" was estimated from the reported number of hours per week that the subject was "around smokers." Occupation was coded by following the 1980 US Census criteria (21), and six occupational categories were defined. Women who
Leukocyte Count Correlates
reported no menstrual periods in the last 2 years were classified as postmenopausal. Three indices of physical activity (work, sports, and leisure) were estimated from a modified version of the questionnaire by Baeckeetal. (21,22). Blood pressure measurements, anthropometry, and venipuncture were carried out after subjects had fasted for at least 12 hours. Sitting blood pressure was measured before venipuncture; systolic pressure was calculated as the average of the second and third of three consecutive measurements with a random zero sphygmomanometer. Anthropometric measurements were carried out with subjects wearing a scrub suit and no shoes. Body mass index was calculated as weight (kg)/height (m)2. Triceps and subscapular skinfolds were measured as the average of two consecutive measurements using a Lange caliper. Heart rate was estimated from the average rate in the 2-minute electrocardiogram rhythm strip. Pulmonary function tests were carried out using a Collins Survey II volume displacement spirometer (William E. Collins Co.). Blood specimens were drawn and processed following a protocol described in detail elsewhere (23). Leukocyte count and the other hematologic parameters were determined in hospital-based, independent laboratories within 24 hours after venipuncture, after storage at 4°C, using automated particle counters (Coulter Diagnostics, Hialeah, Florida): in Jackson, Coulter S+IV, calibration S-Cal; in Washington County, at two laboratories, both using Coulter S+IV, calibration S-Cal; in Minneapolis, at one laboratory using two different counters, Coulter S+III and Coulter S+IV, calibration S-Cal; in Forsyth County, a Technicon H-6000 (Technicon Corporation, Tarrytown, New York), calibration Fisher ("Computrol"). Quality control replicate data for leukocyte count resulted in high reliability coefficients, ranging between 0.96 and 1.00 for all laboratories. All of the remaining aliquots were centrifugated, frozen at —70°C, and shipped to the ARIC central laboratories. Lipids, chemis-
527
try, and hemostatic factors were measured by standard procedures as previously described (20, 24). Statistical methods
Means, standard deviations, and selected percentiles of leukocyte count according to potential covariates, were calculated. For continuous variables, categories were defined as quartiles according to the distribution across the entire study population. Given the strong association between leukocyte count and cigarette smoking, most analyses were carried out in never smokers. Race-specific, age-, sex- and center-adjusted means were obtained by analysis of covariance (25), with leukocyte count as the dependent variable. For ordinal variables and quartiles of continuous variables, a trend test consisted of assessing the significance of the regression coefficient corresponding to the ordinal variable in the model. Variables that showed significant associations with leukocyte count in the previous analyses were included in race-specific multiple linear regression models. These analyses were carried out in three phases: first, including only sociodemographic, life-style, and constitutional factors (model 1); second, adding hematologic, lipid, and biochemical variables (model 2); and finally, including hemostatic factors as well (model 3). Only subjects without missing values for all variables in model 3 were included in these analyses. The statistical significance of regression variables was evaluated using an F test, comparing the model without the given variable (null hypothesis) with the model including the variable (alternative hypothesis) (26). Version 6.06 of the Statistical Analysis System was used for all statistical analyses (27). RESULTS
The distribution of leukocyte count was approximately lognormal. The sex-race-age composition of the study group, as well as the mean, standard deviation, and selected percentiles of leukocyte count according to sex, race, and two age categories are shown
528
Nieto et al.
TABLE 1. Mean, standard deviation (SD), and selected percentiie (P) values of leukocyte count (in thousands per mm1) by sex, race, and age, the Atherosclerosis Risk in Communities Study, 1986-1989 Race end age (years)
Total
No.
14,679
Mean
6.1
SO
P5
P10
P25
P50
P75
P90
P95
(1.8)
3.6
4.0
4.8
5.8
7.0
8.5
9.6
Males White 45-54 55-64 Black 45-54 55-64
2,404 2,519
6.3 6.4
(1.8) (1.8)
4.0 4.1
4.4 4.5
5.1 5.2
6.0 6.2
7.2 7.4
8.7 87
9.9 9.9
833 665
5.6 5.6
(1.9) (1.9)
3.2 3.2
3.6 3.5
4.3 4.3
5.2 5.2
6.5 6.6
8.2 8.1
9.2 9.1
Females White 45-54 55-64 Black 45-54 55-64
3,177 2,643
6.2 6.1
(1.9) (1.7)
3.7 3.8
4.1 4.2
4.9 4.8
5.9 5.8
7.2 7.0
8.6 8.3
9.7 9.4
1,459 979
5.6 5.6
(1.8) (1.8)
3.3 3.2
3.7 3.6
4.4 4.4
5.4 5.4
6.6 6.5
7.9 7.8
9.0 8.9
in table 1. Leukocyte count was higher in whites than in blacks. No age differences were seen in blacks, whereas in whites, leukocyte count increased slightly with age in males and decreased slightly in females. No differences according to sex were evident among blacks, whereas in whites, leukocyte count was higher in males, particularly in the older age group. Because leukocyte count determinations were carried out in different laboratories at each field center and because of the different race distributions by center, these analyses were repeated including only participants from Forsyth County, a cohort that included both blacks and whites. A similar pattern according to race was evident: for males, the mean leukocyte count (in thousands per mm3) was 6.6 for whites and 5.8 for blacks, and for females, it was 6.2 and 5.7 for whites and blacks, respectively. Age-, sex-, and center-adjusted mean leukocyte counts (in thousands per mm3) were 5.7, 5.5, 5.5, and 5.6 for winter, spring, summer, and fall in blacks (p = 0.14) and 6.3, 6.1, 6.3, and 6.3 for winter, spring, summer, and fall in whites (p = 0.01), respectively. Table 2 shows race-specific age-, sex-, and center-adjusted mean leukocyte counts ac-
cording to different cigarette smoking categorizations. Strong statistically significant trends of increasing leukocyte count with increasing amount and duration of smoking were evident for most smoking variables. The trend according to degree of inhalation fell short of statistical significance among blacks, but was in the same direction as that among whites. Among never smokers, no clear trends were observed according to number of hours per week spent around smokers, with the exception of an increased mean leukocyte count in whites in the upper quartile. No consistent trends of leukocyte count according to cigar smoking or pipe smoking were observed (data not shown). Tables 3-7 present age-, sex-, and centeradjusted mean leukocyte counts according to a variety of sociodemographic, life-style, constitutional, and laboratory variables, stratified by race. These analyses were carried out in never smokers only. For socioeconomic and life-style characteristics (table 3), mean leukocyte count was significantly higher in the lowest categories of education and income only among whites. No trend according to occupation was evident. No consistent trends according to alcohol intake were observed. With the exception of the sport activity index in
Leukocyte Count Correlates
TABLE 2. Race-specific age-, sex-, and centeradjustedt mean values of leukocyte count (in thousands per mm1) by cigarette smoking status, the Atherosclerosis Risk in Communities Study, 1986-1989 Whites
Blacks
TABLE 3. Race-specific age-, sex-, and centeradjustedt mean values of leukocyte count (in thousands per mm9) by socioeconomic and lifestyle characteristics, never smokers, the Atherosclerosis Risk in Communities Study, 19861989 Whites (n = 4,832)
Smoking status}: Never Former Current Cigarettes/day§
Vol. 136, No. 5 Printed in U S A.
Leukocyte Count Correlates in Middle-aged Adults: The Atherosclerosis Risk in Communities (ARIC) Study
F. Javier Nieto,1 Moyses Szklo,1 Aaron R. Folsom,2 Robert Rock,3 and Michele Mercuri4 for the ARIC Investigators
Cross-sectional associations between leukocyte count and sociodemographic and cardiovascular risk factors were investigated in 14,679 participants aged 45-64 years in the Atherosclerosis Risk in Communities Study carried out in four US communities in 1986-1989. Leukocyte count was strongly associated with present or past history of cigarette smoking and was higher in males than in females and in white subjects than in black subjects. Among never smokers, no sex differences were evident after adjustment for other risk factors. Race-associated differences were substantially reduced after other factors were taken into account in multivariate analyses. In never smokers, leukocyte count was higher in those who reported poor health, and it was inversely associated with high density lipoprotein cholesterol, forced expiratory volume at 1 second, physical activity, and, among whites, height and socioeconomic indicators. It was directly associated with indices of body weight and body fat, heart rate, Wood pressure, hemoglobin, platelet count, uric acid, fasting insulin and glycemia, triglycerides, fibrinogen, antithrombin III, protein C, factors VII and VIII, and von Willebrand factor. The associations of leukocyte count with cardiovascular risk factors may either represent manifestation of subclinical disease or suggest that leukocyte count is part of the causal chain leading to atherosclerosis. Alternatively, the relation of leukocyte count to cardiovascular disease may be confounded by risk factors and thus be noncausal. Am J Epidemiol 1992;136:525-37. atherosclerosis; cardiovascular diseases; coronary disease; ethnic groups; hemostasis; leukocyte count; risk factors
Since Friedman et al. (I) reported in 1974 that leukocyte count was a significant predictor of myocardial infarction, several studies have shown direct associations between
leukocyte count and death from all causes (2), coronary heart disease incidence and mortality (3-9), cerebral infarction (10), cancer mortality (3, 11, 12), and incidence
Received for publication December 23, 1991, and in final form April 3, 1992. Abbreviations: ARIC, Atherosclerosis Risk in Communities, FEV,, forced expiratory volume in 1 second; HDL cholesterol, high density lipoprotein cholesterol ' Department of Epidemiology, School of Hygiene and Public Health, The Johns Hopkins University, Baltimore, MD. 2 Division of Eptdemiotogy, School of Public Health, University of Minnesota, Minneapolis, MN 3 Department of Laboratory Medicine, The Johns Hopkins Hospital, Baltimore, MD. 4 Division of Vascular Ultrasound Research, Bowman Gray School of Medicine, Winston-Salem, NC. Reprint requests to Dr F. Javier Nieto, Department of Epidemiology, School of Hygiene and Public Health, The Johns Hopkins University, 615 North Wolfe Street, Baltimore, MD 21205 The Atherosclerosis Risk in Communities Study is carried out as a collaborative study supported by National
Heart, Lung, and Blood Institute contracts N01-HC-55015, M01-HC-55016, N01-HC-55018, N01-HC-55019, N01-HC55020, N01-HC-55021, and N01-HC-55022. The authors acknowledge the contributions of the staff at the following Atherosclerosis Risk in Communities centers- Project Office, National Heart, Lung, and Blood Institute, Bethesda, MD—Donna J. Neal and Patricia A. Smith; Coordinating Center, University of North Carolina, Chapel Hill, NC—Cynthia Stahoviak, Myra Carpenter, Deborah Rubin-Williams, and Yi-Hsin Yang; Jackson Field Center, Jackson, MS-Bobbte Alhston, Jane Johnson, Dr. Qinton Smith, and Dr. Robert Smith; Forsyth County Field Center, University of North Carolina, Chapel Hill, NC—Regina DeLacy, Barbara Owens, Carmen Woody, and Billie Young; Minneapolis Field Center, University of Minnesota, Minneapolis, MN—Jennifer Hill, Ellie Justiniano, Byrna Lester, and Gail Murton; Washington County Reid Center, The Johns Hopkins University, Hagerstown, MD—Joyce Chabot, Carol Chnstman, Joel Hill, and Joan Netting.
526
Nieto et al.
of clinical hypertension (13). The biological significance of these associations is controversial. Mechanisms whereby leukocyte count may be a factor in the causal pathway of atherosclerosis and hypertension have been proposed (14). Leukocyte count may also be a marker for another risk factor, cigarette smoking, which is a strong determinant of most of the health outcomes listed above and is also strongly correlated to leukocyte count (15-17). However, most of the current evidence shows an independent value of leukocyte count as a predictor of cardiovascular disease, regardless of smoking status and other risk factors (3-5, 14). Other risk factors for atherosclerosis found to correlate with leukocyte count include high density lipoprotein cholesterol (HDL cholesterol) (inverse association), triglycerides (3, 17, 18), total cholesterol, and uric acid (6). Data from the first National Health and Nutrition Examination Survey (19) show higher leukocyte counts in whites than in blacks, regardless of sex and age, and higher leukocyte counts in young women and in older men, with the crossover occurring around age 40 years. Oral contraceptive use, short stature, body mass index, heart rate, low physical fitness, and systolic blood pressure have also been found to be associated with a higher leukocyte count (17, 18). The objectives of this report were twofold: first, to show the distribution of leukocyte count in adults aged 45-64 years in population samples obtained in four US communities, as reference for comparison with other populations, and second, to further study leukocyte count correlates, which may assist in elucidating the current controversy regarding the pathogenetic mechanisms linking leukocyte count to atherosclerosisrelated diseases. Some of the correlates of leukocyte count reported herein, including hemostatic factors and others, have never been reported before to our knowledge. MATERIALS AND METHODS Study population
The main goals of the Atherosclerosis Risk in Communities (ARIC) Study are to inves-
tigate the etiology and natural history of subclinical and clinical atherosclerosis. The study population consists of 15,800 subjects aged 45-64 years of both sexes, randomly selected from four communities in the US (approximately one fourth from each): selected Minneapolis, Minnesota, suburbs; Washington County, Maryland; Forsyth County, North Carolina; and Jackson, Mississippi. The first two communities are predominantly white. Blacks were exclusively sampled in Jackson and comprised 12 percent of the study sample from Forsyth County. The study objectives, design, sampling procedures, and examination procedures have been described in detail elsewhere (20). Briefly, a baseline interview and clinical examination were carried out between the fall of 1986 and the winter of 1989; participants have been reinterviewed annually by telephone and are currently being reexamined (3 years after the baseline visit). For the present cross-sectional analyses, only data from the baseline examination were used. Subjects of races other than white or black, subjects taking medication which induces hemoconcentration, subjects from whom blood samples were not drawn, and those with a leukocyte count of lower than 2,000 per mm3 or higher than 15,000 per mm3 were excluded. A total of 14,679 subjects were included in the present analyses. Examination procedures
The baseline examination took place in the study clinic at each community after a home interview by trained interviewers. Information on education (highest grade completed), occupation, income, cigarette smoking, ethanol intake, physical activity, the presence of diagnosed cardiovascular disease or diabetes, menopausal status, and estrogen replacement therapy was obtained from the interview. "Passive smoking" was estimated from the reported number of hours per week that the subject was "around smokers." Occupation was coded by following the 1980 US Census criteria (21), and six occupational categories were defined. Women who
Leukocyte Count Correlates
reported no menstrual periods in the last 2 years were classified as postmenopausal. Three indices of physical activity (work, sports, and leisure) were estimated from a modified version of the questionnaire by Baeckeetal. (21,22). Blood pressure measurements, anthropometry, and venipuncture were carried out after subjects had fasted for at least 12 hours. Sitting blood pressure was measured before venipuncture; systolic pressure was calculated as the average of the second and third of three consecutive measurements with a random zero sphygmomanometer. Anthropometric measurements were carried out with subjects wearing a scrub suit and no shoes. Body mass index was calculated as weight (kg)/height (m)2. Triceps and subscapular skinfolds were measured as the average of two consecutive measurements using a Lange caliper. Heart rate was estimated from the average rate in the 2-minute electrocardiogram rhythm strip. Pulmonary function tests were carried out using a Collins Survey II volume displacement spirometer (William E. Collins Co.). Blood specimens were drawn and processed following a protocol described in detail elsewhere (23). Leukocyte count and the other hematologic parameters were determined in hospital-based, independent laboratories within 24 hours after venipuncture, after storage at 4°C, using automated particle counters (Coulter Diagnostics, Hialeah, Florida): in Jackson, Coulter S+IV, calibration S-Cal; in Washington County, at two laboratories, both using Coulter S+IV, calibration S-Cal; in Minneapolis, at one laboratory using two different counters, Coulter S+III and Coulter S+IV, calibration S-Cal; in Forsyth County, a Technicon H-6000 (Technicon Corporation, Tarrytown, New York), calibration Fisher ("Computrol"). Quality control replicate data for leukocyte count resulted in high reliability coefficients, ranging between 0.96 and 1.00 for all laboratories. All of the remaining aliquots were centrifugated, frozen at —70°C, and shipped to the ARIC central laboratories. Lipids, chemis-
527
try, and hemostatic factors were measured by standard procedures as previously described (20, 24). Statistical methods
Means, standard deviations, and selected percentiles of leukocyte count according to potential covariates, were calculated. For continuous variables, categories were defined as quartiles according to the distribution across the entire study population. Given the strong association between leukocyte count and cigarette smoking, most analyses were carried out in never smokers. Race-specific, age-, sex- and center-adjusted means were obtained by analysis of covariance (25), with leukocyte count as the dependent variable. For ordinal variables and quartiles of continuous variables, a trend test consisted of assessing the significance of the regression coefficient corresponding to the ordinal variable in the model. Variables that showed significant associations with leukocyte count in the previous analyses were included in race-specific multiple linear regression models. These analyses were carried out in three phases: first, including only sociodemographic, life-style, and constitutional factors (model 1); second, adding hematologic, lipid, and biochemical variables (model 2); and finally, including hemostatic factors as well (model 3). Only subjects without missing values for all variables in model 3 were included in these analyses. The statistical significance of regression variables was evaluated using an F test, comparing the model without the given variable (null hypothesis) with the model including the variable (alternative hypothesis) (26). Version 6.06 of the Statistical Analysis System was used for all statistical analyses (27). RESULTS
The distribution of leukocyte count was approximately lognormal. The sex-race-age composition of the study group, as well as the mean, standard deviation, and selected percentiles of leukocyte count according to sex, race, and two age categories are shown
528
Nieto et al.
TABLE 1. Mean, standard deviation (SD), and selected percentiie (P) values of leukocyte count (in thousands per mm1) by sex, race, and age, the Atherosclerosis Risk in Communities Study, 1986-1989 Race end age (years)
Total
No.
14,679
Mean
6.1
SO
P5
P10
P25
P50
P75
P90
P95
(1.8)
3.6
4.0
4.8
5.8
7.0
8.5
9.6
Males White 45-54 55-64 Black 45-54 55-64
2,404 2,519
6.3 6.4
(1.8) (1.8)
4.0 4.1
4.4 4.5
5.1 5.2
6.0 6.2
7.2 7.4
8.7 87
9.9 9.9
833 665
5.6 5.6
(1.9) (1.9)
3.2 3.2
3.6 3.5
4.3 4.3
5.2 5.2
6.5 6.6
8.2 8.1
9.2 9.1
Females White 45-54 55-64 Black 45-54 55-64
3,177 2,643
6.2 6.1
(1.9) (1.7)
3.7 3.8
4.1 4.2
4.9 4.8
5.9 5.8
7.2 7.0
8.6 8.3
9.7 9.4
1,459 979
5.6 5.6
(1.8) (1.8)
3.3 3.2
3.7 3.6
4.4 4.4
5.4 5.4
6.6 6.5
7.9 7.8
9.0 8.9
in table 1. Leukocyte count was higher in whites than in blacks. No age differences were seen in blacks, whereas in whites, leukocyte count increased slightly with age in males and decreased slightly in females. No differences according to sex were evident among blacks, whereas in whites, leukocyte count was higher in males, particularly in the older age group. Because leukocyte count determinations were carried out in different laboratories at each field center and because of the different race distributions by center, these analyses were repeated including only participants from Forsyth County, a cohort that included both blacks and whites. A similar pattern according to race was evident: for males, the mean leukocyte count (in thousands per mm3) was 6.6 for whites and 5.8 for blacks, and for females, it was 6.2 and 5.7 for whites and blacks, respectively. Age-, sex-, and center-adjusted mean leukocyte counts (in thousands per mm3) were 5.7, 5.5, 5.5, and 5.6 for winter, spring, summer, and fall in blacks (p = 0.14) and 6.3, 6.1, 6.3, and 6.3 for winter, spring, summer, and fall in whites (p = 0.01), respectively. Table 2 shows race-specific age-, sex-, and center-adjusted mean leukocyte counts ac-
cording to different cigarette smoking categorizations. Strong statistically significant trends of increasing leukocyte count with increasing amount and duration of smoking were evident for most smoking variables. The trend according to degree of inhalation fell short of statistical significance among blacks, but was in the same direction as that among whites. Among never smokers, no clear trends were observed according to number of hours per week spent around smokers, with the exception of an increased mean leukocyte count in whites in the upper quartile. No consistent trends of leukocyte count according to cigar smoking or pipe smoking were observed (data not shown). Tables 3-7 present age-, sex-, and centeradjusted mean leukocyte counts according to a variety of sociodemographic, life-style, constitutional, and laboratory variables, stratified by race. These analyses were carried out in never smokers only. For socioeconomic and life-style characteristics (table 3), mean leukocyte count was significantly higher in the lowest categories of education and income only among whites. No trend according to occupation was evident. No consistent trends according to alcohol intake were observed. With the exception of the sport activity index in
Leukocyte Count Correlates
TABLE 2. Race-specific age-, sex-, and centeradjustedt mean values of leukocyte count (in thousands per mm1) by cigarette smoking status, the Atherosclerosis Risk in Communities Study, 1986-1989 Whites
Blacks
TABLE 3. Race-specific age-, sex-, and centeradjustedt mean values of leukocyte count (in thousands per mm9) by socioeconomic and lifestyle characteristics, never smokers, the Atherosclerosis Risk in Communities Study, 19861989 Whites (n = 4,832)
Smoking status}: Never Former Current Cigarettes/day§
