• Open Access

Transvenous Electrical Cardioversion of Equine Atrial Fibrillation: Patient Factors and Clinical Results in 72 Treatment Episodes


  • All work on this project was performed at the Ontario Veterinary College. This study has been presented in part as an abstract at the ACVIM forum in Charlotte, NC, in June 2003 and at the Cardiostim 2004 conference in Nice, France in June 2004.

Corresponding author: M. Kimberly J. McGurrin, Department of Clinical Studies, Ontario Veterinary College, University of Guelph, Guelph, ON, Canada N1G 2W1; e-mail: mcgurrin@uoguelph.ca.


Background: Transvenous electrical cardioversion (TVEC) has been developed for treatment of atrial fibrillation (AF) in horses. The relationship among patient variables, treatment response, and outcome in a typical referral population has not been evaluated.

Hypothesis: Patient variables such as age, sex, weight, and duration of arrhythmia affect prognosis for response to treatment and the energy level at which cardioversion occurs.

Animals: TVEC was applied to 72 episodes of lone AF in 63 client-owned performance horses, with the majority (54) being Standardbred racehorses.

Methods: Catheterization of the right atrium (RA) and pulmonary artery (PA) through the jugular vein was used for electrode placement before horses were placed under general anesthesia. Biphasic, truncated exponential shock waves were delivered at incremental energy levels until cardioversion was achieved or a maximum single-energy level of 300 J was reached (cumulative energy 50–1,960 J). A multivariate model was constructed to evaluate influence of patient factors on cardioversion energy.

Results: Cardioversion was achieved in 71 of 72 episodes (62 of 63 horses) at a mean energy of 165.43 ± 8.75 J. Cardioversion energy was higher for females than for males, and for interaction terms, weight was negatively related to energy in females and positively related in males. Age was positively related to cardioversion energy in females. No relationship was identified between duration of arrhythmia before treatment and prognosis for response or cardioversion energy.

Conclusions and Clinical Importance: TVEC is highly effective in the treatment of lone AF in horses. Although age and sex influence cardioversion energy level, duration of arrhythmia does not.

Atrial fibrillation (AF) is the most common clinically relevant dysrhythmia in the horse.1–4 Affected horses are incapable of achieving maximum performance. Management has traditionally involved PO or IV administration of a quinidine salt,1–4 and is effective in approximately 80% of cases. However, drug therapy involves a range of dose-related and idiosyncratic toxic responses varying from mild and benign to fatal.1–4

Electrical conversion of AF is a routine procedure in human medicine, with an efficacy often higher than pharmacologic methods for restoration of sinus rhythm (SR).5 Technical and methodologic considerations in the application of direct current shocks to the heart have been reviewed recently.6,7

Preliminary investigation by the authors into cardiac catheterization with purpose-designed cardioversion catheters and application of electrical shock to horses in chronic AF determined that cardioversion using this technique was possible, and that when electrodes were correctly positioned, the technique was safe. Reports describing the 1st 3 successful cardioversion attempts,6 and the technique and its development7,8 have been published. The present report describes the outcome of a clinical trial in which a standardized technique was applied to 72 clinical episodes of AF in 63 horses over 76 treatment events. Because treatment was successful in 62 of 63 horses and 71 of 76 treatment events, the primary outcome of interest for purposes of statistical analysis was conversion energy within clinical episode. The goal of the study was to determine whether patient factors influenced energy level required for successful cardioversion. Each of the patient factors, weight, age, sex, and duration of arrhythmia, and their interactions were statistically evaluated to determine influence on cardioversion energy.

Materials and Methods

With the exception of 1 horse given digoxin, the administration of antiarrhythmic medications at any stage of the procedure was specifically avoided so that effects of treatment could be unambiguously assessed. The protocol used in this study was approved by the Animal Care Committee at the University of Guelph.

Study Subjects

The study group consisted of 63 horses examined over 72 clinical episodes of AF, with median duration of arrhythmia for episodes being 4 weeks (reference range, 1–104 weeks). Horses were enrolled over a period of 3 years. All horses presenting to the hospital for evaluation of AF that were found on examination to have lone AF (72 presentations) were included in the study. Seventy presentations were for a complaint of poor performance and 2 were presentations after the detection by the referring veterinarian of a dysrhythmia without performance effects. Horses in AF and showing clear structural or functional evidence of organic heart disease (eg, chamber enlargement with or without valvular disease, signs of heart failure) were not included in the study. Fifty-four horses were Standardbred racehorses, 2 Thoroughbred racehorses, 1 Clydesdale, and 6 warmbloods. Eighteen horses were mares, 34 were geldings, and 11 were stallions. All horses were client-owned and enrolled in the study after full procedural disclosure to owners. Mean age of horses at time of AF episode was 4.81 ± 0.27 years (reference range, 2–10 years), and mean weight was 486 ± 10 kg (reference range, 375–885 kg). To the extent that group sizes permitted comparison, breed or use and duration of AF did not appear to be related in this group of horses. The overall number of electrical treatment events was 76, with 4 horses undergoing 2 treatment events within the same clinical episode. Eight horses were treated for 2 separate clinical episodes, and 1 for 3 episodes. Fourteen horses had previously undergone unsuccessful quinidine treatment, whereas an additional 6 had experienced adverse responses to quinidine. One horse had previously been treated successfully with quinidine with no adverse effects.

One horse had a left-sided grade II/VI diastolic murmur at the cranial heart base, 1 had a grade II/VI left-sided holosystolic murmur at the heart base, 3 had loud (grade III–IV/VI) right-sided holosystolic murmurs at the heart base. One horse had tachycardia, with a resting rate of 56 beats/min and episodes of severe tachycardia with rates of 100–230 beats/min (lasting from 2 to 30 minutes on an overnight Holter monitor). The remaining 57 horses had no evidence of cardiovascular disease on clinical examination at presentation other than arrhythmia. No clinically relevant abnormalities were detected by echocardiographic examination among these horses. The horse with a diastolic murmur had moderate aortic regurgitation, whereas the horse with a left-sided systolic murmur had moderate left atrioventricular (AV) valvular regurgitation on echocardiographic examination. The 3 horses with loud right-sided murmurs had severe right AV valvular regurgitation.

Stable AF was confirmed by ECGa as the cause of the arrhythmia in all clinical episodes. Heart rate was normal in all but 1 horse, where intermittent tachycardia was present. This horse was digitalized before treatment and tachycardic episodes abated (heart rate, 44 beats/min). All horses were examined by echocardiography.b Cardiac structure was evaluated from standardized right and left long and short axis views. Color-flow Doppler echocardiography was performed to evaluate for valvular regurgitation. The inner edge method was used to assess cardiac dimensions. Aortic and pulmonic dimensions were measured from right long axis views. Left atrial and ventricular dimensions as well as left ventricular functional indices were evaluated from the left. Values were compared with those established for the hospital population, which were adapted from experience with reference to established values.9–11 Cardiac dimensions were within normal limits. Functional indices were within normal limits. Mild right AV valvular regurgitation was present in 50 horses. Trivial to mild left AV valvular regurgitation, without changes in chamber dimensions or function, was present in 16 horses. Increased velocity (>3 m/s) across the right AV valve was detected by continuous wave Doppler evaluation in 2 horses. Trivial regurgitation at the pulmonic valve, aortic valve, or both was detected in 23 horses.

Complete physical examination with particular focus on the cardiovascular system was performed the day before treatment. Baseline blood samples were collected for CBC, serum biochemical profile, plasma fibrinogen assay, and plasma cTnI enzyme immunoassayc or cTnT ELISAd (cTnI detection limits: 0.3–50 ng/mL, %CV = 10% at 0.16 ng/mL; cTnT detection limits: 0.04–25 ng/mL, %CV = 10% at 0.04 ng/mL). Feed was withheld for a minimum of 6 hours.


Catheterization was achieved by a standardized protocol, as previously described.7,8 Briefly, 2 identical single electrode 7.5 F, 150 cm, open lumen catheterse with 10 cm titanium wire coil cardioversion electrodes were placed aseptically through the jugular vein. The right jugular was used in 74 treatment events and the left in 2 because of thrombosis of the right jugular vein. Catheters were placed through 10 g 7.5 cm over-the-needle catheters.f Electrode positioning used a combination of ultrasonographic and pressure profile guidance. Pulmonary artery (PA) catheters were advanced smoothly and without hesitation under ultrasound guidance, with images obtained using a right thoracic window and long axis views, and were guided into the left in preference to the right PA. The left PA was preferentially selected because of evidence in humans for lower energy requirement for successful cardioversion with left versus right PA electrode positions.12–14 Left PA placement could not be obtained in 5 treatment events, and the PA catheter thus was placed in the right PA in 5/76 treatment events and in the left PA in 71/76. In all cases, the proximal end of the electrode was positioned 10–15 cm distal to the pulmonic valve, and thus with the entire electrode deep in the PA. Right atrial catheters were positioned in the right atrium (RA) by advancing them until a right ventricular pressure profile was obtained, and then withdrawing the catheter incrementally until the ventricular profile was lost. Immediately after placement, all cardioversion catheters were fixed in place by suturing them to the skin. All cardioversion catheters were placed in the standing horse, facilitating rapid catheterization and electrode positioning and minimizing duration of general anesthesia.


After catheter placement, general anesthesia was induced. Horses were sedated with xylazineg (0.5–1.2 mg/kg) before induction. Induction was achieved with combinations of IV ketamineh with or without diazepam,i glyceryl guaiacolate,j or both, and maintenance anesthesia was then initiated with isofluranek (1.5–3%) inhaled in 100% O2 delivered via a circle breathing system. Intravenous lactated Ringer's solutionl was administered throughout the procedure. Dobutamine,m if required, was administered IV to maintain arterial blood pressure but was withdrawn a minimum of 5 minutes before administration of electrical shocks. The anesthesia protocols are described in a separate report.15

Electrode Placement Confirmation by Radiography and Ultrasonography

After induction, lateral thoracic radiography was used to confirm electrode placement. Catheters were repositioned as necessary to confirm placement of the RA electrode in the RA and the PA electrode sufficiently deep in the PA that its shadow did not overlie the cardiac silhouette on radiography.

Cardioversion Attempts

A surface, base-apex ECG was attached. Cardioversion catheters then were connected to a biphasic electrical defibrillator,n with the RA electrode comprising the cathode and the PA electrode the anode. The device delivers an asymmetrical biphasic truncated exponential shock wave. To avoid shock-induced ventricular arrhythmias, shocks were delivered synchronously with the R wave. Step-wise increases in shock energy levels were used in all horses. To minimize total delivered energy, steps were not even. Initial energy level was 50 J. Shocks were then applied at 2-minute intervals at incremental increases in delivered energy (50, 70, 100, 125, 150, 175, 200, 250, and 300 J [maximum cumulative energy 1,420 J]) until either cardioversion was achieved or maximum single-shock energy of 300 J was delivered. In 6 instances, catheter position was re-evaluated by radiography after unsuccessful delivery of a 300 J shock, and the PA catheter was repositioned. Two to 3 additional shocks at energies between 200 and 300 J were then administered (maximum total energy 1,960 J). Cardiac rhythm was monitored by continuous ECG throughout the procedure.

After shock delivery, catheters were removed and horses recovered from anesthesia. Radiography was used before catheter removal to reconfirm electrode placement in the 1st 10 horses.

Postprocedural Monitoring

After anesthetic recovery, horses were monitored for 6 hours by hourly heart rates, clinical monitoring, then every 4 hours for 24 hours, then twice daily for 7 days in 3 treatment events, and for 2 days in 73 events. ECG was performed every 4 hours in the 1st 8 events and daily in the remaining events. Echocardiography was performed on postprocedural days 2 and 7 in 3 treatment events and on day 2 only in 73. Echocardiography was performed to evaluate for pericardial effusion, valvular regurgitation, and left ventricular function (ejection fraction and fractional shortening). Both right and left parasternal windows were used and short and long axis views collected. Horses were discharged from the hospital after 2–7 days. Postprocedural heparinized blood samples at 4, 8, 12, 16, 20, 24, 36, and 48 hours in 5 events and at 6 and 24 hours in 71 events were submitted simultaneously with preprocedural samples for troponin I (cTnI) or T (cTnT) concentrations. cTnI was used for the 1st 18 AF events and cTnT for the remainder. Troponin assays were performed at the Laboratory Reference Centre, Hamilton, ON, Canada. Testing was switched to cTnT because of test availability at that laboratory. Serum biochemical profiles were performed 48 hours after cardioversion in 38 treatment events.

All horses were monitored for return to performance and adequacy of performance by evaluation of performance records and client communication.

Statistical Method

Initially, the outcome of interest was whether or not cardioversion was achieved. However, because all but 1 clinical episode of AF responded to treatment and only 5 of 76 treatment events were unsuccessful, analysis on this basis did not yield useful information. Cardioversion energy within clinical episode thus was defined as the outcome of interest for purposes of statistical analysis. This approach allowed significance of explanatory variables specific to treatment event to be assessed. The order in which horses were presented for treatment was included in the analysis to identify evidence of a learning effect in application of the standardized technique.

The relationship between outcome and the 4 explanatory variables (weight, age, sex, and duration of arrhythmia) was examined by starting with a full 2nd-order model. This model had as its components the main effects of the 4 explanatory variables, all interactions between pairs of variables, and, for continuous variables, quadratic terms that allowed for curvilinear effects. The model was simplified using a combination of backward and forward elimination, using an α level of 0.05 except for quadratics, where a level of 0.10 was used. After final simplification, a lack-of-fit test16 was applied to ensure no statistically important terms were excluded from the model. The assumptions of a normal-theory regression were assessed using residual plots, and a formal test of normality of residuals was applied.16 Analyses were conducted using Proc GLM (general linear models) in SAS 8.2. In the results, means are stated with standard errors.


SR was achieved in response to electrical shock in 71 of 72 clinical episodes (71 of 76 treatment events), in 67 on the 1st treatment attempt, and in 4 on a 2nd attempt. Analysis of residuals revealed them to be normally distributed and the assumptions of a normal-theory regression model to be met. The only independent variables entering the model were sex and the interaction terms age by sex and weight by sex. Significant differences between geldings and stallions were not observed and these were grouped together as males. Results indicated that cardioversion energy was higher for females than for males (P= .002), and that for interaction terms, weight was negatively related to energy in mares (P= .014) and positively related in males (P= .0003). Age was positively related to cardioversion energy in females (P= .012, Fig 1A). The relationship between duration of arrhythmia before treatment and cardioversion energy was not statistically significant at the 5% level (P= .403).

Figure 1.

 Relationship between age, weight, and TVEC cardioversion energy for males and females. These graphs show the relationship between cardioversion energy, age, and body weight for females (A) and males (B), together with the regression equations describing the relationships. For females, both age (P= .012) and weight (P= .014) influence energy, with higher energies being required for older, lighter mares. For males, age has little effect (P= .6898) and required energy increases with body weight (P= .0003).

When first placed in the standing horse, catheters were successfully positioned using ultrasonography with the PA electrode in the left PA in 71 of 76 treatment events. In 4, the catheter remained in the right PA despite attempts to reposition it. In the remaining event, the catheter folded within the PA and was repositioned under radiographic and echocardiographic guidance, but final electrode positioning within the left versus right PA could not be confirmed.

Electrode repositioning was performed after induction and radiography in 23 of 76 treatment events, in 19 to reposition the PA electrode, in 1 to correct the RA electrode, and in 3 to correct both. Twenty cases involved advancing the PA catheter deeper into the PA. In 1 case, the catheter was redirected into a different branch of the PA; so its radiographic shadow did not overlie that of the RV, and in 1 case, coiling of the PA catheter within the RV was corrected. The RA catheter had to be withdrawn slightly (<10 cm) in 2 horses and advanced slightly (<10 cm) in 2.

The defibrillator accurately identified the R wave in all cases. In 75%, the defibrillator also intermittently identified the T wave as an R wave. Shock delivery on the T wave could be avoided by pressing the shock button after the observation of a T wave, allowing the shock to be delivered synchronously with the next R wave.

Direct cardioversion to normal SR was achieved in 69 treatment events. In 2 cases, a brief (<2 seconds) episode of atrial flutter was noted immediately after shock application and before onset of SR. No other rhythm abnormalities were associated with application of shocks. In 6 treatment events, shocks at 250, 300 J, or both were repeated after PA catheters had been advanced by 10 cm: cardioversion then was achieved in 3. Energy levels for successful cardioversion ranged from 50 to 300 J (median 150 J, mean 166.6 ± 8.36 J). Energy per kilogram body weight ranged from 0.1 to 0.8 J (median 0.31 J, mean 0.35 ± 0.02 J), whereas cumulative energy delivered (sum of all shocks) ranged from 50 to 1,970 J (median 495 J, mean 697.8 ± 58.8 J). Too few cases had the PA electrode positioned within the right PA to allow comparison of cardioversion energies between right and left PA, but cardioversion was achieved from both positions. Troponin I and T concentrations remained at or below the lower level of detection in all horses.

Isolated premature atrial contractions (PACs; 1–3/minute) were noted in 8 horses after cardioversion to SR, and were observed throughout the remaining anesthetic monitoring period. A single PAC was noted within 5 seconds of successful cardioversion in 5 other horses. PACs were not detected on ECG or auscultation after anesthetic recovery. AF recurred in 1 horse upon withdrawal of the RA catheter. The RA catheter was placed again, a single 250 J shock applied, and SR restored.

Anesthesia was well tolerated in all horses. All horses recovered well from anesthesia. Mild tachycardia, muscle swelling, or lameness attributed to myopathy was noted in 8 horses. All horses were clinically normal at 24 hours after recovery.

No abnormalities were detected on repeated clinical examination during the 2–7 postprocedural days for which horses were closely monitored, except for 1 horse that had periodic severe tachycardia before treatment. This horse had episodes of supraventricular tachycardia postcardioversion. Tachycardic episodes were not detected on ECG evaluation and physical evaluations performed at 6-hour intervals for 48 hours. A 24-hour Holter monitor placed 48 hours after treatment indicated a resting rate of 36–44 beats/min and a rate of 60–108 beats/min during tachycardic periods. Six episodes of tachycardia were recorded. This horse was discharged on a decreasing dose regimen of prednisolone (480 mg PO q12h daily for 7 days, then q24h for 7 days, and then q48h for 14 days). Re-evaluation at 4 weeks postcardioversion indicated normal SR at rest. However, on overnight Holter recordings, ectopic activity persisted and 10 tachycardic episodes were noted. Further examination of this horse was not possible, but the horse has been reported as performing well over the past 2 years. Overnight Holter recordings were evaluated in 3 additional horses. In all of these, SR was maintained throughout. Second degree AV block was noted intermittently in all horses and was considered to be a normal finding. No new abnormalities were detected on echocardiography in any horse, and all dimensions and functional indices were within normal limits in all horses. Pericardial effusion was not observed.

At the time this report was compiled, successfully treated racehorses had returned to full race training in 56/64 treatment events, and 40 of these were actively racing at at least the same level of competition as before the onset of arrhythmia. Three treated racehorses not racing were lost to long-term follow-up; 2 of these had not raced before the AF episode. Five horses had been treated within 2 weeks of this report being compiled, and were in active training but had not raced. All racehorses returned to training within 1 week of treatment. No concerns were reported by any owners. AF recurred in 11 instances, between 3 and 147 weeks postcardioversion (median, 18 weeks). Eight horses were successfully treated with electrical cardioversion a 2nd time, and 1 of these also was successfully treated a 3rd time. Treatment was not pursued in the remaining 2 instances of recurrence.


The catheter placement technique used in all cases in this study allowed electrodes to be placed with confidence. Ultrasonographic visualization of catheter progressed through the heart and past the bifurcation of the PA was achieved without difficulty. Failure to direct the catheter into the left PA in 5 treatment events reflected difficulty steering the catheter tip, and was not caused by inadequate visualization. The importance of electrode position and ready availability of suitable radiographic equipment encouraged routine use of thoracic radiography to confirm electrode location. Radiographic data additionally allowed ongoing evaluation of the relationship among response to shock delivery, delivered energy, and electrode position.

During development of this technique, it was determined that cardioversion was most likely to be successful when the PA catheter was advanced sufficiently far that the electrode did not overlie the cardiac silhouette on a lateral thoracic radiograph. Efforts therefore were made to achieve this positioning when first placing catheters in the horses reported here. Repositioning of the PA electrode after thoracic radiography was elected in 23 of 76 treatment events. Because we did not attempt cardioversion in these events before repositioning, we do not know whether cardioversion might have been achieved from the initial position. Aside from potential impacts on treatment success, however, repositioning may reduce the potential for stunning of the AV node, a complication that might result if the principal current pathway passes through the base of the ventricular chambers.17

Electrode placement, timing of shock delivery, and patterns of underlying atrial electrical activity may be factors contributing to the observed variation in energy required to achieve cardioversion. Patterns of atrial activity were not evaluated in this study; however, direction and rate of atrial activation may influence whether sufficient susceptible atrial myocardium can be depolarized by a set energy level. The clinical relevance of deep positioning of the PA electrode is assumed to relate to the volume and location of myocardial tissue traversed by the depolarizing current, and thus the probability that random atrial electrical activity will be interrupted and SR re-established. Neither radiography nor ultrasound examination can confirm lateral electrode position, and the role played by distance between electrodes in determining ease of cardioversion could not be directly assessed. This factor and the effect of stage of the respiratory cycle at the moment of shock delivery require further investigation. Whether the PA electrode travels dorsally, horizontally, or ventrally does not appear to be important, because horses were successfully cardioverted from all orientations.

The need to apply shocks under general anesthesia may be seen as a disadvantage of transvenous electrical cardioversion (TVEC) when compared with pharmacologic management. However, the outcomes in the present case series, in particular the high response rate, evident absence of any complications or adverse effects, and rapid return to training and active competition, are seen as more than compensating for potential risks of general anesthesia.15 Anesthetic time in this clinical investigation was acceptable in most cardioversion attempts, but was occasionally prolonged by the need to reposition electrodes. Times have been shortened as experience has been gained in electrode placement and could likely be shortened further. More horses will need to be treated before the relative risks of traditional drug versus electrical treatment can be assessed.

High cumulative energy can cause myocardial damage, a source of potential risk to the patient.18 The protocol applied in this study results in a maximum cumulative energy of 1,420 J. Cardioversion was achieved in 3 of the 6 instances where repeated shocks were delivered after catheter repositioning, but these repeated shocks increased cumulative energy. Although no evidence of myocardial damage was identified here, we advise, based on the results of this study, that a cumulative energy of 1,420 J not be exceeded.

Premature atrial activity is reported after electrical cardioversion in humans,19 and was noted in 8 horses immediately after cardioversion and for the 1st 12 hours postrecovery in 1 horse. This activity could represent initial electrical instability within the atria, unidentified myocardial damage, or pre-existing foci of ectopic activity. Further evaluation by Holter monitoring might aid in this determination.

In this study, cardiac troponin assay was used to screen for myocardial injury, and the absence of any detectable increase was encouraging. However, use of this assay in the horse is a recent innovation.20–23 cTnI and cTnT concentrations are considered the preferred biochemical markers of cardiac injury in humans,24–26 and conservation of troponin structure across species allows application in the horse of tests designed for human use.26–28 Antibodies used in the TnT and some TnI tests have been shown to detect equine cTnT or cTnI, respectively,26,27 validating the tests inasmuch as they are able to detect equine troponins.

Issues remain in the use of troponin assays in horses. Each TnI test kit measures different epitopes and fragments of cTnI, and up to a 100-fold difference in reported concentrations can occur depending on the kit used.29,30 Appropriate cut-off values for each assay are unique and results cannot be compared between assays,29 and each published report on horses used a different cTnI assay. The laboratory used in this study switched tests from cTnI to cTnT, necessitating cTnT testing in the majority of horses. Currently, there is only 1 assay for cTnT,29 and in humans, this assay is considered to have cardiac specificity equivalent to the assays for cTnI.29 In the present study, normal values for the horse were extrapolated from the human medical literature as being at or below the lower detection limits for the assays. This approach was justified on the basis that we used each horse as its own control and have found concentrations at or below detection limits in clinically normal horses in our hospital, whereas increases have been observed in diseased horses. We have observed increases in cTnI with aortic root rupture (2.1 and 1.3 μg/mL) and in cTnT with subacute myocarditis (4.96 ng/mL at admission, 0.15 ng/mL on day 6, <0.04 ng/mL on day 8) with the tests used in this study. Increases in cTnT were also detected in a study evaluating cardiac biomarkers in foals.23 Also, in the present study, our primary interest lay in determining whether detectable increases in troponin occurred with TVEC.

Troponin kinetics after myocardial injury also have not been evaluated in the horse. Troponin proteins have a short half-life in plasma, but continued release from damaged cells results in persistently increased concentrations.31 Detectable concentrations often are not present for 3 hours after injury,31 and increased serum troponin is first detected 4–6 hours after myocardial damage, peaks after 12–18 hours, and remains increased for 5–9 days.31 Troponin changes have been investigated after electrical cardioversion in humans,25,32 and increases have been detected in some studies,25 but increases were not detected in the horses of the present study. Because initial sampling at 4-hour intervals disclosed no abnormalities, sampling intervals were increased and cTnI samples at 6 and 24 hours were considered appropriate to evaluate for myocardial damage. Because troponin kinetics in humans are similar for cTnI and cTnT, the sampling interval of 6 and 24 hours was also used when the test was switched to cTnT. It is acknowledged, however, that because cTnI/cTnT concentrations were below the detection limit in this study, slight increases above baseline might not have been detected. Troponin testing was used as an adjunct to clinical, electrocardiographic, and echocardiographic examination. In the absence of abnormalities detected by these means and on the basis of return to full performance, clinically relevant myocardial damage seems unlikely. Regardless, more information on normal troponin concentrations in horses is required to confirm the absence of transient increases.

The biological importance of the observed relationships between patient characteristics and cardioversion energy is unclear. The impact of sex on cardioversion energy is difficult to explain, and the finding that increasing body weight for mares was associated with decreasing cardioversion energy seems counterintuitive. It is possible that variations in precise electrode position and in patterns of atrial electrical activity from event to event could be a source of sufficient variation as to prevent identification of precise relationships between patient characteristics and cardioversion energy. The relationships between patient factors and cardioversion energy as discovered in this study are unlikely to be of clinical importance.

Duration of arrhythmia and size of horse have been identified as prognostic indicators in assessing probable response to treatment with quinidine.1,3,4 However, absence of statistically significant impact of duration of arrhythmia in the present case series suggests duration may not be a reliable prognostic indicator for cardioversion when using TVEC. A larger sample of horses with longstanding AF would be needed to fully evaluate this hypothesis. Ease of conversion conceivably could be related to duration of arrhythmia within the 1st 2 weeks after onset but not thereafter. In other species, electrical remodeling is such that after 2 weeks of AF the atrial myocardium is predisposed to the arrhythmia,33,34 but the range of patient characteristics in the present study population did not allow such a relationship to be identified.

Ease of conversion must be considered together with overall prognosis. Performance records indicated that for the majority of racehorses, prognosis for athletic performance was excellent. Not all horses could be followed, however, and recurrence of arrhythmia may have played a role in withdrawal of some horses from racing. Confirmed recurrence of AF in 10 horses (11 recurrences with 1 horse having 2 recurrences) in this study is consistent with anticipated outcome in horses treated conventionally,1,4 and is not considered necessarily to relate to the use of TVEC. Additional data on long-term outcome need to be gathered and compared with an appropriately selected control population to evaluate if performance is equivalent to that of unaffected horses. Although TVEC represents a therapeutic option for all horses, including those not amenable to other treatments, subsequent outcome is unlikely to differ from that for animals successfully treated conventionally. Because duration of AF did not predict treatment outcome in this case series, it may be possible to achieve cardioversion in horses that have had AF for months to years and for which treatment has not been considered. However, such animals may not necessarily return to work or perform sufficiently well to justify the investment. The impact of duration of AF before treatment on prognosis for long-term maintenance of SR is the subject of ongoing investigation in this hospital.

The findings of the present investigation suggest that when treatment by TVEC is available, none of the variables evaluated here should deter attempts at cardioversion. However, critical evaluation of the success of TVEC awaits results of its application in a broader range of horses, particularly older and heavier animals. Only 3 treatments were carried out in horses >600 kg in this case series. Also, no comments can be made concerning treatment of AF in animals with organic heart disease.


TVEC is an effective procedure for restoration of SR in horses with AF in which there is no evidence of underlying structural heart disease. Successful application of the technique to horses in which treatment with quinidine salts has either been unsuccessful or associated with unacceptable toxicity, and subsequent return of these horses to their previous level of athletic activity indicates that the treatment may be of particular value in animals for which there were previously no therapeutic options. Duration of arrhythmia may not have any impact on the probability that cardioversion can be achieved. However, long-term outcome in animals with a prolonged history of AF is undetermined.


aEK 10 ECG, Burdick, WA

bVivid 7, GE Medical Systems, Saskatoon, SK, Canada

cAxSym-Abbott Laboratories Ltd, Saint-Laurent, QC, Canada

dElecsys system, Boehringer Mannheim, Indianapolis, IN

eCustom catheter, Rhythm Technologies Inc, Irvine, CA

fAngiocath, Becton Dickinson and Company, Franklin Lakes, NJ

gRompun, Bayer Inc, Toronto, ON, Canada

hKetalean, Bimeda-MTC Animal Health Inc, Cambridge, ON, Canada

iDiazepam, Sabex Inc, Boucherville, QC, Canada

jGuaifenesin powder, Rhodia Canada, Mississauga, ON, Canada

kIsoflurane, Bimeda-MTC Animal Health Inc

lLactated Ringer's Injection USP, Baxter Corporation, Toronto, ON, Canada

mDobutrex, Abbott Laboratories Ltd

nLifepak 12 3D, Medtronic Physio-Control, Redmont, WA


The early phase of this project was supported by the Grayson-Jockey Club Foundation and the Ontario Ministry of Agriculture, Food and Rural Affairs.

The authors would like to acknowledge William Sears for his assistance with the statistical analysis, Medtronic Emergency Response System and Dr Fred W. Chapman for loan of the defibrillator used in this project, and Cesar Diaz and Peter Accorti from Cardiac Output Technologies Inc for assistance with cardioversion catheters. The invaluable assistance of clients who agreed to enroll their horses in the study is gratefully acknowledged.