The Clotbust Review

Associations Between Cerebral Amyloid Angiopathy, Cognitive Impairment, and Depressive Symptoms

· DOI: 10.1212/wnl.0000000000218354 · PMC13446197 · stroke deep-dive cerebral amyloid angiopathy cognition depression

Stylized illustration of Associations Between Cerebral Amyloid Angiopathy, Cognitive Impairment, and Depressive Symptoms
AI-generated editorial illustration. Generated for editorial use.

Clinical Question (PICO)

Population: Adults with probable cerebral amyloid angiopathy (CAA) and cognitively normal controls recruited from two Canadian clinics and the community. Exposure/comparator: CAA versus no CAA; this was an observational comparison, not a treatment trial. Outcomes: depressive symptoms measured by the 15-item Geriatric Depression Scale (GDS-15), cognitive performance in episodic memory, executive function, and processing speed, and associations between depressive symptoms and CAA-related MRI markers. The prespecified objectives were to examine the CAA–depression and CAA–cognition relationships, the potential mediating role of depressive symptoms, and whether MRI markers were associated with depressive symptom severity.

Bottom Line

Among 168 complete cases, participants with probable CAA had markedly higher depressive symptom burden than cognitively normal controls: adjusted odds of GDS-15-defined “possible depression” were 15.71 (95% CI 4.26–80.05; p < 0.001), and adjusted GDS-15 scores were 2.71-fold higher (95% CI 2.10–3.51; p < 0.001). Depressive symptoms were associated with worse cognition, but mediation accounted for only 11% of the CAA association with episodic memory, 9% with executive function, and 2% with processing speed. The study supports routine symptom screening in CAA; it does not establish that treating depression will reverse CAA-related cognitive impairment.

Design

  • Trial type: Cross-sectional observational analysis of baseline data from a two-center prospective longitudinal cohort; no randomization or blinding.
  • N: 168 complete cases for the primary analyses (182 initially enrolled)
    • Probable CAA: 85
    • Cognitively normal controls: 83
  • Randomization: None. The comparison was defined by CAA status; sex interaction was explored, and CAA presentations were compared descriptively.
  • Setting: Memory, stroke-prevention, and geriatric clinics in Calgary and Edmonton, Canada; controls were recruited through community advertising.
  • Enrollment: 2011–2020.
  • Mean follow-up: None for this analysis; only baseline cross-sectional measurements were analyzed.
  • Analysis: Complete-case multivariable logistic, Poisson, linear, and negative-binomial regression adjusted for age, sex, and years of education; causal mediation models used 1,000 simulations. MRI-marker analyses included 81 CAA participants.
  • Primary outcome: Association of CAA status with depressive symptoms and the proportion of the CAA–cognition association statistically mediated by depressive symptoms. Cognitive domain scores were age-, sex-, and education-adjusted z-score composites.

Population

Inclusion Criteria

  • Adults aged at least 55 years recruited to the Functional Assessment of Vascular Reactivity cohort.
  • Probable CAA identified from stroke-prevention, memory, or geriatric clinics using Boston Criteria version 2.0.
  • Cognitively normal controls aged 60–85 years, with no stroke or other CNS disease and Montreal Cognitive Assessment score ≥26.

Exclusion Criteria

  • Not fluent in English or French.
  • Ongoing alcohol or drug use affecting cognition.
  • Inability to undergo MRI.

Baseline Characteristics (overall or representative arm)

  • CAA participants: mean age 73.5 ± 7.05 years; 35.3% female; 14.2 ± 3.16 years of education.
  • Controls: mean age 68.8 ± 7.08 years; 62.7% female; 15.5 ± 3.16 years of education.
  • CAA participants had more hypertension (65.9% vs 22.9%) and worse mean domain scores than controls: episodic memory z score −1.05 vs 0.46, executive function −1.18 vs 0.28, and processing speed −0.86 vs 0.48 (all p < 0.001).
  • CAA presentations included lobar ICH (33), cognitive impairment (27), transient focal neurologic episodes (23), and CAA-related inflammation (2). The current analysis was not restricted to patients with prior ICH.

Interventions

  • CAA exposure: Probable CAA defined clinically by MRI markers and Boston Criteria version 2.0; no study treatment was assigned.
  • Control condition: Cognitively normal community controls without stroke or other CNS disease; no study treatment was assigned.
  • Depressive symptom assessment: Self-reported GDS-15, with a score ≥5 labeled “possible depression.” This is a screening threshold, not a DSM diagnosis.

Outcomes

Primary Outcome (depressive symptoms, cognition, and mediation):

  • “Possible depression” occurred in 20 CAA participants versus 3 controls. After adjustment for age, sex, and education, CAA was associated with higher odds of possible depression (OR 15.71, 95% CI 4.26–80.05; p < 0.001) and a higher expected GDS-15 count (CR 2.71, 95% CI 2.10–3.51; p < 0.001).
  • Possible depression was associated with poorer episodic memory (β −1.03, 95% CI −1.53 to −0.53; p < 0.001), executive function (β −1.11, 95% CI −1.65 to −0.58; p < 0.001), and processing speed (β −0.73, 95% CI −1.24 to −0.21; p = 0.006).
  • In mediation using the binary possible-depression definition, the indirect effect accounted for 11% of the CAA effect on episodic memory (β −0.14, 95% CI −0.30 to −0.03; p = 0.014), 9% on executive function (β −0.14, 95% CI −0.30 to −0.02; p = 0.016), and 2% on processing speed (β −0.04, 95% CI −0.18 to 0.08; p = 0.56). Thus, the estimated direct CAA component remained 89%, 91%, and 97%, respectively.
  • In a sensitivity mediation using the continuous GDS-15 count, the indirect effect was significant for episodic memory and represented approximately 12% of its total effect (β −0.17, 95% CI −0.35 to −0.03; p = 0.01), but was not significant for executive function (6%; p = 0.11) or processing speed (7%; p = 0.10).
  • No prespecified treatment subgroup was tested. Median GDS-15 scores did not differ across CAA presentation groups—ICH, transient focal neurologic episodes, and cognitive impairment/inflammation (p = 0.91)—and the sex interactions were not significant, although power for interaction testing was limited.

Secondary Outcomes:

  • Among 81 CAA participants with usable GDS-15 and MRI data, higher depressive symptom counts were associated with lower mean cortical thickness (CR 1.33 per SD decrease, 95% CI 1.09–1.62; p = 0.005), any cortical superficial siderosis (cSS; CR 2.04, 95% CI 1.35–3.09; p < 0.001), and higher CAA small-vessel-disease score (CR 1.16, 95% CI 1.00–1.35; p = 0.047).
  • Associations were not statistically significant for centrum-semiovale perivascular-space grade (CR 0.95, 95% CI 0.76–1.19; p = 0.67), cerebral microbleed count (CR 1.00, 95% CI 0.997–1.001; p = 0.28), or white-matter-hyperintensity volume (CR 1.01, 95% CI 1.00–1.02; p = 0.06).

Adverse Events / Safety:

  • No intervention was administered, so treatment adverse events, symptomatic ICH, and treatment-related mortality were not evaluated. The cohort included 33 participants whose CAA presentation involved lobar ICH.

Figures

Distribution of GDS-15 scores in CAA and controls
Figure 1. Distribution of GDS-15 scores in cognitively normal controls and participants with CAA. Individual scores are shown with jitter, and large dots indicate group means. “Possible depression” was defined as a GDS-15 score ≥5.

Source: PMC PMC13446197 — WNL-2026-300393f1.jpg. Click image to expand.

Mediation using possible depression
Figure 2. Direct and indirect effects of CAA on episodic memory, executive function, and processing speed using “possible depression” as the mediator. Solid arrows indicate statistically significant pathways; dashed arrows indicate nonsignificant pathways. Estimates are observational and do not prove causation.

Source: PMC PMC13446197 — WNL-2026-300393f2.jpg. Click image to expand.

Mediation using GDS-15 total score
Figure 3. Direct and indirect effects of CAA on the three cognitive domains using the continuous GDS-15 total score as the mediator. Solid arrows indicate statistically significant pathways; dashed arrows indicate nonsignificant pathways. Estimates are observational and do not prove causation.

Source: PMC PMC13446197 — WNL-2026-300393f3.jpg. Click image to expand.

Criticisms

  • The cross-sectional design prevents determining whether CAA causes depression, depression contributes to cognitive decline, cognition drives depressive symptoms, or the relationships are bidirectional.
  • “Possible depression” was a GDS-15 screening category, not a structured DSM-5 diagnosis; the authors note that GDS-15 can overestimate depression in cognitively impaired populations.
  • The complete-case primary analysis dropped participants with missing neuropsychological or GDS-15 data, and the MRI analysis was reduced to 81 CAA participants after missing-data and processing exclusions.
  • CAA and control groups differed substantially in age, sex, education, and hypertension. Adjustment covered age, sex, and education but cannot guarantee removal of residual confounding.
  • Controls were volunteers recruited by advertising and may underrepresent people with current or severe depression. Notably, among participants with available history data, prior depression was reported less often in CAA than controls (16.5% vs 41.1%), despite higher current GDS-15 burden in CAA.
  • Neuropsychological tests changed after 2016, and the study did not fully measure fatigue, anxiety, apathy, or medications other than antidepressants. The cohort was recruited through specialized clinics, limiting generalizability.

Funding

The study was funded by the Canadian Institutes of Health Research, Brain Canada, the Heart and Stroke Foundation of Canada, and the Alzheimer Society of Canada. Additional salary or trainee support came from the University of Calgary, the University of Alberta Henri Toupin Chair, the Dr. Sandra Black Centre for Brain Resilience and Recovery, the Hotchkiss Brain Institute, the Vascular Training Platform, the Killam Trust, and the Alzheimer Society of Canada/Canadian Institutes of Health Research. The article does not report a clinical-trial registration; ethics approval was obtained from the University of Calgary and University of Alberta research ethics boards.

The paper

  • Authors. Nukala et al.
  • Title. Associations Between Cerebral Amyloid Angiopathy, Cognitive Impairment, and Depressive Symptoms.
  • Journal. Neurology.
  • Year. 2026.
  • DOI. 10.1212/wnl.0000000000218354
  • PMCID. PMC13446197
Deep Dive — click to expand

What this is

This is a clinic-based, cross-sectional study asking whether depressive symptoms are part of the cognitive phenotype of probable CAA, rather than merely a consequence of a CAA-related ICH. The headline association is large—adjusted odds of GDS-15-defined “possible depression” were 15.71-fold higher in CAA—but the mechanistic claim is deliberately smaller: depressive symptoms statistically accounted for 11% of the CAA–episodic-memory association, 9% of the executive-function association, and 2% of the processing-speed association. It is therefore a strong argument to look for depression in CAA, not proof that depression is the main cause of CAA-related cognitive impairment or that antidepressant treatment will restore cognition.

1. Shadow Audit

The striking OR 15.71 can conceal the denominator and the instrument. Only 20 of 85 CAA participants and 3 of 83 controls crossed the GDS-15 threshold; the wide 95% CI (4.26–80.05) reflects a small number of events. “Possible depression” is not major depressive disorder, and the paper did not use a structured diagnostic interview. The internal pattern is also more complicated than the headline: among participants with available history data, prior depression was reported in 16.5% of CAA participants versus 41.1% of controls, while current symptom scores were higher in CAA. That could reflect disease-related neuropsychiatric symptoms, selective control recruitment, treatment effects, or reporting differences. Finally, 89%–97% of the estimated cognitive effect remained direct to CAA in the binary mediation model. The paper supports a clinically relevant comorbidity, but it does not make depression the hidden master variable of CAA cognition.

2. Inversion Engine

To invert the practical conclusion from “screen for depression” to “treating depression reverses CAA cognitive impairment,” the observed mediation would need to be causal, stable over time, and modifiable. A simple quantitative stress test is that depression would need to explain substantially more than the observed 11% of episodic-memory effect, 9% of executive-function effect, and 2% of processing-speed effect—at least a majority of the relevant deficit for it to become the dominant treatment target. The continuous-score sensitivity analysis did not rescue that interpretation: the indirect effect was significant for episodic memory at about 12%, but not for executive function (6%, p = 0.11) or processing speed (7%, p = 0.10). A longitudinal intervention showing a clinically meaningful cognitive improvement after depressive symptoms fall, with no excess hemorrhage risk, is the missing inversion test.

3. Second-Order Catalyst

The first practice change should be low-cost and operational: CAA, lobar-ICH, and transient-focal-neurologic-episode clinics should add GDS-15 screening to the baseline cognitive and vascular assessment, document antidepressant exposure, and repeat the screen longitudinally. Patients with GDS-15 ≥5 should receive a clinical assessment rather than an automatic depression diagnosis, with attention to apathy, fatigue, anxiety, medication effects, and suicidality, and referral to primary care or geriatric psychiatry when indicated. The imaging signal suggests prioritizing follow-up for patients with cSS, reduced cortical thickness, or high CAA small-vessel-disease burden, but it is not yet a validated risk score. This change could occur at the next clinic template update; antidepressant prescribing should wait for individualized diagnosis and safety review.

4. Asymmetric Leverage

The leverage is not the large OR; it is the inexpensive screen applied across a growing CAA and lobar-ICH denominator. Twenty CAA participants crossed the symptom threshold versus three controls, and depression is potentially actionable even if it explains only a modest portion of cognitive impairment. A small improvement in mood, participation, sleep, and rehabilitation engagement could matter to patients without needing to alter the underlying amyloid vasculopathy. Conversely, the MRI associations are larger in selected subgroups—cSS had a count ratio of 2.04—but the denominator is only 81 CAA participants and the design cannot establish a treatment target. The asymmetric move is therefore broad screening with a narrow claim: identify and treat clinically significant symptoms for wellbeing, while measuring cognition rather than promising a disease-modifying cognitive effect.

5. Paradigm Destroyer

This paper kills the reflex that CAA is only a hemorrhage-risk diagnosis and that cognitive complaints in CAA can be interpreted without asking about mood. It also kills the opposite reflex that a large depression association identifies a reversible cause: most of the cognitive association remained statistically direct to CAA, and processing speed showed essentially no binary-depression mediation. Tomorrow morning’s protocol update could read: “At every CAA or lobar-ICH visit, record GDS-15 and review cSS, cortical thickness, and small-vessel-disease burden. Treat a positive screen as a prompt for diagnostic assessment and supportive care—not as proof that depression caused the cognitive deficit.”

MVP — Minimum Viable Proof

The minimum practice-changing proof would be a prospective, diagnostically confirmed CAA cohort or pragmatic trial showing that a defined depression intervention in patients with GDS-15 ≥5 produces a reproducible improvement in episodic-memory or executive-function z scores—ideally at least 0.2 SD beyond test–retest and practice effects—without increasing ICH or other clinically important harm. This study provides the screening signal and an effect-size hypothesis; it does not provide that intervention evidence.

Best Combination

Combine this study with the established literature showing that CAA-related hemorrhage, white-matter injury, cortical neurodegeneration, Alzheimer pathology, and small-vessel-disease markers can independently impair cognition, plus longitudinal studies of depression after ICH and randomized evidence for depression treatment in older adults. The coherent model is additive and potentially bidirectional: CAA-related brain injury may produce cognitive and affective symptoms, while depression may further reduce cognitive performance and recovery. The bedside synthesis is to screen and treat depression on its own clinical merits while continuing disease-specific hemorrhage prevention, cognitive evaluation, and longitudinal surveillance.

Overvalue Warning

  • Do not translate OR 15.71 into a 15-fold individual probability of major depression; it is an adjusted association with a GDS-15 screening threshold in a small, selected cohort, with a wide confidence interval.
  • Do not interpret 11%–12% mediation as proof that treating depression will improve memory, or interpret cSS and cortical-thickness associations as validated predictive biomarkers. The study is cross-sectional, and the direct CAA component remained dominant for all three cognitive domains.

Educational use notice. The Clotbust Review is dedicated to medical education. Figures from published articles are embedded for teaching purposes, with attribution to the original publication. We do not claim copyright over any embedded figure; all rights remain with the original publisher and authors under the article's published license. If you are a rights holder and would like a figure removed, please contact the editorial team. See our editorial notice for the full policy.