Wildfire Smoke/Dementia Risk Paper Retracted

Medscape

3 days ago

Research published last year linking wildfire smoke exposure to increased dementia risk has been retracted by the study's authors after a single stroke coding error was discovered in the original analysis.
Using amended data, the association between dementia and exposure to fine particulate matter < 2.5 microns in diameter (PM2.5) from wildfire sources is no longer significant, the researchers determined.
Investigators were 'very, very disappointed' by the 'unfortunate error,' lead author Joan A. Casey, PhD, associate professor, School of Public Health, University of Washington, Seattle, told Medscape Medical News.
However, Casey maintains that the updated data do not alter the overall conclusion that wildfire exposure is 'pretty consistently' associated with increased dementia risk. She noted a continued positive association between the 3-year average of wildfire exposure and incident dementia.
The notice of retraction with revisions was published online on June 30 in JAMA Neurology .
A Single Keystroke Error
The original study — whose preliminary findings were presented at the 2024 Alzheimer's Association International Conference and reported by Medscape Medical News at that time — was published online on November 25, 2024, in JAMA Neurology .
That analysis included 1,223,107 million Kaiser Permanente Southern California members, individuals aged 60 years or older who were tracked from 2009 to 2019. Of these, 80,884 developed dementia.
The original study showed an 18% increased risk for dementia for every 1 µg/m3 increase in the 3-year average wildfire PM2.5 concentration (odds ratio [OR], 1.18; 95% CI, 1.03-1.34), after adjusting for age, sex, race and ethnicity, marital status, smoking, and economic status.
For the study, researchers had identified dementia using International Classification of Diseases codesfor Alzheimer's disease, frontotemporal dementia, Lewy body dementia, nonspecific dementia, vascular dementia, and Parkinson's disease (PD) dementia.
Criteria for PD dementia was supposed to require a PD diagnosis and a secondary dementia diagnosis. But in preparing for a follow-up study, researchers realized the original analysis had a coding error.
'Where there should have been an 'and' symbol in our code, there was an 'and/or' symbol,' Casey explained . 'We included people as cases that either had PD or had one of the secondary dementia codes, when we only wanted to include people who had both of those things.'
'It's literally one keystroke in like thousands of lines of code,' said Casey. But the error resulted in the inclusion of 5413 individuals without dementia who were classified as dementia cases.
Researchers also incorrectly excluded 5345 individuals with prevalent PD at baseline, which slightly altered the comparison group, said Casey.
'We have some people who should have been eligible as comparators in the analysis, and they were excluded but shouldn't have been,' she added.
Recalculated Analyses
The investigators have now recalculated all analyses using the corrected total cohort of 1,228,452 individuals and revised the number of people who developed dementia to 75,471.
Not only do the corrections affect the cohort composition but also do affect odds ratios (ORs) and CIs. The previously reported finding of an association between 1-μg/m3 higher 3-year average of wildfire PM2.5 exposure and incident dementia is no longer statistically significant, with a smaller effect size (revised OR, 1.12; revised 95% CI, 0.98-1.28).
However, Casey noted that results for multiple consecutive days of high levels of wildfire smoke exposure 'are unchanged' in the new analysis and the association between wildfire exposure and dementia risk among certain racial and ethnic group, including Native Americans and Alaska Natives, remains statistically significant in the new analysis, although not as robust as in the original reporting, said Casey.
The updated analysis still showed an association between exposure to non–wildfire fine particulate matter and incident dementia with some alterations(original analysis: OR, 1.01; 95% CI, 1.01-1.02; corrected analysis: OR, 1.007; 95% CI, 1.002-1.011).
However, the revised analysis affected results for non-wildfire exposure in certain subgroups. While the original analysis showed a significant association between 3-year mean non-wildfire PM2.5 exposure and incident dementia in both sexes and in high- and low-poverty neighborhoods, the association in men and in high-poverty communities no longer meets the conventional threshold for statistical significance ( P < .05).
New Data, Same Story
Ultimately, the updated data 'don't change things a huge amount,' Casey said.
'As a whole, the story remains quite similar. We still need more research, but to me, this points towards there being a relationship between long–term wildfire smoke exposure and incident dementia,' she added.
The study also continues to highlight the importance of thinking about environmental risk factors for neurological conditions, Casey added.
'We should be having conversations with patients about ways they can protect themselves from these types of exposure, especially if they're coming in already with things like mild cognitive impairment,' she said.
Wildfire smoke, which results from burning organic materials, contains elevated concentrations of polycyclic aromatic hydrocarbons. The smoke may enter the central nervous system via the olfactory nerve and cross the blood-brain barrier, possibly triggering neuroinflammation. However, much of this is speculation.
'For us, the big next steps are trying to understand the biological underpinnings and what is mediating this relationship,' said Casey.
Three years may not be the most relevant exposure window for the onset of dementia, Casey said. 'What's probably happening here is the exposure is hastening a process that was already set in motion and unveiling dementia sooner,' she noted.
The coding error was discovered when the researchers were updating the cohort to investigate a longer duration of exposure.
'We wanted to look back 10 years and see what those relationships looked like,' she said. 'I think that will be really important, especially as we continue to see ongoing and continued wildfire smoke exposure in our population.'
Wildfires have become more severe and prolonged, and they're occurring earlier in the year than in the past. The fires are also now more frequent. Studies suggest exposure to wildfires has increased 27-fold over the past decade in the US.
Experts attribute changes to wildfire characteristics to warmer temperatures, earlier snowmelt, and prolonged dry periods.
'It's so important for neurologists to engage in the conversation about climate change and how it affects the health of their patients,' said Casey. 'To me, the biggest overarching public health message here is that we need strong climate policy, which is going to be the thing that slows these types of exposures.'
As previously reported by Medscape Medical News , wildfire exposure has been linked to an increased risk to several neurological disorders. One study showed exposure to wildfire smoke reduces cognitive performance in adults while other research has linked it to lower academic performance in students.
Additional research has tied air pollution, including wildfire smoke, to an increased risk for attention deficit/hyperactivity disorder in children and youth as well as an increased risk for PD and multiple sclerosis in adults.
An 'Unfortunate' Error
Asked to comment, Anthony White, PhD, senior group leader, Cellular and Molecular Neurodegeneration, QIMR Berghofer Medical Research Institute, Queensland, Australia, said while the retraction is 'unfortunate,' the authors should be commended for applying the necessary scientific rigor to update their work.
'The study obviously involves an enormous level of complex data and assessments, and the error reported should not be seen as any reflection of poor science by the research team,' White said. 'On the contrary, they have done extremely well to pick up the original error and modify the study.'
He noted how difficult it is to obtain statistically relevant results in association studies such as this one as they look at a broad range of dementia types and years after a diagnosis. As well, researchers need to separate a range of pollution sources into wildfire and non-wildfire exposure.
White also noted that the study only assessed the relationship between dementia and PM2.5 levels, which represent just one component of wildfire smoke.
The association between wildfires and dementia is supported by other research, said White and pointed to a 2023 study in JAMA Internal Medicine that showed positive links between dementia and wildfires.
Other studies show strong ties between air pollution and dementia, he added. 'We know that wildfire smoke is a particularly potent form of air pollution.'
Heather M. Snyder, PhD, senior vice president of medical and scientific relations, Alzheimer's Association, also weighed in on the new development.
She, too, noted several other 'credible studies' link wildfire to dementia, as well as overall health including heart and metabolic diseases. 'All of the literature today suggests an association. Whether that's the full extent of the connection, or if there is a cause and effect, is still not known.'