Exercise-Induced Rise in Arterial Potassium in Patients With Chronic Heart Failure*

Transcription

1 Exercise-Induced Rise in Arterial Potassium in Patients With Chronic Heart Failure* Relation to Excessive Exercise Ventilation Yasuhiko Tanabe, MD; Masahiro Ito, MD; Yukio Hosaka, MD; Eiichi Ito, MD; Kaoru Suzuki, MD; and Minoru Takahashi, MD Objectives: In patients with chronic heart failure (), exercise is frequently associated with skeletal muscle fatigue and breathlessness due to heightened ventilatory response. The exerciseinduced rise in potassium, which is released from the exercising skeletal muscle, has been implicated in ventilatory control during exercise. The aim of the present study was to determine whether the exercise-induced rise in arterial potassium is altered in patients with and to examine the relationship between increased exercise ventilation and exercise-induced hyperkalemia in patients with. Methods and results: We evaluated 88 patients with (25 patients were in class I, 35 in class II, and 28 in class III according to the New York Heart Association functional classification) and 14 normal subjects. Subjects performed symptom-limited ergometer exercise while expired gas, arterial blood gas, and arterial potassium were analyzed. The increases in ventilation ( V E), effective alveolar ventilation ( V A), and carbon dioxide output ( V CO2 ) from rest to peak exercise decreased as the severity of advanced. The ratio of V E to V CO2 was significantly elevated in class III patients, although there was no difference in the ratio of V Ato V CO2 among the four groups. Rest and exercise arterial PCO 2 did not differ among the four groups and was controlled within the normal range. The increase in arterial potassium ( K ) from rest to peak exercise was markedly reduced as the severity of advanced: (mean SD) mmol/l in normal subjects; mmol/l in class I patients; mmol/l in class II patients; and mmol/l in class III patients. The ratios of V Aor V CO2 to K were not different among the four groups. The ratio of V E to K, however, was significantly greater in patients in class III than in normal subjects or patients in class I or II. Conclusions: The K from rest to peak exercise was markedly reduced as the severity of advanced. The increased exercise ventilation due to increased physiologic dead space in severe was not accompanied by the corresponding augmentation of exercise-induced hyperkalemia. Exercise-induced hyperkalemia does not contribute to the increased ventilatory drive to keep normal arterial PCO 2 during exercise in the presence of increased physiologic dead space in severe. (CHEST 1999; 116:88 96) Key words: anaerobic threshold; exercise test; heart failure; muscles; oxygen uptake; potassium; respiration Abbreviations: AT anaerobic threshold; chronic heart failure; K arterial potassium; K increase in arterial potassium; NYHA New York Heart Association; V A effective alveolar ventilation; V A increase in effective alveolar ventilation; V co 2 carbon dioxide output; V co 2 increase in carbon dioxide output; V e ventilation; V e increase in ventilation; V o 2 oxygen uptake; WR work rate; WR increase in work rate Early muscle fatigue and increased breathlessness are major symptoms in patients with chronic heart failure (). 1 6 An inadequate blood flow to *From the Department of Internal Medicine (Drs. Tanabe, M. Ito, Hosaka, E. Ito, and Suzuki), Niigata Prefectural Shibata Hospital, Niigata, Japan; and the First Department of Internal Medicine (Dr. Takahashi), Niigata University School of Medicine, Niigata, Japan. Manuscript received August 25, 1998; revision accepted February 18, Correspondence to: Yasuhiko Tanabe, MD, Niigata Prefectural Shibata Hospital, Ohtemachi , Shibata City, Niigata, Japan the working skeletal muscle, abnormalities of skeletal muscle metabolism, and reduced skeletal muscle mass have been recognized in patients with. 1 4 These abnormalities contribute to early muscle fatigue and exercise intolerance. In addition, patients with show a heightened ventilatory response to exercise, which is an important factor causing breathlessness during daily activities. 5,6 It is well known that exercise is associated with hyperkalemia Potassium is released into the circulation from the exercising skeletal muscle. 13,14 88 Clinical Investigations

2 Recently, the exercise-induced rise in arterial potassium has been implicated in ventilatory control during exercise However, the exercise-induced rise in potassium in patients with, especially in patients with severe who have abnormal skeletal muscle function, has not been fully evaluated. Also, the relationship between heightened ventilatory drive in patients with severe and the exercise-induced rise in potassium has not been evaluated. We hypothesized that if potassium is the major humoral factor that stimulates breathing during exercise, the exercise-induced rise in potassium increases as the ventilation during exercise increases in patients with severe. To determine whether the exercise-induced rise in arterial potassium is altered in patients with, and to examine the relationship between increased exercise ventilation and the exercise-induced rise in potassium in severe, we measured ventilation (V e), effective alveolar ventilation (V A), arterial blood gases, and arterial potassium concentration (K ) during incremental exercise in patients with and normal subjects. Subjects Materials and Methods Eighty-eight patients with (49 men and 39 women), ranging in age from 18 to 73 years (mean age, 53 years), and 14 normal volunteer subjects (mean age, 49 years) were included in this study (Table 1). All patients had a history of at least 3 months in duration and had stable symptoms on the same medications at least 1 month before the evaluation of exercise tolerance in the present study. The etiologic basis of their included idiopathic dilated cardiomyopathy (36 patients); dilated cardiomyopathy after valve replacement (4 patients); old myocardial infarction (18 patients); mitral regurgitation (8 patients); aortic regurgitation (5 patients); and mitral stenosis (17 patients). Left ventricular ejection fraction determined by contrast ventriculography or echocardiography was %. All patients, except those with old myocardial infarction, had normal or minimally sclerotic coronary arteries. No patients with old myocardial infarction showed scintigraphic evidence of exerciseinduced myocardial ischemia. Patients were subdivided into three groups according to the New York Heart Association (NYHA) functional classification: 25 patients were in class I, 35 in class II, and 28 in class III. The clinical characteristics of the subjects and their medications are summarized in Tables 1 3. We did not include patients who received potassium chloride supplements, beta-blockers, or beta-stimulants. Exercise Tests Written informed consent was obtained from each subject before the study. All subjects performed a preliminary exercise test with respiratory gas analysis 2 to 14 days before the study to familiarize them with the procedure. All patients and normal subjects underwent maximal symptom-limited exercise test using an electronically braked ergometer (Examiner 400; Lode BV; Groningen, Netherlands). After a 3-min rest period on the ergometer, exercise began with a 3-min warm-up period at 10 W, followed by a continuous ramp protocol corresponding to increments of 10 to 20 W/min (1 W every 6sor1Wevery 3 s) until the subjects could no longer continue. The end point of the exercise tests was leg fatigue with various degrees of breathlessness in all subjects. The ECG was monitored throughout exercise in all subjects, and the 12-lead ECG was recorded at 1-min intervals in all patients. Cuff BP was measured at 1-min intervals. Respiratory gas analysis was performed on a breath-by-breath basis using an Aero Monitor AE-280 (Minato Medical Science; Osaka, Japan). This system consists of a hot-wire spirometer, a zirconium solid O 2 analyzer, and an infrared CO 2 analyzer, and it introduces only a small volume of dead space into the ventilatory circuit. The apparatus was calibrated before each study was performed. Data were processed using an on-line computer system, and the following parameters were calculated: V e, oxygen Table 1 Patient Characteristics and Hemodynamic Data at Rest and Peak Exercise* Variables Normal Subjects (n 14) (n 35) I (n 28) Age, yr Weight, kg Height, cm BMI, kg/m HR at rest, bpm HR at peak exercise, bpm SBP at rest, mm Hg SBP at peak exercise, mm Hg DBP at rest, mm Hg DBP at peak exercise, mm Hg *Values are given as mean SD. bpm beats per minute; BMI body mass index; DBP diastolic blood pressure; HR heart rate; SBP systolic blood pressure. p 0.05 compared with normal subjects. p 0.05 compared with class I. p 0.05 compared with class II. CHEST / 116 / 1/ JULY,

3 Table 2 Pulmonary Function at Rest and Breathing Reserve at Peak Exercise* Variables (n 18) I (n 23) VC, L VC, % predicted FEV 1, L FEV 1 /FEV, % BR, L MVV, L BR/MVV, % *Values given as mean SD. BR breathing reserve; MVV maximum voluntary ventilation; VC vital capacity. p 0.05 compared with class I. p 0.05 compared with class II. Table 3 Medication Taken by the Subjects* Medications (n 35) I (n 28) Digoxin 7 (28) 26 (74) 23 (82) Furosemide 9 (36) 30 (86) 26 (93) Spironolactone 4 (16) 10 (29) 13 (46) Vasodilators 19 (76) 27 (77) 23 (82) *Values given as No. of subjects (%). Vasodilators include angiotensin-converting enzyme inhibitors, calcium antagonists, isosorbide dinitrate, and pimobendan. the K from rest to peak exercise and peak V o 2, and between the K and the V e, V A, or V co 2 during exercise in individual subjects. A value of p 0.05 was considered significant. uptake (V o 2 ), carbon dioxide output (V co 2 ), end-tidal Po 2, end-tidal Pco 2, and the respiratory exchange ratio (V co 2 /V o 2 ). Before exercise, a cannula was introduced into a radial or brachial artery. Arterial blood was sampled at rest, during the warm-up period, and at 1-min intervals throughout the exercise. Samples were analyzed for arterial K with an ion selective electrode analyzer (CIM-104A; Shimadzu; Kyoto, Japan). Arterial ph, arterial Po 2, arterial Pco 2, and calculated HCO 3 were measured with a blood gas analyzer (ABL-2; Radiometer Medical A/S; Copenhagen, Denmark). V A was calculated from the alveolar gas equation (V A V co 2 863/arterial Pco 2 ). The increases in V e ( V e), V A ( V A), V co 2 ( V co 2 ), and K ( K ) from rest to peak exercise were determined. Exercise capacity was assessed on the basis of peak V o 2 and anaerobic threshold (AT). Peak V o 2 was determined as the average of values obtained during the final 15 s of exercise. AT was determined by two experienced reviewers using the V-slope method, as well as the following conventional criteria: (1) the point at which the ratio of V e to V o 2 increases after being stable or decreasing while the ratio of V e to V co 2 remains constant or decreasing; (2) the point at which an increase in end-tidal Po 2 occurs without a decrease in end-tidal Pco 2 ; and (3) the point at which a stable or slowly rising respiratory exchange ratio begins to rise more steeply The slope of the K with respect to the increase in work rate ( K / WR) from rest to peak exercise was calculated as: [(peak K rest K )/(peak WR rest WR)]. The K / WR was calculated below and above the AT as follows: K / WR below the AT [(K at the last blood sampling before AT rest K )/(WR at the last blood sampling before AT WR at rest)]. K / WR above the AT [(K at peak exercise K at the first blood sampling after AT)/(WR at peak exercise WR at the first blood sampling after AT)] The K / WR below and above the AT was calculated only in subjects whose work rate at the AT was higher than 30 W, because blood sampling was performed only a few times before the AT in patients whose work rate at the AT was 30 W. Statistical Analysis All data were expressed as mean SD. Multiple comparisons among groups were performed by one-way analysis of variance combined with Scheffe s test. Comparisons of K / WR below and above the AT was performed by paired t tests. Linear regression analysis was used to assess the relationship between Exercise Tests Results Based on the results of preliminary exercise tests, all 14 normal subjects and 36 patients (20 patients in NYHA class I, 11 in class II, and 5 in class III) exercised using a ramp protocol of 20 W/min. The remaining 52 patients (5 patients in class I, 24 in class II, and 23 in class III) exercised using a ramp protocol of 10 W/min. Pulmonary function tests at rest and breathing reserve at peak exercise are shown in Table 2. Heart rate and BP at rest and at peak exercise are shown in Table 1, and results of respiratory gas analysis at rest and during exercise are shown in Table 4. Exercise duration was only slightly shorter in patients in class III. The peak WR, peak V o 2, AT, V e, V A, and V co 2 from rest to peak exercise decreased as the severity of advanced (Table 4). The V e and the V A at rest and during exercise are shown in Figure 1. The V e and V A at the matched submaximal work rate were significantly greater in class III patients (Fig 1). The ratio of V e to V co 2 ( V e/ V co 2 ) from rest to peak exercise was significantly elevated in patients in class III compared with normal subjects or with class I or II patients, while the ratio of V A to V co 2 ( V A/ V co 2 ) from rest to peak exercise did not differ among the four groups (Fig 2). Table 5 shows changes in arterial Po 2, arterial Pco 2, and ph at rest and during exercise. The arterial Po 2 and arterial Pco 2 at rest and during exercise were not different among the four groups and were controlled within the normal range. K Concentration and Ventilation The K concentration at rest did not differ among the four groups (Table 4). The increase in arterial potassium from the resting value is plotted against 90 Clinical Investigations

4 Table 4 Results of Exercise Tests* Measurements Normal Subjects (n 14) (n 35) I (n 28) Exercise duration, min Peak WR, W Peak V o 2, ml/kg/min Maximal O 2 pulse, ml/beat Work rate at AT, W AT, ml/kg/min K at rest, mmol/l K, mmol/l RER at rest RER at peak exercise HCO 3 at rest, mmol/l HCO 3, mmol/l V e, ml/kg/min V A, ml/kg/min V co 2, ml/kg/min *Values given as mean SD. K, V e, V A, and V co 2 increase in K,V e, V A, and V co 2 from rest to peak exercise; HCO 3 change in HCO 3 from rest to peak exercise; RER respiratory exchange ratio. p 0.05 compared with normal subjects. p 0.05 compared with class I. p 0.05 compared with class II. WR in Figure 3. The K at the matched submaximal WR tended to be greater in patients in NYHA class II or III, but it did not reach statistical significance. The K from rest to peak exercise was progressively reduced as the severity of advanced (Table 4). The K from rest to peak exercise was significantly correlated with peak V o 2 (r 0.75; p 0.001; Fig 4). The ratio of K / WR (mmol/l/10 W) from rest to peak exercise did not differ among the four groups ( mmol/l/10 W in normal subjects; mmol/l/10 W in NYHA class I patients; mmol/l/10 W in class II; and mmol/l/10 W in class III). The K / WR below the AT could not be determined in 1 of 25 class I patients, in 18 of 35 class II patients, and in 23 of 28 class III patients. Therefore, analysis of the data of K / WR below and above the AT could not be done in NYHA class III patients. The K / WR below the AT did not differ by group ( mmol/l/10 W in normal subjects; mmol/l/10 W in class I patients; and mmol/ L/10 W in class II). The K / WR above the AT also did not differ by group ( mmol/ L/10 W in normal subjects; mmol/ L/10 W in class I; and mmol/l/10 W in class II). In each group, the K / WR was significantly greater above the AT than below the AT (p 0.001). In individual subjects, there were strong correlations between the K and the V e (correlation coefficient values: in normal subjects, and in patients with ), between the K and the V A (correlation coefficient values: in normal subjects, and in patients with ), or between the K and the V co 2 (correlation coefficient values: in normal subjects, and in patients with ). The ratio of V A to K from rest to peak exercise or the ratio of V co 2 to K from rest to peak exercise did not differ between normal subjects and patients in NYHA classes I, II, or III. The ratio of V e to K from rest to peak exercise, however, was significantly greater in patients in NYHA class III than in normal subjects or in patients in NYHA class I or II (Fig 5). Discussion The present study revealed that the increase in potassium from rest to peak exercise was markedly reduced in severe and that the increased exercise ventilation due to enlarged physiologic dead space in class III was not accompanied by a corresponding augmentation of exercise-induced hyperkalemia. Reduced Exercise-Induced Rise in Potassium in Patients With It is well known that exercise is associated with hyperkalemia Exercise-induced hyperkalemia is CHEST / 116 / 1/ JULY,

5 Figure 2. Bar graphs showing the ratios of V e (top, A) or V A (bottom, B) to V co 2 in normal subjects (N) and in NYHA class I, II, and III patients. V e is the increase in ventilation from rest to peak exercise (ml/kg/min); V A is the increase in V A from rest to peak exercise (ml/kg/min); V co 2 is the increase in V co 2 from rest to peak exercise (ml/kg/min). * p Figure 1. Plots of ventilation (top, A) and effective alveolar ventilation (bottom, B) at rest and during exercise in normal subjects and in NYHA class I, II, and III patients. Comparisons between groups were made at matched submaximal WR during submaximal exercise. * p 0.05 compared with normal subjects; p 0.05 compared with NYHA class I; 0.05 compared with NYHA class II. to a large extent the result of potassium release from the exercising skeletal muscle. 13,14 Recent research has demonstrated that patients with exhibit significant functional impairment of skeletal muscle. 2 4 The exercise-induced rise in potassium in patients with, however, has not been fully examined. Barlow et al 18 showed that the exercise-induced rise in potassium and ventilation are greater at matched submaximal WR in patients with than in subjects with normal left ventricular function. The present study also showed that the exercise-induced rise in potassium at matched submaximal WR tended to be greater in severe, but it did not reach statistical significance. We showed that the ratio of K / WR was significantly greater above the AT than below the AT. Therefore, patients with a low AT value would exhibit a greater rise in potassium at the near-maximal WR, which is below the AT level of subjects with good exercise capacity. Decreased muscle mass with increased loading of each muscle 92 Clinical Investigations

6 Table 5 Results of Arterial Blood Gas Tests at Rest and During Exercise* Tests Normal Subjects (n 14) (n 35) I (n 28) Arterial Pco 2,mmHg Rest % load % load Peak exercise Arterial Po 2,mmHg Rest % load % load Peak exercise ph Rest % load % load Peak exercise *Values given as mean SD. p 0.05 compared with normal subjects. p 0.05 compared with class I. fiber at a given WR or reduced reuptake of potassium through decreased concentration of sodiumpotassium pumps in skeletal muscle of patients with may also be related to the increased hyperkalemia at matched submaximal WR. The findings of the present study that the exercise-induced rise in K from rest to peak exercise decreased progressively as the severity of advanced have not been reported previously, to our knowledge. The K from rest to peak exercise was closely correlated with peak V o 2. The amount of potassium released into the circulation from the muscle is dependent on the magnitude of muscle contraction, as reflected by absolute workload. 19 Our results also showed that the K was dependent on the workload. Although our subjects attained a max- Figure 3. Plots of the exercise-induced rise in K from the resting value in normal subjects and in NYHA class I, II, and III patients. Comparisons between groups were made at matched WR during submaximal exercise. No significant differences were noted. Figure 4. Scatterplots show relationship between K from rest to peak exercise and peak V o 2 in patients with and normal subjects. CHEST / 116 / 1/ JULY,

7 imal or near-maximal exercise level as indicated by severe leg fatigue, the decrease in HCO 3 from rest to peak exercise, which reflects the amount of lactate production, 20,21 was significantly lower in patients in NYHA class III. Previous studies have also shown that the lactate level at peak exercise is significantly lower in patients with severe. 1,22 Significant intrinsic alteration in skeletal muscle, such as reduced aerobic enzyme activities and muscle atrophy rather than the magnitude of lactic acidosis may lead to early fatigue in patients with. 2 4 Significantly decreased WR at peak exercise because of these skeletal muscle abnormalities and skeletal muscle hypoperfusion may be associated with the markedly reduced rise in potassium at peak exercise through the decreased total number of action potentials of the contracting muscle. Figure 5. Bar graphs showing the ratios of V e (top, A), V A (middle, B), or V co 2 (bottom, C)to K in normal subjects (N) and in NYHA class I, II, and III patients. K is the increase in K from rest to peak exercise (mmol/l). See Figure 2 legend for other abbreviations. * p Increased Ventilation and Potassium in Severe The excessive exercise ventilation recognized in patients with severe, which may be an important factor in exertional dyspnea, results from increased pulmonary physiologic dead space. 5,6 Sullivan et al 5 showed that the ratio of ventilation to CO 2 production is elevated in patients with who have large physiologic dead space, while rest and exercise arterial Pco 2 regulation is normal. The present study also showed that V e/ V co 2 was significantly elevated in patients in NYHA class III, while V A/ V co 2 and arterial Pco 2 did not differ between normal subjects and patients with. Therefore, the neurohumoral ventilatory control mechanisms are intact even in patients with. The increased ventilation during exercise in severe is primarily the consequence of a stimulus to regulate arterial Pco 2 in the presence of increased physiologic dead space. An exercise-induced rise in K has been implicated in the ventilatory control during exercise Several studies have shown a strong correlation between K concentration and ventilation during exercise and recovery Although lactic acid may contribute to exercise hyperpnea through carotid body chemoreceptors, 23 patients with McArdle s disease who cannot produce lactic acid still hyperventilate during exercise. 10 There is also a linear relationship between ventilation and K concentration during and after exercise in patients with McArdle s disease. 10 When two successive exercise tests are performed in normal subjects, there is a close relationship between ventilation and potassium, ph, and lactic acid during the first exercise test. In the second exercise test starting with exercise-induced lactic acidosis, ventilation is related only to K concentration, but is unrelated to ph or lactic acid. 11,12 In an animal model, carotid body chemoreceptor discharge and ventilation increase when potassium is raised to the levels similar to those seen during exercise. 24,25 Although there is no direct proof that potassium causes an increase in ventilation during 94 Clinical Investigations

8 exercise, all of these studies suggest that the potassium contributes to the control of exercise ventilation. We hypothesized that if potassium acts as the major humoral factor stimulating ventilation during exercise, the exercise-induced rise in potassium would be augmented as the ventilation during exercise increased in patients with severe who had increased physiologic dead space. The results of the present study showing a strong correlation between K and V co 2, V e, or V A during exercise are not inconsistent with previous reports suggesting that potassium acts as a ventilatory stimulus. However, the increased exercise ventilation due to increased physiologic dead space in patients with severe was not accompanied by a corresponding augmentation of exercise-induced hyperkalemia. The exercise-induced rise in potassium was markedly reduced in patients with severe. The exerciseinduced rise in potassium may not contribute to the increased ventilatory drive to keep normal arterial Pco 2 during exercise in the presence of increased physiologic dead space. Limitations This study included patients with various etiologies of. However, in all subjects, the end point of the exercise was leg fatigue with various degrees of breathlessness. We did not include patients whose exercise was limited by angina, respiratory failure, or other medical problems. Adequately elevated respiratory exchange ratios in each subject suggested that the determinant of exercise tolerance was nearly the same in our subjects. Therefore, the potassium kinetics may not be influenced by the etiology of in the present study. The sensitivity of the ventilatory control system may be enhanced by hypoxia, catecholamines, or lactic acid. 7,26 However, arterial Po 2 during exercise did not differ between normal subjects and patients with. The plasma catecholamine response to exercise is not augmented in severe. 27 The decrease in HCO 3, which reflected lactic acid production, was not augmented in patients with severe. Therefore, hypoxia, catecholamines, and lactic acid do not contribute to our results. Cardiovascular drugs such as beta-blockers, betastimulants, digoxin, and diuretics may influence potassium homeostasis We did not include patients who received beta-blockers or beta-stimulants. In the present study, significantly more patients in NYHA class III took diuretics or digoxin. Diuretic therapy induces decreased concentration of sodiumpotassium pumps. 31 Digoxin inhibits sodium-potassium pumps. 32 Reuptake of potassium is dependent on the sodium-potassium pumps of the skeletal muscles. 14 Digoxin or diuretics may induce augmented exercise-induced hyperkalemia through reduced reuptake of potassium by skeletal muscle. Therefore, the tendency of greater rise in potassium at matched submaximal WR in severe may be related to the use of these drugs. However, the markedly smaller rise in potassium at peak exercise in patients with severe cannot be explained by the effect of these drugs. We did not intend to clarify the exact mechanism of ventilatory control during exercise; rather, we examined the K response to exercise in, as well as the relationship between increased exercise ventilation due to increased physiologic dead space and exercise-induced hyperkalemia in patients with. The role of potassium in the control of ventilation is only conjectural and not clearly proven. Therefore, our finding that the increased exercise ventilation in severe was not accompanied by a corresponding augmentation of exercise-induced hyperkalemia does little to further understanding of ventilatory control mechanisms during exercise. Exercise ventilation must be regulated redundantly by a combination of multiple neural and humoral factors. 23,33 Further studies are needed to clarify the exact ventilatory control mechanisms that maintain normal arterial Pco 2 during exercise. Conclusions The exercise-induced rise in potassium from rest to peak exercise was markedly reduced as the severity of advanced. The increased exercise ventilation due to increased physiologic dead space in patients with severe was not accompanied by a corresponding augmentation of exercise-induced hyperkalemia. The exercise-induced rise in potassium may not contribute to the increased ventilatory drive to keep arterial Pco 2 normal during exercise in the presence of increased physiologic dead space. References 1 Willson JR, Martin JL, Schwartz D, et al. Exercise intolerance in patients with chronic heart failure: role of impaired nutritive flow to skeletal muscle. Circulation 1984; 69: Sullivan MJ, Green HJ, Cobb FR. Altered skeletal muscle metabolic response to exercise in chronic heart failure: relation to skeletal muscle aerobic enzyme activity. Circulation 1991; 84: Drexler H, Riede U, Münzel T, et al. Alteration of skeletal muscle in chronic heart failure. Circulation 1992; 85: Mancini DM, Walter G, Reichek N, et al. Contribution of skeletal muscle atrophy to exercise intolerance and altered muscle metabolism in heart failure. Circulation 1992; 85: CHEST / 116 / 1/ JULY,

9 5 Sullivan MJ, Higginbotham MB, Cobb FR. Increased exercise ventilation in patients with chronic heart failure: intact ventilatory control despite hemodynamic and pulmonary abnormalities. Circulation 1988; 77: Buller NP, Poole-Wilson PA. Mechanism of the increased ventilatory response to exercise in patients with chronic heart failure. Br Heart J 1990; 63: Paterson DJ. Potassium and ventilation in exercise. J Appl Physiol 1992; 72: Paterson DJ, Robbins PA, Conway J. Changes in arterial plasma potassium and ventilation during exercise in man. Respir Physiol 1989; 78: Yoshida T, Chida M, Ichioka M, et al. Relationship between ventilation and arterial potassium concentration during incremental exercise and recovery. Eur J Appl Physiol 1990; 61: Paterson DJ, Friedland JS, Bascom DA, et al. Changes in arterial K and ventilation during exercise in normal subjects and subjects with McArdle s syndrome. J Physiol 1990; 429: Busse MW, Maassen N, Konrad H. Relation between plasma K and ventilation during incremental exercise after glycogen depletion and repletion in man. J Physiol 1991; 443: Busse MW, Scholz J, Saxler F, et al. Relationship between plasma potassium and ventilation during successive periods of exercise in men. Eur J Appl Physiol 1992; 64: Fenn WO. Loss of potassium in voluntary contraction. Am J Physiol 1937; 120: Sjøgaad G. Role of exercise-induced potassium fluxes underlying muscle fatigue: a brief review. Can J Physiol Pharmacol 1937; 69: Beaver WL, Wasserman K, Whipp BJ. A new method for detecting anaerobic threshold by gas exchange. J Appl Physiol 1986; 60: Matsumura N, Nishijima H, Kojima S, et al. Determination of anaerobic threshold for assessment of functional state in patients with chronic heart failure. Circulation 1983; 68: Wasserman K, Whipp BJ, Koyal SN, et al. Anaerobic threshold and respiratory gas exchange during exercise. J Appl Physiol 1973; 35: Barlow CW, Qayyum MS, Davey PP, et al. Effect of physical training on exercise-induced hyperkalemia in chronic heart failure: relation with ventilation and catecholamines. Circulation 1994; 89: Vøllestad NK, Hallén J, Sejersted OM. Effect of exercise intensity on potassium balance in muscle and blood in man. J Physiol 1994; 475: Wasserman K. Determinants and detection of anaerobic threshold and consequences of exercise above it. Circulation 1987; 76(Suppl VI):VI-29 VI Naimark A, Wasserman K, MacIlroy MB. Continuous measurement of ventilatory exchange ratio during exercise. J Appl Physiol 1964; 19: Roubin GS, Anderson SD, Shen WF, et al. Hemodynamic and metabolic basis of impaired exercise tolerance in patients with severe left ventricular dysfunction. J Am Coll Cardiol 1990; 15: Wasserman K, Whipp BJ, Casaburi R. Respiratory control during exercise. In: Fishman AP, Cherniack NS, Widdicombe JG, eds. Handbook of physiology. Section 3: The respiratory system, vol. 2. Bethesda, MD: American Physiological Society, 1986; Linton RAF, Band DM. The effect of potassium on carotid chemoreceptor activity and ventilation in the cat. Respir Physiol 1985; 59: Band DM, Linton RAF. The effect of potassium on carotid body chemoreceptor discharge in the anesthetized cat. J Physiol 1986; 381: Burger RE, Estavillo JA, Kumar P, et al. Effect of potassium, oxygen and carbon dioxide on the steady-state discharge of cat carotid body chemoreceptors. J Physiol 1988; 401: Francis GS, Goldsmith SR, Ziesche S, et al. Relative attenuation of sympathetic drive during exercise in patients with congestive heart failure. J Am Coll Cardiol 1985; 5: Paterson DJ, Nye PCG. The effect of beta adrenergic blockade on the carotid body response to hyperkalaemia in the cat. Respir Physiol 1988; 74: Williams ME, Rosa RM, Silva P, et al. Impairment of extrarenal potassium disposal by alpha-adrenergic stimulation. N Engl J Med 1984; 311: Williams ME, Gervino EV, Rosa RM, et al. Catecholamine modulation of rapid potassium shifts during exercise. N Engl J Med 1985; 312: Dørup I, Skajaa K, Clausen T, et al. Reduced concentrations of potassium, magnesium, and sodium-potassium pumps in human skeletal muscle during treatment with diuretics. BMJ 1988; 296: Smith TW, Haber E. Digitalis (first of four parts). N Engl J Med 1973; 289: Forster HV, Pan LG. Exercise hyperpnea: its characteristics and control. In: Crystal RG, West JB, eds. The lung: scientific foundations. New York, NY: Raven Press, 1991; Clinical Investigations