American Journal of Epidemiology
Copyright © 2004 by the Johns Hopkins Bloomberg School of Public Health
All rights reserved
Vol. 160, No. 1
Printed in U.S.A.
DOI: 10.1093/aje/kwh179
Prospective Study of Cigarette Smoking and Amyotrophic Lateral Sclerosis
M. G. Weisskopf1,2, M. L. McCullough3, E. E. Calle3, M. J. Thun3, M. Cudkowicz4, and
A. Ascherio1,5,6
1
Department of Nutrition, Harvard School of Public Health, Boston, MA.
Department of Environmental Health, Harvard School of Public Health, Boston, MA.
3 Epidemiology and Surveillance Research, American Cancer Society, Atlanta, GA.
4 Neurology Clinical Trial Unit, Massachusetts General Hospital, Harvard Medical School, Boston, MA.
5 Department of Epidemiology, Harvard School of Public Health, Boston, MA.
6 The Channing Laboratory, Department of Medicine, Harvard Medical School and Brigham and Women’s Hospital, Boston, MA.
2
Received for publication October 8, 2003; accepted for publication February 4, 2004.
Cigarette smoking has been proposed as a risk factor for amyotrophic lateral sclerosis (ALS), but because of
the low incidence of ALS this association has been examined only with case-control methods. The authors
prospectively assessed the relation between cigarette smoking and ALS mortality among participants in the
Cancer Prevention Study II cohort of the American Cancer Society, a cohort of over 1 million people enrolled in
1982 who completed a lifestyle questionnaire including a detailed smoking history at baseline. Causes of deaths
were ascertained through death certificates; ALS was not identified separately until 1989. From January 1, 1989,
through 1998, 291 women and 330 men died from ALS. The relative risk of ALS among current smokers
compared with never smokers was 1.67 (95% confidence interval: 1.24, 2.24; p = 0.002) in women and 0.69 (95%
confidence interval: 0.49, 0.99; p = 0.04) in men. The difference in the relative risk estimates between the sexes
was statistically significant (p < 0.0003). This large prospective study provides limited evidence that current
cigarette smoking may be associated with increased death rates from ALS in women but not in men.
amyotrophic lateral sclerosis; mortality; motor neuron disease; prospective studies
Abbreviations: ALS, amyotrophic lateral sclerosis; CI, confidence interval; CPS II, Cancer Prevention Study II; ICD-9,
International Classification of Diseases, Ninth Revision.
Cigarette smoke contains numerous toxic chemicals
including many that induce oxidative stress (11–13), which
has been suggested to play a role in the pathogenesis of ALS.
A significant positive association between cigarette smoking
and risk of ALS has been recently reported in two casecontrol studies (14, 15). Although no relation was found in
some previous studies (16–21), the null findings may have
been the result of several limitations, including small sample
size and biased control selection.
Data from prospective cohort studies could bring stronger
evidence to the assessment of the relation between cigarette
smoking and ALS. The relatively low frequency of the
disease, however, means that a very large cohort would be
necessary. The Cancer Prevention Study II (CPS II) cohort
of the American Cancer Society, comprising over 1 million
Amyotrophic lateral sclerosis (ALS) is a neurodegenerative disorder of motor neurons that afflicts an estimated
30,000 Americans (1). The consequences of ALS are
dramatic, as it rapidly compromises motor function. The
median survival time from diagnosis with ALS is an estimated 1.5–3 years, and mean survival times are only minimally longer (2–6). Although there are familial cases of
ALS, and both autosomal dominant and autosomal recessive
transmissions have been documented (7), about 90 percent
of cases are sporadic (8) and of unknown etiology. Several
risk factors for sporadic ALS have been examined epidemiologically—including aspects of diet, farm work, and exposure to metals or solvents—but evidence of associations has
remained generally inconsistent, in part because analyses
have been limited to case-control studies (9, 10).
Reprint requests to Dr. Marc G. Weisskopf, Department of Environmental Health, Occupational Health Program, Harvard School of Public
Health, Landmark Center, 3rd Floor, P.O. Box 15697, Boston, MA 02215 (e-mail: [email protected]).
26
Am J Epidemiol 2004;160:26–33
Cigarette Smoking and Risk of ALS 27
Americans (22), provided the opportunity to assess prospectively the relation between cigarette smoking and ALS
mortality among men and women.
also asked the age at which they quit smoking. Data on the
covariates considered in the analyses were also obtained
from responses to the questionnaire in 1982.
MATERIALS AND METHODS
Statistical analyses
Study population
Participants contributed follow-up time from January 1,
1989, to the date of death from ALS or any other cause, or
December 31, 1998, whichever came first. Age-specific
mortality rates were calculated as the number of ALS deaths
divided by the person-time of follow-up in each 5-year age
group. Age-adjusted (in 5-year age groups) relative risks were
calculated by dividing the incidence of ALS among participants in each category of smoking variable by the corresponding incidence in never smokers, using Mantel-Haenszel
weights. We used Cox proportional hazards regression to estimate relative risk and 95 percent confidence intervals when
adjusting for, or stratifying by, additional variables. To obtain
a better age adjustment, we stratified the Cox models by age
in single years. The significance of trends was assessed using
the method of Breslow and Day (25), by assigning medians to
each category and modeling as a continuous variable. In the
comparison of never with past and current smoking, a single
variable with values of 1, 2, and 3 was modeled. Interactions
were entered as multiplicative terms (described in Results) in
the Cox models, and their significance was ascertained on the
basis of the Wald test or the likelihood ratio test. SAS version
6.12 software (SAS Institute, Inc., Cary, North Carolina) was
used for all analyses.
The CPS II is a prospective cohort study of nearly 1.2
million US men and women, begun in 1982. Participants
were recruited by American Cancer Society volunteers in 50
states, the District of Columbia, and Puerto Rico (22). Families with at least one member over the age of 45 years and
other family members over the age of 30 years were invited
to participate. The median age at cohort entry in 1982 was 57
years for men and 56 years for women. In total, 508,334 men
and 676,288 women completed a four-page questionnaire.
Because deaths from ALS before 1989 were not coded separately, we included in the study only 459,360 men and
638,849 women who were still alive as of January 1, 1989.
The follow-up period extended from January 1, 1989, to
December 31, 1998.
Case ascertainment
Vital status of the study participants has been determined
by automated linkage with the National Death Index through
December 31, 1998 (23). Death certificates or codes for
cause of death have been obtained for over 98 percent of
known deaths. The underlying cause of death was coded
from death certificates according to the International Classification of Diseases, Ninth Revision (ICD-9) (24). Prior to
and through 1988, deaths from ALS were coded together
with rare causes of death and, thus, cannot be identified.
Deaths from ALS occurring after 1988 are defined as ICD-9
code 335.2 (motor neuron disease) as either the underlying
or a contributing cause of death. We had available for review
183 death certificates originally coded with ICD-9 code
335.2 as the underlying cause of death in CPS II. Among
these 183 certificates, 164 (89.6 percent) were specified as
ALS (ICD-9 code 335.20), 13 (7.1 percent) were not specified beyond motor neuron disease (ICD-9 code 335.2), and
for three certificates (1.6 percent) the text was illegible. Of
the remaining three certificates (1.6 percent), two (1.1
percent) had a diagnosis of bulbar palsy, and one (0.6
percent) had a diagnosis of progressive muscular dystrophy.
The results were similar when stratified by gender, as 86.1
percent and 94.6 percent of men (n = 108) and women (n =
75), respectively, were specified as having ALS. These
results suggest that few death certificates with code 335.2
did not have a diagnosis of ALS.
Assessment of exposure
Cigarette smoking status was ascertained in 1982 by the
question, “Do you now or have you ever smoked cigarettes,
at least one a day for one year’s time?” Ever smokers were
then asked questions on the average number of cigarettes
smoked per day, the age when they started smoking, and the
total number of years they smoked. Former smokers were
Am J Epidemiol 2004;160:26–33
RESULTS
Between 1989 and 1998, we documented 330 deaths from
ALS during 4,135,897 person-years among men and 291
deaths from ALS during 6,018,646 person-years among
women. ALS mortality rates were higher in men than in
women in every age group and increased steadily with age
up to age 75–79 years in men and age 80–84 years in women
(figure 1).
Selected characteristics of men and women included in the
study by smoking status are shown in table 1. The level of
education among men was highest among never smokers and
lowest among current smokers. Among women, educational
status was highest among former smokers. In both men and
women, alcohol consumption was associated with cigarette
smoking.
Among men, the age-adjusted relative risk for ALS
mortality was 0.69 (95 percent confidence interval (CI):
0.48, 0.98) for current smokers compared with never
smokers (table 2). No gradation of risk was seen with cigarettes smoked per day, years of smoking, or age when quit
smoking. Results were similar after adjustment for education
and alcohol consumption, neither of which was significantly
associated with risk of ALS (table 2). Risk of ALS among
men was also not associated with ever smoking of cigars or
pipes (relative risk = 0.94, 95 percent CI: 0.69, 1.29; not
shown in table 2).
Among women, current cigarette smoking was associated
with an increased risk of death from ALS (table 3). The ageadjusted relative risk for current smokers compared with
28 Weisskopf et al.
FIGURE 1. Amyotrophic lateral sclerosis (ALS) death rates in US men and women from the American Cancer Society’s Cancer Prevention
Study II cohort, 1989–1998.
never smokers was 1.66 (95 percent CI: 1.24, 2.22). The estimate was virtually unchanged with adjustment for education
and alcohol consumption. As among men, neither education
nor alcohol consumption was associated with ALS. No clear
gradation of risk was seen with cigarettes smoked per day in
analyses restricted to current smokers, nor with age when
quit smoking in analyses restricted to former smokers (table
3). There was a suggestion of decreasing risk with increasing
duration of smoking or cigarettes smoked per day among
current smokers. The association between current cigarette
smoking and death from ALS differed significantly between
men and women (p for interaction = 0.0003).
Our results for both men and women were not appreciably
affected by further adjustment for occupation type (white or
blue collar or housewife); coffee consumption; or history of
diabetes, heart disease, hypertension, or stroke as reported in
the baseline questionnaire. Because of the possibility that
some CPS II participants who were included in our analysis
(alive in 1989) may have had ALS when they responded to
the baseline questionnaire in 1982 despite the exclusion of
the first 6 years of follow-up, we repeated the analyses
excluding the 44,218 men and 65,836 women who reported
a major illness at baseline. These analyses included 211
deaths and 242 deaths attributed to ALS among men and
women, respectively. The results were not materially
different from the results with the entire cohort. Comorbidities might lead to some underascertainment of ALS on death
certificates. Because comorbidities would be more likely at
older ages, we repeated the analyses with follow-up only
until the age of 75 years. These analyses included 165 deaths
and 172 deaths attributed to ALS among men and women,
respectively. Compared with never smokers, the relative
risks for former and current smoking men were 0.88 (95
percent CI: 0.62, 1.26) and 0.85 (95 percent CI: 0.57, 1.26),
respectively. The comparable relative risks for former and
current smoking women were 1.08 (95 percent CI: 0.72,
1.61) and 1.79 (95 percent CI: 1.26, 2.54), respectively.
Because of the different results obtained in men and
women, we explored whether, among women, the association between smoking and ALS was modified by factors
related to hormonal levels, such as menopausal status, use of
oral contraceptives, or postmenopausal hormones. In separate analyses, we found significantly stronger positive associations among premenopausal women or women with a
history of use of oral contraceptives. In contrast, use of postmenopausal hormones did not appear to modify the association between smoking and risk of ALS. None of these
analyses was based on an a priori hypothesis, however, and
the results should thus be interpreted cautiously.
DISCUSSION
In this large prospective study, we found a positive association between current cigarette smoking and increased death
rate from ALS in women but not in men. The increased risk
in women was seen only in current smokers. There was no
evidence of a dose-response relation.
The influence of cigarette smoking on the risk of ALS has
been previously examined only in case-control studies (14–
21, 26). In many of these studies, no association was seen
(16–21, 26), but methodological considerations limit the
interpretation of these studies. Most investigations had small
sample sizes (less than 100 cases) (16, 17, 19, 21) and used
controls that are likely to have smoking habits similar to
those of the cases, such as friends (16, 21), spouses (20), or
controls matched on lifestyle (17) or socioeconomic status
(21). In one study (26), determination of smoking status was
based on medical records, which are generally incomplete
with respect to smoking histories. In others, the reported
control selection methods (18) and details of smoking exposure (18–20, 26) were minimal.
More recently, two case-control studies have attempted to
address some of these methodological issues and have found
increased risk of ALS among cigarette smokers (14, 15). In a
Am J Epidemiol 2004;160:26–33
Cigarette Smoking and Risk of ALS 29
TABLE 1. Baseline (1982) characteristics* among men and women according to smoking status, American
Cancer Society’s Cancer Prevention Study II
Smoking status (men)
Smoking status (women)
Never
Former
Current
Never
Former
Current
118,398
135,233
93,078
337,043
132,248
126,859
White
93.7
95.5
92.2
92.9
95.2
92.3
Black
3.5
2.5
5.2
4.5
3.2
5.8
Other
2.8
2.1
2.6
2.6
1.6
2.0
Some high school or less
12.4
15.5
21.9
14.3
9.7
15.6
High school graduate
18.7
20.2
22.6
32.1
25.5
31.1
Vocational/trade school
5.7
7.0
7.1
5.9
5.7
6.3
Some college
17.7
21.9
22.1
22.8
27.7
25.8
College graduate or more
45.4
35.4
26.3
25.1
31.5
21.1
Missing
1.2
1.1
1.6
1.3
1.1
1.5
10.3
18.1
23.8
5.4
10.9
13.3
25.9
26.2
25.4
25.0
24.6
23.7
56.1
57.6
53.9
57.0
55.2
53.4
Age when started smoking
(mean, years)†
18.2
18.8
20.6
23.2
Years smoked (mean)†
23.2
37.7
19.3
32.7
Cigarettes smoked/day (mean)†
25.6
24.7
16.0
19.3
Pack-years smoked (mean)†
31.6
46.1
17.9
32.0
Age when quit smoking (mean,
years)†
41.3
40.0
Years since last cigarette (mean)†
15.4
16.1
No.
Race (%)
Education (%)
Alcohol (mean, g/day in 1982)†
Body mass index (mean,
kg/m2)†
Age (mean, years)
Menopausal status (%)†
Premenopausal
29.4
28.9
26.2
Postmenopausal, no hormone
use
38.8
34.2
39.6
Postmenopausal, hormone use
31.8
37.0
34.2
Oral contraceptive use (ever) (%)
26.3
31.4
29.2
5.3
4.8
5.3
Years of oral contraceptive use
(mean)
* All variables, except age, are age adjusted by direct standardization to the entire cohort.
† Among those with nonmissing data.
population-based case-control study in western Washington
State, with 161 cases and 321 controls selected through
random digit dialing and Medicare lists, Nelson et al. (15)
found an odds ratio of 3.7 (95 percent CI: 1.6, 8.4) among
men and of 3.2 (95 percent CI: 1.3, 8.0) among women when
comparing current smokers with never smokers. That study
also found significant increases in risk of ALS with
increasing pack-years of smoking and duration of smoking.
A second population-based case-control study in New
England, with 109 cases and 256 controls selected through
random digit dialing, found an odds ratio of 1.7 (95 percent
CI: 1.0, 2.8) for ever smokers compared with never smokers
(14). This study, however, did not find dose-response relations by pack-years or years smoked. Only Nelson et al. (15)
Am J Epidemiol 2004;160:26–33
and Kondo and Tsubaki (20) reported results separately for
men and women, although all previous studies included both
sexes. Nelson et al. (15) found an increased risk of ALS
among both male and female current smokers, but past
smoking was associated with increased ALS risk only
among women. There were two substudies in the report by
Kondo and Tsubaki (20). The first involved cases identified
from mortality records (n = 458 for men, with 216 controls;
n = 254 for women, with 421 controls) and found relative
risks for smokers of 0.95 among men and 0.92 among
women. The second involved hospitalized ALS cases (104
men, 54 women) and single matched controls. In this
substudy, the relative risk for smokers among men was 1.05,
while the relative risk among women was undefined because
30 Weisskopf et al.
TABLE 2. Baseline age- and multivariate-adjusted relative risk of amyotrophic lateral sclerosis by smoking
category among men, American Cancer Society’s Cancer Prevention Study II, 1989–1998
Multivariate†
95%
confidence
interval
No. of
cases*
No. of
person-years
Relative
risk
Never
99
1,099,896
Referent
Ever
150
2,083,797
0.79
0.61, 1.01
0.77
0.59, 1.00
Former
99
1,213,698
0.84
0.64, 1.11
0.83
0.62, 1.10
Current
48
817,741
0.69
0.48, 0.98
0.69
0.49, 0.99
Relative
risk
95%
confidence
interval
Smoking status
ptrend
Referent
0.09
0.04
Duration of smoking (years)‡
0
99
1,099,896
1–19
2
59,750
Referent
0.95
0.25, 3.68
Referent
1.30
0.30, 5.62
20–34
18
337,442
0.87
0.50, 1.51
1.33
0.70, 2.55
≥35
28
411,422
0.61
0.40, 0.95
0.79
0.63, 0.98
ptrend§
0.03
0.03
ptrend¶
0.49
0.24
Cigarettes per day‡
0
99
1,099,896
<1 pack
10
190,073
Referent
0.60
0.31, 1.16
Referent
0.61
0.32, 1.17
1 pack
10
232,047
0.50
0.26, 0.97
0.50
0.26, 0.96
>1 pack
28
379,939
0.91
0.59, 1.41
0.89
0.57, 1.39
ptrend§
0.29
0.31
ptrend¶
0.20
0.32
Age when quit smoking (years)#
Never smoked
99
1,099,896
≤30
15
243,372
0.79
0.46, 1.37
0.81
0.47, 1.40
31–40
24
368,713
0.73
0.47, 1.14
0.75
0.48, 1.17
41–50
33
359,461
0.91
0.61, 1.35
0.93
0.62, 1.39
>50
26
232,370
0.93
0.60, 1.45
0.92
0.59, 1.45
ptrend¶
Referent
Referent
0.57
0.68
Referent
Referent
Education
Graduated college or more
118
1,601,701
Some college
77
848,427
1.16
0.87, 1.55
1.33
0.95, 1.86
Vocational training
21
267,510
0.98
0.62, 1.55
0.98
0.56, 1.70
Graduated high school
67
814,200
1.02
0.76, 1.39
1.14
0.81, 1.62
Some high school
43
550,686
0.86
0.61, 1.23
0.93
0.61, 1.42
Nondrinker
77
1,045,254
1st quartile
31
444,316
0.99
0.65, 1.51
1.02
0.63, 1.65
2nd quartile
26
417,369
0.87
0.56, 1.36
0.91
0.54, 1.53
3rd quartile
29
401,480
0.91
0.60, 1.40
1.08
0.66, 1.75
4th quartile
33
443,273
1.16
0.77, 1.74
1.29
0.80, 2.07
Alcohol intake
ptrend
Referent
0.52
Referent
0.37
* The number of cases does not sum to the total because of missing data.
† Also adjusted for education and alcohol intake. Models for education and alcohol also adjusted for smoking status.
‡ Among never smokers and current smokers only. Trends are weighted by the median of each category.
§ Including referent category.
¶ Excluding referent category.
# Excluding current smokers. Trends are weighted by the median of each category.
Am J Epidemiol 2004;160:26–33
Cigarette Smoking and Risk of ALS 31
TABLE 3. Baseline age-adjusted relative risk of amyotrophic lateral sclerosis by smoking category among
women, American Cancer Society’s Cancer Prevention Study II, 1989–1998
Multivariate†
No. of
cases*
No. of
person-years
Relative
risk
95%
confidence
interval
Relative
risk
95%
confidence
interval
Smoking status
Never
138
3,186,602
Referent
Ever
136
2,550,111
1.35
1.06, 1.71
1.32
1.03, 1.69
Former
56
1,258,807
1.09
0.80, 1.49
1.04
0.76, 1.43
Current
73
1,179,898
1.66
1.24, 2.22
1.67
1.24, 2.24
ptrend
Referent
0.003
0.002
Duration of smoking (years)‡
0
138
3,186,602
1–19
9
159,363
Referent
2.69
1.43, 5.46
Referent
2.39
1.20, 4.78
20–34
33
588,825
2.06
1.35, 3.15
1.44
0.85, 2.42
≥35
31
419,745
1.29
0.87, 1.92
1.18
0.96, 1.44
ptrend§
0.004
0.004
ptrend¶
0.64
0.20
Cigarettes per day‡
0
138
3,186,602
<1 pack
29
443,537
Referent
1.69
1.13, 2.54
Referent
1.70
1.13, 2.55
1 pack
22
373,574
1.58
1.00, 2.50
1.61
1.02, 2.55
>1 pack
13
303,643
1.35
0.76, 2.40
1.36
0.76, 2.43
ptrend§
0.03
0.02
ptrend¶
0.34
0.36
Age when quit smoking (years)#
Never smoked
138
3,186,602
≤30
13
377,359
1.19
0.66, 2.12
1.14
0.64, 2.03
31–40
17
351,108
1.40
0.84, 2.32
1.31
0.79, 2.20
41–50
11
301,153
0.80
0.43, 1.48
0.77
0.41, 1.42
>50
14
212,430
1.03
0.59, 1.80
1.02
0.58, 1.78
ptrend¶
Referent
Referent
0.57
0.53
Referent
Education
Graduated college or more
76
1,532,711
Referent
Some college
73
1,474,333
0.96
0.70, 1.33
0.96
0.68, 1.34
Vocational training
19
356,881
1.00
0.60, 1.66
0.98
0.58, 1.67
Graduated high school
87
1,835,934
0.89
0.65, 1.21
0.90
0.65, 1.25
Some high school
36
729,737
0.75
0.50, 1.12
0.86
0.56, 1.32
Alcohol intake
Nondrinker
81
1,710,933
1st quartile
20
360,152
1.31
0.80, 2.14
1.37
0.83, 2.28
2nd quartile
18
412,380
0.99
0.59, 1.65
0.91
0.53, 1.56
3rd quartile
26
405,702
1.38
0.89, 2.15
1.32
0.83, 2.10
4th quartile
19
399,460
1.07
0.65, 1.76
0.96
0.57, 1.62
ptrend
Referent
0.71
Referent
0.75
* The number of cases does not sum to the total because of missing data.
† Also adjusted for education and alcohol intake. Models for education and alcohol also adjusted for smoking status.
‡ Among never smokers and current smokers only. Trends are weighted by the median of each category.
§ Including referent category.
¶ Excluding referent category.
# Excluding current smokers. Trends are weighted by the median of each category.
Am J Epidemiol 2004;160:26–33
32 Weisskopf et al.
none of the controls smoked (1.9 percent of the cases
reported smoking).
Our study has several limitations. Follow-up of the CPS II
cohort was limited to mortality and did not directly measure
ALS incidence. The median survival with ALS is short (1.5–
3 years) (2–6), so mortality should be a reasonable surrogate
for incidence. Further, death certificate data have been estimated to accurately identify 70–90 percent of ALS or motor
neuron disease cases (27–30). Thus, a small number of ALS
deaths will have been attributed to other causes in CPS II.
However, this would not materially bias the relative risk estimates unless the misclassification of the cause of death was
strongly related to cigarette smoking. The general validity of
our ALS assessment is supported by our finding of an
increasing risk of ALS mortality with age in both men and
women except at the oldest ages when the risk decreases
again, which is in good agreement with results from previous
US studies with respect to both the rate and the trend over
age (31–33). In addition, the age-specific rates in our study
predict a cumulative ALS mortality (assuming no competing
risks) between the ages of 55 and 74 years of 1.48 per thousand in men and of 0.94 per thousand in women, which are
similar to those found in incidence studies. For example, in a
study in Washington State (34), the predicted cumulative
incidence per thousand over the same age period assuming
no competing risks (our calculation) would be 1.66 in men
and 1.35 in women. Further, the fact that results did not
materially change by excluding deaths after the age of 75
years, which are probably more prone to misclassification,
suggests that diagnostic errors are unlikely to fully explain
our results. An additional limitation is that smoking status
was determined in 1982 only. Cessation of smoking during
follow-up could bias the risk estimates for current smokers
toward the null. While this would not account for the opposite results observed in men and women, it would also make
our results more similar to those of Nelson et al. (15), who
found an increased risk of ALS among women who were
former smokers but not among men who were former
smokers. Finally, CPS II participants are, on average, more
educated and affluent than the US population as a whole
(35). Although these differences may influence comparisons
of absolute rates between the CPS II and US populations,
they should not affect internal validity.
If the relation between smoking and ALS is causal, it is
unclear why smoking would increase risk in women but not
in men. There was some suggestion in our data of possible
effect modification by menopausal status or oral contraceptive use but, because of the small number of cases in these
subanalyses, this result should be interpreted cautiously. It
remains possible that the positive association in women is
due to chance or residual confounding rather than to a causal
relation with smoking.
Cigarette smoking might influence the risk of ALS by
either a direct neurotoxic effect on motor neurons or
increasing oxidative stress. Cigarette smoke contains lead
and other heavy metals (11), and body burdens of many of
these compounds are significantly higher in smokers than in
nonsmokers (11, 36, 37). Although the results have been
questioned to some extent (9), there are several case-control
studies that have shown a positive association between
heavy metal and other chemical exposures and the development of ALS (10, 38). In a recent large case-control study,
exposure to agricultural chemicals was found to be significantly associated with a higher incidence of ALS (39).
Before restrictions on organochlorine pesticide use, cigarette
smoke was found to be a significant source of exposure to
dichlorodiphenyltrichloroethane (DDT) and related
compounds (40). Although the concentrations of individual
organochlorine pesticides in tobacco have declined dramatically since the early 1970s, cigarette smoke may remain a
source of exposure to other agricultural chemicals that could
directly injure motor neurons. Cigarette smoke may also
influence the development of ALS by inducing oxidative
stress. There are many oxidant compounds in cigarette
smoke (12), and cigarette smokers have been found to have
higher blood levels of products of lipid peroxidation (13). In
addition, cigarette smoke has been found to inhibit vascular
endothelial growth factor in hypoxic conditions in vitro and
in vivo (41). Low circulating vascular endothelial growth
factor levels, as found in individuals homozygous for a
mutation in the gene promotor, have been recently found to
be associated with higher risk of ALS (42).
In summary, we found an increased risk of ALS mortality
among female smokers but not among male smokers. The
possibility of effect modification, not only by hormones and
oral contraceptive use but also by other factors that might
relate to the absence of effect among the men in our cohort,
should be explored.
REFERENCES
1. Walling AD. Amyotrophic lateral sclerosis: Lou Gehrig’s disease. Am Fam Physician 1999;59:1489–96.
2. del Aguila MA, Longstreth WT Jr, McGuire V, et al. Prognosis
in amyotrophic lateral sclerosis: a population-based study.
Neurology 2003;60:813–19.
3. Magnus T, Beck M, Giess R, et al. Disease progression in amyotrophic lateral sclerosis: predictors of survival. Muscle Nerve
2002;25:709–14.
4. Sorenson EJ, Stalker AP, Kurland LT, et al. Amyotrophic lateral sclerosis in Olmsted County, Minnesota, 1925 to 1998.
Neurology 2002;59:280–2.
5. Drory VE, Birnbaum M, Korczyn AD, et al. Association of
APOE ε4 allele with survival in amyotrophic lateral sclerosis. J
Neurol Sci 2001;190:17–20.
6. Louwerse ES, Visser CE, Bossuyt PM, et al. Amyotrophic lateral sclerosis: mortality risk during the course of the disease and
prognostic factors. The Netherlands ALS Consortium. J Neurol
Sci 1997;152(suppl 1):S10–17.
7. Siddique T, Nijhawan D, Hentati A. Familial amyotrophic lateral sclerosis. J Neural Transm Suppl 1997;49:219–33.
8. Armon C. Motor neuron disease. In: Gorelick PB, Alter M, eds.
Handbook of neuroepidemiology. New York, NY: Dekker,
1994:407–56.
9. Armon C. Environmental risk factors for amyotrophic lateral
sclerosis. Neuroepidemiology 2001;20:2–6.
10. Nelson LM. Epidemiology of ALS. Clin Neurosci 1996;3:327–
31.
11. Chiba M, Masironi R. Toxic and trace elements in tobacco and
tobacco smoke. Bull World Health Organ 1992;70:269–75.
12. Church DF, Pryor WA. Free-radical chemistry of cigarette
Am J Epidemiol 2004;160:26–33
Cigarette Smoking and Risk of ALS 33
smoke and its toxicological implications. Environ Health Perspect 1985;64:111–26.
13. Morrow JD, Frei B, Longmire AW, et al. Increase in circulating
products of lipid peroxidation (F2-isoprostanes) in smokers—
smoking as a cause of oxidative damage. N Engl J Med 1995;
332:1198–203.
14. Kamel F, Umbach DM, Munsat TL, et al. Association of cigarette
smoking with amyotrophic lateral sclerosis. Neuroepidemiology
1999;18:194–202.
15. Nelson LM, McGuire V, Longstreth WT Jr, et al. Populationbased case-control study of amyotrophic lateral sclerosis in
western Washington State. I. Cigarette smoking and alcohol
consumption. Am J Epidemiol 2000;151:156–63.
16. Pierce-Ruhland R, Patten BM. Repeat study of antecedent
events in motor neuron disease. Ann Clin Res 1981;13:102–7.
17. Provinciali L, Giovagnoli AR. Antecedent events in amyotrophic lateral sclerosis: do they influence clinical onset and
progression? Neuroepidemiology 1990;9:255–62.
18. Mitchell JD, Davies RB, al-Hamad A, et al. MND risk factors:
an epidemiological study in the north west of England. J Neurol
Sci 1995;129(suppl):61–4.
19. Granieri E, Carreras M, Tola R, et al. Motor neuron disease in
the province of Ferrara, Italy, in 1964–1982. Neurology 1988;
38:1604–8.
20. Kondo K, Tsubaki T. Case-control studies of motor neuron disease: association with mechanical injuries. Arch Neurol 1981;
38:220–6.
21. Savettieri G, Salemi G, Arcara A, et al. A case-control study of
amyotrophic lateral sclerosis. Neuroepidemiology 1991;10:
242–5.
22. Thun MJ, Calle EE, Rodriguez C, et al. Epidemiological
research at the American Cancer Society. Cancer Epidemiol
Biomarkers Prev 2000;9:861–8.
23. Calle EE, Terrell DD. Utility of the National Death Index for
ascertainment of mortality among Cancer Prevention Study II
participants. Am J Epidemiol 1993;137:235–41.
24. World Health Organization. International classification of diseases for oncology. 2nd ed. Geneva, Switzerland: World Health
Organization, 1990.
25. Breslow NE, Day NE. Statistical methods in cancer research.
Vol II. The design and analysis of cohort studies. Lyon, France:
International Agency for Research on Cancer, 1987. (IARC scientific publication no. 82).
26. Li TM, Alberman E, Swash M. Clinical features and associations of 560 cases of motor neuron disease. J Neurol Neurosurg
Psychiatry 1990;53:1043–5.
27. Buckley J, Warlow C, Smith P, et al. Motor neuron disease in
England and Wales, 1959–1979. J Neurol Neurosurg Psychiatry 1983;46:197–205.
Am J Epidemiol 2004;160:26–33
28. Chio A, Magnani C, Oddenino E, et al. Accuracy of death certificate diagnosis of amyotrophic lateral sclerosis. J Epidemiol
Community Health 1992;46:517–18.
29. Hoffman PM, Brody JA. The reliability of death certificate
reporting for amyotrophic lateral sclerosis. J Chronic Dis 1971;
24:5–8.
30. O’Malley F, Dean G, Elian M. Multiple sclerosis and motor
neurone disease: survival and how certified after death. J Epidemiol Community Health 1987;41:14–17.
31. Bharucha NE, Schoenberg BS, Raven RH, et al. Geographic
distribution of motor neuron disease and correlation with possible etiologic factors. Neurology 1983;33:911–15.
32. Lilienfeld DE, Chan E, Ehland J, et al. Rising mortality from
motoneuron disease in the USA, 1962–84. Lancet 1989;1:710–
13.
33. Riggs JE. Longitudinal Gompertzian analysis of amyotrophic
lateral sclerosis mortality in the U.S., 1977–1986: evidence for
an inherently susceptible population subset. Mech Ageing Dev
1990;55:207–20.
34. McGuire V, Longstreth WT Jr, Koepsell TD, et al. Incidence of
amyotrophic lateral sclerosis in three counties in western Washington State. Neurology 1996;47:571–3.
35. Watkins ML, Erickson JD, Thun MJ, et al. Multivitamin use
and mortality in a large prospective study. Am J Epidemiol
2000;152:149–62.
36. Hu H, Payton M, Korrick S, et al. Determinants of bone and
blood lead levels among community-exposed middle-aged to
elderly men. The Normative Aging Study. Am J Epidemiol
1996;144:749–59.
37. Mussalo-Rauhamaa H, Leppanen A, Salmela SS, et al. Cigarettes as a source of some trace and heavy metals and pesticides
in man. Arch Environ Health 1986;41:49–55.
38. Mitchell JD. Amyotrophic lateral sclerosis: toxins and environment. Amyotroph Lateral Scler Other Motor Neuron Disord
2000;1:235–50.
39. McGuire V, Longstreth WT Jr, Nelson LM, et al. Occupational
exposures and amyotrophic lateral sclerosis. A populationbased case-control study. Am J Epidemiol 1997;145:1076–88.
40. Djordjevic MV, Fan J, Hoffmann D. Assessment of chlorinated
pesticide residues in cigarette tobacco based on supercritical
fluid extraction and GC-ECD. Carcinogenesis 1995;16:2627–
32.
41. Michaud SE, Menard C, Guy LG, et al. Inhibition of hypoxiainduced angiogenesis by cigarette smoke exposure: impairment
of the HIF-1α/VEGF pathway. FASEB J 2003;17:1150–2.
42. Lambrechts D, Storkebaum E, Morimoto M, et al. VEGF is a
modifier of amyotrophic lateral sclerosis in mice and humans
and protects motoneurons against ischemic death. Nat Genet
2003;34:383–94.