Arrhythmogenic Marker for the Sudden Unexplained Death Syndrome in Thai Men
Background Between 1981 and 1988, the Centers for Disease Control and Prevention reported a very high incidence of sudden death among young male Southeast Asians who died unexpectedly during sleep. The pattern of death has long been prevalent in Southeast Asia. We carried out a study to identify the clinical markers for patients at high risk of developing sudden unexplained death syndrome (SUDS) and long-term outcomes.
Methods and Results We studied 27 Thai men (mean age, 39.7±11 years) referred because they had cardiac arrest due to ventricular fibrillation, usually occurring at night while asleep (n=17), or were suspected to have had symptoms similar to the clinical presentation of SUDS (n=10). We performed cardiac testing, including EPS and cardiac catheterization. The patients were then followed at ≈3-month intervals; our primary end points were death, ventricular fibrillation, or cardiac arrest. A distinct ECG abnormality divided our patients who had no structural heart disease (except 3 patients with mild left ventricular hypertrophy) into two groups: group 1 (n=16) patients had right bundle-branch block and ST-segment elevation in V1 through V3, and group 2 (n=11) had a normal ECG. Group 1 patients had well-defined electrophysiological abnormalities: group 1 had an abnormally prolonged His-Purkinje conduction time (HV interval, 63±11 versus 49±6 ms; P=.007). Group 1 had a higher incidence of inducible ventricular fibrillation (93% for group 1 versus 11% for group 2; P=.0002) and a positive signal-averaged ECG (92% for group 1 versus 11% for group 2; P=.002), which was associated with a higher incidence of ventricular fibrillation or death (P=.047). The life-table analysis showed that the group 1 patients had a much greater risk of dying suddenly (P=.05).
Conclusions Right bundle-branch block and precordial injury pattern in V1 through V3 is common in SUDS patients and represents an arrhythmogenic marker that identifies patients who face an inordinate risk of ventricular fibrillation or sudden death.
In the 1980s, the CDC reported that the number of young male Southeast Asian refugees in the United States who died suddenly and unexpectedly had risen dramatically.1 2 The victims had all been active and apparently healthy before dying suddenly while sleeping at night. The time of death ranged between 10:00 pm and 8:00 am in 82 documented cases. All the decedents, except 1, were men between the ages of 16 and 63 years (median, 32 years). This syndrome, which came to be known as SUDS, had been prevalent for many years in Asia, particularly in the Southeastern region.3 4 5 6 7 8 The native populations of this part of the world have long been familiar with this pattern of death: SUDS was known by the name Lai Tai (“died during sleep”) in northeast Thailand,3 5 6 Bangungut (“moaning and dying during sleep”) in the Philippines,4 and Pokkuri (“sudden unexpected death at night”) in Japan.7 8 In Thailand, SUDS first attracted public attention when accounts were published of young male Thai construction workers in Singapore who died suddenly in 1990.9 The pattern of death in these young Thai male workers was similar to that described in the CDC report. A subsequent epidemiological survey of young Thai men living along the northeastern border of Thailand abutting Laos and Cambodia found that the annual SUDS death rate was 26 to 38 per 100 000 men (range, 20 to 49 years old).5 6 This figure confirmed that SUDS was the leading natural cause of death of young Thai men, and, as such, posed a critical medical and social threat to Thai society.
Although researchers had determined that VF was the rhythm leading to cardiac arrest or death10 11 12 —a characteristic that placed SUDS in the constellation of idiopathic VF13 —SUDS presented a particularly vexing problem for physicians. There were no precipitating factors or premonitory symptoms before death occurred. SUDS victims had no overt structural heart disease. The clinical markers that would identify patients at high risk for SUDS had not been discovered. We consequently carried out a study to determine the prevalence of cardiac abnormalities in patients who run a great risk of developing SUDS, prospectively evaluate their clinical outcome, and identify the clinical markers for SUDS.
From January 1994 through September 1996, we prospectively studied 27 Thai men who were thought by referring physicians to have survived SUDS-like episodes and thus had been identified as possible Lai Tai patients. The protocol required the following inclusion criteria for the study patients: (1) SUDS victims were patients who had been apparently healthy before suddenly developing sudden cardiac arrest due to VF but had been successfully resuscitated, and (2) probable SUDS patients experienced symptoms that reflected the clinical presentation of SUDS3 6 : agonal respiration, unresponsiveness after labored respiration during sleep, transient symptoms of distress (eg, moaning, thrashing, grimacing), and syncope or seizure-like symptoms. Probable SUDS patients did not have documented cardiac arrest or VF before being referred. We excluded patients who had structural heart disease or identifiable causes of VF causing cardiac arrests, such as prolonged QT syndrome, myocardial ischemia, or drug-induced life-threatening arrhythmias.
All patients were transferred to our hospitals for a routine physical examination. Cardiac tests performed included ECG, EPS, exercise treadmill test, Holter monitoring, echocardiographic study, SAECG, and cardiac catheterization, including coronary angiography and left and right ventricular angiography. MRI of the heart was obtained if possible. The patients were followed in the outpatient clinics and were treated with amiodarone, propranolol, or an ICD according to the physician’s discretion. The ICDs that were used were model CPI-P3 (Guidant-CPI, Inc) with a storage memory and shock E-gram (far-field electrogram), which allowed us to determine the precise rhythms before, during, and after defibrillation.
Stimulation Protocol for VT Induction
The stimulation protocol included ventricular stimulation at the right ventricular apex and three cycle length driving trains (normal sinus rhythm or 600-ms ventricular pacing, 500-ms ventricular pacing, and 400-ms ventricular pacing). If VT was not induced at the right ventricular apex, then the right ventricular outflow tract was used for induction in the same manner. Isoproterenol was not used for the induction. Induced arrhythmias were defined as (1) VF (≥300 bpm), (2) polymorphic VT (<300 bpm), or (3) monomorphic VT (<300 bpm). “Sustained” was defined as lasting 30 seconds or requiring earlier termination because of hemodynamic compromise. “Nonsustained” was defined as inducible arrhythmias of at least 15 beats but less than 30 seconds.
SAECG was analyzed by the time domain method at the 40-Hz filter. Abnormal values are (1) total duration of the SAECG-QRS complex of >114 ms, (2) RMS voltage in the terminal 40 ms of the SAECG-QRS complex (v40) of <20 μV, and (3) duration of low-amplitude signals, with 40 μV in the terminal portion of the SAECG-QRS complex of >38 ms. Our protocol required at least two of these three criteria for a positive SAECG.
Data and Statistical Analysis
Our primary variables for clinical outcomes during the follow-up period include the development of VF, cardiac arrest, or death. Two-sample t test was used to compare the differences in the mean values of the two groups, and a Kaplan-Meier life-table analysis was used to determine the differences in event-free survival rates between the two groups. The Fisher exact test was used for comparing the incidence of inducible VT/VF, positive SAECG, and clinical occurrence of VF or cardiac arrest between the two groups.
We studied 27 male patients (mean age, 39.7±11 years), which included 17 SUDS survivors and 10 probable SUDS patients. All patients had normal cardiac function (mean left ventricular ejection fraction, 64±9%) and no evidence of structural heart disease. The mean interventricular septal and posterior wall thicknesses were 1.1±0.2 and 0.99±0.1 cm, respectively; only 3 patients had a small increase in the wall thickness (>1.1 cm), suggesting mild LVH. All patients had normal coronary arteries and a normal right ventricle. All patients had normal exercise treadmill tests. Five patients, all of whom were SUDS survivors, had normal cardiac MRI results.
Of the 17 SUDS survivors who developed cardiac arrest due to VF before the study, 12 developed VF while sleeping between 8:00 pm and 6:00 am. Two patients suffered VF arrest while awake at 10:00 pm and 1:00 am, respectively. The other 3 patients developed VF in the late afternoon. There was no evidence of identifiable secondary causes of VF. Four SUDS survivors and two probable SUDS patients had serum potassium levels of <3.5 mEq/L. All patients had their potassium level repleted to normal before undergoing EPS.
ECG and Electrophysiological Abnormalities
Sixteen of the 27 patients had unique ECG abnormalities that manifested as a RBBB with a significant precordial injury as demonstrated by ST-segment elevation in V1 through V3 (Fig 1⇓). These ECG abnormalities are identical to those described by Brugada and Brugada in patients who have idiopathic VF.14 However, the RBBB-like pattern in V1 may not represent a true RBBB but have been caused instead by a marked J-junction elevation in V1, which would produce the RBBB-like pattern. This configuration waxed and waned over time in several patients (Fig 2⇓). In 6 patients, the abnormal ECG pattern normalized during exercise but reappeared after they stopped exercising during the recovery period.
Only 23 of the 27 patients underwent EPS; 2 patients died before undergoing EPS, and 2 refused to participate. Of the 23 patients undergoing EPS, 14 had inducible sustained VF necessitating cardioversion (Fig 3⇓), and 9 patients had no inducible arrhythmias. SAECG was performed in 22 of the 23 patients undergoing EPS (1 patient did not undergo SAECG testing); 12 had a positive SAECG, and 10 had a negative SAECG.
Clinical Differences Between Patients With and Without RBBB and Precordial Injury as Demonstrated by ST-Segment Elevation in V1 Through V3
The Table⇓ shows that the ECG abnormalities divided our patient population into two distinct groups: patients who had RBBB and a precordial injury pattern in V1 through V3 (group 1) and those who did not (group 2). The group 1 patients who had RBBB with a precordial injury pattern in V1 through V3 also had well-defined electrophysiological abnormalities. The group 1 patients had an abnormally prolonged His-Purkinje conduction time: the mean HV interval was 63±11 ms in the group 1 patients compared with 49±6 in the group 2 patients (P=.007). However, the AV nodal conduction time was normal in both groups: the mean AH intervals were 91±10 (group 1) and 85±9 ms (group 2) (not statistically different). Thirteen of the 14 group 1 patients had inducible VF (93%) compared with only 1 of the 9 group 2 patients (11%; relative risk, 8.35; P=.0002). Moreover, 92% of the group 1 patients had a positive SAECG (11 of the 13 patients) compared with 1 of the 9 group 2 patients (11%; relative risk, 7.62; P=.002). All these findings signified a primary electrical instability that represented an electrophysiological substrate for VF.
VF and Clinical Outcome
VF did occur more often in the group 1 patients. Those patients who displayed the SUDS markers—RBBB and precordial injury pattern in V1 through V3—developed VF more frequently than the group 2 patients. Fourteen of the 16 group 1 patients were SUDS survivors (87.5%) who survived an episode of VF compared with only 3 of the 11 (27%) group 2 patients who were SUDS survivors (relative risk, 3.2; P=.003). During the follow-up period (mean, 11.8±7 months; range, 3 to 25 months), 7 patients died suddenly and unexpectedly: 6 were group 1 patients (2 were on propranolol, 1 was on amiodarone, and the other 3 were not on any antiarrhythmics) and 1 was a group 2 patient (on propranolol at the time of death). A total of 10 of the 16 group 1 patients developed either sudden death (n=6) or VF as displayed on the shock E-gram of the ICD (n=3) or the ECG rhythm strip recorded in the emergency department (1 patient who survived a recurring episode) compared with only 2 of the 11 group 2 patients (1 had sudden death and the other developed VF and was resuscitated; relative risk, 3.4; P=.047). All episodes occurred between 9:00 pm and 7:00 am during sleep except for 2 patients: 1 had VF in the early afternoon at ≈2:00 pm and the other died suddenly at ≈10:00 am while awake. The time of death or VF during the follow-up period is remarkably similar to the time of VF episodes observed before the study. One of the 2 group 1 patients who had only been identified as a probable SUDS patient died after having gone to sleep at 10:00 pm. This patient, who was only 17 years old, had not previously experienced VF, but he had manifested the clinical markers for SUDS (Fig 2⇑).
Eight patients (6 group 1 and 2 group 2 patients) were treated with an ICD, 8 (5 group 1 and 3 group 2 patients) with propranolol, and 4 with amiodarone (2 group 1 and 2 group 2 patients), and 7 (3 group 1 patients and 4 group 2 patients) were not treated with antiarrhythmic therapy. Three patients who had received ICDs developed spontaneous VF during sleep; thus far, neither of the 2 group 2 patients with an ICD has experienced a defibrillation shock. An example of VF episodes as detected by the shock-E gram is shown in Fig 4⇓. This patient, who had RBBB and a precordial injury pattern V1 through V3, had survived an episode of cardiac arrest before ICD implantation. Fig 4⇓ shows a VF episode detected by the ICD. The first episode (Fig 4A⇓) shows a nonsustained VF episode at ≈5:00 am while he was asleep and asymptomatic. The second episode (Fig 4B⇓) shows a sustained episode triggering the ICD discharge. This episode occurred 2 months later during sleep after midnight; the patient was found to have agonal respiration by his wife before the ICD discharged. In fact, the patient did not know that the ICD had discharged at that time until he came for a follow-up appointment. Similarly, the VF episodes detected by the ICD occurring during sleep in the other 2 SUDS survivors were associated with labored respiration and groaning. Moreover, their spouses were unable to arouse these patients until shortly after the defibrillation; 1 of the other patients also did not know that the ICD had discharged. These symptoms and narratives vividly mirror those of SUDS victims described in the literature.
The VF episodes detected by the ICD emphasize two important points. The ICD confirmed that VF is the culprit rhythm causing SUDS. And the ICD effectively terminates the VF; it prevents both sudden death and the adverse consequences of cardiac arrest. A life-table analysis showing a comparison of the cumulative proportion of either VF or sudden death between the two groups is shown in Fig 5⇓. The group 1 patients had a much higher mortality rate and incidence of VF recurrence.
Our data demonstrate that SUDS patients face a significantly high incidence of recurrent VF, most of whom—especially SUDS survivors—manifest distinct ECG abnormalities. The group 1 patients had the RBBB pattern and a precordial injury pattern in V1 through V3; these ECG abnormalities correlated with a higher incidence of inducible and spontaneous VF as well as a higher death rate, electrophysiological abnormalities, and a positive SAECG. The group 2 patients did not manifest these ECG or electrophysiological abnormalities and had a significantly lower VF and death rate. All deaths in our study were sudden death; one can speculate that the death rate may have been higher if the 3 patients who had recurrent VF did not receive ICDs. These findings underscore the fact that the abnormal ECG pattern of the RBBB and precordial injury in V1 through V3 is an important arrhythmogenic marker for VF in our patient population.
These ECG abnormalities are identical to those described by Brugada and Brugada in their studies of idiopathic VF14 ; our patients share other similarities. Both groups have no organic heart disease, are composed almost exclusively of men, have a high incidence of inducible polymorphic VT degenerating to VF, and have a high morality rate.
Our patient population is also identical to the SUDS patients described in the CDC report.1 2 In both our study and the CDC report, the patients were quite young and the event rate was significantly high. These findings are in line with the SUDS death rate of 26 to 38 per 100 000 reported in the study of the young Thai men (between 20 and 49 years old). The time of VF occurrence and sudden death in our population is reproducible when comparing the time that the episodes occurred before enrollment in the study with that during the follow-up period. Most VF episodes or sudden death in our patients occurred at night between 8:00 pm and 7:00 am, suggesting that both the pattern and mechanism of death were similar in our patient population and that of the CDC.1 2 The only difference between the CDC study population and our patient population is ethnicity: there were no Thais in the CDC population because the Thai people were not part of the refugee population coming to the United States at the end of the Vietnam War. However, a similar pattern of death occurred among the Thai construction workers who had come to Singapore.9 15
The abnormal ECG pattern of the RBBB and right precordial injury pattern in V1 through V3 can be dynamic, waxing, and waning over time (Fig 2⇑). This observation becomes relevant in the case of the 3 SUDS survivors who had a normal ECG: it could signify either that the patients truly had no ECG abnormalities or that the recording was simply reflecting an ebbing phase of the cycle. Our finding that exercise normalized the ECG abnormalities suggests that sympathetic stimulation plays a major role in correcting the ECG abnormalities. Our data confirm those of Miyazaki et al,16 who recently published a study about the autonomic influence on similar ECG patterns in 5 patients who had the Brugada syndrome. An isoproterenol infusion normalized the ECG pattern, a finding that parallels our observation that exercise normalized the ECG abnormalities. Our observations and those of Miyazaki et al support the hypothesis offered by Yan and Antzelevitch17 that these ECG abnormalities were possibly caused by the outward shift in the ionic currents active at the end of phase 1 of the action potential. Increased Ito or decreased ICa may cause loss of the dome of phase 2 repolarization of the cardiac action potential in the right epicardium. Stimulation of ICa with a β-agonist could restore the lost dome; acetylcholine facilitated the loss of the action potential dome by suppressing ICa. The loss of the action potential dome in the epicardium, and not in the endocardium, would cause the precordial injury pattern, which may lead to phase 2 reentry- induced VF.18 Whether this mechanism is operative in our patients remains speculative and deserves study.
All our patients had no structural heart disease, except for 3 patients who had mild LVH. However, the group 1 patients had an abnormally prolonged His-Purkinje conduction time associated with a positive SAECG. These two abnormalities are usually linked to structural changes in the heart and conduction system. A correlation can be made between our findings and those of Kirschner et al,19 who performed autopsy studies of Laotian and Cambodian refugees in the United States.
The autopsy studies revealed conduction system anomalies, specifically, persistent fetal dispersion of the AV node and His bundle, in 14 of the 18 hearts examined of patients who had died suddenly during sleep. Whether fetal dispersion existed in our patients is not known, but if it did, it would account for the prolonged His-Purkinje conduction time. Persistent fetal dispersion causes sudden cardiac death in the young, as first described by James and Marshall.20
The poignant message that emerged from our study is that SUDS patients face a grave risk of VF recurrence and sudden cardiac death. Our finding is the first prospective study to demonstrate that SUDS patients who have idiopathic VF run an inordinate risk of sudden cardiac death and confirm the conclusion made by the Brugadas that patients with RBBB and ST-segment elevation in V1 through V3 face a grave prognosis. For many years, investigators viewed the phenomenon of sudden cardiac death in Southeast Asian men as a regional subset of the idiopathic VF population (separate and enigmatic), but the fact that patients from Europe who have idiopathic VF and SUDS patients from Asia share essentially the same ECG abnormalities, as well as other clinical characteristics, necessitates a reconsideration: The phenomenon of SUDS is not contained within one geographic region but instead spans across regions, from Southeast Asia (among SUDS patients) to Europe and North America (among patients with idiopathic VF).
Although current knowledge of these syndromes is still in its nascency and the best therapy for these patients has yet to be determined, our preliminary data suggest that the ICD may be logical treatment of choice. Meanwhile, we must educate physicians and cardiologists about these clinical markers and potential SUDS patients to avoid the devastation of young patients who may die prematurely.
Selected Abbreviations and Acronyms
|CDC||=||Centers for Disease Control and Prevention|
|ICD||=||implantable cardiac defibrillator|
|LVH||=||left ventricular hypertrophy|
|MRI||=||magnetic resonance imaging|
|RBBB||=||right bundle-branch block|
|SUDS||=||sudden unexplained death syndrome|
The authors thank Mary Kolb, BA, for editorial assistance, and Chung-Yin Chiu for data and statistical analysis. The CPI-P3s used in this study were kindly provided by Guidant-CPI, Inc.
Reprint requests to Dr Koonlawee Nademanee, University of Southern California, Division of Cardiology, Ambulatory Health Center, 1355 San Pablo St, Room 117, Los Angeles, CA 90033.
- Received January 20, 1997.
- Revision received May 5, 1997.
- Accepted May 19, 1997.
- Copyright © 1997 by American Heart Association
Parish RG, Tucker M, Ing R, Encarnacion C, Eberhardt M. Sudden unexplained death syndrome in southeast Asian refugees: a review of CDC surveillance. MMWR. 1987;36(no. 1 SS):43SS-53SS.
Aponte GE. The enigma of bangungut. Ann Intern Med. 1960;52:1258-1263.
Tungsanga K, Sriboonlue P. Sudden unexplained death syndrome in north-east Thailand. Int J Epidemiol. 1993;22:81-87.
Tatsanavivat P, Chirawatkul A, Klungboonkrong V, Munger R, Chaisiri S, Charerntanyarak L, Saowakontha S. Sudden and unexplained deaths in sleep (Lai Tai) of young men in rural northeastern Thailand. Int J Epidemiol. 1992;21:904-910.
Sugai M. A pathological study on sudden and unexpected death, especially on the cardiac death autopsied by medical examiners in Tokyo. Acta Pathol Jpn. 1959;9(suppl):723-752.
Goh KT, Chao TC, Chew CH. Sudden nocturnal deaths among Thai construction workers in Singapore. Lancet. 1990;335:1154.
Otto CM, Tauxe RV, Cobb LA, Green HL, Gross BW, Werner JA, Burroughs RW, Samson WE, Weaver WD, Trobaugh GB. Ventricular fibrillation causes sudden death in southeast Asian immigrants. Ann Intern Med. 1984;100:45-47.
Gilbert J, Gold RL, Haffajee CI, Alpert JS. Sudden cardiac death in a southeast Asian immigrant: clinical, electrophysiologic, and biopsy characteristics. PACE. 1986;9:912-914.
Yan GX, Antzelevitch C. Cellular basis for the electrocardiographic j wave. Circulation. 1996;93:372-379.
Antzelevitch C, Sicouri S, Lukas A, Di Diego JM, Nesterenko VV, Liu DW, Roubache JF, Zygmunt AC, Zhang Z, Iodice A. Clinical implications of electrical heterogeneity in the heart: the electrophysiology and pharmacology of epicardial, m and endocardial cells. In: Podrid PJ, Kowey PR, eds. Cardiac Arrhythmias: Mechanisms, Diagnosis and Management. Baltimore, Md: Williams & Wilkins; 1994:88-107.
James TH, Marshall TK. Persistent fetal dispersion of the atrioventricular node and His-bundle within the central fibrous body. Circulation. 1976;53:1026-1034.