American Journal of Epidemiology
Copyright O 1996 by The Johns Hopkins University School of Hygiene and Public Health
All rights reserved
Vol 144, No. 8
Printed In U.S A
Longitudinal Relation between Smoking and White Blood Cells
Jordi Sunyer, 12 Alvaro Munoz,1 Yun Peng,1 Joseph Margolick,1 Joan S. Chmiel,3 John Oishi,4
Lawrence Kingsley,3 and Jonathan M. Samet1
Higher white blood cell counts in smokers compared with nonsmokers have been well documented, but the
longitudinal relation between changes in smoking and changes in white blood cells has not been well
described. Since 1984, data have been collected semiannually by the Multicenter AIDS Cohort Study (MACS),
a four-center prospective cohort study of acquired immunodeficiency syndrome (AIDS) in homosexual men.
The study population includes 2,435 participants who were human immunodeficiency virus (HIV) seronegative
as of September 1994 and who contributed 20,918 person-visits for this analysis. For individuals who modified
their smoking behavior, changes in white blood cell counts occurred primarily during the first 6 months
following changes in the amount of cigarettes smoked. Among former smokers who resumed smoking, the
extent of the increase in white blood cell count depended on the number of cigarettes smoked. Specifically,
increases of 241,340, and 740 cells/pJiter were observed for smokers who resumed smoking < 1 , 1 to <2, and
^ 2 packs/day, respectively. Conversely, smokers who quit smoking had a decrease of white blood cell count:
- 3 2 , - 6 2 9 , and -1,122 cells//xliter for men who previously smoked < 1 , 1 to <2, and a 2 packs/day,
respectively. Long-term ex-smokers, however, still had higher white blood cell counts than did never smokers.
There was a high within-individual correlation of white blood cell count in persons who reported a consistent
level of smoking (i.e., average correlations between two white blood cell counts 6 years apart were 0.51 for
never smokers, 0.48 for ex-smokers, 0.56 for men who smoked <1 pack/day, and 0.43 for men who smoked
^ 1 pack/day). These analyses indicate an acute effect of changes in smoking on changes in white blood cell
count, a residual effect of having been a smoker, and high long-term tracking for white blood cell count. Am J
Epidemiol 1996;144:734-^1.
leukocytes; longitudinal studies; smoking
White blood cell count has been associated with
coronary heart disease (1-5), respiratory function and
symptoms (6, 7), and cancer (2). Mechanisms underlying these associations have been proposed (2). The
total leukocyte count and counts of specific cell subtypes are plausible biomarkers of active atherogenesis
and other oxidative damage at the tissue level (8-11).
White blood cell count may also be a marker of
exposure to toxic substances. A relation between
smoking and white blood cell count has been long
established (12, 13); smokers have higher white blood
cell counts than nonsmokers, and the extent of the
increase rises with the number of cigarettes smoked.
The strength of the association has led to the suggestion that the white blood cell count provides a better
indicator of cigarette smoking than the self-reported
number of cigarettes (14). However, smoking explains
only part of the association between white blood cell
count and myocardial infarction (2, 3).
To use white blood cell count as a biomarker of
exposure and response, information is needed on its
variation over time and on the factors that determine
this variation. Short-term variation in white blood cell
count has been documented (15), but there is a lack of
information on long-term variation. For example, although experimental studies have shown that white
blood cell counts returned to the levels of never smokers less than 6 weeks after quitting smoking (16, 17),
cross-sectional studies, by contrast, have shown
slightly higher white blood cell counts in ex-smokers
than in never smokers (18, 19).
Received for publication October 31, 1995, and accepted for
publication May 8, 1996.
Abbreviations: AIDS, acquired Immunodeficiency syndrome; HIV,
human immunodeficiency vims; MACS, Multicenter AIDS Cohort
Study
1
Departments of Epidemiology and of Molecular Microbiology
and Immunology, The Johns Hopkins School of Public Health, Baltimore, MD.
2
Department of Epidemiology and Public Hearth, Institut Municipal Investgaci6 Medica, IMIM, Barcelona, Spain.
3
Department of Preventive Medicine, Northwestern University
Medical School, Chicago, I L
4
UCLA School of Public Hearth, Los Angeles, CA.
'Graduate School of Public Health, University of Pittsburgh,
Pittsburgh, PA.
Reprint requests to Dr. Alvaro Mufioz, Dept. of Epidemiology,
Johns Hopkins University School of Hygiene and Public Hearth, 624
N. Broadway, Baltimore, MD 21205.
734
Longitudinal Relation between Smoking and White Blood Cells
Studies on the dynamics of changes in smoking and
changes in leukocyte counts require repeat assessment
of both variables in the same individual over time. Our
objective was to characterize the temporal variation
in white blood cell count with changes in smoking
among individuals not infected with human immunodeficiency virus (HTV) in the Multicenter AIDS Cohort Study (MACS) (20), a prospective cohort study
of acquired immunodeficiency syndrome (AIDS) in
homosexual men. We characterize these dynamic relations taking into account the dependency of withinindividual measurements of white blood cell count.
Furthermore, we compare the intrinsic withinindividual correlation (tracking) of the repeated white
blood cell count observations according to patterns of
smoking.
MATERIALS AND METHODS
Study population and variables
A detailed description of the MACS population and
study design has been provided elsewhere (21).
Briefly, a cohort of homosexual/bisexual volunteer
men from four US sites, ATDS-free and older than age
18 years at entry, was followed semiannually. Seropositivity to HTV was determined at each visit by
enzyme-linked immunosorbent assay and confirmed
with Western blot. At entry, 2,191 of the participants
were seropositive, and of the 3,388 seronegative men,
485 had seroconverted as of September 1994. In this
paper, we focus on the participants who remained
seronegative during follow-up to September 1994.
Median age of this study population in 1994 was 43
years and 91 percent were white.
Although the first of the semiannual visits in the
MACS took place in 1984, we restricted attention to
data collected between visits 8 and 20 when a uniform
questionnaire on smoking was used. Hence, time of
follow-up spanned 6.5 years, from visit 8 in 1987—
1988 to the end of visit 20 in 1994. At each visit, blood
samples were collected for hematologic studies,
including complete blood counts and automated
differential, and for determination of percent and number of T-lymphocytes by flow cytometry. A comprehensive quality control program for flow cytometry
exists at the four study sites (22). Of the 2,519 HTVseronegative persons who attended visits 8-20, 2,435
provided at least one measurement of white blood cell
count and answered the smoking questionnaire. After
excluding white blood cell count values > 15,000
cells/jiliter, which are likely to be associated with
occurrence of acute infections, hematologic information was available for 20,918 person-visits over the
6.5-year period of follow-up. Specifically, the median
Am J Epidemiol
Vol. 144, No. 8, 1996
735
number of individuals seen at each visit was 1,849
(range from 1,892 at visit 9 to 1,121 at visit 20). At
each visit, smoking was categorized as never or ever.
For men who had smoked, daily cigarette consumption
was classified in six categories: not currently smoking
(as of one month ago), occasionally smoke (<1 cigarette/day), more than occasionally smoke but <0.5
pack/day, 0.5 to < 1 pack/day, 1 to < 2 packs/day, and
>2 packs/day.
Statistical analysis
Due to the small number of observations in the
categories of more than occasionally smoke to <1
pack/day, we combined them. Thus, we categorized
cigarette smoking as never, smoked in the past but not
currently (i.e., ex-smoker), occasional smoker, and
current smoker: < 1 pack/day, 1 to <2 packs/day, and
S:2 packs/day. Prior to the analysis, we used the longitudinal data on a given individual to adjust for inconsistencies in smoking history. In particular, we
made the following three adjustments: 1) a report of
never having smoked after a previous report of cigarette consumption was recoded as ex-smoker, 2) missing information on visits prior to a report of never
having smoked was recoded as never smoker, and 3) if
in two consecutive visits there was a change from
never to ex-smoker, then the ex-smoker was recoded
as an occasional smoker. Of the 20,918 person-visits,
we made adjustment 1 for 2,529 subjects, adjustment
2 for 27 subjects, and adjustment 3 for 151 subjects.
The longitudinal information on smoking minimizes
misclassification of the smoking status.
First, we assessed the relations between current
smoking, white blood cell count, and the composition
of white blood cells (i.e., neutrophils, basophils, eosinophils, monocytes, CD4 + T-cells, CD8 T-cells,
and non-T-lymphocytes calculated as the subtraction
of CD4 + and CD8 + T-cell counts from the lymphocyte counts) for all person-visits. Next, we determined
the net cross-sectional effects of smoking on the total
white blood cell count and components by analyzing
data from the participants with the same smoking
status over the full follow-up period by using linear
regression models for within-individual correlation of
repeated observations (23).
To assess the longitudinal association between
smoking and white blood cell count, we let changes in
white blood cell count between consecutive visit pairs
be the outcome measures and changes in smoking
status be the predictor variables (24). Because the
correlation structure of differences is complex and the
within-individual correlation of the differences was of
no prior interest, we used the generalized estimating
equation (25) method, with unstructured correlation
736
Sunyer et al.
matrix, for analysis. A model was fit for each smoking
category at the previous visit, and die model included
a covariate indicating the smoking category of me next
visit. To determine if changes in white blood cell
count persisted beyond the first 6 months after stopping smoking, data from men who successfully quit
smoking (i.e., men who always reported ex-smoking
after last report of smoking) were selected and their
average white blood cell levels were described over
time.
Analysis of the tracking of white blood cell count
was carried out in men who did not change their
smoking habit in die follow-up period. For the / individuals in a given smoking category, we denote by
Ylt the white blood cell count of die ith individual, 1 <
i < /, at time t, 1 =£ t < Tt. The within-individual
sample mean and variance are given by
and s2 =
m, =
-
1).
The variance between individuals in a given smoking
category is given by
s2 = S , K - m)2/(I - 1),
with m being die mean of the mr The tracking, r!t of
die widiin-person measurements of white blood cell
count was described as die ratio of die betweenindividual variation of the corresponding smoking
group divided by the sum of the between-individual
variance and die widiin-individual variance, i.e.,
r, = s2/(s2
+ s,2).
In addition, using the regression methods widi damped
exponential correlation (23), we assessed change in
within-individual correlation as the time span between
measures lengthened.
RESULTS
Table 1 describes smoking and white blood cell
counts at all visits. Smoking (i.e., occasionally or more
often) was reported in 5,790 of 20,918 (27.7 percent)
person-visits. During the study period, the proportion
of men who had quit smoking increased, while the
proportions of those who had never smoked and who
smoked ^ 1 pack/day decreased (figure 1).
Cross-secdonally, a dose-response relationship was
observed between white blood cell count and smoking
(table 1). With respect to the composition of white
blood cells, smokers had elevated percentages of neutrophils and eosinophils and lower percentages of
CD8 + T-cells and non-T-lymphocytes.
Men who stayed in the same smoking category
during the entire follow-up period showed the following mean ± standard error of white blood cell
count//iliter: 6,287 ± 42 in never smokers, 6,502 ±
51 in ex-smokers, 6,610 ± 221 in occasional smokers, and 6,806 ± 153, 7,939 ± 124, and 8,107 ±
136 in smokers of < 1 , 1-2, and S 2 packs/day,
respectively. Long-term ex-smokers had significantly higher white blood cell counts than never
smokers (a difference of 215 ± 59 cells/jiliter, p <
0.01). Because never smokers, ex-smokers, and
smokers of 1 to < 2 packs/day showed a stable
smoking behavior, results for all person-visits were
similar to those obtained by selecting only those
who stayed in a smoking category. The number of
men who persisted in the other smoking categories
was lower and the data differed slightly.
Figure 2 depicts die frequency of the widiinindividual correlation of white blood cell count for
individuals who remained in die same smoking category. The median of the widiin-individual correlation
of white blood cell counts was 0.72, 0.70, 0.74, and
0.63 for never smokers, ex-smokers, smokers of <1
pack/day, and smokers of ^ 1 pack/day, respectively.
Heavy smokers (^1 pack/day) had a significantly
lower within-individual correlation (p < 0.05) compared witii never smokers. Average correlations (95
TABLE 1. Descrtptlve statistics of number and components of white blood cells (WBC), by smoldng status reported at all visits
by 2,435 seronegative participants In the MuKlcenter AIDS Cohort Study, 1987-1994
status
Never smoked
Ex-smoker
Occasionally smoke
<1 pack/day
1-2 packs/day
22 packs/day
N0.t
No. of
personvtsitsi
980
1,108
361
553
493
210
7,709
7,419
639
2,060
2,185
906
WBC count
(mean ± SE§)
6,265 ± 1 8
6,501 ± 19*
6,412
7,113
7,922
8,279
± 60*
±40*
± 42*
± 64*
WBC composition (% of total WBC)
Neutrophil
Basophfl
EoslnophU
Monocyte
CD4*
T-cefl
CO8*
T-cefl
NortT-cefl
55.8
56.1
55.1
56.2
58.6*
59.5*
0.70
0.75*
0.69
0.72
0.76*
0.72
2.21
2.24*
2.20
2.32*
2.35*
2.64*
5.9
6.1*
5.9
5.9
5.6*
5.6*
15.8
16.1
16.1
16.6*
16.1
15.6
9.9
9.7
9.8
9.4*
9.0*
8.5*
9.8
9.3*
10.0
9.0*
8.2*
7.6*
* p < 0.05 (reference category: never).
t Number of individuals (note: the same Individual may appear In several smoldng catsgories).
t Number of person-visits that provided data on WBC.
Am J Epidemiol
Vol. 144, No. 8, 1996
Longitudinal Relation between Smoking and White Blood Cells
737
CD
a.
0-
i
1987-88
r
1988-89
i
1989-90
1990-91
i
i
1991-92
1992-93
r
1993-94
Year
FIGURE 1. Percent at each smoking category (n, never smokers; e, ex-smokers; 1, smokers of <1 pack/day, 2, smokers of ^ 1 pack/day),
by calendar year the Multicenter AIDS Cohort Study, 1987-1994.
percent confidence intervals) between two white blood
cell counts 6 years apart, estimated from the regression
models, were 0.51 (0.50-0.52) for never smokers,
0.48 (0.46-0.50) for ex-smokers, 0.56 (0.49-0.62)
for men who smoked < 1 pack/day, and 0.43 (0.400.46) for men who smoked ^ 1 packs/day. Dampening
was moderate, although statistically significant at all
the smoking categories (p < 0.05).
Changes in white blood cell count among individuals who stop smoking are depicted in figure 3. White
blood cell counts in these men decreased after quitting
smoking, reaching white blood cell counts similar to
the levels in those who entered as and remained exsmokers. The steepest decline in the white blood cell
count after quitting smoking occurred in the first 6
months for those who smoked ^ 1 pack/day. A stable
level of white blood cell count was attained at approximately 1.5 years after quitting.
Finally, we assessed changes in white blood cell
count at the individual level. A total of 16,982 pairs of
person-visits provided information on changes in
white blood cell counts between two visits. Of these,
36.9 percent and 33.1 percent were in never and exAm J Epidemiol
Vol. 144, No. 8, 1996
smokers, respectively, who remained in the same
smoking category. Table 2 shows changes in white
blood cell count for those who smoked. Participants
who did not change their smoking status showed small
variations in white blood cell count, similar to those
observed in never smokers (mean ± standard error,
19 ± 7 cells//xliter) and in persistent ex-smokers (8 ±
9). Changes from nonsmoking to smoking were
mainly in persons who previously had stopped smoking. Among men who restarted smoking (n = 394
pairs), the higher the number of cigarettes first reported after they restarted smoking, the greater the
white blood cell count increase. Conversely, smokers
who quit smoking (n = 571 pairs) had a monotonic
decrease of white blood cell count according to the
amount of cigarettes previously smoked. Changes between smoking categories (n = 547 and 554 pairs who
increased and decreased, respectively) were also consistent with the previous findings, i.e., the increase
(decrease) of cigarettes was followed by an increase
(decrease) in the white blood cell count, except for
changes between smoking 1 to < 2 packs/day and ^ 2
packs/day.
738
Sunyer et al.
CO
o
o
1
o
O
CD
o
•g
c
d
Never
ppd
Ex
>1 ppd
Smoking Status
FIGURE 2. Frequency of the wrthln-lndividual variation In white blood cell count standardized by the between-individual vanation by
smoking status (never, never smokers; ex, ex-smokers; <1 ppd, smokers of <1 pack/day, 2:1 ppd, smokers of 2 1 pack/day): the MuWcerrter
AIDS Cohort Study, 1987-1994. Boxplots depict percentlles 5, 25, 50, 75, and 95.
DISCUSSION
Our analysis confirms the previously described
dose-response relationship between smoking and
white blood cell count, and provides new findings on
changes in smoking and changes in white blood cell
count. Longitudinal data have the advantage that they
allow characterization of change at the individual level
in contrast to comparison of groups of individuals in
cross-sectional analyses. The longitudinal information
on smoking also facilitated quality control of the data,
because inconsistencies over time could be examined
and misclassification of smoking status minimized.
Leukocytosis in smokers has been well established
based on cross-sectional (12) and longitudinal studies
in the general population (2). In contrast, information
on changes in cell subtypes, particularly T-cell subtypes, has been limited to small laboratory studies
(16, 17) and cross-sectional studies (26-28). We corroborated a moderate increase of neutrophils and eosinophils and decrease of lymphocytes, as percent of
total white blood cells (12, 29, 30). Discrepant findings have been reported on the effect of smoking on
monocytes. In our study, monocytes were not increased, as has been previously reported (29). We
observed a moderate decrease of CD8 + T-cells and
non-T-lymphocyte cells and no changes in CD4 + Tcells in percent over total leukocytes in smokers of ^ 1
pack/day. These findings agree with previous observations of an increase of CD4 + T-cells and no changes
in CD8 + T-cells in smokers of s i pack/day when
lymphocyte subtypes were analyzed as percent over
total lymphocytes (20, 26, 27).
The greatest change in white blood cell count after
a change in smoking occurred within the first 6
months. Changes in shorter time intervals could not be
assessed in our study. Changes after this time period
were smaller, and a significant residual smoking effect
on white blood cell count after quitting smoking was
observed in our follow-up of 6.5 years. Thus, participants who entered the study as ex-smokers and who
remained in that category had white blood cell counts
an average of approximately 200 white blood cells/
/xliter higher than those who entered as never smokers.
Similarly, participants who quit smoking during the
Am J Epidemiol
Vol. 144, No. 8, 1996
Longitudinal Relation between Smoking and White Blood Cells
739
Semesters after Quitting
FIGURE 3. White blood cell count averages and 95 percent confidence intervals for successive visits after quitting smoking by smoking
status at semester 0 (1 smoking <1 pack/day, 2: smoking 1-2 packs/day, 3: smoking &2 packs/day): the Multicenter AIDS Cohort Study,
1987-1994 Number of observations = >15 at each point.
study period did not decrease to the lower count of
never smokers, but did reach that of subjects who
entered as ex-smokers and remained in that category.
The findings are consistent with previous observations
of an intermediate level for ex-smokers between those
of never and current smokers (1, 2, 18, 19). In addition, we found that this plateau is reached in less than
2 years, in contrast to previous studies that found
longer times to reach a plateau (18, 19). Biologic
mechanisms for a persistent effect of smoking in white
blood cell counts have been proposed, such as a residual chronic inflammation of the bronchial tree due
to smoking (18). Although we corrected inconsistencies on repeated measurements of smoking status, a
higher misclassification in ex-smokers could not be
excluded, because we were not able to biologically
confirm the smoking status, as a previous study did
(2).
The white blood cell count showed high tracking
whatever the smoking status, although less so in heavy
smokers (in particular, among subjects who smoked
s 2 packs/day, who likely were more heterogeneous in
Am J Epidemiol
Vol. 144, No. 8, 1996
the amount of cigarettes smoked). This high level of
tracking was seen even without taking into account
factors that acutely modify the white blood cell count,
such as acute infections or hour of the day (15, 31), or
host factors that change over time and that have been
related with white blood cell count, such as age or
body mass index (18, 32). If white blood cell count is
a marker of the oxidant load, the high tracking in white
blood cell count after removal of the effect of smoking
suggests an individual predisposition to the oxidative
damage, invariant over time. On the other hand, if
white blood cell count is a marker of exposure to toxic
substances, the long-term high tracking suggests the
persistence of environmental exposures, possibly related to white blood cell count, other than smoking
(e.g., diet).
Race has been related to white blood cell count (27).
In this study, because 91 percent of the subjects were
white, we did not have high power to detect differences according to race; however, race did not appear
to modify the relation between smoking and white
blood cell count (data not shown). Although we stud-
740
Sunyer et al.
TABLE 2. Estimated average change ± standard error for differences In white blood cell count
(cells/filter) for participants who smoked at some time during the HuHicenter AIDS Cohort Study,
1987-1994
Type ct change
tn smoking status
No change
Increase
From ex-smoker
From occasional smoker
From smoking <1 pack/day
From 1-2 packs/day
Decrease
To quit
To occasional smoker
To <1 pack/day
To 1-2 packs/day
SmoWng (packs/day)
<1
27 ± 2 7
241 ±131
135 ±126
-32±119
-196 ±122
1-2
7 ±28
i2
-6 ±49
340±200t
1,866 ± 7 2 1 *
202 ± 9 4 *
740 ± 524
-*
382 ±357
-123 ±144
-629 ±168*
-172 ± 9 4 t
-1,122*412
-
-295±265
90 ±161
• p < 0.05.
t 0.05 <p< 0.10.
t Dash given when no. of pairs <5.
ied only men, the present findings contribute to the
general understanding of the dynamic relation between
smoking and white blood cell count and provide new
information about the high level of memory of white
blood cell count at the individual level. This may be
relevant for understanding the underlying mechanisms
of the relation between smoking and disease, and
evidence, if smoking behavior is constant, of the utility
of a single measurement of white blood cell count as
a biomarker in epidemiologic studies.
Morgenstern, Parunag Nishanian, and Jeremy Taylor. Pittsburgh: University of Pittsburgh Graduate School of Public
Health, Pittsburgh, PA: Dr. Charles R. Rinaldo, Principal
Investigator; Drs. Lawrence Kingsley, James T. Becker,
Phalguni Gupta, and Monto Ho. Data Coordinating Center:
The Johns Hopkins University School of Hygiene and Public Health, Baltimore, MD: Dr. Alvaro Mufloz, Principal
Investigator; Drs. Cheryl Enger, Stephen Gange, Lisa P.
Jacobson, Robert Lyles, Yun Peng, Steven Piantadosi, and
Sol Su. National Institutes of Health, Bethesda, MD: National Institute of Allergy and Infectious Diseases: Dr.
Lewis Schrager, Project Officer; National Cancer Institute:
Dr. Daniela Seminara.
ACKNOWLEDGMENTS
Supported by grants from the National Institute of Allergy and Infectious Diseases, National Institutes of Health
(cooperative agreements nos. UO1-AI-35042, 35043,
35039, 35040, and 35041). The work of Dr. Sunyer was
partially funded by the Fondo Investigaciones Sanitarias of
Spain, BAE/FIS 95/5259.
The Multicenter AIDS Cohort Study includes the following sites and institutions. Baltimore: The Johns Hopkins
University School of Hygiene and Public Health, Baltimore,
MD: Dr. Alfred J. Saah, Principal Investigator; Drs. Haroutune Armenian, Homayoon Farzadegan, Neil Graham,
Donald R. Hoover, Nancy Kass, Joseph Margolick, and
Justin McArthur, and Ellen Taylor. Chicago: Howard
Brown Health Center-Northwestern University Medical
School, Chicago, IL: Dr. John P. Phair, Principal Investigator; Drs. Joan S. Chmiel, Bruce Cohen, Maurice O'Gorman,
Daina Variakojis, Jerry Wesch, and Steven M. Wolinsky.
Los Angeles: University of California, UCLA Schools of
Public Health and Medicine, Los Angeles, CA: Dr. Roger
Detels, Principal Investigator, John Oishi, Dr. Barbara R.
Visscher, Janice P. Dudley, and Drs. John L. Fahey, Janis V.
Giorgi, Moon Lee, Oto Martfnez-Maza, Eric N. Miller, Hal
REFERENCES
1. Friedman GD, Klatsky AL, Siegelaub AB. The leukocyte
count as a predictor of myocardial infarction. N Engl J Med
1974;290:1275-8.
2. Gnmm RH Jr, Neaton JD, Ludwig W, for the Multiple Risk
Factor Intervention Trial Research Group. Prognostic importance of the white blood cell count for coronary, cancer, and
all-cause mortality. JAMA 1985;254:1932-7.
3. Yamell JWG, Baker IA, Sweetnam PM, et al. Fibrinogen,
viscosity, and white blood cell count are major risk factors for
ischemic heart disease. Circulation 1991;83:836-44
4. Kannel WB, Anderson K, Wilson PWF. White blood cell
count and cardiovascular disease. JAMA 1992;267:1253-6.
5. Phillips AN, Neaton JD, Cook DG, et al. Leukocyte count and
risk of major coronary heart disease events. Am J Epidemiol
1992;136.59-70.
6. Schwartz J, Weiss ST. Peripheral blood leukocyte count and
respiratory symptoms. Ann Epidemiol 1993;3:57-63.
7. Gillum RF. Peripheral leukocyte count and pulmonary function in a national sample of women and men. Ann Epidemiol
1991;1:525-31.
8. Malech HL, Gallin JI. Neutrophils in human disease. N Engl
J Med 1987:317:687-94.
9. Weiss SJ. Tissue destruction by neutrophils. N Engl J Med
1989:320:365-76.
Am J Epidemiol Vol. 144, No. 8, 1996
Longitudinal Relation between Smoking and White Blood Cells
10. Crowell RE, Samet JM. Why does the white blood cell count
predict mortality? Am J Epidemiol 1995;142:493-8.
11. Morrow JD, Frei B, Longmire AT, et al. Increase in circulating products on lipid peroxidation (F2-isoprostanes) in smokers. Smoking as a cause of oxidative damage. N Engl J Med
1995;332:1198-1203
12. Corre F, Lellouch J, Schwartz D. Smoking and leukocytecounts: results of an epidemiological survey. Lancet 1971;2:
632-4.
13 Friedman GD, Stegelaub AB, Seltzer CC, et al Smoking
habits and the leukocyte count Arch Environ Health 1973;26:
137-43.
14. Zalokar JB, Richard JL, Claude JR. Leukocyte count, smoking, and myocardial infarction. N Engl J Med 1981,304:
465-8.
15. Pocock SJ, Ashby D, Shaper AG, et al. Diurnal variations in
serum biochemical and haematological measurements. J Cbn
Pathol 1989;42:172-9.
16 Miller LG, Goldstein G, Murphy M, et al. Reversible alterations in immunoregulatory T cells in smoking Chest 1982;
5:527-9.
17. Hughes DA, Haslam PI, Townsend PJ, et al. Numerical and
functional alterations in circulatory lymphocytes in cigarette
smokers. Clin Exp Immunol 1985;61:459-66.
18. Petitti DB, Kipp H The leukocyte count: associations with
intensity of smoking and persistence of effect after quitting.
Am J Epidemiol 1986; 123:89-95.
19. Schwartz J, Weiss ST. Host and environmental factors influencing the peripheral blood leukocyte count Am J Epidemiol
1991;1341402-9.
20. Park LP, Margolick JB, Giorgi JV, et al. Influence of HTV-1
infection and cigarette smoking on leukocyte profiles in homosexual men J Acquir Immune Defic Syndr 1992,5:
1124-30.
21. Kaslow RA, Ostrow DG, Detels R, et al. The Multicenter
AIDS Cohort Study: rationale, organization, and selected
Am J Epidemiol
Vol. 144, No. 8, 1996
22.
23.
24.
25
26.
27.
28.
29.
30.
31.
32.
741
characteristics of the participants. Am J Epidemiol 1987; 126:
310-18.
Giorgi JV, Cheng HL, Margolick JB, et al. Quality control in
the flow cytometric measurement of T-lymphocyte subsets:
the Multicenter AIDS Cohort Study experience Clin Immunol
Immunopathol 1990;55:173-86.
Mufloz A, Carey V, Shouten JP, et al. A parametric family of
correlation structures for the analysis of longitudinal data.
Biometrics 1992;48733-42.
Ware J, Dockery DW, Louis TA, et al Longitudinal and
cross-sectional estimates of pulmonary function decline in
never-smoking adults. Am J Epidemiol 1990; 132:685-700.
Diggle P, Liang K, Zeger SL. Analysis of longitudinal data.
New York: Oxford Science Publications, 1994143-5
Tollerud DJ, Clark JW, Morris Brown L, et al. The effects of
cigarette smoking on T cell subsets. Am Rev Respir Dis
1989;139:1446-51.
Milli F, Flanders WD, Boring JR, et al The associations of
race, cigarette smoking, and smoking cessation to measures of
immune system in middle-aged men. Clin Immunol Immunopathol 1991;59:187-200.
Chavance M, Perrot JY, Annesi I. Smoking, CD45RO"1"
(memory) and CD45RA + (naive) CD4 + T cells. Am Rev
Respir Dis 1993; 148:237-40
Noble RC, Penny BB. Comparison of leukocyte count and
function in smoking and nonsmoking young men. Infect Immun 1975,12:550-5.
Taylor RG, Gross E, Joyce H, et al. Smoking, allergy, and the
differential white blood cell count. Thorax 1985 ;40:17-22.
Chamberlain AC, Turner FM, Williams EK. The evaluation of
white-cell counting in radiation protection Br J Radiol 1952;
25:169-76.
Nieto FJ, Szklo M, Folsom AR, et al. Leukocyte count correlates in middle-aged adults, the Atherosclerosis Risk in
Communities (ARIC) Study. Am J Epidemiol 1992;136:
525-37.