The ventricular action potential (AP) shortens exponentially upon a progressive reduction of the preceding diastolic interval. Steep electrical restitution slopes have been shown to promote wavebreaks, thus contributing to electrical instability. The present study was designed to assess the predictive value of electrical restitution in hypokalemia-induced arrhythmogenicity. We recorded monophasic APs and measured effective refractory periods (ERP) at distinct ventricular epicardial and endocardial sites and monitored volume-conducted ECG at baseline and after hypokalemic perfusion (2.5 mM K+ for 30 min) in isolated guinea pig heart preparations. The restitution of AP duration measured at 90% repolarization (APD90) was assessed after premature extrastimulus application at variable coupling stimulation intervals, and ERP restitution was assessed by measuring refractoriness over a wide range of pacing rates. Hypokalemia increased the amplitude of stimulation-evoked repolarization alternans and the inducibility of tachyarrhythmias and reduced ventricular fibrillation threshold. Nevertheless, these changes were associated with flattened rather than steepened APD90 restitution slopes and slowed restitution kinetics. In contrast, ERP restitution slopes were significantly increased in hypokalemic hearts. Although epicardial APD90 measured during steady-state pacing (S1-S1 = 250 ms) was prolonged in hypokalemic hearts, the left ventricular ERP was shortened. Consistently, the epicardial ERP measured at the shortest diastolic interval achieved upon a progressive increase in pacing rate was reduced in the hypokalemic left ventricle. In conclusion, this study highlights the superiority of ERP restitution at predicting increased arrhythmogenicity in the hypokalemic myocardium. The lack of predictive value of APD90 restitution is presumably related to different mode of changes in ventricular repolarization and refractoriness in a hypokalemic setting, whereby APD90 prolongation may be associated with shortened ERP.
- cardiac arrhythmia
- ventricular refractoriness
the ventricular action potential (AP) shortens immediately upon an abrupt reduction in cardiac cycle length, an effect referred to as restitution (3). AP restitution has an important physiological significance as it allows the heart to preserve a sufficiently long diastolic interval (DI) at fast beating rates, thus maintaining an appropriate ventricular filling and coronary flow. Over the last decade, electrical restitution analysis, based on plotting the AP duration (APD) as a function of the preceding DI, has been widely implicated in cardiac arrhythmia research. Experimental and computational modeling studies have shown that a steep (greater than unity) slope of electrical restitution promotes persistent oscillations in APD (repolarization alternans) and conduction block, which precipitate a spontaneous breakup and multiple fragmentation of the reentrant wavefront, thereby initiating ventricular fibrillation (VF) (5, 11, 44, 45). Steep APD restitution slopes were found to be an important attribute of structural heart disease and may represent a substrate for cardiac electrical instability in human patients (20, 29, 38). In contrast, flattening of the APD restitution slope after antiarrhythmic drug administration may prevent or terminate VF, an effect ascribed to increased stability of activation wavefronts (11, 17, 47). Apart from the steepness of electrical restitution, the heterogeneity of APD restitution slopes determined at multiple ventricular recording sites is thought to be an important factor promoting VF due to creating large repolarization gradients (26, 29).
The predictive value of electrical restitution nevertheless has been challenged in studies demonstrating that sustained ventricular arrhythmias could be induced in the absence of steep APD restitution slopes (25, 47). Likewise, heart preparations showing greater than unity APD restitution slopes might not necessarily exhibit an increased vulnerability to VF (2, 6, 32). Numerical simulations have shown that steep APD restitution slopes reduce the vulnerable window of reentry via increased wavefront-waveback interactions (51). Optical mapping studies on rabbit ventricular muscle have revealed that a breakup of the reentrant excitation wavefront could occur in the absence of steep (≥1) AP restitution slopes (50). Taken together, these findings highlight the complexity of the mechanisms that influence the dynamic stability of the excitation wavefront, including preexisting anatomic heterogeneities (32, 51), wavefront curvature (4), the size of the central core of spiral wave (35), conduction velocity restitution (6, 47), and abnormal Ca2+ cycling (13, 31, 42).
Hypokalemia is a side effect of diuretic therapy contributing to increased ventricular arrhythmogenicity and sudden cardiac death in hypertensive patients (1, 7, 8). In the present study, we used the isolated, perfused guinea pig heart model to test the hypothesis that the arrhythmogenic potential of hypokalemia is linked to steeper electrical restitution slopes and/or increased heterogeneities in the distribution of restitution slopes across ventricular tissue. We also sought to determine if hypokalemia-induced arrhythmogenicity is predicted by APD restitution or effective refractory period (ERP) restitution or both. Indeed, the intermittent conduction blocks in VF occur primarily when the excitation wavefront encounters refractory tissue (44, 45), thus raising the possibility that ERP restitution may be superior to APD restitution at predicting dynamic instability. In the clinical setting, ERP restitution is easier to assess as it may be based on cardiac electrogram or surface ECG recordings without using a monophasic AP (MAP) catheter. Finally, hypokalemia has been shown to produce dissimilar changes in ventricular repolarization and refractoriness (33), thereby indicating that measuring both APD and ERP restitutions is necessary for the objective evaluation of arrhythmic susceptibility in the hypokalemic setting.
The present study complied with European Community Guidelines for the Care and Use of Experimental Animals and was approved by the Animal Ethics Screening Committee of the Panum Institute (clearance no.: 2007/561-1341). Female Dunkin-Hartley guinea pigs weighing 400–500 g were allowed to acclimatize to the housing conditions with free access to food and tap water for at least 7 days before being entered into the study.
Guinea pigs were anesthetized with pentobarbital sodium (50 mg/kg ip) and anticoagulated with heparin (1,000 IU/kg ip). The chest was opened, and the hearts were immediately excised and mounted on a Langendorff perfusion setup (Hugo Sachs Elektronik-Harvard Apparatus, March-Hugstetten, Germany). Hearts were retrogradely perfused via the aorta at a constant pressure of 55–60 mmHg with carefully filtered, warmed physiological saline solution saturated with 95% O2 and 5% CO2. The normokalemic perfusion solution contained (in mM) 118.0 NaCl, 4.7 KCl, 2.5 CaCl2, 25 NaHCO3, 1.2 KH2PO4, 1.2 MgSO4, and 10.0 glucose and had a pH of 7.4. Hypokalemic (2.5 mM K+) solutions used in experiments were prepared by reducing the quantity of KCl added.
Aortic perfusion pressure was measured with an ISOTEC pressure transducer connected to the aorta block of the setup. An ultrasonic flowmeter probe (Transonic Systems) and a thermocouple microprobe (Harvard Apparatus) were placed just above the aortic cannula to monitor the coronary flow rate (10–15 ml/min) and the temperature of the perfusion solution (37 ± 0.2°C), respectively. The electrical activity of the heart preparations was assessed from the volume-conducted ECG as well as MAP recordings. Throughout the experiments, heart preparations were kept immersed in the temperature-controlled, perfusate-filled chamber to enable ECG recording and to minimize thermal loss.
A platform carrying four Ag-AgCl ECG electrodes was fitted in a mounting ring placed in the perfusate-filled chamber just below the heart preparation. The volume-conducted ECG was recorded from simulated bipolar Einthoven leads I, II, and III and unipolar Goldberger leads aVL, aVR, and aVF. Electrical stimulation-induced VF was defined as tachyarrhythmia exhibiting completely disorganized ventricular electric activity with a mean cycle of <70 ms (43).
90) during steady-state pacing and programmed ventricular stimulation and at 50% repolarization (APD50) during repolarization alternans. APD0 was defined as the peak of the upstroke, and APD100 was measured at the diastolic potential level. LV and RV epicardial APD90 were calculated as the mean values of three simultaneous MAP recordings taken from each ventricular chamber. The mean ventricular epicardial APD90 was determined as the mean value of all six simultaneous MAP recordings.
To assess the refractoriness and arrhythmic susceptibility of cardiac preparations, electrical stimulations were performed using custom-made bipolar needle electrodes placed on the anterolateral surface of the basal LV and RV epicardium. Endocardial stimulations were accomplished using a bipolar stick-shaped stimulating electrode advanced toward the LV endocardial base via a small incision made in the left atrium. Heart preparations were stimulated with 2-ms rectangular pulses delivered at twice diastolic threshold current using the programmable stimulator (Hugo Sachs Electronik-Harvard Apparatus).
To measure the VF threshold, a burst of 50 pulses and an interpulse interval of 20 ms was applied while the current intensity was progressively increased from an initial value of 5 mA. In successive stimulations, the current intensity was increased with an increment of 1 mA until VF was induced. The VF threshold was defined as the minimal current intensity inducing VF during normokalemic and hypokalemic perfusion as measured at both LV epicardial and LV endocardial stimulation sites. To allow repeated measurements, VF episodes persisting >20 s were terminated by a bolus injection of highly concentrated KCl solution (2 ml, 50 mM) just above the aortic cannula.
Rapid cardiac pacing.
A series of LV epicardial and LV endocardial stimulations was performed to apply a burst of 100 pulses (current intensity of twice diastolic threshold) at a progressively increased pacing rate. For this purpose, the interpulse pacing interval in successive stimulations was reduced by 10 ms starting from the initial value of 250 ms. The protocol was continued until VF was induced or the minimum value of the interpulse interval (10 ms) was achieved. The repolarization alternans preceding VF induction was defined as regular beat-to-beat oscillations in APD whereby the APD50 difference in each pair of long-short APs was ≥ 5 ms. To determine the arrhythmic susceptibility of heart preparations, we assessed the inducibility of repolarization alternans and VF, the amplitude of repolarization alternans, and the degree of shortening in pacing cycle length needed to induce repolarization alternans and VF in normokalemic and hypokalemic conditions.
Electrical restitution stimulation protocols.
The AP restitution and ERP restitution protocols used in the present study are shown in the Supplemental Material.1
APD90 restitution was determined at discrete ventricular epicardial and endocardial recording sites using the standard S1-S2 epicardial pacing protocol. The LV was continuously paced at a cycle length of 250 ms (S1-S1 interval), and a premature extrastimulus (S2) was delivered to the pacing site after each 50 beats while the S1-S2 coupling interval was progressively reduced in 2- to 10-ms decrements from 240 ms until refractoriness was reached (Supplemental Material, section A).
ERP restitution was assessed by measuring refractoriness over wide range of pacing intervals at LV epicardial, LV endocardial, and RV epicardial stimulation sites (Supplemental Material, section B). To enable measurements at long cycle lengths, the right atrium was removed, and the atrioventricular (AV) node was mechanically crashed with forceps to slow down the intrinsic beating rate. The presence of complete AV block was verified by ECG recordings (n = 20) showing prolongation of the R-R interval above 500 ms (basal: 314 ± 5 ms and AV blocked: 543 ± 10 ms, P < 0.0001) and the absence of ventricular responses after left atrial stimulation. The ERP was measured by an extrastimulus method whereby a burst of 15 regular (S1) pulses was followed by an extrastimulus (S2) applied at variable coupling stimulation intervals. The measurements were started with a S1-S1 cycle length of 500 ms, which was then reduced in 20-ms steps to 200 ms and thereafter with a decrement of 5–10 ms until the capture was lost at pacing intervals of 140–160 ms. At variable pacing cycle lengths, an extrastimulus application in late diastole was avoided to keep the whole ERP restitution protocol reasonably short. For this purpose, after each step of S1-S1 reduction, an extrastimulus application was started with a S1-S2 coupling stimulation interval equal to ERP measured at the previous pacing cycle length. If an extrastimulus failed to elicit a propagating response, the S1-S2 interval was increased by 2–5 ms to ensure capture at longer coupling stimulation intervals. Otherwise, the S1-S2 interval was reduced by 2–5 ms to determine if capture was lost at shorter coupling stimulation intervals. At all pacing cycle lengths used, the ERP was defined as the longest S1-S2 interval at which a premature extrastimulus failed to elicit a propagating response (Supplemental Material, section B).
Electrical restitution analysis.
Once the S1-S2 stimulation protocol was completed, appropriate DIs were calculated as the difference between S1-S2 coupling stimulation intervals and APD90 determined at each ventricular recording site during steady-state pacing. For ERP restitution, DIs were found as a difference between the pacing cycle length used and ERP measured. APD90 and ERP were then plotted as a function of the preceding DI, and the restitution curves were fitted using the following double-exponential function: y = y0 + A1e(−DI/τ1)+ A2e(−DI/τ2), where y is APD90, y0 is a free-fitting variable, and A1 and A2 are the amplitudes and τ1 and τ2 are the time constants of the fast (A1 and τ1) and slow (A2 and τ2) exponential components obtained by a least-squares fit. Curve fitting was performed using Igor Pro 6.0 software (WaveMetrics, Portland, OR).
In each experiment, the restitution curves based on either APD90 measurements at distinct ventricular recording sites or ERP measurements at distinct ventricular stimulation sites after normokalemic and hypokalemic perfusion were analyzed to determine the maximal restitution slope and the range of APD90 or ERP changes (ΔAPD90 and ΔERP, the maximum to minimum change) yielded by the stimulation protocol (Supplemental Material, sections C and D). The maximal slope of electrical restitution was found by analyzing the first derivative of the exponential fit (Supplemental Material, section D). Electrical restitution kinetics were assessed by an empirical rate constant (21) calculated as a ratio between ΔAPD90 or ΔERP and the range of DIs covered by the stimulation protocol (ΔDI, the difference between the longest and shortest DI).
Spatial variability of electrical restitution slopes.
The variability of APD90 restitution slopes across the ventricular epicardium was quantified by a coefficient calculated as a percent ratio of SD and the average restitution slope value determined at six ventricular recording sites in each experiment. The transepicardial variability of ERP restitution slopes was assessed as the difference between the maximal slope values determined at RV epicardial and LV epicardial stimulation sites.
The variability of APD90 restitution slopes across the LV wall was assessed as the difference in maximum slope values determined at LV endocardial and opposite LV epicardial recording sites. The variability of ERP restitution slopes across the LV wall was assessed as the difference in maximum slope values determined at LV endocardial and opposite LV epicardial stimulation sites.
+) perfusion. Three groups of experiments were performed to assess 1) VF threshold and the inducibility of repolarization alternans and VF by tachypacing (n = 11), 2) APD90 restitution (n = 12), and 3) ERP restitution (n = 20) immediately before and at the end of hypokalemic perfusion.
t-tests were used to compare the variables determined at baseline and after 30 min of hypokalemic perfusion. Unpaired t-tests were used to compare the maximum restitution slopes and restitution rate constants determined at distinct ventricular recording sites (APD90 restitution) or stimulation sites (ERP restitution). The inducibility of repolarization alternans and VF by rapid cardiac pacing at baseline and after hypokalemic perfusion was compared using a two-tailed Fisher's exact test. P values of <0.05 were considered to be significant.
RV-to-LV transepicardial heterogeneities of electrical restitution in normokalemic hearts.
A representative example of electrical restitution after LV epicardial stimulation is shown in the Supplemental Material, and the composite data are shown in Fig. 1. A reduction of DIs achieved either by moving an extrastimulus application toward earlier time points in diastole (Supplemental Material, section A) or progressive shortening of regular pacing intervals (Supplemental Material, section B) was associated with exponential abbreviation of APD90 (Fig. 1A) and ERP (Fig. 1E). However, the electrical restitution properties as assessed by these stimulation protocols were nonuniform at distinct ventricular sites, thereby suggesting spatial heterogeneities determined both in RV-to-LV transepicardial and LV transmural (epicardial to endocardial) planes.
During steady-state pacing, RV epicardial APD90 was longer than LV APD90 [LV (n = 12): 114 ± 2 ms and RV (n = 12): 122 ± 2 ms, P = 0.008]. Consistently, the RV APD90 restitution curve was shifted upward, suggesting longer APD90 at comparable DIs in the RV epicardium than in the LV epicardium (Fig. 1A). RV epicardial recording sites showed a greater amplitude of APD90 change during the restitution protocol [LV ΔAPD90 (n = 12): 32 ± 1 ms and RV ΔAPD90 (n = 12): 39 ± 1 ms, P = 0.0003], greater maximum restitution slopes (Fig. 1, B and C), and faster restitution kinetics (Fig. 1D).
The RV-to-LV transepicardial heterogeneities in electrical restitution were also revealed by ERP restitution protocol (Fig. 1, E–H). The RV epicardial stimulation site showed a greater amplitude of ERP change [LV ΔERP (n = 10): 35 ± 4 ms and RV ΔERP (n = 10): 46 ± 3 ms, P = 0.04], greater maximum restitution slopes (Fig. 1, F and G), and faster restitution kinetics (Fig. 1H).
LV epicardial-to-LV endocardial heterogeneities of electrical restitution in normokalemic hearts.
During steady-state pacing, similar APD90 values were determined at LV epicardial and LV endocardial recording sites [LV epicardial (n = 12): 117 ± 3 ms and LV endocardial (n = 12): 115 ± 3 ms, P = 0.53]. Nevertheless, the maximum restitution slope was significantly higher at the LV endocardial site than at the LV epicardial recording site (Fig. 2, B and C). Consistently, the maximum ERP restitution slope was greater at the LV endocardial site than at the LV epicardial stimulation site (Fig. 2, F and G). The LV endocardial stimulation site showed a greater amplitude of ERP change during the restitution protocol [LV epicardial ΔERP (n = 10): 35 ± 4 ms and LV endocardial ΔERP (n = 10): 51 ± 2 ms, P = 0.005], an effect contributing to the greater restitution rate constant at the LV endocardium (Fig. 2H).
Hypokalemia-induced arrhythmogenicity as assessed by tachypacing.
At baseline, a progressive reduction in pacing cycle length was associated with a gradual shortening of the ventricular APD until 2:1 conduction block was elicited at interpulse pacing intervals shorter than the ERP (Fig. 3A). In 40–50% of the heart preparations, conduction failure was preceded by repolarization alternans whereby APD50 exhibited regular beat-to-beat oscillations once the LV was subjected to fast (S1-S1 of <100 ms) pacing rates (Fig. 3A, fragment 5). Hypokalemia increased the inducibility of repolarization alternans (Fig. 3C) and increased the amplitude of APD50 change in each pair of long-short APs after LV epicardial [basal (n = 4): 8 ± 1 ms and hypokalemic (n = 10): 11 ± 1 ms, P = 0.04] and LV endocardial [basal (n = 6): 8 ± 2 ms and hypokalemic (n = 11): 14 ± 1 ms, P = 0.02] stimulations. Furthermore, slower pacing rates were sufficient to induce repolarization alternans in hypokalemic compared with normokalemic conditions. The longest pacing interval at which repolarization alternans was induced was significantly increased in hypokalemic heart preparations after LV epicardial [basal (n = 4): 90 ± 4 ms and hypokalemic (n = 10): 99 ± 1 ms, P = 0.04] and LV endocardial [basal (n = 6): 57 ± 14 ms and hypokalemic (n = 11): 97 ± 8 ms, P = 0.03] stimulation.
Importantly, an enhancement in the amplitude of repolarization alternans was followed by VF upon a further increase in the LV pacing rate (Fig. 3B, fragments 4–6). Hypokalemia increased VF inducibility (Fig. 3D) and enabled VF induction at a slower pacing stimulation rate. The pacing intervals to induce tachyarrhythmia were significantly longer in hypokalemic heart preparations after LV epicardial [basal (n = 2): 73 ± 3 ms and hypokalemic (n = 8): 84 ± 2 ms, P = 0.02] and LV endocardial [basal (n = 5): 40 ± 11 ms and hypokalemic (n = 10): 77 ± 8 ms, P = 0.02] stimulation.
Hypokalemia reduced VF thresholds as measured at LV epicardial [basal (n = 10): 15 ± 2 mA and hypokalemic (n = 10): 10 ± 2 mA, P = 0.04] and LV endocardial [basal (n = 10): 7 ± 1 mA and Hypokalemic (n = 10): 4 ± 1 mA, P = 0.04] stimulation sites.
Epicardial APD90 restitution.
During steady-state pacing, mean ventricular epicardial APD90 was significantly prolonged after hypokalemic perfusion [basal (n = 12): 117 ± 2 ms and hypokalemic (n = 12): 124 ± 2 ms, P = 0.01]. Nevertheless, the LV epicardial ERP measured at the pacing cycle length (250 ms) used for the APD90 restitution protocol was reduced in hypokalemic heart preparations [basal (n = 12): 100 ± 2 ms and hypokalemic (n = 12): 89 ± 3 ms, P = 0.006].
Hypokalemic perfusion reduced the degree of epicardial APD90 shortening after a premature extrastimulus application, an effect especially pronounced at short DIs preceding the ERP (Fig. 4, A and G). Consequently, the APD90 restitution curves determined at the LV epicardium and RV epicardium were shifted upward and flattened. The flattening of the curves in hypokalemic heart preparations was evidenced by a reduction of APD90 restitution slopes over a wide range of DIs [LV (Fig. 4B) and RV (Fig. 4H)]. Importantly, the maximal slope of epicardial APD90 restitution was reduced in both ventricular chambers [LV (Fig. 4C) and RV (Fig. 4I)]. Hypokalemia slowed the APD90 restitution kinetics as evidenced by a reduced restitution rate constant at the LV epicardium [basal (n = 12): 0.25 ± 0.01 and hypokalemic (n = 12): 0.22 ± 0.01, P = 0.02] and RV epicardium [basal (n = 12): 0.31 ± 0.02 and hypokalemic (n = 12): 0.24 ± 0.02, P = 0.01].
LV endocardial APD90 restitution.
LV endocardial APD90 as measured during steady-state pacing was not changed after hypokalemic perfusion [basal (n = 12): 115 ± 3 ms and hypokalemic (n = 12): 116 ± 4 ms, P = 0.42]. Nevertheless, the degree of LV endocardial APD90 shortening after an extrastimulus application at short DIs preceding the ERP was reduced in hypokalemic heart preparations (Fig. 4D). This change contributed to the flattening of the restitution curve as evidenced by reduced APD90 restitution slopes over DIs of <40 ms (Fig. 4E). Consistently, both the maximum LV endocardial APD90 restitution slope (Fig. 4F) and restitution rate constant [basal (n = 12): 0.27 ± 0.03 and hypokalemic (n = 12): 0.20 ± 0.03, P = 0.04] were reduced after hypokalemic perfusion.
In contrast to the flattened APD90 restitution, ERP restitution slopes were significantly increased in hypokalemic heart preparations (Fig. 5). This change was consistently observed at LV epicardial (Fig. 5, B and C), LV endocardial (Fig. 5, E and F), and RV epicardial (Fig. 5, H and I) stimulation sites.
After a progressive reduction in the LV epicardial pacing cycle length, the minimal pacing interval at which the capture was preserved was significantly reduced after hypokalemic perfusion [basal (n = 10): 161 ± 3 ms and hypokalemic (n = 10): 143 ± 4 ms, P = 0.0003], an effect indicating ERP reduction in the hypokalemic LV. Consistently, the minimal LV epicardial ERP [i.e., the ERP achieved at the shortest DI used (Fig. 5A)] was significantly reduced in hypokalemic heart preparations [basal (n = 10): 124 ± 2 ms and hypokalemic (n = 10): 112 ± 2 ms, P = 0.002]. This effect was associated with an increased amplitude of LV epicardial ERP change during the restitution protocol [basal ΔERP (n = 10): 36 ± 3 ms and hypokalemic ΔERP (n = 10): 46 ± 2 ms, P = 0.01] and increased restitution rate constant [basal (n = 10): 0.13 ± 0.01 and hypokalemic (n = 10): 0.15 ± 0.007, P = 0.04].
Neither the amplitude of ERP change nor the restitution rate constant was changed upon hypokalemic perfusion at LV endocardial and RV epicardial stimulation sites (data not shown).
Spatial variability of electrical restitution slopes.
Hypokalemia proportionately reduced the mean epicardial APD90 restitution slope and its SD, thereby producing no effect on transepicardial slope variability (Table 1). Consistently, the difference in ERP restitution slopes measured at RV epicardial and LV epicardial stimulation sites was not changed after hypokalemic perfusion [basal (n = 10): 0.19 ± 0.11 and hypokalemic (n = 10): 0.14 ± 0.17, P = 0.91].
The LV endocardial-to-LV epicardial difference in the maximal APD90 restitution slope was not changed after hypokalemic perfusion [basal (n = 12): 0.57 ± 0.23 and hypokalemic (n = 12): 0.31 ± 0.12, P = 0.28]. With ERP restitution, the relative increase in maximum restitution slope was greater in the hypokalemic LV epicardium compared with the LV endocardium (Fig. 5, C and F), an effect contributing to the reduced endocardial-to-epicardial difference in ERP restitution slope in the hypokalemic LV [basal (n = 10): 0.38 ± 0.1 and hypokalemic (n = 10): −0.18 ± 0.27, P = 0.04].
Previous studies have shown no consistent effect of hypokalemia on ventricular AP restitution. In canine endocardial muscle, the dynamic protocol of APD90 restitution revealed no change of restitution properties while the extracellular K+ concentration was varied between 2.7 and 4.0 mM (19). The inducibility of tachyarrhythmias in the hypokalemic setting was not determined in that study. On the other hand, hypokalemia was found to increase the inducibility of tachyarrhythmias by rapid cardiac pacing and increase the maximal slope of AP restitution in isolated, perfused murine hearts (34). These findings, however, may not be extrapolated to the hearts of large mammalian species due to fast beating rates and differences in ionic currents contributing to ventricular repolarization in mice (14).
In the present study, we show that the predictive value of electrical restitution in hypokalemia-induced arrhythmogenicity critically depends on whether APD90 or ERP is chosen as the variable reflecting the dynamic changes of ventricular repolarization upon a progressive reduction of DI. Although hypokalemia contributed to the flattened restitution slopes and slowed restitution kinetics after APD90 restitution assessments, ERP restitution exhibited steeper slopes and faster kinetics. The latter changes were most pronounced at the LV epicardial stimulation site, and they may account for the increased inducibility of repolarization alternans and VF by rapid cardiac pacing, increased amplitude of repolarization alternans, and reduced VF threshold in the hypokalemic LV. Dissimilar APD90 and ERP restitution changes in the hypokalemic setting may not be attributed to the inherent technical limitations of the stimulation protocols used [e.g., to impact of past S1-S1 pacing history on S2-evoked APD90 change (18)] as both restitution protocols were found to have an equal value at detecting the preexisting spatial heterogeneities in electrical restitution in normal hearts. Instead, the lack of similarities in APD90 and ERP restitution kinetics in the hypokalemic setting presumably reflects the different nature of APD90 and ERP changes in hypokalemia whereby prolongation of the ventricular epicardial AP may be associated with shortened LV refractoriness.
Computer modeling studies have shown that a breakup and multiply fragmentation of the spiral wave is promoted by oscillations of the APD (repolarization alternans) once their amplitude reaches a critical value at which a local conduction block is induced at some point of the excitation wavefront (5). According to the restitution hypothesis, persistent repolarization alternans leading to VF may be induced only in heart preparations showing sufficiently steep (≥1) electrical restitution slopes (44, 45). Once the heart is paced at a constant cycle length, a sudden shortening of the AP in one beat contributes to the prolongation of the following DI and, therefore, the increased APD in the next beat. This, in turn, is followed by a reduced DI and hence shortened AP, thereby initiating repolarization alternans. If the relationship between APD and the preceding DI is characterized by a steep (≥1) restitution slope, then the amplitude of repolarization alternans is incrementally amplified over successive cardiac cycles, which finally precipitates VF. In contrast, low (<1) restitution slopes prevent the induction of persistent repolarization alternans because the initial (e.g., stimulation evoked) APD change will be dampened over next cardiac cycles due to a flat relationship between the APD and preceding DI.
Alternatively, the genesis of repolarization alternans has been linked to restitution-independent mechanisms such as abnormal Ca2+ handling at fast beating rates (13, 31, 42). Depletion of sarcoplasmic reticulum Ca2+ stores by thapsigargin and ryanodine or chelation of intracellular Ca2+ by BAPTA was found to raise the threshold for alternans induction (13, 42). Nevertheless, the amplitude of repolarization alternans is significantly reduced once stimulation protocols that prevent beat-to-beat oscillations in the DI are used (constant DI pacing) (48), thereby highlighting the importance of restitution-related APD changes.
Consistent with the restitution hypothesis, we found that an increased inducibility of tachypacing-evoked VF in the hypokalemic LV is associated with a greater amplitude of repolarization alternans as well as highly increased electrical restitution slopes, at least when assessed by ERP dynamics. Ng et al. (27) found a strong negative correlation between VF threshold values and the steepness of electrical restitution slope after autonomic nerve stimulations in isolated, perfused rabbit hearts. In line with these findings, we found that reduced VF thresholds as measured at epicardial and endocardial stimulation sites in the hypokalemic LV may be accounted for by an increased steepness of electrical restitution, at least when assessed by ERP dynamics.
Intrinsic spatial heterogeneities in APD90 restitution.
Global assessment of AP restitution using multiple ventricular recording sites revealed preexisting spatial heterogeneities in the guinea pig (21), rabbit (30), and human heart (26, 52). High-resolution optical mapping of the guinea pig heart has demonstrated that over the anterior LV wall, epicardial recording sites showing longer APD90 during steady-state pacing exhibit steeper restitution slopes and faster restitution kinetics after a premature stimulation compared with recording sites with shorter basal APD90 (21). In the present study, we extend these observations to show that spatial restitution heterogeneities in the guinea pig heart exist in both the RV-to-LV transepicardial plane and transmurally across the LV wall. Steeper electrical restitution slopes and faster restitution kinetics were found at the RV epicardium compared with the LV epicardium as well as at the LV endocardium compared with the LV epicardium. These differences were consistently revealed both by APD90 and ERP restitution protocols. Spatial heterogeneities in electrical restitution could be attributed to a nonuniform distribution of K+ channels in ventricular tissue. In the guinea pig, the density of the inward rectifier K+ current (IK1) was found to be greater in the LV myocardium compared with the RV myocardium (40, 43). On the other hand, peak outward K+ current and the density of the slow component of the delayed rectifier K+ current (IKs) and transient outward K+ current (Ito) are greater in the RV compared with the LV in the dog (10, 41). Transmurally, the greater density of the rapid component of the delayed rectifier K+ current (IKr) has been found in epicardial myocytes compared with midmyocardial and endocardial myocytes dissociated from the guinea pig LV (23).
Increased spatial heterogeneities in APD90 restitution are thought to promote ventricular arrhythmogenesis irrespectively from the steepness of restitution slopes determined at each individual ventricular site (26, 29). This mechanism, however, could not account for the hypokalemia-induced arrhythmogenicity in the present study. Indeed, both APD90 and ERP restitution protocols revealed no change in the RV-to-LV transepicardial variability of restitution slopes in hypokalemic heart preparations, whereas the LV epicardial-to-LV endocardial slope difference was either unchanged (APD90 restitution) or even reduced (ERP restitution) when assessed by the two different electrical restitution protocols.
A premature depolarization recruits less inward currents due to incomplete recovery of Na+ current and Ca2+ current from inactivation caused by previous excitation. Furthermore, due to the slow time decay of outward K+ currents contributing to ventricular repolarization, an extrastimulus applied shortly after previous regular beat activates more IKr and IKs. Taken together, these changes account for the abbreviation of the ventricular AP after a premature excitation (3). Hypokalemia has been shown to inhibit outward K+ currents such as IKr and IK1 (16, 36–37), an effect contributing to the prolongation of ventricular APD90 during steady-state pacing and less shortening of APD90 after an extrastimulus application at a given DI. The reduced shortening of the premature AP was especially pronounced over the range of short coupling intervals preceding refractoriness, thereby flattening the slope of the initial portion of the restitution curve (Fig. 4).
Computational modeling studies have shown that in simulated guinea pig ventricular myocytes, the activation of the delayed rectifier K+ current is the most important determinant of APD90 restitution over a range of short coupling intervals whereby the rate of APD90 change depends on the IKr-to-IKs ratio (53). An increase in the IKr-to-IKs ratio contributes to the greater shortening of ventricular repolarization in response to a given reduction of the preceding DI, thereby increasing the maximal APD90 restitution slope (53). Importantly, hypokalemia inhibits IKr while increasing IKs in isolated guinea pig ventricular myocytes (36), changes contributing to the reduced IKr-to-IKs ratio. This effect presumably accounts for the reduced maximum slopes of APD90 restitution in hypokalemic heart preparations observed in the present study.
In VF, local conduction block precipitating a wavebreak occurs primarily when the excitation wavefront encounters refractory tissue (44, 45). Steep ERP restitution precipitates tachypacing-induced VF due to spatiotemporal oscillations in cardiac activation, which progressively increases the amplitude upon a reduction in the pacing interval (5). Computer modeling studies have shown that ERP restitution parallels APD90 restitution after simulated changes in the dynamic stability of the spiral wave in cardiac tissue (49). Consistently, in the present study, we found that APD90 and ERP restitution protocols are equally effective at detecting preexisting transepicardial and LV transmural heterogeneities in electrical restitution in the normokalemic guinea pig heart. Nevertheless, these protocols yielded different outcomes once the arrhythmic susceptibility of heart preparations was increased upon hypokalemic perfusion. In contrast to the flattened APD90 restitution, ERP restitution slopes were consistently increased at distinct ventricular stimulation sites.
These discrepancies are presumably related to dissimilar changes in ventricular repolarization and refractoriness in hypokalemic heart preparations. Indeed, the standard protocol of APD90 restitution (S1-S1 = 250 ms) has revealed that although mean ventricular epicardial APD90 is prolonged, the LV epicardial ERP is reduced in hypokalemic heart preparations. Consistently, the minimal epicardial ERP (i.e., the ERP achieved at the shortest DI used) was significantly shorter in the hypokalemic LV compared with the normokalemic LV (Fig. 5A). These changes enabled capture at shorter S1-S1 pacing intervals while the ERP restitution was assessed in the hypokalemic LV, an effect contributing to the increased amplitude of the ERP restitution curve and hence faster restitution kinetics.
Dissimilar changes in APD and refractoriness have been previously shown in isolated, perfused murine hearts (33), whereby APD90 prolongation was associated with shortened ERP in the hypokalemic setting. These changes in ERP may be accounted for by the hypokalemia effect on the Na+-carrying system, which determines the duration of ventricular refractoriness. Hypokalemia has been shown to hyperpolarize the resting membrane potential in isolated guinea pig ventricular myocytes (46), an effect associated with an increased maximal velocity of the AP upstroke (Vmax) and enhanced amplitude of the ventricular AP (12, 24). These changes suggest an increased availability of fast Na+ channels for activation. In rabbit ventricular tissue, the threshold concentration of lidocaine, a Na+ channel blocker, required to reduce the Vmax is increased by 10-fold in the hypokalemic setting (39). The increased availability of fast Na+ channels for activation may contribute to the recurrence of excitability at earlier time points during ventricular repolarization. In support of this argument, a recovery from refractoriness was found to occur at less negative potentials in hypokalemic guinea pig papillary muscle, an effect abolished by a Na+ channel blocker (22). Taken together, these changes could account for the reduced ERP duration at short DIs and hence the increased steepness of ERP restitution in the hypokalemic setting. Alternatively, the reduced ERP duration in the hypokalemic LV may be attributed to alterations in the voltage profile of the ventricular AP (triangulation) (28) as well as changes in ventricular excitability resulting from an inhibition of background outward K+ current (9) or a low K+ effect on inactivation of fast Na+ channels (15).
Steep ERP restitution slopes and faster restitution kinetics may account for the increased amplitude of repolarization alternans, increased inducibility of stimulation-evoked ventricular tachyarrhythmias, and reduced VF threshold in hypokalemic heart preparations. These changes, however, are overlooked when the APD90 restitution protocol was used. The lack of similarities in the predictive value of these restitution protocols is presumably attributed to different nature of changes in ventricular repolarization and refractoriness in the hypokalemic setting, whereby a prolongation of APD90 is associated with shortened ERP.
This work was supported by the Danish National Research Foundation.
No conflicts of interest are declared by the author(s).
↵ 1 Supplemental Material for this article is available online at the American Journal of Physiology-Heart and Circulatory Physiology website.
- Copyright © 2010 the American Physiological Society