Abstract
Objectives. We studied the effect of smoking on cognitive decline over a 5-year period at middle age (43 to 70 years).
Methods. In the Doetinchem Cohort Study, 1964 men and women in the Netherlands were examined for cognitive function at baseline and 5 years later. The association between smoking status and memory function, speed of cognitive processes, cognitive flexibility, and global cognitive function were assessed.
Results. At baseline, smokers scored lower than never smokers in global cognitive function, speed, and flexibility. At 5-year follow-up, decline among smokers was 1.9 times greater for memory function, 2.4 times greater for cognitive flexibility, and 1.7 times greater for global cognitive function than among never smokers. Among ever smokers, the declines in all cognitive domains were larger with increasing number of pack-years smoked.
Conclusions. Interventions to prevent or stop people from smoking may postpone cognitive decline in middle-aged persons.
With the aging of populations in Western societies, the number of people with dementia, one of the most common neurodegenerative disorders, is expected to rise. Decline in cognitive function is one of the major symptoms of dementia. A preclinical or subclinical phase of reduced cognitive function precedes the appearance of diagnosed Alzheimer's disease by at least 10 years.1 Persons with mild cognitive impairment progress to dementia or Alzheimer's disease at a rate of 10% to 15% per year. Among these persons, those with a faster rate of cognitive decline have a greater chance of developing dementia.2
The pathophysiology of dementia is complex and still largely unclear, and there is no treatment that can stop its progress. Because age and genetics cannot be controlled, it is important to examine risk factors for dementia that can be modified. Lifestyle factors (e.g., smoking) are suitable for interventions.
Most studies have been done with elderly participants (aged 65 years and older). Based on the results of several case–control studies among the elderly, smoking has been reported to have a protective effect on the risk of Alzheimer's disease.3,4 Other studies were inconclusive about the effect of smoking on cognitive function.5 More recently, prospective studies have shown that smoking increases the risk of cognitive decline and dementia among the elderly,6–9 which was confirmed in a recent meta-analysis by Anstey et al.10
In studies of the effects of smoking among the elderly, there can be a selection bias caused by differential mortality among smokers.11 Furthermore, to postpone or prevent cognitive decline and, eventually, dementia at old age, intervention is required by middle age. In addition, to determine cause and effect, it is essential to study associations between potential determinants and cognitive decline in a longitudinal design. Very few studies have examined the relationship between smoking and cognitive function at middle age.12–14 The only longitudinal study conducted showed that heavy smoking is associated with memory decline.14
In 2002, we showed in a cross-sectional study that smoking was associated with lower cognitive function in several cognitive domains among middle-aged men and women.13 We now report the effect of smoking on cognitive decline over a 5-year period among men and women in the same age group.
METHODS
Population
The Doetinchem Cohort Study15 is a prospective study that, during the first examination round (1987–1991), was carried out among a representative sample of 12 405 men and women aged 20 to 59 years in Doetinchem, a town in the eastern part of the Netherlands. Since then, 3 additional examination rounds have been completed (in 1993–1997, 1998–2002, and 2003–2007).
Since 1995, all participants of the Doetinchem Cohort Study aged 45 years and older have been eligible to take part in a cognitive function test consisting of several subtests: the 15-Word Verbal Learning Test,16 the Stroop Color–Word Test,17 the Animal Naming Verbal Fluency test,18 and the Letter–Digit Substitution Test.19 When participants’ time was taken up with other testing (which was especially the case in the years 1995 to 1997), they were not invited for the cognitive testing. Sampling for other testing occurred at random. As a result, 71% of the eligible population took part in the cognitive testing at baseline. In later years of baseline measurements (1998–2000) and during follow-up measurements, over 90% of the eligible participants were invited and took part in the cognitive testing.
Between 1995 and January 2000, a total of 2434 men and women aged 43 to 70 years (48% were younger than 55 years and 86% were younger than 65 years) took part in the baseline measurement of cognitive function. From 2000 to 2005, a total of 1964 of these participants again took part in the cognitive testing. Participants who reported (at baseline or at follow-up) that they had had a cerebrovascular accident (n = 60) were excluded from the analyses because a cerebrovascular accident has direct effects on brain functions and cognition. A total of 1904 participants without a cerebrovascular accident and with cognitive measurements at baseline and 5-year follow-up (923 men and 981 women) were included in the analyses.
Measurements
Cognitive tests.
The neuropsychological test battery that was used measured specific cognitive domains (i.e., memory function, speed of cognitive processes, cognitive flexibility, and global cognitive function). Cognitive tests, which were carried out by trained investigators, took about 20 minutes to complete. The tests are sensitive to calendar age, have no ceiling effect, are robust in detecting age-related impairment (even at middle age), and are sensitive to subcortical dysfunction.20 They have also been used in other large-scale studies on cognitive function.21,22
In the 15-Word Verbal Learning Test (VLT), 15 monosyllabic words were presented visually on paper, one by one, in 3 subsequent trials, with a free recall procedure immediately following each presentation (immediate recall) and an additional free recall trial after a delay of 15 minutes (delayed recall).16 The number of correctly recalled words was recorded. The VLT total represents the total number of correctly recalled words on the 3 immediate recalls. The VLT maximal represents the number of correctly recalled words in the best of the 3 immediate recall performances. The VLT delayed recall is the score on the single delayed recall test.
In the Stroop Color–Word Test,17 3 conditions were tested: the participant was asked to (1) read aloud the names of 40 colors written down, (2) name the color of 40 colored patches, and (3) name the color of the ink in which 40 incongruously named color words were printed (e.g., if the word “red” is printed in yellow ink, the color yellow should be named). Participants were asked to act as quickly as possible without making mistakes. The time used per test condition was recorded.
In the Animal Naming Verbal Fluency Test,18 the participant was asked to name as many animals as possible within 1 minute. The number of uniquely named animals was recorded. In the Letter–Digit Substitution Test (LDST),19 9 letters of the alphabet were linked to a unique digit code (1–9) shown at the top of the form given to each participant. The participant was asked to match the letters shown farther down on the form with the correct corresponding digit. The number of digits that were filled in correctly within 1 minute was recorded.
Results of the different cognitive tests at baseline were transformed, after normalization of scores involving participant's response time (the Stroop Color–Word Test only; plotted distributions were unimodal and skewed to the right of the graph), into standardized z scores. Results of the different cognitive tests at follow-up were transformed—also after normalization for scores involving participant's response time—into standardized scores on the basis of the means and standard deviations of the corresponding test scores at baseline. In this way, the mean standardized score at follow-up was below zero because of cognitive decline with aging. Standard deviations at follow-up were similar to standard deviations at baseline.
Standardized scores of the Stroop Color–Word Test were inverted, so that higher scores always represent better cognition. All standardized test scores were then combined based on their conceptual coherence, derived from neuropsychological practice in performance testing, to form 3 specific cognitive domains and a score for global cognitive function by the following formulae:
 |
 |
 |
 |
Smoking status.
Smoking status was categorized based on the answer to a question about current cigarette smoking at baseline and follow-up as persistent nonsmoker (nonsmoker at baseline and at follow-up; n = 618), ex-smoker (ex-smoker at baseline and at follow-up; n = 796), persistent smoker (smoker at baseline and at follow-up; n = 287), or recent quitter (smoker at baseline but ex-smoker at follow-up; n = 119). Participants who started or resumed smoking (nonsmoker at baseline and smoker at follow-up; n = 37), those who reported conflicting combinations of smoking status between baseline and follow-up (n = 36), and those who did not report smoking status at baseline or at follow-up (n = 11) were excluded from the analyses on effects of smoking status on cognitive decline. For ever smokers, the reported number of cigarettes smoked was calculated into pack-years by dividing the average number of cigarettes smoked daily by 20 and then multiplying that by the number of years they had smoked. In this way, 1 pack-year corresponds to 1 year of smoking 20 cigarettes per day (or 20 years of smoking 1 cigarette per day). The number of pack-years could be calculated for 1173 of 1286 ever smokers.
Other measures.
Demographic characteristics and medical history of chronic diseases were collected through standardized questionnaires at baseline and follow-up; these questionnaires included items about age, educational level, marital status, lifestyle factors (e.g., smoking, alcohol consumption, and physical activity), personal and family history of diseases, and medication use. Educational level, assessed as the highest level reached, was classified into 5 categories. Marital status was classified as married or not married (including never married, divorced, and widowed). Alcohol consumption was classified into 5 categories: (1) no alcohol use, (2) 0 to 1 glass per day, (3) 1 to 2 glasses per day, (4) 2 to 4 glasses per day, and (5) more than 4 glasses per day. Reported level of physical activity was dichotomized as less versus more than half an hour per day of at least moderate-intensity physical activities.
In addition, a self-administered food-frequency questionnaire was used to assess the habitual consumption of 178 food items during the previous year.23 For our study, data on frequency of fish consumption (times per week), coffee consumption (cups per day), total energy intake (megajoules per day), total fat intake (grams per day), and dietary intake of the antioxidants beta-carotene (micrograms per day) and vitamins C and E (milligrams per day) were used.
During a physical examination at the research center, a participant's height and weight were measured, blood pressure was measured twice, and nonfasting blood samples were obtained to determine serum total and high-density lipoprotein cholesterol (millimoles per liter). Body mass index was determined as weight in kilograms divided by height in meters squared. Hypertension was defined as a diastolic blood pressure of 90 mm Hg or higher, a systolic blood pressure of 140 mm Hg or higher, or use of blood pressure–lowering medication. Elevated serum cholesterol was defined as a level of 6.5 mmol/L or higher or use of cholesterol-lowering medication.
Statistical Analyses
In multivariate linear regression analyses, smoking behavior was related to cognitive decline over the 5 years of follow-up. Declines in cognitive domains were analyzed as continuous measures. Smoking was analyzed either in 4 categories (never smokers, ex-smokers, quitters, and consistent smokers) or as a continuous measure (number of pack-years of cigarettes smoked, log transformed, for ever smokers only). Confounders (assessed at baseline) that were considered were age,16 gender,16 level of education,16 alcohol consumption,13 cardiovascular risk factors (hypertension, serum total and high-density lipoprotein cholesterol, body mass index, diabetes or cardiovascular disease, and physical activity), energy intake,24 total fat intake (as energy residual), coffee consumption,25 fish consumption,26 antioxidant intake27 (beta-carotene, vitamin C, and vitamin E as energy residuals), estrogen use28 (for women only), and marital status.29 Potential confounders were included in the models only when the variable was associated with both smoking status and with change in at least 1 domain of cognitive function at P less than .10.
Intake of vitamin E, marital status, estrogen use, and fish consumption were not associated with smoking status. Coffee consumption and total serum cholesterol were not associated with cognitive change. These variables were therefore not included in the analyses.
To adjust for possible regression to the mean, analyses on cognitive decline were adjusted for the baseline level of the specific cognitive domain. Educational level, age, and gender were also tested for interaction with smoking, but no interaction effects were observed. To exclude potentially cognitive-impaired participants, analyses were also performed on data excluding the lowest 5% scores on cognitive domains at baseline, but the results were unaffected. In addition, analyses were performed that excluded persons 65 years and older at baseline, but the results did not change.
RESULTS
Table 1 shows the general characteristics of the study population. The average age at baseline was 56 years, with about equal proportions of men and women. At baseline, 34.9% of the participants reported that they had never smoked cigarettes, 43.8% were ex-smokers, and 21.3% were current smokers. Compared with never smokers, persistent smokers were somewhat younger, more often male, more often lower educated, had a lower serum high-density lipoprotein cholesterol level, were less often hypertensive, had a lower body mass index, were less physically active, and had a higher alcohol intake, a higher energy intake, a higher fat intake, and a lower antioxidant intake.
TABLE 1.
General Characteristics of the Study Population, by Smoking Status: Doetinchem Cohort Study, Doetinchem, Netherlands, 1995–2005
| Totala (N = 1904) | Never Smokers (n = 618) | Ex-Smokers (n = 796) | Recent Quitters (n = 119) | Smokers (n = 287) | Pb | |
| Age, y, mean (SD) | 56.0 (7.0) | 56.3 (7.1)z | 56.6 (7.1)y,z | 55.1 (6.6)x | 54.1 (6.4)w,x | |
| Men, % | 48.5 | 35.6 | 60.2 | 47.1 | 44.9 | <.01 |
| Married, % | 85.1 | 83.7 | 87.0 | 80.7 | 84.2 | .15 |
| Lifelong cigarette smoking, pack-years,c % | <.01 | |||||
| 0 | 36.4 | 100 | 0.9 | 0 | 0 | |
| 0–20 | 38.9 | 0 | 71.3 | 55.1 | 35.7 | |
| > 20 | 24.8 | 0 | 27.8 | 44.9 | 64.3 | |
| Level of education, % | <.01 | |||||
| Primary school | 7.6 | 7.9 | 5.8 | 4.2 | 12.2 | |
| Lower vocational | 26.4 | 29.8 | 23.5 | 22.7 | 28.2 | |
| Intermediate secondary | 17.5 | 16.3 | 18.0 | 23.5 | 17.8 | |
| Intermediate vocational/higher secondary | 24.3 | 19.4 | 27.1 | 28.6 | 24.7 | |
| Higher vocational/university | 24.2 | 26.5 | 25.6 | 21.0 | 17.1 | |
| Cardiovascular risk factors | ||||||
| Total cholesterol, mmol/L, mean (SD) | 5.9 (1.0) | 5.95 (1.03)x | 5.82 (0.98)w,z | 5.88 (0.95) | 5.99 (1.1)x | |
| Elevated serum cholesterol level,d % | 30.0 | 30.6 | 28.0 | 26.9 | 32.8 | .39 |
| HDL cholesterol, mmol/L, mean (SD) | 1.38 (0.39) | 1.43 (0.38)x,z | 1.36 (0.39)w | 1.39 (0.47) | 1.33 (0.38)w | |
| Systolic blood pressure, mm Hg, mean (SD) | 131.1 (17.8) | 130.9 (17.7) | 132.2 (17.4)z | 129.9 (18.9) | 129.2 (18.7)x | |
| Hypertension,e % | 35.6 | 36.6 | 37.4 | 30.3 | 28.9 | .04 |
| Body mass index, kg/m2, mean (SD) | 26.3 (3.8) | 26.5 (3.9)y,z | 26.6 (3.5)y,z | 25.4 (3.3)w,x | 25.4 (3.7)w,x | |
| Self-reported diabetes or cardiovascular disease, % | 4.1 | 2.6 | 5.8 | 2.5 | 3.5 | .02 |
| Physically active, % | 56.2 | 57.1 | 60.1 | 53.8 | 44.3 | <.01 |
| Alcohol consumption, % | <.01 | |||||
| No alcohol use | 30.7 | 45.5 | 23.7 | 16.8 | 26.5 | |
| 0–1 glass/day | 27.9 | 29.0 | 29.0 | 26.9 | 21.6 | |
| 1–2 glasses/day | 20.0 | 14.1 | 22.1 | 27.7 | 23.0 | |
| 2–4 glasses/day | 16.4 | 9.2 | 19.1 | 21.0 | 22.0 | |
| > 4 glasses/day | 5.0 | 2.3 | 6.0 | 7.6 | 7.0 | |
| Dietary intake, mean (SD) | ||||||
| Total energy intake, MJ/day | 8.9 (2.3) | 8.7 (2.3)x,z | 9.1 (2.2)w | 8.9 (2.2) | 9.1 (2.5)w | |
| Total fat intake, g/day | 84 (26) | 81 (27)x,z | 85 (25)w | 86 (25) | 88 (29)w | |
| Beta-carotene, μg/day | 1485 (609) | 1550 (600)x,z | 1460 (555)w | 1458 (619) | 1448 (756)w | |
| Vitamin C, mg/day | 112 (44) | 118 (43)x,y,z | 112 (44)w,y,z | 102 (48)w,x | 102 (45)w,x |
Note. HDL = high-density lipoprotein. Statistical significant difference (P < .05) for a smoking group compared with: never smokers (w), ex-smokers (x), recent quitters (y), and smokers (z). For classification of smoking groups, see “Methods” section.
Includes 84 participants who could not be classified by smoking category.
Determined by the χ2 test, for percentages only.
Pack-years were calculated by dividing the average number of cigarettes smoked daily by 20 and then multiplying that by the number of years the participant had smoked.
Elevated serum cholesterol level was defined as serum total cholesterol of 6.5 mmol/L or higher or use of cholesterol-lowering medication.
Hypertension was defined as diastolic blood pressure of 90 mm Hg or higher, systolic blood pressure of 140 mm Hg or higher, or use of blood pressure–lowering medication.
Figure 1 shows the effects of smoking status on cognitive function at baseline and follow-up. Compared with never smokers, smokers performed worse at baseline on speed of cognitive processes, cognitive flexibility, and global cognition. On average, all smoking subgroups showed a decline in cognitive domains during follow-up. At follow-up, smokers performed significantly worse than did never smokers on all cognitive domains.
FIGURE 1.
Average cognitive functions at baseline and at 5-year follow-up for never smokers, ex-smokers, recent quitters, and consistent smokers: Doetinchem Cohort Study, Doetinchem, Netherlands, 1995–2005.
Note. Effects as shown are adjusted for baseline age, gender, level of education, body mass index, hypertension, cardiovascular disease, high-density lipoprotein cholesterol, physical activity, alcohol consumption, total energy intake, and intake of total fat, beta-carotene, and vitamin C (as energy residuals). The cognitive function scores at baseline of the group of never smokers were set as reference and therefore have a mean score of zero.
*P < .05, for significant cross-sectional difference compared with never smokers.
Among smokers, the declines in memory function (P < .05), cognitive flexibility (P < .05), and global cognitive function (P < .05) over the 5-year follow-up period were respectively 1.9, 2.4, and 1.7 times greater than among never smokers (Table 2). For ex-smokers and recent quitters, no statistically significant differences were observed in cognitive decline compared with never smokers, although the point estimates of the decline increased steadily from never smokers to ex-smokers to recent quitters and to smokers.
TABLE 2.
Cognitive Functions at Baseline and Change in Cognitive Functions at 5-Year Follow-Up, by Smoking Status: Doetinchem Cohort Study, Doetinchem, Netherlands, 1995–2005
| Memory Function, Mean (Pdiff) | Speed of Cognitive Processes, Mean (Pdiff) | Cognitive Flexibility, Mean (Pdiff) | Global Cognitive Function, Mean (Pdiff) | |
| Baseline score | ||||
| Never smokers | 0 (Ref) | 0 (Ref) | 0 (Ref) | 0 (Ref) |
| Ex-smokers | −0.04 (.39) | −0.04 (.31) | −0.08 (.13) | −0.04 (.18) |
| Recent quitters | −0.05 (.57) | −0.10 (.19) | −0.05 (.58) | −0.04 (.50) |
| Smokers | −0.03 (.63) | −0.22 (<.01) | −0.23 (<.01) | −0.09 (.04) |
| Change at follow-up | ||||
| Never smokers | −0.13 (Ref) | −0.14 (Ref) | −0.07 (Ref) | −0.09 (Ref) |
| Ex-smokers | −0.14 (.75) | −0.12 (.44) | −0.09 (.65) | −0.09 (.85) |
| Recent quitters | −0.20 (.38) | −0.13 (.79) | −0.16 (.16) | −0.13 (.27) |
| Smokers | −0.25 (.02) | −0.15 (.72) | −0.17 (.03) | −0.16 (.03) |
Note. The mean cognitive function scores at baseline and the changes in the mean cognitive function scores at 5-year follow-up are expressed in standardized scores. A negative value means worse cognitive function or cognitive decline over follow-up. Pdiff represents the significance level of the difference between the mean changes in cognitive domain functions in contrast to the group of never smokers. The cognitive function scores at baseline of the group of never smokers were set as the reference and therefore have a mean score of zero. The effects of smoking are adjusted for age (continuous), gender, level of education (5 classes), body mass index, hypertension, serum high-density lipoprotein cholesterol, cardiovascular disease, total energy intake, physical activity, alcohol consumption (5 classes), and (energy residuals of) intake of total fat, beta-carotene and vitamin C. Effects on change in cognitive functions are additionally adjusted for the baseline level of cognitive function.
Apart from current smoking status, lifelong exposure to cigarettes showed a dose–response group relationship to cognitive decline: in a fully adjusted model, the decline in memory function, speed of cognitive processes, and cognitive flexibility increased with increasing pack-years of cigarettes (Table 3). The unstandardized parameter estimates denote the additional change in standardized cognitive function score over 5 years of follow-up when the number of pack years smoked was 2.7 (or e) times higher. For global cognitive function, the effect was borderline statistically significant after full adjustment.
TABLE 3.
Effect of the Number of Cigarettes Smoked on Change in Cognitive Functions Over a 5-Year Period Among Ever Smokers: Doetinchem Cohort Study, Doetinchem, Netherlands, 1995–2005
| b | P | |
| Memory function (n = 1162) | −0.04 | .03 |
| Speed of cognitive processes (n = 1161) | −0.02 | .03 |
| Cognitive flexibility (n = 1165) | −0.03 | .04 |
| Global cognitive function (n = 1146) | −0.02 | .06 |
Note. The changes in domains of cognitive function are expressed in standardized scores per log-transformed pack-year (pack-years were calculated by dividing the average number of cigarettes smoked daily by 20 and then multiplying that by the number of years the participant had smoked). A negative value means more cognitive decline with more cigarettes smoked. Effects of smoking are adjusted for age (continuous), gender, level of education (5 classes), body mass index, hypertension, serum high-density lipoprotein cholesterol, cardiovascular disease, total energy intake, physical activity, alcohol consumption (5 classes), intake of total fat, beta-carotene and vitamin C (as energy residuals), and the baseline level of cognitive function.
DISCUSSION
Internal Validity
In our study of middle-aged men and women, cigarette smokers exhibited poorer cognitive performance (speed of cognitive processes, cognitive flexibility, and global cognitive function) at baseline and a larger decline in memory function, cognitive flexibility, and global cognitive function over time compared with never smokers. Overall, at follow-up, smokers had significantly lower cognitive function on all assessed cognitive domains compared with never smokers.
The major strengths of our study are the prospective design and the relative youth of participants (86% were younger than 65 years). Although the participants were still relatively young and healthy when they entered the study, a clear decline in cognitive function, with a wide range of the extent of cognitive decline, was detected with the cognitive test battery used.
In addition, the Doetinchem Cohort Study has an extensive spectrum of variables assessed during the measurements. We were therefore able to adjust for a wide range of potential confounders, but this hardly changed the observed rough associations between smoking and cognitive change (data not shown). This indicates that the independent effect of smoking on cognitive decline is a very robust one.
Other conditions that were not assessed may have influenced test results (e.g., sleeping prior to testing, fatigue, medication usage, stress, caffeine use, and mood). In our study, smokers scored worse then never smokers on mental health (including feelings of depression and nervousness) and vitality (including feelings of energy and fatigue), both assessed with the Rand 36-Item Health Survey.30 In addition, mental health and vitality scores were associated with change in cognitive function. However, additional adjustments for mental health and vitality hardly affected associations between smoking status and change in cognitive functions (data not shown). Further, because cognitive testing was performed under nonfasting conditions, we assume that coffee consumers drank coffee prior to testing. In our data, however, coffee consumption was not associated with change in cognitive function and therefore may not influence our results.
As in all longitudinal studies, not all participants who entered our study returned for a follow-up measurement. The persons who participated at baseline but not at follow-up (n = 459 without cerebrovascular accident), in contrast to those who participated twice (n = 1904 without cerebrovascular accident), were on average almost 2 years older, were more often smokers, and scored lower on all 3 cognitive domains and global cognitive function at baseline. Baseline associations between smoking status (never smoker, ex-smoker, and smoker) and cognitive functions were different for the groups of participants who did and did not participate at follow-up; no clear associations were found in the latter group. On the basis of this observation, we cannot assume that associations between smoking and cognitive decline in the group that was lost to follow-up would have been similar to the effects found in the group with complete data.
The observed change in cognitive function from baseline to follow-up among those who scored low at baseline and those who were relatively old at baseline offers the best available prediction for how those lost to follow-up might have scored. Based on these results, there was no indication that in the group of persons who were lost to follow-up, the associations between smoking status and cognitive decline were different from the associations observed in the complete data.
Another possible drawback of our study is a potential learning effect among participants. If there was a learning effect, it might be expected that participants would perform higher at follow-up than at baseline, because the tests at baseline were exactly the same as the ones used 5 years later. If a learning effect was present, however, it should have affected all participants and therefore should not have influenced the observed association between smoking status and change in cognitive function. In addition, we observed an overall decline in cognitive function, as would be expected with aging.
Quitters were defined as participants who reported they were smokers at baseline and ex-smokers at follow-up. Not all quitters stopped smoking at the same time, and some may have stopped just after baseline and others just before follow-up, making them more equivalent to ex-smokers and smokers, respectively. Ex-smokers, on the other hand, quit smoking some time before baseline, most of them between the ages of 20 and 40 years. Thus, compared with ex-smokers, all quitters in our study stopped smoking relatively recently.
Comparison With Other Studies
In a study among 60-year-olds, smoking was 1 of the risk factors that accelerated cerebral degenerative changes: in the group with cognitive decline, the percentage who were smokers was more than 20% higher than in the group who maintained a higher level of cognition.6 And in a recent cross-sectional study on the effect of smoking (relative to that of nonsmoking and former smoking) on 66-year-olds, smokers scored 4% to 8% lower on memory and information processing speed tests, indicating an adverse effect of smoking on these functions.31 These results are in line with our findings at follow-up, when participants were on average aged 61 years. As far as we know, only 1 other longitudinal study on the effect of smoking on cognitive decline has been performed with middle-aged participants.14 That study found that among participants aged 43 to 53 years, smoking more than 20 cigarettes per day was associated with stronger memory decline. Together with those earlier studies, ours contributes to the body of evidence on the adverse effect of smoking on cognitive function and cognitive decline in healthy middle-aged people.
Interpretations of Results
We found no difference between smokers and never smokers in rates of decline in the speed of cognitive processes during follow-up. However, smokers scored significantly worse than did never smokers on speed of cognitive processes at baseline, a significant difference that persisted over follow-up. Regarding memory function, no differences were observed between smokers and nonsmokers at baseline, whereas during follow-up, smokers’ decline in function was significantly larger, resulting in a significant difference between the 2 groups. For cognitive flexibility and global cognitive function, smokers scored worse at baseline and showed a larger decline over follow-up compared with never smokers. These results could imply that smoking induces effects on different cognitive domains at different ages, with an early effect on speed of cognitive processes, cognitive flexibility, and global cognition, whereas memory function is affected later on.
Our results suggest that smoking has an effect on cognitive decline. We found a dose–response group effect for the number of cigarettes smoked during the lifetime and the degree of cognitive decline over follow-up. Also, although the differences were not always significant, the point estimates of the decline increased steadily from never smokers to ex-smokers to recent quitters and to smokers. In addition, reporting bias for smoking behavior because of cognitive impairment is not likely, because the study population was relatively young and healthy. Our results indicate that giving up smoking at any age may prevent further smoking-induced cognitive decline. We stress the need for stop-smoking interventions in order to postpone cognitive decline among middle-aged persons.
Acknowledgments
The Doetinchem Cohort Study was financially supported by the Ministry of Health, Welfare and Sport of the Netherlands and the National Institute for Public Health and the Environment.
We thank the epidemiologists and fieldworkers of the Municipal Health Service in Doetinchem for their contribution to the data collection for this study. The project director is W. M. M. Verschuren. Logistic management was provided by J. Steenbrink and P. Vissink and administrative support by E. P. van der Wolf. Data management was provided by A. Blokstra, A. W. D. van Kessel, and P. E. Steinberger. We also thank S. Kalmijn, M. T. Schram, and M. Angevaren for their useful contributions concerning data input, practical experience, and content and H. C. G. Boshuizen for her statistical input.
Human Participant Protection
This study was approved according to the guidelines of the Helsinki Declaration by the external Medical Ethics Committee of the Netherlands Organization of Applied Scientific Research Institute. All participants gave written informed consent.
References
- 1.Geschwind DH, Robidoux J, Alarcon M, et al. Dementia and neurodevelopmental predisposition: cognitive dysfunction in presymptomatic subjects precedes dementia by decades in frontotemporal dementia. Ann Neurol 2001;50:741–746 [DOI] [PubMed] [Google Scholar]
- 2.Petersen RC, Smith GE, Waring SC, Ivnik RJ, Tangalos EG, Kokmen E. Mild cognitive impairment. Clinical characterization and outcome. Arch Neurol 1999;56:303–308 [DOI] [PubMed] [Google Scholar]
- 3.Brenner DE, Kukull WA, van Belle G, et al. Relationship between cigarette smoking and Alzheimer's disease in a population-based case–control study. Neurology 1993;43:293–300 [DOI] [PubMed] [Google Scholar]
- 4.Almeida OP, Hulse GK, Lawrence D, Flicker L. Smoking as a risk factor for Alzheimer's disease: contrasting evidence from a systematic review of case–control and cohort studies. Addiction 2002;97:5–28 [DOI] [PubMed] [Google Scholar]
- 5.Tyas SL, Pederson LL, Koval JJ. Is smoking associated with the risk of developing Alzheimer's disease? Results from three Canadian data sets. Ann Epidemiol 2000;10:409–416 [DOI] [PubMed] [Google Scholar]
- 6.Meyer JS, Rauch GM, Crawford K, et al. Risk factors accelerating cerebral degenerative changes, cognitive decline and dementia. Int J Geriatr Psychiatry 1999;14:1050–1061 [DOI] [PubMed] [Google Scholar]
- 7.Ott A, Andersen K, Dewey ME, et al. Effect of smoking on global cognitive function in nondemented elderly. Neurology 2004;62:920–924 [DOI] [PubMed] [Google Scholar]
- 8.Reitz C, Luchsinger J, Tang MX, Mayeux R. Effect of smoking and time on cognitive function in the elderly without dementia. Neurology 2005;65:870–875 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Barnes DE, Cauley JA, Lui LY, et al. Women who maintain optimal cognitive function into old age. J Am Geriatr Soc 2007;55:259–264 [DOI] [PubMed] [Google Scholar]
- 10.Anstey KJ, von Sanden C, Salim A, O'Kearney R. Smoking as a risk factor for dementia and cognitive decline: a meta-analysis of prospective studies. Am J Epidemiol 2007;166:367–378 [DOI] [PubMed] [Google Scholar]
- 11.Fratiglioni L, Wang H-X. Smoking and Parkinson's and Alzheimer's disease: review of the epidemiological studies. Behav Brain Res 2000;113:117–120 [DOI] [PubMed] [Google Scholar]
- 12.Cerhan JR, Folsom AR, Mortimer JA, et al. Correlates of cognitive function in middle-aged adults. Gerontology 1998;44:95–105 [DOI] [PubMed] [Google Scholar]
- 13.Kalmijn S, van Boxtel MPJ, Verschuren WMM, et al. Smoking and alcohol consumption in relation to cognitive performance at middle age. Am J Epidemiol 2002;156:936–944 [DOI] [PubMed] [Google Scholar]
- 14.Richards M, Jarvis MJ, Thompson N, Wadsworth MEJ. Cigarette smoking and cognitive decline in midlife: evidence from a prospective birth cohort study. Am J Public Health 2003;93:994–998 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Verschuren WMM, Blokstra A, Picavet HSJ, Smit HA. Cohort profile: the Doetinchem Cohort Study. Int J Epidemiol. doi: 10.1093/ije/dym292. [published online ahead of print June 3, 2008] doi:10.1093/ije/dym292. [DOI] [PubMed] [Google Scholar]
- 16.Van der Elst W, Van Boxtel MP, Van Breukelen GJ, Jolles J. Rey's Verbal Learning Test: normative data for 1855 healthy participants aged 24–81 years and the influence of age, sex, education, and mode of presentation. J Int Neuropsychol Soc 2005;11:290–302 [DOI] [PubMed] [Google Scholar]
- 17.Van der Elst W, Van Boxtel MP, Van Breukelen GJ, Jolles J. The Stroop Color–Word Test: influence of age, sex, and education; and normative data for a large sample across the adult age range. Assessment 2006;13:62–79 [DOI] [PubMed] [Google Scholar]
- 18.Van der Elst W, Van Boxtel MP, Van Breukelen GJ, Jolles J. Normative data for the Animal, Profession and Letter M Naming verbal fluency tests for Dutch speaking participants and the effects of age, education, and sex. J Int Neuropsychol Soc 2006;12:80–89 [DOI] [PubMed] [Google Scholar]
- 19.Van der Elst W, Van Boxtel MP, Van Breukelen GJ, Jolles J. The Letter—Digit Substitution Test: normative data for 1858 healthy participants aged 24–81 from the Maastricht Aging Study (MAAS): influence of age, education, and sex. J Clin Exp Neuropsychol 2006;28:998–1009 [DOI] [PubMed] [Google Scholar]
- 20.Brand N, Jolles J. Information processing in depression and anxiety. Psychol Med 1987;17:145–153 [DOI] [PubMed] [Google Scholar]
- 21.De Groot JC, de Leeuw FE, Oudkerk M, et al. Cerebral white matter lesions and cognitive function: the Rotterdam Scan Study. Ann Neurol 2000;47:145–151 [DOI] [PubMed] [Google Scholar]
- 22.Møller JT, Cluitmand P, Rasmussen LS, et al. Long-term postoperative cognitive dysfunction in the elderly: ISPOCD1 Study. Lancet 1998;351:857–861 [DOI] [PubMed] [Google Scholar]
- 23.Ocké MC, Bueno-de-Mesquita HB, Goddijn HE, et al. The Dutch EPIC food frequency questionnaire, I: description of the questionnaire, and relative validity and reproducibility for food groups. Int J Epidemiol 1997;26:S37–S48 [DOI] [PubMed] [Google Scholar]
- 24.Mattson MP, Chan SL, Duan W. Modification of brain aging and neurodegenerative disorders by genes, diet and behaviour. Physiol Rev 2002;82:637–672 [DOI] [PubMed] [Google Scholar]
- 25.Van Gelder BM, Buijsse B, Tijhuis M, et al. Coffee consumption is inversely associated with cognitive decline in elderly European men: the FINE Study. Eur J Clin Nutr 2007;61:226–232 [DOI] [PubMed] [Google Scholar]
- 26.Van Gelder BM, Tijhuis M, Kalmijn S, Kromhout D. Fish consumption, n-3 fatty acids, and subsequent 5-y cognitive decline in elderly men: the Zutphen Elderly Study. Am J Clin Nutr 2007;85:1142–1147 [DOI] [PubMed] [Google Scholar]
- 27.Engelhart MJ, Geerlings MI, Ruitenberg A, et al. Dietary intake of antioxidants and risk of Alzheimer's disease. JAMA 2002;287:3223–3229 [DOI] [PubMed] [Google Scholar]
- 28.Rice MM, Graves AB, McCurry SM, et al. Postmenopausal estrogen and estrogen-progestin use and 2 year rate of cognitive change in a cohort of older Japanese American women: the Kane Project. Arch Intern Med 2000;160:1641–1649 [DOI] [PubMed] [Google Scholar]
- 29.Van Gelder BM, Tijhuis M, Kalmijn S, Giampaoli S, Nissinen A, Kromhout D. Marital status and living situation during a 5-year period are associated with a subsequent 10-year cognitive decline in older men: the FINE Study. J Gerontol B Psychol Sci Soc Sci 2006;61:213–219 [DOI] [PubMed] [Google Scholar]
- 30.Hays RD, Morales LS. The RAND-36 measure of health-related quality of life. Ann Med 2001;33:350–357 [DOI] [PubMed] [Google Scholar]
- 31.Starr JM, Deary IJ, Fox HC, Whalley LJ. Smoking and cognitive change from age 11 to 66 years: a confirmatory investigation. Addict Behav 2007;32:63–68 [DOI] [PubMed] [Google Scholar]