PDF: Atrial Pouch
(2025)Objective
To determine whether left atrial (LA) diverticula detected on acute cardiac CT in patients with acute ischemic stroke (AIS) are associated with increased risk of recurrent ischemic stroke and differences in functional outcome over 2 years.
Study Summary
• Recurrent ischemic stroke occurred in 15% with diverticula vs 8% without.
• Larger diverticulum volume increased recurrence risk
Intervention
Prospective, single-center observational cohort (post hoc substudy of Mind the Heart) comparing outcomes between AIS patients with vs without LA diverticula detected on ECG-gated cardiac CT performed during acute stroke imaging.
Inclusion Criteria
Adults with AIS who underwent ECG-gated cardiac CT as part of acute stroke imaging; informed consent obtained.
Study Design
Arms: LA diverticulum present; No LA diverticulum
Patients per Arm: LA diverticulum: 126; No LA diverticulum: 321; Total: 447
Outcome
• Diverticulum volume and recurrence: adjusted HR 1.02 per 10 mm³ (P=0.001)
• Functional outcome (mRS) at 2 years: median 2 (IQR 1–3) vs 3 (IQR 2–4), adjusted common OR 0.62 (95% CI 0.42–0.92; P=0.02)
Bottom Line
LA diverticula were found in 28% of AIS patients on cardiac CT and were associated with a 2-fold increase in recurrent ischemic stroke at 2 years (15% vs 8%; adjusted HR 2.01; 95% CI 1.08-3.77; P=0.03). Risk increased with larger diverticulum volume (adjusted HR 1.02 per 10 mm³; P=0.001). Paradoxically, patients with diverticula had better functional outcomes (median mRS 2 vs 3), likely from residual confounding.
Major Points
- LA diverticula prevalence: 126/447 (28%) of AIS patients on cardiac CT.
- Doubled recurrent stroke risk: 15% vs 8% at 2 years (adjusted HR 2.01; 95% CI 1.08-3.77; P=0.03).
- Volume-dependent risk: each 10 mm³ increase → adjusted HR 1.02 (95% CI 1.01-1.03; P=0.001). Median volume 169 mm³ in recurrence vs 101 mm³ without.
- Paradoxical better functional outcome: median mRS 2 vs 3 (adjusted cOR 0.62; 95% CI 0.42-0.92; P=0.02) — likely residual confounding.
- No thrombi detected within diverticula (0/126), though microthrombi or prior embolization plausible.
- New AF not different: 20% with diverticula vs 13% without (P=0.10).
- Morphology: 83% right superior location, 82% cystic, 83% smooth surface. Median length 6mm, width 5mm, volume 113 mm³.
- Anticoagulation imbalance: 10% vs 21% at baseline (P=0.007) — key confounder adjusted for.
- Recurrent strokes largely cryptogenic/cardioembolic: 44% cardioembolic + 44% unknown in diverticula group.
- Single-center observational (Amsterdam UMC); hypothesis-generating only.
Study Design
- Study Type
- Post hoc analysis of prospective observational cohort (Mind the Heart)
- Randomization
- No
- Blinding
- None (observational)
- Sample Size
- 447
- Follow-up
- 2 years
- Centers
- 1
- Countries
- Netherlands
Primary Outcome
Definition: Recurrent ischemic stroke at 2 years
| Control | Intervention | HR/OR | P-value |
|---|---|---|---|
| 24/314 (8%) | 18/124 (15%) | 2.01 (1.08-3.77) | 0.03 |
Limitations & Criticisms
- No universally accepted definition of LA diverticula — pragmatic definition (length/ostium >0.5, length ≥3mm) may not be reproducible.
- Single-center, small sample, low events — hypothesis-generating only.
- Volume underestimation: cylinder formula used (median 113 mm³) vs voxel-wise segmentation (mean 415 mm³ in prior studies).
- Residual confounding for functional outcome — diverticula group had less anticoagulation, hypertension, malignancy.
- No thrombi detected within diverticula — mechanism of embolism unproven.
- Observational design — no causal inference possible.
- 2-year follow-up only.
Citation
Stroke. 2025;56:00-00.