Jump to comment:

• Baseline artery diameter confounds flow-mediated dilation only if a percentage change is selected as the size-scaling index
  Greg Atkinson
  Published on: 23 November 2015

• Published on: 23 November 2015

  Baseline artery diameter confounds flow-mediated dilation only if a percentage change is selected as the size-scaling index

  • Greg Atkinson, Professor of Health Sciences and Biostatistics

  Dear Editor, Maruhashi et al.[1] meticulously measured the flow-mediated changes in brachial artery diameter with a large sample of participants. The changes in diameter were quantified using the conventional percentage-based index (FMD%). As usual, baseline artery diameter (Dbase) was found to be substantially and negatively correlated with FMD%. So it seems that even the most robust protocols and precise measurements o...

  Show More

  Dear Editor, Maruhashi et al.[1] meticulously measured the flow-mediated changes in brachial artery diameter with a large sample of participants. The changes in diameter were quantified using the conventional percentage-based index (FMD%). As usual, baseline artery diameter (Dbase) was found to be substantially and negatively correlated with FMD%. So it seems that even the most robust protocols and precise measurements of arterial diameters cannot eradicate the FMD%-Dbase correlation. This is not surprising because it is the FMD% index itself that causes this Dbase-dependency [2].

  It seems illogical to persist in quantifying endothelial "function" with FMD% when it is so erroneously dependent on the structural/morphological variable that is its denominator. The FMD% index was presumably selected to "normalise" the flow-mediated response for variability in Dbase. But the flow-mediated response itself (measured in mm) tends to be uncorrelated to Dbase, or is even inversely proportional to Dbase[2]. The overriding problem is that FMD% "works" only if this flow -mediated response is consistently, substantially and positively proportional to Dbase. This incongruity between the nature of the physiological change and the index used to describe that change means that FMD% itself creates the substantial negative dependency of FMD% on Dbase [2].

  Physiological explanations have been forwarded for the FMD%-Dbase correlation [3] but such explanations are putting the "cart before the horse". Because the output of the protocol, the FMD% index, is the source of the Dbase-dependency problem rather than the physiological changes that are occurring during the protocol, any further physiological explanations for FMD%-Dbase dependency are unsatisfactory[2]. For example, the shear rate explanation for Dbase-dependency forwarded by Pyke et al.[3] and repeated by Maruhashi et al.[1] is inconsistent with the fact that FMD% depends on Dbase even when calculated from randomly-generated data with no physiological basis at all [2]. If the flow-mediated response was scaled to Dbase properly in the first place, there would be no substantial negative correlation between these two variables, and so no need to explain and correct with other variables. Such circular logic is avoided by appropriate size scaling in the first place. Then the influence of other important variables, e.g., shear rate, on the flow-mediated response per se can be quantified more precisely.

  I agree that Dbase should be taken into consideration as a confounder of FMD%. A full consideration of this problem has surfaced recently and, as in the study by Maruhashi et al.[1], been applied to age- and sex- differences in the flow-mediated response [2]. The FMD% index is higher in women than in men simply because FMD% does not scale properly for the lower Dbase in women[2]. Similarly, Dbase increases with age, thus biasing age-related changes when FMD% is selected as the outcome of interest [2]. As long as FMD% is the index used to quantify the relative flow-mediated response, such inferential errors will occur, and these errors cannot be corrected parsimoniously with any physiological mechanisms. If FMD% was replaced with a more accurate scaling index, Dbase-dependency, and the associated interpretive problems would be eradicated.

  References

  1. Maruhashi T, Soga J, Fujimura N, et al. Relationship between flow- mediated vasodilation and cardiovascular risk factors in a large community -based study. Heart 2013;991837-1842.

  2. Atkinson G, Batterham AM. Allometric scaling of diameter change in the original flow-mediated dilation protocol. Atherosclerosis 2013;226:425 -427

  3. Pyke KE, Dwyer EM, Tschakovsky ME. (2004) Impact of controlling shear rate on flow-mediated dilation responses in the brachial artery of humans. J Appl Physiol 97: 499-508

  Conflict of Interest:

  None declared

  Show Less

  Conflict of Interest:
  None declared.

  • Back to top