Supine, prone, right and left gravitational effects on human pulmonary circulation
Body position can be optimized for pulmonary ventilation/perfusion matching during surgery and intensive care. However, positional effects upon distribution of pulmonary blood flow and vascular distensibility measured as the pulmonary blood volume variation have not been quantitatively characterized. In order to explore the potential clinical utility of body position as a modulator of pulmonary hemodynamics, we aimed to characterize gravitational effects upon distribution of pulmonary blood flow, pulmonary vascular distension, and pulmonary vascular distensibility.
Healthy subjects (n = 10) underwent phase contrast cardiovascular magnetic resonance (CMR) pulmonary artery and vein flow measurements in the supine, prone, and right/left lateral decubitus positions. For each lung, blood volume variation was calculated by subtracting venous from arterial flow per time frame.
Body position did not change cardiac output (p = 0.84). There was no difference in blood flow between the superior and inferior pulmonary veins in the supine (p = 0.92) or prone body positions (p = 0.43). Compared to supine, pulmonary blood flow increased to the dependent lung in the lateral positions (16–33%, p = 0.002 for both). Venous but not arterial cross-sectional vessel area increased in both lungs when dependent compared to when non-dependent in the lateral positions (22–27%, p ≤ 0.01 for both). In contrast, compared to supine, distensibility increased in the non-dependent lung in the lateral positions (68–113%, p = 0.002 for both).
CMR demonstrates that in the lateral position, there is a shift in blood flow distribution, and venous but not arterial blood volume, from the non-dependent to the dependent lung. The non-dependent lung has a sizable pulmonary vascular distensibility reserve, possibly related to left atrial pressure. These results support the physiological basis for positioning patients with unilateral pulmonary pathology with the “good lung down” in the context of intensive care. Future studies are warranted to evaluate the diagnostic potential of these physiological insights into pulmonary hemodynamics, particularly in the context of non-invasively characterizing pulmonary hypertension.
KeywordsCardiovascular magnetic resonance imaging Pulmonary circulation Posture
Acute respiratory distress syndrome
Body mass index
Body surface area
Balanced steady-state free precession
Cardiovascular magnetic resonance
Left ventricle/left ventricular
Left ventricular end-diastolic volume
Left ventricular end-diastolic volume index
Left ventricular ejection fraction
Left ventricular mass
Left ventricular mass index
Pulmonary blood volume variation
Positron emission tomography
Pulmonary vascular distensibility reserve
The influence of body position on the pulmonary circulation is incompletely understood , and can have important implications both for diagnosis and risk stratification in cardiopulmonary disease [2, 3], as well as therapeutic applications in surgery, anesthesia, and intensive care [1, 3, 4, 5]. Prior knowledge about right/left pulmonary blood flow distribution in the lateral position [6, 7] comes primarily from small studies in non-healthy anesthetized human subjects using intravenous injection of Xenon133 which may miss shunted blood flow [8, 9, 10], and animal studies performed on small sample sizes [1, 11]. Notably, these studies have yielded conflicting results [12, 13].
Non-invasive characterization of the pulmonary circulation by phase contrast (PC) cardiovascular magnetic resonance (CMR) has potential for improving screening, diagnosis, characterization, and risk stratification in pulmonary hypertension . While right heart catheterization remains the reference standard, previous results suggest that mean pulmonary arterial pressure can be accurately estimated using 4D flow PC-CMR . Also, CMR is ideally suited to quantify vessel cross-sectional area as a measure of pulmonary vascular distension.
2D PC-CMR has been used to measure the pulmonary blood volume variation (PBVV) during the cardiac cycle, as a surrogate of pulmonary vascular distensibility . PBVV has been shown to be unchanged in systemic sclerosis patients who often have pulmonary arterial hypertension , but is reduced in left ventricular (LV) dysfunction in swine after experimental acute myocardial infarction . This suggests that PBVV primarily reflects pulmonary venous or capillary distensibility, which may be affected in left-sided heart disease, whereas pulmonary arterial distensibility may be affected in pulmonary arterial hypertension. Consequently, PC-CMR may be able to non-invasively detect pulmonary hypertension and differentiate between pre- and post-capillary etiologies. Furthermore, the influence of body position upon PBVV is unknown. Altering body position can potentially serve as a “gravitational stress test”, which may unmask further information about the pulmonary circulatory hemodynamics.
In order to explore the potential clinical utility of body position as a modulator of pulmonary hemodynamics, we aimed to investigate the influence of body position on the distribution of pulmonary arterial and venous flow, pulmonary vascular distension, and pulmonary vascular distensibility using PC-CMR in healthy subjects.
Some of the results of these studies have been previously reported in the form of an abstract .
This prospective study was conducted in non-smoking healthy subjects. Inclusion criteria were an age of 18–65 years, absence of prior cardiopulmonary or systemic disease, and absence of contraindications for CMR. Exclusion criteria were abnormal electrocardiography (ECG) findings, systolic blood pressure outside 90–140 mmHg, and pregnancy. Approval was obtained from the local ethics committee on human subject research, and all participants provided written informed consent.
Immediately prior to the CMR scan, all patients underwent screening including standard 12-lead ECG (Marguette, General Electric Healthcare, Little Chalfont, United Kingdom), blood pressure measurement, and a patient history to exclude cardiopulmonary or systemic disease and contraindications to CMR.
CMR was performed on a 1.5 T scanner (Aera, Siemens Healthineers, Erlangen, Germany) with ECG gating and phased array receiver coils, without any contrast agent. Participants sequentially underwent CMR scanning in four positions: supine, prone, and right, and left lateral decubitus. Flow measurements were obtained using a free-breathing velocity-encoded PC sequence with number of excitations set to three in order to reduce the impact of respirophasic variation. The total acquisition time was approximately 2.5 min depending on heart rate, representing approximately 30–40 respiratory cycles. Typical image acquisition parameters included 256 × 192–216 matrix size, 1.4 × 1.4 mm in-plane resolution, 5 mm slice thickness, TR/TE 30/2.7 ms. For all four body positions, flow was measured in the main pulmonary artery (velocity encoding 100 cm/s), right and left pulmonary arteries (velocity encoding 125 cm/s), and all four pulmonary veins (velocity encoding 80 cm/s). Phase contrast images were acquired during normal breathing and had a temporal resolution of 35 frames per cardiac cycle, roughly corresponding to 25–30 ms depending on heart rate. In order to image through-plane flow perpendicular to each vessel, a balanced steady state free precession (bSSFP) sequence was used to image the long axis of each vessel as previously described . Upon acquisition, phase contrast images were screened for quality and image plane placement. Additional attempts were made until images were of acceptable quality and imaging plane. The full protocol for all sides including supine cine images and a two-minute break between body positions required approximately two hours. Typical scan duration for all pulmonary blood vessels in one body position was approximately 25 min. This included time for a localizer sequence, flow measurements in the main pulmonary artery, both pulmonary artery branches and all four pulmonary veins as well as, on average, one repeated attempt at imaging a pulmonary vein that was difficult to visualize. We estimate that unilateral measurements including one artery and two pulmonary veins can be obtained in ~ 8–10 min for one side in a clinical setting.
Furthermore, the imaging protocol performed in the supine position included 2-, 3-, 4-chamber long axis bSSFP cine images of the LV as well as a short-axis stack. Typical image parameters for bSSFP images included 256 × 216 matrix size with 1.5 × 1.5 mm in-plane resolution, TR/TE 36/1.16 ms.
Female gender, n (%)
27 ± 1
177 ± 12
72 ± 12
23.0 ± 0.5
1.9 ± 0.1
QRS duration, ms
96 ± 3
Heart rate, beats/min
61 ± 3
Cardiac output, l/min
6.2 ± 0.3
Cardiac index, l/min/m2
3.3 ± 0.2
60 ± 2
105 ± 7
55 ± 2
76 ± 8
39 ± 3
187 ± 12
98 ± 4
When a subject is lying in the lateral position, the term gravity-dependent or simply dependent lung refers to the lowermost lung, which is also referred to as being in the dependent position.
When a subject is lying in the lateral position, the term gravity-non-dependent or simply non-dependent lung refers to the uppermost lung, which is also referred to as being in the non-dependent position.
Generally refers to an enlargement, dilation, or ballooning effect. In this context we use distension to refer to an increase in intra-vessel blood volume. We used the changes in vascular cross-sectional area as a measure of changes in vessel distension.
Generally refers to the capacity to distend as a result of pressure from inside. In cardiovascular physiology, vascular distensibility, or compliance, refers to the change in intravascular volume per unit of change in intravascular pressure. In the current study, we used PBVV as a marker of distensibility.
Pulmonary blood volume variation (PBVV)
The cumulative change in intravascular blood volume in the pulmonary circulation that occurs during the cardiac cycle. This is measured in ml, by using CMR flow imaging to measure the net inflow and net outflow of blood into and out of the pulmonary circulation. PBVV can be measured either combined for both lungs, or independently for each respective lung.
The PBVV in ml, divided by the arterial flow volume per heartbeat in ml. Thus, relative PBVV is a fraction which in this study is expressed as a percentage.
Pulmonary vascular distensibility reserve (PVDR)
The pulmonary vascular distensibility reserve (PVDR) was defined as the percent change in unilateral relative PBVV when changing between body positions.
If supine left lung PBVV = 25 ml and left pulmonary artery net flow = 50 ml, then left relative PBVV = 50%.
If in the same subject, in the right lateral position, left lung PBVV = 32 ml and left pulmonary artery net flow = 40 ml, then left relative PBVV = 80%.
Left PVDR is then calculated as:
∆ relative PBVV / supine relative PBVV = (80–50)/50 = 60%.
Statistical testing was performed using SPSS Statistics (version 24, Statistical Package for the Social Sciences (SSPS), International Business Machines, Inc., Armonk, New York, USA). Continuous variables were reported as mean ± standard error of the mean (SEM). Categorical variables were presented as percentages. Differences in continuous variables were tested using the Wilcoxon signed-rank test or the Mann-Whitney U test, as appropriate. Differences in repeated measures between multiple groups was tested with the Friedman test. A p-value less than 0.05 was considered statistically significant. Due to the uncertainty of effect magnitudes and measurement variation in the impact of body position of PBVV and relative PBVV, no a priori power calculations were made. We measured the interobserver variability for PBVV, relative PBVV for the left lung in the supine and the right lateral positions, as well as for left lung PVDR, by having a second observer perform independent image analysis on the same raw data. The interobserver variability is reported as the mean percent difference and standard deviation from the average of the two observers, as the Dahlberg error, relative Dahlberg error and as the intra-class correlation coefficient (ICC).
The study participants were 27 ± 1 years (range: 22–31 years), and were equally distributed between the sexes. Subject characteristics are summarized in Table 1. There were no signs of atelectasis in any subject in any body position. We found no pulmonary vein abnormalities or unusual variations in our cohort, including the presence of any separate right middle lobe vein.
Impact of body position on cardiac output
Impact of body position on pulmonary blood flow distribution
Impact of body position on pulmonary venous flow distribution
The mean ± SD difference between main pulmonary arterial flow and summed pulmonary venous net flow ranged between 0 ± 7 ml (0 ± 7%) and 4 ± 4 ml (4.0 ± 3.8%) across the four body positions.
Impact of body position on cross-sectional vessel areas
Impact of body position on absolute PBVV
Impact of body position on relative PBVV
Positional effects upon the pulmonary circulation
We found a right/left distribution of pulmonary blood flow of 54/46% in the supine position, shifting to 63/37% in the right lateral position and 41/59% in the left lateral position. We further found a larger increase in blood flow to the left lung in the dependent position, possibly due to anatomical right/left differences in vessel tree architecture and/or lung size.
Anesthesiology textbooks [6, 7] cite known measures of right/left pulmonary blood flow distribution according to body position in humans. These textbooks state that the right/left distribution of pulmonary blood flow is 55/45% in upright/supine, 65/35% in right lateral position and 45/55% in the left lateral position [6, 7], with references to studies in dogs, primates or humans [8, 10, 11, 20, 21]. These studies used varying respiratory conditions with regards to subjects breathing spontaneously, or being anesthetized and intubated with varying positive airway pressures. Importantly, these studies utilized varying methods for measuring blood flow including the use of radiotracers or extrapolations from oxygen consumption (VO2) measurements in small sample sizes, and show somewhat diverging results. In concordance with the results of the present study, one study found a 43/57% right/left distribution of pulmonary perfusion in the left lateral position using intravenous injection of Xe133 and measurement of the exhaled concentration of Xe133 in a double lumen endotracheal tube in patients without cardiopulmonary disease . In contrast, a different study found a close to balanced distribution between the dependent and non-dependent lungs in both lateral positions, with most subjects having a majority of blood flow to the dependent lung with a few extreme outliers . Those authors speculated that this may be due to Xe133 flowing through shunts in atelectases in the dependent lung not being exhaled. Other studies have been performed to evaluate the effects of gravity in the lateral position on the pulmonary circulation using positron emission tomography (PET) or pulmonary scintigraphy [23, 24]. These studies indicate that blood flow tends to favor dependent regions of the lungs, but showed that gravity has a lesser impact particularly in the prone position. Taken together, prior results have been inconclusive with varying results in limited sample sizes and potentially imprecise methods. The current study is the first study to use robust methods to provide definitive data on the magnitude of the effects of the lateral position on pulmonary blood flow distribution in spontaneously breathing awake humans.
Positional impact on superior/inferior venous flow
Our results show that the supine and prone positions did not affect the balance between flow in the ventral upper lung lobes drained by the superior pulmonary veins, and the more dorsal lower lung lobes drained by the inferior pulmonary veins. This is in agreement with the above-mentioned nuclear imaging studies showing no meaningful perfusion gradient between dorsal and ventral lobes in the prone position [23, 24, 25]. This suggests that the improved gas exchange observed in the prone position compared to supine  is unrelated to gravity-induced redistribution of blood flow from the lobes drained by the superior and inferior pulmonary veins, respectively. Although the difference in superior and inferior venous flow volume approached reaching statistical significance in the left lateral position (p = 0.05), we see no reasonable physiological explanation behind this finding. Furthermore, the magnitude of the superior/inferior venous flow difference did not differ in the right lateral position, indicating that the observed difference in the left lateral position may be due to statistical noise.
Mechanism of positional effect on pulmonary blood flow distribution
The three-zone model initially developed by West, et al. , describes the impact of gravity on the pulmonary circulation. One mechanism governing this model is the gravity-induced differences in hydrostatic pressures which affect vessel distension and therefore regional vascular resistance and blood flow. Further studies have demonstrated large heterogeneity in blood flow within isogravitational planes . This has led to the conclusion that vessel anatomy is the major determinant of regional pulmonary blood flow, and that gravity is an important but secondary determinant . Our results strongly suggest that gravity plays a role in determining the distribution of blood flow between the dependent and non-dependent lungs in the lateral positions. Importantly, our study is the first to provide data to support the notion that there is a redistribution of blood volume in the pulmonary venous circulation that coincides with the changes in blood flow.
Reproducibility of PBVV
We found the combined PBVV in both lungs to be roughly 50 ml and 50% of stroke volume (46 ± 4 ml, 44 ± 2%), and this agreed excellently with a previous study of healthy subjects in the supine position (48 ± 5 ml, 46 ± 6%) . This suggests that the method to determine PBVV by CMR is reproducible across different healthy populations, sites, CMR scanner vendors, and operators. Our findings regarding inter-observer variability were found to be consistent with the findings of a previous study that reported a 1 ± 6% inter-observer variability for PBVV .
There are several sources for measurement error in using 2D flow measurements which are important to minimize in order to avoid large compounded measurement errors from several vessels. We attempted to minimize error through carefully placed imaging planes perpendicular to the vessel course, eddy current compensation with careful static tissue masking and manual review of automated vessel delineation in all time frames. This resulted in an average difference between pulmonary arterial and summed venous flow in our cohort that we believe is acceptable for all body positions.
Physiology and potential diagnostic utility of PBVV
Our results demonstrate that PBVV varies with body position to a greater extent than does ipsilateral blood flow, and that this appears to be related to gravity-induced variations in distension of the pulmonary veins rather than the pulmonary arteries. The determinants of PBVV and their relative importance have yet to be explored, since PBVV is a relatively recently described measure. We hypothesize that important determinants of PBVV include the combined arterial and venous vessel wall compliance, circulating blood volume and its distribution measured as vessel distension, and blood flow volume per heartbeat. The non-dependent lung should experience decreased hydrostatic pressure due to gravity, and receive less blood, translating into decreased venous distension. This is supported by our finding of decreased venous cross-sectional area in the non-dependent position compared to the same lung in the dependent position. Due to the relative lack of smooth muscle in the walls of the pulmonary veins, the vessel wall compliance of pulmonary veins can be expected to be high, and thus pulmonary venous blood volume should be sensitive to pressure changes induced by the effects of gravity on blood volume and blood flow distribution related to body position. Taken together, our findings support the notion that PBVV is a marker inversely related to pulmonary venous distension. This is consistent the previous finding that LV dysfunction in swine reduces PBVV , possibly through increased left atrial pressure leading to congestion and distension of the pulmonary veins. Notably, in that study of experimental infarction , pulmonary wedge pressure was increased after infarction, and this is consistent with increased left atrial pressure contributing to a decrease in PBVV. This implies that PBVV could potentially be used to non-invasively gauge left atrial pressure. Specifically, we speculate that increased left atrial pressure may lead to a smaller change in relative PBVV with body position. Thus, left atrial pressure would be inversely related to PVDR, in concordance with the expectation that an increase in left atrial pressure would increase pulmonary venous distension and decrease the distensibility reserve. Notably, this hypothesis remains to be evaluated and tested prospectively.
The potential clinical role of PC-CMR in characterizing pulmonary hypertension is tantalizing in light of the findings that 4D flow PC-CMR can accurately, precisely, and non-invasively measure mean pulmonary arterial pressure . Taken together, PC-CMR could potentially be used to both identify pulmonary hypertension as such, and also possibly to characterize it as pre- or post-capillary in nature, based on estimation of left atrial pressure by measuring PVDR.
It remains to be determined whether a 4D flow approach could be useful in measuring PBVV and PVDR in a way that is less dependent on the operator of the CMR scanner, and equally as accurate and precise as for 2D PC imaging.
An important limitation of this study was the limited sample size which was adequate to glean results with regards to physiology, but insufficient to measure normal references ranges for PBVV and PVDR in healthy subjects. Another limitation was that although our inclusion criteria permitted a variety of ages, the subjects most accessible to us were young and healthy which is reflected in our cohort. Thus, the resulting homogenous cohort did not permit studying the effects of normal aging and disease on PBVV and PVDR, which remain to be explored. The presence of atelectasis may potentially influence both PBVV and PVDR. In future studies exploring the utility of PBVV and PVDR in subjects with pulmonary disease, it may be necessary to develop strategies to account for atelectasis. Furthermore, an important limitation is that conditions causing irregular heart rate, such as atrial fibrillation or excessive premature contractions, preclude the accurate measurement of PBVV and PVDR due to their detrimental effect on the reliability of CMR flow measurements.
In addition, the test-retest repeatability of PBVV, relative PBVV and PVDR remain to be studied in order to quantify the differences in these measures that can be accurately measured. An additional limitation is that observed venous vessel areas can be influenced by the placement of the image plane closer to or further away from the left atrium, due to the pronounced tapering of pulmonary veins.
In case of pulmonary venous abnormalities and individual variations, these would have to be taken into account when measuring PBVV. Additional pulmonary veins may require separate imaging planes, while certain shunt variations may make the measurement of PBVV unreliable for some patients.
Clinical implications and future directions
Our results provide quantitative data on blood flow distribution that help explain the drop in partial oxygen pressure seen in patients with acute respiratory distress syndrome (ARDS) in the lateral position, with the most affected lung in the dependent position [26, 28]. Lateral position PBVV, and PVDR, are novel measures which may theoretically correlate with left atrial pressure through venous distension. Measurement of left atrial pressure, LV end-diastolic pressure, or pulmonary wedge pressure are invasive, associated with risk, and difficult to motivate measuring in healthy volunteers. Therefore, the use of PVDR to non-invasively distinguish between pre- and postcapillary pulmonary hypertension by estimating left atrial pressure should be investigated in patients undergoing clinical invasive pressure measurements, and such studies are underway.
This study is the first to provide definitive data confirming widespread notions of right/left pulmonary blood flow distribution in the lateral position. Furthermore, we corroborate previous findings of pulmonary vascular distensibility, measured as PBVV in healthy subjects, suggesting robust reproducibility of the method. We conclude that there is an increase in PBVV in the non-dependent lung compared to when the same lung is in the supine position. The positional increase in PBVV in the non-dependent lung is more pronounced when the blood flow volume to the same lung is taken into account. We introduce the novel measure PVDR to describe this positional change in relative PBVV. Also, positional changes in cross-sectional vessel areas suggest that PVDR is due to gravity-induced differences in venous, but not arterial, distension between the dependent and non-dependent lung. Thus, PVDR may be related to left atrial pressure. Future studies are warranted to evaluate the potential clinical role of these novel measures, primarily in the context of characterizing pulmonary hypertension.
Study design: BW, JGR, MU. Subject recruitment: BW. Data acquisition: BW, JGR. Data analysis: BW, JGR. Statistical analysis: BW, JGR. Results interpretation: BW, JGR, MU, JP, MA. Manuscript preparation and critical revision: BW, JGR, JP, MA, MU. All authors read and approved the final manuscript.
Dr. Wieslander, Dr. Ramos and Dr. Ugander were supported in part by the Swedish Research Council, the Swedish Heart and Lung Foundation, Stockholm County Council, and Karolinska Institutet.
Ethics approval and consent to participate
This study was approved by the Regional Ethical Review Board in Stockholm and all subject provided written informed consent.
Consent for publication
Dr. Ugander is principal investigator for an institutional research and development agreement regarding cCMR between Karolinska University Hospital and Siemens.
- 6.Benumof JL. Anesthesia for thoracic surgery. W. B. Saunders Company; 1995.Google Scholar
- 7.Aitkenhead AR, Moppett IK, Thompson JP. Smith & Aitkenhead's Textbook of Anaesthesia. 6th edition ed: Churchill Livingstone, Elsevier 2013.Google Scholar
- 16.Kanski M, Arheden H, Wuttge DM, Bozovic G, Hesselstrand R, Ugander M. Pulmonary blood volume indexed to lung volume is reduced in newly diagnosed systemic sclerosis compared to normals--a prospective clinical cardiovascular magnetic resonance study addressing pulmonary vascular changes. J Cardiovasc Magn Reson. 2013;15:86.CrossRefGoogle Scholar
- 17.Wieslander B, Ramos JG, Ax M, Petersson J, Ugander M. Supine, prone, right and left gravitational effects on human pulmonary vascular physiology (abstract). Paper presented at: CMR 2018 - the joint EuroCMR/SCMR meeting 2018; Barcelona. Spain. .Google Scholar
- 28.Hewitt N, Bucknall T, Faraone NM. Lateral positioning for critically ill adult patients. Cochrane Database Syst Rev. 2016;5:CD007205.Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.