Available online at www.sciencedirect.com

ScienceDirect Journal of Electrocardiology 47 (2014) 316 – 323 www.jecgonline.com

Rapid slowing of the atrial fibrillatory rate after administration of AZD7009 predicts conversion of atrial fibrillation☆,☆☆,★ Maria Aunes, MD, a,⁎, 1 Kenneth Egstrup, MD, PhD, b, 1 Lars Frison, PhD, a, 1 Anders Berggren, MD, PhD, a, 1 Martin Stridh, PhD, c, 1 Leif Sörnmo, PhD, c, 1 Nils Edvardsson, MD, PhD d, 1 a

AstraZeneca R&D, Mölndal, Sweden OUH Svendborg Sygehus, Svendborg, Denmark c Department of Electrical and Information Technology and Center for Integrative Electrocardiology, Lund University, Lund, Sweden d Sahlgrenska Academy at the Sahlgrenska University Hospital, Göteborg, Sweden b

Abstract

Background: Effects on the atrial fibrillatory rate (AFR) were studied during infusion with the combined potassium and sodium channel blocker AZD7009. Methods and Results: Patients with persistent atrial fibrillation (AF) were randomized to AZD7009 or placebo. Thirty-five patients converted to sinus rhythm (SR) and were matched to 35 nonconverters. The mean AFR before conversion was 231 fibrillations per minute (fpm), having decreased by 41%; in non-converters, it was 296 fpm at the end of infusion, having decreased by 26%. The rate of decrease was greater in converters at 5 min, − 88 vs. − 66 fpm (p = 0.02), and at 10 min, − 133 vs. − 111 fpm (p = 0.048). The AFR-SD and the exponential decay decreased. A small left atrial area was the only baseline predictor of conversion to SR. Conclusions: AZD7009 produced a significantly more rapid decrease of the AFR in converters than in non-converters, but the AFR at baseline was not predictive of conversion. © 2014 Elsevier Inc. All rights reserved.

Keywords:

Atrial fibrillation; Atrial fibrillatory rate; AZD7009; Conversion

Introduction The duration of ongoing atrial fibrillation, AF, [1,2] and underlying comorbidities are among the factors that influence atrial remodeling [3,4] and the atrial fibrillatory rate, AFR. The AFR is high, typically between 350 and 600 fibrillations per minute (fpm). The AFR is in the lower range in short-lasting AF and in the higher range after longer periods of AF, and may therefore be predictive of conversion to sinus rhythm (SR), both spontaneous [5–7], pharmacological [8–11] and during catheter ablation [12,13]. A low ☆

The study was funded by AstraZeneca R&D, Mölndal, Sweden. Conflict of interest disclosures: M.A., L.F., and A.B. are employees of AstraZeneca R&D, Mölndal, Sweden. K.E., M.S., L.S., and N.E. have nothing to disclose. ★ Clinical Trial Registration: ClinicalTrial.gov NCT00255281. ⁎ Corresponding author at: AstraZeneca R&D, S-431 83 Mölndal, Sweden. E-mail address: [email protected] 1 This author takes responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation. ☆☆

0022-0736/$ – see front matter © 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jelectrocard.2013.12.008

AFR at baseline would suggest conversion, while a high AFR would indicate a low likelihood of restoration and maintenance of SR [12,13]. If this is a consistent finding, the AFR could play an important role in the selection of treatment in individual patients. A lower AFR in short-lasting AF has been suggested to be related to an initially low number of functioning wavelets, while atrial remodeling during long-lasting AF would gradually increase the number of wavelets and thereby the AFR. Theoretically, any treatment that eliminates a sufficient number of wavelets could decrease the AFR so that AF terminates and no longer can be sustained. A decrease of the AFR may be associated with a gradual organization of the atrial rhythm, which may facilitate conversion [9,14,15]. In a small crossover study in patients with permanent AF, the mean AFR rapidly and significantly decreased on AZD7009, but not on placebo [16]. Our aim was to analyze the effects of the combined potassium and sodium current blocker AZD7009 on the AFR, AF rhythm organization, and its relationship to various clinical variables and its predictivity of pharmacological conversion in patients with persistent AF.

M. Aunes et al. / Journal of Electrocardiology 47 (2014) 316–323

Patients and methods Eligible patients were between 18 and 80 years of age, adequately anticoagulated according to guidelines, with a clinical indication for cardioversion of AF and a duration of the current AF episode between 48 h and 90 days. Exclusion criteria included the following: cardioversion of AF/atrial flutter (AFL) in the last 4 weeks or unsuccessful cardioversion of the current episode; left ventricular ejection fraction (LVEF) b 0.35; clinically significant valvular heart disease; hypotension (systolic blood pressure b 90 mm Hg); QTc (Bazett) N 450 ms; any QRS N 150 ms; clinically significant sinus and/or atrioventricular node dysfunction; bradycardia b 40 bpm; history of torsade de pointes, other polymorphic ventricular tachycardia (VT) or sustained monomorphic VT; history or family history of long QT syndrome; concomitant medication with QT-prolonging drugs; and serum potassium concentration outside 3.8– 5.0 mmol/L. Drugs known to prolong the QT interval were to be stopped at least 5 half-life values before the study (3 months for amiodarone). The present study is a retrospective analysis of patients from a double-blind, randomized, placebo-controlled, parallel-group

317

study [17]. In the main prospective study three AZD7009 infusion regimens were designed to reach different target plasma concentrations: a 30-min infusion at 3.25 mg/min (total dose 98 mg, target maximum plasma concentration 2.5 μmol/L); a 15-min infusion at 3.25 mg/min infusion (total dose 49 mg, target maximum plasma concentration 1.8 μmol/L); and a 15-min infusion at 4.4 mg/ml (total dose 66 mg, target maximum plasma concentration 2.5 μmol/L). There were two corresponding 30- and 15-min placebo infusion regimens. The infusion was stopped if successful conversion occurred, defined as termination of AF followed by at least 2 min of stable SR. If patients did not convert to SR within 90 min of the start of infusion, they underwent biphasic DC cardioversion. The study protocol was developed in accordance with the ACC/AHA/ESC 2006 guidelines [18] but is also compatible with the latest versions of the guidelines for the management of AF [19,20]. The study was carried out in accordance with the 1975 Declaration of Helsinki, and was approved by appropriate local and national ethics review boards. All patients provided written informed consent. The results of the main study, including the proportion of patients who converted to SR within 90 min from the start of infusion and the time to conversion were presented

Main study (n = 159*) Treatment regimens

Total number of patients (n)

Converters (n)

Nonconverters (n)

30-min inf usion of AZD7009 3.25 mg/min

42

21

21

15-min inf usion of AZD7009 3.25 mg/min

39

7

32

15-min inf usion of AZD7009 4.40 mg/min

36

7

29

30-min inf usion of Placebo

22

0

22

15 min inf usion of Placebo

20

0

20

Study on atrial fibrillatory rate (n = 92)

Treatment regimens

Total number of patients (n)

Converters (n)

Nonconverters (n)

30-min inf usion of AZD7009 3.25 mg/min

42

21

21

15-min inf usion of AZD7009 3.25 mg/min

14

7

7 (matched)

15-min inf usion of AZD7009 4.40 mg/min

14

7

7 (matched)

30-min inf usion of Placebo

22

0

22

Fig. 1. Study population.

318

M. Aunes et al. / Journal of Electrocardiology 47 (2014) 316–323

Table 1 Demographic characteristics (n, mean ± SD or %) at baseline in AZD7009 converters and non-converters and patients receiving placebo. Variable

n Age, years Gender, male % AF episode duration, days AF rate, fpm AF rate SD Exponential decay LVEF, % LA area, cm2 Beta-blocker, % ACE inhibitor, % Angiotensin ATI receptor antagonist, % Statin, % Angina pectoris, % Myocardial infarction, % Heart failure, % Diabetes mellitus, % Hypertension, % CHADS2 score CHA2 DS2-VASc score

AZD7009 3.25 mg/min/30 min

AZD7009, all patients

Converters

Non-converters

Converters

Non-converters

21 61 ± 14 67 28 ± 23 403 ± 63 26 ± 7 1.3 ± 0.2 59 ± 11 22 ± 6 90 48 0 24 5 5 14 14 62 1.3 ± 1.3 2.1 ± 1.8

21 66 ± 9 76 48 ± 23 394 ± 58 23 ± 9 1.2 ± 0.4 56 ± 9 28 ± 7 95 52 24 29 5 5 14 19 76 1.6 ± 1.1 2.5 ± 1.5

35 63 ± 12 63 33 ± 24 393 ± 59 27 ± 7 1.2 ± 0.3 57 ± 11 22 ± 6 94 60 0 26 6 9 17 14 74 1.4 ± 1.2 2.4 ± 1.8

35 64 ± 9 77 44 ± 23 400 ± 56 26 ± 10 1.2 ± 0.3 56 ± 8 27 ± 6 89 51 17 23 9 3 26 20 71 1.5 ± 1.0 2.3 ± 1.5

Placebo

22 58 ± 9 86 47 ± 23 410 ± 33 23 ± 6 1.4 ± 0.3 57 ± 8 26 ± 5 64 36 18 14 5 0 14 5 68 0.9 ± 0.6 1.3 ± 1.0

AF = atrial fibrillation; SD = standard deviation; LVEF = left ventricular ejection fraction; LA = left atrium; ACE = angiotensin-converting enzyme.

elsewhere [17]. The aim of the present study was to assess changes in the AFR in patients receiving AZD7009 and placebo, in particular what characterized patients who converted and who did not convert to SR on AZD7009. The study population consisted of 92 patients; 35 who converted to SR on AZD7009, 35 patients who did not convert on AZD7009 and 22 patients who received placebo. Among them were all 42 patients receiving the 30-min infusion regimen of AZD7009 3.25 mg/min (21 converters and 21 non-converters), 14 patients receiving the 15-min infusion regimen of AZD7009 3.25 mg/min (7 converters and 7 matched non-converters), 14 patients receiving the 15min infusion regimen of AZD7009 4.40 mg/min (7 converters and 7 matched non-converters) and all 22 patients receiving the 30-min infusion regimen of placebo (Fig. 1). The matching between each pair of converter/non-converter was made such that the current episode of AF duration always differed by b 1 week in conjunction with differences in atrial size b 10 cm 2.

ECG recording and analysis The 12-lead ECG was recorded continuously from 5 min before until 3 h after the end of infusion in converters and up to 5 h in patients who did not convert pharmacologically. Five-minute recordings were also performed at 3.5, 4, 4.5, 5, 5.5, 6, 9 and 24 h after the end of infusion. The ECG signal (digitized at 1-kHz sampling rate) was transferred to a computer for off-line processing. The ECG was processed using techniques described in detail elsewhere [21–23]. QRS detection and beat classification were performed so that averages of similar-shaped beats could be computed (one set of averages for each 5-min segment). Atrial activity in lead V1 was extracted using spatiotemporal

QRST cancellation in which the average beats of adjacent leads were mathematically combined with the average beat of the analyzed lead in order to produce optimal cancellation [21]. Since the AFR and harmonics are confined to frequencies below 25 Hz, the atrial signal was down-sampled to 50 Hz and subjected to time–frequency analysis [22]. A spectral estimate of the atrial signal was determined every second in overlapping 2.56-s windows using a short-term, non-uniform (logarithmic) Fourier transform. Before spectral analysis each window was evaluated with respect to its total power and maximum absolute amplitude for the purpose of identifying and excluding large QRST-related residuals and other disturbances. The exponential decay (ED) of the harmonic magnitudes was determined from a running spectrum (“spectral profile”) representing a measure of atrial signal organization. Organized rhythms have stronger harmonics and a lower ED while disorganized rhythms have a higher ED due to the broader spectral content with weaker harmonics. The mean AFR, the AFR standard deviation (AFR-SD), the ED, the RR interval and heart rate were determined in 1min segments, except the baseline AFR which was determined as the mean of the 5 min preceding the start of infusion. Each consecutive minute in AF was compared to that of the baseline. The AFR-SD at baseline and during each consecutive minute in AF was taken as a measure of the regularity of the atrial rhythm.

Statistical methods Student's t-test was used for comparing continuous variables between treatment groups, while Fisher's exact test was adopted for binary variables. Analyses of potential predictors for a reduction in AFR were based on multivariate linear regression, while corresponding analyses with

M. Aunes et al. / Journal of Electrocardiology 47 (2014) 316–323

319

more often had beta-blockers (91% versus 64%, p = 0.004), while the left and the right atrial areas and the LVEF did not differ between the groups (Table 1). There were 35 conversions to SR in the AZD7009 group and no conversion on placebo. The demographics of AZD7009 converters and non-converters are shown in Table 1. The median time to conversion was 19 (range 8– 64) min: 21 (10–64) min in the AZD7009 3.25 mg/30 min, 12 (range 8–43) min in the 3.25 mg/15 min and 16 (range 8–31) min in the 4.4 mg/15 min groups, respectively. The median doses of AZD7009 were 78 (range 35–98, theoretical maximum 98) mg, 49 (range 35–49, theoretical maximum 49) mg and 66 (range 53–66, theoretical maximum 66) mg, respectively.

Fig. 2. Converters to SR (upper panel, n = 21) and non-converters (middle panel, n = 21) in the 30-min 3.25 mg/min AZD7009 infusion group, and patients randomized to placebo (lower panel, n = 22). The blue dotted vertical line at 30 min marks the end of infusion. The red diamonds in the upper panel signify the time of conversion.

drug-induced conversion to SR as the dependent variable utilized multivariate logistic regression. The analysis of potential predictors was based on the 42 patients receiving the 3.25-mg/min 30-min group. A p value b 0.05 (two-sided) was considered statistically significant. Results Demographics Patients who received AZD7009 were older than those who received placebo (63 [SD 11] [range 20–78) vs. 58 [SD 9] [range 45–72] years, p = 0.025), and 70% versus 86% of the patients were men. Patients randomized to AZD7009

Fig. 3. Effects on the AFR in relation to the AF episode duration in patients who converted (in green, the diamond sign shows the time of conversion) and did not convert to SR (in red) on AZD7009. The blue dotted vertical line at 30 min marks the end of infusion. Note the more rapid and pronounced decrease of the AFR in the patients with the shortest AF episode durations.

320

M. Aunes et al. / Journal of Electrocardiology 47 (2014) 316–323

Table 2 Mean changes in relation to the current AF episode duration in patients receiving AZD7009. Variable

n ΔAFR ΔAFR-SD ΔED

Time after start of infusion

5 10 5 10 5 10

min min min min min min

Subgroups by the estimated AF episode duration

Relationship to AF episode duration

=7 d

8–30 d

31–60 d

61–90 d

p Valuea

4 − 115 − 158 − 4.7 − 5.1 − 0.3 − 0.6

15 − 77 − 132 − 10.7 − 16.2 − 0.2 − 0.4

30 − 90 − 129 − 9.1 − 11.9 − 0.3 − 0.4

21 − 72 − 118 − 7.4 − 10.6 − 0.2 − 0.4

0.27 0.16 0.76 0.48 0.84 0.71

In spite of small numbers there was a tendency toward a decreasing effect on the AFR by increasing episode duration, but no change was statistically significant. AF = atrial fibrillation; AFR = atrial fibrillatory rate; SD = standard deviation; ED = exponential decay. a Linear regression between ΔAFR (and the same for Δ AFR-SD and ΔED) and the logarithm of AF episode duration.

The AFR and conversion to SR within 90 min after the start of infusion The mean AFR at baseline was 397 ± 57 (range 253– 584) versus 410 ± 33 (range 363–469) fpm, in patients who were randomized to AZD7009 and placebo, respectively. Within 2 min after the start of infusion with AZD7009 the AFR decreased significantly, and the mean AFR before conversion was 231 fpm, corresponding to a decrease of − 162 fpm (41%) from the baseline level. In non-converters on AZD7009 the mean AFR at the end of infusion was 296 fpm, having decreased by−104 fpm (26%). In the 3.25 mg/min 30 min group the rate of decrease of the AFR was significantly greater in converters, showing a mean reduction of − 88 versus − 66 fpm (p = 0.02) after 5 min and − 133 versus − 111 fpm (p = 0.048) after 10 min of the AZD7009 infusion, respectively (Fig. 2). The mean of the last AFR recorded before conversion was 235 (range 183– 327), 230 (range 185–326) and 222 (range 175–278) in the three AZD7009 groups, respectively. Among patients who received AZD7009 the conversion rate was 100% (4/4 patients) in AF episodes lasting ≤ 1 week, 82% (9/11 patients) ≤ 30 days, 55% (16/29 patients) ≤ 60 days and 50% (21/42 patients) ≤ 90 days. The mean baseline AFR levels were 428 fpm (b 7 days), 417 fpm (8–30 days), 382 fpm (31–60 days) and 402 fpm (61–90 days), respectively (Fig. 3). There was no statistically significant difference in the mean change of the AFR at 5 and 10 min according to duration of the AF episode (Table 2). Twenty-one, 7 and 7 patients from the 30-min 3.25 mg/ min, 15-min 3.25 mg/min and 15-min 4.40 mg/min infusion groups, respectively, did not convert on AZD7009 and 17, 6 and 3 of them were successfully DC cardioverted to SR. All 22 placebo patients underwent DC cardioversion, which was successful in 16 (73%) of them. Among patients who underwent successful and unsuccessful DC cardioversion after AZD7009 infusion, the mean AFR at baseline was 395 fpm (n = 25) and 414 fpm (n = 9), respectively, and at the end of infusion, it was 296 fpm and 296 fpm, respectively. The corresponding values for patients undergoing DC cardioversion after placebo infusion were 419 fpm (n = 14) and 389 fpm (n = 6) at baseline and 415 fpm and 377 fpm at the end of infusion.

If the last AFR before conversion was not carried forward, a response to increase in exposure to AZD7009 was demonstrated in non-converters. The AFR returned to the baseline level within 6 h in patients who received the full infusion and did not at all convert to SR. In patients who received placebo, the AFR remained unchanged at around 410 fpm before, during and after the infusion. The atrial rate variability (AFR-SD) The mean AFR-SD at baseline was similar in patients randomized to AZD7009 and those randomized to placebo, 26 versus 23. In the AZD7009 3.25 mg/min 30 min group the AFR-SD decreased to 19 within 4 min and reached its

Fig. 4. Mean levels for AFR-SD (upper panel) and ED (lower panel) in converters and non-converters on AZD7009 and on placebo. Note, that the effects on the AFR-SD and the ED were not different between converters and non-converters.

M. Aunes et al. / Journal of Electrocardiology 47 (2014) 316–323

minimum value of 12 after about 18 min and stayed there for the rest of the 30-min period in those who did not convert (Fig. 4). The AFR-SD did not change during placebo infusion. The exponential decay The ED was somewhat lower at baseline in patients randomized to AZD7009 than those randomized to placebo, 1.2 versus 1.4, i.e. more organized, although the AFR values were comparable. During the 30-min infusion of AZD7009, the ED decreased and reached its lowest mean level of 0.7 after 12 min, which was maintained through the infusion in non-converters, while it remained unchanged in the placebo group (Fig. 4). Thus the ED reached its lowest level sooner than the AFR and the AFR-SD. There was a correlation between delta AFR and delta ED (10 min after start of infusion), for the 70 AZD7009 patients, (− 0.24, p = 0.051). If the last value before conversion was not carried forward, the lowest ED was 0.6 after 31 min (n = 40). The ED returned to the baseline level within 4 h after the start of the infusion. The ED did not change during placebo infusion. Predictor analysis All baseline variables presented in Table 3 were assessed as potentially prognostic for the effects on the AFR during the first 5 and 10 min during the AZD7009 infusion in the AZD7009 3.25 mg/30 min group. The estimated change, expressed as percentage reduction in AFR by each unit increase in the predictor. Four variables were associated with the decrease of AFR during the first 5 min with p values b 0.05: higher heart rate SD at baseline and higher ED at baseline with a greater change, and previous myocardial infarction (MI) and the use of calcium channel blocker with a smaller change (Table 4). In a multivariate linear regression analysis including all the above variables only heart rate SD at baseline was maintained in the final model (with p b 0.05); no other variables remained significant. Six variables were associated with a decrease of AFR during the first 10 min with p values b 0.05: higher AFR at baseline, higher AFR-SD at baseline, higher heart rate at baseline, higher heart rate SD at baseline, and higher ED at baseline with a greater change, and previous MI with a smaller change (Table 4). In a multivariate linear regression analysis including all the above variables heart rate SD at baseline (− 0.47 [− 0.77, − 0.16], p = 0.0037) and AFR-SD at baseline (− 0.43 [− 0.63, − 0.22], p = 0.0001) were maintained in the final model. Logistic regression was performed using 21 variables at baseline (Table 3) for being potential predictors for intravenous conversion to SR. A small left atrial area was the only variable that predicted conversion to SR by AZD7009 versus placebo (OR 0.85; p = 0.002). Discussion The highest conversion rate was seen in the AZD7009 3.25 mg/30 min group, preceded by a rapid and significantly more pronounced decrease in the AFR, occurring at 5 and 10 min into the infusion, than in non-converters. A similar

321

pattern was seen in the other two groups of AZD7009 dose regimens. A short-duration AF episode predicted an effect on the AFR, but a small left atrial area was the only baseline variable that predicted conversion to SR [17]. All three AZD7009 treatments resulted in highly significant decreases in the AFR and increased organization of the atrial rhythm as reflected by the ED and the AFR-SD. Pharmacological cardioversion occurs because an agent interferes with the arrhythmia mechanism by means of its mode of action. However, a very pronounced decrease in AFR was no guarantee for conversion, implying that the substrate becomes unresponsive at some point in time, probably because of continuing electromechanical remodeling. Considering the decrease in AFR and the role of the estimated AF episode duration, the effect seemed to be greater the shorter the duration of the episode, even within the short duration range in our study population. It has been suggested that patients with a low AFR at baseline would be more prone to cardioversion, assuming that the AFR reflects the duration of the AF episode [8–11]. Thus, an AFR less than 360 fpm at baseline was highly sensitive and predictive of AF termination following intravenous ibutilide [24], and an AF cycle length N 145 ms (corresponding to an AFR less than 405 fpm) was associated with conversion by pilsicainide [9]. In the present study, the AFR at baseline ranged between 253 and 584 fpm with 62% (n = 59/89) of the patients having an AFR less than 405 fpm and only 19% (n = 17/89) of the patients less than 360 fpm, and conversions by AZD7009 occurred irrespective of the baseline AFR. In our study population of patients with AF episodes shorter than 90 days we did not find the baseline AFR to be useful in selecting

Table 3 Baseline variables investigated as potential predictors for the reduction in AFR and for the probability of AZD7009 induced conversion. Variable

Units

Age Gender AFR, mean AFR-SD Exponential decay Estimated AF episode duration Time since first diagnosed AF episode Left ventricular ejection fraction, LVEF Left atrial area Right atrial area Heart rate, mean Heart rate, SD Serum potassium ACE inhibitor Angiotensin AT1 inhibitor Beta-blocker Statin Digoxin Calcium channel blocker Angina pectoris Heart failure

years male/female fpm fpm ratio days years % cm2 cm2 bpm bpm mmol/L no/yes no/yes no/yes no/yes no/yes no/yes no/yes no/yes

All variables relate to information available at baseline. AFR = atrial fibrillatory rate; SD = standard deviation; ACE = angiotensin-converting enzyme.

322

M. Aunes et al. / Journal of Electrocardiology 47 (2014) 316–323

Table 4 Univariate analysis of potential predictors for the reduction in AFR during the first 5 and 10 min of AZD7009 infusion in the AZD7009 3.25 mg/30 min group. Variable

Heart rate, SD Previous MI Exponential decay CCB usage AFR-S D Heart rate, SD Previous MI AFR, mean Exponential decay Heart rate, mean

Unit

bpm no/yes ratio no/yes fpm bpm no/yes fpm ratio bpm

Time after start of infusion

Estimated change (percent reduction) Estimate

95% CI

p Value

5 5 5 5 10 10 10 10 10 10

− 0.60 10.1 − 7.0 5.5 − 0.46 − 0.53 12.1 − 0.04 − 7.7 − 0.13

− 0.94 to − 0.26 1.3 to 18.9 − 13.1 to − 0.8 0.0 to 10.9 − 0.68 to − 0.24 − 0.90 to − 0.17 3.3 to 20.9 − 0.08 to − 0.01 − 13.9 to − 1.5 − 0.25 to − 0.01

0.001 0.025 0.027 0.049 0.0002 0.005 0.008 0.010 0.017 0.044

min min min min min min min min min min

The estimated change expressed as percent reduction in AFR by each unit increase in the predictor is reported. Only variables associated with p b 0.05 are included in the table. AF = atrial fibrillation; AFR = atrial fibrillatory rate; SD = standard deviation; MI = myocardial infarction; CCB = calcium channel blocker.

individuals for a pharmacological conversion attempt. Nevertheless, the magnitude of the effect on the AFR seemed to be greater in patients with the shortest AF episode duration, implying a role of remodeling and that early treatment has a greater chance of success. In comparison with other agents, the conversion rate with AZD7009 was high, especially in patients with duration of the AF episode longer than 1 week [17]. Unfortunately, the development of AZD7009 was discontinued because of unexplained flu-like symptoms appearing after repeated administration with an extended release formulation in elderly patients. A decrease in the AFR and/or an increased organization of the atrial rhythm was observed with intravenous flecainide [14], intravenous ibutilide [24], oral flecainide [12,25,26], oral amiodarone [12], oral bepridil [8–10,26], and oral pilsicainide [9,26]. In the only study that compared agents with class I (flecainide) and class III action (bepridil) in paroxysmal AF, the AFR decreased only with flecainide, suggesting that bepridil was effective mainly via its class III mode of action [26]. This could in turn imply that the decrease in AFR caused by AZD7009 was achieved by means of sodium rather than potassium channel blockade. In the present study, beta-blockers, ACE inhibitors or statins did not predict an effect on the AFR or conversion to SR. The AFR-SD and the ED reflect the degree of organization of the atrial rhythm, and both responded quickly and significantly within minutes after the start of any of the AZD7009 infusions. However, ED was not correlated with the AFR, not even at the highest rates of about 580 fpm. Furthermore, there was no indication that the AFR-SD or the ED at baseline was a reliable discriminator between patients suitable for active treatment or not. The AFR at baseline was not different in patients who converted to SR by AZD7009, who were DC cardioverted after failing to convert on AZD7009, who were DC cardioverted on placebo, or who did not convert at all. Even though the AFR at baseline was a predictor of decrease during the first 10 min of infusion, the only predictor of conversion to SR was a small left atrial area. It was not possible, in this study population, to identify an AFR cut-off limit at baseline that would be useful as a single criterion for

decision about treatment. A report of successful catheter ablation in a patient with drug-resistant persistent AF and a median AF cycle length of 120 ms (corresponding to an AFR of 500 fpm) also supports that treatment, at least in selected patients, may be meaningful in spite of a high AFR [27]. However, the degree by which the AFR decreased, but not the AFR at baseline, seems to reflect aspects of atrial remodeling and may be helpful in clinical decision making, together with other available patient information.

Limitations The results are representative for the selected patient population and we cannot exclude that the effects of AZD7009 on the AFR could be different in patients with much shorter episodes of AF. The different dosing regimens of AZD7009 reflect the dose-finding purpose of the main study and may be seen as a limitation in the current study. Whenever relevant all 70 AZD7009 patients were included in the analyses, but in the predictor analyses only the patients in the AZD7009 3.25 mg/30 min group were included, while 28 patients from the 15-min groups were excluded, due to the matched-pairs selection.

Conclusions The AFR consistently decreased in response to AZD7009 but the rate of the decrease was significantly higher in converters to SR. The AFR at baseline was not predictive of conversion by AZD7009. An estimated short duration of the AF episode correlated significantly with an AZD7009induced decrease of the AFR, but a small atrial area was the only baseline variable that predicted AZD7009-induced conversion to SR. Acknowledgments We are grateful to Emma Samuelsson, MSc, for providing statistical analysis. The authors adhere to the principles of ethics and publishing of scientific articles [28].

M. Aunes et al. / Journal of Electrocardiology 47 (2014) 316–323

References [1] Husser D, Cannom DS, Bhandari AK, Stridh M, Sörnmo L, Olsson SB, et al. Electrocardiographic characteristics of fibrillatory waves in newonset atrial fibrillation. Europace 2007;9:638–42. [2] Ravelli F, Masé M, Del Greco M, Faes L, Disertori M. Deterioration of organisation in the first minutes of atrial fibrillation: a beat-to-beat analysis of cycle length and wave similarity. J Cardiovasc Electrophysiol 2007;18:60–5. [3] Bollmann A, Husser D, Mainardi L, Lombardi F, Langley P, Murray A, et al. Analysis of surface electrocardiogram in atrial fibrillation: techniques, research, and clinical applications. Europace 2006;8:911–26. [4] Drewitz I, Willems S, Salukhe TV, Steven D, Hoffmann BA, Servatius H, et al. Atrial fibrillation cycle length is a sole independent predictor of a substrate for consecutive arrhythmias in patients with persistent atrial fibrillation. Circ Arrhythm Electrophysiol 2010;3:351–60. [5] Petrutiu S, Sahakian AV, Swiryn S. Abrupt changes in fibrillatory wave characteristics at the termination of paroxysmal atrial fibrillation in humans. Europace 2007;9:466–70. [6] Nilsson F, Stridh M, Bollmann A, Sörnmo L. Predicting spontaneous termination of atrial fibrillation using the surface ECG. Med Eng Phys 2006;28:802–8. [7] Choudhary MB, Holmqvist F, Carlson J, Nilsson HJ, Roijer A, Platonov PG. Low atrial fibrillatory rate is associated with spontaneous conversion of recent-onset atrial fibrillation. Europace 2013;15:1445–52. [8] Fujiki A, Sakamoto T, Iwamoto J, Nishida K, Nagasawa H, Mixumaki K, et al. Pharmacological conversion of persistent atrial fibrillation with and without a history of drug-resistant paroxysmal atrial fibrillation. Circ J 2006;70:1138–41. [9] Fujiki A, Sakabe M, Nishida K, Sugao M, Tsuneda T, Iwamoto J, et al. Drug-induced changes in fibrillation cycle length and organization index can predict chemical cardioversion of long-lasting atrial fibrillation with bepridil alone or in combination with aprindine. Circ J 2004;68:1139–45. [10] Niwano S, Sasaki T, Kurokawa S, Kiryu M, Fukaya H, Hatakeyama Y, et al. Predicting the efficacy of antiarrhythmic agents for interrupting persistent atrial fibrillation according to spectral analysis of the fibrillation waves on the surface ECG. Circ J 2009;73:1210–8. [11] Aoyama Y, Niwano S, Niwano H, Satoh A, Kishihara J, Ishikawa S, et al. Repetitive evaluation of fibrillation cycle length predicts the efficacy of bepridil for interruption of long-lasting persistent atrial fibrillation. Int Heart J 2011;52:353–8. [12] Matsuo S, Lellouche N, Wright M, Bevilacqua M, Knecht S, Nault I, et al. Clinical predictors of termination and clinical outcome of catheter ablation for persistent atrial fibrillation. J Am Coll Cardiol 2009;54:788–95. [13] Haissaguerre M, Sanders P, Hocini M, Hsu L-F, Shah D, Scavée C, et al. Changes in atrial fibrillation cycle length and inducibility during catheter ablation and their relation to outcome. Circulation 2004;109:3007–13. [14] Tuan J, Osman F, Jeilan M, Kundu S, Mantravadi R, Stafford PJ, et al. Increase in organization index predicts atrial fibrillation termination

[15]

[16]

[17]

[18]

[19]

[20]

[21] [22]

[23]

[24]

[25]

[26]

[27]

[28]

323

with flecainide post-ablation: spectral analysis of intracardiac electrograms. Europace 2010;12:488–93. Holmqvist F, Stridh M, Waktare JEP, Roijer A, Sörnmo L, Platonov PG, et al. Atrial fibrillation signal organization predicts sinus rhythm maintenance in patients undergoing cardioversion of atrial fibrillation. Europace 2006;8:559–65. Aunes-Jansson M, Edvardsson N, Stridh M, Sörnmo L, Frison L, Berggren A. Decrease of the atrial fibrillatory rate, increased organization of the atrial rhythm and termination of atrial fibrillation by AZD7009. J Electrocardiol 2013;46(1):29–35. Geller JC, Egstrup K, Kułakowski P, Rosenqvist M, Jansson MA, Berggren A, et al. Rapid conversion of persistent atrial fibrillation to sinus rhythm by intravenous AZD7009. J Clin Pharmacol 2009;49: 312–22. Fuster V, Rydén LE, Cannom DS, Crijns HJ, Curtis AB, Ellenbogen KA, et al. ACC/AHA/ESC 2006 guidelines for the management of patients with atrial fibrillation: full text: a report of the American College of Cardiology/American Heart Association Task Force on practice guidelines and the European Society of Cardiology Committee for Practice Guidelines. Europace 2006;8:651–745. Camm AJ, Kirchhof P, Lip GYH, Schotten U, Savelieva I, Ernst S, et al. Guidelines for the management of atrial fibrillation. Europace 2010;12:1360–420. Camm AJ, Lip GYH, De Caterina R, Savelieva I, Atar D, Hohnloser SH, et al. 2012 focused update of the ESC Guidelines for the management of atrial fibrillation. Europace 2012;14:1385–413. Stridh M, Sörnmo L. Spatiotemporal QRST cancellation techniques for analysis of atrial fibrillation. IEEE Trans Biomed Eng 2001;48:105–11. Stridh M, Sörnmo L, Meurling CJ, Olsson SB. Sequential characterization of atrial tachyarrhythmias based on ECG time–frequency analysis. IEEE Trans Biomed Eng 2004;51:100–14. Husser D, Stridh M, Sörnmo L, Geller C, Klein HU, Olsson SB, et al. Time–frequency analysis of the surface electrocardiogram for monitoring antiarrhythmic drug effects in atrial fibrillation. Am J Cardiol 2005;95:526–8. Bollmann A, Kanuru NK, McTeague KK, Walter PF, DeLurgio DB, Langberg JJ. Frequency analysis of human atrial fibrillation using the surface electrocardiogram and its response to ibutilide. Am J Cardiol 1998;81:1439–45. Husser D, Binias K-H, Stridh M, Sörnmo L, Olsson SB, Molling J, et al. Pilot study: noninvasive monitoring of oral flecainide's effects on atrial electrophysiology during persistent human atrial fibrillation using the surface electrocardiogram. ANE 2005;10:206–10. Yoshida T, Niwano S, Inuo K, Saito J, Kojima, Ikeda-Murakami K, et al. Evaluation of the effect of bepridil on paroxysmal atrial fibrilation. Circ J 2003;67:11–5. Diemberger I, McCready J, Nunn L, Chow AWC. Is atrial fibrillation with very short cycle length suitable for ablation? A case report. Europace 2008;10:1336–9. Shewan LG, Coats AJS. Ethics in the authorship and publishing of scientific articles. Int J Cardiol 2010;144:1–2.

Rapid slowing of the atrial fibrillatory rate after administration of AZD7009 predicts conversion of atrial fibrillation.

Effects on the atrial fibrillatory rate (AFR) were studied during infusion with the combined potassium and sodium channel blocker AZD7009...
635KB Sizes 0 Downloads 0 Views