Edward Zamrini

Objective: Effective strategies to prevent Alzheimer’s disease and related dementias (ADRD) are limited. Implementing our Medication-Wide Association Study (MWAS+), we highlighted varenicline as a candidate. This study evaluated the association between varenicline use and risk of ADRD relative to nicotine replacement therapy (NRT). Methods: Using the All of Us Research Database, we identified adults aged 50 years or older with available lifestyle survey responses regarding tobacco use history who initiated smoking cessation therapy with either varenicline or NRT between 2006 and 2023. We limited our cohort to incident monotherapy users and designated the index date as the first prescription. Individuals with baseline ADRD or no follow-up were excluded. Baseline characteristics included demographics, comorbidities, BMI, and prior medication exposures. Outcomes were incident ADRD and a composite outcome of ADRD or death. Findings: The cohort consisted of 9,584 individuals (1,801 varenicline users and 7,783 NRT users). After propensity score matching, there were 1,758 individuals in each group with balanced baseline characteristics. Varenicline use was associated with a reduced risk of ADRD (HR 0.516, 95% CI: 0.353–0.754) and the composite outcome ADRD or death (HR 0.657, 95% CI: 0.501–0.861). Kaplan–Meier curves also demonstrated varenicline’s sustained benefits through the study period. Implications: Varenicline use is associated with a lower risk of ADRD among older adults receiving smoking cessation therapy. Further investigation of varenicline as a potential strategy for ADRD prevention is warranted. These findings underscored the value of explainable AI-guided approaches to drug repurposing. Highlight: Artificial Intelligence (AI) Medication-Wide Association Study (MWAS+) identified varenicline as candidate for ADRD prevention. Varenicline linked to lower ADRD risk vs NRT in propensity score matched cohort. Consistent results for ADRD and composite outcome ADRD or death.

Background: Alzheimer’s disease and related dementias (ADRD) affect over 7.2 million Americans aged 65 and older, with the APOE ε4 allele representing the strongest known genetic risk factor. Physical activity (PA) has been associated with reduced dementia risk, but its interaction with APOE genotype remains poorly characterized in large, genomically informed cohorts. Methods: We conducted a retrospective cohort analysis using linked genomic, survey, and longitudinal electronic health record data from the VA Million Veteran Program (MVP). Veterans aged ≥65 at enrollment with available APOE genotype data, at least 10 years of prior VA medical history, and a completed Lifestyle Survey were included. Individuals with a pre-existing ADRD diagnosis were excluded, yielding a final cohort of 137,593 veterans. Self-reported vigorous physical activity (PA) was ascertained from the MVP Lifestyle Survey and coded as both a four-level ordinal variable and a binary active/inactive classification. The primary composite outcome was incident ADRD or all-cause mortality. Cox proportional hazards models were fitted adjusting for age, sex, race/ethnicity, and baseline comorbidities. A multiplicative interaction term between APOE ε4 allele count and PA level was included to formally test for effect modification. Results: Over a median follow-up of 84 months, 65,628 participants (47.7%) experienced the composite outcome. Each additional APOE ε4 allele was associated with a 19% increase in risk (aHR = 1.19, 95% CI 1.17–1.21), while active individuals had a 33% lower risk compared to inactive individuals (aHR = 0.67, 95% CI 0.66–0.68). A statistically significant interaction between APOE ε4 burden and PA was identified (p < 0.005). Stratified analyses demonstrated that the protective association of PA was present across all genotype groups, with ε4 homozygotes demonstrating a 22% risk reduction among active individuals. Conclusions: In this large veteran cohort, vigorous PA was independently and significantly associated with reduced risk of incident ADRD or death across all APOE genotype groups, with the greatest absolute benefit observed among individuals at highest genetic risk. These findings support the prioritization of PA as a targeted preventive strategy, particularly for APOE ε4 carriers.

Mid‐life high blood pressure (BP) is a risk factor for Alzheimer's disease (AD). CERAD amyloid β (Aβ) plaques, Braak tau neurofibrillary tangles, and Thal Aβ plaque location are major scoring systems for quantifying neuropathological features of AD. We examined the association of late‐life systolic BP (SBP) with CERAD, Braak, and Thal in the National Alzheimer's Coordinating Center (NACC) Neuropathology Dataset. Of 1978 participants with data on CERAD, 762 had scores 0–1 (none to sparse) and 1216 had 2–3 (moderate to frequent). Of 1947 with data on Braak, 411 had stages 0–II (normal to mild) and 1536 had III–VI (moderately to very severe). Of 2132 with data on Thal, 438 had phases 0–I, 428 II–III, and 1266 IV–V. Using the mean of the last four SBP before death, SBP was categorized into P = 0.03) and 1.26 (0.89, 1.78, P = 0.20), respectively. Similar association was observed for Thal II–III and IV–V. These associations essentially remained unchanged after additional adjustment for APOE and Lewy Body pathology. These findings suggest that higher late‐life SBP is associated with markers of presence and severity of neuropathological features of AD. Further studies with larger sample sizes are necessary to confirm the findings.

BACKGROUND: Hypertension and physical inactivity are risk factors for stroke. The effect of cardiorespiratory fitness (CRF) on stroke risk in patients with hypertension has not been assessed. We evaluated stroke incidence in patients with hypertension according to CRF and changes in CRF. METHODS: We included 483 379 patients with hypertension (mean age±SD; 59.4±9.0 years) and no evidence of unstable cardiovascular disease as indicated by a standardized exercise treadmill test. Patients were assigned to 5 age- and sex-specific CRF categories based on peak metabolic equivalents achieved at the initial exercise treadmill test and in 4 categories based on metabolic equivalent changes over time (n=110 576). Multivariable Cox models, adjusted for age, and comorbidities were used to estimate hazard ratios and 95% CIs for stroke risk. RESULTS: During a median follow-up of 10.6 (interquartile range, 6.6–14.6) years, 15 925 patients developed stroke with an average yearly rate of 3.1 events/1000 person-years. Stroke risk declined progressively with higher CRF and was 55% lower for the High-fit individuals (hazard ratio, 0.45 [95% CI, 0.42–0.48]) compared with the Least-fit. Similar associations were observed across the race, sex, and age spectra. Poor CRF was the strongest predictor of stroke risk of all comorbidities studied (hazard ratio, 2.24 [95% CI, 2.10–2.40]). Changes in CRF reflected inverse and proportional changes in stroke risk. CONCLUSIONS: Poor CRF carried a greater risk than any of the cardiac risk factors in patients with hypertension, regardless of age, race, or sex. The lower stroke risk associated with improved CRF suggests that increasing physical activity, even later in life, may reduce stroke risk.

High blood pressure (BP) and cerebral amyloid angiopathy (CAA) are two common risk factors for intracranial hemorrhage, potentially leading to cognitive impairment. Less is known about the relationship between BP and CAA, the examination of which was the objective of this study. We analyzed data from 2510 participants in the National Alzheimer’s Coordinating Center (NACC) who had information on longitudinal BP measurements before death and on CAA from autopsy. Using the average of four systolic BPs (SBPs) prior to death, SBP was categorized into three groups: n = 435), 120–139 mmHg (n = 1335), and ≥140 mmHg (n = 740). CAA was diagnosed using immunohistochemistry in 1580 participants and categorized as mild (n = 759), moderate (n = 529), or severe (n = 292). When adjusted for age at death, sex, APOE genotype, Braak, CERAD, antihypertensive medication use, and microinfarcts, the odds ratios (95% CIs) for CAA associated with SBPs of 120–139 and ≥140 mmHg were 0.91 (0.74–1.12) and 1.00 (0.80–1.26), respectively. Findings from predictor effect plots show no variation in the probability of CAA between the three SBP categories. Microbleeds had no association with CAA, but among those with SBP ≥ 130 mmHg, the proportion of those with microbleeds was numerically greater in those with more severe CAA (p for trend, 0.084). In conclusion, we found no evidence of an association between SBP and CAA. Future studies need to develop non-invasive laboratory tests to diagnose CAA and prospectively examine this association and its implication on the pathophysiology and outcome of Alzheimer’s disease.

Anti-amyloid therapies (AATs), such as anti-amyloid monoclonal antibodies, are emerging treatments for people with early Alzheimer’s disease (AD). AATs target amyloid β plaques in the brain. Amyloid-related imaging abnormalities (ARIA), abnormal signals seen on magnetic resonance imaging (MRI) of the brain in patients with AD, may occur spontaneously but occur more frequently as side effects of AATs. Cerebral amyloid angiopathy (CAA) is a major risk factor for ARIA. Amyloid β plays a key role in the pathogenesis of AD and of CAA. Amyloid β accumulation in the brain parenchyma as plaques is a pathological hallmark of AD, whereas amyloid β accumulation in cerebral vessels leads to CAA. A better understanding of the pathophysiology of ARIA is necessary for early detection of those at highest risk. This could lead to improved risk stratification and the ultimate reduction of symptomatic ARIA. Histopathological confirmation of CAA by brain biopsy or autopsy is the gold standard but is not clinically feasible. MRI is an available in vivo tool for detecting CAA. Cerebrospinal fluid amyloid β level testing and amyloid PET imaging are available but do not offer specificity for CAA vs amyloid plaques in AD. Thus, developing and testing biomarkers as reliable and sensitive screening tools for the presence and severity of CAA is a priority to minimize ARIA complications.

Mid-life high blood pressure (BP) is a risk factor for cerebral microinfarcts. Less is known about the relationship between late-life BP and cerebral microinfarcts, the examination of which is the objective of the current study. This case–control study analyzed data from 551 participants (94.6% aged ≥80 years; 58.6% women) in the Adult Changes in Thought (ACT) study who had autopsy data on microinfarcts and four values of systolic and diastolic blood pressure (SBP and DBP) before death. Using the average of four values, SBP was categorized using 10 mmHg intervals; a trend was defined as a ≥10 mmHg rise or fall from the first to fourth values (average gap of 6.5 years). Multivariable-adjusted regression models were used to examine the associations of BP and microinfarcts, adjusting for age, sex, last BP-to-death time, APOE genotype, and antihypertensive medication use. Microinfarcts were present in 274 (49.7%) participants; there were multiple in 51.8% of the participants, and they were located in cortical areas in 40.5%, subcortical areas in 29.6%, and both areas in 29.9% of the participants. All SBP categories (reference of 100–119 mmHg) and both SBP trends were associated with higher odds of both the presence and number of microinfarcts. The magnitude of these associations was numerically greater for subcortical than cortical microinfarcts. Similar associations were observed with DBP. These hypothesis-generating findings provide new information about the overall relationship between BP and cerebral microinfarcts in octogenarians.

Cerebral microinfarcts are associated with cognitive impairment and dementia. Small vessel diseases such as cerebral arteriolosclerosis and cerebral amyloid angiography (CAA) have been found to be associated with microinfarcts. Less is known about the associations of these vasculopathies with the presence, numbers, and location of microinfarcts. These associations were examined in the clinical and autopsy data of 842 participants in the Adult Changes in Thought (ACT) study. Both vasculopathies were categorized by severity (none, mild, moderate, and severe) and region (cortical and subcortical). Odds ratios (OR) and 95% CIs for microinfarcts associated with arteriolosclerosis and CAA adjusted for possible modifying covariates such as age at death, sex, blood pressure, APOE genotype, Braak, and CERAD were estimated. 417 (49.5%) had microinfarcts (cortical, 301; subcortical, 249), 708 (84.1%) had cerebral arteriolosclerosis, 320 (38%) had CAA, and 284 (34%) had both. Ors (95% CI) for any microinfarct were 2.16 (1.46–3.18) and 4.63 (2.90–7.40) for those with moderate (n = 183) and severe (n = 124) arteriolosclerosis, respectively. Respective Ors (95% CI) for the number of microinfarcts were 2.25 (1.54–3.30) and 4.91 (3.18–7.60). Similar associations were observed for cortical and subcortical microinfarcts. Ors (95% Cis) for the number of microinfarcts associated with mild (n = 75), moderate (n = 73), and severe (n = 15) amyloid angiopathy were 0.95 (0.66–1.35), 1.04 (0.71–1.52), and 2.05 (0.94–4.45), respectively. Respective Ors (95% Cis) for cortical microinfarcts were 1.05 (0.71–1.56), 1.50 (0.99–2.27), and 1.69 (0.73–3.91). Respective Ors (95% Cis) for subcortical microinfarcts were 0.84 (0.55–1.28), 0.72 (0.46–1.14), and 0.92 (0.37–2.28). These findings suggest a significant association of cerebral arteriolosclerosis with the presence, number, and location (cortical and subcortical) of microinfarcts, and a weak and non-significant association of CAA with each microinfarct, highlighting the need for future research to better understand the role of small vessel diseases in the pathogenesis of cerebral microinfarcts.

Deep neural network (DNN) is a powerful technology that is being utilized by a growing number and range of research projects, including disease risk prediction models. One of the key strengths of DNN is its ability to model non-linear relationships, which include covariate interactions. We developed a novel method called interaction scores for measuring the covariate interactions captured by DNN models. As the method is model-agnostic, it can also be applied to other types of machine learning models. It is designed to be a generalization of the coefficient of the interaction term in a logistic regression; hence, its values are easily interpretable. The interaction score can be calculated at both an individual level and population level. The individual-level score provides an individualized explanation for covariate interactions. We applied this method to two simulated datasets and a real-world clinical dataset on Alzheimer’s disease and related dementia (ADRD). We also applied two existing interaction measurement methods to those datasets for comparison. The results on the simulated datasets showed that the interaction score method can explain the underlying interaction effects, there are strong correlations between the population-level interaction scores and the ground truth values, and the individual-level interaction scores vary when the interaction was designed to be non-uniform. Another validation of our new method is that the interactions discovered from the ADRD data included both known and novel relationships.