No Link Between Smoking and Increased Dementia Risk?

Batya Swift Yasgur, MA, LSW

March 26, 2019

There appears to be no causal link between smoking and dementia, new research suggests.

Investigators followed over 500 senior adults for an average of 11 years, analyzing the potential association between smoking and dementia and adjusting for the competing risk of death without dementia.

Although smoking was associated with earlier death, it was not associated with dementia — findings that were corroborated by neuropathological studies showing no higher prevalence of Alzheimer's disease (AD) pathology in smokers.

"We are not suggesting that smoking protects the brain from dementia, but rather that it is a risk to health on so many levels that death is the more likely outcome, relative to dementia," Erin L. Abner, MD, associate professor of epidemiology and biostatistics at the University of Kentucky's Sanders-Brown Center on Aging in Lexington, told Medscape Medical News.

"Keep telling your patients not to smoke — death is not an ideal way to prevent dementia," she emphasized.

The study was published online March 2 in the Journal of Alzheimer's Disease.

Modifiable Risk Factor?

"Tobacco smoking has been described as a modifiable risk factor for dementia," the authors write.

"Smoking is one of the lifestyle factors that have been suggested to increase the risk of dementia," Abner said, citing a "very good paper, based on the method used, suggesting that a 25% reduction in smoking prevalence could prevent 1 million cases of dementia worldwide each year."

"Given that smoking is such a strong risk factor for mortality, we wanted to do an analysis that accounted for that mortality risk when considering the risk of dementia that might be due to smoking," she recounted.

To investigate the question, the researchers studied 531 initially cognitively intact older adults participating in the Biologically Resilient Adults in Neurological Studies (BRAiNS) study, who enrolled between 1989 and 2003, for an average of 11.5 years.

Participants underwent annual cognitive assessments and autopsies following death.

The researchers adjusted for confounders such as APOE ε4 carrier status, female sex, low education, type 2 diabetes, head injury, hypertension, overweight, family history of dementia, and use of hormone replacement therapy.

"Smoking exposure," which was determined at baseline, was categorized in pack-years and included data on pipes, cigars, and cigarettes.

Competing Risk

Abner recounted the reason that competing analysis was used as the statistical model to analyze the data.

"Survival analysis is used when we are interested not just in whether or not an event occurred — like the onset of dementia — but also in how long it takes for that event to occur," she said.

"The usual assumption in survival analysis is that the fact that volunteers left the study without developing dementia doesn't tell us anything about their future risk of dementia — this is called 'uninformative censoring,' " she continued.

A competing risk is "an event that occurs, which makes it impossible for the competing event of interest to occur, so if a volunteer dies without being diagnosed with dementia, there is zero probability that dementia will be observed for that volunteer in the future [in contrast to] a volunteer who is censored because they moved away."

To look at risk factors for dementia that also carry a strong mortality risk, the researchers therefore used a Fine-Gray competing risk model for survival analysis.

The final cohort consisted of 531 participants (mean age at enrollment 73.2 [±7.4] years, 63.1% female, and highly educated) divided into three categories: current smokers, former smokers, and "never smokers."

A large percentage (40.3%) had a positive family history of dementia and other risk factors for dementia, including diabetes (8.3%), hypertension (53.1%), head injury (15.4%), and at least one APOE ε4 allele (30.3%).

Smoking exposure was "common," with 49 current smokers (median pack-years, 47.3) and 231 former smokers (median pack-years, 24.5) at baseline.

Former smokers were found to be significantly older than current smokers and significantly less likely to be female, while never-smokers were significantly more likely to report a family history of dementia.

At the end of follow-up, mean age for the overall cohort was 84.7 (7.3) years.

For the subset of those participants (n = 111) who transitioned to dementia vs those who died without dementia, the mean age was 85.5 (7.0) vs 86.3 (7.5) years, respectively.

Higher Mortality

Mild cognitive impairment (MCI) was diagnosed in 30.9% of participants who did not die before or develop dementia and of these, 21.8% died with a diagnosis of MCI.

When the researchers used a Cox model, they found that smoking was a risk for incident dementia for former smokers vs never-smokers (adjusted hazard ratio [HR] 1.64; 95% confidence interval [CI]: 1.09, 2.46] and for former smokers vs current smokers [HR, 1.20; 95% CI: 0.50, 2.87).

Having 10 or more pack-years of smoking history (n = 215) vs never smoking (n = 251) was also associated with dementia (HR 1.65; 95% CI: 1.08, 2.53), while less than 10 pack-years (n = 65) vs never smoking yielded an adjusted HR of 1.39 (95% CI: 0.76, 2.58).

However, when the researchers applied the competing risk analysis of death without dementia, current smoking was no longer a significant risk for dementia.

The adjusted subdistribution hazard ratio (sHR) was 1.21 (95% CI: 0.81, 1.80) for former smokers and 0.70 (95% CI: 0.30, 1.64) for current smokers, while the adjusted sHR for at least 10 pack-years of exposure was 1.14 (95% CI: 0.76, 1.71).

By contrast, baseline current smoking increased the incidence of death without dementia (sHR= 2.38; 95% CI: 1.52, 3.72).

Factors that significantly increased the incidence of death before dementia for baseline current smokers were female sex, no APOE ε4 allele, and low educational attainment.

When examining cognitive impairment, the researchers found that the adjusted HR derived from the Cox model for former smokers vs never-smokers was 1.33 (95% CI: 0.96, 1.83) and 0.80 (95% CI: 0.39, 1.61) for current vs never-smokers.

However, in the competing risk analysis, the sHR for former smokers vs never-smokers was 0.97 (95% CI: 0.70, 1.34) and 0.51 (95% CI: 0.26, 0.99) for current smokers vs never-smokers.

The apparent "protective effect" of smoking among current smokers was actually increased by a higher incidence of mortality in smokers as compared with never-smokers.

No Increased Neuropathology

Of the 531 participants in the study, 350 died and 302 (86.3%) were autopsied.

The average age at death among autopsied never-smokers vs "ever-smokers" (current or former) was 88.6 (±7.1) years and 86.2 (±7.5) years, respectively.

Compared to never-smokers, ever-smokers were less likely to have higher levels neuropathology, with the exception of lacunar infarcts.

Adjustment for age at death, sex, and APOE ε4 allele did not alter these associations.

The top four causes of death were heart disease, cancer, pneumonia, and stroke, representing 72% of all reported causes of death.

Among those cases whose cause of death was reported, 76.6% with a smoking history died of the top 4 causes, compared with 65.9% with no smoking history (odds ratio, 1.70; 95% CI: 0.90, 3.19).

Abner stated that she was not surprised by the results of the competing risk analysis, since her group has "done a fair amount of work in the area of competing risks" and that she was also not surprised that "Alzheimer-type pathology was not increased in the smokers."

However, she acknowledged being surprised by the other aspects of the neuropathological data analysis, since she expected to see more cerebrovascular pathology in the current smokers and former smokers, compared to never-smokers.

"I would like to see more autopsy-based studies of smoking before drawing firm conclusions," she added.

Healthy Choices

Commenting on the study for Medscape Medical News, Keith Fargo, PhD, director of scientific programs and outreach at the Alzheimer's Association, cautioned, "Smoking is bad for you [because] it leads to earlier death and a wide variety of other negative and fatal outcomes."

Fargo, who was not involved with the current study, emphasized that even "if one is fortunate to escape smoking-related heart disease and cancer, you are nonetheless more likely than non-smokers to develop dementia."

The Alzheimer's Association "encourages everyone who smokes to stop, and everyone who doesn't [smoke] not to start," he said.

He noted that additional brain-healthy lifestyle choices are included in the Alzheimer's Association's 10 Ways to Love Your Brain .

Abner added that her group plans to "conduct similar analyses in other study populations that also have robust autopsy data on their cohort participants."

They are also interested in other risk factors that have a strong mortality risk, such as hypertension and diabetes, she said.

This research was partially funded with support from the National Institute on Aging, the National Institute of Nursing Research, and the National Center for Advancing Translational Sciences. Abner and Fargo have disclosed no relevant financial relationships. The other study authors' disclosures are listed at:

J Alzheimers Dis. Published online March 2, 2019. Abstract

For more Medscape Neurology news, join us on Facebook and Twitter


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.