Left ventricular (LV) systolic function increases with passive heat stress (HS); however, less is known about diastolic function. Eight healthy subjects (24.0 ± 2.0 yr of age) underwent whole body passive heating ∼1°C above baseline (BL). Cardiac magnetic resonance imaging was used to measure biventricular volumes, function, filling velocities, volumetric flow rates, and LV twist and strain at BL and after 45 min of HS. Passive heating reduced left atrial volume (−17.6 ± 11.7 ml, P < 0.05), right and LV end-diastolic volumes (−22.7 ± 11.0 and −25.7 ± 24.9 ml, respectively; P < 0.05), and LV stroke volume (−6.7 ± 6.8 ml, P < 0.05) from BL. LV ejection fraction (EF), end-systolic elastance, septal and lateral mitral annular systolic velocities, circumferential strain, and peak LV twist increased with HS (P < 0.05). Right ventricular stroke volume, EF, and systolic tissue velocities were unchanged with HS (P > 0.05). Early LV diastolic tissue and blood velocities and strain rates were maintained with HS, whereas untwisting rate increased significantly from 166.4 ± 46.9 to 268.7 ± 76.8°/s (P < 0.05). The major novel finding of this study was that, secondary to an increase in peak LV twist and untwisting rate, early diastolic blood and tissue velocities and strain rates are maintained despite a reduction in filling pressure.
- cardiac magnetic resonance imaging
- biventricular function
during passive heat stress (HS), cutaneous vascular conductance increases to augment skin blood flow to facilitate heat loss (3, 34). To maintain blood pressure and tissue oxygen delivery, cardiac output rises (17, 18, 33, 34, 41, 42), and efferent sympathetic nerve activity increases to noncutaneous vascular beds (32), redistributing blood flow away from skeletal muscle (21) and splanchnic vascular beds (7, 18, 34, 35) toward the skin. Secondary to this redistribution of blood flow, thoracic blood volume and central venous pressure is also reduced significantly (17, 28, 34, 41, 42).
The effects of reduced cardiac filling pressure on ventricular function remain unclear. A recent study reported that early diastolic filling and mitral tissue velocities are maintained during HS (3), suggesting that early diastolic function is maintained despite reduced cardiac filling pressures. However, the effect on left ventricular (LV) end-diastolic volume remains equivocal, since previous studies have shown it decreases (41) or remains unchanged (7) during passive HS.
The development of minimum LV pressure is a function of active relaxation and elastic recoil of the LV (13, 22). During systole, basal-to-apical rotation of the LV (twisting) serves to augment stroke volume, whereas, in addition, systolic torsional deformation produces potential energy that is stored in myocardium (1, 2) and released as kinetic energy that facilitates LV diastolic untwisting. LV untwisting is therefore closely associated with the generation of LV diastolic suction (4, 24, 38). Increased twist and untwisting may provide the mechanistic explanation for previous demonstration of maintained diastolic function under reduced filling pressures; however, this has not been studied to date.
The purpose of this investigation was to examine the mechanisms that maintain early diastolic function despite a reduction in filling pressures with HS. We hypothesized that LV untwisting would increase with HS as a result of increased systolic twist facilitating the maintenance of early diastolic filling through increased ventricular suction.
Eight healthy, sedentary males participated in this investigation. All subjects provided written informed consent before study participation, which was approved by the University of Alberta Health Research Ethics Board. Before each part of the experiment, each subject was asked to abstain from caffeine, alcohol, and strenuous physical activity for at least 24 h.
Day 1. Incremental exercise test.
Each subject completed an incremental exercise test on a cycle ergometer to measure maximal oxygen consumption (V̇o2 max).
Day 2. Passive HS.
Participants dressed in a tube-lined suit that covered the entire body, with the exception of the face, hands, and feet (Allen-Vanguard Technologies, Ottawa, ON). The subject lay in the supine position, covered with a heavy blanket to reduce heat loss to the environment while thermoneutral water (34°C) was circulated throughout the suit for 30 min to establish a physiological baseline (BL). Immediately after BL, 50°C water was circulated throughout the suit for 60 min (whole body HS). This trial took place a minimum of 7 days before the magnetic resonance imaging (MRI) trial to avoid the potential for acclimation. Core and skin temperatures could not be monitored during MRI because of the presence of the ferrous metal in the telemetry pill. Therefore, this test was used to establish the physiological response to passive heating on day 3.
Day 3. Cardiac MRI and passive HS.
After arriving at the Peter S. Allen MRI center at the University of Alberta, subjects dressed in the tube-lined suit, and the experimental protocol described for day 2 was repeated while cardiac MRI were acquired.
Maximal oxygen consumption.
V̇o2 max was assessed with an incremental protocol on a cycle ergometer (Monark 894E, Varberg, Sweden), with the workload set at 50 W increasing by 25 W every 2 min until ventilatory threshold, after which the workload was increased by 25 W every minute until volitional exhaustion. Ventilatory threshold was defined as a systematic rise in the ventilatory equivalent for O2, without an increase in the ventilatory equivalent for CO2 (8). Expired gases were collected and analyzed with a metabolic measurement cart (TrueOne 2400; ParvoMedics), calibrated with known gas concentrations before each test. Flow was calibrated using a 3.0-liter syringe.
At least 4 h before the beginning of the familiarization trial (day 2), subjects ingested a thermistor pill and were fitted with four dermal patches for the telemetric measurement of core and skin temperature (VitalSense; Mini Mitter, Bend, OR). Mean skin temperature was calculated as previously described (31).
Arterial blood pressure was obtained from an automatic blood pressure cuff, and mean arterial blood pressure was calculated as one-third pulse pressure plus diastolic pressure. Heart rate was continuously monitored by electrocardiogram.
Short axis cine images covering the length of the right ventricle (RV) and LV were acquired to assess end-diastolic and end-systolic volume, stroke volume, and ejection fraction. Image acquisition parameters were as follows: repetition time = 3.0 ms; echo time = 1.5 ms; flip angle = 78°; slice thickness = 8 mm with a 2-mm gap between slices; matrix = 256 × 192; field of view = 300–380 mm; 25 ms temporal resolution with 64 reconstructed phases. Short axis through-plane phase contrast (velocity) images, prescribed at the level of the mitral leaflet tips, were acquired to measure early (E) and atrial (A) filling velocities and volumetric flow rates. Peak early diastolic and systolic annular tissue velocities were assessed from a four-chamber cine image. Myocardial tissue tagging was applied in four to six short axis slices covering the entire LV to assess peak twist and peak untwisting rate (using only apical and basal slices), as well as global peak circumferential and radial strain (using all short axis slices). Tags were applied ∼150 ms after the R wave to ensure tag persistence and clarity throughout diastole. All MRI data were cardiac gated based on an electrocardiogram and acquired during end-expiratory breath holds. The order of image acquisition was kept consistent, with total image acquisition time between 15 and 20 min during both BL and following 45 min of passive HS.
RV and LV volumes were assessed by manual endocardial tracing of short-axis cine images at end-diastole and end-systole by a single nonblinded observer (M. D. Nelson). The papillary muscles were included in all tracings. End systole was defined visually as the phase with the smallest LV and RV volumes. Left atrial volumes were also calculated using short axis slices, at a phase just before mitral valve opening (as determined from a 3-chamber cine image), to capture the largest volume over the cardiac cycle. Using our custom software, long-axis cine images were also used to delineate basal and apical cardiac borders, for both ventricular and atrial volumes. Left ventricular end-systolic cavity area (ESCA) and end-systolic myocardial area (ESMA, epicardial border) at the level of the papillary muscles was used to calculate end-systolic wall stress (1.33 × SBP × ESCA/ESMA, where SBP is systolic blood pressure) as previously described (14). Total peripheral resistance was calculated as mean arterial blood pressure divided by cardiac output. End-systolic elastance was calculated as systolic arterial blood pressure divided by end-systolic volume. Our in-laboratory intrarater and interrater reliability for measuring LV volumes, reported as a coefficient of variation, were determined to be 2.6 and 2.1%, respectively.
Volume flow rates were calculated as the sum of velocities over the filling orifice multiplied by the pixel cross-sectional area (16, 37). Peak early diastolic and systolic annular tissue velocities were assessed from a four-chamber cine image, using a manual tracking program for calculation of annular position and velocity over the entire cardiac cycle. In addition to conventional assessment of E and A waveforms, a correction of the A wave at higher heart rates was performed to estimate the contribution from the overlapping E wave, as shown in Fig. 1.
The time of aortic valve closure and mitral valve opening was determined using short axis blood velocity images that were oriented to contain the leaflet planes. Cardiac events were referenced to aortic valve closure (%systolic duration) and mitral valve opening (ms). Systolic duration started from the time of ventricular depolarization (R wave, electrocardiogram) and ended with the time of aortic valve closure. All subsequent events were referenced to a percentage of this duration.
Tag analysis was fully automated, using image morphing software developed in house, to determine the spatial deformation field for the myocardium as a function of cardiac phase, relative to a reference cardiac phase. The tag processing software was developed and is used in the MATLAB environment using an open source algorithm for the image morphing calculations (http://elastix.isi.uu.nl/index.php). User input was limited to tracing the endo- and epicardium at a single reference cardiac phase for each slice. The angle of rotation for each short axis slice, Φ, was calculated as the average rotation of each point relative to the reference phase. The reference phase, relative to which all deformations are referenced, was selected in diastasis, before atrial contraction. All values were averaged over all positions, spanning endo- to epicardium and all circumferential positions, and over all slices for strain measurements. The global LV twist, θ, was calculated as the difference between the counterclockwise rotation at the apex and clockwise rotation at the base (viewed from apex to base), θ = Φapex − Φbase. The rate of untwisting was calculated as the discrete time derivative of the twist vs. time curve. Twist was defined as the peak value on the curve, and peak untwisting rate was calculated as the minimum value of the time derivative of the twist curve directly following the time of peak twist (23). The peak rates of rotation were also calculated separately for the apex and base. As shown in Fig. 2, strain and strain rates were also defined as the peak value on the corresponding strain and strain rate curves (27).
Data are presented as a mean ± SD. Paired sample t-tests were used for statistical comparison between thermal conditions. Simple linear regression was used to identify the relationships between selected variables. The results of statistical analyses were considered to be significant when P < 0.05.
Participant characteristics were as follows (mean ± SD): age: 24.0 ± 3.0 yr; height: 181.9 ± 4.8 cm; body mass: 81.8 ± 12.6 kg; V̇o2 max: 42.3 ± 2.6 ml·kg−1·min−1 (range: 39.7 −48.2 ml·kg−1·min−1). The aerobic fitness level of our subjects was consistent with untrained members of this age category (43).
Whole body passive HS significantly elevated skin temperature from 33.8 ± 1.0°C at BL to 38.8 ± 0.5°C, and core body temperature increased from 37.1 ± 0.3°C at BL to 38.0 ± 0.5°C.
Passive HS reduced (P < 0.05) RV and LV end-diastolic and end-systolic volumes and left atrial volume (Figs. 3 and 4). LV systolic annular velocities at the septum (BL 7.1 ± 1.0 cm/s vs. HS 8.8 ± 1.5 cm/s; P < 0.05) and lateral wall (BL 8.5 ± 2.0 cm/s vs. HS 12.1 ± 2.8 cm/s; P < 0.05) and LV end-systolic elastance (BL 1.7 ± 0.4 mmHg/ml vs. HS 2.3 ± 0.6 mmHg/ml) significantly increased with HS (P < 0.05), whereas LV end-systolic wall stress (BL 79.5 ± 7.3 mmHg·cm2 vs. 66.8 ± 9.2 mmHg·cm2; P < 0.05) was reduced, resulting in a significant increase (6.5%) in LV ejection fraction (Fig. 3). Peak twist and peak systolic circumferential strain was greater during HS compared with BL, whereas radial strain was unchanged (Table 2). Despite the significant improvement in systolic function, LV stroke volume was reduced with HS (Fig. 3). As a result, cardiac output was augmented entirely by an increase in heart rate (BL 69.8 ± 3.9 beats/min vs. HS 108.1 ± 9.8 beats/min; P < 0.05).
RV stroke volume tended to decline in response to HS, but the changes were not statistically significant (Fig. 4, P = 0.08). Both RV ejection fraction (Fig. 4) and systolic annular tissue velocity (BL 12.3 ± 1.6 cm/s vs. HS 14.3 ± 2.9 cm/s; P = 0.164) increased with HS (P > 0.05).
LV untwisting rate increased significantly with HS (see Table 2), with no change in early diastolic filling velocity (Table 1). Peak early diastolic filling rates were reduced with HS while the late filling rates and filling velocities were unchanged (Table 1). All other diastolic functional parameters were unchanged with HS, including the early diastolic annular tissue velocities at the septal and lateral wall (Table 1) and the RV free wall (12.4 ± 1.7 vs. 13.5 ± 5.8 cm/s), E/Esept′ (6.7 ± 1.1 vs. 7.1 ± 2.3), E/Elat′ (6.4 ± 1.5 vs. 6.2 ± 4.8), and the circumferential and radial diastolic strain rates (Table 2).
Tables 1 and 2 report selected cardiac events in both absolute time (ms) and as a percentage of systolic duration. At both BL and with HS, the sequence of events was as follows: peak twist, followed by the peak untwisting rate and finally the peak blood and tissue velocities associated with early ventricular filling.
Timing Changes With HS
Peak twist was delayed with respect to aortic valve closure with HS. The early diastolic annular tissue velocities, Elat′ and Esep′, that normally precede the peak filling E wave velocity are significantly delayed (P < 0.05) and follow the E wave peak velocity with HS. The isovolumic relaxation time decreased by 13.6 ms with HS (P < 0.05).
The primary findings of this investigation were that whole body passive HS significantly reduced biventricular preload and left atrial volume while early diastolic functional parameters were maintained. It is proposed that increased LV recoil and suction, associated with augmented LV twist and subsequent untwisting rates, offset the reduction in diastolic function that normally accompanies reductions in central blood volume and filling pressures.
Biventricular Function During Passive HS
Whole body HS is associated with a large increase in cutaneous vascular conductance and concomitant decreases in central venous pressure (6, 7, 17, 28, 36, 42) and LV filling pressure (41, 42). The effect of HS on LV end-diastolic volume has not been definitively established, since the available literature is conflicting. Crandall et al. (7), using multiple-gated acquisition imaging, demonstrated that LV end-diastolic volume was unaltered by passive heating (despite reduced central venous pressure). In contrast, Wilson et al. (41), using two-dimensional echocardiography, found that passive heating reduced LV end-diastolic volume and pulmonary capillary wedge pressure by ∼11% and ∼29%, respectively. Our results confirm that HS reduces LV end-diastolic volume (11.3%) and expand on prior work by showing a 13.5% reduction in RV end-diastolic volume. The difference between these studies remains difficult to explain; however, measurement sensitivity is a likely explanation for the differing results.
During periods of reduced preload, the maintenance of stroke volume is largely dependent on enhanced LV contractility and/or reduced afterload. Consistent with previous reports (3, 7), passive HS significantly increased LV systolic function, as evidenced by an increase in LV ejection fraction, end-systolic elastance, and peak septal and lateral mitral annular velocities, which were coupled with reduced LV end-systolic wall stress. Basal-to-apical systolic rotation (twist) was also found to increase with HS, contributing to the reduction in end-systolic volume. To our knowledge, this is the first study to report that passive HS increases LV twist and circumferential strain. Notwithstanding enhanced contractility, we observed a small but significant decrease in stroke volume (4%). Previous studies have reported stroke volume to be maintained or even slightly elevated during HS (34, 41, 42). However, our results suggest that, during passive HS, the reduction in end-systolic volume does not completely compensate for the decrease in end-diastolic volume.
LV Diastolic Function During Passive HS
It is well known that reductions in venous return and thus filling pressures will significantly reduce both early mitral inflow velocity and mitral annular velocities (11, 20, 26, 29). Mild unloading (−30 mmHg of lower body negative pressure) has been shown to reduce E wave velocities from 77 to 53 cm/s and E′ in the lateral wall from 15.0 to 10.1 cm/s (11). Similar unloading, at −40 mmHg of lower body negative pressure, in healthy young volunteers reduced LV end-diastolic volume by 14.8 ml with a reduction of E wave velocities by 15 cm/s (10). The reduction in these velocities is associated with reduced atrial pressure, particularly at the time of mitral valve opening (5, 15). In the present study, we found that passive HS yielded a significant reduction in LV end-diastolic volume, similarly to these unloading studies, by 20.3 ± 11.7 ml (P < 0.05), with a reduction in left atrial volumes, by −17.6 ± 11.7 ml (P < 0.05). However, in contrast to unloading studies, early mitral inflow and mitral annular velocities were unchanged. The maintenance of early diastolic blood and tissue velocities, as well as circumferential and diastolic strain rates, given the significant volume unloading suggests that the ventricular contribution to early diastolic function is augmented with passive heating. Our results therefore suggest that the increased LV untwisting rate, which is strongly associated with increased recoil and LV suction (9, 12, 19, 24, 25, 30), is responsible for maintaining early diastolic function, counteracting the effects of the reduction in left atrial volume and, likely, filling pressures.
The relationships between twist, untwisting rates, and early filling have been documented previously. End-systolic volume and the extent of LV systolic twist are inversely related to the magnitude of untwisting generated during diastole (22, 39, 44). Nearly half of untwisting occurs before mitral valve opening, during isovolumic relaxation (19, 24, 30), and the extent of this recoil is regarded as an important contributor to LV suction (12, 19, 24, 25, 30), with the greatest portion of LV pressure decay occurring during the isovolumic relaxation period. We propose that, despite a marked reduction in left atrial volume with HS, augmented LV untwisting is responsible for maintaining the forces that drive early filling in the present experiment, as observed via the maintained filling velocities, tissue velocities, and strain rates.
Similar to Brothers (3), early filling blood velocity was not significantly changed with HS, whereas the late filling velocity was significantly increased. Volumetric filling rates during the A wave also increased significantly with HS, but the early filling rate was slightly reduced. Although the total A wave filling blood velocities and flow rates are significantly increased with HS, the relative contributions from the existing E wave inertia and the A wave contraction cannot be directly measured because of the significant fusion of the two waveforms. Thus, although there are clearly increased velocities and filling rates at the time of the A wave, no conclusions regarding the magnitude of the atrial contraction can be drawn. By estimating and subtracting the residual E wave velocities (or volume flow rates) at the time of the atrial contraction, as outlined in Fig. 1, modified peak A wave values were approximated, which were shown to be similar to the BL A wave values, suggesting that the contractile force generated by the atrium is unchanged with HS. Also, given the dependence of atrial muscle contraction on the initial fiber length [analogous to the Frank-Starling mechanism of the ventricular myocardium (40)], and the reduction in venous return to the atrium with HS (associated with the reduction in central blood volume), it is thus likely that the enhanced A wave velocities and flow rates with HS, and changes in E-to-A wave ratios, are purely a heart rate phenomenon.
Several limitations must be considered when interpreting the present results. Core temperature could not be measured on the same day as MRI for safety reasons (i.e., telemetry pill contains ferrous metal). We are confident, however, that we can predict core temperature from prior HS exposures within a very small margin of error. Pilot work in heat-stressed subjects confirmed a low between-day variation in the relative change in core and skin temperature from BL (0.1 and 0.6°C, respectively), equivalent to 12.4 and 10.1% coefficient of variation, respectively. Central venous pressure or pulmonary capillary wedge pressure was not measured in the present experiment; however, numerous investigations using similar protocols have previously documented this response (7, 41, 42). A complete cardiac scan took ∼10–15 min, which corresponded to a 0.4 ± 0.1°C change in core temperature during the passive HS condition. It is therefore possible that measurements near the beginning of the scan (e.g., volumes) may not correspond precisely with measurements taken near the end of the scan (e.g., tissue velocities). However, any error made would be systematic and therefore not selectively influence one condition over the other. Finally, the reported average changes in the timing of some events with HS, such as the reduction in isovolumic relaxation time, are less than the temporal resolution of the imaging studies. It is possible that there are systematic errors in these values because of limitations of temporal resolution.
In conclusion, our study demonstrates that whole body passive HS: 1) reduces cardiac preload, as evidenced by reduced LV and RV end-diastolic volumes, 2) increases LV contractility, as demonstrated by an increase in mitral systolic tissue velocities, increased LV end-systolic elastance and ejection fraction, and augmented LV twist, and 3) that increased LV twisting gives rise to increased untwisting rates during HS, and maintains indexes of global diastolic function despite decreases in filling pressures, as reflected by reductions in left atrial volumes. LV untwisting has previously been associated with increased LV suction (9, 24, 38); as such, we reason that, without an increase in LV untwisting rate in the presence of reduced left atrial volume, LV end-diastolic volume would be reduced to a greater extent during passive heating, leading to larger reductions in stroke volume.
M. D. Nelson and J. Cheng-Baron were supported by doctoral scholarships from the Natural Sciences and Engineering Research Council of Canada. M. J. Haykowsky has a career award from the Canadian Institutes of Health Research. S. R. Petersen was supported by the Department of National Defense. R. B. Thompson is an Alberta Heritage Foundation for Medical Research Scholar.
No conflicts of interest are declared by the authors.
We acknowledge Zoltan Kenwell for technical assistance, Brad Welch and Meredith Giroux for assistance with data collection, and Gordon Sleivert and the Canadian Sport Centre, Pacific, for the use of equipment.
- Copyright © 2010 the American Physiological Society