Reply from Jonathan E. Elliott, Joseph W. Duke, Jerold A. Hawn, John R. Halliwill and Andrew T. Lovering

In their letter (Frise & Dorrington, 2015), Drs Frise and Dorrington have provided a number of critiques on our recent study, entitled ‘Increased cardiac output, not pulmonary artery systolic pressure, increases intrapulmonary shunt in healthy humans breathing room air and 40% O2’ (Elliott et al. 2014). Specifically, they have concerns, ‘that the title and content of [our] manuscript raise two fundamental issues related to fluid mechanics.’ First, it was most appropriate to specify ‘pulmonary artery systolic pressure’ (PASP) in the title because Doppler ultrasound was used to determine pulmonary pressure and we did not want readers to think that we had measured anything but PASP. This technique measures the velocity of retrograde blood flow across the tricuspid valve during systole, and thus the peak blood pressure developed in the pulmonary artery during systole can be calculated using the modified Bernoulli equation. The potential that ‘systolic [pulmonary] pressure’ per se, is the period of the cardiac cycle responsible for increasing intrapulmonary arteriovenous anastomoses (IPAVAs) blood flow () is neither implied from the title, nor discussed in our manuscript. Second, Drs Frise and Dorrington state: ‘cardiac output cannot directly increase an intrapulmonary shunt, since the latter is simply a flow of blood down some tubes.’ In our opinion, this is an oversimplification that ignores the many physiological intricacies of the regulation of pulmonary blood flow. For example, pulmonary blood flow is dependent upon a myriad of factors, including, but not limited to, gravity, architecture of the pulmonary vascular tree, oxygen tension, and may even be determined by precapillary sphincters associated with supernumerary vessels that are still poorly understood. Clearly pressure at the beginning of the tube needs to be greater than the pressure at the end of the tube to facilitate orthograde blood flow. However, in one of our experimental conditions, that combined atropine (ATR) and intravenous (i.v.) adrenaline (ADR) at 80 ng kg−1 min−1 (ATR + 80 ADR), PASP was not statistically increased, and therefore we could not state that there was an increase in pulmonary pressure; had we done so, we would have been lying and likely to be challenged by the reviewers, and rightly so, for over-interpreting or misinterpreting our data. Additionally, when cardiac output () increases, as occurs during low intensity exercise, there is recruitment of under- and/or non-perfused pulmonary blood vessels, which may include IPAVAs. In this way, increases in may be able to recruit IPAVAs without significant increases in pulmonary pressure and subsequently result in a shunt, if blood flow through these pathways does not participate in gas exchange. Several other studies, referenced in our manuscript, demonstrate an increase in right-to-left shunt secondary to increases in . Third, Drs Frise and Dorrington state: ‘It would seem desirable to avoid the use of vasoactive drugs, which might directly alter IPAVA tone independent of an effect on cardiac output.’ Central to this concern is the assumption that ATR and ADR specifically affect the vascular tone of IPAVAs, and do not alter the tone of the rest of the pulmonary vasculature. For example, the observation of increased without significant increases in PASP could be explained independent of any change in the vascular tone of IPAVAs if either ATR caused pulmonary vasodilatation of the conventional pulmonary circulation (not IPAVAs), or the lower dose of ADR provided less of a vasoconstrictor influence on the rest of the conventional pulmonary circulation – possibilities that we detailed in our manuscript. Our data neither imply, nor do we claim, that there was a ‘constant drop in pressure across the tubes [IPAVAs]’ during ATR + 80 ADR. We also addressed this issue by providing references to work where other authors concluded that, ‘…the increase in shunt from pharmacologically increased is probably not due to vasoconstrictor or vasodilator activity of these drugs as the increase in shunt is observed when is increased via mechanical means as well.’ A final statement from Drs Frise and Dorrington was that the complexity of our study ‘…makes it difficult to draw any conclusions about mechanisms determining flow through IPAVA…’ and quoted us as follows: ‘considering the unique experimental conditions used in the present study, it is difficult to extrapolate these findings to their potential physiological relevance.’ The entire quote from our work was: ‘Considering the unique experimental conditions used in the present study, it is difficult to extrapolate these findings to their potential physiological relevance during other conditions known to increase blood flow through IPAVAs, such as exercise.’ Again, we acknowledged the complexities of the regulation of pulmonary blood flow while being careful not to misinterpret or over-interpret our data. We recognize the complexities of quantifying pressure, flow and in humans. As such, we specifically commented in our paper that: ‘Nevertheless, future work in this area will be helpful in clarifying the roles of pressure and flow under other experimental conditions such as hypoxia and exercise.’ This statement very clearly acknowledges that our findings might not be applicable to other conditions. In summary, our manuscript measured PASP and and provides evidence that: (1) is associated with an impairment in pulmonary gas exchange efficiency, i.e. an increase in the alveolar-to-arterial difference (A-aDO2), and calculated volume of venous admixture during unique and tightly controlled experimental conditions that reduce/eliminate the potential contribution of alveolar ventilation-to-perfusion inequality and diffusion limitation to the A-aDO2, and (2) primarily depends on increases in , as it occurred independent of significant increases in PASP. These data provide the impetus to investigate the relative importance of and mean pulmonary arterial pressure or pulmonary vascular resistance with respect to , and we are actively working on this. For more information on this exciting area of physiology, and a description of the various physiological conditions in which is, and is not, detected, we refer the reader to very recent reviews (Duke et al. 2014; Lovering et al. 2015) of the relevant literature to date on this topic.