Specific methods

Eight healthy volunteers were recruited. Brachial artery dilatation was assessed in response to reactive hyperaemic blood flow induced by cuff occlusion times of 2, 3, 5 and 8 minutes. Each scan was performed by the same operator but in a random order using an edge detection algorithm to measure the complete time course of FMD (section 2.2.4.3).

For each scan maximal FMD (FMDmax), absolute FMD (FMDabs)the time to maximal dilatation (t^ax), the area under the curve of dilatation (F M D a u c ) and the slope of the dilatation/time curve (Ddil/Dt) were calculated. Blood flow velocity was measured at baseline at 5 and 10 seconds after release of the occluding cuff and thereafter at 15 second intervals to a maximum of 90 seconds. The peak blood flow velocity VTI^ax and the ratio of blood flow for each time point was calculated with reference to baseline blood flow, the AUC of absolute blood flow velocity (V T I a u c ) and o f the ratio of blood flow velocity was calculated for each scan. The slope o f the degradation of blood flow velocity was calculated (DVTI/Dt). Pearson bivariate correlation coefficients were used to assess which flow variables were most closely associated with parameters of FMD.

2.3.3 Results

Complete data was available for all scans on all subjects. The reactive hyperaemia flow envelopes are shown in Figure 2.3.1. Reactive hyperaemic blood flow was maximal at 5 seconds after the cuff release 2 or 3 minutes of distal forearm ischemia nad at 10 seconds after 5 or 8 minutes. Increasing duration of forearm ischemia, above 3 minutes, did not significantly increase peak reactive hyperaemic blood flow (Figure 2.3.2). However, there was a dose dependent increase in the duration of hyperaemic blood flow such that blood flow velocity was increased up to 75 seconds after cuff release and the AUC of the flow envelope was significantly increased (P = 0.01) and the slope of flow degradation significantly decreased (P = 0.001) (Figure 2.3.2).

Similarly increased occlusion time was associated with a stepwise increase in FMDmax (P < 0.001) (Figure 2.3.3), Ddil/dt (P < 0.001) and F M D a u c (P < 0.001) but had little influence on the tmax (P = 0.3) or the FMDabs (P = 0.4).

The correlation coefficients between flow variables and parameters of dilatation are shown in Table 2.3.1. FMDmax was significantly correlated with VTI at 10 seconds after cuff release (r = 0.4, P = 0.02), DVTl/Dt (r = -0.4, P = 0.04), V T I a u c (r = 0.3, P = 0.03). Although FMDmax was not associated w ith baseline blood flow or VTImax, there was a significant association with the percentage peak reactive hyperaemia (r = 0.4, P = 0.03). Similar associations were seen between F M D a u c and reactive hyperaemia. (Table 2.3.1) but there was no assocaition between tmax- and any flow variable. Correction of VTI for subjects heart rate did not significantly improve the relationship between the assessment of blood flow and FMD.

E F

>

Cuff occlusion time (minutes)

75 90

15 30 45 60

0

Tim e (seconds)

Figure 2.3.1 Flow envelope of reactive hyperaemia during different lengths of forearm ischaemia.

0 .5 - ^ 0 .4 -

>

o

<

Cuff occlusion tim e (minutes)

1000 800 600 ^ 400 200 0

Cuff occlusion tim e (minutes)

Figure 2.3.2 Effect of different lengths of forearm ischaemia on peak reactive hyperaemia (upper panel) and the area-under the time/VTI curve (AUC) (lower panel). Beyond two minutes there was little increase in peak reactive hyperaemia but a significant increase in the AUC of time/VTI curve

c o s CD

10 _

Cuff occlusion tim e (minutes)

Figure 2.3.3 Effect of different lengths of forearm ischaemia on maximum flow- mediated dilatation.

Table 2.3.1 Pearson correlation coefficients for reactive hyperaemic and dilatation parameters in 8 healthy subjects.

xFMDabs _xFMDmax FMDauc Ddil/Dt tmax

FMDmax 0.97 t FMDauc 0.95 Î 0.97 Î Ddil/Dt & 8 8 I 0.89 { 0 .8 4 } tmax 0.11 0.14 0.17 0.25 VTIbl -0.18 -0.15 -0.25 -0.02 0.03 VTImax 0.25 0.26 0.16 0.30 0.24 VTIIO 0.41 0.48 t 0 .4 0 * 0.43 * 0.21 VTIauc 0 .4 2 * 0 .3 9 * 0.32 0 .4 2 * 0.32 DVTI/Dt -0.46 t -0.44 * -0.41 * -0.32 -0.07 RH% _{0.49 t} 0 .4 9 * & 5 2 } 0.30 0.06 * P < 0.05; t P < 0.01; %?< 0.001

Abbreviations: xFMDabs - absolute FMD; FMDmax = maximal flow mediated dilatation; F M Dauc = area under the curve for dilatation; Ddil/Dt = gradient of the

time/dilatation curve; tmax = time to maximum dilatation after release of the cuff; VTImax = maximal blood flow velocity after release of the cuff; VTI 10 = blood flow velocity at 10 seconds after release of the cuff; V T Iauc = area under the curve

of the blood flow envelope; DVTI/Dt = gradient of flow degradation during reactive hyperaemia; RH% = peak percentage increase in reactive hyperaemic blood flow from baseline.

.3.4 Discussion

These data demonstrate a significant dose dependent relationship between parameters of reactive hyperaemic blood flow and subsequent brachial artery dilatation. Increased blood flow was associated with enhanced FMDmax, ^ increased Ddil/Dt gradient and the AUC of FMD. These parameters were closely related and further studies will be required to determine which most sensitively differentiates between risk factor groups. Increasing duration of forearm ischaemia had relatively little effect on the peak intensity of blood flow following release of the cuff and

consequently VTImax- did not significantly predict FMD. In contrast, the duration of hyperaemic blood flow and the AUC were significantly enhanced. This is consistent with the poor correlation between peak blood flow and FMD that was seen and has previously been reported (Leeson et al, 1997) and suggests that endothelial activation is dependent, at least in part, on the integral of blood flow over time rather than detecting rapid changes in shear stress as occur with reactive hyperaemia. From a physiological point of view this might be appropriate as high gradients in shear stress which occur with the transmission of a pulse wave would not be sufficient to stimulate the endothelium to synthesise vasodilators. The mechanisms whereby the endothelium integrates the flow stimulus are unknown however. Although there were no significant correlations between any parameter of FMD and either baseline blood flow or VTImax there was a significant association between FM Dmax and F M Dauc

and the percentage increase in blood flow (i.e. the ratio of VTImax and baseline VTI). This suggests that the relative increase in blood flow might also be an important stimulus to the endothelium and it is possible that a number of different mechanisms are operating.

These findings have practical implications for the study of endothelial function. Firstly, they indicate that differences in FMD between individuals and risk factor groups might be mediated by differences in the blood flow stimulus to the endothelium. Factors such as the size and muscularity of the arm might be important in determining such responses. In the experimental setting, therefore, it will be necessary to adjust FMD for any inter-individual or inter-cohort differences in the reactive hyperaemic stimulus in order to correctly assess differences in endothelial function. Secondly, these data indicate that peak hyperaemic blood flow itself is a poor determinant o f the subsequent dilatation and adequate assessment of the flow stimulus to the endothelium will require description of the hyperaemic flow envelope over a longer period of time.