Kidney International, Vol. 63 (2003), pp. 283–290

Impaired K� regulation contributes to exercise limitation inend-stage renal failure

TERMBOON SANGKABUTRA, DAVID P. CRANKSHAW, CLAUDIA SCHNEIDER, STEVE F. FRASER,SIMON SOSTARIC, KIM MASON, CAROLINE M. BURGE, SANDFORD L. SKINNER,LAWRENCE P. MCMAHON, and MICHAEL J. MCKENNA

School of Human Movement, Recreation and Performance, Centre for Rehabilitation, Exercise and Sports Science,Victoria University of Technology, Melbourne; Departments of Surgery and Nephrology, The Royal Melbourne Hospital,Parkville; and Department of Physiology, and Anaesthetic Research and Education Unit, Department of Pharmacology,The University of Melbourne, Parkville, Victoria, Australia

Impaired K� regulation contributes to exercise limitation in ing incremental exercise testing, reaching only 50 to 60%end-stage renal failure. of healthy control levels [reviewed in 1]. It is likely that

Background. Patients with end-stage renal failure (ESRF) impairments in muscle O2 delivery, due to anemia [2–4],exhibit grossly impaired maximal exercise performance. Thisdepressed muscle blood flow [5] and capillary rarefactionstudy investigated whether K� regulation during exercise is[6], contribute to the depressed VO2 peak during exerciseimpaired in ESRF and whether this is related to reduced exer-

cise performance. in ESRF. However, improvements in VO2 peak withMethods. Nine stable hemodialysis patients and eight con- epoetin (EPO) are considerably less than expected from

trols (CON) performed incremental cycling exercise to voli- increases in hemoglobin concentration ([Hb]) [1, 2, 5],tional fatigue, with measurement of peak oxygen consumption

while normalization of [Hb] to 14 g · dL�1 with EPO in(VO2 peak). Arterial blood was sampled during and followingESRF does not restore normal VO2 peak values [7]. Inexercise and analyzed for plasma [K�] (PK).

Results. The VO2 peak was approximately 44% less in ESRF addition, muscle mitochondrial O2 utilization probablythan in CON (P � 0.001), whereas peak exercise PK was greater is reduced also, as indicated by abnormalities in mito-(7.23 � 0.38 vs. 6.23 � 0.14 mmol · L�1, respectively, P � 0.001). chondrial structure [8] and impaired post-contractileIn ESRF, the rate of rise in PK during exercise was twofold metabolic recovery [6]. Thus, numerous factors in thegreater (0.43 � 0.05 vs. 0.23 � 0.03 mmol · L�1 ·min�1, P �

oxygen transport and utilization chain may limit exercise0.005) and the ratio of rise in PK relative to work performedperformance in ESRF, but additional factors are alsowas 3.7-fold higher (90.1 � 13.5 vs. 24.7 � 3.3 nmol · L�1 · J�1,

P � 0.001). A strong inverse relationship was found between important.VO2 peak and the �PK · work�1 ratio (r � �0.80, N � 17, P � Muscle fatigue is consistently reported as a major or0.001). sole subjective factor causing exercise termination inConclusions. Patients with ESRF exhibit grossly impaired

ESRF [7, 9]. The mechanisms of fatigue are multifacto-extrarenal K� regulation during exercise, demonstrated by anrial, but include disturbances in muscle sodium ([Na�])excessive rise in PK relative to work performed. We further show

that K� regulation during exercise was correlated with aerobic and potassium ([K�]) concentrations and reduced maxi-exercise performance. These results suggest that disturbed K�

mal muscle Na�,K�-ATPase activity [10, 11]. In ESRF,regulation in ESRF contributes to early muscle fatigue during recent evidence is suggestive of defective muscle Na�

exercise, thus causing reduced exercise performance.and K� regulation, suggesting a link with increased mus-cle fatigue. In uremic rats, the skeletal muscle maximalNa�,K�-ATPase activity is compromised [12–15]. In hu-

Patients with end-stage renal failure (ESRF) on main- mans, erythrocytic Na�,K�-ATPase activity also is re-tenance hemodialysis typically demonstrate markedly re- duced in uremia [16, 17], with membrane depolarizationduced peak pulmonary oxygen uptake (VO2 peak) dur- evident in resting skeletal muscle, together with an ele-

vated or labile myocytic intracellular [Na�] [18–20]. Im-paired muscle Na�,K�-ATPase activity in ESRF, throughKey words: fatigue, extrarenal potassium, aerobic exercise, hemodialy-

sis, pulmonary oxygen uptake, epoetin, hemoglobin. excessive increases in myocytic intracellular [Na�] andextracellular [K�] during exercise, may result in early-Received for publication November 15, 2001onset muscle fatigue.and in revised form August 8, 2002

Accepted for publication September 6, 2002 Due to the complexity of measuring muscle ion con-centrations during exercise in humans [21], changes in 2003 by the International Society of Nephrology

283

Sangkabutra et al: K� and exercise limitation in renal failure284

plasma [K�] (PK) are often used as a marker of K� Incremental exercise testregulation [11, 22–24]. Few studies have measured PK All subjects reported to the laboratory in the after-during exercise in ESRF and these all suffer major meth- noon at least two hours after a light meal and confirmedodological limitations. Several studies have lacked com- that they had not performed strenuous physical activityparable controls [7, 9, 25, 26], which preclude identifica- in the previous 24 hours, nor taken caffeine or nicotinetion of any abnormal PK responses. Another study failed in the two hours prior to the exercise test. The test wasto detect the usual significant performance deficit in performed from 6 to 42 hours after their last dialysisESRF [27]. Methodological flaws in blood sampling in- session (27.3 � 13.1 h, mean � SD). Heart rate (HR) andvalidate findings in several studies. Collection of blood

rhythm were monitored using a 12-lead electrocardiogramsamples following cessation of exercise [28] underesti-

(Mortara, Boston, MA, USA). Subjects sat on an electri-mates the peak exercise PK due to the rapid post-exercisecally braked cycle ergometer (Lode, Groningen, Hol-decline that occurs in PK [22, 24]. Measurement of PK inland) for 15 minutes before commencing pedaling at 25 W,venous blood antecubital [27, 29] may underestimate PKwith progressive increments of 25 W each minute. Theduring low intensity leg exercise due to extraction of K�

pedal cadence was 70 rpm and the test was continuedby inactive forearm muscle [30]. In contrast, during highuntil volitional fatigue, defined as an inability to maintainintensity leg exercise, such as at the VO2 peak [27], ante-pedal cadence above 55 rpm. Mixed expired O2 and CO2cubital venous sampling may exaggerate PK due to K�

fractions were continuously measured during exercise byrelease from contracting forearm muscles [11]. The limi-rapidly responding analyzers (Ametek S-3A/II; Ametektations in each of these studies and the possible impor-CD-3A, Pittsburgh, PA, USA), with expired volume de-tance of defective muscle K� regulation in exercise limi-termined using a flow transducer (K520; KL Engi-tation make it important to clarify whether arterial K�

neering, Sylmar, CA, USA). The VO2 was calculated andregulation during exercise is abnormal in ESRF. There-displayed (Turbofit, Vacumetrics, Ventura, CA, USA)fore, to our knowledge for the first time, we measuredevery 15 seconds on a personal computer. The gas ana-PK in arterial blood, sampled during exercise, at well-lyzers and ventilometer were calibrated before and re-controlled work rates, and compared this to the PK re-

sponses during exercise in a matched control group. We checked after each test with precision reference gasestested two hypotheses: (1) that patients with ESRF and a standard syringe, respectively.would display an abnormally elevated arterial PK during

Blood sampling and analysesand following incremental exercise compared to controlsubjects; and (2) that altered K� regulation is related to A 16G needle (Kawasumi Laboratories, Tokyo, Ja-impaired VO2 peak. pan) was inserted into the arterial side of the forearm

arteriovenous fistula, allowing sampling of arterial bloodin the ESRF patients. In CON, a catheter (Jelco 20GMETHODSor 22G) was inserted into a radial artery under localSubjectsanesthesia (2% lignocaine injection). Resting blood sam-

All subjects gave written informed consent. All proce- ples were obtained after the subject had been seated fordures were approved by the Victoria University of Tech- 12 minutes on the cycle ergometer. Subsequent bloodnology Human Research Ethics Committee and the Royal samples were drawn during the last 15 seconds of eachMelbourne Hospital Board of Medical Research Advi-

minute of incremental exercise, at the point of fatiguesory Committee. Nine young untrained patients (5 males,

and at 1, 2, 5, 10, 20 and 30 minutes post-exercise, with4 females) with ESRF with no cardiovascular, respiratorysubjects maintaining their seated posture. The totalor skeletal muscle disease and eight (4 males, 4 females)quantity of blood withdrawn was approximately 75 mL.healthy, untrained, age-matched subjects participated asBlood samples were analyzed in duplicate for PK (Cibacontrols (CON). No significant differences were foundCorning 865, Essex, UK) and [Hb] (Radiometer OSM2;between the groups (mean � SD) for age (ESRF 25.8 �Radiometer, Copenhagen, Denmark), and in triplicate4.1 vs. CON 24.3 � 6.1 years), height (ESRF 169.3 � 6.7for hematocrit (Hct) using heparinized micro-hematocritvs. CON 171.2 � 7.7 cm) or body mass (ESRF 69.4 � 16.7tubes centrifuged for four minutes at 13,000 rpm. Ana-vs. CON 67.0 � 7.7 kg). The patients had received regularlyzers were calibrated immediately before and during thehemodialysis treatment for periods of 18 to 58 monthsanalyses with precision standards. Calculations for the(34.3 � 12.5 months, mean � SD). Causes of ESRFpercentage decline in plasma volume from resting condi-included glomerulonephritis (N � 6), reflux nephropathytions (�PV), the ratio of rise in PK at peak exercise relative(N � 2) and Alport syndrome (N � 1). Six patients wereto total work output (�PK · work�1, nmol · L�1 · J�1),treated for hypertension with an angiotensin-convertingand the correction of PK for the decline in PV were asenzyme (ACE) inhibitor and/or a Ca2� channel blocker,

while four patients were receiving treatment with epoetin. previously described [10, 22, 31].

Sangkabutra et al: K� and exercise limitation in renal failure 285

Fig. 1. Oxygen consumption (VO2) at common submaximal and atpeak work rates during incremental cycle exercise, in patients with end-stage renal failure (�; ESRF, N � 9) and in healthy controls (�; CON,N � 8). Data are presented as mean � SEM. ***ESRF � CON, P �0.05; #slope of the VO2 – work rate relationship, ESRF � CON, P �0.005.

Total blood volume

For patients with ESRF, Hb mass was measured, andblood volume (BV) and plasma volume (PV) then de-rived, using the carbon monoxide (CO) rebreathing tech-nique [32]. Patients were seated for 20 minutes prior to Fig. 2. Change in plasma volume during and following incrementalthis estimation. Tests were conducted two days prior to cycle exercise (A ) and during exercise expressed as a proportion of

VO2 peak (B ), for patients with ESRF (�; N � 9) and healthy controlsthe exercise test and at the same time of day and interval(�; N � 8). Data are mean � SEM. *�PV, ESRF � CON, P � 0.05.from dialysis. For control subjects, resting BV and PV

were assumed equivalent to results previously obtainedusing the same technique, in a group of healthy controls ESRF 24.5 � 2.3 vs. CON 43.7 � 1.5 mL · kg�1 · min�1)of similar age, size and gender in our laboratory (ab- than CON (Fig. 1). The slope of the linear relationshipstract; Haukka et al, Clin Sci 87:176, 1994) [32]. The rise between the submaximal VO2 and work rate was less inin arterial K� content (mmol) with peak exercise was ESRF (ESRF 0.12 � 0.01 vs. CON 0.15 � 0.01 mL · J�1,calculated from the product of PK and PV, after correct- P � 0.005; Fig. 1). The heart rate in ESRF was highering for the calculated decline in PV with exercise. at rest (ESRF 94 � 4 vs. CON 77 � 5 beats · min�1)

and at all common submaximal work rates (25 to 100 WStatistical analysesinclusive, P � 0.05), but was lower at peak exercise

Results are reported as mean � SEM, with P � 0.05 compared with CON (ESRF 168 � 7 vs. CON 181 � 3accepted for significance. PK at rest, during exercise at beats · min�1, P � 0.05).submaximal work rates common for both ESRF and

Hematology and plasma volumeCON (25, 50, 75 and 100 W), at peak exercise and inrecovery for both groups were compared by analysis of In ESRF patients, [Hb] ranged from 6.1 to 9.2 g · dL�1

variance (ANOVA) with repeated measures, with post- (ESRF 7.8 � 0.4 vs. CON 14.3 � 0.5 g · dL�1, P � 0.001)hoc Newman-Keuls analyses. and Hct from 19.6 to 31.1% (ESRF 25.3 � 1.2 vs. CON

41.3 � 1.5%, P � 0.001). Resting plasma volume (PV)in ESRF patients was 3.70 � 0.30 L, or 54.4 � 5.1 mL ·RESULTSkg�1. This was 32% (L) and 30% (mL · kg�1) greater

Exercise performance and oxygen consumption than the PV previously measured for equivalent young,The reason given for stopping the exercise test was healthy controls, at 2.81 � 0.17 L, or 41.9 � 1.4 mL ·

leg fatigue for all ESRF patients, who attained a lower kg�1, respectively (abstract; ibid) [32]. The PV declinedprogressively with increased exercise intensity, with a(P � 0.001) peak work rate (49%) and VO2 peak (44%,

Sangkabutra et al: K� and exercise limitation in renal failure286

Fig. 4. Rate of the rise in arterial PK and the rise in PK relative to workdone for patients with end-stage renal failure (�; N � 9) and healthycontrols (�; N � 8). Data expressed as mean � SEM. ***ESRF �CON, P � 0.001; #ESRF � CON, P � 0.005.

Fig. 3. Arterial plasma [K�] (PK) at rest, during and following incre-mental cycle exercise (A ) and the rise in PK during exercise (B ), forpatients with end-stage renal failure (�; N � 9) and healthy controls(�; N � 8). Data are mean � SEM. *ESRF � CON, P � 0.05; φESRF �CON, P � 0.005.

greater �PV in ESRF evident at 100 W (P � 0.05), Fig. 5. Arterial PK expressed as a proportion of VO2 peak, for patientsbut not at peak exercise (Fig. 2A). The resting PV was with end-stage renal failure (�; N � 9) and healthy controls (�; N � 8).

Data are mean � SEM.regained after 10 minutes of recovery in ESRF, but notuntil 30 minutes of recovery in CON (Fig. 2A). Whenexpressed relative to the percentage of the peak exerciseVO2, the �PV was similar in both groups (Fig. 2B). ESRF than in CON for the first two minutes of recovery

(P � 0.05).Arterial plasma potassium concentration and contentThe absolute rise in PK above rest (�PK) was greaterduring exercise

in ESRF than in CON during exercise at work rates ofResting arterial PK was approximately 0.9 mmol · L�1

50, 75 and 100 W (P � 0.05), but was similar at peakhigher in ESRF than in CON (P � 0.005; Fig. 3A). PK exercise (ESRF 2.35 � 0.26 vs. CON 2.25 � 0.13 mmol ·increased during exercise in both groups (P � 0.001), L�1; Fig. 3B). During exercise the rate of rise in PK inwith greater PK found during submaximal and peak exer- ESRF (�PK · min�1) was twice that of CON (P � 0.005),cise for ESRF than for CON (P � 0.05; Fig. 3A). PK at while at peak exercise, the �PK · work�1 ratio was 3.6-peak exercise ranged from 5.91 to 8.98 mmol · L�1 in fold higher in ESRF (P � 0.001; Fig. 4).ESRF, including �8 mmol · L�1 in three patients, and When expressed as a proportion of peak exercise VO2,in CON ranged from 5.72 to 6.92 mmol · L�1. PK fell PK remained higher in ESRF at rest and throughoutrapidly after exercise, reaching pre-exercise values by exercise than in CON (Fig. 5), whereas the �PK wastwo minutes post-exercise; a similar rate of decline over similar for both groups (data not shown). To determinethe initial two-minute recovery period was seen also whether the higher PK during exercise in ESRF might(ESRF 1.14 � 0.14 vs. CON 1.00 � 0.06 mmol · L�1 · result from greater hemoconcentration, the PK was cor-

rected for the decline in plasma volume from rest (�PV).min�1, NS; Fig. 3A). However, PK remained higher in

Sangkabutra et al: K� and exercise limitation in renal failure 287

mance in ESRF, not simply mediated by an elevated PK,but rather, reflecting some related associated abnormal-ity, possibly in muscle K� regulation.

When data from the ESRF and controls were pooled(N � 17), a significant association was found betweenVO2 peak and resting [Hb] (r � 0.81, P � 0.0001). Whenthe two groups were considered separately, no significantrelationship was found between VO2 peak and resting[Hb] in either the control or ESRF group.

DISCUSSION

This study investigates abnormalities in PK regulationduring exercise in ESRF and provides two major andnovel findings. We provide clear evidence that arterialPK is abnormally elevated and K� regulation severelyFig. 6. Relationships between VO2 peak and the rise in PK relative to

work done for patients with end-stage renal failure (�; N � 9) and impaired, during and following incremental exercise inhealthy controls (�; N � 8). Regression lines indicate controls (solid

ESRF, compared to healthy controls. We also show, toline, r � �0.73, P � 0.05), ESRF (dash-dot line, r � �0.50, NS) andpooled data (dashed line, r � �0.80, P � 0.05). Data are mean � SEM. our knowledge for the first time, that K� regulation was

related to the maximal incremental exercise performance(VO2 peak). Therefore, it is highly likely that impairedK� regulation in ESRF contributes to their reduced exer-

After correction, PK remained higher in ESRF at allcise performance. This probably reflects deficiencies in

exercise work rates and in recovery compared to controlsskeletal muscle Na� and K� regulation, which precipitate

(P � 0.01) and the corrected �PK at peak exercise didmuscle fatigue and lead to early exercise cessation.not differ between groups (ESRF 1.27 � 0.22 vs. CON

1.49 � 0.10 mmol · L�1, NS). Abnormal extrarenal K� regulation duringSince the resting absolute PV in ESRF was higher than exercise in ESRF

the value previously determined in controls, changes inThis study provides clear evidence of grossly abnormalthe total plasma K� content with exercise also were cal-

arterial PK before, during and following exercise inculated in ESRF and estimated in controls. A similarESRF, together with severely compromised maximal ex-rise in PK content at peak exercise was found in bothercise performance. The higher resting PK in ESRF isgroups (ESRF 4.18 � 0.90 vs. CON �4.19 mmol). How-dependent on the time post-dialysis and might be ex-ever, the �PK content · work�1 ratio at peak exercise waspected to contribute to the systematically greater PK ob-nearly fourfold higher in ESRF (ESRF 163.09 � 29.26served during submaximal and peak exercise and in re-vs. CON, �42.48 nmol · J�1), consistent with findings forcovery. However, other factors are clearly importantthe �PK · work�1 ratio. Therefore, the exaggerated PKsince a greater rise in PK (that is, �PK) during submaximalresponse in ESRF relative to work output cannot beexercise was found in ESRF. Although the �PK at peakexplained simply by different fluid volumes at rest orexercise was similar in ESRF and controls, this occurredduring exercise.at drastically compromised peak exercise work rates and

Relationships between VO2 peak, [Hb] VO2 in ESRF. The similar �PK at maximal exercise in theand K� variables two groups does not indicate a normal response in ESRF,

since this would imply an identical rate of rise in arterialWhen the ESRF and the control groups were consid-PK during exercise (�PK), as well �PK at peak exerciseered separately, a significant inverse relationship waswhen normalized for work output (�PK · work�1 ratio)found between VO2 peak and the �PK · work�1 ratioin ESRF and controls. Abnormality is evidenced since(Fig. 6) in controls (r � �0.73, N � 8, P � 0.05), butESRF patients exhibited a twofold more rapid rate ofnot in ESRF (r � �0.50, N � 9; NS). When all data�PK and a fourfold higher �PK · work�1 ratio. The four-were pooled (N � 17), a strong inverse relationship wasfold higher �PK · work�1 ratio in ESRF is much higherfound between VO2 peak and the �PK work�1 ratio (rthan the 22% and 78% higher ratios that can be esti-� �0.80, P � 0.001; Fig. 6). No significant relationshipsmated from data in previous studies [27, 29]. This proba-were found between VO2 peak and the peak exercise PK

bly reflects their underestimation of PK due to venousfor either the separate groups or for pooled data (r �blood sampling and further emphasizes the importance�0.34). These findings suggest that impaired K� regula-

tion may play a direct role in reduced exercise perfor- of arterial rather than venous blood sampling.

Sangkabutra et al: K� and exercise limitation in renal failure288

Possible mechanisms of impaired K� regulation during A possible impairment in muscle Na�,K�-ATPase ac-tivity most likely would affect all skeletal muscle. Inac-exercise in uremiative skeletal muscle plays an important role in plasmaDuring exercise, PK reflects the balance between K�

K� clearance during exercise [30]. Thus, impaired musclereleased into plasma from contracting muscles and K�

Na�,K�-ATPase activity in ESRF would reduce PK clear-cleared from plasma by other tissues, including inactiveance by inactive muscle and directly contribute to theskeletal muscle, red blood cells and the kidneys [11,elevated PK that we have observed. While the role of33–35]. The sharp rise in PK during exercise (Fig. 3)red blood cells in plasma K� clearance during exerciseindicates that K� released by contracting skeletal muscleremains controversial [11, 34, 39], depressed erythrocyticexceeds the PK clearance rate [11]. Thus, greater PK dur-Na�,K�-ATPase activity in ESRF [16, 17] also might being exercise in ESRF likely reflects an augmented K�

important in impaired PK regulation in ESRF [16, 17].release from contracting muscles. The rate of K� releaseTherefore, impaired Na�,K�-ATPase activity in bothfrom contracting muscle is dependent upon the balancecontracting and inactive skeletal muscle in ESRF wouldbetween the rates of cellular K� efflux and K� re-uptakeboth exacerbate muscle K� release and directly reduce(which occurs via Na�,K�-ATPase activity, discussedPK clearance.later in this article), multiplied by the muscle blood flow

It is likely that the greatly diminished or lack of renal[11, 33]. Cellular K� efflux is primarily dependent uponK� clearance in ESRF would make only a very smallmembrane excitation, and thus upon motor unit recruit-contribution to their exaggerated plasma K� responsement [23, 36, 37]. Any absolute submaximal work rateduring exercise. Renal K� clearance in healthy controlsrepresents a greater relative intensity (that is, higher %may double during moderate exercise [35, 40], but thisVO2 peak) in ESRF, which would demand increasedeffect disappears at higher intensity submaximal exercisemotor unit recruitment and lead to greater cellular K�

[35] and K� clearance falls during sprint exercise [41].efflux [33]. Interestingly, the �PK with exercise in ESRFHowever, even a doubling of renal K� clearance aboveand controls was similar when exercise intensity wasresting values (�1.5 �mol · kg�1 · min�1) [41] representsexpressed relative to peak exercise VO2. This indicatesonly a minor fraction of the muscle K� efflux rates,that the greater rise in PK during exercise in ESRF wasestimated at �38 �mol · kg�1 · min�1 during submaximallargely due to their greater relative exercise intensity.cycling exercise [33]. Thus, extrarenal mechanisms ac-This suggests that an exacerbated muscle K� release atcount for almost all of the PK clearance during exercise,very low work rates may be a major factor underlyingand the lack of renal K� excretion during exercise cannotthe greater PK response in ESRF. Muscle blood flow isaccount for our observation of impaired PK regulationreduced in ESRF [5], which by itself is inconsistent withwith exercise in ESRF. It is possible that the highera hypothesized greater muscle K� release (A-V [K�]resting PK in ESRF also may affect the rate of K� clear-difference flow). However, reductions in skeletal mus-ance, as shown at rest in a small group of hemodialysiscle blood flow itself may have elevated muscle venouspatients [42]. However, this has not been investigated inK� efflux [38]. Thus, reduced blood flow in ESRF doesexercise.not preclude the likelihood of a greater K� release from

The elevated PK response in ESRF was only partiallycontracting muscle.due to a greater hemoconcentration (about 25%), sinceImpaired Na�,K�-ATPase activity in skeletal musclePK remained higher after correction for fluid shifts, atcould reduce K� re-uptake in contracting muscle cellssubmaximal and peak (12%) exercise work rates, as welland thereby increase venous [K�], exacerbate muscleas in recovery. Finally, chronic deconditioning may con-K� release and directly contribute to elevated PK duringtribute to impaired K� regulation in ESRF, since thisexercise in ESRF. There are no published data support-results in an elevated �PK · work�1 ratio during exerciseing defective skeletal muscle Na�,K�-ATPase in uremic[43]. Since training increases the skeletal muscle Na�,K�-patients. However, 30 to 50% lower maximal in vitroATPase content and reduces the �PK · work�1 ratio dur-muscle Na�,K�-ATPase activity has been reported ining exercise in healthy subjects [31], improved K� regula-uremic rats [12–14]. Further, impaired muscle Na�,K�-tion also may contribute to the improved exercise perfor-ATPase activity is suggested also by lowered rates ofmance in ESRF after physical conditioning [1].plasma rubidium clearance, in both predialysis and he-

modialysis patients [17]. Such a premise is also entirelyImplications of abnormal K� regulation forconsistent with the reported muscle membrane depolar-muscle fatigueization and disturbed muscle [Na�] in uremic patients

These study results show, to our knowledge for the first[18–20] and with impaired erythrocytic Na�,K�-ATPasetime, a highly significant inverse relationship between theactivity [16, 17]. Consistent with these findings, we have�PK · work�1 ratio and VO2 peak in the pooled ESRFrecently found muscle Na�,K�-ATPase activity to beand control data. Further, we have recently found that39% less in a small group of uremic patients compared

to healthy controls (unpublished data). EPO-induced increases in [Hb] from 10 to 14 g · dL�1

Sangkabutra et al: K� and exercise limitation in renal failure 289

Reprint requests to M.J. McKenna, Ph.D., School of Human Move-in ESRF resulted in a reduced �PK · work�1 ratio andment, Recreation and Performance (FO22), Victoria University of Tech-

an increased VO2 peak [7]. Together these data suggest nology, PO Box 14428, MCMC, Melbourne, Victoria, 8001, Australia.E-mail: michael.mckenna@vu.edu.authat impaired K� regulation may play an important role

in the poor exercise performance in ESRF. The impairedREFERENCESK� handling properties may reflect Na�,K�-ATPase ac-

tivity abnormalities within skeletal muscle that predis- 1. Painter P, Moore GE: The impact of recombinant human erythro-poietin on exercise capacity in hemodialysis patients. Adv Renalpose the muscle to early muscle fatigue, thereby leadingRepl Ther 1:55–65, 1994to early exercise cessation and reduced VO2 peak.

2. Grunze M, Kohlmann M, Mulligan M, et al: Mechanisms ofThe proposed underlying mechanism for such an effect improved physical performance of chronic hemodialysis patients

after erythropoietin treatment. Am J Nephrol 10(Suppl 2):S15–S23,is dramatic disturbances in muscle transsarcolemmal [K�]1990and [Na�] gradients during exercise. At peak exercise,

3. Mayer G, Thum J, Cada EM, et al: Working capacity is increasedthe mean PK exceeded 7 mmol · L�1 in ESRF, indicating following treatment with recombinant human erythropoietin. Kid-

ney Int 34:525–528, 1988an even higher [K�] in femoral venous plasma [22] and4. McMahon LP, Johns JA, McKenzie A, et al: Haemodynamicin muscle interstitial [K�] [44]. A recent study using micro-

changes and physical performance at comparative levels of haemo-dialysis probes in healthy subjects has demonstrated in- globin after long-term treatment with recombinant erythropoietin.

Nephrol Dial Transplant 7:1199–1206, 1992creases in interstitial [K�] in contracting skeletal muscle5. Marrades RM, Roca J, Campistol JM, et al: Effects of erythropoi-to in excess of 11 mmol · L�1 [44]. An even greater increase etin on muscle O2 transport during exercise in patients with chronic

could be expected in ESRF that, together with a possible renal failure. J Clin Invest 97:2092–2100, 19966. Moore GE, Bertocci LA, Painter PL: 31P-magnetic resonancereduction in muscle t-tubular [Na�], would substantially

spectroscopy assessment of subnormal oxidative metabolism inreduce muscle tetanic force and cause muscle fatigue skeletal muscle of renal failure patients. J Clin Invest 91:420–424,[45–47]. We have recently demonstrated that the muscle 1993

7. McMahon LP, McKenna MJ, Sangkabutra T, et al: Physical per-maximal Na�,K�-ATPase activity declines during fatigu-formance and associated electrolyte changes after haemoglobining contractions in healthy controls, which may exacer- normalization: A comparative study in haemodialysis patients.

bate muscle K� and Na� perturbations [10]. Thus, any Nephrol Dial Transplant 14:1182–1187, 19998. Diesel W, Emms M, Knight BK, et al: Morphologic features ofimpairment in muscle Na�,K�-ATPase activity in ESRF

the myopathy associated with chronic renal failure. Am J Kidneycould compound this effect, precipitating early fatigue. Dis 22:677–684, 19939. Lundin AP, Stein RA, Brown CD, et al: Fatigue, acid-base andAnemia is perhaps the most critical additional, but

electrolyte changes with exhaustive treadmill exercise in hemodial-not exclusive, factor in the severe exercise and peak VO2 ysis patients. Nephron 46:57–62, 1987limitation in patients with ESRF [3] and we also found 10. Fraser SF, Li JL, Carey MF, et al: Fatigue depresses maximal in

vitro skeletal muscle Na�,K�-ATPase activity in untrained anda significant correlation between resting [Hb] and peaktrained individuals. J Appl Physiol 93:1650–1659, 2002VO2. However, normalization of [Hb] with EPO did not

11. Sejersted OM, Sjøgaard G: Dynamics and consequences of potas-fully restore muscle performance or the VO2 peak [7], sium shifts in skeletal muscle and heart during exercise. Physiol

Rev 80:1411–1481, 2000and others also have reported that increases in exercise12. Bonilla S, Goecke IA, Bozzo S, et al: Effect of chronic renalperformance are less than those expected with an increased failure on Na�, K�-ATPase alpha 1 and alpha 2 mRNA transcrip-

[Hb] [1, 2, 5]. This indicates that factors other than ane- tion in rat skeletal muscle. J Clin Invest 88:2137–2141, 199113. Druml WR, Kelly A, May RC, Mitch WE: Abnormal cationmia must contribute to the exercise limitation in ESRF.

transport in uremia. Mechanisms in adipocytes and skeletal muscleIn conclusion, we have demonstrated that extrarenal from uremic rats. J Clin Invest 81:1197–1203, 1988K� regulation is grossly impaired during incremental 14. Goecke IA, Bonilla S, Marusic ET, Alvo M: Enhanced insulin

sensitivity in extrarenal potassium handling in uremic rats. Kidneyexercise in patients with ESRF. Compared to controls,Int 39:39–43, 1991these patients exhibit an abnormally high PK at rest, 15. Greiber S, England BK, Price SR, et al: Na� pump defects in

during exercise and in recovery; a fourfold higher rise chronic uremia cannot be attributed to changes in Na�-K�-ATPasemRNA or protein. Am J Physiol 266:F536–F542, 1994in PK at peak exercise when expressed relative to work

16. Cole CH: Decreased ouabain-sensitive adenosine triphosphataseoutput; and severely reduced maximal exercise perfor- activity in the erythrocytes membrane of patients with chronic

renal disease. Clin Sci 45:775–784, 1973mance and VO2 peak. Such abnormal K� homeostasis17. Brearley CJ, Aronson JK, Boon NA, Raine EG: Effects of hemo-with exercise might be due to defective muscle Na� and

dialysis and continuous ambulatory peritoneal dialysis on abnor-K� regulation. Impaired K� regulation likely contributes malities of ion transport in vivo in patients with chronic renal

failure. Clin Sci 85:725–731, 1993directly to the early muscle fatigue and poor exercise18. Bergstrom J, Hultman E: Water, electrolyte and glycogen contentperformance exhibited by these patients.

of muscle tissue in patients undergoing regular dialysis therapy.Clin Nephrol 2:24–34, 1974

19. Cotton JR, Woodard T, Carter NW, Knochel JP: Resting skele-ACKNOWLEDGMENTStal muscle membrane potential as an index of uremic toxicity. A

This study was supported in part by the Australian Kidney Founda- proposed new method to assess adequacy of hemodialysis. J Clintion. A portion of the findings were published in abstract form [Sangka- Invest 63:501–506, 1979butra et al, Nephrology 3(Suppl 1):S304, 1997]. We are grateful to the 20. Cunningham JN, Carter NW, Rector FC, Selden DW: Restingpatients and healthy volunteers for their enthusiastic cooperation in transmembrane potential difference of skeletal muscle in normal

subjects and severely ill patients. J Clin Invest 50:49–59, 1971this study.

Sangkabutra et al: K� and exercise limitation in renal failure290

21. Sjøgaard G, Adams RP, Saltin B: Water and ion shifts in skeletal kalakis T: Renal responses to bicycle exercise in trained athletes:Influence of exercise intensity. Int J Sports Med 7:86–88, 1986muscle of humans with intense dynamic knee extension. Am J

36. Everts ME, Clausen T: Excitation-induced activation of the Na�-Physiol 248:R190–R196, 1985K� pump in rat skeletal muscle. Am J Physiol 266:C925–C934,22. McKenna MJ, Heigenhauser GJ, McKelvie RS, et al: Sprint1994training enhances ionic regulation during intense exercise in men.

37. McKenna MJ: Role of the skeletal muscle Na�,K�-pump duringJ Physiol 501:687–702, 1997exercise, in Biochemistry of Exercise X, edited by Hargreaves M,23. Medbø JI, Sejersted OM: Plasma potassium changes with highThompson MW, Champaign, Human Kinetics Publishers, 1998, ppintensity exercise. J Physiol 421:105–122, 199071–9724. Vøllestad NK, Hallen J, Sejersted OM: Effect of exercise inten-

38. Lindinger MI, Hawke TJ: Increased flow rate and papaverinesity on potassium balance in muscle and blood of man. J Physiolincrease K� exchange in perfused rat hindlimb skeletal muscle.475:359–368, 1994Can J Physiol Pharmacol 77:536–543, 199925. Castellino P, Bia MJ, DeFronzo RA: Adrenergic modulation of 39. Lindinger MI, Horn PL, Grudzien SP: Exercise-induced stimula-

potassium metabolism in uremia. Kidney Int 37:793–798, 1990 tion of K� transport in human erythrocytes. J Appl Physiol 87:2157–26. Newstead CG: Exercise-induced arterial hyperkalaemia in haemo- 2167, 1999

dialysis patients. (Letter) Nephrol Dial Transplant 5:400–402, 1990 40. Castenfors J: Renal function during prolonged exercise. Ann N Y27. Clark BA, Shannon C, Brown RS, Gervino EV: Extrarenal Acad Sci 301:151–159, 1977

potassium homeostasis with maximal exercise in end-stage renal 41. McKelvie RS, Lindinger MI, Heigenhauser GJ, et al: Renaldisease. J Am Soc Nephrol 7:1223–1227, 1996 responses to exercise-induced lactic acidosis. Am J Physiol 257:

28. Latos DL, Strimel D, Drews MH, Allison TG: Acid-base and R102–R108, 1989electrolyte changes following maximal and submaximal exercise 42. Sterns RH, Feig PU, Pring M, et al: Disposition of intravenousin hemodialysis patients. Am J Kidney Dis 10:439–445, 1987 potassium in anuric man: A kinetic analysis. Kidney Int 15:651–660,

29. Huber W, Marquard E: Plasma potassium and blood pH following 197943. Hall MJ, Snell GI, Side EA, et al: Exercise, potassium, andphysical exercise in dialysis patients. Nephron 40:383–384, 1985

muscle deconditioning post-thoracic transplantation. J Appl Phys-30. Kowalchuk JM, Heigenhauser GJF, Lindinger MI, et al: Roleiol 77:2784–2790, 1994of lungs and inactive muscle in acid-base control after maximal

44. Green S, Langberg H, Skovgaard D, et al: Interstitial and arterial-exercise. J Appl Physiol 65:2090–2096, 1988venous [K�] in human calf muscle during dynamic exercise: Effect31. McKenna MJ, Schmidt TA, Hargreaves M, et al: Sprint trainingof ischaemia and relation to muscle pain. J Physiol 529:849–861,increases human skeletal muscle Na�, K�-ATPase concentration2000and improves K� regulation. J Appl Physiol 75:173–180, 1993

45. Bouclin R, Charbonneau E, Renaud J-M: Na� and K� effect32. Burge CM, Skinner SL: Determination of hemoglobin mass andon contractility of frog sartorius muscle: Implication for the mecha-blood volume with CO: Evaluation and application of a method. nism of fatigue. Am J Physiol 268:C1528–C1536, 1995

J Appl Physiol 79:623–631, 1995 46. Nielsen OB, Clausen T: The Na�/K�-pump protects muscle excit-33. Hallen J: K� balance in humans during exercise. Acta Physiol ability and contractility during exercise. Exerc Sport Sci Rev 28:159–

Scand 156:279–286, 1996 164, 200034. Lindinger MI, McKelvie RS, Heigenhauser GJF: Potassium and 47. Overgaard K, Nielsen OB, Flatman JA, Clausen T: Relations

lactate distribution in humans during intense exercise: Role in between excitability and contractility in rat soleus muscle: Role ofmuscle fatigue attenuation? J Appl Physiol 78:765–777, 1995 the Na�,K�-pump and Na� /K� gradients. J Physiol 518:215–225,

199935. Virvidakis C, Loukas A, Mayopoulou-Symvoulidou D, Mounto-