Reproducibility of laser Doppler flowmetry in the human optic nerve head

We sought to demonstrate reproducibility of the noninvasive laser Doppler flowmetry technique for measuring microvascular hemodynamics of the optic nerve head in human subjects.

Laser Doppler flowmetry was performed by four examiners on four human optic nerve heads during multiple sessions. Intersubject, interexaminer, intersession, and intrasession variabilities were calculated for velocity, blood volume, and blood flow (flux).

Initial variability of measurements improved after 3 months' experience. The standard deviation of the intrasession variation for single readings was 18% of the velocity mean value and 24% of the flow (flux) mean value. Based on this, a sample size of 17 (power beta = 0.8, alpha = 0.05, paired t test) will demonstrate a 10% physiologic change in flux within a session if the average of five readings is used before and after the experimental perturbation. The intersession variation was 12% and 32% for velocity and flux, respectively. Because of this, a larger sample size or additional measurement sessions are required for detection of an experimental influence between two sessions. The coefficient of variation among subjects for single readings was 7% and 10% for velocity and flux, respectively.

Reproducibility of velocity measurements is better than reproducibility of flux measurements. After experience with the method, laser Doppler flowmetry is sufficiently precise to permit feasible studies of human optic nerve head microvascular hemodynamics when physiologic perturbations are applied.