Strategies for reducing variance in laser Doppler flowmetry measurements.

Background: Scattering of blood flow data as assessed with laser Doppler flowmetry (LDF) in humans is a problem in many studies using this technique. We set out to reduce variability in LDF data by eliminating the effect of the total returning light level (DC) on LDF parameters in the choroid through partial regression analysis.
Methods: In 20 healthy subjects, choroidal blood flow parameters were measured at different DC values using a portable confocal LDF device. We used two different strategies to reduce scattering of data eliminating the effect of yield, which is defined as DC/gain. On the one hand, we used a previously described method based on a third-order polynomial fit, which combines all obtained data. On the other hand, we applied a new method based on a linear fit for each individual subject.
Results: Variability of data during changes in DC is higher for LDF parameters volume and flow than for velocity. Both methods were successful in reducing scattering of LDF parameters with varying DC.
Conclusions: The present study indicates that both methods to correct for changes in yield were successful in reducing the variability of LDF measurements. When systematic changes in DC occur after an intervention, one needs to be careful in interpreting the obtained data and it remains to be shown if either of the two techniques is capable of correcting for this effect. The approach presented here may, however, represent an effective, easily applicable and valid approach to reduce scattering of data from using LDF to assess blood flow in the posterior pole of the human eye.