To our knowledge, this is the first study to directly investigate the effect of time taken to inflate the venous occlusion cuff on FBF values calculated using strain-gauge VOP. In view of evidence that FBF should be calculated over the 1st cardiac cycle following cuff inflation, because FBF progressively decreases from the 2nd cardiac cycle onwards (Tschakovsky et al. 1995; Wood and Stewart 2010), our primary objective was to compare FBF values calculated from the 1st cardiac cycle following slower and rapid inflation, using manual and automated inflation, respectively, to achieve this. With slower inflation, peak FBF following isometric contraction at 70% MVC was considerably less than that calculated with rapid inflation. Furthermore, whereas FBF calculated following rapid inflation remained significantly raised until the 7th min following contraction, FBF calculated following slower inflation was greater than baseline only until the 4th min. Thus, slower inflation greatly underestimated both the peak and time course of post-contraction hyperaemia relative to rapid inflation.
The fact that it took almost four times longer to achieve the desired occluding pressure of 50 mmHg with slower manual inflation than rapid automated inflation (0.23 vs. 0.92 s), inevitably introduced a time delay before FBF could be calculated. The delay itself could not explain the discrepancy between the values obtained. For example, the difference between FBF measured at peak hyperaemia and 15 s later with rapid inflation was only ~ 12% of the peak value; far less than the 40% discrepancy between peak values calculated following slower manual and rapid automated inflation. The explanation is much more likely to lie in the additional venous filling allowed during inflation by the slower rate of inflation (see Brown et al. 1966), such that the perfusion pressure gradient was already reduced when the occluding pressure of 50 mmHg was reached.
The comparisons we made between FBF values calculated after slower manual inflation and those made over successive cardiac cycles following rapid inflation agree with that interpretation. For example, following rapid inflation, peak FBF following 70% MVC fell progressively when calculated over the 2nd and 3rd, rather than the 1st cardiac cycle, in agreement with Wood and Stewart (2010) who used the same automated inflation system as us and estimated peak FBFs at 15–60% MVC. Furthermore, following rapid inflation, FBF values calculated over the 1st, 2nd, and 3rd cardiac cycles from peak hyperaemia until return to baseline each correlated well with those calculated over the 1st cardiac cycle following slower inflation. However, the regression line was closest to unity for FBF values calculated over the 3rd cardiac cycle with rapid automated inflation and the 1st cardiac cycle with slower manual inflation. These findings can be reconciled, because slower manual inflation took almost 1 cardiac cycle at rest, but ~ 2 cardiac cycles when HR and FBF were raised following exercise, whereas rapid automated inflation took only a small portion of 1 cardiac cycle under both conditions. Thus, the 1st complete cardiac cycle over which it was possible to calculate FBF following slower inflation corresponded with the 2nd or 3rd cardiac cycle from the end of rapid inflation. Indeed, our results allow us to argue that cuff inflation time must be ≤ 0.3 s to ensure that high FBFs can be accurately calculated from the 1st cardiac cycle, even when HR increases to 180–200 beats min− 1.
The present findings add to existing evidence that any time delay in the period over which FBF is calculated when using VOP leads to underestimation, particularly at high flows. Such delays include calculating FBF over the 2nd, 3rd, or 4th cardiac cycles rather than the 1st (Tschakovsky et al. 1995; Wood and Stewart 2010), using the cumulative slope over several cardiac cycles (Wood and Stewart 2010) and, as we now show, using cuff inflation that takes longer than one cardiac cycle. Although we compared manual and automated inflation to achieve two different rates of inflation, it is reasonable to assume our results reflect the duration, rather than the method of inflation. Furthermore, there is no obvious reason why these same principles should not apply when using VOP for measurement of CBF. Certainly, when the venous occlusion cuff was kept inflated during rhythmic calf muscle contractions at 30–70% MVC, so avoiding cuff inflation time, CBF measured over the 1st cardiac cycle during the relaxation phases, compared very favourably with values estimated by Doppler ultrasound (Green et al. 2011; Murphy et al. 2018).
These points are highly relevant, because even in some studies published since the key study of Tschakovsky et al. (1995), a 1–2 s delay was allowed after completing venous occlusion and before limb flow was calculated, and/or flow was calculated over 4–6 s (e.g., Groothuis et al. 2003; Thijssen et al. 2005; Kooijman et al. 2007). Indeed, Wythe et al. (2015) recently recommended in a methodological study on VOP that a delay as long as 4 s should be allowed before calculating FBF over a period of 2 s, i.e., at least 2 cardiac cycles. Furthermore, their statement that the same rapid inflation system as we used, with a manufacturer’s quoted inflation time of 0.3 s, took 2.5 s to inflate the occluding cuff to 50 mmHg, suggests actual inflation time may reflect factors other than the inflation system per se. As we indicated in the Introduction, periods of up to 0.75 s and < 0.3 s have been quoted for commercially available, rapid inflation systems (Tschakovsky et al. 1995; Wood and Stewart 2010), while some studies used manual inflation, and others provided no information at all on the method of cuff inflation.
In view of this variability in the application of VOP, it seems that changes in absolute FBF must have been underestimated in many published studies, especially during vasodilator stimuli such as exercise and reactive hyperaemia, mental stress, and heating. This does not devalue the physiological significance of these findings, or the contribution VOP has made to understanding of peripheral vascular regulation (see Joyner et al. 2001), and it simply questions the absolute values reported. Moreover, any underestimation of high limb flows in pharmacological studies (see Wilkinson and Webb 2001) must have made the effects of antagonists on vasodilator responses more difficult to differentiate. Looking back, it is paradoxical that in early studies with VOP, and it was recognised that the inflation system should be optimised using a cuff of small distended volume, large pressure reservoir, and short, wide-bore connexions so as to achieve venous occlusion pressure as rapidly as possible. Furthermore, it was accepted that recordings in which apparent rate of inflow falls from beat to beat, i.e., the slope from which flow is calculated, are difficult to interpret and should be analysed by joining corresponding points on pulse waves that “lie in a straight line” (Greenfield et al. 1963; Roddie and Wallace 1979). We still seem to be re-learning and exploring these issues.
Looking forward, we argue that in any new study involving VOP, information should be provided on cuff inflation time, as well as on the timing and duration of the slope used to calculate FBF, or CBF. Such information has wider application for venous occlusion is often combined with near-infrared spectroscopy (NIRS) to allow calculation of limb muscle blood flow, or oxygen consumption (VO2) from the rate of increase in total haemoglobin (THb), or deoxygenated haemoglobin (HHb) following venous occlusion (Homma et al. 1996; Van Beekvelt et al. 2001; Malagoni et al. 2010). Cuff inflation times of up to 4 s and slope durations of 3–80 s were used for calculation in these studies. However, recently, when Cross and Sabapathy (2017) used automated rapid cuff inflation they found that muscle blood flow calculated from rate of increase in THb decreased progressively from the 1st cardiac cycle onwards, the decrement being greater at high flows, just like FBF calculated from the rate of increase in limb volume when using VOP. Thus, it is reasonable to deduce cuff inflation time is just as important when calculating muscle blood flow or VO2 with NIRS, as we show for VOP.
Clearly, it was impossible to simultaneously inflate the occluding cuff in two different ways in the same arm. Thus, we inevitably made comparisons between FBF values calculated with slower and rapid inflation for two different periods of isometric contraction in each subject. Although this must have contributed to the variability, it is unlikely it contributed to the marked disparity between FBF values calculated with the two methods given we randomised the order in which they were used. Our results relate to FBF values calculated when the venous occlusion cuff pressure was set at 50 mmHg; this value was chosen, because it is the widely used in studies involving venous occlusion plethysmography (Lorentsen et al. 1970; Tschakovsky et al. 1995; Wood and Stewart 2010). Had we used a lower pressure, for example, 30 or 40 mmHg, the time taken by manual inflation would have been shorter and the disparity between FBF measurements made over the 1st cardiac cycle with the two techniques would probably have been smaller. However, the principle that inflation time affects FBF values calculated at high flows would still hold. Furthermore, we only calculated FBF values before and following isometric contraction at 70% MVC, not during other stimuli. However, it is unlikely the mechanisms that cause vasodilatation affect the mechanical factors that determine the FBF values calculated with cuff inflation at different rates.
In conclusion, we have shown that when using VOP, a venous cuff inflation time of ~ 0.2, or 0.9 s, within the range used in studies over the last 70–80 years, has considerable impact on calculated absolute FBF, especially when FBF is high. It had already been established that absolute FBF decreases from the 1st cardiac cycle onwards following venous occlusion; therefore, the 1st cycle gives the most accurate assessment of FBF (Tschakovsky et al. 1995; Wood and Stewart 2010). We now show that an inflation time of just 0.9 s removes the benefit of calculating FBF over the 1st cardiac cycle, leading to FBF values equivalent to those calculated over the 3rd cardiac cycle with an inflation time of ~ 0.2 s. Taken together, these results indicate that cuff inflation time as well as the period over which FBF is calculated are of crucial importance in studies involving VOP. Without this information, the results cannot be fully evaluated.