Sudden unexplained death during sleep occurred commonly in the ...

Journal of Clinical Epidemiology 60 (2007) 567e571

Sudden unexplained death during sleep occurred commonly in the general population in the Philippines: a sub study of the National Nutrition and Health Survey Giselle Gervacio-Domingo*, Felix Eduardo Punzalan, Ma. Lourdes Amarillo, Antonio Dans University of the Philippines-Philippine General Hospital, Section of Cardiology, 1000 Manila, Phillippines Accepted 9 October 2006

Abstract Objective: Sudden unexplained death during sleep (SUDS) is found frequently among Asians. The nationwide incidence of SUDS in the Philippines was measured using a questionnaire, validated in a previous study versus autopsy. Study Design and Setting: The questionnaire was administered as part of the 2003 National Nutrition and Health Survey. A total of 4,747 households were sampled in a stratified randomized manner. Household members were interviewed regarding the occurrence of presumptive SUDS within the last 5 years. Presumptive SUDS was death in a young (!40 years) healthy individual with no reasonable alternative explanation for death. Results: After adjustment for age and sampling weight, the 5-year incidence of sudden death during sleep was 380 (95% CI 210e640) per 100,000, whereas that of SUDS was 110 (95% CI 29e540) per 100,000 in the 20e39 year age group. The computed annualized incidence of sudden death during sleep in the 20e39 year age group was 76 per 100,000, that of SUDS 22 was per 100,000. Computed annualized incidence of SUDS based on the questionnaire accuracy was 43 per 100,000. Conclusion: SUDS occurs commonly among young Filipinos affecting 43 per 100,000 per year of which most are young males. Ó 2007 Elsevier Inc. All rights reserved. Keywords: Sudden unexplained death syndrome; SUDS; Asian SUDS; Bangungut

1. Introduction Sudden unexplained death syndrome (SUDS) is known as bangungut in the Philippines, lai tai in Thailand, and pokkuri disease in Japan. It is a condition, which appears to occur more frequently among Asians [1e3]. The condition gained worldwide attention when it was described in a group of Southeast Asian refugees in the United States [4]. The typical victim is a young, healthy male who dies suddenly during sleep. Estimates of the incidence of SUDS in the general population have been unreliable because the sampling methods used were not probabilistic, and in many cases were based on spontaneous reports. In Northeastern Thailand, mailed questionnaires requiring prospective reporting of sudden death by village health workers were complimented by a retrospective interview by the authors. The incidence of * Corresponding author. University of the Phillippines-Phillippine General Hospital, Section of Cardiology, Taft Avenue, Ermita, 1000 Manila, Phillipines. Tel.: þ632-5232010; fax: þ632-523-2010. E-mail address: [email protected] (G. Gervacio-Domingo). 0895-4356/07/$ e see front matter Ó 2007 Elsevier Inc. All rights reserved. doi: 10.1016/j.jclinepi.2006.10.003

SUDS in the 20e49 year age group was 38 per 100,000 per year in this series [5]. Death certificates and autopsy records reviewed at the Manila Health Department between 1948 and 1982 revealed a peak incidence of 26.3 per 100,000 per year [6]. In a study among Southeast Asian refugees in the United States from 1977 to 1982, the incidence of SUDS between 25 and 44 years was 92, 82, and 59 per 100,000 among Hmong, Laoatian, and Kampuchean refugees, respectively [2]. To date, there are no data on incidence of SUDS on a nationwide scale. There is a need to determine the incidence because this condition affects the most productive segment of society. The study was carried out to 1) measure the incidence of sudden death during sleep in the general population in the Philippines and 2) measure the annualized incidence of SUDS in the 20e39-year-old age group. The meticulous sampling method used in this study, allowed random selection of households nationwide and ensured that the prevalence measured is truly reflective of the general population. This is the strength of this study that

568

G. Gervacio-Domingo et al. / Journal of Clinical Epidemiology 60 (2007) 567e571

eliminates referral bias found in all studies that measure incidences based on voluntary reporting from local health officials or a report on a group of refugees. The gold standard of sudden unexplained death is autopsy. This would be very difficult to perform on all cases in the general population; hence, the use of a surrogate tool, that is a questionnaire tool, which in a previous study had been validated against the gold standard (autopsy) allowed measurement of the incidence of SUDS in the general population.

Ten ancillary questions were included in the questionnaire. These questions were on associated symptoms, family history, and eating and drinking patterns prior to demise. The full questionnaire including its development and validation is detailed elsewhere [8] (Table 2). For the death to qualify as presumptive of SUDS, four conditions had to be fulfilled:

2. Methods

The questionnaires were administered by one of 14 trained research assistants to one adult member of each sampled household.

2.1. National survey This study involved a randomly selected sample from the general population of the Philippines. The 2003 Master Sample of the National Statistics Office for the 2003 Family Income and Expenditure Survey (FIES), which used a stratified multistage sampling design covering all the regions and provinces in the country, was used. The 6th National Nutrition Survey (6th NNS) used a randomly selected sample, which corresponds to half of the FIES sampled households. The National Nutrition Health Survey (NNHES) then used a randomly selected sample, which corresponds to one-fourth of the 6th NNS sampled households. The primary sampling units (PSUs), which were described as barangays (villages) or a group of contiguous small barangays containing at least 500 households, were selected within a set of strata using probability proportional to estimated size sampling. The second stage was the selection of Enumeration Areas (EAs) or contiguous areas in a PSU with at least 150e200 households. The third stage was the selection of households with equal probability in each sampled EA. The detailed description of the sampling design is found in the NNHES report [7]. 2.2. SUDS questionnaire SUDS is defined as death within an hour of onset of symptoms, in a previously healthy individual. On autopsy, victims of this condition show normal autopsy findings. Since autopsy data are not available for victims who are chosen randomly from the general population, a four-point questionnaire validated versus autopsy in a previous study was used to diagnose SUDS [8]. Using this questionnaire we have defined presumptive SUDS using four-points: Question (Q) 1: sudden, unexpected death during sleep, Q2: age 20e39 years, Q3: no cardiac, pulmonary, or neurologic symptoms suggesting an alternative cause of death, and Q4: no previously known or suspected cardiac, pulmonary, or neurologic disease (Table 1).

1. 2. 3. 4.

A ‘‘Yes’’ answer to Q1. Age less than 40 years in answer to Q2. A ‘‘No’’ answer to Q3. A ‘‘No’’ answer to Q4.

2.3. Statistical analysis Crude incidences were calculated using the following equation f =SN where f is the number of victims counted and SN is the sum of the numbers of residents in each of the sampled households during the period queried. Two adjustments of the crude rate were made. The first adjustment was made based on the sampling weights and age composition of the Philippine population based on 2003 projections. The second adjustment corrected for false positives and false negatives of the questionnaire using the Marchevsky equation:

P5

A FPR ; TPR FPR

where A is the observed prevalence based on a marker for the surveyed condition (in this case, the SUDS questionnaire), FPR and TPR are the false and true positive rates respectively, and P is the true prevalence of disease [8,9]. Results were expressed in incidence per 100,000 population. Table 1 Questionnaire items related to four-point definition of SUDS 1. Has any member of your household died unexpectedly during sleep during the last 5 years? 2. How old was the victim at the time of death? 3. Did victim complain of any of the following during last day of life: Severe chest pain Severe difficulty of breathing Sudden one-sided weakness Severe headache 4. Has the victim ever been diagnosed to have: Heart disease Epilepsy or seizure Stroke Substance abuse or poisoning

G. Gervacio-Domingo et al. / Journal of Clinical Epidemiology 60 (2007) 567e571 Table 2 Ancillary questionnaire items 5. How many years ago did victim die? 6. How many members of the household were there at the time of death of the victim? 7. Was the victim male or female? 8. Did victim have siblings or first-degree cousins who died during sleep? 9. Did victim ever complain of palpitations? 10. Did victim complain of abdominal pain before going to bed on the day of demise? 11. Did the victim eat heavily before going to bed during the day of demise? 12. Did the victim drink heavily before going to bed on the night of demise? 13. Did the victim moan in his sleep before dying? 14. Was the victim autopsied?

3. Results 3.1. Demographics Two thousand seven hundred forty-seven households were included in the 6th NNHES Survey. The estimated number of residents in the sampled households for the last 5 years was 4,745. There were 22 victims who died suddenly during sleep, 18 of them were males. Nearly half (10 out of 22) were in the 20e39 year age bracket while the remaining were 40 years of age or older (Table 3). 3.2. Reasons for excluding victims from the SUDS group Of 22 victims of sudden death, 14 did not meet the criteria for SUDS. The reasons for exclusion were chest pain (5/14 victims) one-sided weakness (5/14). Difficulty of breathing (3/14 victims), previously known or suspected heart disease (4/14 victims) stroke (1/14), and substance abuse (1/14). None of the victims had known or suspected seizure disorder. 3.3. Adjusted rates After adjustment for age and sampling weight, the 5year incidence of sudden death during sleep was 380 (95% CI 210e640) per 100,000, whereas that of SUDS is 110 (95% CI 29e540) per 100,000 in the 20e39-year-old age group. The annualized incidence of sudden death during sleep is 76 per 100,000 per year. The annualized incidence of SUDS is 22 per 100,000 per year in the 20e39-year-old age group. Table 3 Age and gender distribution of sudden death victims Age group

Males

Females

Total

20e39 40e59 60 & over Total

8 5 5 18

2 1 1 4

10 6 6 22

569

After adjustment based on false positive and false negative rates of the questionnaire, the annualized incidence of SUDS was 43 per 100,000 (Table 4). 3.4. Comparison of SUDS versus nonSUDS victims Patients with SUDS were younger (29.75 years) than nonSUDS victims (52.4 years). Both groups were predominantly male. The characteristic ‘‘moaning’’ in sleep before demise was not observed in either group. Affected first-degree relatives were reported only in one out of the 14 nonSUDS patients and in none of the SUDS patients. None of the SUDS victims partook of a heavy meal or had an alcoholic binge before going to bed on the day of demise. None reported severe abdominal pain. Among nonSUDS patients, only one out of 14 ate heavily, one of 14 binged on alcohol, and one of 14 had premorbid abdominal pain Table 5.

4. Discussion There is a need to document the magnitude of SUDS in the Philippines, a country where SUDS was first reported, and whose nationals are frequently afflicted with the condition [1,10,11]. The adjusted annualized incidence of 43 per 100,000 in the 20e40 age group is similar to the incidence of 26 per 100,000 in Manila [5]. Our data are derived using a validated questionnaire on a randomly selected sample from the general population, whereas the latter was derived from a retrospective review of death certificates of which filing and data entry are incomplete. Similar figures of 38 per 100,000 were reported in Northeastern Thailand [4], a region where SUDS is likewise said to be endemic. SUDS has been documented in both Filipino and Thai migrant workers [3,12,13]. This Northeastern Thai study was, however, limited to one region in Thailand as compared to our study, which covered all regions in the Philippines. In the Thai study, the presumptive definition of SUDS was similar to our definition [4]. Table 4 Adjusted incidence of sudden death Adjusted statistic

Incidence (per 100,000)

5-year incidence of sudden death during sleep 5-year incidence of SUDS among 20e40 year age group 1-year incidence of sudden death during sleep 1-year incidence of SUDS among 20e40 year age group 1-year incidence of SUDS adjusted based on questionnaire sensitivity and specificity

380 (95% CI 210e670) 110 (95% CI 29e540) 76 22 43

570


Table 5 Comparison of SUDS versus nonSUDS

Number Mean age (years) Male/female Family history of sudden death in sleep Heavy meal before death Alcoholic binge before death (Q12) Abdominal pain before death (Q10) Palpitations (Q9) Moaning before death (Q13) No. of households w/multiple (more than 1) victims per household

Group A (SUDS)

Group B (nonSUDS)

8 29.75 6/8 (0.75) 0/8 0/8 0/8 0/8 0/8 0/8 0/8

14 52.4 12/14 1/14 0/14 1/14 1/14 1/14 0/14 0/14

(0.86) (0.07) (0.07) (0.07) (0.07)

The use of a validated questionnaire made it possible to measure the nationwide incidence of SUDS. Performing autopsy in all suspected cases is not feasible. Furthermore, validating the questionnaire definition of SUDS makes it possible to 1) provide an easy to use tool, which may be applied o different populations and 2) correct the measured incidence based on the sensitivity and specificity of the definition. For this purpose, the Marchevsky equation was used to calculate the true incidence based on the measured sensitivity and specificity of the questionnaire. The adjusted annualized incidence based on the measured sensitivity and specificity of the questionnaire is 43 per 100,000. This figure confirms that SUDS occurs relatively frequently. This figure is the first true populationbased estimate of the incidence of SUDS. The strength of this estimate lies in the meticulousness of the sampling as well as the use of a validated tool that allowed a survey of this magnitude to be conducted. Ancillary findings of this study include a comparison of characteristics of SUDS and nonSUDS patients. SUDS patients were younger (mean 29.8 years) compared to nonSUDS patients (mean 56.2 years). This is expected as coronary artery disease, which accounts for as much as two-thirds of all sudden deaths, affects older individuals [12]. Male predominance was expected and observed in both the SUDS (75% male) and nonSUDS groups (86% male) congruent with reports in the literature [2e5,12]. The familial clustering of cases observed in other studies [3,4] was not observed in ours. Brugada and Nademanee have identified an electrocardiographic marker present among Asian SUDS patient and have identified both sporadic and familial patterns of the syndrome [15]. Interestingly, the moaning in sleep before death reported by others [1e5] and from which the term bangungut originated was not observed among the SUDS patients or in controls. It may be that the witnesses were awoken from sleep late in the course of the arrest during which time agonal respirations may have ceased. For several decades in the Philippines, SUDS was attributed to acute hemorrhagic pancreatitis, a belief that originated from local pathologic reports beginning in the

1960’s and persisting till now [5]. However, autopsy series performed on Southeast Asian refugees in the United States who died of SUDS did not confirm this finding [2,16]. A study comparing autopsies of SUDS victims with controls is ongoing in the Philippines. To support or refute this belief, we queried whether victims had any form of abdominal pain on the day of demise. None of the respondents reported this. Hence, in this survey there was no pattern of premorbid abdominal pain observed and this seems to be inconsistent with the typical presentation of severe pancreatitis where a more protracted course with prominent abdominal pain is expected [17,18]. Other factors, which classically were associated with bangungut, were not documented in our survey including the history of a heavy meal before going to bed on the evening of demise. In the series by Tungsanga, death occurred within 3 hours of a meal in 42% of cases [4]. An alcoholic binge by the victim on the evening of demise, another factor associated with pancreatitis, was denied by housemates of the victims. Ventricular fibrillation is the observed mechanism of death among Asian SUDS patients [14,17]. In our survey, there was no report of palpitations, chest pain, or shortness of breath before demise. Of note, one 23-year-old victim had a previous episode of loss of consciousness 1 year before demise for which he was brought to the Emergency Room (ER). He regained consciousness upon arrival at the ER. He subsequently was discharged and was later found dead in bed. He allegedly did not complain of any symptom before going to bed on the evening of demise. Such episode of aborted sudden death could be explained by spontaneous conversion of ventricular fibrillation [15,19]. Such phenomena have been recorded in patients who received therapy with an implantable cardioverter defibrillator (ICD) in the DEBUT study [20]. In the same study, episodes of aborted SUDS correlated with recordings of ventricular fibrillation and no further deaths were observed in the ICD group in contrast to controls [19]. In this national survey, electrocardiograms (ECGs) were taken of all members of households surveyed. A study is underway on whether the ECG marker identified in other Asian SUDS victims, known as the Brugada sign, characterized by J point and ST elevation in the right precordial leads and is said to be an arrhythmogenic marker for sudden death, is found more commonly among relatives of SUDS victims compared to controls.

5. Conclusion Our study confirms that SUDS occurs commonly among Filipinos at a rate of 110 (95% CI 29e540) per 100,000 per 5 years in the 20e39-year-old age group and an estimated 43 per 100,000 per year in the same age group. SUDS victims are commonly young asymptomatic males.


Acknowledgment Funding for this study was provided by the Philippine Department of Health, the Food and Nutrition Research Institute, and the Philippine Heart Association. The authors wish to acknowledge the help of the NNHES study group headed by Dr. Dante Morales; Dr. Annette Borromeo, Dr. Mariano Lopez, Ms. Fely Velandria, Ms. Charmaine Duante, Ms. Josephine Sanchez.

[9]

[10] [11] [12] [13]

References [1] Guazon MPM. Algunas notas sobre ‘‘Bangungot’’. Rev Fil Med y Farm 1917;8:437e42. [2] Baron RC, Thacker SB, Gorelkin L, Vernon A, Taylor WR, Choi CK. Sudden death among southeast Asian refugees: an unexplained nocturnal phenomenon. JAMA 1983;250:2947e51. [3] Goh KT, Chao TC, Heng BH, Poh SC. Epidemiology of sudden unexpected death syndrome among Thai migrant workers in Singapore. Int J Epidemiol 1993;22:88e95. [4] Sudden, unexpected, nocturnal deaths among Southeast Asian refugees. MMWR Morb Mortal Wkly Rep 1981;30:581e4. 589. [5] Tungsanga K, Sriboonlue P. Sudden unexplained death syndrome in northeastern Thailand. Int J Epidemiol 1993;22:81e95. [6] Munger RG, Booton EA. Bangungut in Manila: sudden and unexplained death in the sleep of adult Filipinos. Int J Epidemiol 1998;27:677e84. [7] Dans AL, Morales DD, Velandria FV, Abola TB, et al. National Nutrition and Health Survey (NNHES): Atherosclerosis-related diseases and risk factors. Philipp J Int Med 2005;43:103e15. [8] Gervacio-Domingo G, De Castro B, Daez ML, Lim C, Punzalan FE. Development of a questionnaire to measure the incidence of sudden

[14]

[15] [16] [17]

[18]

[19]

[20]

571

unexplained death during sleep in the general population. Philipp J Cardiol 2005;33:33e7. Marchevsky N. Errors in prevalence estimates in population studies: a practical method for calculating real prevalence. Zoonosis 1974;16: 98e109. Rogan WJ, Gladen B. Estimating prevalence from the results of a screening test. Am J Epidemiol 1978;107:71e6. Cruz JC. The pathology of ‘‘bangungut’’. J Phil Med Assoc 1951;27: 476e81. Majoska AV. Sudden death in Filipino men: an unexplained syndrome. Hawaii Med J 1948;7:469e73. Pollanen MS, Chiasson Dam Cairns J, Young JG. Sudden unexplained death in Asian immigrants: recognition of a syndrome in Metropolitan Toronto. CMAJ 1996;155:537e40. Park HY, Weinstein SR. Sudden unexpected nocturnal death syndrome in the Mariana Islands. Am J Forensic Med Pathol 1990;11: 205e7. Zipes D, Wellens H. Sudden cardiac death. Circulation 1998;98: 2334e51. Camm AJ, editor. The Brugada syndrome. New York: Futura Publishing Co.; 1999. Kirschner R, Eckner FO, Baron RC. The cardiac pathology of sudden unexplained nocturnal death in Southeast Asian refugees. JAMA 1986;256:2700e5. Wrobleski DM, Barth MM, Oyen LJ. Necrotizing pancreatitis: pathophysiology, diagnosis, and acute care management. AACN Clin Issues 1999;10:464e77. Nademanee K, Veerakul G, Nimmannit S, Chaowakul V, Bhuripanyo K, Likittanasombat K, et al. Arrhythmogenic marker for sudden unexplained death syndrome in Thai men. Circulation 1997;96:2595e600. Nademanee K, Veerakul G, Mower M, Likittanasombat K, Krittayapong R, Bhuripanyo K, et al. Defibrillator versus beta-blockers for unexplained death in Thailand (DEBUT). Circulation 2003;107:2221e6.