Variability of flow mediated dilation: consequences for clinical application

BACKGROUND: Flow mediated vasodilatation (FMD), a non-invasive tool to assess endothelial function, has been shown to have prognostic value for the development of cardiovascular disease. Conventional B-mode ultrasonography has been criticised for its 'limited' resolution in vivo, which complicates reliable detection of the minute diameter changes during reactive hyperaemia. In the present study we evaluated the physical resolution, reproducibility and the capability to detect FMD impairment of a wall tracking system (WTS). METHODS: The resolution of WTS was compared with that of intravascular ultrasound (IVUS) in pig femoral arteries in vivo. Subsequently, intra- and interobserver variability of FMD testing with WTS was evaluated in 75 healthy volunteers. Finally, the effect of smoking as single risk factor for atherosclerosis on FMD in vivo was assessed. RESULTS: WTS and IVUS readings were not different (difference in arterial cross sectional area 1.97 mm(2), r=0.87). Intrasession coefficient of variation in baseline diameter was 1.1% (extremes 0.06--2.0%). Inter-session baseline diameter variation was 3.6 and 3.8% for each observer and 4.1% between observers. Intra-individual variability in FMD between sessions was considerable with coefficients of variation of 13.9% for FMD and 9.3% for NTG. Smokers had impaired FMD responses compared with matched non-smokers (4.7+/-2.4 vs. 9.6+/-4.4%, P<0.001), whereas NTG induced vasodilatation did not differ (13.4+/-6.2 vs. 15.4+/-5.1%; p=ns). CONCLUSION: WTS is a suitable technique for reproducibly assessing the brachial artery diameter in vivo with a accuracy comparable to that of IVUS. Using this sensitive technique the reproducibility of FMD in vivo proves to be poor mainly due to physiological factors. Whereas this seriously limits the use of FMD as follow-up parameter for individual subjects, FMD is demonstrated to be a useful research tool at group level.