Original Article (Clinical Investigation)

Original article (Clinical investigation)

Mechanical alternans in human idiopathic dilated cardiomyopathy is caused with impaired force-frequency relationship and enhanced poststimulation potentiation.

Takeshi Kashimura, Makoto Kodama, Komei Tanaka, Keiko Sonoda, Satoru Watanabe,Yukako Ohno,Makoto Tomita,Hiroaki Obata,Wataru Mitsuma, Masahiro Ito,Satoru Hirono,Haruo Hanawa,and Yoshifusa Aizawa.

Division of Cardiology, Niigata University Graduate School of Medical and Dental Sciences, Niigata, Japan

Corresponding author:

Takeshi Kashimura, Division of Cardiology, Niigata University Graduate School of Medical and Dental Sciences, 1-757 Asahimachi, Niigata, 951-8510 Japan

Tel: +81-25-227-2185

Fax: +81-25-227-0774

Email:

Abstract

Mechanical alternans (MA) is frequently observed in patients with heart failure, and is a predictor of cardiac events. However there have been controversies regarding the conditions and mechanisms of MA.In order to clarify heart rate-dependent contractile properties related to MA, we performed incremental right atrial pacing in 17idiopathic dilated cardiomyopathy (DCM)patients and in 6 control patients.The maximal increase in left ventricular dP/dt during pacing-induced tachycardia wasassessed as the force gain in the force-frequency relationship (FG-FFR) and the maximal increase in left ventricular dP/dt of the first post-pacing beats was examined as the force gain in poststimulation potentiation (FG-PSP). As a result, MA was induced in 9 DCM patients (DCM MA(+))but not in the other 8 DCM (DCM MA(-)) and not in any of the control patients. DCM MA(+) had significantly lower FG-FFR (34.7 ±40.9 mmHg/secvs.159.4 ±103.9 mmHg/sec, p=0.0091) and higher FG-PSP (500.0± 96.8 mmHg/sec vs. 321.9 ± 94.9 mmHg/sec, p=0.0017),accordingly wider gap between FG-PSP and FG-FFR (465.3 ± 119.4 mmHg/sec vs.162.5 ± 123.6, p=0.0001) than DCM MA(-) patients. These characteristics of DCM MA(+) showed clear contrasts to those of the control patients. In conclusion, MA is caused with impaired FFR despite significant PSP, suggesting that MA reflects ineffective utilizationof the potentiated intrinsic force during tachycardia.

Key Words: alternans;dilated cardiomyopathy;tachycardia; contractility; calcium

Introduction

Heart rate dependent hemodynamic changes in heart failure are subjects of great interest [1].Tachycardia frequently induces mechanical alternans (MA) in patients with heart failure [2,3] and tachycardia-induced MA has been reported to be a predictor of cardiac events [2].Although MA has been reported in patients with heart failure for more than a century [4-6], the mechanism causing MA is still to be clarified.Recent findings of cellular and computational studies suggest MA is caused by calcium transient alternans in cardiac myocyteswith impaired calcium handling [7-10]. Therefore it is considered that the calcium handling in failing hearts cannot catch up with rapid cardiac cycles and leads toMA. But relationships between MA and slow contractile properties have not been shown in in vivo human hearts.

Another well-known tachycardia-induced response in patients with heart failure is the attenuated force-frequency relationship (FFR). In normal hearts, an increased heart rate progressively enhances the force of ventricular contraction, that is, FFR. In failing hearts, tachycardia increases the contractile force to a lesser extent or can decrease it in some severe cases [11,12]. The mechanism of attenuated FFR is considered to be impaired calcium handling and indeed, FFR and MA were induced at the same time in rodents by manipulating calcium handling [13,14], but there has been no human data on the relationship between MA and FFR.Poststimulation potentiation (PSP) is also a rate-dependent increase in contractile force shown in the first beat after cessation of pacing-induced tachycardia [15,16] and has been reported to depend on calcium handling [17]. But again its relationship to MA has not been studied in patients with heart failure.In this study, we examined how MA correlates with FFR and PSP in the human heart and discussed whether it is consistent with previous clinical and recent experimental findings.

Methods

Subjects

Left ventricular contractile properties and the occurrence of mechanical alternans (MA) were assessed by right atrial incremental pacing in 23idiopathic dilated cardiomyopathy (DCM) patients with sinus rhythm who underwent diagnostic cardiac catheterization at the Niigata University Medical and Dental Hospital. The aim of this study was to assess the link between MA and myocardial contractile properties, therefore the patient population did not include those with significant coronary stenosis of 75% or more according to the American Heart Association classification by coronary angiography, those with localized left ventricular dysfunction by left ventriculography, those with significant mitral or aortic regurgitation of second degree or more according to the Sellor’s classification, or significant mitral or aortic stenosis by pressure measurements, because tachycardia induced ischemia, asymmetrical wall motion, or attenuated or enhanced left ventricular pressure development could influence the evaluation of myocardial contractile properties. Five patients whose Wenchebach point of atrioventricular conduction was 110 per minute or less were excluded to eliminate patients who could showMA at a higher heart rate. Another patient was excluded because of frequent premature ventricular contractions during pacing. As a consequence, we examined the remaining 17DCM patients.

Control data from hearts with preserved left ventricular function were taken from patients whose left ventricular ejection fraction was 60% or more after diagnostic catheterization for symptoms suggesting stable angina pectoris.Out of 12 patients examined, 4 patients with Wenckebach point of atrioventricular block at 110/min or less and 2 patients with frequent premature ventricular contractions were excluded, and data from the remaining 6 were used.Diagnoses of those patients were stable angina pectoris in 3 and chest pain syndrome without significant coronary artery stenosis in the other 3.

Right Atrial Incremental Pacing

After diagnostic procedures including right heart catheterization, coronary angiography and left ventriculography, a 7-French micromanometer-tipped pig tail catheter (Millar Instruments Incorporation, Houston, TX) was placed in the left ventricle and a pacing catheter was placed in the right atrium. Subsequent to measurements at a basal heart rate, the right atrium was paced at a rate just above the basal heart rate for at least 20 seconds while the left ventricular pressure was continuously recorded with its first derivative (dP/dt) until dP/dt became stable(Fig. 1a,c,e). At least 10 secondsafter cessation of pacing at a previous pacing rate,the right atrium was paced again at a rate 10 per minute faster than the previous rate for at least 20 seconds. Thus the pacing rate was increased by the increment of 10 per minute until the pacing rate reached 150 per minute.Other end points included patient’s complaint of discomfort, Wenckebach point of atrioventricular conduction, remarkable decline in left ventricular systolic pressure by 25% or down to below 70 mmHg.

All procedures in this study were approved by the ethical committee of Niigata University Graduate School of Medical and Dental Sciences and adhered to the principles of the declaration of Helsinki. Written informed consent was obtained from each patient beforehand.

Definition of Terms

Mechanical Alternans (MA): We defined MA as alternans in left ventricular systolic pressure of 4mmHg or more in our previous studies [3,18,19]. However this study focused onthe relationship between MA and FFR or PSP, which were measured by dP/dt, therefore MA should be defined by dP/dt. Then we defined MA as a phenomenon in which dP/dt alternans exceeded 100mmHg/sec or more for at least 20 consecutive beats based on our previous comparison of pressure alternans and dP/dt alternans [3].

Alternans Amplitude (AA) and Maximal AA (max AA):To quantify size of alternans during a steady state, the difference between dP/dt of a beat and that of the subsequent beat was defined as alternans amplitude (AA).The maximal AA (max AA) of each patient was determined by incremental right atrial pacing (Fig 1b,d,f).

Force Gain in Force- Frequency Relationship (FG-FFR):The force of left ventricular contraction was measured by the peak dP/dt. Typically, as the heart rate increases incrementally to a certain point the force keeps increasing, and when the heart rate increases further the force starts decreasing. This is known as the force-frequency relationship (FFR). The maximal force was assessed by incremental atrial pacing in each case and the increase from the force during the basal condition was defined as the force gain in FFR (FG-FFR) (Fig. 1b,d,f). In this study, many cases showed MA during steady-state pacing, therefore averages of two consecutive beats were used as the force. The rationale of using the average was that during MA, extrasystole at a phase reversal point abolishes MA and force is newly set in between that of a strong beat and that of a weak beat [20], and clinical studies on left ventricular contractility of heart failure patients haveused the average of dP/dt, and do not mention the presence or absence of MA [21,22].

Force Gain in Poststimulation Potentiation (FG-PSP): Animal and human data have shown that the contractile force of the first beat after pacing is positively dependenton increases in the pacing rate [15,16]. This is known as poststimulation potentiation (PSP). In this study, PSP was assessed in the first poststimulation spontaneous beat.The maximal force was assessed by incremental atrial pacing in each case and the increase from the force in the basal condition was defined as the force gain in PSP (FG-PSP) (Fig. 1b,d,f).

Statistical Analysis

Data analysis was performed with JMP 5.0.1J (SAS Institute, Cary, NC). All data are presented as means ± standard deviation (SD). Homoscedasticity between groups was tested by the F-test. The Shapiro-Wilk W test was used to assess whether the values were distributed normally in each group. Differencesfor normally distributed values between groups with homoscedasticitywereanalyzed using Student’s t-tests.Otherwise, the Wilcoxon rank-sum test was usedfor values, and Pearson’s chi-square test was used for categorical variables. P < 0.05 was considered significant.

Results

Induction of Mechanical Alternans (MA)

The average basal heart rate of the 17DCM patientswas 79.4 ± 16.1/min and the maximal paced heart rate was 134.7 ± 15.9 /min. Right atrial incremental pacing was terminated with the target heart rate of 150/min in 5 patients, with Wenchebach type atrioventricular block in 6, with low LV systolic pressure in 5, and with atrial tachycardia in 1. Noone complained of remarkable discomfort during pacing.

Nine out of the 17DCM patients showed mechanical alternans (MA) during pacing at least at a pacing rate (Fig. 1a arrows) (DCM MA(+)), but the other 8 did not (Fig. 1c arrow) (DCM MA(-)). The average heart rate at which maximal alternans amplitude (max AA) was obtained was 122.8 ± 17.5/min for the 9 DCM MA(+) patients. In eight out of the 9 cases, their AA reached the definition of MA (100 mmHg/sec) with a heart rate of 120/min or less. The other had a basal heart rate of 114 /min and AA rose over 100mmH/sec at 150/min. In 5 out of the 9 DCM MA(+) patients, AA did not reach its peak because of a continuous AA increase even at 150/min in 3 cases and because of Wenchebach type atrioventiricular block in 2 cases. On the other hand, the other 4 DCM MA(+) patients showed a decline of AA with an excessive increase of heart rate (Fig.1b open circles), indeed 3 cases lost MA during the incremental pacing. One patient each lost MA at 120/min, 130/min, and 140/min.In DCM MA(+) patients at maximal AA, LVEDP of a strong beats, LVEDP of a weak beat, and the differencebetween them were 8.6 ± 6.5 mmHg, 6.3 ± 6.4, and 2.2 ± 3.5 mmHg, respectively.

The average basal heart rate of the 6control patients was 75.5 ± 12.4/min and the maximal paced heart rate was 131.7 ± 9.8/min. None of the control patients showed MA.

Patient Characteristics and MA

Table 1 showsthe background characteristics of each group, and those of DCM MA(+) and of DCM MA(-) were compared. Digoxin had been prescribed to 3 out of the 9 DCM MA(+) patients, but to none of the DCM MA(-) patients, even though there was no statistical difference (p=0.090). The DCM MA(+) patients had lower pulmonary capillary wedge pressure (p=0.011) and lower left ventricular end diastolic pressure (p=0.009). Other characteristics including LVEFand basal left ventricular dP/dtdid not differ between the two DCM groups(p=0.79, and p=0.16, respectively) (Table 1, Fig. 2a,b).

Force Gain in Force-Frequency Relationship (FG-FFR) and MA

Left ventricular dP/dt increased to some extent as the pacing rate increased in most of the patients (Fig. 1a,c,e, arrows) and the force-frequency relationship (FFR) was obtained from each patient as shown in Fig.1b, d, and f with closed squares. The absolute values of the maximal dP/dt during incremental pacing did not differ between DCM MA(+) and DCM MA(-) patients (963 ± 199vs.973 ± 397 mmHg/sec, p=0.94) (Fig.2c). However the maximal increase of dP/dt (the force gain in FFR: FG-FFR) was significantly less in DCM MA(+) patients (35 ± 41vs.159 ± 104 mmHg, p=0.0091) (Fig.2d).The average LVEDP at FG-FFR measurement was 9.2 ± 5.9 mmHg in MA(+) patients, and 14.4 ± 8.3 mmHg in MA(-) patients (p=0.15).

Force Gain in Poststimulation Potentiation (FG-PSP) and MA

Poststimulation potentiation (PSP) (Fig. 1a,c,e, asterisks)was enhanced to some extent as the pacing rate increased in most of the patients (Fig. 1b,d,f, open triangles). The absolute value of maximaldP/dtdid not differ between DCM MA(+) and DCM MA(-) patients (1433 ± 263 vs. 1141 ± 367 mmHg/sec, p=0.085). But the increase from the basal condition (the force gain in PSP: FG-PSP) was significantly elevated in DCM MA(+) patientscompared withDCM MA(-) patients (500 ± 97 vs.325 ± 131 mmHg/sec, p=0.0017) (Fig.3a).The average LVEDP at FG-PSP measurement was 11.8 ± 2.4 mmHg in MA(+) patients, and 19.4 ± 9.3 mmHg in MA(-) patients (p=0.043).

Gap between FG-FFR and FG-PSP

As shown in Fig.1b, DCM MA(+) patientsseemed to have a wider gap between a high FG-PSP and a low FG-FFR, therefore this gap was examined. The gap was wider in DCM MA(+) patientscompared withDCM MA(-) patients (465 ± 119vs.163 ± 124 mmHg/sec, p=0.0001) (Fig.3b). Fig.3c shows that DCM MA(+) patients had both high FG-PSP and low FG-FFR levels. These properties of DCM MA(+) patients were quite different from those of the control patients.

The occurrence of MA at a single rapid pacing rate

We also examined whether the gap between dP/dt in PSP and dP/dt in FFR reflected the occurence of AA at a single rapid pacing rate, because FG-FFR,FG-PSP, andmax AAwere obtained separately during incremental pacing at different pacing rates.For example, FG-FFR and FG-PSP were obtained at different heart rates in 11 out of 17DCM cases, andin these 11 patientsdP/dt in FFR started to decline at a heart rate during incremental pacing, while dP/dt in PSP kept increasing until a higher pacing rate was reached.

The number of patients who showed MA at 110/min and 120/min were 5 and 6, respectively, and these numbers weremore than those at other heart rates. The 5 DCM patients with MA at 110/minhad a wider gap between dP/dt in PSP and dP/dt in FFR at 110/min than those without(375 ± 64 vs. 123 ± 98 mmHg/sec, p=0.0002, Fig.4a). The 6 DCM patients with MA at 120/min had a wider gap between dP/dt in PSP and dP/dt in FFR at 120/min than those without(423 ± 121 vs. 165 ± 133, p=0.0017, Fig.4b).

Discussion

Background characteristics of patients with MA

Although MA has been reported in patients with severe heart failure or LV dysfunction [2-6], a recent study ofDCM patients showed no statistical difference in LVEF between patients with MA and those without [2]. Our data also showed no correlation between LVEF and MAin DCM patients (Fig. 2a). It was obvious that among patients with low LVEF some had MA and some did not, and low LVEF was not sufficient to cause MA. In the present study, PCWP and LVEDP were significantly lower in DCM MA(+)than in DCM MA(-) patients. This seems inconsistent with the concept that MA is caused by heart failure and in fact other studies have shown that patients with MA had higher PCWP and LVEDP [2,3]. However, MA is also known to be induced by the standing posture [23] or by inferior vena caval occlusion [24] and can disappear with exacerbation of heart failure [5,6]. Thus far, the effect of preload on MA is still controversial. Effects of inotropic agents on MA are also controversial. Digoxin had been prescribed only for patients with MA in this study and dobutamine has been reported to increase the occurrence of MA [3,25]. On the other hand, attenuation or elimination of MA with digoxin and isoproterenol has been reported in patients and in dogs in the 1950’s and 1960’s [5,26,27]. These results show that inotropic agents affect MA but their effects are not straightforward.Considering that MA is induced by tachycardia, we can expect that heart rate-dependent parameters are more likely to correlate with MA than those basal hemodynamics or patient characteristics.

Rate-dependent contractile parameters of patients with MA

We showed that DCM MA(+) patients had smallerFG-FFR, largerFG-PSP, and a wider gap between themthan DCM MA (-) patients did. The relationship between FG-FFR and FG-PSP in patients with MA showeda clearcontrast compared with thecontrol cases (Fig. 3c). This implies that not only each of the two rate-dependent contractile properties, but also the balance between them play important roles in theoccurrence of MA. This may explain why the conditions that cause MA are not straightforward.

One may say that the smaller FG-FFR and the larger FG-PSP in MA(+) patients might have been caused by insufficient baseline preload, indicated by their significantly lower LVEDP and PCWP, and although not statistically significant, by theirsmaller LVEDV and LVESV than those of MA(-) patients. However, LVEDP at baseline and at FG-FFR measurement were 5.3± 1.9 mmHg and 9.2 ± 5.9 mmHg in MA(+) patients, and 13.5 ±10.6 mmHg and 14.4 ± 8.3 mmHg in MA(-) patients.The larger increase in LVEDP could not explain the smaller FG-FFR. LVEDP at FG-PSP measurement were 11.8 ± 4.2 mmHg (6.4 ± 4.6 mmHg increase from baseline) in MA(+) patients and 19.4 ± 9.3 mmHg (5.9 ± 4.6 mmHgincrease from base line) in MA(-) patients. The similar increases could not explain the larger FG-PSP in MA(+) patients. For precise evaluation of preload, left ventricular volume measurement is required in future studies.

One advantage of using the gap between the two types of FGs is that it offsets influences of the basal heart rate. When the basal heart rate is low, there will be ahigher chance to increase the force before incremental pacing reaches its endpoint, and as a result, to increase both FG-PSP and FG-FFR. Butthis bias is offset by using the gap. Furthermore as shown in Fig.4, the gap between dP/dt in PSP and dP/dt in FFR at a single rapid pacing rate can reveal a contractile property prone to cause MA without the need of incremental pacing.