Anxiety and aging: A marker of brain changes and potential treatment target to promote brain health

Almost 7 million Americans have Alzheimer's dementia. Approximately 20% of middle-aged women and 10% of middle-aged men will eventually develop Alzheimer's dementia in their lifetime, usually occurring after age 65.¹ Given the aging of the population, absent incident prevention or efforts that slow the course of the illness there will be close to 14 million Americans living with the disease by 2060.¹ Although new treatments exist, there is no currently available scalable cure that is cost-effective (accounting for Quality-adjusted life-years (QALYs), incremental cost-effectiveness ratios, and costs), nor will one likely be launched within the next decade. Thus, there needs to be a greater focus on modifiable risk factors (12 of which account for 40% of worldwide dementias)² to prevent new cases in the United States and across the globe.

Anxiety disorders are common: they have a lifetime prevalence of approximately 34% in the United States³ and are the second most common neuropsychiatric disease after depression.⁴ Anxiety is linked with higher rates of both depression⁵ and addiction⁶; reflects a state of both psychic and physical stress⁷; and is linked with pro-inflammatory⁸ states, cognitive impairment,⁹ cardiovascular disease,¹⁰ and all-cause mortality.¹¹ It is also treatable. Elucidating the interaction between having anxiety in late-life and rates of incident dementia may add to the list of modifiable risk factors for cognitive decline and neurodegenerative diseases.

This issue of the journal includes a study by Khaing and colleagues entitled “The effect of anxiety on all-cause dementia: a longitudinal analysis from the Hunter Community Study.”¹² Their communication describes the association of chronic versus resolved versus new onset anxiety on subsequent diagnosis of dementia. The investigators hypothesized that both the (1) chronicity of anxiety and (2) age of exposure to anxiety would be linked with all-cause dementia risk. The sample (n = 2132, mean age = 76) was an existing cohort of community-dwelling Australians who were recruited between 2004 and 2007. Wave 2 and Wave 3 assessments were completed at 5-year intervals after Wave 1. The natural history of anxiety was categorized as (1) Chronic Anxiety (present at Wave 1 and Wave 2); (2) Resolved Anxiety (present only at Wave 1); and (3) New Anxiety (present only at Wave 2). The primary outcome was incident all-cause dementia up to 13 years after Wave 1. Sixty-four participants (3%) were diagnosed with dementia with the average onset at year 10. Chronic Anxiety (HR = 2.80) and New Anxiety (HR = 3.20) at Wave 2 were both associated with increased risk of all-cause dementia. Resolved Anxiety was not linked with increased risk of all-cause dementia. When the analysis was stratified by age, those aged 60–70 with both Chronic Anxiety (HR = 4.58) and New Anxiety (HR = 7.21) experienced higher risk of dementia. This was not observed in the two other age groups: 71–80 years and >80 years.

The authors interpret and describe these observations with equipoise and care. Sensitivity analyses accounted for (1) participants censored during the first 5 years from baseline and (2) missing data, using multiple imputation and observed case analysis, revealed similar observations from the main analysis, although the missing data sensitivity analysis had an attenuated (but still significant) effect size. This approach strengthens the rigor and stability of the main observations, and the authors thoughtfully describe the five major methodological limitations of the study.

Perhaps the biggest question is: What does anxiety in the years before the diagnosis of all-cause dementia represent? Is it a prodrome of the neurodegenerative disease (meaning an early behavioral marker of impending cognitive decline and subsequent neuropsychiatric changes) or is it a true risk factor that is mechanistically linked with the development or hastening of neurodegeneration and cognitive and behavioral changes? A cohort study such as this can establish an association but cannot disambiguate prodrome from risk factor. As eloquently described by Geoffroy and Scott, “The key difference between a ‘prodrome’ and ‘risk syndrome’ is that the former is primarily a predictor of the onset of an episode of the mental disorder under examination, while the latter is primarily a predictor of the overall likelihood that someone will experience a first onset of a disorder (compared to no disorder or another disorder).”¹³ While the methodology of Khaing et al uses both a prospective and case versus control approach, it is challenging to assign causality (e.g., as a risk factor) without including a translational probe of brain changes that includes a longer window of surveillance with more frequent assessments.

Regardless of the prodrome or risk factor conundrum, Khaing et al. add to the emerging science linking anxiety, neurodegeneration, and dementia. We agree with the conservative closing statement of the article: “Therefore, these findings support anxiety as a potential modifiable risk factor for dementia and point to the possible role of managing anxiety in middle-aged and “young” older adults to reduce the risk of dementia in later life.¹²” Readers may consider middle-aged and late-life anxiety as a putative (and prevalent) behavioral marker of age-related brain changes—much as we consider depression. The next question—does treating anxiety in late-life preserve cognition—may be best answered using a placebo-controlled randomized clinical trial in which patients with chronic anxiety or late-onset anxiety are randomized to evidence-based pharmacological and psychosocial interventions¹⁴ for anxiety versus placebo to determine if successful treatment of anxiety attenuates the risk for dementia.

Dr. Karp and Dr. Lenze conceived of the work and participated equally in the writing of this editorial.

JFK: research support from Janssen, potential for equity in Aifred Health, and scientific advising for Biogen. EJL: receipt of grant support (nonfederal) from COVID Early Treatment Fund, Mercatus Center Emergent Ventures, Balvi Philanthropic Fund, the Patient-Centered Outcomes Research Institute, Janssen, MagStim (support in kind). He also receives consulting fees from Merck, IngenioRx, Boeringer-Ingelheim, Prodeo, and Pritikin ICR. He has a recent patent applied for the use of fluvoxamine in the treatment of COVID-19.

None.