Reduced Blood flow in nonexudative AMD
Posted inInteresting Papers

Reduced Blood flow in nonexudative AMD

General Overview

  • Study Purpose: To assess whether laser speckle flowgraphy (LSFG)-derived indices of ocular blood flow (choroidal and inner retinal) are reduced in nonexudative AMD compared to age-matched controls and to explore associated factors.
  • Design: Retrospective case-control study conducted at the University of Iowa and Iowa City Veterans Health Administration (2017–2023).
  • Cohort: 39 eyes from 24 AMD patients (early, intermediate, advanced nonexudative AMD) and 41 eyes from 21 healthy controls.
  • Key Relevance: Reduced ocular blood flow may contribute to AMD pathogenesis, and LSFG offers a novel, non-invasive method to quantify flow dynamics, potentially aiding in diagnosis, risk stratification, and therapeutic monitoring.

Study Design and Methods

  • Population:
    • AMD Group: 24 patients (39 eyes; 5 early, 14 intermediate, 5 advanced with geographic atrophy [GA]). Excluded exudative AMD, diabetic retinopathy, ischemic optic neuropathy, glaucoma, or other conditions affecting blood flow.
    • Control Group: 21 age- and sex-matched healthy subjects (41 eyes).
    • Demographics: Predominantly male (90–92%), Caucasian (100%), mean age 73.1 (controls) vs. 76.7 (AMD) years (P=.152). High prevalence of vascular risk factors (smoking, hypertension, hyperlipidemia, diabetes) in both groups, with no significant differences.
  • LSFG Technique:
    • Device: LSFG-NAVI (Softcare Co., Ltd.) uses an 830 nm diode laser to measure relative erythrocyte velocity via speckle contrast, reported as mean blur rate (MBR) in arbitrary units (AU).
    • Measurements:
      • Choroidal blood flow (MT): MBR in areas without major retinal vessels, dominated by choriocapillaris and choroid.
      • Inner retinal blood flow (MV-MT): MBR in retinal vessels minus underlying tissue flow, reflecting central retinal artery perfusion.
    • Segmentation:
      • Binary threshold masking (LSFG Analyzer software) to separate vascular (MV) and tissue (MT) segments.
      • Superpixel segmentation (~1500 superpixels/image) to map areas of very low flow (MBR 20 AU).
    • Validation: Inter-rater reliability for manual thresholding was excellent (mean difference 0.043 AU, P=.61). Total Retinal Artery and Vein Analysis (TRAVA) confirmed inner retinal flow measurements.
  • Additional Metrics:
    • Ocular perfusion pressure (OPP): Calculated as OPP = 2/3(mean arterial pressure) – IOP, using automated blood pressure and tonometry.
    • Choroidal thickness: Measured via OCT when available.
    • AMD staging: Classified as early (medium drusen), intermediate (large drusen/pigment changes), or advanced (GA) per Ferris et al. (2013).
  • Statistical Analysis:
    • Mixed-effects linear regression to account for inter-eye correlation.
    • Multivariable logistic regression to assess AMD risk factors and blood flow associations.
    • Two-way ANOVA with Tukey’s post-hoc for AMD stage comparisons.

Key Findings

  • Choroidal Blood Flow (MT):
    • Reduced by 33% in AMD vs controls (5.3 ± 0.3 AU vs 7.9 ± 0.5 AU, P=.00005).
    • Significant reduction across all AMD stages (early: β=-2.4, P=.005; intermediate: β=-2.3, P=.0001; advanced: β=-3.87, P=.000002).
    • Odds ratio (OR) for AMD: 2.27 (95% CI: 1.33–6.47, P=.027) per 10% reduction in MT from normal (8.0 AU). A 50% reduction yields OR=60.2.
  • Inner Retinal Blood Flow (MV-MT):
    • Reduced by 20% in AMD vs controls (12.5 ± 0.6 AU vs 15.6 ± 0.5 AU, P=.004).
    • Significant reduction across all stages (early: β=-2.8, P=.033; intermediate: β=-2.45, P=.005; advanced: β=-4.03, P=.0009).
    • Associated with AMD in univariate analysis but not included in final multivariable model to avoid multicollinearity.
  • Superpixel Analysis:
    • AMD eyes had a higher percentage of superpixels with very low flow (
    • Low-flow areas were diffuse, often exceeding regions of visible retinal pathology (e.g., GA or drusen).
  • Ocular Perfusion Pressure:
    • No significant difference between AMD (50 ± 5.5 mm Hg) and controls (53 ± 6.7 mm Hg, P=.17), indicating reduced blood flow is not due to systemic hypotension or elevated IOP.
  • Choroidal Thickness:
    • No significant correlation with choroidal blood flow in AMD or controls, suggesting flow reductions are independent of structural thinning.
  • Other Factors:
    • No significant associations between blood flow and demographics (age, sex), vascular risk factors (smoking, hypertension, diabetes, hyperlipidemia), BMI, or macular structural features (pigment changes, subretinal drusenoid deposits [SDDs], GA area).

Clinical Implications

  • Pathophysiology:
    • Reduced choroidal and inner retinal blood flow supports the hemodynamic model of AMD, where choriocapillaris hypoperfusion precedes retinal pigment epithelium (RPE) and photoreceptor loss.
    • Diffuse low-flow areas suggest widespread vascular dysfunction, even in early AMD, potentially contributing to disease progression.
  • Diagnostic Utility:
    • LSFG is a promising biomarker for detecting reduced ocular blood flow in nonexudative AMD, with potential to differentiate AMD from healthy eyes.
    • High ORs indicate that even modest flow reductions (10%) significantly increase AMD risk, useful for early identification.
  • Therapeutic Potential:
    • Interventions targeting ocular blood flow (e.g., astaxanthin, ophthalmic artery angioplasty) may slow AMD progression, warranting further study.
    • LSFG could monitor treatment efficacy in clinical trials.
  • Comparison to OCT-A:
    • LSFG advantages: Quantifies blood flow velocity on a continuous scale, sensitive to moderate reductions, unlike OCT-A, which primarily detects structural flow deficits.
    • LSFG limitations: Not depth-resolved, less precise for isolating choroidal vs. retinal contributions compared to OCT-A.

Limitations:

  • Selection Bias: Predominantly male, Caucasian veteran population limits generalizability (underrepresentation of females).
  • Cross-Sectional Design: Cannot establish causation between reduced blood flow and AMD progression.
  • Manual Thresholding: Potential bias in binary segmentation, though mitigated by masked graders and TRAVA validation.
  • Limited Field of View: Standard LSFG protocol excluded some temporal macula; future studies plan montage imaging.
  • Choroidal Flow Validation: MBR’s dominance by choroidal flow is supported but not fully validated.
  • Sample Size: Relatively small (24 AMD, 21 controls), though statistically significant results were achieved.

Pearls

  • LSFG Definition: Non-invasive imaging using 830 nm laser to measure relative erythrocyte velocity (MBR) in choroid (MT) and inner retina (MV-MT).
  • AMD Blood Flow:
    • Choroidal flow: Reduced 33% in AMD (5.3 vs 7.9 AU, P=.00005), OR=2.27 per 10% reduction.
    • Inner retinal flow: Reduced 20% (12.5 vs 15.6 AU, P=.004).
    • Significant at all AMD stages (early, intermediate, advanced).
  • No OPP Difference: Rules out systemic BP or IOP as causes of reduced flow; suggests intrinsic vascular resistance.
  • Superpixel Analysis: Quantifies low-flow areas (
  • Clinical Relevance:
    • Supports hemodynamic model of AMD.
    • LSFG may aid early diagnosis and monitor therapies targeting perfusion.
  • Management:
    • No proven treatment for nonexudative AMD; AREDS2 supplements for intermediate AMD; monitor for GA or neovascular conversion.
    • Future therapies may target blood flow (e.g., vasodilators, angioplasty).
  • Prognosis: Reduced flow may predict GA progression; longitudinal studies needed.

Citation

Linton EF, Ahmad N-US, Filister R, Wang J-K, Sohn EH, Kardon RH. Laser Speckle Flowgraphy Reveals Widespread Reductions in Ocular Blood Flow in Nonexudative Age-Related Macular Degeneration. Am J Ophthalmol. 2025;273:91–106. doi:10.1016/j.ajo.2025.01.014