Originai Paper

NEPHRON Nephron 1999;82:122-126 Accepted: January 15, 1999

Potassium Removallncreases the QTcInterval Dispersion during Hemodialysis

Adamasco Cupistia Fabio Galettaa Raffaele Capriolib Ester MorelliaGian Carlo Tintoria Ferdinando Franzonia Alberto Lippib Mario MeolaaPaolo Rindib Giuliano Barsottia

aDipartimento di Medicina Interna, Università di Pisa, and bDivisione Nefrologica, A.O. Pisana, Pisa, Italia

Key WordsOTc dispersion . Hemodialysis. Potassium

Abstract

This study was planned to clarify the mechanism(s) bywhich hemodialysis increases the OTc dispersion, amarker of risk of ventricular arrhythmias. To this aim, 10uremic patients, without any relevant heart diseases,underwent two different types of hemodialysis sched-ules. In the first, 1 h of isolated high rate ultrafiltrationpreceded the standard diffusive procedure. In the sec-ond, during the first hour of standard bicarbonate hemo-dialysis, the decrease of plasma potassium concentra-tion was prevented by increasing K+concentration in thedialysate, according to its pre dialysis plasma levels. Dur-ing the high rate ultrafiltration period, together with ECGsigns of increased sympathetic nervous system activityand catecholamines secretion, the OTc dispersion didnot change significantly. Instead, an evident incrementwas observed 1 h after the start of the diffusive hemodi-

alysis, then slowly progressing until the end of the dialy-sis and finally returning to the pre dialysis values within2 h after the end of the session. To the contrary, theincrease ofthe OTc dispersion was totally blunted duringa standard hemodialysis procedure in absence of plasmaK+decrease, but appeared again when the K+dialysate

KARGER @ 1999 S. Karger AG, Basel0028-2766/99/0822-0122$17.50

Fax+41 61 306 1234E-Mailkarger@karger.chwww.karger.com

Accessible online al:

http://BioMedNet.com/karger

fluid concentration was restored to 2 mmol/I. This studyprovides evidence that the increase of OTc dispersionoccurring on hemodialysis is mainly related to the diffu-sive process, more precisely to the K+removal. This isone more reason to focus attention on K+removal rate

especially when hemodialysis treatment is given in ure-mics affected by cardiac diseases with high risk ofarrhythmias.

Introduction

The interlead variability ofthe QT interval in the sur-faeeeleetroeardiogram is ealled QT dispersion, defined asthe differeneebetween maximal and minimal QT intervalduration. It refleets regional differenees in ventrieularreeovery time [1].

Evidenee exists that hemodialysis treatment eauses aninereased dispersion of QTe interval in a 12-1eadeleetro-eardiogram [2-4]. This suggestsan inereased regional het-erogeneityof repolarization times in adjaeent areas of themyoeardium [5-7] and it is assoeiated with higher risk ofventrieular taehyarrhythmias in various eonditions [8-11].

The ehanges of the intra- and extraeellular eleetrolyteeoneentrations and of the autonomie nervous system or

" Prof. Giuliano Barsotti

Dipartimento di Medicina Interna, Università degli Studi di Pisavia Roma, 671-56100 Pisa (Italia)Te!. +39 50 553 005, Fax +3905 502 617

catecholamines secretion have been proposed to explainthe well-known arrhythmogenic effect of the standardhemodialysisprocedure [12, 13].In particular, the suddenreduction of plasma potassium concentration as well asthe elevation of ionized calcium plasma levels [14-16]can increase the susceptibility to ventricular arrhythmias.In fact, the intra- and extracellular concentrations ofpotassium, calcium and magnesium are important factorsfor the electrical stability of the myocardium, since theyare involved in creating normal cellular excitability, im-pulse propagation and regular ventricular recovery. Ofconsequence, it is likely that the hemodialysis-inducedelectrolyte 'disequilibrium' could be the main cause oftheincreased QTc dispersion, but it has not been proven.

Moreover, other factors may potentially affectthe QTcdispersion since the anatomic-functional heterogeneityamong different areas of the ventricular myocardium canbe developed by the several and complex modificationson the electrolytic status and on the autonomic nervoussystembalance occurring during the hemodialysissession.The present study was aimed to individually identify thefactors responsible for the increased QTc dispersion fol-lowingthe hemodialysis procedure.

Patients and Methods

Patients

Ten chronic uremics (5 males and 5 females, aged 50 :t 13years)on maintenance hemodialysis treatment entered the study.

Patients were maintained on thrice-weekly bicarbonate hemodi-alysis treatment lasting 4 h. Polysulphone holIowfiber dialyzers (sur-face 1.8 m2) were used. The standard dialysate composition was thefolIowing: potassium 2 mmol/l, calcium 1.75 mmolll, magnesium0.37 mmolll, sodium 139 mmol/l, chloride 109 mmolll, acetate 3mmol/l and bicarbonate 32 mmol/I.

Patients affected by diabetes, heart failure, ischemic heart dis-ease, atrial fibrillation or conduction disturbances were exc1udedaswelI as those with proven intra dialysis cardiovascular instability.

Two dimensional echocardiography examination was performedin alI the studied patients: the detection of global or regional defectsof contractility, or of an ejection fraction lower than 60%, wereregarded as exc1usioncriteria.

No patient was on pharmacological treatment known to affect theQTc intervai. Informed consent from patients was obtained and thestudy was conducted in accordance with the principles of the Dec1a-ration ofHelsinki.

Study DesignTo separately evaluate the possible effects of the hemodialysis-

induced changes of extracelluiar volume, of catecholamines, of auto-nomic nervous system activity and of electrolytes (namely potas-sium) on the QTc dispersion values, two different dialysis scheduleswere planned.

QTc Dispersion and Potassium Removal

Study A. High rate isolated ultrafiltration was performed duringthe first hour of the session, then the ultrafiltration rate was reducedand a standard diffusive hemodialysis was begun and performed for3 h.

Surface l2-lead ECG recordings were taken before and afterultrafiltration, 1 h after the beginning of diffusive hemodialysis, atthe end and 2 h after the end of the hemodialysis session. At the sametimes, arterial blood pressure was recorded and blood samples weretaken from the afferent line for determinations of electrolytes, ureaand creatinine plasma levels.

Seven days later, the same hemodialysis schedule was planned asabove specified, and a 24-hour dynamic ECG monitoring was per-formed to assess the incidence of arrhythmias and the heart rate vari-ability during and out of the hemodialysis session. Epinephrine andnorepinephrine plasma levelswere determined at baseline and at theend ofthe isolated ultrafiltration periodo

Study B. During the first hour of a standard bicarbonate hemodi-alysis, the K+ dialysate concentration was increased to obtain thesame value of K+ plasma level at the beginning of the session, thenavoiding any plasma-dialysate K+gradient to prevent plasma K+lev-els from lowering. Then it was restored to a fixed concentration of2 mmol/l for the folIowing3 h of hemodialysis.

Samples for plasma K+ determinations and ECG recordings forevaluation ofQTc dispersion were taken before and at the end ofthehemodialysis period with high K+dialysate, then after 1 h and at theend of the folIowing period of hemodialysis with low K+ dialysateconcentration.

Methods

The 12-lead electrocardiogram recordings were performed byCardiovit CS-l 00 ECG Recorder (SchillerAG, Baar, Switzerland), at50 mm/s paper speed (gain lOmm/mV).

The measurements of QT interval in alI the possible leads wereblindly performed by a single observer. The QT interval was takenfrom the onset ofthe QRS complex to the end ofthe T wave. In thepresence of a U wave the nadir between T and U was taken as finalpoint.

For each ECG recording we took measurements ofthree consecu-tive QT interval from each lead and calculated the arithmetic means;these QT values were then corrected (QTc) by the heart rate accord-ing to the Bazett's formula. The QTc dispersion was determined asthe difference between the maximum and the minimum value ofthe QTc interval in different leads (ten at least) ofthe same ECG rec-ording.

The arithmetic mean of the QTc from the 12 leads of each ECGrecording was assumed as the QTc intervallength.

Heart rate variability provides an index of cardiac sympathetic-vagaI balance: it was evaluated from 24 h continuous ECG record-ings (Diagnostic Monitoring System USA), by measuring the stan-dard deviation of alI normal R-R intervals in the time domain(SDNN). The SDNN data during 2 h preceding the session and dur-ing the isolated ultrafiltration period were recorded.

The samples for plasma catecholamine assays were drawn by theafferent line, maintained on ice until refrigerated centrifugation wasperformed, then plasma was stored at -70 o C until assay was per-formed by the high-performance liquid chromatography method.The electrolytes and the other studied parameters were determinedusing the standard methods of our laboratory.

.. Nephron 1999;82:122-126 123

Table 1. StudyA:changes ofthe QTc dispersion and ofthe plasma concentrations ofurea, creatinine and electrolytes,before and at the end of isolated ultrafiltration (UF), I h after, at the end and 2 h after the end of standard hemodialy-sis (HD)

Table 2. Study A: mean QTc length, arterial blood pressure and heart rate values before andat the end of the isolated ultrafiltration period, 1h after and at the end of the bicarbonatehemodia/ysis

QTc length, msSystolic BP, mm HgDiastolic BP, mm HgHeart rate, bpm

416:t13130:t 2485:t 976:t 8

420:t22122:t208l:t 1374:t 11

430:t21118:t2678:tll78:t17

418:t17129:t 3384:t 1686:t 15

Statistical AnalysisAli the results are expressed as mean :t SD. Statistical evaluation

has been performed using the analysis of variance (ANOVA) forrepeated measurements with the Bonferroni's multiple comparisontest, and the Student's t test for paired data.

Differences were considered as statistically significant when p <0.05.

Results

At the end ofthe hemodialysis sessionthe QTc disper-sion invariably increased but retumed to the basaI valueswithin 2 h after the end ofthe hemodialysis (table 1).Dur-ing the isolated high rate ultrafiltration (1.3 :t 0.6 kg/h)period the QTc dispersion did not change whereas a sig-nificant increase occurred already 1h after the beginningofthe hemodialysis procedure (table 1)with a lowultrafil-tration rate (0.4 :t 0.2 kg/h). Then the QTc dispersionmore slowly increased until the end of the session whenthe QTc dispersion reached the highest values, that was

124 Nephron 1999;82:122-126

almost two times higher than at baseline. Finally, the QTcdispersion retumed to the baseline values within 2 h afterthe end ofthe hemodialysis session (table 1).

The expected changes ofthe plasma biochemistry dur-ing the different phases of study A are shown in table 1. Itemerges that the increment of QTc dispersion is associat-ed with the decrease of potassium and magnesium andwith the increase of ca1cium plasma levels.

Table 2 shows the values of the mean QTc length, ar-terial blood pressure and heart rate through the study A.

Norepinephrine circulating levels increased at the endof the high rate ultrafiltration period (fig. 1). Moreover,the SDNN values decreased during isolated ultrafiltration(fig. 1), reflecting a relative high sympathetic tone in thisperiodo

However, study B c1ear1yshows that the QTc disper-sion does not change when a standard hemodialysis pro-cedure is performed, provided that the potassium plasmaconcentration remains unchanged (fig. 2). Conversely,when potassium removal is restored, a rapid increase of

Cupisti et al.

QTc dispersion, ms 30:t 7 34:t 5 47:t 9b 58:t15a 3H 11Urea, mgldi 132:t 38 127:t 35 72 :t 20a 47:t 15a 63 :t 21"Creatinine, mgldl 10.H 1.78 1O.0:t 1.7 6.6:t 2.0a 4.3:t I.oa 5.6:t 1.1"Sodium, mEq/1 137:t 3 136:t 4 139:t 3 139:t 4 138:t 4Potassium, mEq/1 5.4:t 0.5 5.6:t 0.9 4.0:t O.4a 3.4:t 0.3a 4.0:t 0.5aChloride, mEq/1 100:t 4 99:t 5 102:t 3 102:t 4 97:t 3Calcium, mgldl 9.8:t0.8 10.5:t 1.1 11.0:t0.7b 11.6:t0.7a 11.2:t 0.5bMagnesium, mEq/1 2.4:t0.7 2.5:t0.7 1.9:t 0.6 1.7:t 0.5e 2.0:t0.6Phosphorus, mgldl 4.9:t 1.9 4.22:t 1.5 2.5:t 0.9b 2.2:t O.7a 3.6:t 1.2e

a p < 0.001, bP < 0.01, e p < 0.05 vs. before UFo

Fig. 1. Heart rate variability defined as the SD of aH normal R-Rintervals in the time domain (SDNN) values (dotted bars) and nor-epinephrine circulating levels (empty bars) before and during the iso-lated high rate ultrafiltration (UF) period (study A). Mean :t SD.* p < 0.05 vs. basaI value.

the QTc dispersion invariably occurred (fig.2), similarlyto that observed since the beginning of diffusive hemodi-alysisprocedure in study A.

No c1inically relevant arrhythmia was recorded byECG Holter monitoring, neither during the time of thehemodialysis, nor during the time preceding or followingit. In the studied patients, the incidence of ventricular orsupra ventricular premature contractions was very low(asa mean, 1 and lO per hour, respectively), and no signifi-cant abnormality related to the hemodialysis time periodwas observed.

Discussion

This study confirms that the standard bicarbonatehemodialysis procedure increases the QTc dispersionwhich seems to be independent of the technique used,since this effect was reported also when a soft hemodia-filtration, as it is the acetate-free biofiltration, was per-formed [4].

The increase ofthe QTc dispersion seemsto be limitedto the hemodialysis session, because it occurs from thefirst hour of dialysis, it is roughly two times greater at theend of the session,but it returns to basaIvalues within 2 hafter the end of the treatment.

QTc Dispersion and Potassium Removal

Fig. 2. Changes ofthe QTc dispersion values (solid line) and oftheK+plasma levels (dotted line) in the patients given hemodialysis ses-sion with high K+ dialysate concentration during the first hour, andwith fixed low K+ dialysate concentration during the foHowing 3 h(study B). ***p < 0.001 vs. basaI and l-hour values.

As far as the meehanism by whieh hemodialysis eausesthe inerease of the QTe dispersion is eoneerned, thepresent studyhas foeusedthe ehanges indueed by the dial-ysis treatment on extraeellular volume, autonomie ner-vous system aetivity, eleetrolytes and partieularly on po-tassium extraeellulareoneentrations.

Our results demonstrate that potassium removal is thecriticaI faetor indueing the enhaneement of the QTe dis-persion values during the hemodialysis session. This is inkeeping with the authors c1aimingthe role of potassium,and of its removal rate, as the main arrhythmias indueingfaetor assoeiated with the hemodialysis treatment. Ae-tually, Redaelli et al. [16] demonstrated that the arrhyth-mogenie effeetofhemodialysis isblunted when a eonstantplasma-dialysate K+ gradient is used instead of a fixed lowK+dialysate eoneentration. In addition, low intraeellularK+ eoneentrations were found to be associated with aninereased risk ofventrieular arrhythmias [15].

The present study suggests a minor role of the auto-nomie nervous system and ofthe fluid volume ehanges inenhaneing QTe dispersion during hemodialysis, in agree-ment with previous observations demonstrating no rela-tion between QT dispersion and autonomie abnormalitiesin patients with left-ventrieular hypertrophy and heartfailure[17].Indeed,duringthe isolatedhighrate ultrafil-tration period, in the presenee of strong sympathetie tone

'" Nephron 1999;82:122-126 125

activation and increased catecholaminessecretion,wedidnot observea significant increment ofthe QTc dispersion.Instead, the QTc dispersion invariably increased oncehemodialysis procedure was commenced, and its incre-ment was maximal during the first hour of treatmentwhen electrolytic changes, namely potassium decrease,were more rapido

Ventricular arrhythmias are one of the most severecomplications in hemodialysis patients, although they donot seem to predict their overall mortality [18]. The dis-persion of the QT interval reflects regional differences ofthe myocardial repolarization [l, 6, 7] which in tUfficanfavor arrhythmias due to re-entry mechanisms [5].There-by, the increased inter-Iead QT dispersion has been pro-posed as an indicator of arrhythmogenic risk, able to pre-dict severe ventricular arrhythmias or sudden death inpatients with hypertrophic cardiomyopathy [8, 9], myo-cardial infarction [lO]or chronic heart failure [Il].

Nevertheless, this study has failed to demonstrate adirect linkage between increased QTc dispersion andappearance of ventricular arrhythmias. It must be consid-ered that this study was carried out in stable chronic ure-

References

l Van de Loo A, Arendts W, Hohnloser SH:Variability of QT dispersion measurements inthe surface electrocardiogram in patients withacute myocardial infarction and in normal sub-jects. Am l CardioI1994;74:1113-1118.

2 Cupisti A, Galetta F, Tintori GC, Sibilia G,Morelli E, Meola M, Barsotti G: Effetto deltrattamento emodialitico sulla dispersionedell'intervallo QTc; in Timio M, Wizemann V,Venanzi S (eds): Cardionephrology 4. Cosenza,Bios Ed., 1997, pp 187-188.

3 Locati EH, Bagliani G, Saronio P, Stramba-Badiale M, Timio M, Schwartz Pl: lncreasedQT interval dispersion following potassium de-crease in nephropathic patients receiving he-modialysis; in Timio M, Wizemann V, Venan-zi S (eds): Cardionephrology 4. Cosenza, BiosEd., 1997,pp 173-177.

4 Cupisti A, Galetta F, Morelli E, Tintori GC,Sibilia G; Meola M, Barsotti G: Effect ofhemo-dialysis on the dispersion of the QTc interval.Nephron 1998;78:429-432.

5 Burgess Ml: Relation of ventricular repolariza-tion to electrocardiographic T-wave and ar-rhythmia vulnerability. Am l Physiol 1979;5:H391-H402.

mics on maintenance hemodialysis,without c1inicallyrel-evant cardiac diseases,just in order to separately analyzethe effects of the main changes occurring during a stan-dard dialysis session and potentially affecting the QTcdispersion. Actually, the values ofQTc dispersion at base-line and at the end of the hemodialysis are lower thanthose found to be associatedwith severe arrhythmic com-plications [lO, Il]. However, since hemodialysis per se isable to increase the QTc dispersion through lowering theK+ plasma concentration,a specialattention is neededwhen hemodialysis is given in uremics suffering fromsevere left-ventricular hypertrophy, heart failure, isch-emic heart disease, or arrhythmic or conduction distur-bances. This is one more reason to suggesthigher potas-sium concentration in the dialysate fluid or constant K+gradient hemodialysis techniques for the patients at highrisk of severeventricular arrhythmias [16, 19].

In conc1usion,the dispersion of the QTc interval doesincrease during hemodialysisby the loweringof the potas-sium plasma levels. Changes of the intradialysis potas-sium kinetics must be considered to reduce the risk ofarrhythmias in hemodialysis patients.

6 Higham PD, Hilton Cl, Aitcheson DA, FurnissSS, Bourke lP, Campbell RWF: QT dispersiondoes reflect regional variation in ventricularrecovery (abstract). Circulation 1992;86(suppl1):392.

7 Higham PD, Campbell RWF: QT dispersion.Br Heart l 1994;71 :508-51 O.

8 Mayet l, Shahi M, McGrath K, Poulter NR,Sever PS, Foale RA, Thom SA: Left Ventricu-lar Hypertrophy and QT dispersion in hyper-tension. Hypertension 1996;28:791-796.

9 Buja G, Miorelli M, Turrini P, Melacini P,Nava A: Comparison of QT dispersion in hy-pertrophic cardiomyopathy between patientswith and without ventricular arrhythmias andsudden death. Am l CardioI1993;72:973-976.

IO Mirvis DM: Spatial variation of QT intervalsin normal persons and patients with acute myo-cardial infarction. l Am Coli Cardiol 1985;5:625-631.

II Barr CS, Naas A, Freeman M, Lang CC,Struthers AD: QT dispersion and sudden unex-pected death in chronic heart failure. Lancet1994;343:327-329.

12 Gruppo Emodialisi e Patologie Cardiovasco-lari: Multicentre, cross-sectional study of ven-tricular arrhythmias in chronically hemodia-Iysed patients. Lancet 1988;ii:305-309.

13 Abe S, YoshizawaM, Nakanishi N, Yazawa T,Yokota K, Honda M, Sioman G: Electrocardio-graphic abnormalities in patients receiving he-modialysis. Am Heart l 1996;131:1137-1144.

126 Nephron 1999;82:122-126

14 Nishimura M, Nakanishi T, Yasui A, Tsuji Y,Kunishige H, Hirabayashi M, Takahashi H,Yoshimura M: Serum ca1cium increases the

incidence of arrhythmias during acetate hemo-dialysis. Am l Kidney Dis 1992;19:149-155.

15 Rombolà G, Colussi G, De Ferrari ME, Fron-tini A, Minetti L: Cardiac arrhythmias andelectrolyte changes during hemodialysis. Neph-rol Dial Transplant 1992;7:318-322.

16 Redaelli B, Locatelli F, Limido D, Andrulli S,Signorini MG, Sforzini S, Bonoldi L, VincentiA, Cerutti S, Orlandini G: Effect of a new mod-el of hemodialysis potassium removal on thecontrol of ventricular arrhythmias. Kidney lnt1996;50:609-617.

17 Davey PP, Bateman l, Mulligan lP, Forfar C,Barlow C, Hart G: QT interval dispersion inchronic heart failure and left ventricular hyper-trophy: Relation to autonomic nervous systemand Holter tape abnormalities. Br Heart l1994;71:268-273.

18 Sforzini S, Latini R, Mingardi G, Vincenti A,Redaelli B: Ventricular arrhythmias and four-year mortality in hemodialysis patients. Lancet1992;339:212-213.

19 Redaelli B: Electrolyte modelling in haemodi-alysis - potassium. Nephrol Dial Transplant1996; l I (suppl 2):39-41.

Cupisti et al.