Mechanisms by Which Adenosine Restores Conduction in Dormant Canine Pulmonary Veins
Background— Adenosine acutely reconnects pulmonary veins (PVs) after radiofrequency application, revealing “dormant conduction” and identifying PVs at risk of reconnection, but the underlying mechanisms are unknown.
Methods and Results— Canine PV and left-atrial (LA) action potentials were recorded with standard microelectrodes and ionic currents with whole-cell patch clamp before and after adenosine perfusion. PVs were isolated with radiofrequency current application in coronary-perfused LA-PV preparations. Adenosine abbreviated action potential duration similarly in PV and LA but significantly hyperpolarized resting potential (by 3.9±0.5%; P<0.05) and increased dV/dtmax (by 34±10%) only in PV. Increased dV/dtmax was not due to direct effects on INa, which was reduced similarly by adenosine in LA and PV but correlated with resting-potential hyperpolarization (r=0.80). Adenosine induced larger inward rectifier K+current (IKAdo) in PV (eg, −2.28±0.04 pA/pF; −100 mV) versus LA (−1.28±0.16 pA/pF). Radiofrequency ablation isolated PVs by depolarizing resting potential to voltages positive to −60 mV. Adenosine restored conduction in 5 dormant PVs, which had significantly more negative resting potentials (−57±6 mV) versus nondormant (−46±5 mV, n=6; P<0.001) before adenosine. Adenosine hyperpolarized both, but more negative resting-potential values after adenosine in dormant PVs (−66±6 mV versus −56±6 mV in nondormant; P<0.001) were sufficient to restore excitability. Adenosine effects on resting potential and conduction reversed on washout. Spontaneous recovery of conduction occurring in dormant PVs after 30 to 60 minutes was predicted by the adenosine response.
Conclusions— Adenosine selectively hyperpolarizes canine PVs by increasing IKAdo. PVs with dormant conduction show less radiofrequency-induced depolarization than nondormant veins, allowing adenosine-induced hyperpolarization to restore excitability by removing voltage-dependent INa inactivation and explaining the restoration of conduction in dormant PVs.
Received July 9, 2009; accepted January 4, 2010.
Pulmonary vein isolation (PVI) is an effective treatment for atrial fibrillation (AF).1,2 Nevertheless, many patients require repeated ablation procedures because of AF recurrence, which in most cases are associated with reconnection of previously isolated PVs.3,4 It has recently been noted that intravenous purinergic agonists such as adenosine can transiently restore conduction through a previously isolated PV, a phenomenon called “dormant conduction.”5–8 The demonstration of dormant conduction has predictive value for eventual reconnection, and additional radiofrequency (RF) applications to veins showing dormant conduction at the time of initial PVI may prevent reconnection and AF recurrence.6–8 The mechanisms by which adenosine restores conduction to dormant PVs are unknown. The objectives of this study were to (1) explore the effects of adenosine on ionic currents and action potentials (APs) in canine left-atrial (LA) and PV cardiomyocytes, and (2) relate these effects to changes in conduction between the PV and LA after RF ablation in an in vitro model.
Clinical Perspective on p 972
Materials and Methods
See the online-only Data Supplement for the complete Materials and Methods section. The following text summarizes the complete section.
Animals and Tissues
Forty-seven adult mongrel dogs were anesthetized with pentobarbital (30 mg kg−1 intravenously) and artificially ventilated. Hearts were excised and immersed in oxygenated Tyrode solution.
For standard microelectrode experiments, intact tissue preparations, including LA and PVs, were perfused via the left-circumflex coronary artery with oxygenated Krebs solution at 35°±0.5°C, and APs were recorded as previously described.9 Measurements included resting membrane potential (RMP), AP amplitude, and AP duration (APD) at 90% of repolarization (APD90). Conduction time was measured during LA pacing between peaks of differentiated (dV/dt) signals for APs recorded with 2 stable microelectrodes ≈1-cm apart, one in a PV and the other in the closest adjacent LA region. Cell isolation was performed as previously described.10,11 After isolation, cells were stored (4°C) and studied within 12 hours.
PVI was conducted in LA-PV preparations subjected to microelectrode AP recordings. One PV was isolated in each dog by ablating in the antral region as close as possible to the PV-LA junction. Bipolar electrodes were attached to the epicardial surface of the LA and target PV to evaluate PV-LA conduction. RF energy was delivered epicardially in the unipolar mode between the standard 4-mm tip of a 7F quadripolar ablation catheter and an indifferent peripheral electrode, with a power limit of 25 to 35 W, for only 10 seconds at each site to control damage. The endpoint of PVI was a bidirectional conduction block (both from LA to PV and from the PV at 4 different quadrants to LA, with pacing stimuli up to maximum possible stimulation strength).
When PVI was achieved, APs were recorded in PV sleeves immediately above the ablation line for 15 minutes after the final RF application. If conduction recovered, then additional RF was delivered to complete PVI. Adenosine was then added, and APs were recorded for an additional 15 minutes in 11 preparations (6 left superior and 5 left inferior PVs). In preparations that recovered conduction with adenosine, APs were recorded during a final 15-minute adenosine washout period. In 6 preparations (5 left superior and 1 left inferior PVs) studied to establish the time course of spontaneous changes after PVI, the same PVI protocol was conducted without subsequently adding adenosine, and preparations were monitored for up to 4 hours (average monitoring time 3.4±0.3 hours). In 9 other preparations (6 left superior and 3 left inferior PVs), prolonged monitored was obtained after adenosine washout to evaluate whether adenosine reconnection predicts subsequent spontaneous reconnection.
Adenosine Receptor Protein Expression
Protein-enriched fractions were obtained from cardiomyocytes isolated as described above, with paired PV and LA samples from each of 5 dogs studied. Antiadenosine A1- and A2A-receptor antibodies were obtained from ABCAM (ab3460 and ab3461); antiadenosine A2B- and A3-receptor antibodies (A2bR23 and A3R32) were from Alpha Diagnostic (San Antonio, TX). Donkey anti-rabbit secondary antibodies conjugated to horseradish peroxidase were used for detection. All of the expression data are provided relative to GAPDH staining for the same samples on the same gels.
Details of solution contents for AP recording and voltage-clamp studies are provided in the online-only Data Supplement. Adenosine was freshly prepared as a 1-mmol/L solution before each experiment.
Electrophysiology Data Acquisition
The whole-cell patch-clamp technique was used to record currents. Electrode tip resistances were 1 to 4 Megohms, with 1 to 2 Megohm pipettes used for INarecording. Cell capacitances averaged 97±9 pF for PV and 99±7 pF for LA cardiomyocytes. Rs averaged 2.9±0.3 Megohms and 279±18 μs after compensation in LA; 2.6±0.2 Megohms and 242±13 μs in PV. Currents are expressed as densities (pA/pF). K+-current recording was performed at 35°±0.5°C. INa was recorded at room temperature. The maximum phase 0 upstroke velocity (dV/dtmax) of standard microelectrode recordings was determined by electronic differentiation.
Data are expressed as mean±SEM. Repeated-measures 1-way ANOVA was used to compare RMP at different times after PVI. Repeated-measures 2-way ANOVA was used to study the interaction between location (PV versus LA) and adenosine for RMP, APD90, AP amplitude, and dV/dtmax; the interaction between test potential and treatment (adenosine versus no adenosine) for K+current and INa; and the interaction between test potential and location (PV versus LA) for IKAdo. Bonferroni-adjusted comparisons were used to compare group means when ANOVA was significant. Repeated-measures 2-way analysis was used to study interactions between dormant/nondormant state and adenosine effect on RMP after PVI. An unpaired Student t test was used to compare ablation time between dormant and nondormant PVs and between control and adenosine-treated PVs. All of the data satisfied criteria for normality of distribution (Kolmogrov-Smirnov) except for conduction time. The effect of drug on conduction time was studied with Wilcoxon signed rank test because it was not normally distributed. Pearson correlation coefficients were used to study the relation between RMP and dV/dtmax. Fisher exact test was used to compare dormant-conduction prevalence in different PVs. Whenever more than 1 cell was obtained per region for a dog, all of the cell values in each region were averaged to obtain single representative values for that dog for statistical comparison to avoid weighting values unevenly for different dogs. A 2-tailed P<0.05 was considered statistically significant. The authors had full access to and take responsibility for the integrity of the data. All of the authors have read and agree to the article as written.
AP and Conduction-Time Changes
AP effects of adenosine were explored in 6 dogs. AP recordings before and after adenosine are illustrated in Figure 1 (Top). Adenosine shortened APD in both PV (Figure 1A) and LA (Figure 1B) cells but significantly hyperpolarized RMP and increased dV/dtmax only in PV cells. Complete results and statistics are provided in Table 1, and mean data are illustrated in Figure 1 (bottom). In PV sleeves but not in LA, adenosine significantly increased RMP (Figure 1C) and dV/dtmax (Figure 1D). Adenosine abbreviated APD similarly in PV and LA (Figure 1E). AP amplitude was increased only in PV (Figure 1F). RMP and dV/dtmax correlated (Figure 2A), with similar and strong correlation for PV data alone (Figure 2B) and LA data alone (Figure 2C), consistent with known voltage dependence of INaavailability. Adenosine also reduced conduction time between LA and PV electrodes from 17.4±3.2 to 14.9±3.1 ms (P=0.031).
Adenosine-Induced Ion-Current Changes
IKAdo recordings from a PV and LA cardiomyocyte are illustrated in Figure 3A. Larger adenosine-induced K+currents were consistently seen in PVs. Overall results are shown in Figure 3B. The reversal potential was approximately −70 mV (when corrected for the 15-mV mean junction potential). Location was a highly significant determinant of IKAdo, which was larger in PV than LA cardiomyocytes. For example, at −100 mV, IKAdo averaged −2.28±0.04 pA/pF in PV cells versus −1.28±0.16 pA/pF in LA cells.
INa recordings from LA and PV cardiomyocytes are shown in Figure 4A and 4B. Panels labeled “a” show control currents, panels labeled “b” show currents from the same cells in the presence of adenosine, and panels labeled “c” show corresponding current-voltage relations. Overall, adenosine reduced peak INa by 26±4% in PV cells versus 26±8% in LA cells (P=NS). Washout resulted in up to 99.8% reversal of effect, indicating that the response was due to adenosine. These results indicate that the adenosine-induced dV/dtmaxincreases in PV cardiomyocytes are not due to direct INa-enhancing effects.
Recordings After PVI
Figure 5A shows a typical LA-PV preparation, with the left superior PV attached, before PVI. Figure 5B shows the same preparation after PVI. Figure 5C shows PV and LA electrograms during PV pacing before PVI (before ablation) and recordings from the same sites after PVI (after ablation). Adenosine led to restoration of 1:1 PV-LA conduction. Adenosine washout was followed initially by loss of 1:1 conduction and then by complete PV-LA conduction block. Figure 5D shows corresponding recordings from a vein without dormant conduction. In contrast to the dormant vein, adenosine did not restore PV-LA conduction. Adenosine revealed dormant conduction in 5 of 11 PVs (46%). Reconnection with adenosine occurred after 1 minute of perfusion initiation in 2 cases and after 2, 4, and 8 minutes in the other 3 dormant PVs. Reconnection with adenosine was neither related to the type of PV (3 of 5 left inferior PVs recovered conduction versus 2 of 6 left superior; P=0.567), nor to the total RF time required to achieve PVI (57±21 seconds in dormant and 54±21 seconds in nondormant PVs; P=0.925).
Microelectrode recordings just above the ablation line in a PV without dormant conduction are illustrated in Figure 6A and in a PV with dormant conduction in Figure 6B. RF-induced PVI depolarized RMP, causing inexcitability. Adenosine hyperpolarized RMP comparably (by ≈10 mV) in PVs without (Figure 6C) or with (Figure 6D) dormant conduction. However, PVs with dormant conduction had significantly more negative RMPs before adenosine exposure (−57±6 versus −46±5 mV in nondormant veins; P<0.001), so that after adenosine-induced hyperpolarization, the RMP in dormant veins became negative to −60 mV (mean −66±6 mV), whereas in nondormant veins the RMP remained more depolarized (−56±6 mV; P<0.001). Adenosine effects in dormant PVs disappeared on washout, with RMP depolarization and return of block in 4 of the 5 PVs with dormant conduction (80%).
To examine another agent sometimes used to reveal dormant conduction, we administered isoproterenol (1 μmol/L) to 5 veins subjected to PVI. Figure I in the online-only Data Supplement compares spontaneous RMP changes after PVI (panel A), changes caused by adenosine (panel B), and those occurring with isoproterenol (panel C). Although hyperpolarization occurred with isoproterenol, its magnitude was significantly smaller than with adenosine (panel D), and no cases of isoproterenol-induced reconnection occurred.
Time Course of RMP Changes After PVI
To assess the time course of spontaneous RMP changes after PVI (in the absence of adenosine), RF ablation was applied in 6 preparations, and APs were then recorded over time during continued perfusion. In these PVs, RMP averaged −73±2 mV before ablation versus −49±4 mV after ablation (P<0.001). No statistically significant hyperpolarization was seen from 15 to 25 minutes after ablation (Figure 7A), in contrast to the clear hyperpolarization after adenosine administration during the same interval (Figure 7B). Statistically significant spontaneous hyperpolarization began after 30 minutes and progressed slowly thereafter (Figure 7C). Spontaneous reconnection occurred in 2 of 6 PVs (33%), 1 at 34 minutes, and the other at 51 minutes after PVI. These data suggest that spontaneous recovery of RMP occurs gradually after PVI and can lead to reconnection. However, the time course of such changes in the absence of adenosine is slow compared with the rapid hyperpolarization and reconnection of dormant veins seen after adenosine administration.
We then assessed whether the response to adenosine could predict spontaneous reconnection. Adenosine reconnection occurred in 6 of 9 additional PVI preparations monitored for 90 minutes after adenosine perfusion and washout (within 1 minute of adenosine perfusion in 2, 2 minutes in 1, 5 minutes in 2, and 10 minutes in 1). Adenosine hyperpolarized RMP in these veins (Figure 7D) with reversal of hyperpolarization, and return of block in 5 of 6 PVs, on adenosine washout. Thereafter, slow spontaneous hyperpolarization followed, and 4 of 5 veins reconnected after 10 to 40 minutes. In all 3 PVs that failed to reconnect with adenosine, no spontaneous late reconnection occurred. Thus, 5 of 6 PVs (83%) with adenosine-exposed dormant conduction showed reconnection, versus none of 3 PVs (0%) without adenosine-exposed dormant conduction (P=0.048).
Figure 8 shows expression data for adenosine receptor types 1, 2A, 2B, and 3. No PV-LA differences were seen.
In the present study, we assessed the mechanisms by which adenosine restores conduction to PVs that are isolated by RF ablation. We found that adenosine selectively hyperpolarizes PV cardiomyocytes and elicits larger IKAdo in PV cells versus LA cells. Furthermore, we found that RF-induced PVI depolarizes PV cells to voltages positive to −60 mV, at which Na+channels are known to be inactivated, thereby inducing inexcitability. By hyperpolarizing cells to voltages negative to −60 mV, adenosine restores excitability to dormant PVs.
Ionic Mechanisms of Adenosine Action on PVs
The majority of the principal electrophysiological effects of adenosine, including hyperpolarization and repolarization acceleration in atrial cells and conduction slowing and refractoriness prolongation in AV nodal tissues,12 are attributable to the activation of a G protein-coupled K+current, IKAdo.13 IKAdo is mediated by the same G protein system and coupled K+channels (Kir3.1/3.4) as acetylcholine-regulated K+current (IKACh).13 We found that adenosine reduced LA and PV APD but significantly hyperpolarized RMP only in PV cells. This differential action on RMP may have been due to 2 factors: (1) IKAdo was larger in PV than in LA cardiomyocytes, and (2) PV cells have smaller IK1 (and therefore less negative RMPs) than LA cells,14 which would increase the membrane input resistance and increase the effect on transmembrane potential of a given amount of change in ionic current.15 A similar phenomenon may also account for the larger degree of adenosine-induced hyperpolarization in PVs after PVI (≈10 mV), compared with the effect in PVs not subjected to PVI (≈3 mV), because PVI causes substantial PV cell depolarization. The mechanistic basis for larger IKAdo in PVs is unclear, but interestingly constitutive IKACh, which shares similar Kir3.1/3.4 ion-channel subunits and regulatory G protein pathways with IKAdo, is also larger in PV than LA.16 These differences are not due to differential Kir3.x or inhibitory G protein expression, which are similar in PV and LA cells,16 nor is LA-PV differential adenosine receptor expression involved (Figure 8). Further work is needed to clarify the molecular basis of differential constitutive IKACh and IKAdo in PV versus LA cardiomyocytes.
Two mechanisms could account for the adenosine-induced dV/dtmax increase in PV cells. The first possibility is removal of voltage-dependent INainactivation by membrane hyperpolarization, which is expected to be particularly important at the relatively depolarized RMP of PV cells. This idea is consistent with the correlation between RMP and dV/dtmax shown in Figure 2. Another possibility would be direct adenosine-induced increases in INa. However, on direct measurement, adenosine reduced rather than increased INa. Acetylcholine inhibits INa via inhibitory G protein–dependent mechanisms in the presence of adenylate-cyclase activation by isoproterenol or forskolin,17 an action that could presumably occur with adenosine because it activates the same inhibitory G proteins as cholinergic agonists.13
Clinical Significance of PV Reconnection
The effectiveness of PVI to treat AF is well established.1,2 Considerable evidence supports the relationship between electric isolation of PVs and the success of catheter ablation procedures. The majority of the patients who fail an initial ablation procedure have resumption of PV conduction.3,4 The most convincing evidence for the crucial role of successful PV-LA disconnection in curing AF comes from reports that describe a dramatic difference in the PV reconnection rate between patients cured of AF (few of whom show recurrent PV conduction) and those with recurrences (most showing PV reconnection).18,19 Repeated procedures to ensure PVI significantly improve long-term outcomes in patients who have recurrent AF after an initial procedure.2,20
Early spontaneous recurrence of PV conduction has been observed in 24% to 50% of isolated PVs after a waiting period of 30 to 60 minutes.21,22 Additional RF lesions to ensure isolation of PVs showing acutely recovered conduction provides similar long-term AF control to that seen in patients without early reconnection23 and better AF control than in cases in which early reconnection was not explored.24 Based on these observations, some authors recommend a 60-minute waiting period after initial PVI to detect early recurrences of conduction.21 Similar to clinical observations, we noted that 33% of PVs reconnected spontaneously during the 30- to 60-minute window after PVI. Adenosine has been used to assess rapidly the ability of PVs to reconnect, with 25% to 36% of PVs showing acute reconnection on adenosine infusion and repeated RF delivery to adenosine-reconnected veins apparently improving long-term outcome.5–8 In the present study, we observed adenosine-induced reconnection in 46% of PVs. We noted that RF lesions caused PV conduction block in association with extreme membrane depolarization that produced cellular inexcitability. We also observed recovery of membrane potential rapidly on adenosine administration and more slowly in its absence to the point that excitability returned in some veins. Our study raises 2 potential explanations for the relationship between adenosine-induced restoration of excitability and long-term success of PVI. One is that adenosine simply mimics the hyperpolarizing effect that occurs spontaneously with time, allowing for more rapid identification of PVs in which spontaneous reconnection would be observed if they were followed long enough in the hours after the initial procedure. The second possible explanation is that PVs that are more strongly depolarized (and therefore fail to reconnect with adenosine) are more severely damaged and less likely eventually to recover conduction. Further work is indicated to define more clearly the relationship between acute and long-term PV reconnection and to understand better the significance of the extent of PV depolarization.
Novelty and Potential Clinical Relevance
Our study is the first to examine systematically the effects of adenosine on PV cellular electrophysiology, ion currents, conduction, and the response to PVI. Previously, authors have suggested that membrane hyperpolarization and APD shortening by adenosine may facilitate electrotonic conduction.5,7 Adenosine-induced changes in autonomic tone have also been implicated7,25 but were clearly not involved in the present study because adenosine-induced reconnection was observed in the absence of autonomic innervation in our in vitro preparations. In humans, dormant conduction demonstrated during sinus rhythm could have been indirectly promoted by adenosine-induced sinus cycle-length prolongation if disconnection was due to rate-dependent LA-PV conduction block.6 However, in the present study, all preparations were continuously paced at 2 Hz, indicating that heart rate slowing is not necessary for adenosine-induced reconnection.
Tissue injury by RF ablation is recognized to be thermally mediated.26 Hyperthermia significantly changes cardiomyocyte electrophysiological properties, producing potentially reversible (but irreversible when more severe) membrane depolarization and loss of cellular excitability.27 This observation is consistent with our finding of a key role for membrane depolarization in the determination of conduction block and reconnection.
Our observations may have relevance for the design of improved approaches to identify PVs at risk of reconnection. The primary mechanism for adenosine effects on dormant conduction appears to be IKAdo activation in PV cardiomyocytes, which occurs via stimulation of adenosine type-1 (A1) receptors. Adenosine also stimulates A2A, A2B, and A3 receptors, which may not contribute to the drug’s effect on PV excitability and conduction, but rather to the drug’s adverse-effect profile.28 Novel A1 selective receptor agonists are presently under development. Some, such as tecadenoson, selodenoson, and PJ-875, are being developed for the treatment of supraventricular tachyarrhythmias.29,30 Theoretically, these drugs may have several advantages over adenosine, including more sustained and selective cardiac A1 receptor effect and fewer adverse effects (eg, hypotension, flushing, bronchoconstriction, chest discomfort) in the assessment of PVs at risk of reconnection after PVI.
Our findings may also have relevance for understanding adenosine-induced proarrhythmia. Purinergic agonists such as adenosine are well recognized to induce AF episodes in some patients after intravenous administration.31 One potential mechanism is acceleration and stabilization of atrial reentrant rotors by increased inward-rectifier current.32 However, another possibility, based on our studies, would be improved conduction through potential PV drivers that fail to induce AF in the absence of adenosine because of poor PV-LA coupling.
PVI was achieved with RF delivered epicardially to multicellular preparations. This technique is different from clinical practice, in which RF is delivered by endocardially positioned transvenous catheters. Because pulmonary veins are smaller in dogs versus humans, and because we applied RF energy directly to the epicardial surface, we limited RF applications to 25 to 35 W, for only 10 seconds at each site, to control damage. Although in humans RF applications are longer and with higher power, the end point is similar: complete electrical isolation of PVs from LA. In addition, we isolated single PVs with epicardial antral lesions adjacent to the PV-LA junction, different from the circumferential endocardial antral lesions most commonly created in humans. Nevertheless, we did observe conduction block between PV and LA and recovery with adenosine, with properties similar to those noted after PVI in humans.
Although we observed that adenosine acutely restores PV-LA conduction by hyperpolarizing PV cells and thereby enhancing Na+-current availability, other potential mechanisms that we did not study could also be involved. RF-induced tissue edema or inflammation could be reversible and may not be well tracked in all cases by RMP changes. Reversible edema or inflammation could take longer to reverse than RMP, with eventual gaps developing in the RF-induced scar that allows recovery of conduction. Such recurrences may not be well predicted by the acute response to adenosine and might account for cases in which late reconnection occurs despite the absence of an acute adenosine response. Long-term in vivo studies of recurrent PV conduction in experimental models to assess mechanisms of restored PV conduction and the relationship to the acute adenosine response would be of interest.
Ionic currents are sensitive to cell-isolation technique. Great care was therefore taken to record the same ionic currents from similar numbers of PV and LA cells from each dog so that any influence of the isolation procedure would be equally distributed. When recordings could be obtained from cells of only one of PV or LA, they were rejected for analysis. Furthermore, ionic current recordings can change over time due to rundown. Therefore, all of the currents were recorded after the same time intervals and with protocols applied in the same order for both cell types.
There is wide species variability in adenosine sensitivity, and the dog is relatively insensitive,32 so we had to use a relatively high adenosine concentration (1 mmol/L) to reproducibly achieve significant effects. We chose to work with dogs because of the similarity of their PV cardiomyocytes sleeve to humans and because of the appropriateness of their PVs for RF isolation and conduction assessment. The electrophysiological effects and IKAdo properties we observed for adenosine were typical for the compound across a range of species.12,13,28,32 Nevertheless, the differential adenosine sensitivity of canine versus human atrium could affect the applicability of our results.
Adenosine selectively hyperpolarizes canine PV cardiomyocytes compared with LA cells, apparently by selectively increasing IKAdo. Adenosine-induced hyperpolarization increases dV/dtmax by removing voltage-dependent INa inactivation. RF energy isolation of PVs is associated with severe depolarization, and adenosine significantly hyperpolarizes PV cells after PVI. RMP is less depolarized by RF application in PVs with dormant conduction versus PVs with fixed conduction block, allowing adenosine-induced hyperpolarization to voltages negative to −60 mV to restore excitability and PV-LA conduction by removing voltage-dependent INa inactivation. The effects observed in our animal model could explain the restoration of conduction in damaged but viable PVs, thereby potentially accounting for the clinical phenomenon of adenosine-revealed dormant conduction.
The authors thank Nathalie L'Heureux and Chantal St-Cyr for technical assistance, Roxanne Gallery and Ana Fernández for secretarial help, France Thériault for excellent manuscript preparation, and Annik Fortier for statistical support.
This article was supported by grants from the Canadian Institutes of Health Research (Awards MOP 44365, MGP 6957), the Quebec Heart and Stroke Foundation, the Mathematics of Information Technology and Complex Systems (MITACS) Network of Centers of Excellence, the European-North American Atrial Fibrillation Research Alliance (ENAFRA) network award from Fondation Leducq (07-CVD-03), the Ministerio Español de Sanidad y Consumo, Instituto de Salud Carlos III (in collaboration with Fundación para la Investigación Biomédica del Hospital Gegorio Marañón), Red RECAVA, and the Spanish Society of Cardiology.
Haissaguerre M, Shah DC, Jais P, Hocini M, Yamane T, Deisenhofer I, Chauvin M, Garrigue S, Clementy J. Electrophysiologic breakthroughs from the left atrium to the pulmonary veins. Circulation. 2000; 102: 2463–2465.
Oral H, Knight BP, Tada H, Ozaydin M, Chugh A, Hassan S, Scharf C, Lai SW, Greenstein R, Pelosi F Jr, Strickberger SA, Morady F. Pulmonary vein isolation for paroxysmal and persistent atrial fibrillation. Circulation. 2002; 105: 1077–1081.
Cappato R, Negroni S, Pecora D, Bentivegna S, Lupo PP, Carolei A, Esposito C, Furlanello F, De Ambroggi L. Prospective assessment of late conduction recurrence across radiofrequency lesions producing electrical disconnection at the pulmonary vein ostium in patients with atrial fibrillation. Circulation. 2003; 108: 1599–1604.
Nanthakumar K, Plumb VJ, Epstein AE, Veenhuyzen GD, Link D, Kay GN. Resumption of electrical conduction in previously isolated pulmonary veins: rationale for a different strategy? Circulation. 2004; 109: 1226–1229.
Tritto M, De Ponti R, Salerno-Uriarte JA, Spadacini G, Marazzi R, Moretti P, Lanzotti M. Adenosine restores atrio-venous conduction alter apparently successful ostial isolation of the pulmonary veins. Eur Heart J. 2004; 25: 2155–2163.
Hachiya H, Hirao K, Takahashi A, Nagata Y, Suzuki K, Maeda S, Sasaki T, Kawabata M, Isobe M, Iesaka Y. Clinical implications of reconnection between the left atrium and isolated pulmonary veins provoked by adenosine triphosphate after extensive encircling pulmonary vein isolation. J Cardiovasc Electrophysiol. 2007; 18: 392–398.
Matsuo S, Yamane T, Date T, Inada K, Kanzaki Y, Tokuda M, Shibayama K, Miyanaga S, Miyazaki H, Sugimoto K, Mochizuki S. Reduction of AF recurrence after pulmonary vein isolation by eliminating atp-induced transient venous re-conduction. J Cardiovasc Electrophysiol. 2007; 18: 704–708.
Kneller J, Ramirez RJ, Chartier D, Courtemanche M, Nattel S. Time-dependent transients in an ionically based mathematical model of the canine atrial action potential. Am J Physiol Heart Circ Physiol. 2002; 282: H1437–H1451.
Li D, Zhang L, Kneller J, Nattel S. Potential ionic mechanism for repolarization differences between canine right and left atrium. Circ Res. 2001; 88: 1168–1175.
Yue L, Feng J, Gaspo R, Li GR, Wang Z, Nattel S. Ionic remodeling underlying action potential changes in a canine model of atrial fibrillation. Circ Res. 1997; 81: 512–525.
Workman AJ, Kane KA, Rankin AC. Ionic basis of a differential effect of adenosine on refractoriness in rabbit AV nodal and atrial isolated myocytes. Cardiovasc Res. 1999; 43: 974–984.
Belardinelli L, Shryock JC, Song Y, Wang D, Srinivas M. Ionic basis of the electrophysiological actions of adenosine on cardiomyocytes. FASEB J. 1995; 9: 359–365.
Nattel S, Maguy A, Le Bouter S, Yeh Y-H. Arrhythmogenic ion-channel remodeling in the heart: heart failure, myocardial infarction, and atrial fibrillation. Physiol Rev. 2007; 87: 425–456.
Matsuda JJ, Lee HC, Shibata EF. Acetylcholine reversal of isoproterenol-stimulated sodium currents in rabbit ventricular myocytes. Circ Res. 1993; 72: 517–525.
Verma A, Kilicaslan F, Pisano E, Marrouche NF, Fanelli R, Brachmann J, Geunther J, Potenza D, Martin DO, Cummings J, Burkhardt JD, Saliba W, Schweikert RA, Natale A. Response of atrial fibrillation to pulmonary vein antrum isolation is directly related to resumption and delay of pulmonary vein conduction. Circulation. 2005; 112: 627–635.
Ouyang F, Antz M, Ernst S, Hachiya H, Mavrakis H, Deger FT, Schaumann A, Chun J, Falk P, Hennig D, Liu X, Bansch D, Kuck KH. Recovered pulmonary vein conduction as a dominant factor for recurrent atrial tachyarrhythmias after complete circular isolation of the pulmonary veins: lessons from double Lasso technique. Circulation. 2005; 111: 127–135.
Cheema A, Dong J, Dalal D, Marine JE, Henrikson CA, Spragg D, Cheng A, Nazarian S, Bilchick K, Sinha S, Scherr D, Almasry I, Halperin H, Berger R, Calkins H. Incidence and time course of early recovery of pulmonary vein conduction after catheter ablation of atrial fibrillation. J Cardiovasc Electrophysiol. 2007; 18: 387–391.
Wang XH, Liu X, Sun YM, Gu JN, Shi HF, Zhou L, Hu W. Early identification and treatment of PV re-connections: role of observation time and impact on clinical results of atrial fibrillation ablation. Europace. 2007; 9: 481–486.
Nath S, Lynch C III, Whayne JG, Haines DE. Cellular electrophysiological effects of hyperthermia on isolated guinea pig papillary muscle: implications for catheter ablation. Circulation. l993; 88: 1826–1831.
Ellenbogen KA, O'Neill GO, Prystowsky EN, Camm JA, Meng L, Lieu HD, Jerling M, Shreeniwas R, Belardinnelli L, Wolff AA, TEMPEST study group. Trial to evaluate the management of paroxysmal supraventricular tachycardia during an electrophysiology study with tecadenoson. Circulation. 2005; 111: 3202–3208.
Froldi G, Belardinelli L. Species-dependent effects of adenosine on heart rate and atrioventricular nodal conduction: mechanism and physiological implications. Circ Res. 1990; 67: 960–978.
Ablation procedures are becoming an increasingly important therapeutic option for atrial fibrillation, particularly the paroxysmal form. A major limitation to such procedures is arrhythmia recurrence, often attributable to recurrent pulmonary-vein conduction, or so-called “reconnection.” Intravenous adenosine administration at the time of an ablation procedure can cause acute pulmonary-vein reconnection, revealing “dormant” conduction, and is sometimes used as a test of the ability of veins to undergo later reconnection to guide additional ablation lesions. This experimental study aimed to clarify the unknown mechanism by which adenosine reveals dormant conduction. Arterially perfused canine left atrial/pulmonary vein preparations were exposed to adenosine and radiofrequency ablation during action-potential and extracellular-electrode monitoring. Adenosine alone selectively hyperpolarized pulmonary-vein cardiomyocytes sleeves, increased their phase 0 upstroke velocity, and accelerated pulmonary vein/left atrial conduction. Radiofrequency ablation to the antral region adjacent to the pulmonary vein/left atrial junction produced pulmonary vein/left atrial conduction block and depolarized pulmonary-vein cardiomyocytes near the ablation line. Adenosine caused acute reconnection by hyperpolarizing damaged pulmonary-vein cells sufficiently to restore excitability. Similar phenomena occurred spontaneously in some veins during a prolonged (mean 3-hour) observation period. Adenosine effects were reversible on washout, and adenosine-exposed dormant conduction was a good predictor of subsequent spontaneous reconnection in the same vein. Voltage-clamp experiments suggested that the primary ionic-current mediator of adenosine’s actions was adenosine-induced inward-rectifier potassium current. These studies provide insights into the mechanisms by which adenosine acutely restores pulmonary vein/left atrial conduction after radiofrequency ablation, with potential implications for our understanding of mechanisms of pulmonary-vein disconnection and reconnection.
The online-only Data Supplement is available with this article at http://circ.ahajournals.org/cgi/content/full/CIRCULATIONAHA.109. 893107/DC1.