Experimental Physiology (1993), 78, 639-648 Printed in Great Britain POTASSIUM LOSS FROM SKELETAL MUSCLE DURING EXERCISE IN MAN: A RADIOISOTOPE STUDY MOHAMMED S. QAYYUM, CHRISTOPHER A. J. FREEMANTLE*, CHRISTOPHER J. CAREY*, BRIAN C. PAGE*, NIGEL SOPERt, DAVID J. PATERSON AND PETER A. ROBBINS University Laboratory of Physiology, Parks Road, Oxford OX] 3PT, the * MRC Cyclotron Unit, Hammersmith Hospital, Du Cane Road, London W12 OHS and t Medical Physics, John Radcliffe Hospital, Headington, Oxford OX3 9DU (MANUSCRIPT RECEIVED 15 OCTOBER 1992, ACCEPTED 23 APRIL 1993) SUMMARY Muscle potassium (K+) content decreases during exercise. Previous studies, in humans, have used measurements of arteriovenous plasma potassium concentration differences (AV A[K+]) and/or muscle biopsy to measure the loss of muscle K+ during exercise. In the current study a non-invasive method was developed to measure skeletal muscle K+ before and after exercise, using an isotope of K+, potassium-43 (43K+). Twelve subjects performed single-leg extension exercise for 2 h at 50 % of their maximum predicted heart rate. The level of radioactivity from the quadriceps femoris was determined before exercise and during two periods post-exercise. After correction for counts arising outside the exercised muscle, we estimate a decrease in muscle K+ content of 3-2 + 155 % (mean + S.E.M.) following exercise. The muscle K+ was not restored following 75 min of recovery. The decrease in muscle K+ following exercise in our study is considerably less than that suggested by previous studies using AV A[K+] measurements but not so dissimilar from results obtained using muscle biopsy. We conclude that a small but significant loss of K+ occurs following prolonged dynamic exercise, and that complete recovery of muscle K+ is slow. INTRODUCTION During exercise, potassium (K+) is released from skeletal muscle predominantly during repolarization of the muscle cell membrane. During prolonged dynamic exercise there is a decrease in muscle K+ content, which may be due to incomplete reuptake by the muscle Na+-K+-ATPase (Clausen, 1990). This net loss of muscle K+ during exercise results in an increase in plasma K+ concentration (Kilburn, 1966). Measurements of femoral arteriovenous differences in plasma K+ concentration (AV A[K+]) have been used to study the net loss of K+ from leg skeletal muscle during sustained periods (1-2 h) of dynamic exercise (Sj0gaard, 1986; Sahlin & Broberg, 1989). The results of these two studies were rather different. Sj0gaard (1986) reported a 22% decrease in muscle K+ content and Sahlin & Broberg (1989) reported a 2% fall. A particular problem with this technique is that small errors in the measurement of arterial and venous [K+] result in large variations in estimated K+ efflux. Muscle biopsy has also been used to assess the decrease in muscle K+ content during sustained (2 h) periods of dynamic exercise (Ahlborg, Bergstrom, Ekelund & Hultman, 1967; Sj0gaard, 1986). Both studies found a decrease in muscle K+ content. Ahlborg et al. * Laboratory at which experimental work was carried out.
640 M. S. QAYYUM AND OTHERS (1967) reported a 4-4 % decrease and Sj0gaard (1986) an 11 % decrease. A problem with the analyses of tissue from muscle biopsy specimens following prolonged exercise, is that the muscle can become oedematous (Bergstrom, Guarnieri & Hultman, 1971) and hyperaemic (Edwards, Maunder, Jones & Batra, 1975) which may complicate the determination of muscle electrolyte content. In addition, muscle biopsy determines the K+ content in a very small fraction of the total muscle mass. This may cause error if the losses of K+ from the muscle are not homogenous. In addition, repeat biopsies may not yield samples with the same mix of fibre types, as these may vary within a particular muscle (Johnson, Polgar, Weightman & Appleton, 1973). A further limitation to the methods described above is that both are highly invasive. We therefore developed an alternative non-invasive method for measuring changes in muscle K+ following prolonged exercise. This would also allow K+ changes in other tissue, apart from the exercising muscle, to be measured. Some of these results have been presented previously in abstract form (Qayyum, Freemantle, Carey, Page, Paterson & Robbins, 1993). METHOD S Development of technique A radioisotope of potassium, potassium-43 (43K+, Medical Research Council Cyclotron Unit, Hammersmith Hospital, London) was used and the changes in the radioactive counts over skeletal muscle before and after exercise were measured. 43K+ has gamma energies that are too high to be measured using standard equipment in clinical use. Instead, a sodium iodide scintillation detector from a whole body counter (Spinks, Bewley, Ranicar & Joplin, 1977) was used. The crystal size of 15 cm diameter by 10 cm depth gives the detector a high sensitivity, but a low spatial resolution. To improve the spatial resolution, a seven-hole converging lead collimator was constructed. Overall the collimator was of 10 cm diameter at the face distal to the detector, 15 cm diameter at the proximal face, and 12 cm in length. At the distal face, each of the holes was of 1 cm diameter, arranged in a hexagonal configuration, with 0-5 cm septa, giving an effective diameter of 4 cm. At the proximal face each hole was of 3 cm diameter, giving an effective diameter of 10 cm. The collimator was fitted to the detector face which could then be placed over muscle mass to count 43K+. The attachment of the collimator to the scintillation detector caused a ca 96 % reduction in counts. Because 43K+ has a short physical half-life of 22-4 h, all the counts were corrected for radioactive decay. Time zero was taken as the time of the first measured experimental count. All counts were measured over 1 min periods (counts min-') and there was a ca 30 s delay between each 1 min period, for data to be transferred to a computer. During this delay no further counts were made. Collimating properties of the system. To investigate the collimating properties of the system, a line spread function was determined using a capillary tube filled with distilled water and ca 0 75 MBq of 43K+. The line spread was measured by moving the capillary tube from a position directly in front of the collimator to, in turn, either side of the collimator, in a plane parallel to the face of the collimator. The number of counts detected per minute at each of the different positions was measured. Evaluation ofsystem using a phantom. To investigate the amount of radiation detected from outside the expected field of view of the collimator a number of experiments were performed using a phantom man. The phantom man consisted of separate, hollow, body segments of a similar volume for a standard man (Bush, 1946). The first experiment, protocol 1, involved putting fixed proportions of a 1 MBq dose of 43K+ into all the body segments of the phantom man. The proportion of the 1 MBq dose used for each segment was based upon the normal distribution of potassium in man (Bewley, 1988). The collimated detector was positioned over the lateral aspect of the left thigh segment and five sets of counts, of 1 min duration, were taken. A second set of five 1 min counts, protocol 2, was made with 43K+ in the left thigh replaced with distilled water, but with 43K+ in all the other phantom segments. Finally, a third set of five 1 min counts, protocol 3, was made with 43K+ only in the left thigh segment and distilled water in all the other segments. In each of these three protocols the appropriate activity of 43K+ in any segment, when present, was the same. Accurate repositioning was ensured by using fixed markers aligned along the side of the phantom segments.
SKELETAL MUSCLE K+ LOSS DURING EXERCISE The second phantom man study started with a repeat of protocol 1 to provide a baseline value for the left thigh segment. The activity in the left thigh was then increased in fixed increments based upon the initial volume of 43K+ added to the left thigh segment. The increments were drawn from the same reservoir of 43K+ that was used for the initial 1 MBq of 43K+ added to the whole phantom. After each increment in volume, another period of counting was undertaken. Physiological experiments Subjects. Twelve healthy male subjects aged 20-26 years participated in the experiment. Subjects gave informed consent after a detailed verbal and written explanation of the study. Permission to inject the 43K+ into human subjects was obtained from the Administration of Radioactive Substances Advisory Committee (ARSAC). Ethical permission was obtained from the Central Oxford Research Ethics Committee. The individual subject characteristics are shown in Table 1. Experimental protocol Twenty-four hours before the experiment, the subjects received a 1 MBq dose of 43K+ intravenously which allowed the 43K+ time to equilibrate adequately within the body compartments, (Jasani & Edmonds, 1971). On the day of the experiment, the subject was positioned supine on a platform within the whole body counter. The subject's skin was marked with a pen to indicate anatomical landmarks. External pointers on the apparatus were fixed in position to overlie these landmarks so that the subject could be repositioned accurately on the platform. The collimated detector was positioned over the lateral aspect of the anterior half of the left thigh. To determine whether there were any significant effects of subject repositioning, the pre-exercise counting period was divided into four subperiods. In each subperiod a set of five counts was made, each of 1 min duration. Between each subperiod the subject was moved and repositioned into the original experimental position. The subject then performed 2 h of single-leg extension exercise at 50 % of the subjects maximum predicted heart rate (Astrand & Rodahl, 1986), which was determined in a prior set of experiments. The exercise involved extension of the left leg which had weights wrapped around the ankle. The subject would fully extend the leg every second in time with a metronome and then allow the leg to swing back passively. The subject was asked to indicate the level of discomfort in his exercising thigh by pointing to a Borg scale (Borg, 1982), every 5 min during exercise. Heart rate was measured from an electrocardiogram recording from a standard, bipolar limb lead. Following exercise the subject was repositioned within the whole body counter in the same manner as for the pre-exercise counts. The time from the end of exercise to the start of counting was kept to less than 1 5 min. The counting period consisted of twenty counts, each of 1 min duration. Following a half-hour rest the subject was again repositioned on the couch and a further twenty counts of 1 min duration were obtained. RESULTS Development of method Collimation. The line spread function for the collimator is shown in Fig. 1. As would be expected, the maximum number of counts is detected when the source is directly over the central axis of the collimator. As the source is moved there is a sharp drop in the counts which reach a minimum 5 cm either side of the central axis. As the source is moved further there is a small secondary asymmetric increase in the counts either side of the collimator. This increase in the counts is probably due to radioactivity getting past the lead shielding in the area where the collimator is attached to the detector. Evaluation ofsystem using a phantom. Figure 2 shows the results from the three separate protocols. There are 471 (100%) counts detected from protocol 1 the control experiment, 96 (20 %) from protocol 2 and 358 (76%) from protocol 3. The difference of 24% between protocols 1 and 3 represents the calculated counts originating from phantom segments other than the left thigh. This difference is similar in magnitude to the measured counts originating from segments other than the left thigh, protocol 2. The sum of protocols 2 and 3 (96%) which represents all the phantom segments, is similar in magnitude to the measured counts from all the segments, protocol 1. 641
642 M. S. QAYYUM AND OTHERS Table 1. Physical characteristics of subjects Subject Age Height Weight number (years) (metres) (kg) 802 20 1-79 70 3 835 20 1-76 71-5 836 22 1-81 73-5 837 21 1-85 78-0 843 26 1-76 75-8 853 20 1-81 74-3 860 24 1-70 58 6 865 25 1 71 53-0 876 22 1-76 76-0 887 20 1-87 68-0 909 22 1-86 76-5 911 25 1-85 74 0 32000 r 24000 F E 16000 I 8000k o.~~~~~~~, -80-40 0 40 80 Distance from centre (cm) Fig. 1. The line spread function. Figure 3 shows the linear relationship between the counts detected and the concentration of 43K+ present. Extrapolation of the results to 0 % gives a value of 24 % for the counts originating from phantom segments other than the left thigh. This figure is similar to the previous values.
SKELETAL MUSCLE K+ LOSS DURING EXERCISE 643 500 400 7300 -c 0 ~200- c 100T 0 1 2 3 Protocol Fig. 2. The mean counts (± 2 S.E.M.) from the three protocols of the first phantom man study. Protocol 1, all the segments contain 43K+. Protocol 2, all segments apart from the left thigh contain 43K+. Protocol 3, only left thigh segment contains 43K+. The difference between protocol 1 and protocol 3 represents counts originating from phantom segments other than the left thigh, which is similar to the result from Protocol 2. 600 500 400 r. X 300 0 200 100 0 0 20 40 60 80 100 I 140 Percentage increase of 43K+ from 100%/ Fig. 3. The counts (± 2 S.E.M.) measured from the phantom left thigh as the concentration of 43K+ in the thigh is increased in proportions of the original 100%. The remaining segments contain 43K+ based on the normal distribution of K+ in man.
644 M. S. QAYYUM AND OTHERS 150 r r_ e (4 Cd 0.0 (U 4- Cd +1 6. Lld 4) x 100 p 50 I x f x I Z--.& x x x T +f++4 JL a) o 0 m I-I I II I 0 L I~~~~~~~~~~~~~~~ 0 20 40 60 80 100 120 20 18 F 16 F 14F,r-,~ I. - - - / ',/ / I %.- - i.' I / _-I /N, - - - N / I / 12- -I itaj IV Il ti O 20 40 60 Time (min) 80 100 120 Fig. 4. The mean heart rate (±2 S.E.M.) and perceived exertion response (PER, dashed lines indicate range of values obtained) for all 12 subjects during exercise. The heart rate in all the individual subjects remained fairly constant with increasing leg fatigue. The majority of the subjects exercised at heart rates of around 100 beats min-', with one subject (802) exercising at 140 beats min-'. Physiological experiments The measured 1 min counts in these experiments are normalized for each subject by the mean of the pre-exercise counts for the respective subject. Repositioning. The means of the normalized counts for each of the four pre-exercise subperiods for each subject were used to assess whether there was any significant variation produced by repositioning of the subjects. A one-way analysis of variance showed that the variation in the subperiods arising from the repositioning of a subject (mean square error for individual observations 218 %), was not significantly greater (P> 005) than the intrinsic variation between the 1 min counts within each subperiod (mean square error 18.2 %). Heart rate and perceived exertion response. Heart rate during exercise (Fig. 4) rapidly increased to a plateau where it remained for the duration of exercise. Each of the subjects reached ca 50% of their predicted maximum heart rate, apart from two, who exercised at
SKELETAL MUSCLE K+ LOSS DURING EXERCISE 645 Table 2. Measured counts during control period (1), weight lifted during exercise and normalized counts from periods 1, 2 and 3 Weight Subject Control lifted Period I Period 2 Period 3 number (counts min-') (kg) (%) (%) (%) 802 844 5 00 100 00 87-98 90-63 835 648 8-80 100 00 96-08 94 39 836 615 7 50 100 00 99 30 97-15 837 646 13-8 100 00 96-60 98 56 843 648 13-8 100 00 95 55 99 24 853 639 6-00 100 00 102-33 100-59 860 620 5-00 100 00 102 03 101-76 865 583 6-00 100 00 95 09 86-34 876 777 15-0 100 00 103-17 96 52 887 543 13-8 100 00 96-46 97-08 909 441 17-0 100 00 97 39 101-36 911 506 5 00 100 00 98-79 103-20 Mean 100 00 97 57 97-24 S.E.M. 1-18 1-41 60% (subject 835) and 70% (subject 802) of their predicted maximum heart rate. The subject who exercised at 70% of his maximum predicted heart rate produced the largest immediate decrease in muscle K+ following exercise. All of the subjects reached perceived exertion response (PER) scores of 16 or greater (very hard) on the Borg scale (Fig. 4). Pre- and post-exercise counts. The mean values for the normalized counts, for each subject, in each period are shown in Table 2. On average, there was a small decrease in counts following exercise of 2 4 % for period 2 and 2-8 % for period 3, compared with the pre-exercise period (P < 0 05, one-tailed Student's t test). There was no significant difference between the average counts for periods 2 and 3. DISCUSSION Accuracy of 43K+ methodfor measuring K+ loss A number of possible sources of error may be present in this method. An even distribution of 43K+ around the subject's body prior to the experiment is important. The period of 24 h between injection and the study is certainly sufficient to equilibrate 43K+ around most of the body stores (Jasani & Edmonds, 1971). The radioactive decay of 43K+ during the experiment is accounted for by correcting the non-background component of the count rate observed for physical decay from the start of the experiment using the half-life of 43K+. In addition to radioactive decay, there would also be a biological loss of 43K+. Typically, the body turns over 80 mmol of K+ per day out of a total body content of 3600 mmol of K+ (International Committee for Radiation Protection, 1975). This gives a turnover rate of ca 2 % per day, and so should not be very important during our experiments. This has been confirmed by performing whole body counts of 43K+ before and after the experiment on ten of the subjects. The average change in whole body 43K+ was 0-02 + 0-66 % (mean + S.E.M.). The attachment of the collimator to the detector reduced the count rate to ca 4% of its original value, but there remains a proportion of counts detected which do not originate
646 M. S. QAYYUM AND OTHERS from the subject's left thigh. The two sets of experiments using the phantom segments both suggest that between 20 and 24% of the counts are due to activity from sources other than the thigh segment under study. If 24% of the counts arise from outside the thigh, then the percentage loss of K+ from the thigh (percentage difference between periods 1 and 2) may be corrected from 2-43 + 1-18 % to 3-20+ 1-55% (means+s.e.m.). There will also be counts within the collimated field of view that do not arise from the quadriceps femoris muscle. In order to estimate how significant these may be, a ray diagram was drawn on a cross-section of the mid-thigh region (Carter, Morehead, Wolpert, Hammerschlag, Griffiths & Kahn, 1977). Almost all the muscle within the field of view was quadriceps femoris, although a part of sartorius was also included. The other structures within the field were skin, fat and blood; the field of view excluded the femur. Using reference values for the K+ content of skin, fat and blood (International Committee for Radiation Protection, 1975), a value of 6% was calculated for the 43K+ counts arising from non-muscle sources. This value suggests that counts from non-muscle sources within the collimated field of view are rather unimportant, and will not have a large effect on the estimate of K+ loss. One further potential problem with the method is that intracellular and extracellular volume changes associated with exercise might affect the amount of 43K+ within the collimated field of view. For exercise of similar intensity and duration to that employed in this study, it is unclear whether these volumes increase (Sj0gaard, 1986) or remain unchanged (Ahlborg et al. 1967). Comparison of results with other studies of similar work rate and exercise duration The single-leg extension exercise used in this study was devised to match closely that used by Sj0gaard (1986). The duration of 2 h, the work rate and the form of exercise, were similar in each study. Sj0gaard (1986) used two different techniques for estimating K' loss from the quadriceps muscle. Measurements of AV A[K+] suggested a 60 mmol loss of K+. Using Sj0gaard's figure of 3 kg for the wet weight of the quadriceps, and a value of 92-2 mmol (kg wet weight)-' for the amount of K+ in muscle (Lentner, 1981), the 60 mmol loss of K+ is equivalent to a 22% decrease in muscle content. Data from muscle biopsy before and after exercise yield a substantially smaller estimate of K+ loss of 11 %. It seems more likely that the biopsy result will be closer to the true K+ loss, since the arteriovenous difference is very small, and any errors in measurement can produce very large variations in calculated K+ loss. Indeed, the AV A[K+] shown by Sj0gaard (1986) for the subject at rest of 0.1 mm, which is presumably measurement error, would result in about half the K+ loss recorded during exercise. In addition, it is possible that there may be small shifts in K+ concentration between the red cell and plasma as the blood traverses the muscle capillary which could cause artifactual losses or gains of K+ from the exercising muscle. Other studies of K+ loss during sustained exercise have used cycle ergometery as the (1967) reported an average fall in muscle K+ content of means of exercise. Ahlborg et al. 444% in nine subjects who undertook cycle ergometery at an average work rate of 116 W for an average duration of 2 h. This figure matches that of our current study more closely. Sahlin & Broberg (1989) studied cycle exercise at 67 % of maximum 02 consumption rate (V0 max) for 60-70 min in eight subjects. Using an arteriovenous difference technique, Sahlin & Broberg (1989) estimated the K+ loss from the whole exercising leg as 22 mmol. Using a figure of 2 kg for the dry weight of muscle per leg and assuming an even loss of K+ throughout this muscle, Sahlin & Broberg (1989) calculated a 2-3 % decrease in K+ content. This figure is probably an underestimate in the sense that it assumes an even loss
SKELETAL MUSCLE K+ LOSS DURING EXERCISE of K+ from all muscles within the leg. If the K+ loss was solely from the quadriceps (presumably an overestimate) then the figure would become 8 %. Comparison with studies of exercise of shorter duration Skeletal muscle biopsy has been used to study K+ changes following short periods of intense, dynamic exercise. These studies have shown a decrease of muscle K+ content of between 58-10%. (Bergstrom et al. 1971; Sahlin, Alvestrand, Brandt & Hultman, 1978; Sj0gaard, 1986). Bergstrom et al. (1971) used cycle ergometer exercise at 163 W to exhaustion (17-20 min). Muscle biopsy of the vastus lateralis taken immediately following exercise showed a decrease of 8 % in muscle K+ content. Sahlin et al. (1978) used cycle ergometer exercise at a mean work load of 180 W for 5 min followed by a mean work rate of 260 W to exhaustion (5-6 min). Muscle biopsy of the vastus lateralis taken within 1 min of exercise showed a 5-8 % decrease in muscle K+ content. Sj0gaard (1986) also investigated intense single-leg extension exercise of short duration. Exercise was performed at 55 % of knee extensor V02 max (Andersen, Adams, Sj0gaard, Thorboe & Saltin, 1985), for 10 min and at 100% knee extensor V02, max to exhaustion (6-7 min). Biopsy of the vastus lateralis muscle following exercise showed a decrease of muscle K+ content of 10%. The kinetics of K+ loss and re-uptake from muscle At the onset of exercise there is a rapid rise in plasma [K+] and at the end of exercise this concentration falls abruptly (Laurell & Pernow, 1966). For single-leg extension exercise of a similar intensity to that used in this paper, Sj0gaard (1986) reports a rise in arterial [K+] of ca 0 5 mm. Assuming this is evenly distributed through 15 1 of extracellular fluid, this is equivalent to a loss of 7-5 mmol from the exercising muscle. For a quadriceps muscle of 3 kg wet weight, this is ca 3 % of the total muscle K+ content. At the cessation of exercise, the plasma K+ falls rapidly to control levels, presumably because of reuptake into the previously exercising muscle (Rolett, Strange, Sj0gaard, Kiens & Saltin, 1990), consequently this amount of K+ is returned to the muscle rapidly. Our experiments would not be expected to register this K+ loss because of the delay after the cessation of exercise in obtaining counts. Once the [K+] has risen at the start of exercise, it remains relatively stable during the course of exercise. During this period, any on-going loss of K+ from the muscle must be matched by K+ uptake in other tissues or excretion by the kidneys. Indeed, it may be the case that the rate of loss of K+ from exercising muscle in this period may be determined in part by the rate of uptake of K+ by other tissues in response to the rise in [K+]. The current study gives a measure of how much K+may have been taken up by the other resting tissues. Including the time between each 1 min count when no counts are being measured, the study gives an average period of 75 min between the two measures of K+ after exercise. Interestingly, no increase in count rate was detected in the second period. This suggests that the process of recovering K+which has been taken up by other tissues during exercise may be rather slow. This is in marked contrast to the rapid uptake of K+ remaining in the extracellular space. In conclusion, we find that a 2 h period of single-leg extension exercise is associated with approximately a 3-2+1-55% fall in K+ content of the quadriceps muscle. This value is rather lower than reported by Sj0gaard (1986) who used a similar exercise protocol, but is broadly in line with results from other studies of sustained leg exercise. The results also suggest that, apart from a rapid reuptake of K+ from the interstitial space at the cessation of exercise, the recovery of intracellular muscle K+ content is slow. 647
648 M. S. QAYYUM AND OTHERS This work was supported by a grant from the British Heart Foundation. We thank Dr B. J. Shepstone for his help in administration of radioactive isotopes. REFERENCES AHLBORG, B., BERGSTROM, J., EKELUND, L.-G. & HULTMAN, E. (1967). Muscle glycogen and muscle electrolytes during prolonged physical exercise. Acta Physiologica Scandinavica 70, 129-142. ANDERSEN, P., ADAMS, R. P., SJOGAARD, G., THORBOE, A. & SALTIN, B. (1985). Dynamic knee extension as model for study of isolated exercising muscle in humans. Journal ofapplied Physiology 59, 1647-1653. ASTRAND, P.-O. & RODAHL, K. (1986). Textbook of Work Physiology. McGraw-Hill, New York. BERGSTROM, J., GUARNERI, G. & HULTMAN, E. (1971). Carbohydrate metabolism and electrolyte changes in human muscle tissue during heavy work. Journal of Applied Physiology 30, 122-125. BEWLEY, D. K. (1988). Anthropomorphic models for checking the calibration of whole body counters and activation analysis systems. Physics in Medicine and Biology 33, 805-813. BORG, G. A. V. (1982). Psychophysical bases of perceived exertion. Medicine and Science in Sports and Exercise 14, 377-381. BUSH, F. (1946). Energy absorption in radium therapy. British Journal of Radiology 19, 14-21. CARTER, B. L., MOREHEAD, J., WOLPERT, S. M., HAMMERSCHLAG, S. B., GRIFFITHS, H. J. & KAHN, P. C. (1977). Cross-sectional Anatomy, Computed Tomography and Ultrasound Correlation. Appleton-Century-Crofts, New York. CLAUSEN, T. (1990). Significance of Na+-K+ pump regulation in skeletal muscle. News in Physiological Sciences 5, 148-151. EDWARDS, R. H. T., MAUNDER, C., JONES, D. A. & BATRA, G. J. (1975). Needle biopsy for muscle chemistry. Lancet ii, 736-740. INTERNATIONAL COMMITTEE FOR RADIATION PROTECTION. (1975). Reference Man: Anatomical, Physiological and Metabolic Characteristics, Publication 23. Pergamon Press, Oxford. JASANI, B. M. & EDMONDS, C. J. (1971). Kinetics of potassium distribution in man using isotope dilution and whole-body counting. Metabolism 20, 1099-1106. JOHNSON, M. A., POLGAR, J., WEIGHTMAN, D. & APPLETON, D. (1973). Data on the distribution of fiber types in thirty-six human muscles. An autopsy study. Journal of Neurological Science 18, 111-129. KILBURN, K. H. (1966). Muscular origin of elevated plasma potassium during exercise. Journal of Applied Physiology 21, 675-678. LAURELL, H. & PERNOW, B. (1966). Effect of exercise on plasma potassium in man. Acta Physiologica Scandinavica 66, 241-242. LENTNER, C. (1981). Geigy Scientific Tables, vol. 1. Ciba-Geigy, Basel. QAYYUM, M. S., FREEMANTLE, C. A. J., CAREY, C. J., PAGE, B. C., PATERSON, D. J. & ROBBINS, P. A. (1993). Changes in skeletal muscle K' content during prolonged exercise in humans. Journal of Physiology 459, 155P. ROLETT, E. L., STRANGE, S., SJOGAARD, G., KIENS, B. & SALTIN, B. (1990). /32-Adrenergic stimulation does not prevent potassium loss from exercising quadriceps muscle. American Journal of Physiology 258, RI 192-1200. SAHLIN, K., ALVESTRAND, A., BRANDT, R. & HULTMAN, E. (1978). Intracellular ph and bicarbonate concentration in human muscle during recovery from exercise. Journal of Applied Physiology 45, 474-480. SAHLIN, K. & BROBERG, S. (1989). Release of K' from muscle during prolonged dynamic exercise. Acta Physiologica Scandinavica 136, 293-294. SJOGAARD, G. (1986). Water and electrolyte fluxes during exercise and their relation to muscle fatigue. Acta Physiologica Scandinavica 128, supplement 556, 129-136. SPINKS, T. J., BEWLEY, D. K., RANICAR, A. S. 0. & JOPLIN, G. F. (1977). Measurement of total body calcium in bone disease. Journal of Radioanalytical Chemistry 37, 345-355.