Tachy-brady arrhythmias: The critical role of adenosine-inducedsinoatrial conduction block in post-tachycardia pauses

Qing Lou, PhD,* Alexey V. Glukhov, PhD,* Brian Hansen,* Lori Hage, BS,* Pedro Vargas-Pinto, DVM, MSc,†

George E. Billman, PhD, FHRS,* Cynthia A. Carnes, PharmD, PhD, FHRS,† Vadim V. Fedorov, PhD*

From the *Department of Physiology and Cell Biology, The Ohio State University, Columbus, Ohio and yCollege of Pharmacy,Davis Heart and Lung Research Institute, The Ohio State University, Columbus, Ohio.

BACKGROUND In patients with sinoatrial nodal (SAN) dysfunc-tion, atrial pauses lasting several seconds may follow rapid atrialpacing or paroxysmal tachycardia (tachy-brady arrhythmias).Clinical studies suggest that adenosine may play an important rolein SAN dysfunction, but the mechanism remains unclear.

OBJECTIVE To define the mechanism of SAN dysfunctioninduced by the combination of adenosine and tachycardia.

METHODS We studied the mechanism of SAN dysfunction producedby a combination of adenosine and rapid atrial pacing in isolatedcoronary-perfused canine atrial preparations by using high-resolution optical mapping (n ¼ 9). Sinus cycle length andsinoatrial conduction time (SACT) were measured during adenosine(1–100 mM) and DPCPX (1 mM; A1 receptor antagonist; n ¼ 7)perfusion. Sinoatrial node recovery time was measured after 1minute of “slow” pacing (3.3 Hz) or tachypacing (7–9 Hz).

RESULTS Adenosine significantly increased sinus cycle length(477 � 62 ms vs 778 � 114 ms; P o .01) and SACT during sinusrhythm (41 � 11 ms vs 86 � 16 ms; P o .01) in a dose-dependentmanner. Adenosine dramatically affected SACT of the first SAN beatafter tachypacing (41 � 5 ms vs 221 � 98 ms; P o .01). Moreover,at high concentrations of adenosine (10–100 mM), terminationof tachypacing or atrial flutter/fibrillation produced atrialpauses of 4.2 � 3.4 seconds (n ¼ 5) owing to conduction

This study was supported by Davis Heart and Lung Research Institute (toDr Fedorov) and National Institutes of Health (grants R01 NHLBI/HL089836 [to Dr Carnes] and NHLBI/HL086700 [to Dr Billman]). Addressreprint requests and correspondence: Dr Vadim V. Fedorov, PhD,Department of Physiology and Cell Biology, College of Medicine,The Ohio State University, 306 Hamilton Hall, 1645 Neil Avenue,Columbus, OH 43210-1218. E-mail address: vadim.fedorov@osumc.edu;fedorov.2@osu.edu.

The first 2 authors contributed equally to the study.

1547-5271/$-see front matter B 2013 The Heart Rhythm Society. All rights reser

block between the SAN and the atria, despite a stable SANintrinsic rate. Conduction block was preferentially relatedto depressed excitability in SAN conduction pathways.Adenosine-induced changes were reversible on washout orDPCPX treatment.

CONCLUSIONS These data directly demonstrate that adenosinecontributes to post-tachycardia atrial pauses through SAN exitblock rather than slowed pacemaker automaticity. Thus, these datasuggest an important modulatory role of adenosine in tachy-bradysyndrome.

KEYWORDS Sinoatrial node; Optical mapping; Adenosine; Atrialflutter/fibrillation; Tachy-brady syndrome

ABBREVIATIONS AF¼ atrial fibrillation; AFL¼ atrial flutter; AP¼action potential; APD¼ action potential duration; APD80%¼ APDat 80% repolarization; cSNRT ¼ corrected sinoatrial node recoverytime; DF ¼ dominant frequency; OAP ¼ optical action potential;SACP ¼ sinoatrial node conduction pathway; SACT ¼ sinoatrialconduction time; SAN¼ sinoatrial node; SCL ¼ sinus cycle length;SNRT ¼ sinoatrial node recovery time; SNRTi ¼ indirect SNRT;SNRTd ¼ direct SNRT; SNRTr ¼ real SNRT; SR ¼ sinus rhythm

(Heart Rhythm 2013;10:110–118) I 2013 The Heart RhythmSociety. All rights reserved.

IntroductionSinoatrial node (SAN), the primary pacemaker of the humanheart, is a specialized and complex structure.1–3 Dysfunctionof the SAN leads to more than 50% (4100,000) of theannual pacemaker implants in the United States.4 One of themain manifestations of SAN dysfunction is tachy-bradysyndrome, characterized as the heart rate alternating between

too fast and too slow.5 In patients with SAN dysfunction,termination of rapid pacing or paroxysmal tachycardia maybe followed by long atrial pauses lasting several seconds,6–8

which can provoke another tachyarrhythmia paroxysm.However, the cause of the post-tachycardia pause remainselusive.

One possible reason for the pause could be the effect ofadenosine, an endogenous metabolite of the heart.9 In 1929,Drury and Szent-Gyorgyi10 demonstrated for the first timethat adenosine significantly slowed sinus rhythm (SR),produced atrioventricular block, and facilitated both atrialflutter (AFL) and atrial fibrillation (AF) by shortening therefractory period. In 1985, Watt11 hypothesized thatincreased endogenous production of adenosine and/or hyper-sensitivity to adenosine could result in SAN dysfunction,particularly tachy-brady syndrome. This is supported by the

ved. http://dx.doi.org/10.1016/j.hrthm.2012.09.012

Lou et al Adenosine and Tachy-Brady Syndrome 111

clinical findings that (1) bolus injection of adenosinesuppresses SAN function and produces pauses especiallyin patients with SAN dysfunction12,13 and (2) orally admini-strated theophylline, a potent nonselective antagonist ofadenosine receptors, reduces both the frequency and theduration of the pauses in patients with sick sinus syndrome.14

Despite abundant evidence of a suppressive effect ofadenosine in the SAN, the causal relationship betweenadenosine and post-tachycardia pauses has not been experi-mentally demonstrated. Furthermore, owing to a lack ofdirect clinical mapping data from the SAN, the mechanismby which adenosine leads to atrial pauses has not beendetermined. The atrial pauses during tachy-brady syndromecould result from (1) poor function of the SAN as an impulsegenerator (reduced automaticity and sinus arrest) or (2)conduction block of the generated pulses from the SAN tothe atria (SAN exit block). In the present study, we proposethat adenosine induces post-tachycardia atrial pauses viasuppression of SAN conduction rather than by slowingpacemaker automaticity. To test this hypothesis, we usedhigh-resolution multistructural near-infrared fluorescenceoptical imaging to map functionally the coronary-perfusedcanine SAN preparation.

MethodsAll animal procedures and protocols (n ¼ 9 dogs; 9.9 � 1.1months of age; 4 males and 5 females) were approved by theOhio State University’s Institutional Animal Care and UseCommittee. Detailed description of the heart isolationprocedure, data analysis, histology, and immunohistochem-istry can be found in the online supplemental materials.

Experimental protocolThe optical mapping of the canine SAN has been describedpreviously.15,16 The excitation-contraction uncoupler bleb-bistatin17 (10–20 mM) and the near-infrared voltage-sensitivedye di-4-ANBDQBS18 (10–40 mM) were added to theperfusate. Optical mapping was performed at a rate of1000 frames/s with the MiCam Ultima-L CMOS camera(SciMedia, Costa Mesa, CA) with an optical field of view of25 � 25 mm2 (250 mm/pixel).

After control measurements, the preparations (n ¼ 7)were perfused with 1, 10, and 100 mM adenosine for 10–30minutes, followed by perfusion with the selective A1antagonist DPCPX (1 mM; n ¼ 5) for an additional 30minutes and/or washout of all drugs (n ¼ 7). Sinus cyclelength (SCL), direct sinoatrial conduction time (SACT), andSAN recovery time (SNRT) were measured.16 SNRT wasmeasured after 1 minute of slow atrial pacing (3.3 Hz) ortachypacing (7–9 Hz). To assess the potential effects ofendogenous adenosine, we conducted 2 experiments whereDPCPX (1 mM) was applied before the perfusion ofadenosine (100 mM) and no significant effects were observed(online supplemental results and online Table 1). Histologywas performed, and anatomic structures of the canine

SAN pacemaker complex were identified as previouslydescribed.15,19,20

Optical mapping data analysis and interpretationsSince canine SAN is surrounded by atrial tissue layers, thenear-infrared optical recordings were weighted averages ofsignals from both atria and SAN structures. The analysis ofthese multicomponent intramural optical action potentials(OAPs) has been previously described.15,16,21 To determinethe SAN activities during pacing, we developed a newmethod to extract SAN signals (online Figure 1). The detailsof the method are described in online supplementalmaterials.

Three different methods were used to measure SNRT (seeonline Figure 1): indirect SNRT (SNRTi), direct SNRT(SNRTd), and real SNRT (SNRTr). SNRTi was calculatedas the interval from the last atrial pacing to the firstpostpacing atrial beat, which is the traditional way of SNRTmeasurement in the clinical setting. SNRTd was calculatedas the interval from the last atrial pacing to the firstspontaneous SAN activation as described previously byGomes et al22 in 1984 (online Table 2). SNRTr wascalculated as the interval from the last SAN beat duringpacing and the first postpacing SAN beat. Corrected SNRT(cSNRT) was measured by subtracting preceding SCL fromthe measured SNRT.

StatisticsQuantitative data are shown as mean � SD. Hypothesistesting was carried out by using an unpaired Student t test orrepeated measurements analysis of variance (Minitab 16),where appropriate. Following analysis of variance, signifi-cance of the pairwise difference between SAN compartments(head, center, and tail) was determined by using a post hocTukey test. A value of P of o.05 was considered to bestatistically significant.

ResultsExperimental preparations and anatomy of thecanine SAN pacemaker complexFigure 1 illustrates that the functionally defined SANcorrelates precisely with the SAN structure, which is definedby cell morphology, fiber organization, and a lack ofconnexin 43 expression in the head of the SAN, aspreviously reported.15 It has been previously recognized thatthe SAN is insulated from the atria except at specialized SANconduction pathways (SACPs), via which SAN activationexits to the atrial myocardium.15,16,21

Effect of adenosine on SAN complex during SRAdenosine (1–100 mM) increased SCL and SACT in a dose-dependent manner (Table 1). Adenosine (10 mM) slowsSR and sinoatrial conduction, as evident from increasedSCL and SACT in Figure 2. This depression in conduction

Figure 1 Experimental preparation and histology of the canine sinoatrialnode (SAN). A: An epicardial view of a canine SAN preparation. The SAN isdemarcated by a red line. The black square shows the mapped area containingthe SAN. The adjacent black-and-white image shows the same area seen from acamera. B: Parallel histology section of SAN close to the epicardial surface.Arrows indicate locations of SAN conduction pathways. Black dots indicatelocations of pins in panel A. C: Immunolabeling of connexin 43 (Cx43) of SANconfirms the boundary of functionally defined SAN. D: Graphic representationshowing the SAN complex and the superior and inferior SACPs. FP¼ fat pad;IAS ¼ interatrial septum; IVC ¼ inferior vena cava; PV ¼ pulmonary vein;RAA ¼ right atrial appendage; SVC ¼ superior vena cava.

112 Heart Rhythm, Vol 10, No 1, January 2013

may be partially explained by the depression of slowdiastolic depolarization and SAN OAP upstroke (Figure 2).Moreover, adenosine might have depressed conduction more

Table 1 Effects of adenosine on canine SAN

Adenosi

Parameters Control 1 lM

SCL (ms) 477 � 62 551 � 8SACT (ms) during sinus rhythm 41 � 11 52 � 2cSNRTi at 3.3 Hz (ms) 65 � 28 143 � 2cSNRTr at 3.3 Hz (ms) 25 � 20 60 � 3cSNRTi at 7.5 Hz (ms) 95 � 73 124 � 4cSNRTr at 7.5 Hz (ms) 89 � 37 39 � 3SACT of first sinus beat after 7.5 Hz SNRT (ms) 41 � 5 64 � 4Atrial pause duration (s) NA NASustained AF/AFL 0/7 NAPost-tachycardia pauses via SAN exit block 0/7 NAAPD80% (ms) 164 � 6 165 � 6

Note: cSNRTs and SACT in this table do not contain SAN exit block, which are

control.

AF¼ atrial fibrillation; AFL¼ atrial flutter; APD80%¼ APD at 80% repolarizati

corrected sinoatrial node recovery time; SACT ¼ sinoatrial conduction time; SA

recovery time.

in the SAN head than in the SAN tail, leading to a switch ofpreferential conduction from the right superior to the inferiorSACP and a significant change in the atrial activation pattern(Figure 2). A switch in the preferential conduction pathwaywas observed in 4 of 7 preparations during perfusion withadenosine (10–100 mM). Moreover, inferior pacemaker shiftinside the SAN complex (from head-center to center-tail)was observed in 3 of 7 preparations. Pacemaker shift outsidethe SAN complex was not observed during regular sponta-neous rhythm.

Effects of adenosine on postpacing atrial pauses:role of SAN exit blockAdenosine (1–100 mM) significantly prolonged the cSNRTiafter “slow” atrial pacing in a dose-dependent manner(Table 1). Figure 3 shows the effect of adenosine (10 mM)on cSNRTi. After slow atrial pacing (3.3 Hz), cSNRTi wasprolonged by adenosine owing to a slower recovery of SANpacemaker activity and conduction (Figure 3A). Atrialtachypacing during adenosine exposure produced a longatrial pause of 3.8 seconds, while the SAN was not quiescentbut maintained its regular rhythm (Figure 3B). These datastrongly suggest that the post-tachypacing atrial pause is dueto the SAN exit block rather than due to reduced SANautomaticity. Atrial pauses of 4.2 � 3.4 seconds wererecorded in 5 of 7 preparations after tachypacing duringthe highest concentration of adenosine (100 mM).

While it is evident in Figure 3B that abnormally longSNRTi (or atrial pauses) is mainly due to suppressed SANconduction, it is possible that reduced automaticity of SANmay still contribute. To quantify the specific changes in theautomaticity of SAN after tachypacing, we introduced a newSNRT measurement—SNRTr (Figure 3B and online Fig-ure 1). The results (Table 1) show that cSNRTr at 3.3Hzpacing first increased and then decreased with the increasingdose of adenosine and that at 7.5 Hz pacing, cSNRTrdecreased with the increasing dose of adenosine. Thus, the

ne

10 lM 100 lM DPCPX (1 lM) Washout

2† 661 � 120† 778 � 114† 447 � 58 440 � 421 75 � 25* 86 � 16† 44 � 11 50 � 191‡ 277 � 86‡ 230 � 46‡ 120 � 49 135 � 871† 53 � 27* 2 � 15* 19 � 32 2 � 22 198 � 110 184 � 182 136 � 72 108 � 1379 34 � 20† 12 � 19‡ 8 � 35‡ 29 � 20*

2 104 � 16‡ 221 � 98† 120 � 52* 58 � 173.8 4.2 � 3.4 NA NA0/7 3/7 0/5 NA1/7 5/7 2/5 NA

161 � 8 141 � 10† 162 � 12 164 � 16

summarized in atrial pause duration(s). *P o .05; †P o .01; ‡P o .001 vs

on; cSNRTi¼ indirect corrected sinoatrial node recovery time; cSNRTr¼ real

N ¼ sinoatrial node; SCL ¼ sinus cycle length; SNRT ¼ sinoatrial node

Figure 2 Adenosine depresses sinoatrial node (SAN) conduction during sinus rhythm. On the left are the optical action potentials (OAPs) and atrial pseudo-ECG in control condition (A) and during perfusion of 10 mM adenosine (B). The blue OAP (1) is from the SAN; the green (2) is from the atria. Optical recordingsfrom the SAN exhibit the slow diastolic depolarization (SDD), slowly rising upstroke of the SAN (SAN component), and rapidly rising upstroke of the atrialmyocardium (atrial component). SAN activation time (SACT) is the delay between the SAN activation and the atrial activation. On the right are maps for SANactivation and subsequent atrial activation in control and 10 mM adenosine. Black arrows indicate the pathways through which the activation exits from the SANto the atria. CT ¼ crista terminalis; IAS ¼ interatrial septum; RAA ¼ right atrial appendage; SVC ¼ superior vena cava.

Lou et al Adenosine and Tachy-Brady Syndrome 113

inhibitory effects of rapid pacing on the SAN pacemakerautomaticity was decreased by adenosine owing to a reducedfrequency of paced SAN (entrance block) under adenosine(Figures 3B and 4) that prevented SAN automaticityoverdrive suppression. These results further emphasized thatreduced automaticity is not the mechanism for atrial pausesin our study.

The extracted SAN signals in Figure 3 show that SANwas paced 1:1 at the slow atrial pacing rate but not at therapid pacing in control. This suggests that the last atrialpacing beat often does not correspond to the onset of the lastpaced SAN beat during tachypacing. Therefore, tachypacingcSNRTi (Table 1) and cSNRTd (online Table 2) are highlyvariable parameters that depend on the timing of the lastSAN beat relative to the last atrial pacing beat.

Adenosine and tachy-brady arrhythmiaAdenosine induced significant shortening of the atrial actionpotential duration (APD) and refractoriness at 100 mMadenosine (Table 1), which allowed the induction of AFLby rapid pacing in 3 preparations. In tachy-brady syndrome,termination of AFL is followed by atrial pauses. Figure 4 is anexample that recapitulates this phenomenon and shows a post-tachycardia atrial pause of more than 7 seconds. During thislong atrial pause, the SAN remains active, suggesting SANexit block as the primary mechanism for post-tachycardiaatrial pauses. Similar to the acetylcholine-induced AFL

reported previously,16 the AFL induced in isolated RApreparations in this study was consistently found to be reentryaround SAN (Figure 4B). DPCPX (selective adenosinereceptor A1 antagonist) or washout recovered adenosine-induced changes in SCL, SACT, cSNRTi, and atrial APDback to values no different from those for control (Table 1).

SAN activity during atrial pacingAs shown in Figure 3B and previously,16 the frequency inthe SAN during tachypacing was lower than that in the atriabut faster than its spontaneous rate in control conditions,16

suggesting that the atrial waves entered the SAN withvariable entrance block (from 2:1 to 5:1). Online Figure 2shows a detailed example of how atrial excitation entered theSAN at the end of atrial tachycardia.

Dominant frequency (DF)21 was used to quantify adeno-sine effects on SAN activity during atrial tachypacing (7–9Hz). The DF was significantly decreased by adenosine (10–100 mM) in all 3 SAN compartments: head, center, and tail(Figures 5A and 5C). This is consistent with the adenosine-induced reduced SAN frequency during rapid pacing shownin the extracted SAN signals in Figure 3. These resultsconfirm that adenosine depressed the conduction from theatria to the SAN. Furthermore, the decrease in the DF withadenosine was heterogeneous, with the most significanteffects on the head and tail of SAN compared with thoseon the center of the SAN (Figure 5D).

Figure 3 Adenosine increases the corrected indirect sinoatrial node(SAN) recovery time (cSNRTi). Optical action potential (OAP) and atrialelectrocardiogram (ECG) during and after the termination of pacing at 3.3Hz (A) and 7 Hz (B) are shown. The blue OAPs are from the low center partof the SAN; the green atrial OAPs are from the crista terminalis. Theextracted SAN signals (dark blue) are indicated as “extracted SAN.”Corrected indirect SAN recovery time (cSNRTi) and sinus cycle length(SCL) are shown for each recording. Real SNRT (cSNRTr) is shown for thelast recording where SAN exit block was present. “Paced SAN” means thatSAN was paced by atria.

114 Heart Rhythm, Vol 10, No 1, January 2013

Depression of excitability in SAN conductionpathways by adenosineTo determine the spatial distribution of depression of SANexcitability after the termination of tachypacing, we exam-ined the post-tachypacing changes in SAN action potential(AP) amplitude, AP upstrokes (dV/dtmax), and hyperpolar-ization. Figure 6 shows the inhibition of excitability in thehead of SAN near the superior SACP induced by atrialtachypacing during adenosine perfusion. This depression inexcitability is reflected by decreased SAN AP amplitude(Figure 6A and 6C), reduced upstroke velocity (dV/dt;Figures 6A and 6D), and hyperpolarization (Figure 6A).This suppression is more significant at the head and tail ofSAN compared to the central part and is therefore hetero-geneous. Online Figure 3 also shows that tachycardia

nonuniformly suppressed SAN excitability during post-tachycardia pauses.

Figure 6 demonstrates that after the termination of atrialpacing, the leading pacemaker in the center-inferior part ofthe SAN continues to generate regular activity with almostthe same frequency as before pacing. However, the SANactivation waves (E1-E5) could not exit to the atrialmyocardium owing to complete inexcitability of the SANregion next to the SACPs. The significant depression in thesuperior SAN is correlated with the larger reduction of DF inthe SAN head by adenosine during tachypacing (Figure 5).More than 3 seconds were required for the recovery ofexcitability at the SAN head, which leads to the recovery ofconduction through the superior SACP, restoring SR(Figure 6B, S1).

We also successfully measured the APD of various SANcompartments owing to the complete separation of SANOAPs from atrial signals during adenosine-induced exitblock (Figure 7). Figure 7B shows 2 representative examples(data from Figures 4 and 6) that indicate that the APD in thecenter part of SAN (277 � 10 ms) was significantly longerthan the APD in the head (224 � 15 ms) and tail (201 � 28ms) of the SAN (P o .001 for head vs center vs tail APDs).Moreover, the shortest APD areas of the SAN complex(Figure 7, middle) were connected to the preferentialentrance conduction pathways through which the SAN waspaced by atria (online Figure 2). Furthermore, drift of theleading pacemaker sites was usually observed during the first3–5 beats of post-tachycardia pauses (SAN exit block) aswell as after recovery of sinoatrial conduction and SRrestoration. The leading pacemaker would drift toward eitherthe superior part (Figure 6) or the inferior part (OnlineFigure 3) by 1–10 mm. Figure 7C shows a summary ofleading pacemaker sites in these 2 SANs.

DiscussionIn this study of a canine SAN preparation, we found thatadenosine heterogeneously depressed SAN conduction andexcitability, leading to post-tachycardia atrial pauses up toseveral seconds and an apparent tachy-brady syndromepattern. We observed that the post-tachycardia atrial pausesduring adenosine resulted from SAN exit block rather thanthe inhibition of pacemaker automaticity.

Electrophysiological consequences of adenosine in SANpacemaker cells include activation of a potassium outwardcurrent as well as suppression of inward calcium current andhyperpolarization-activated current (“funny” current).9,23–25

These effects could explain our observations of reducedexcitability of the SAN. West and Belardinelli9 demonstratedthat the response to adenosine was heterogeneous in therabbit SAN. In the same study, this heterogeneous responsewas suggested to lead to the pacemaker shift duringadenosine treatment9 and it echoes the heterogeneous con-duction depression and pacemaker shift during SR and post-tachycardia atrial pauses observed in our canine study. Themechanism for this heterogeneous suppression is less well

Figure 4 Adenosine-induced tachy-brady arrhythmias. A: Optical action potential (OAP) recordings and atrial electrocardiogram (ECG) recording during anepisode of the tachy-brady pattern during adenosine (100 mM). Atrial flutter (AFL) was converted to atrial fibrillation (AFib) by rapid atrial pacing. AFibspontaneously terminated and was followed by an atrial pause, which is evident from the ECG and atrial OAP (green). B: Activation maps during AFL, sinoatrialnode (SAN) exit block, and the first sinus rhythm (SR) beat of the same recording in panel A. Numbers 1 and 2 in the maps indicate the signal origins of OAPs inpanel A.

Lou et al Adenosine and Tachy-Brady Syndrome 115

defined. It could be related to variable expression ofadenosine receptors and their downstream effectors, suchas the potassium outward current channels, in differentcompartments of the SAN complex. The heterogeneoussensitivity to adenosine might also be explained by thecomplex specialized structure of the SAN. SAN pacemakertissue near the superior and inferior SACPs might be moresusceptible to overdrive suppression by rapid atrial pacingowing to their proximity to the atria and shorter repolariza-tion time (and presumably shorter refractory periods; seeFigure 7). This would potentially filter the rapid frequency ofstimulation and thus attenuate the overdrive suppression ofthe SAN center. This may explain why rapid atrial pacing didnot suppress SAN automaticity during higher concentrationsof adenosine (Figure 3B; cSNRTr in Table 1).

The endogenous production of adenosine in vivo isrelatively low in a normal heart at resting physiologicalconditions without ischemia or hypoxia (o1 mM in coronarysinus plasma and interstitial fluid26). The absence of theDPCPX effect on the basic SAN function (online Table 1)supports our assumption that the effects of endogenousadenosine are minor in our experimental conditions. Thus,in our study, we mimicked the conditions with the highproduction of adenosine (such as ischemia26 and hypoxia27)by adding adenosine to the perfusate. However, only thecombination of adenosine and rapid atrial rates (pacing ortachycardia) led to pauses. The observation that adenosine

alone did not induce pauses might be due to the absence ofany preexisting SAN dysfunction in normal dogs. In patientswith SAN dysfunction, it has been shown that either pacingalone or bolus injection of adenosine alone can produce atrialpauses.12,13 In the absence of SAN dysfunction in normaldogs, adenosine alone or rapid atrial rates15,16 alone onlyincrease susceptibility to atrial pauses but are insufficient toreach the threshold for the induction of exit block and atrialpauses.

SAN dysfunction could result from intrinsic changes inthe node itself or from extrinsic causes.4 Our results supportadenosine as an extrinsic cause of SAN dysfunction andprovide experimental evidence to support Watt’s hypoth-esis11 on the causal relationship between adenosine andtachy-brady syndrome. As we have shown, adenosinecontributes to tachy-brady activation patterns by bothproducing post-tachycardia pauses and facilitating AFL/AFby shortening the APD, the latter of which is consistent withthe findings reported in previous pioneering studies.10,24

Ideally, the role of adenosine in SAN dysfunction shouldbe studied in the human heart with SAN dysfunction. Indeed,the normal canine heart is different from the diseased humanheart with SAN dysfunction. However, access to intactexplanted human hearts is limited.21 Both anatomical1,28,29

and functional20,30 studies suggest that the canine SANpreparation is a realistic model for the human SAN.15,21

Therefore, information gained from the canine SAN

Figure 5 Adenosine slows the activation frequency of sinoatrial node (SAN) complex during atrial tachypacing (7.5 Hz). A: Dominant frequency (DF) mapsduring rapid pacing. SAN is outlined by a dashed line. Note that the frequency decreased from the head and tail of SAN toward the center of the SAN and that 10mM adenosine reduced the frequencies in the SAN. B: Optical action potentials (OAPs) during pacing under control conditions and 10 mM adenosine. The blueOAP (1) is from the center of SAN; the green OAP (2) is from the crista terminalis (CT) in the atria. The exact origins of these OAPs are shown in panel A.C: Adenosine-induced SAN frequency shift. Frequency power spectra for 4 locations are shown, including SAN center (1), atrial CT (2), SAN tail (3), and SANhead (4). The exact locations are indicated in the maps in panel A. Note that 10 mM adenosine significantly decreased DF, especially in the head of SAN.D: Summary of DF and DF change during adenosine perfusion. Left: DF during sinus rhythm (SR) and during rapid atrial pacing at control, 10 mM adenosine,and 100 mM adenosine. *P o .05 vs SR. Right: DF change during adenosine perfusion relative to control condition. #P o .05 between the ends (head and tail)and the center of SAN.

116 Heart Rhythm, Vol 10, No 1, January 2013

preparation improves our mechanistic understanding ofhuman SAN dysfunction.

ConclusionsThese data directly demonstrate that adenosine-induced post-tachycardia atrial pauses resulted from depressed conductionbetween the SAN and the atria (exit block) rather than from

slowed pacemaker automaticity in canine atria. Conductionblock was preferentially related to adenosine-induced excit-ability depression within the SAN superior and inferiorconduction pathways. We suggest that the local targetingof adenosine receptors in SACPs could be sufficient toprevent SAN arrhythmias by restoring proper conductionbetween the SAN and the atrial myocardium withoutaffecting other cardiac functions.

Figure 6 Adenosine-induced inhibition of sinoatrial node (SAN) excitability by rapid pacing. A: Optical action potentials (OAPs) from the center (blue) andthe head (green) of SAN during an episode of SAN exit block induced by 10 mM adenosine and tachypacing. Below the OAPs are a close-up view of 5 beats ofSAN exit block (E1-E5) and signal derivatives (maximum derivative of each beat indicates the excitability). B: Activation maps showing the restoration of SANconduction toward the head of SAN and the conduction through the superior SAN conduction pathway (black arrow) to the atria. SAN OAP amplitude maps (C)and maximum OAP derivative (dV/dtmax) maps (D) show the recovery of excitability in the head of SAN.

Lou et al Adenosine and Tachy-Brady Syndrome 117

Figure 7 Sinoatrial node (SAN) activation and repolarization duringSAN exit block. A and B: Activation maps and action potential duration(APD) maps at 80% recovery (APD80%) for 2 canine SANs. Note thatconduction in the head of SAN is slower than that in the tail of SAN,consistent with more connexin 43 (Cx43) expression in the inferior part ofSAN compared to the superior part. Far right, histology overlays display agraphic representation of SAN complexes. Blue small ovals indicatelocations of leading pacemaker sites during adenosine-induced exit blockand recovery from it. IAS ¼ interatrial septum; RAA ¼ right atrialappendage; SVC ¼ superior vena cava.

118 Heart Rhythm, Vol 10, No 1, January 2013

AcknowledgmentsWe thank Dr Stanislav Zakharkin for his help in statisticaldata analysis and Jeanne Green for her technical assistance.

Supplementary InformationSupplementary data associated with this article can be foundin the online version at doi:10.1016/j.hrthm.2012.09.012.

References1. James TN. Anatomy of the human sinus node. Anat Rec 1961;141:109–139.2. Boyett MR, Honjo H, Kodama I. The sinoatrial node, a heterogeneous pacemaker

structure. Cardiovasc Res 2000;47:658–687.3. Opthof T. The mammalian sinoatrial node. Cardiovasc Drugs Ther 1988;1:

573–597.4. Mangrum JM, DiMarco JP. The evaluation and management of bradycardia. N

Engl J Med 2000;342:703–709.5. Short DS. The syndrome of alternating bradycardia and tachycardia. Br Heart J

1954;16:208–214.6. Asseman P, Berzin B, Desry D, et al. Persistent sinus nodal electrograms during

abnormally prolonged postpacing atrial pauses in sick sinus syndrome in humans:sinoatrial block vs overdrive suppression. Circulation 1983;68:33–41.

7. Mandel W, Hayakawa H, Danzig R, Marcus HS. Evaluation of sino-atrial nodefunction in man by overdrive suppression. Circulation 1971;44:59–66.

8. Yeh SJ, Lin FC, Wu D. Complete sinoatrial block in two patients withbradycardia-tachycardia syndrome. J Am Coll Cardiol 1987;9:1184–1188.

9. West GA, Belardinelli L. Sinus slowing and pacemaker shift caused by adenosinein rabbit SA node. Pflugers Arch 1985;403:66–74.

10. Drury AN, Szent-Gyorgyi A. The physiological activity of adenine compoundswith especial reference to their action upon the mammalian heart. J Physiol1929;68:213–237.

11. Watt AH. Sick sinus syndrome: an adenosine-mediated disease. Lancet 1985;1:786–788.

12. Lin JM, Lin JL, Lai LP, Huang SK. Usefulness of single-bolus adenosine test forconfirming sinus node dysfunction and correlation with atrial overdrive suppres-sion test. Am J Cardiol 2004;94:1569–1572.

13. Fragakis N, Antoniadis AP, Korantzopoulos P, Kyriakou P, Koskinas KC,Geleris P. Sinus nodal response to adenosine relates to the severity of sinus nodedysfunction. Europace 2012;14:859–864.

14. Saito D, Matsubara K, Yamanari H, et al. Effects of oral theophylline on sicksinus syndrome. J Am Coll Cardiol 1993;21:1199–1204.

15. Fedorov VV, Schuessler RB, Hemphill M, et al. Structural and functionalevidence for discrete exit pathways that connect the canine sinoatrial node andatria. Circ Res 2009;104:915–923.

16. Fedorov VV, Chang R, Glukhov AV, et al. Complex interactions between thesinoatrial node and atrium during reentrant arrhythmias in the canine heart.Circulation 2010;122:782–789.

17. Fedorov VV, Lozinsky IT, Sosunov EA, et al. Application of blebbistatin as anexcitation-contraction uncoupler for electrophysiologic study of rat and rabbithearts. Heart Rhythm 2007;4:619–626.

18. Matiukas A, Mitrea BG, Qin M, et al. Near-infrared voltage-sensitive fluorescentdyes optimized for optical mapping in blood-perfused myocardium. HeartRhythm 2007;4:1441–1451.

19. Fedorov VV, Hucker WJ, Dobrzynski H, Rosenshtraukh LV, Efimov IR.Postganglionic nerve stimulation induces temporal inhibition of excitability inthe rabbit sinoatrial node. Am J Physiol 2006;291:H612–H623.

20. Fedorov VV, Glukhov AV, Chang R. Conduction barriers and pathways of thesinoatrial pacemaker complex: their role in normal rhythm and atrial arrhythmias.Am J Physiol Heart Circ Physiol 2012;302:H1773–H1783.

21. Fedorov VV, Glukhov AV, Chang R, et al. Optical mapping of the isolatedcoronary-perfused human sinus node. J Am Coll Cardiol 2010;56:1386–1394.

22. Gomes JA, Hariman RI, Chowdry IA. New application of direct sinus noderecordings in man: assessment of sinus node recovery time. Circulation 1984;70:663–671.

23. West GA, Belardinelli L. Correlation of sinus slowing and hyperpolarizationcaused by adenosine in sinus node. Pflugers Arch 1985;403:75–81.

24. Belardinelli L, Giles WR, West A. Ionic mechanisms of adenosine actions inpacemaker cells from rabbit heart. J Physiol 1988;405:615–633.

25. Zaza A, Rocchetti M, DiFrancesco D. Modulation of the hyperpolarization-activated current (I(f)) by adenosine in rabbit sinoatrial myocytes. Circulation1996;94:734–741.

26. Lasley RD, Hegge JO, Noble MA, Mentzer RM, Jr. Comparison of interstitialfluid and coronary venous adenosine levels in in vivo porcine myocardium. J MolCell Cardiol 1998;30:1137–1147.

27. Jenkins JR, Belardinelli L. Atrioventricular nodal accommodation in isolatedguinea pig hearts: physiological significance and role of adenosine. Circ Res1988;63:97–116.

28. James TN, Sherf L, Fine G, Morales AR. Comparative ultrastructure of the sinusnode in man and dog. Circulation 1966;34:139–163.

29. Truex RC, Smythe MQ, Taylor MJ. Reconstruction of the human sinoatrial node.Anat Rec 1967;159:371–378.

30. Boineau JP, Schuessler RB, Hackel DB, Miller CB, Brockus CW, Wylds AC.Widespread distribution and rate differentiation of the atrial pacemaker complex.Am J Physiol 1980;239:H406–H415.