Myotonic dystrophy (MD) is a serious autosomal-dominant hereditary disease with an estimated incidence of 1 in 8,000 births (1- 3). Death is a major concern in MD, although its reported incidence varies widely, and may be related to muscular weakness or respiratory failure, or sometimes it may be sudden (4- 13). The most often proposed mechanisms of sudden death in MD are atrioventricular (AV) conduction disorders and ventricular tachyarrhythmias (4,6,8- 9,11,14- 21). Practice guidelines recommend the implantation of a cardiac pacemaker (PM) in patients with MD in case of acquired third-degree AV block (class I indication) or if there is the incidental finding, during electrophysiologic (EP) study, of a markedly prolonged HV interval (>100 ms) in asymptomatic patients (class IIa indication) (22).

In our study, we put forward the hypothesis that implantation of a PM in patients with MD with an HV interval ≥70 ms, even asymptomatic, may protect them against the risk of sudden death related to AV block. The aims of this multicenter study were: 1) to prospectively justify the PM indication by automatic recording of spontaneous AV block episodes in patients with MD without previously documented bradycardia; and 2) to use the PM to accurately record the occurrence of cardiac bradyarrhythmias and tachyarrhythmias that may remain undetected during conventional clinical follow-up.

The study was approved by the Ethical Review Committee of the Cochin School of Medicine. After having given their written, informed consent to participate, the patients were enrolled into the study in eight medical centers, between 1993 and 1996. The Appendix lists the institutions and investigators participating in this study. Molecular genetic testing was performed as previously described (23). The diagnosis of MD was confirmed by the finding, on deoxyribonucleic acid extracted from peripheral blood samples, of an expansion of the cytosine-thymine-guanine triplet over 50.

Electrophysiologic testing was performed, as previously described (24), in the presence of symptomatic or asymptomatic AV conduction abnormalities (first-degree or higher AV block and/or widening of the QRS complex >100 ms), palpitations or documented arrhythmias, or near syncope or syncope, or preoperatively in preparation for major surgery under general anesthesia.

The main study inclusion criteria consisted of an HV interval ≥70 ms, even in the absence of related symptoms. Patients with documented bradycardia requiring cardiac pacing could also be included, with the sole objective of recording spontaneous tachyarrhythmias.

Before implantation of a diagnostic dual-chamber PM (Chorus RM 7034, Ela medical, Le Plessis Robinson, France), a 12-lead rest electrocardiogram (ECG), 24-h ambulatory ECG, signal-averaged ECG, transthoracic echocardiogram, and right and left ventricular radionuclide angiograms were obtained.

A specific diagnostic algorithm was downloaded to the random-access memory of the PM, allowing the diagnosis of bradycardias and tachycardias at the ventricular level during the VDI mode (ventricular pacing and sensing plus atrial sensing). Two types of bradycardic events—“pauses” and “bradycardias”—were defined (Figure 1). The diagnosis of “tachycardia,” based on ventricular rate analysis, was made when the following criteria were present: 1) a beat-to-beat heart rate acceleration ≥25% at the initiation of the tachycardia; 2) a heart rate >154 beats/min; and 3) for more than five consecutive cycles. With each event, the time and date were memorized, along with marker chains, all retrievable by PM interrogation. Each marker chain included atrial and ventricular events preceding and following the diagnostic cycle, allowing clarification of the nature of the event (Figure 2).

Follow-up visits were scheduled every six months, at which time the patients’ clinical status and surface ECG were recorded, and a complete PM interrogation was performed, including retrieval of the diagnostic algorithm data. Important clinical events were recorded by interviewing the patient, the family, or the patient’s primary physician. In case of death, if possible, the PM was interrogated to determine whether the death was arrhythmic.

Data are reported as the mean value ± SD. The rates of AV block, sino-atrial (SA) block, and atrial and ventricular tachyarrhythmias were analyzed for the entire follow-up period by life-table analysis. Kaplan-Meier survival curves were constructed; the patients who died or were no longer being followed up were censored. The prognostic values of factors measured at baseline, that could be related to the occurrence of AV block, SA block, and tachyarrhythmias, were tested by the log-rank test. Cox proportional hazards models were used to study the effect of several factors simultaneously and to adjust for the potential confounding effect of age. All computations were performed with the SAS package (SAS User’s Guide Statistics 1990, version 6, SAS Institute Inc., Cary, North Carolina).

The study population consisted of 49 patients (24 men and 25 women; 45.5 ± 8.9 years old). Their main baseline clinical characteristics are summarized in (Table le1). Forty-five patients had no history of documented bradycardia before study entry. In this group, 11 complained of palpitations, 16 experienced syncope or fainting episodes before EP testing, and 25 had no cardiac symptoms and underwent EP testing because of conduction disturbances on the surface ECG. Only two patients had a normal ECG and underwent EP testing for investigation of syncope in one case and for palpitations and near syncope in the other.

Four patients had documented bradycardia before study inclusion: asymptomatic infrahisian 2:1 AV block (n = 1), infrahisian complete AV block with syncope (n = 1), asymptomatic paroxysmal third-degree AV block on the 24-h ambulatory ECG (n = 1), and complete absence of atrial activity, without ventricular escape, following electrical cardioversion of atrial flutter (n = 1).

Over a mean observation period of 53.5 ± 27.2 months (median 57.5 months, maximum 101.4 months), 10 patients died and 4 patients left the study on their physician’s recommendation (n = 2), because of PM memory failure (n = 1), or because of study protocol violation (n = 1). The circumstances of the 10 deaths, including 4 sudden deaths, were clarified in all but 2 cases (Table le2). In two patients who died suddenly, PM interrogation performed shortly after death excluded an arrhythmic death, because no arrhythmia was stored by the diagnostic algorithm. Their heart rate trend only showed a moderate acceleration of sinus rhythm a few minutes before death.

During 425 follow-up visits (range 1 to 18 follow-up visits per patient), 21 patients (43%) reported symptoms consisting of dyspnea (n = 6), palpitations (n = 9), fainting (n = 9), or syncope (n = 1). In 18 patients, no rythmologic explanation was found for symptoms sometimes atypical; in 3 patients, the symptoms coincided with an arrhythmia memorized by the PM: palpitations related to atrial fibrillation (n = 1) or bitachycardia (n = 1) and dyspnea due to 2:1 AV block (n = 1).

In 41 (83.7%) of the 49 patients, marker chains indicative of arrhythmias were retrieved from the PM’s memory. A marker chain of “pause” was present in 23 (51.1%) of 45 patients without previously documented bradycardia (Figure 3) due to high-degree paroxysmal AV block (n = 21) and/or paroxysmal complete SA block (n = 4). The cumulative incidences of pauses at 6, 12, 24, and 48 months were 22.2%, 37.8%, 46.7%, and 48.9%, respectively. Pauses were exclusively nocturnal in 35% of cases of AV block and in all cases of SA block.

Algorithm criteria of “bradycardia” due to marked sinus bradycardia were confirmed in 24 patients (53%), mostly (75.1%) during the night.

In 28 patients (57.1%), “tachycardia” (Figure 4) was recorded due to atrial arrhythmias with rapid AV conduction in 25 patients (51%) and/or ventricular tachycardia in 13 patients (26.5%). The cumulative incidences of tachycardias at 6, 12, 24 and 48 months were 22.4%, 32.7%, 40.8%, and 51%, respectively. Atrial and ventricular arrhythmias were strictly diurnal in 54.5% and 72.7%, respectively, and strictly nocturnal in only 13.6% and 18.2%, respectively.

No statistical association was found between the occurrence of AV or SA block and several variables, including age, gender, severity of muscle disease, rest heart rate, conduction intervals (PR, QRS, and HV, or H duration), corrected sinus node recovery time, SA conduction time, signal-averaged ECG, anti-arrhythmic treatment, or history of syncope or near syncope.

After adjustment for age, the occurrence of atrial arrhythmias was associated with the number of supraventricular extrasystoles recorded on the 24-h ambulatory ECG, with a relative risk of 4.2 (95% confidence interval 1.3 to 12.9) beyond 1,000 extrasystoles per 24 h. Atrial arrhythmias were not associated with age, gender, severity of muscular disease, corrected sinus node recovery time, SA conduction time, right atrial refractory period, induction of atrial arrhythmia by programmed stimulation, left atrial diameter, vital capacity, or history of palpitations.

No direct or age-adjusted statistical association was found between the incidence of ventricular arrhythmias and age, gender, muscular weakness, anti-arrhythmic treatment, ventricular extrasystoles on the 24-h ambulatory ECG, right ventricular refractory period, induction of ventricular arrhythmia by programmed stimulation, results of the signal-averaged ECG, left or right ventricular ejection fraction, or history of palpitations, syncope, or near syncope.

The occurrence of death was positively associated with age (p = 0.005, r = 1.1) and the degree of muscular weakness (p = 0.01, r = 1.76), and negatively correlated with the pulmonary vital capacity (p = 0.01, r = 0.93) and anti-arrhythmic drug treatment with amiodarone or beta-blockers (p = 0.03, r = 0.18).

Among the 45 patients without documented bradycardia before study entry, surface ECG recordings during follow-up visits revealed the development of high-degree AV block in 5 patients at 25, 51, 56, 64, and 88 months of follow-up (Figure 5). Of these five patients, two patients had no history of syncope or fainting; two patients had fainting episodes; and one patient reported syncope before implantation of the PM. In three of five cases, paroxysmal third-degree AV block was detected by the PM algorithm during the first month after implantation. Two patients had no previous marker chain recording of AV block because, against the study protocol, their device had been reprogrammed 11 to 20 months earlier to the DDD mode, which is a mode that does not allow the identification of bradycardic events.

One patient developed permanent SA block on the surface ECG at 81 months, as predicted by the PM that stored a marker chain of complete SA block 16 months after implantation.

Seven patients had atrial fibrillation on routine ECG recordings, becoming chronic in 5 cases after 7 to 77 months of follow-up. In all cases, an episode of paroxysmal atrial fibrillation with a rapid ventricular rate had been previously diagnosed by the PM.

One patient complained of dizziness 71 months after implantation, related to a sustained bitachycardia (monomorphic ventricular tachycardia and atrial flutter) present both on the surface ECG and in the PM’s memory.

To our knowledge, this study is the first to use an implantable diagnostic device providing an accurate diagnosis of spontaneous bradycardias and tachyarrhythmias, even when the patient is asymptomatic, associated with antibradycardia support. This diagnostic tool was applied to patients with MD (Steinert’s disease) in the largest series published thus far of paced patients with this disorder, which is associated with a high risk of arrhythmias and sudden death.

A high incidence of arrhythmic events was observed. Indeed, 41 (83.7%) of the 49 patients had an arrhythmic event recorded by the PM’s diagnostic algorithm during follow-up. High-degree paroxysmal AV block was recorded in 46.7% of patients who had no known bradycardia on study entry. All but one of the bradycardic events were asymptomatic, as they were brief or associated with pauses limited by backup pacing. The PM protected our 49 patients from the consequences of profound bradycardia. Reprogramming the PM to the DDD mode (sequential atrioventricular sensing and pacing) after documentation of paroxysmal AV block prevented subsequent symptomatic bradyarrhythmias. The strictly nocturnal occurrence of some of the episodes of AV block may suggest a functional precipitant—for example, sleep apnea, a disorder often associated with MD (25- 26). However, the diurnal occurrence of most episodes of AV block, as well as prolongation of the HV interval at study entry, points to the importance of infrahisian disease. In addition, the observation of fixed 2:1 or third-degree AV block on the surface ECG in five patients, months after the recording of paroxysmal third-degree AV block by the PM, supports the hypothesis of an organic etiology. The data stored by the PM during follow-up confirm the validity of PM implantation. Considering these results, prophylactic permanent pacing should be considered in patients with MD when the HV interval is ≥70 ms, even without related symptoms, and not when the HV interval exceeds 100 ms, as previously recommended (22).

Severe sinus node dysfunction with paroxysmal third-degree SA block, later observed on the surface ECG in one case, was documented in only 9% of patients, exclusively at night. This observation militates in favor of a functional precipitant, perhaps sleep apnea, although one of the four patients had an abnormal sinus node recovery time on study entry.

The high number of paroxysmal atrial arrhythmias recorded confirms their frequent occurrence in patients with MD (24). Although no statistically significant relationship was established in this study, the high incidence of atrial arrhythmias recorded during follow-up was consistent with the patients’ history of atrial arrhythmias and the results of programmed atrial stimulation before PM implantation. A density of atrial extrasystoles >1,000 per 24 h on the 24-h ambulatory ECG was the only variable associated with the subsequent occurrence of atrial arrhythmias. It is noteworthy that the actual incidence of atrial arrhythmias was probably underestimated, because their detection was limited to episodes associated with rapid AV node conduction, as the diagnostic algorithm accounts only for the ventricular rate. The importance of these atrial arrhythmias is illustrated in the development of chronic atrial fibrillation in 5 patients between 0 and 28 months after the first recorded arrhythmic episode.

Of the several variables examined, the number of atrial extrasystoles on the 24-h ambulatory ECG was the sole predictor of atrial arrhythmias. In view of the small size of our study population, we were able to identify relative risks of 2 and 3 with powers of 20% and 75%, respectively, by statistical analysis, except for ventricular arrhythmias, which occurred so infrequently as to preclude a power calculation. Consequently, predictive factors of an only moderate relative risk may have been missed.

Spontaneous ventricular tachycardia was recorded in 13 patients. Although highly symptomatic ventricular arrhythmias have been described in MD (4,6,8- 9,11,15- 19), no poorly tolerated episode was observed among our patients. Only one patient complained of palpitations during bitachycardia. The predominantly diurnal incidence (72.7%) of these arrhythmias suggests their facilitation by sympathetic activity. Although a negative correlation has been found between age and arrhythmia induction by programmed ventricular stimulation (24), no correlation was found in this study between the results of EP testing and subsequent recordings of ventricular arrhythmias.

All tachyarrhythmias were identified by the PM algorithm before being diagnosed by conventional follow-up methods. Furthermore, when anti-arrhythmic drug treatment was indicated, the PM allowed us to monitor the efficacy of the therapy.

No death related to AV block occurred. Four of the 10 deaths, one of which was sudden, were due to acute respiratory insufficiency. This 40% rate is comparable to the 42% death rate from respiratory failure in the Canadian population described by Mathieu et al. (12) in a cohort of unselected patients with a common ancestor. It is consistent with the finding, in our study, of a statistical link between the risk of death and a lower vital capacity or more severe muscular weakness. As in the Canadian study, we observed a higher incidence of sudden death than that expected in a young population. However, this incidence was considerably higher in our population, reaching 8.2% at 53.5 months, as compared with 2.2% at 120 months in the Canadian study. Furthermore, mortality may have been limited by permanent pacing, preventing death from AV block, as well as by the early treatment of asymptomatic paroxysmal tachyarrhythmias, reducing the risk of a spontaneous adverse evolution. Anti-arrhythmic drug treatment with amiodarone or beta-blockers demonstrated a protective effect in our study.

No arrhythmic cause was identified in three of our four patients who died suddenly. This demonstrates that typical sudden death can be nonarrhythmic in MD.

As expected, we found a high incidence of paroxysmal complete AV block in patients with MD with infrahisian conduction abnormalities, even without history of related symptoms. The prophylactic implantation of a pacing system when the HV interval reaches 70 ms should then be recommended to protect patients against the clinical consequences of paroxysmal profound bradycardia. A PM including detailed diagnostic functions also facilitates the diagnosis and management of frequent paroxysmal tachyarrhythmias that may remain undetected during conventional clinical follow-up.

The authors wish to express their gratitude to Ela Medical for their technical assistance and Cardioscript International for manuscript translation.

FRANCE: Hopital d’Angers, Angers (J. Victor and F. Dupuis); Hopital Gabriel Montpied, Clermont-Ferrand (D. Lamaison); Hopital Louis Pradel, Lyon (H. Bussillet); Hopital Cochin, Paris (A. Lazarus, J. Varin, H. M. Bécane, P. Laforet, B. Eymard, S. Weber, and D. Duboc); Hopital Pontchaillou, Rennes (P. Mabo and J. C. Daubert); Hopital Charles Nicolle, Rouen (F. Anselme and N. Saoudi); Hopital Trousseau, Tours (D. Babuty); and Hopital Rangueil, Toulouse (M. Salvador-Mazenq).