Because of technical challenges very little is known about absolute myocardial perfusion in humans in vivo during physical exercise. In the present study we applied positron emission tomography (PET) in order to 1) investigate the effects of dynamic bicycle exercise on myocardial perfusion and 2) clarify the possible effects of endurance training on myocardial perfusion during exercise. Myocardial perfusion was measured in endurance-trained and healthy untrained subjects at rest and during absolutely the same (150 W) and relatively similar [70% maximal power output (Wmax)] bicycle exercise intensities. On average, the absolute myocardial perfusion was 3.4-fold higher during 150 W (P < 0.001) and 4.9-fold higher during 70% Wmax (P < 0.001) than at rest. At 150 W myocardial perfusion was 46% lower in endurance-trained than in untrained subjects (1.67 ± 0.45 vs. 3.00 ± 0.75 ml·g−1·min−1; P < 0.05), whereas during 70% Wmax perfusion was not significantly different between groups (P = not significant). When myocardial perfusion was normalized with rate-pressure product, the results were similar. Thus, according to the present results, myocardial perfusion increases in parallel with the increase in working intensity and in myocardial work rate. Endurance training seems to affect myocardial blood flow pattern during submaximal exercise and leads to more efficient myocardial pump function.
- myocardial blood flow
- endurance training
- positron emission tomography
because of technical challenges very little is known about absolute myocardial perfusion in humans in vivo during physical exercise. Myocardial oxygen uptake is strikingly increased during heavy exercise compared with the resting state (11, 22, 29). The increases in myocardial oxygen uptake and demand are mainly met by decreasing coronary resistance and augmenting the myocardial blood flow (24). The myocardial oxygen extraction fraction (OEF) at resting state is already at a high level (70–80%) (39), and therefore the further increases in oxygen uptake are tightly coupled with the changes in myocardial blood flow (39).
The effects of endurance training on both cardiac structure and myocardial function (see e.g., Refs. 27, 35) and blood flow (see, e.g., Refs. 7, 12) are widely studied. To measure myocardial blood flow or perfusion, invasive coronary techniques have been used in many studies, but only estimates of flow per gram of tissue can be given. In general, endurance training has been shown to be associated with reduced sympathetic activation (4, 7, 12) that decreases the heart rate at rest (25) and at any given submaximal workload (31). This has been shown to result in a lower resting myocardial blood flow per gram of myocardium in the trained heart (23, 24). However, in some (20, 36, 38) but not all (5, 11, 12, 23) recent human studies the basal myocardial perfusion was found to be similar between endurance-trained and untrained subjects. During exercise coronary blood flow increases in proportion to the heart rate (11, 15, 22, 26, 34, 39), and thus it has been proposed that the decreased heart rate during submaximal exercise due to endurance training is accompanied by parallel decreases in coronary blood flow compared with untrained animals (2, 39). However, studies on the effects of endurance training on myocardial blood flow during exercise in humans are sparse. Ekblom et al. (7) previously showed that myocardial blood flow increases from resting values by ∼100% and by ∼200% in endurance athletes and sedentary subjects, respectively, during exercise at 65% maximal oxygen uptake (V̇o2 max). Similar changes were also found in the study by Heiss et al. (11). These findings suggest that endurance training might lead to lesser increase in myocardial blood flow, at least during near-maximal exercise.
Myocardial blood flow during exercise has mainly been studied in animal models or by invasive catheter methods in humans. Although several techniques are available (e.g., dilation techniques or intravascular ultrasound Doppler), these methods are often relatively invasive or have the ability to measure only the coronary blood flow instead of myocardial perfusion. Positron emission tomography (PET) allows direct quantitative measurements of tissue metabolism and perfusion noninvasively, and it has been used to examine myocardial perfusion at rest and during pharmacological stress. Thus the present study was designed to 1) investigate the effects of exercise on myocardial perfusion noninvasively in vivo in humans and 2) investigate whether myocardial perfusion is different in endurance-trained athletes and untrained subjects during exercise.
Seven healthy endurance-trained (ET) and seven healthy untrained (UT) men matched for age, body mass index, and blood pressure volunteered for the study (Table 1). The subjects had no clinical or laboratory evidence of coronary heart disease, diabetes, or systemic hypertension. They were not currently taking any medication, and all were nonsmokers. The endurance athletes had a history of 5 yr of aerobic training for at least five times a week. Two of the seven ET subjects were runners, one was a rower, three were orienteers, and one had a combined running and cross-country skiing background. The UT subjects were not currently participating in any systematic exercise training. The purpose and potential risks of the study were explained to the subjects, and written informed consent to participate in the study was obtained. The study was performed according to the Declaration of Helsinki, and the Ethical Committee of the Hospital District of Southwest Finland approved the study protocol.
First, maximal power output (Wmax) and V̇o2 max were determined. Two weeks later the subjects were instructed to maintain a regular diet but to avoid caffeinated drinks and alcohol during the last 24 h before PET studies. Subjects were also instructed to avoid any kind of strenuous physical activity 2 days before the PET experiment. Echocardiographic measurements and PET studies were performed at least 3 h after a light breakfast. Left ventricular (LV) structure and function were assessed with two-dimensional guided M-mode echocardiography. In addition, a catheter was inserted in an antecubital vein for injection of radioactive tracer ([15O]H2O) and blood sampling. Subjects then laid down on a vacuum base and were thereafter adjusted and fastened to the cycle ergometer in a semisupine position with three broad straps over the upper body. Simultaneously, the subjects were allowed to familiarize themselves with the cycle ergometer. The ergometer was inserted into the PET scanner as shown in Fig. 1, and the position of the subject inside the scanner was carefully marked and repeatedly checked during the experiment with help of laser lights. After transmission scan myocardial perfusion was measured at rest (Basal) and during two bicycle exercise loads with radiowater ([15O]H2O). The first exercise intensity was absolutely the same in all subjects (150 W), and the second intensity was individually selected to be relatively similar (70% Wmax) (ET 69 ± 2% vs. UT 70 ± 4% Wmax; P = 0.593) for both groups (Fig. 1).
Measurements of LV structure and function were assessed with two-dimensional M-mode echocardiography (Acuson 128XP, Acuson, Mountain View, CA) to determine end-diastolic and end-systolic wall thicknesses, LV dimensions, and volumes. All echocardiographic recordings and analyses were performed by the same investigator. The study subjects rested for at least 15 min before they were examined in a left lateral decubitus position. LV dimensions and volumes were measured in a standardized way. LV mass was calculated according to the Penn convention: LV mass = 1.04 × [(end-diastolic diameter + posterior wall thickness + septal thickness)3 − end-diastolic diameter3] − 13.6 g (6). LV work was calculated as LV work = cardiac output (CO) × mean arterial pressure (MAP), and forward LV work per gram of tissue was calculated as LV work/LV mass.
Measurements of myocardial perfusion.
The positron-emitting tracer, [15O]H2O, was produced as previously described (37). An ECAT 931/08 tomograph (Siemens/CTI, Knoxville, TN) was used for image acquisition. Before the emission scans, subjects rested at least 20 min, during which a transmission scan for the correction of photon attenuation was performed. Thereafter, [15O]H2O was injected intravenously for 2 min, and dynamic PET scanning was performed for 6 min (Basal). Subjects then started bicycle exercise for a total duration of 25 min. Simultaneously, myocardial perfusion was measured at the intensities of 150 W and 70% of Wmax (Fig. 1). The exercise protocol is described in detail below. All PET data were corrected for dead time, decay, and measured photon attenuation. Images were processed with the iterative reconstruction algorithm (1).
Calculation of myocardial perfusion.
Regions of interest (ROIs) were drawn on four representative transaxial mid-LV slices covering the anterior and lateral myocardial walls (whole) of the LV as previously reported (20). ROIs were drawn on the images obtained at rest and were copied to the images obtained during both exercise intensities. The input function was obtained from the LV time-activity curve with a previously validated method (13). Myocardial perfusion was then calculated with the previously introduced method employing the single-compartment model (14), and mean perfusion values at rest and during both exercise intensities were obtained. Perfusion resistance was calculated by dividing MAP by the respective perfusion value. Myocardial perfusion values during exercise were also individually normalized to correspond to the same myocardial work with rate-pressure product (RPP) and the equation RPP-normalized perfusion = measured individual myocardial perfusion at 150 W or 70% Wmax × (RPP at rest/RPP at 150 W or 70% Wmax). When the myocardial perfusion was analyzed, the data of two subjects (1 ET and 1 UT) obtained during both exercise bouts were observed to be too noisy, and thus these two subjects were excluded from the perfusion results.
Exercise during PET.
After Basal measurements subjects started the exercise, using a cycle ergometer (Bosch ERG 555, Geschäftsbereich Elektronik, Germany). First, subjects cycled 2 min at the level of 40% of measured Wmax (ET 133 ± 5 vs. UT 93 ± 6 W; P < 0.001). Thereafter, the exercise intensity was increased to 150 W, and after 3-min exercise myocardial perfusion was measured at this intensity. This absolute exercise intensity corresponded to the relative intensity of 42 ± 2% of Wmax for ET and 59 ± 4% of Wmax for UT subjects. Immediately after the perfusion measurement at the 150-W exercise level, the exercise intensity was lowered back to the level of 40% of Wmax for 5 min. Thereafter, the exercise intensity was increased to 70% of Wmax (ET 250 ± 14 vs. UT 180 ± 13 W; P < 0.001) for 9 min to repeat the perfusion measurement (Fig. 1). The pedaling rate was 50–55 rpm. Such a low pedaling rate was chosen in order to minimize body motion during PET scanning.
Plasma glucose concentration was analyzed in duplicate by the glucose oxidase method with an Analox GM7 glucose analyzer (Analox Instruments, London, UK). Serum free fatty acid (FFA) concentrations were measured with an enzymatic method (ACS-ACOD Method, Wako Chemicals, Neuss, Germany). Serum total cholesterol, high-density lipoprotein cholesterol, and triglyceride concentrations were measured with standard enzymatic methods (Boehringer Mannheim, Mannheim, Germany) and a fully automated analyzer (Hitachi 704, Hitachi, Tokyo, Japan). The low-density lipoprotein cholesterol concentration was calculated by the formula of Friedewald et al. (8). Blood lactate concentrations were measured with standard enzymatic methods (Boehringer Mannheim) and a Hitachi 917 automatic analyzer (Hitachi).
Heart rate (HR) during PET experiment was automatically recorded with a ECG recorder (MAC 5000; GE Marquette Medical Systems, Milwaukee, WI). Blood pressure (BP) was monitored with an automatic oscillometric BP analyzer (Omrom, Tokyo, Japan) throughout the entire PET procedure. MAP was calculated as: MAP = BPdiast + [⅓ × (BPsyst − BPdiast)], where BPdiast is diastolic BP and BPsyst is systolic BP. The RPP was then calculated as MAP × HR. V̇o2 max and Wmax were determined by bicycle ergometer test in the sitting position with direct respiratory measurements 2 wk before the PET procedure. The criteria used to establish the V̇o2 max were a plateau in V̇o2 despite increase in working intensity and a respiratory quotient of >1.1.
Wmax and HR were also measured in five additional healthy sedentary subjects (age 28 ± 1 yr) during two incremental bicycle exercise tests (2-min exercise steps) in randomized sitting and semisupine positions in order to obtain the normalizing factor for submaximal exercise loads between Wmax test and bicycle exercise during the PET procedure. Wmax in the semisupine position was 7 ± 4% lower (275 ± 35 vs. 296 ± 35 W; P = 0.02) than in the sitting position. Thus, to equalize the submaximal bicycle exercise load used during the PET experiment, the exercise loads in the Wmax test were individually multiplied by a factor of 0.93. The HRmax in the semisupine position was similar to that in the sitting position (semisupine 184 ± 10 vs. sitting 189 ± 10 beats/min; P = 0.168).
Statistical analysis was performed with SAS/STAT statistical software release 8.2 (SAS Institute, Cary, NC). Normal distribution of parameters was tested with the Kolmogorov-Smirnov test. Analysis of variance for repeated measurements with Tukey's post hoc procedure was used when testing the differences between groups and different exercise intensities. Student's t-test was used when appropriate. Linear relationship between parameters was tested by Pearson's correlation coefficient. The significance level was set at P < 0.05. Data are presented as means ± SD.
Cardiac structure and function.
Both the absolute and body surface area-normalized LV mass were higher in ET than UT subjects (Table 1). Resting LV ejection fraction, E-to-A ratio, and CO were similar between groups, whereas stroke volume was higher in ET compared with UT subjects. Resting LV work was 41% lower in the ET group.
Hemodynamic and biochemical parameters.
HR was significantly lower in ET than UT subjects during all measurements (Table 2 and Fig. 2). In contrast, no difference between the groups was observed in BPsyst, BPdiast, or MAP (Table 2 and Fig. 2). RPP was significantly lower in ET compared with UT subjects in all other measurements but not during 70% Wmax exercise load (Table 2). The only biochemical parameter in which differences were observed was blood lactate, which increased during exercise and was lower in ET than UT subjects during both exercise bouts (Fig. 3).
Myocardial perfusion and perfusion resistance.
Resting myocardial perfusion was similar between groups (Table 3 and Fig. 4A). Exercise at 150 W increased myocardial perfusion 3.4 ± 1.5-fold from resting values (Fig. 4A). During exercise at 70% Wmax myocardial perfusion was 4.9 ± 1.8- and 1.6 ± 0.8-fold higher than at rest and during exercise at 150 W, respectively (Fig. 4A). Myocardial perfusion normalized by RPP was not different from rest at either exercise intensity (Fig. 4B). At the exercise intensity of 150 W, both absolute and RPP-normalized myocardial perfusion were lower in ET subjects, but at 70% Wmax no statistically significant difference between groups was observed (Fig. 4). Perfusion resistance was similar between the groups at rest and decreased during exercise (Fig. 5). In the entire study population myocardial perfusion was linearly related to workload (Fig. 6A) and HR (Fig. 6B).
There are two main physiological observations in the present study. First, myocardial perfusion increases linearly with increases in exercise intensity. Second, endurance-trained athletes have ∼35% lower myocardial perfusion at the same absolute exercise intensity when myocardial work is also taken into account. However, during higher and relatively similar exercise intensity (70% Wmax) there is no significant difference between ET and UT subjects.
Cardiac structure and function.
ET subjects demonstrated significant cardiac hypertrophy (∼60% higher LV mass) compared with UT subjects. However, we did not observe any difference in systolic or diastolic function between the groups. Thus our results for LV mass and cardiac function are in accordance with earlier studies (23, 35). Stroke volume was significantly higher and LV work lower in ET than UT subjects, which in turn demonstrates more efficient pump function at rest in the trained heart, as proposed earlier (11, 20).
The resting myocardial perfusion in both groups was comparable with our (19, 20) and other (36, 38) earlier studies performed with the PET technique. However, Kjaer et al. (23) recently found lower myocardial perfusion in rowers than that in the present study. The observation of similar resting myocardial perfusion between ET and UT subjects also agrees with previous cross-sectional studies (19, 36, 38) but not with one longitudinal study of Czernin et al. (5), who found decreased resting perfusion in the endurance-trained state. Similarly, our results disagree with the some studies performed with other techniques in which lower basal blood flow in coronary arteries of endurance-trained subjects has been observed (11, 12). Variety in training backgrounds among study subjects could contribute to this discrepancy between present and previous findings since different sports have been shown to induce different structural adaptations in the heart (27). Whether these unequal adaptations might explain the differences in resting coronary blood flow or myocardial perfusion remains still unclear.
We found that myocardial perfusion in ET subjects during 150-W exercise intensity (∼44% of Wmax) was 2.5-fold higher than at rest, while in UT subjects the increase from rest to 150 W (∼59% of Wmax) was 4.2-fold. Furthermore, during 70% Wmax exercise load myocardial perfusion in the entire study population was 4.9-fold higher than at rest (Fig. 6A). We also observed that the changes in myocardial perfusion were associated with the changes in HR in the entire study population (Fig. 6B). These results are in accordance with earlier studies showing that myocardial perfusion increases in parallel with the increase in workload and the increase in myocardial work rate (7, 11).
It has been proposed that lower HR during exercise due to endurance training might lead to smaller increase in myocardial blood flow (23, 24). We found lower HR in ET subjects than in UT subjects during the entire experiment, and thus our results of lower myocardial perfusion at 150-W exercise intensity in ET subjects agree with these earlier findings (23, 24). Myocardial perfusion was also lower in ET subjects than in UT subjects when myocardial perfusion values were normalized by myocardial work (RPP). This finding also confirms the results of Heiss et al. (11), who found lower coronary blood flow in endurance athletes at the working intensity of 65% of V̇o2 max. It is noteworthy that Heiss et al. (11) used a workload relative to V̇o2 max but in our study the workload was relative to Wmax.
The finding that ET subjects had lower myocardial perfusion at 150 W implies that a trained myocardium consumes less oxygen than an untrained myocardium. Thus one possible explanation might be training-induced changes in energy metabolism in the heart. It has been shown that exercise training increases the basal rates of FFA and glucose uptake in myocardium (3, 18, 32). When exercise intensity is increased from 30% to 55% of V̇o2 max myocardial glucose uptake is doubled compared with resting conditions (9, 21), which in turn suppresses the FFA uptake (glucose-FFA cycle) (9, 21, 40, 41). In addition, the amount of ATP produced per oxygen is more energy efficient through glucose oxidation compared with FFA oxidation. These factors together could suggest that exercise-stimulated myocardial glucose uptake is enhanced by training, which indeed has been shown with rat models (16–18). Although the exact mechanisms behind training-associated improvements in glucose metabolism (3, 28) and in perfusion-related changes in the present study are still unclear, it is likely that lower myocardial perfusion during 150-W exercise intensity is caused by more efficient energy metabolism in the trained heart, as proposed earlier (11).
As a consequence of less efficient substrate metabolism, higher myocardial perfusion in UT subjects at 150-W exercise intensity could also be related to higher CO2 content in the circulation. As a mediator of metabolic regulator of perfusion, elevated CO2 levels likely lead to increased myocardial perfusion (30). However, changes in circulating CO2 levels were not measured in the present study.
During the highest exercise intensity (70% Wmax) the difference in absolute myocardial perfusion diminishes between ET and UT subjects. Interestingly, the difference in RPP-normalized myocardial perfusion between groups is similar to that with 150-W exercise intensity. However, this difference was not statistically significant because of high variation in the product of BPsyst and HR in UT subjects. In the light of our findings, endurance training induces changes in the myocardial perfusion pattern during submaximal exercise, but these changes diminish during near-maximal exercise. Future studies are warranted to investigate whether endurance training associates with lower myocardial perfusion also at higher exercise intensities compared with untrained subjects.
In animals, peak coronary blood flow values during maximal dynamic exercise are typically three to five times higher compared with the resting level (26, 34, 39). In humans, Grubbstrom et al. (10) found with coronary sinus catheterization and thermodilution measurement that maximal coronary blood flow was 3.5-fold higher than at rest in sedentary subjects, although further increase was observed when maximal exercise was performed during hypoxia (10). However, by extrapolating from the results of the present study peak myocardial perfusion in humans would be ∼5 ml·g−1·min−1 (∼7-fold higher than resting myocardial perfusion), assuming that the relationship between myocardial perfusion and workload remains linear (Fig. 6).
In the present study the myocardial perfusion values were found to be slightly higher than in some other studies (10, 11). One possible explanation could be methodological differences between studies. In human studies (10, 11) and in several animal experiments (e.g., Refs. 26, 33, 39) blood flow in the coronary arteries was measured, and only in some cases were the estimates of flow per gram of tissue given. Thus this may have affected myocardial perfusion values and can partly explain the differences between studies. However, we were able to measure myocardial perfusion (blood flow per mass of tissue) directly from the tissue noninvasively during bicycle exercise by using PET. Another likely explanation is the different training backgrounds within study subjects in different studies as mentioned above. Rowing, cycling, swimming, and cross-country skiing have been shown to have the highest hypertrophic effect on LV cavity dimension and wall thickness (27). Interestingly, in our study six of seven athletic subjects were runners or orienteers, but body surface area-normalized LV mass was much higher than, for example, in rowers in the study by Kjaer et al. (23). Moreover, the only subject with a rowing background in our study did not have a different LV mass (324 g) compared with the other ET subjects (range 225–384 g).
Finally, the number of study subjects in the present study was relatively small, and this may have a direct effect on the results of our study. Therefore, additional studies are needed to confirm our findings.
To the best of our knowledge this is the first experiment in which the PET technique has been used to measure the effects of endurance training on myocardial perfusion during exercise. Myocardial perfusion was found to increase in parallel with increases in workload. Resting myocardial perfusion was similar between endurance athletes and untrained subjects, whereas endurance athletes were demonstrated to have ∼35% lower myocardial perfusion at the same absolute bicycle exercise intensity (150 W) even when myocardial work was also taken into account. However, during high-intensity exercise (70% Wmax) there was no difference between trained and untrained subjects either in myocardial perfusion or in perfusion resistance. In conclusion, endurance training seems to induce changes in myocardial perfusion pattern, which indicates more efficient myocardial energy metabolism and pump function during moderate exercise intensity.
This study was supported by grants from the Academy of Finland (Grants 206970 and 204240), the Finnish Cultural Foundation, The Finnish Foundation for Cardiovascular Research, and the Ministry of Education (Grants 143/722/2002, 51/722/2003, and 40/627/2005).
We thank the personnel of the Turku PET Centre for their excellent assistance.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society