Prognostic Implications of Asymptomatic Ventricular Arrhythmias: The Framingham Heart Study Mahesh Bikkina, MD, MPH; Martin G. Larson, ScD; and Daniel Levy, MD
• Objective: To evaluate the prevalence and prognostic significance of asymptomatic complex or frequent ventricular premature beats detected during ambulatory electrocardiographic (ECG) monitoring. • Design: Cohort study with a follow-up period of 4 to 6 years. • Setting: Population-based. • Participants: Surviving patients of the original Framingham Heart Study cohort and offspring of original cohort members (2727 men and 3306 women). • Measurements: One-hour ambulatory electrocardiography. • Results: The age-adjusted prevalence of complex or frequent arrhythmia (more than 30 ventricular premature complexes per hour or multiform premature complexes, ventricular couplets, ventricular tachycardia, or R-on-T ventricular premature complexes) was 12% (95% CI, 11 % to 13%) in the 2425 men without clinically evident coronary heart disease and 33% (CI, 24% to 42%) in the 302 men with coronary heart disease. The corresponding values in women (3064 without disease and 242 with disease) were 12% (CI, 1 1 % to 13%) and 26% (CI, 9% to 43%). After adjusting for age and traditional risk factors for coronary heart disease in a Cox proportional hazards model, men without coronary heart disease who had complex or frequent ventricular arrhythmias were at increased risk for both all-cause mortality (relative risk, 2.30; CI, 1.65 to 3.20) and the occurrence of myocardial infarction or death from coronary heart disease (relative risk, 2.12; CI, 1.33 to 3.38). In men with coronary heart disease and in women with and without coronary heart disease, complex or frequent arrhythmias were not associated with an increased risk for either outcome. • Conclusions: In men who do not have clinically apparent coronary heart disease, the incidental detection of ventricular arrhythmias is associated with a twofold increase in the risk for all-cause mortality and myocardial infarction or death due to coronary heart disease. The preventive and therapeutic implications of these findings await further investigation.
Annals of Internal Medicine. 1992;117:990-996. From the Framingham Heart Study, Framingham, Massachusetts; the National Heart, Lung, and Blood Institute, Bethesda, Maryland; and Boston University School of Medicine and Beth Israel Hospital, Boston, Massachusetts. For current author addresses, see end of text. 990
Although the presence of complex or frequent ventricular arrhythmia has been shown to be associated with a poor prognosis in patients with coronary heart disease (1-9), uncertainty persists about the importance of ventricular ectopy in persons without clinically apparent coronary heart disease. Some earlier studies suggested that ventricular premature beats, even in apparently healthy persons, were associated with an increased risk for sudden death (10, 11) and an increased incidence of ischemic heart disease (12, 13), but more recent studies have suggested that the presence of ventricular premature beats in apparently healthy persons is not associated with an increased risk for death (14-16). The Framingham Heart Study, because of its large free-living population, was in a unique position to examine this issue. We studied the prevalence and prognostic significance of complex or frequent ventricular arrhythmias in Framingham Heart Study participants who routinely underwent ambulatory electrocardiographic (ECG) monitoring. Methods Study Sample In 1948, 5209 residents of Framingham, Massachusetts, 28 to 62 years of age, were selected to undergo biennial examinations in a prospective study. In 1971, 5135 offspring of the original cohort (and spouses of these offspring) were entered into a prospective study. Selection criteria and study design have been reported previously (17-20). From 1979 to 1983, echocardiograms, 12-lead resting ECGs, and 1-hour ambulatory ECGs were obtained in surviving patients of the original cohort undergoing their sixteenth biennial examination and in offspring participants undergoing their second examination. Each participant gave informed consent before the study. Ambulatory ECG data were available for 97% of the 6218 Framingham Heart Study participants who attended the index examination (Table 1). Events Related to Coronary Heart Disease The follow-up period for our study was approximately 6 years for persons in the original cohort and 4 years for offspring participants. At each follow-up examination (2, 4, and 6 years later for the original cohort; 4 years later for offspring), interim events related to coronary heart disease were assessed by medical history, physical examination, and 12-lead ECG. In addition, medical records were obtained for participants who did not attend an examination and were evaluated for evidence of interim coronary heart disease. All suspected coronary heart disease events were reviewed by a committee of three physicians who evaluated pertinent medical records, hospitalization records, and pathology reports. Two primary outcomes were considered in our study: all-cause mortality and myocardial infarction or death from coronary heart disease. Criteria for various coronary heart disease events have been reported previously (21). The coronary heart disease events included myocardial infarction, coronary insufficiency (unstable angina), angina pectoris, and death from coronary heart disease (both sudden and nonsudden). Unwitnessed deaths occurring during
15 December 1992 • Annals of Internal Medicine • Volume 117 • Number 12
Downloaded From: http://annals.org/ by a University of Otago User on 12/08/2016
Table 1. Evaluable Participants Variable
Participants, n
Study population (cohort and offspring) Died before the index examination* Did not attend index examination Attended index examination No ambulatory electrocardiographic data Evaluable cases
10 344 2109 2017 6218
185 6033
* Index examination was examination 16 for members of the original cohort and examination 2 for offspring participants.
sleep did not meet criteria for sudden death caused by coronary heart disease (21). The diagnosis of congestive heart failure was based on fulfillment of two major criteria or one major criterion plus two minor criteria as described previously (21). Ambulatory Electrocardiographic Methods Two-lead, frequency-modulated recorders (simultaneous VI and V5, Clinical Data Inc.) were used for ambulatory recording. Analysis of the continuous ambulatory ECG records was aided by the use of a high-speed digital computer system (Clinical Data Inc.). Premature depolarizations and arrhythmias were identified and interpreted by a technician. All representative tracings were reviewed by a cardiologist. Classification of ventricular arrhythmias was done using a method similar to that of Lown and Wolf (22) and was based on the presence of the following arrhythmia types: ventricular premature complexes, multiform ventricular premature complexes, ventricular couplets, ventricular tachycardia, and R-on-T ventricular premature complexes. The summary arrhythmia-complex or frequent ventricular premature complexes (more than 30 ventricular premature complexes per hour or multiform premature complexes, ventricular couplets, ventricular tachycardia, or R-on-T ventricular premature complexes)-was used as a single dichotomous variable in all outcome analyses. Ventricular tachycardia was defined as three or more consecutive ventricular premature complexes. Echocardiography The echocardiographic methods used in the Framingham Heart Study have been described previously (23). Participants were studied using a standard M-mode echocardiographic technique (24). Measurements were made according to the methods outlined by Devereux and Reichek (25). Left ventricular internal dimension and the ventricular septum and left ventricular posterior wall thickness were measured at end diastole. Left
ventricular mass (in grams) was calculated from the following formula (24): 1.04 [(LVID + VST + PWT)3 - (LVID)3] 13.6, where LVID = left ventricular internal diameter, VST » ventricular septal thickness, and PWT = posterior wall thickness. Left ventricular mass was normalized according to height (in meters) as described previously (23). Left ventricular hypertrophy was defined as a left ventricular mass/height ratio two standard deviations or more above the mean for a healthy reference population. The cutoff values for left ventricular hypertrophy were 143 g/m in men and 102 g/m in women (23). Adequate echocardiographic data were available in about 80% of the sample. No echocardiographic criteria for eligibility were applied in this study.
Statistical Methods Separate analyses were carried out for men and women, as well as for participants with and without clinically evident coronary heart disease. Clinical characteristics were compared between disease-status groups after adjustment for age (age classification was done at the index examination). Percents were adjusted directly to the appropriate sex-specific age distribution, and the Cochran-Mantel-Haenszel statistic was used in between-group comparisons. Age adjustments for continuous variables were done using 'least squares' means (26), and between-group comparisons for these variables were done using t-tests (27). Prevalence rates for participants classified by arrhythmia grade and stratified by sex and coronary heart disease status were directly adjusted for age. The 95% CIs are provided where appropriate. The two outcome events studied were all-cause mortality and myocardial infarction or death from coronary heart disease. For each outcome, age- and risk-factor-adjusted relative risks associated with the presence of arrhythmia were analyzed with age-stratified proportional hazards regression models (28, 29). A result was deemed significant if P < 0.05. The covariates in these analyses were chosen because of their established potential roles as risk factors for coronary heart disease and death, or because of their possible confounding role with regard to ventricular arrhythmias and death. They included age, sex, high blood pressure (systolic pressure > 160 mm Hg or diastolic pressure > 95 mm Hg), number of cigarettes smoked per day, ratio of total cholesterol to high-density lipoprotein cholesterol, diabetes (yes = 1; no = 0), body mass index (weight in kilograms divided by the square of height in meters), hypertension therapy (those currently taking prescribed medication for the treatment of hypertension, excluding beta-blockers, were assigned a value of 1, and all others were assigned a value of 0), beta-blocker therapy (yes = 1; no = 0), the use of antiarrhythmic medications excluding betablockers (yes = 1; no = 0), and congestive heart failure (yes =
Table 2. Mean Age-adjusted Values for Selected Characteristics* Variable
Age,v Total cholesterol/HDL cholesterol Cigarettes/day Body mass index, kg/m2 Systolic blood pressure, mm Hg Diastolic blood pressure, mm Hg Hypertension, % Hypertension therapy, % Beta-blockers, % Diabetes, % Congestive heart failure, % Antiarrhythmic therapy, % LV mass/height, g/m LV fractional shortening, %
NoCHD (n = 2425)
Men CHD (n = 302)
P Valuet
NoCHD (n = 3064)
Women CHD (n = 242)
P Valuet
51.3 5.1 7.0 26.6 134 80.6 24 16 4 5 0.2 0.3 118 36
66.0 5.8 7.2 27.2 133 79.2 35 22 29 7 4 3 139 34
< 0.001 < 0.001 > 0.2 0.01 >0.2 0.02 < 0.001 >0.2 < 0.0001 0.01 < 0.001 < 0.001 < 0.001 < 0.001
53.8 4.4 5.6 25.5 131 76.8 26 21 4 4 0.8 0.2 89.4 39
70.4 5.1 4.8 26.4 134 76.1 55 48 18 19 5 1 103.4 38
< 0.001 < 0.001 > 0.2 0.007 0.004 >0.2 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 0.03
* CHD = coronary heart disease; HDL = high-density lipoprotein; LV = left ventricular. Risk factor data, excepting echocardiographic data | available), were in about 97% of participants. t The P value reflects age-adjusted comparisons between participants with and without coronary heart disease. 15 December 1992 • Annals of Internal Medicine • Volume 117 • Number 12
Downloaded From: http://annals.org/ by a University of Otago User on 12/08/2016
991
Table 3. Age-adjusted Prevalence of Ventricular Arrhythmias on Monitoring LMI 1-hour 1 - I 1 U U I Ambulatory / IL i u u u i a i u i j Electrocardiographic LUMllUI l l l g A I C V U u v a i ui Variable
> 1 ventricular premature complex/hour, % Frequent ventricular premature complexes, %t Complex arrhythmia, % Multiform Ventricular couplets Ventricular tachycardia^ R-on-T complex Complex or frequent arrhythmia, %
NoCHD (n = 2425)
Men CHD (n = 302)
33 7 10 10 3 0.7 0.4 12
58 17 32 31 12 2 0.9 33
P Value < < < <
0.2 < 0.001
NoCHD (n = 3064)
Women CHD in m 242)
P Value
32 7 9 8 3 0.4 0.5 12
49 19 14 13 5 0.3 0.6 26
0.009 0.03 0.03 < 0.001 < 0.001 >0.2 > 0.2 < 0.001
* CHD = coronary heart disease. t Frequent ventricular premature complex defined as more than 30/h. t Ventricular tachycardia defined as 3 or more consecutive ventricular premature complexes.
1; no = 0). Beta-blocker use was considered as a separate dichotomous variable in multivariate analyses because this class of medications possesses antihypertensive, anti-ischemic, and antiarrhythmic properties. Echocardiographic variables considered in subsequent analyses included left ventricular mass and left ventricular fractional shortening (the difference between diastolic and systolic left ventricular diameters divided by diastolic diameter x 100). Results Study Sample Of the 6033 eligible participants, 2425 men (mean age, 51.3 years) and 3064 women (mean age, 53.8 years) showed no clinical evidence of coronary heart disease; 302 men (mean age, 66.0 years) and 242 women (mean age, 70.4 years) had clinically apparent coronary heart disease at the index examination. Age-adjusted values for various characteristics are summarized according to coronary heart disease status in Table 2. For both sexes, participants who did not have coronary heart disease were younger than those with coronary heart disease, had a lower ratio of total cholesterolrHDL cholesterol, lower prevalences of diabetes, hypertension, congestive heart failure, and lower left ventricular
mass. In the coronary heart disease group, 195 of 302 men (65%) and 77 of 242 women (32%) had myocardial infarctions before the baseline examination. Prevalence of Ventricular Premature Beats Age-adjusted prevalence rates for the various arrhythmia grades according to coronary heart disease status are presented in Table 3. In men, the age-adjusted prevalence of 1 or more ventricular premature beats per hour was 33% in those without coronary heart disease and 58% in those with coronary heart disease (P < 0.001). Corresponding rates in women were 32% and 49% (P = 0.009). The age-adjusted prevalence of complex or frequent ventricular premature beats was 12% (CI, 11% to 13%) in men without coronary heart disease and 33% (CI, 24% to 42%) in those with coronary heart disease group (P < 0.001). Corresponding values in women were 12% (CI, 11% to 13%) and 26% (CI, 9% to 43%) (P < 0.001). Age-adjusted prevalence of frequent ventricular arrhythmias (> 30 ventricular premature complexes per hour) was 7% in men without coronary heart disease and 17% in men with coronary heart disease (P < 0.001). Corresponding values in women were 7% and 19% (P = 0.03). In men, the age-adjusted prevalence of complex ventricular arrhythmias (multiform premature ventricular complexes, couplets, ventricular tachycardia, or R-on-T ventricular premature complexes) was 10% in those without coronary heart disease and 32% in those with coronary heart disease (P < 0.001). Corresponding values in women were 9% and 14% (P < 0.001). Age-specific prevalences of complex or frequent arrhythmia are shown in Figure 1. The prevalence of arrhythmia increases with advancing age in men and women, regardless of coronary heart disease status. In both men and women without coronary heart disease, the prevalence roughly doubled in each successive age group. Association of Arrhythmia and Outcome
Figure 1. Age-specific prevalence of complex or frequent ventricular arrhythmias in men and women by coronary heart disease status. CHD = coronary heart disease.
992
Age- and covariate-adjusted analyses of complex or frequent arrhythmia in relation to outcome are shown in Table 4. In men without clinically apparent coronary heart disease, the age- and covariate-adjusted relative
15 December 1992 • Annals of Internal Medicine • Volume 117 • Number 12
Downloaded From: http://annals.org/ by a University of Otago User on 12/08/2016
cause mortality (relative risk, 2.26; CI, 1.39 to 3.66; P = 0.001) and for myocardial infarction or death from coronary heart disease (2.59; CI, 1.40 to 4.79; P = 0.002).
risk associated with complex or frequent arrhythmia was increased both for all-cause mortality (relative risk, 2.30; CI, 1.65 to 3.20; P < 0.0001) and for myocardial infarction or death from coronary heart disease (relative risk, 2.12; CI, 1.33 to 3.38; P = 0.002). In men with coronary heart disease, complex or frequent arrhythmia was not significantly associated with all-cause mortality (relative risk, 1.27; CI, 0.80 to 2.01; P > 0.2) or myocardial infarction or death from coronary heart disease (relative risk, 0.95; 95% CI, 0.56 to 1.60; P > 0.2). In women, complex or frequent arrhythmia was not associated with either outcome, regardless of coronary heart disease status (Table 4). The survival curves of the participants without clinically apparent coronary heart disease, presented in Figure 2 according to arrhythmia status, were predicted by the proportional hazards model. Estimates were made at observed failure times, rounded to the nearest 1/100 year; therefore, the curves do not appear as smooth functions. Men without coronary heart disease who had complex or frequent arrhythmia showed reduced survival compared with those who did not have complex or frequent arrhythmia (Figure 2, panels A and B). No significant difference in survival was noted in women without coronary heart disease (Figure 2, panels C and D). Among both men and women, those with and those without coronary heart disease showed no significant difference in outcome according to arrhythmia status. In a separate multivariate model, the association of complex or frequent arrhythmia with outcome in men who did not have clinically evident coronary heart disease was adjusted for echocardiographic left ventricular mass and left ventricular fractional shortening, as well as for age and other covariates. The presence of arrhythmia was associated with an increased risk for all-
Discussion Our study documents a significant association between asymptomatic ventricular arrhythmias and the risk for all-cause mortality and myocardial infarction or death from coronary heart disease in men without clinically evident coronary heart disease. Moreover, when considered alongside traditional risk factors for coronary heart disease, including age, blood lipid levels, cigarette smoking, obesity, and hypertension, ventricular arrhythmias remained predictive of both outcomes. In women without coronary heart disease and in persons of both sexes with coronary heart disease, no significant associations between arrhythmia and outcome were noted. Complex or frequent ventricular arrhythmias occur commonly in the general population. In both men and women, the prevalence of ventricular arrhythmia was higher in those with coronary heart disease than in those without it. Among participants with coronary heart disease, the prevalence of ventricular arrhythmia was higher in men than in women. Previous studies of the prognostic importance of ventricular premature beats on routine ECGs in selected, apparently healthy persons have shown conflicting results (10-16): Earlier studies found an increased risk for mortality in patients with ventricular premature beats (10-13), but later studies did not (14-16). Our findings of an increased risk for all-cause mortality and for myocardial infarction or death from coronary heart disease
Table 4. Relation of Complex or Frequent Ventricular Arrhythmia to All-Cause Mortality and to Myocardial Infarction or Death from Coronary Heart Disease Variable
All deaths Age-adjusted relative risl: (95% CI) Deaths/participants at ris k, n/n P value Age- and covariate-adjusited relative risk (95% CI)t Deaths/subjects at risk, n In P value Myocardial infarction or CI ID death Age-adjusted relative risli (95% CI) Myocardial infarctions arid CHD deaths/subjects at risk, n/n P value Age- and covariate-adjusted relative risk (95% CI)t Myocardial infarctions and CHD deaths/subjects at risk, n/n P value
Woimen
M en No CHD
CHD
No CHD
CHD
2.36 (1.72 to 3..24) 186/2423 < 0.0001
1.64 (1.10 to 2.45) 100/302 0.01
1.31 (0.94 to 1 .83) 188/3062 0.11
1.61 (1.00 to 2.60) 71/242 0.05
2.30 (1.65 to 3..20)
1.27 (0.80 to 2.01)
1.23 (0.85 to 1 .77)
1.20 (0.69 to 2.09)
173/2368 < 0.0001
90/281 > 0.2
164/2889 > 0.2
60/214 > 0.2
2.17 (1.40 to 3..37)
1.25 (0.78 to 1.19)
1.13 (0.57 to 2 .21)
0.87 (0.44 to 1.71)
111/2423 0.0005
73/302 > 0.2
52/3062 >0.2
41/242 > 0.2
2.12 (1.33 to 3..38)
0.95 (0.56 to 1.60)
1.01 (0.49 to 2 .07)
0.70 (0.33 to 1.48)
106/2368 0.002
68/281 > 0.2
48/2889 > 0.2
37/214 > 0.2
* CHD = coronary heart disease. t Covariates included the ratio of total cholesterol to high-density lipoprotein cholesterol, smoking, body mass index, hypertension or antihypertensive therapy, systolic blood pressure, diabetes, congestive heart failure, beta-blocker use, and antiarrhythmic therapy (coronary heart disease group only).
15 December 1992 • Annals of Internal Medicine • Volume 117 • Number 12
Downloaded From: http://annals.org/ by a University of Otago User on 12/08/2016
993
Figure 2. Survival curves. Survival function is shown according to the presence or absence of arrhythmia. Each panel shows the number of study participants still under observation at 2, 4, and 6 months. Panel A. Survival curves for myocardial infarction or death from coronary heart disease in men without coronary heart disease. The difference between arrhythmia groups for myocardial infarction (Ml) or death from coronary heart disease (CHD) was significant (P = 0.002). Panel B. Survival curves for all-cause mortality in men without coronary heart disease. The difference between arrhythmia groups for all-cause mortality was significant (P < 0.0001). Panel C. Survival curves for myocardial infarction or death from coronary heart disease in women without coronary heart disease. The curves were superimposed on each other. The difference between arrhythmia groups for myocardial infarction or coronary heart disease was not significant (P > 0.2). Panel D. Survival curves for all-cause mortality in women without coronary heart disease. The difference between arrhythmia groups for all-cause mortality was not significant (P > 0.2).
in apparently healthy men with complex or frequent ventricular arrhythmia are consistent with results from the Manitoba study (12), in which 3983 Canadian air pilots without overt coronary heart disease were prospectively followed for 10 years. However, that study failed to explore the role of potentially important covariables, such as hypertension and cigarette smoking, factors that may be associated with both arrhythmia and outcome. Kennedy and coworkers (30-32) also investigated the prognostic importance of complex or frequent ventricular arrhythmias in apparently healthy persons. They reported no increase in all-cause mortality or myocardial infarction or death from coronary heart disease in persons without coronary heart disease who had complex or frequent arrhythmia. The differences between our findings and those of Kennedy and colleagues (32) may be due to differences in sample size (659 994
participants with complex or frequent arrhythmia in our study compared with 73 such persons in the study by Kennedy and colleagues) or to the influence of betablocker use (27% of Kennedy and colleagues' study sample were receiving beta-blockers compared with 4% of men without coronary heart disease in our study). Other differences in study samples may have played a role, with selection bias more likely in the earlier clinical series. In persons with coronary heart disease, no statistically significant association was observed between arrhythmia and either outcome (all-cause mortality or myocardial infarction or death from coronary heart disease) (see Table 4). This finding is consistent with the results of two previous studies (33, 34). Nevertheless, several clinical studies of survivors of myocardial infarction have shown an increased risk for cardiac death
15 December 1992 • Annals of Internal Medicine • Volume 117 • Number 12
Downloaded From: http://annals.org/ by a University of Otago User on 12/08/2016
in those with ventricular arrhythmias, particularly when complex or frequent forms are present (1-7). The difference between our findings and those of these other studies may be attributable to the different selection processes, particularly in relation to the interval between onset of coronary heart disease and ECG monitoring. In our study, participants with coronary heart disease survived long enough to attend a routine screening examination. Unstable patients with recent myocardial infarction and those with coronary heart disease of new onset in whom arrhythmias may contribute to early morbidity or mortality may have been under-represented in our study. Complex or frequent arrhythmia may be a marker for occult coronary heart disease or silent ischemia. However, if occult coronary heart disease or silent ischemia was a contributor to both arrhythmia expression and adverse outcome, we would expect to find similar, if not stronger, associations between arrhythmia and outcome in patients with overt coronary heart disease. Because we did not observe a significant association between arrhythmia and outcome in patients with coronary heart disease, it is possible that a nonischemic mechanism underlies our findings in men without clinically apparent coronary heart disease. Previous studies suggested that hypertensive patients treated with diuretics may be at an increased risk for fatal cardiac events, perhaps because of an increased frequency of ventricular premature beats (35-38). The antiarrhythmic properties of beta-blockers might, in part, be responsible for improved survival among patients with myocardial infarction who received these drugs in several secondary prevention trials (39-41). The significance of antiarrhythmic therapy in patients with ventricular arrhythmias remains controversial (42, 43). Because of the potential effect of diuretics, beta-blockers, and other antiarrhythmic drugs on arrhythmia expression and outcome, we included these variables in our multivariate model. Congestive heart failure was used as a covariate because Swerdlow and colleagues (44) reported that the presence and severity of heart failure were the most powerful independent predictors of both cardiac death and sudden death in patients with ventricular tachyarrhythmias. We included left ventricular mass in the multivariate model because previous studies (45-47) showed an increased risk for ventricular arrhythmias in patients with left ventricular hypertrophy and increased mortality and an increased incidence of coronary heart disease in patients with increased left ventricular mass (48-50). Additionally, left ventricular fractional shortening was also entered in multivariate model because a previous study showed that patients with left ventricular systolic dysfunction had increased mortality (51). After adjustments were made for left ventricular mass and fractional shortening, the increased risk for all-cause mortality and myocardial infarction or death from coronary heart disease in men without coronary heart disease who had complex or frequent arrhythmia persisted with only minimal change. This finding would suggest that complex or frequent ventricular arrhythmia is not merely a marker for occult left ventricular dysfunction or hypertrophy.
Because our study was conducted in a large population-based sample, we avoided referral biases inherent in earlier clinical studies. Ambulatory ECG monitoring was successfully done in about 97% of the persons who attended the index examination. In addition, we carried out extensive analyses to explore the roles of numerous potential confounding variables that may influence outcome and the association between arrhythmia and outcome. Our sample consisted predominantly of white persons, and therefore our data should be interpreted with caution when considering other races. In our study, 1-hour ambulatory ECG monitoring was done in 6033 persons during a routine examination. Because of substantial variation in the occurrence of ventricular arrhythmia, the use of 1-hour ambulatory ECG monitoring may have resulted in an underestimation of the number of participants predisposed to arrhythmia. However, the large sample size would minimize the effect of variation. Even though the use of 24-hour monitoring would have resulted in a higher prevalence of arrhythmia, the 1-hour monitoring was sufficient to predict both outcomes in men without coronary heart disease who had complex or frequent arrhythmia. Because stress testing and angiographic studies were not done in our study, the possibility of occult coronary heart disease cannot be excluded in participants without clinically apparent coronary heart disease. Our results in the group with coronary heart disease should be interpreted with caution because this group was small and included patients with angina pectoris, coronary insufficiency, or myocardial infarction who survived their acute coronary event and were able to attend a routine examination, often several years after disease onset. Thus, we studied a relatively healthy group with coronary heart disease. Our results should not be extrapolated to patients with unstable or recent-onset coronary heart disease. Last, the arrhythmias detected in our study participants were asymptomatic and low grade. Our findings may not be applicable to patients with symptomatic or higher-grade ventricular arrhythmias. Currently, no evidence exists that antiarrhythmic therapy in patients with asymptomatic ventricular arrhythmias improves prognosis, regardless of whether or not coronary heart disease is present. The mechanism of the observed increased risk associated with ventricular arrhythmias in men without clinically apparent coronary heart disease is unknown. The extent to which 1-hour ambulatory ECG monitoring has utility and any therapeutic implications in the clinical setting are also unknown. Further prospective studies are needed to identify the mechanisms of increased risk in apparently healthy subjects with ventricular arrhythmias. Grant Support: In part by an educational grant from Parke-Davis (Dr. Bikkina). Requests for Reprints: Daniel Levy, MD, Framingham Heart Study, 5 Thurber Street, Framingham, MA 01701. Current Author Addresses: Dr. Bikkina: University of South Alabama, Department of Cardiology, 2451 Fillingim Street, Tenth Floor, Mobile, AL 36617. Drs. Larson and Levy: Framingham Heart Study, 5 Thurber Street, Framingham, MA 01701.
15 December 1992 • Annals of Internal Medicine • Volume 117 • Number 12
Downloaded From: http://annals.org/ by a University of Otago User on 12/08/2016
995
References 1. Kotler MN, Tabatznik B, Mower MM, Tominaga S. Prognostic significance of ventricular ectopic beats with respect to sudden death in the late post infarction period. Circulation. 1973;47:959-66. 2. Prognostic importance of premature beats following myocardial infarction. Experience in the coronary drug project. JAMA. 1973;223: 1116-24. 3. Ruberman W, Wernblatt M, Goldberg JD, Frank CW, Shapiro S. Ventricular premature beats and mortality after myocardial infarction. N Engl J Med. 1977;297:750-7. 4. Moss AJ, DeCamilla J, Engstrom F, Hoffman W, Odoroff C, Davis H. The posthospital phase of myocardial infarction: identification of patients with increased mortality risk. Circulation 1974;49:460-6. 5. Bigger JT Jr, Dresdale RJ, Heissenbuttel RH, Weld FM, Wit AL. Ventricular arrhythmias in ischemic heart disease: mechanism, prevalence, significance, and management. Prog Cardiovasc Dis. 1977;19:255-300. 6. Lie KI, Wellens HJJ, Downar E, Durrer D. Observations on patients with primary ventricular fibrillation complicating acute myocardial infarction. Circulation. 1975;52:755-9. 7. Conolly SJ, Carins JA. Comparison of one-, six- and 24-hour ambulatory electrocardiographic monitoring for ventricular arrhythmia as a predictor of mortality in survivors of acute myocardial infarction. CAMIAT Pilot Study Group. Canadian Amiodarone Myocardial Arrhythmia Trial. Am J Cardiol. 1992;69:308-13. 8. Moss AJ. Clinical significance of ventricular arrhythmias in patients with and without coronary artery disease. Prog Cardiovasc Dis. 1980;23:33-52. 9. Horan MJ, Kennedy HL. Ventricular ectopy. History, epidemiology, and clinical implications. JAMA. 1984;251:380-6. 10. Chiang RN, Perlman LV, Ostrander LD Jr, Epstein FM. Relationship of premature systoles to coronary heart disease and sudden death in the Tecumseh epidemiologic study. Ann Intern Med. 1969; 70:1159-69. 11. Hinkle LE Jr, Carver ST, Stevens M. The frequency of asymptomatic disturbances of cardiac rhythm and conduction in middle-aged men. Am J Cardiol. 1969;24:629-50. 12. Pell S, D'Alonzo CA. A three year study of myocardial infraction in a large employed population. JAMA. 1961;175:463-70. 13. Rabkin SW, Mathewson FA, Tate RB. Relationship of ventricular ectopy in men without apparent heart disease to occurrence of ischemic heart disease and sudden death. Am Heart J. 1981 ;101: 135-42. 14. Rodstein M, Wolloch L, Gubner RS. Mortality study of the significance of extrasystoles in an insured population. Circulation. 1971; 44:617-25. 15. Fisher FD, Tyroler HA. Relationship between ventricular premature contractions on routine electrocardiography and subsequent sudden death from coronary heart disease. Circulation. 1973;47:712-9. 16. Elkon KB, Swerdlow TA, Myburgh DP. Persistent ventricular ectopic beats: a long-term study. S Afr Med J. 1977;52:564-6. 17. Dawber TR, Meadors GF, Moore FE. Epidemiologic approaches to heart disease: the Framingham study. Am J Public Health. 1951;41: 279-86. 18. Dawber TR, Kannel WB, Lyell LP. An approach to longitudinal studies in a community: the Framingham Study. Ann NY Acad Sci. 1963;107:539-56. 19. Gordon T, Moore TE, Shurtleff D, Dawber TR. Some methodologic problems in the long-term study of cardiovascular disease: observations on the Framingham study. J Chronic Dis. 1959;10:186-206. 20. Kannel WB, Feinleib M, McNamara PM, Garrison RJ, Castelli WP. An investigation of coronary heart disease in families. The Framingham Offspring Study. Am J Epidemiol. 1979;110:281-90. 21. Shurtleff D. Some characteristics related to the incidence of cardiovascular disease and death: Framingham Heart Study 18 follow-up. In: Kannel WB, Gordon T, eds. The Framingham Study. Section 30. Washington, DC: U.S. Government Printing Office; 1974 (DHEW publication No. [NIH] 74-599): 17-25. 22. Lown B, Wolf M. Approaches to sudden death from coronary heart disease. Circulation 1971;44:130-42. 23. Levy D, Savage DD, Garrison RJ, Anderson KM, Kannel WB, Castelli WP. Echocardiographic criteria for left ventricular hypertrophy: The Framingham Heart Study. Am J Cardiol. 1987;59:956-60. 24. Feigenbaum H. Echocardiography. Third edition. Philadelphia: Lea & Febiger, 1981:51-118. 25. Devereux RB, Reichek N. Echocardiographic determination of left ventricular mass in man. Anatomic validation of the method. Circulation. 1977;55:613-8. 26. SAS Institute Inc. SAS/STAT User's Guide, Version 6, Fourth Edition, Volume 1. Cary, North Carolina: SAS Institute Inc.; 1989. 27. SAS Institute Inc.. SAS/STAT User's Guide, Version 6, Fourth Edition, Volume 2. Cary, North Carolina. SAS Institute Inc.; 1989.
996
28. Kalbfleisch JD, Prentice RL. The Statistical Analysis of Failure Time Data. New York: John Wiley & Sons.; 1980. 29. SAS Institute Inc. SAS Technical Report P-217, SAS/STAT Software: The PHREG Procedure, Version 6. Cary, North Carolina: SAS Institute Inc.; 1991. 30. Kennedy HL, Underbill SJ. Frequent and complex ventricular ectopy in apparently healthy subjects: a clinical study of 25 cases Am J Cardiol. 1976;38:141-8. 31. Kennedy HL, Pescarmona JE, Bouchard RJ, Goldberg RJ. Coronary artery status of apparently healthy subjects with frequent or complex ventricular ectopy. Ann Intern Med. 1980;92:179-85. 32. Kennedy HL, Whitlock JA, Sprague MK, Kennedy LJ, Bucklingham TA, Goldberg RJ. Long-term follow-up of asymptomatic healthy subjects with frequent and complex ventricular ectopy. N Engl J Med. 1985;312:193-7. 33. de Soyza N, Bennett FA, Murphy ML, Bissett JK, Kane JJ. The relationship of paroxysmal ventricular tachycardia complicating the acute phase of ventricular arrhythmia during the late hospital phase of myocardial infarction to long-term survival. Am J Med. 1978; 64:377-81. 34. Cats VM, Lie KI, Van Capelle FJ, Durrer O. Limitations of 24 hour ambulatory electrocardiographic recording in predicting coronary events after acute myocardial infarction. Am J Cardiol. 1979;44: 1257-62. 35. Multiple Risk Factor Intervention Trial: risk factor changes and mortality results. Multiple Risk Factor Intervention Trial Research Group. JAMA. 1982;248:1465-77. 36. MRC trial of treatment of mild hypertension: principal results. Br Med J (Clin Res Ed). Medical Research Council Working Party. 1985;291:97-104. 37. Levy D, Anderson KM, Christiansen JC, Campanile G, Stokes J 3d. Antihypertensive drug therapy and arrhythmia risk. Am J Cardiol. 1988;62:147-9. 38. Ventricular extrasystoles during thiazide treatment: sub-study of MRC mild hypertension trial. The Medical Research Council Working Party on Mild to Moderate Hypertension. Br Med J. 1983;287: 1249-53. 39. Koppes GM, Beckmann CH, Jones FG. Propranolol therapy for ventricular arrhythmias two months after acute myocardial infarction. Am J Cardiol. 1980;46:322-8. 40. Uchstein E, Morganroth J, Harrist Hubble E, for the BHAT study group. Effect of propranolol on ventricular arrhythmia. The betablocker heart attack trial experience. Circulation. 1983;67(Suppl 1): 5-10. 41. Olsson G, Rehnquist N. Ventricular arrhythmias during the first year after myocardial infraction: influence of long-term treatment with metaprolol. Circulation. 1984;69:1129-34. 42. Preliminary report: Effect of encanide and flecanide on mortality in a randomized trial of arrhythmia suppression after myocardial infarction. The Cardiac Arrhythmia Suppression Trial (CAST) Investigators. N Engl J Med. 1989; 321:406-17. 43. Talajic M. Long term management of ventricular arrhythmias-to treat or not to treat? Can J Cardiol. 1991;7:96-7. 44. Swerdlow CD, Winkle RA, Mason WJ. Determinants of survival in patients with ventricular tachyarrhythmias. N Engl J Med. 1983; 308:1436-42. 45. Levy D, Anderson KM, Savage DD, Balkus SA, Kannel WB, Castelli WP. Risk of ventricular arrhythmias in left ventricular hypertrophy: The Framingham Heart Study. Am J Cardiol. 1987;60:560-5. 46. Levy D, Anderson KM, Plehn J, Savage DD, Christiansen JC, Castelli WP. Echocardiographically determined left ventricular structural and functional correlates of complex or frequent ventricular arrhythmias on one-hour ambulatory electrocardiographic monitoring. Am J Cardiol. 1987;59:836-40. 47. Messerili FH, Ventura HD, Elizardi DJ, Dunn FG, Frohlich ED. Hypertension and sudden death. Increased ventricular ectopic activity in left ventricular hypertrophy. Am J Med. 1984;77:18-22. 48. Kannel WB, Gordon T, Castelli WP, Margolis Jr. Electrocardiographic left ventricular hypertrophy and risk of coronary heart disease. The Framingham Study. Ann Intern Med. 1970; 72:813-22. 49. Levy D, Garrison RJ, Savage DD, Kannel WB, Castelli WP. Left ventricular mass and incidence of coronary heart disease in elderly cohort. The Framingham Heart Study. Ann Intern Med. 1989;110: 101-7. 50. Casale PN, Devereux RB, Milner M, Zullo G, Harshfield GA, Pickering TG. Value of echocardiographic left ventricular mass in predicting cardiovascular morbidity events in hypertensive men. Ann Intern Med. 1986;105:173-8. 51. Risk stratification and survival after myocardial infarction. N Engl J Med. 1983;309:331-6.
15 December 1992 • Annals of Internal Medicine • Volume 117 • Number 12
Downloaded From: http://annals.org/ by a University of Otago User on 12/08/2016
• Objective: To evaluate the prevalence and prognostic significance of asymptomatic complex or frequent ventricular premature beats detected during ambulatory electrocardiographic (ECG) monitoring. • Design: Cohort study with a follow-up period of 4 to 6 years. • Setting: Population-based. • Participants: Surviving patients of the original Framingham Heart Study cohort and offspring of original cohort members (2727 men and 3306 women). • Measurements: One-hour ambulatory electrocardiography. • Results: The age-adjusted prevalence of complex or frequent arrhythmia (more than 30 ventricular premature complexes per hour or multiform premature complexes, ventricular couplets, ventricular tachycardia, or R-on-T ventricular premature complexes) was 12% (95% CI, 11 % to 13%) in the 2425 men without clinically evident coronary heart disease and 33% (CI, 24% to 42%) in the 302 men with coronary heart disease. The corresponding values in women (3064 without disease and 242 with disease) were 12% (CI, 1 1 % to 13%) and 26% (CI, 9% to 43%). After adjusting for age and traditional risk factors for coronary heart disease in a Cox proportional hazards model, men without coronary heart disease who had complex or frequent ventricular arrhythmias were at increased risk for both all-cause mortality (relative risk, 2.30; CI, 1.65 to 3.20) and the occurrence of myocardial infarction or death from coronary heart disease (relative risk, 2.12; CI, 1.33 to 3.38). In men with coronary heart disease and in women with and without coronary heart disease, complex or frequent arrhythmias were not associated with an increased risk for either outcome. • Conclusions: In men who do not have clinically apparent coronary heart disease, the incidental detection of ventricular arrhythmias is associated with a twofold increase in the risk for all-cause mortality and myocardial infarction or death due to coronary heart disease. The preventive and therapeutic implications of these findings await further investigation.
Annals of Internal Medicine. 1992;117:990-996. From the Framingham Heart Study, Framingham, Massachusetts; the National Heart, Lung, and Blood Institute, Bethesda, Maryland; and Boston University School of Medicine and Beth Israel Hospital, Boston, Massachusetts. For current author addresses, see end of text. 990
Although the presence of complex or frequent ventricular arrhythmia has been shown to be associated with a poor prognosis in patients with coronary heart disease (1-9), uncertainty persists about the importance of ventricular ectopy in persons without clinically apparent coronary heart disease. Some earlier studies suggested that ventricular premature beats, even in apparently healthy persons, were associated with an increased risk for sudden death (10, 11) and an increased incidence of ischemic heart disease (12, 13), but more recent studies have suggested that the presence of ventricular premature beats in apparently healthy persons is not associated with an increased risk for death (14-16). The Framingham Heart Study, because of its large free-living population, was in a unique position to examine this issue. We studied the prevalence and prognostic significance of complex or frequent ventricular arrhythmias in Framingham Heart Study participants who routinely underwent ambulatory electrocardiographic (ECG) monitoring. Methods Study Sample In 1948, 5209 residents of Framingham, Massachusetts, 28 to 62 years of age, were selected to undergo biennial examinations in a prospective study. In 1971, 5135 offspring of the original cohort (and spouses of these offspring) were entered into a prospective study. Selection criteria and study design have been reported previously (17-20). From 1979 to 1983, echocardiograms, 12-lead resting ECGs, and 1-hour ambulatory ECGs were obtained in surviving patients of the original cohort undergoing their sixteenth biennial examination and in offspring participants undergoing their second examination. Each participant gave informed consent before the study. Ambulatory ECG data were available for 97% of the 6218 Framingham Heart Study participants who attended the index examination (Table 1). Events Related to Coronary Heart Disease The follow-up period for our study was approximately 6 years for persons in the original cohort and 4 years for offspring participants. At each follow-up examination (2, 4, and 6 years later for the original cohort; 4 years later for offspring), interim events related to coronary heart disease were assessed by medical history, physical examination, and 12-lead ECG. In addition, medical records were obtained for participants who did not attend an examination and were evaluated for evidence of interim coronary heart disease. All suspected coronary heart disease events were reviewed by a committee of three physicians who evaluated pertinent medical records, hospitalization records, and pathology reports. Two primary outcomes were considered in our study: all-cause mortality and myocardial infarction or death from coronary heart disease. Criteria for various coronary heart disease events have been reported previously (21). The coronary heart disease events included myocardial infarction, coronary insufficiency (unstable angina), angina pectoris, and death from coronary heart disease (both sudden and nonsudden). Unwitnessed deaths occurring during
15 December 1992 • Annals of Internal Medicine • Volume 117 • Number 12
Downloaded From: http://annals.org/ by a University of Otago User on 12/08/2016
Table 1. Evaluable Participants Variable
Participants, n
Study population (cohort and offspring) Died before the index examination* Did not attend index examination Attended index examination No ambulatory electrocardiographic data Evaluable cases
10 344 2109 2017 6218
185 6033
* Index examination was examination 16 for members of the original cohort and examination 2 for offspring participants.
sleep did not meet criteria for sudden death caused by coronary heart disease (21). The diagnosis of congestive heart failure was based on fulfillment of two major criteria or one major criterion plus two minor criteria as described previously (21). Ambulatory Electrocardiographic Methods Two-lead, frequency-modulated recorders (simultaneous VI and V5, Clinical Data Inc.) were used for ambulatory recording. Analysis of the continuous ambulatory ECG records was aided by the use of a high-speed digital computer system (Clinical Data Inc.). Premature depolarizations and arrhythmias were identified and interpreted by a technician. All representative tracings were reviewed by a cardiologist. Classification of ventricular arrhythmias was done using a method similar to that of Lown and Wolf (22) and was based on the presence of the following arrhythmia types: ventricular premature complexes, multiform ventricular premature complexes, ventricular couplets, ventricular tachycardia, and R-on-T ventricular premature complexes. The summary arrhythmia-complex or frequent ventricular premature complexes (more than 30 ventricular premature complexes per hour or multiform premature complexes, ventricular couplets, ventricular tachycardia, or R-on-T ventricular premature complexes)-was used as a single dichotomous variable in all outcome analyses. Ventricular tachycardia was defined as three or more consecutive ventricular premature complexes. Echocardiography The echocardiographic methods used in the Framingham Heart Study have been described previously (23). Participants were studied using a standard M-mode echocardiographic technique (24). Measurements were made according to the methods outlined by Devereux and Reichek (25). Left ventricular internal dimension and the ventricular septum and left ventricular posterior wall thickness were measured at end diastole. Left
ventricular mass (in grams) was calculated from the following formula (24): 1.04 [(LVID + VST + PWT)3 - (LVID)3] 13.6, where LVID = left ventricular internal diameter, VST » ventricular septal thickness, and PWT = posterior wall thickness. Left ventricular mass was normalized according to height (in meters) as described previously (23). Left ventricular hypertrophy was defined as a left ventricular mass/height ratio two standard deviations or more above the mean for a healthy reference population. The cutoff values for left ventricular hypertrophy were 143 g/m in men and 102 g/m in women (23). Adequate echocardiographic data were available in about 80% of the sample. No echocardiographic criteria for eligibility were applied in this study.
Statistical Methods Separate analyses were carried out for men and women, as well as for participants with and without clinically evident coronary heart disease. Clinical characteristics were compared between disease-status groups after adjustment for age (age classification was done at the index examination). Percents were adjusted directly to the appropriate sex-specific age distribution, and the Cochran-Mantel-Haenszel statistic was used in between-group comparisons. Age adjustments for continuous variables were done using 'least squares' means (26), and between-group comparisons for these variables were done using t-tests (27). Prevalence rates for participants classified by arrhythmia grade and stratified by sex and coronary heart disease status were directly adjusted for age. The 95% CIs are provided where appropriate. The two outcome events studied were all-cause mortality and myocardial infarction or death from coronary heart disease. For each outcome, age- and risk-factor-adjusted relative risks associated with the presence of arrhythmia were analyzed with age-stratified proportional hazards regression models (28, 29). A result was deemed significant if P < 0.05. The covariates in these analyses were chosen because of their established potential roles as risk factors for coronary heart disease and death, or because of their possible confounding role with regard to ventricular arrhythmias and death. They included age, sex, high blood pressure (systolic pressure > 160 mm Hg or diastolic pressure > 95 mm Hg), number of cigarettes smoked per day, ratio of total cholesterol to high-density lipoprotein cholesterol, diabetes (yes = 1; no = 0), body mass index (weight in kilograms divided by the square of height in meters), hypertension therapy (those currently taking prescribed medication for the treatment of hypertension, excluding beta-blockers, were assigned a value of 1, and all others were assigned a value of 0), beta-blocker therapy (yes = 1; no = 0), the use of antiarrhythmic medications excluding betablockers (yes = 1; no = 0), and congestive heart failure (yes =
Table 2. Mean Age-adjusted Values for Selected Characteristics* Variable
Age,v Total cholesterol/HDL cholesterol Cigarettes/day Body mass index, kg/m2 Systolic blood pressure, mm Hg Diastolic blood pressure, mm Hg Hypertension, % Hypertension therapy, % Beta-blockers, % Diabetes, % Congestive heart failure, % Antiarrhythmic therapy, % LV mass/height, g/m LV fractional shortening, %
NoCHD (n = 2425)
Men CHD (n = 302)
P Valuet
NoCHD (n = 3064)
Women CHD (n = 242)
P Valuet
51.3 5.1 7.0 26.6 134 80.6 24 16 4 5 0.2 0.3 118 36
66.0 5.8 7.2 27.2 133 79.2 35 22 29 7 4 3 139 34
< 0.001 < 0.001 > 0.2 0.01 >0.2 0.02 < 0.001 >0.2 < 0.0001 0.01 < 0.001 < 0.001 < 0.001 < 0.001
53.8 4.4 5.6 25.5 131 76.8 26 21 4 4 0.8 0.2 89.4 39
70.4 5.1 4.8 26.4 134 76.1 55 48 18 19 5 1 103.4 38
< 0.001 < 0.001 > 0.2 0.007 0.004 >0.2 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 0.03
* CHD = coronary heart disease; HDL = high-density lipoprotein; LV = left ventricular. Risk factor data, excepting echocardiographic data | available), were in about 97% of participants. t The P value reflects age-adjusted comparisons between participants with and without coronary heart disease. 15 December 1992 • Annals of Internal Medicine • Volume 117 • Number 12
Downloaded From: http://annals.org/ by a University of Otago User on 12/08/2016
991
Table 3. Age-adjusted Prevalence of Ventricular Arrhythmias on Monitoring LMI 1-hour 1 - I 1 U U I Ambulatory / IL i u u u i a i u i j Electrocardiographic LUMllUI l l l g A I C V U u v a i ui Variable
> 1 ventricular premature complex/hour, % Frequent ventricular premature complexes, %t Complex arrhythmia, % Multiform Ventricular couplets Ventricular tachycardia^ R-on-T complex Complex or frequent arrhythmia, %
NoCHD (n = 2425)
Men CHD (n = 302)
33 7 10 10 3 0.7 0.4 12
58 17 32 31 12 2 0.9 33
P Value < < < <
0.2 < 0.001
NoCHD (n = 3064)
Women CHD in m 242)
P Value
32 7 9 8 3 0.4 0.5 12
49 19 14 13 5 0.3 0.6 26
0.009 0.03 0.03 < 0.001 < 0.001 >0.2 > 0.2 < 0.001
* CHD = coronary heart disease. t Frequent ventricular premature complex defined as more than 30/h. t Ventricular tachycardia defined as 3 or more consecutive ventricular premature complexes.
1; no = 0). Beta-blocker use was considered as a separate dichotomous variable in multivariate analyses because this class of medications possesses antihypertensive, anti-ischemic, and antiarrhythmic properties. Echocardiographic variables considered in subsequent analyses included left ventricular mass and left ventricular fractional shortening (the difference between diastolic and systolic left ventricular diameters divided by diastolic diameter x 100). Results Study Sample Of the 6033 eligible participants, 2425 men (mean age, 51.3 years) and 3064 women (mean age, 53.8 years) showed no clinical evidence of coronary heart disease; 302 men (mean age, 66.0 years) and 242 women (mean age, 70.4 years) had clinically apparent coronary heart disease at the index examination. Age-adjusted values for various characteristics are summarized according to coronary heart disease status in Table 2. For both sexes, participants who did not have coronary heart disease were younger than those with coronary heart disease, had a lower ratio of total cholesterolrHDL cholesterol, lower prevalences of diabetes, hypertension, congestive heart failure, and lower left ventricular
mass. In the coronary heart disease group, 195 of 302 men (65%) and 77 of 242 women (32%) had myocardial infarctions before the baseline examination. Prevalence of Ventricular Premature Beats Age-adjusted prevalence rates for the various arrhythmia grades according to coronary heart disease status are presented in Table 3. In men, the age-adjusted prevalence of 1 or more ventricular premature beats per hour was 33% in those without coronary heart disease and 58% in those with coronary heart disease (P < 0.001). Corresponding rates in women were 32% and 49% (P = 0.009). The age-adjusted prevalence of complex or frequent ventricular premature beats was 12% (CI, 11% to 13%) in men without coronary heart disease and 33% (CI, 24% to 42%) in those with coronary heart disease group (P < 0.001). Corresponding values in women were 12% (CI, 11% to 13%) and 26% (CI, 9% to 43%) (P < 0.001). Age-adjusted prevalence of frequent ventricular arrhythmias (> 30 ventricular premature complexes per hour) was 7% in men without coronary heart disease and 17% in men with coronary heart disease (P < 0.001). Corresponding values in women were 7% and 19% (P = 0.03). In men, the age-adjusted prevalence of complex ventricular arrhythmias (multiform premature ventricular complexes, couplets, ventricular tachycardia, or R-on-T ventricular premature complexes) was 10% in those without coronary heart disease and 32% in those with coronary heart disease (P < 0.001). Corresponding values in women were 9% and 14% (P < 0.001). Age-specific prevalences of complex or frequent arrhythmia are shown in Figure 1. The prevalence of arrhythmia increases with advancing age in men and women, regardless of coronary heart disease status. In both men and women without coronary heart disease, the prevalence roughly doubled in each successive age group. Association of Arrhythmia and Outcome
Figure 1. Age-specific prevalence of complex or frequent ventricular arrhythmias in men and women by coronary heart disease status. CHD = coronary heart disease.
992
Age- and covariate-adjusted analyses of complex or frequent arrhythmia in relation to outcome are shown in Table 4. In men without clinically apparent coronary heart disease, the age- and covariate-adjusted relative
15 December 1992 • Annals of Internal Medicine • Volume 117 • Number 12
Downloaded From: http://annals.org/ by a University of Otago User on 12/08/2016
cause mortality (relative risk, 2.26; CI, 1.39 to 3.66; P = 0.001) and for myocardial infarction or death from coronary heart disease (2.59; CI, 1.40 to 4.79; P = 0.002).
risk associated with complex or frequent arrhythmia was increased both for all-cause mortality (relative risk, 2.30; CI, 1.65 to 3.20; P < 0.0001) and for myocardial infarction or death from coronary heart disease (relative risk, 2.12; CI, 1.33 to 3.38; P = 0.002). In men with coronary heart disease, complex or frequent arrhythmia was not significantly associated with all-cause mortality (relative risk, 1.27; CI, 0.80 to 2.01; P > 0.2) or myocardial infarction or death from coronary heart disease (relative risk, 0.95; 95% CI, 0.56 to 1.60; P > 0.2). In women, complex or frequent arrhythmia was not associated with either outcome, regardless of coronary heart disease status (Table 4). The survival curves of the participants without clinically apparent coronary heart disease, presented in Figure 2 according to arrhythmia status, were predicted by the proportional hazards model. Estimates were made at observed failure times, rounded to the nearest 1/100 year; therefore, the curves do not appear as smooth functions. Men without coronary heart disease who had complex or frequent arrhythmia showed reduced survival compared with those who did not have complex or frequent arrhythmia (Figure 2, panels A and B). No significant difference in survival was noted in women without coronary heart disease (Figure 2, panels C and D). Among both men and women, those with and those without coronary heart disease showed no significant difference in outcome according to arrhythmia status. In a separate multivariate model, the association of complex or frequent arrhythmia with outcome in men who did not have clinically evident coronary heart disease was adjusted for echocardiographic left ventricular mass and left ventricular fractional shortening, as well as for age and other covariates. The presence of arrhythmia was associated with an increased risk for all-
Discussion Our study documents a significant association between asymptomatic ventricular arrhythmias and the risk for all-cause mortality and myocardial infarction or death from coronary heart disease in men without clinically evident coronary heart disease. Moreover, when considered alongside traditional risk factors for coronary heart disease, including age, blood lipid levels, cigarette smoking, obesity, and hypertension, ventricular arrhythmias remained predictive of both outcomes. In women without coronary heart disease and in persons of both sexes with coronary heart disease, no significant associations between arrhythmia and outcome were noted. Complex or frequent ventricular arrhythmias occur commonly in the general population. In both men and women, the prevalence of ventricular arrhythmia was higher in those with coronary heart disease than in those without it. Among participants with coronary heart disease, the prevalence of ventricular arrhythmia was higher in men than in women. Previous studies of the prognostic importance of ventricular premature beats on routine ECGs in selected, apparently healthy persons have shown conflicting results (10-16): Earlier studies found an increased risk for mortality in patients with ventricular premature beats (10-13), but later studies did not (14-16). Our findings of an increased risk for all-cause mortality and for myocardial infarction or death from coronary heart disease
Table 4. Relation of Complex or Frequent Ventricular Arrhythmia to All-Cause Mortality and to Myocardial Infarction or Death from Coronary Heart Disease Variable
All deaths Age-adjusted relative risl: (95% CI) Deaths/participants at ris k, n/n P value Age- and covariate-adjusited relative risk (95% CI)t Deaths/subjects at risk, n In P value Myocardial infarction or CI ID death Age-adjusted relative risli (95% CI) Myocardial infarctions arid CHD deaths/subjects at risk, n/n P value Age- and covariate-adjusted relative risk (95% CI)t Myocardial infarctions and CHD deaths/subjects at risk, n/n P value
Woimen
M en No CHD
CHD
No CHD
CHD
2.36 (1.72 to 3..24) 186/2423 < 0.0001
1.64 (1.10 to 2.45) 100/302 0.01
1.31 (0.94 to 1 .83) 188/3062 0.11
1.61 (1.00 to 2.60) 71/242 0.05
2.30 (1.65 to 3..20)
1.27 (0.80 to 2.01)
1.23 (0.85 to 1 .77)
1.20 (0.69 to 2.09)
173/2368 < 0.0001
90/281 > 0.2
164/2889 > 0.2
60/214 > 0.2
2.17 (1.40 to 3..37)
1.25 (0.78 to 1.19)
1.13 (0.57 to 2 .21)
0.87 (0.44 to 1.71)
111/2423 0.0005
73/302 > 0.2
52/3062 >0.2
41/242 > 0.2
2.12 (1.33 to 3..38)
0.95 (0.56 to 1.60)
1.01 (0.49 to 2 .07)
0.70 (0.33 to 1.48)
106/2368 0.002
68/281 > 0.2
48/2889 > 0.2
37/214 > 0.2
* CHD = coronary heart disease. t Covariates included the ratio of total cholesterol to high-density lipoprotein cholesterol, smoking, body mass index, hypertension or antihypertensive therapy, systolic blood pressure, diabetes, congestive heart failure, beta-blocker use, and antiarrhythmic therapy (coronary heart disease group only).
15 December 1992 • Annals of Internal Medicine • Volume 117 • Number 12
Downloaded From: http://annals.org/ by a University of Otago User on 12/08/2016
993
Figure 2. Survival curves. Survival function is shown according to the presence or absence of arrhythmia. Each panel shows the number of study participants still under observation at 2, 4, and 6 months. Panel A. Survival curves for myocardial infarction or death from coronary heart disease in men without coronary heart disease. The difference between arrhythmia groups for myocardial infarction (Ml) or death from coronary heart disease (CHD) was significant (P = 0.002). Panel B. Survival curves for all-cause mortality in men without coronary heart disease. The difference between arrhythmia groups for all-cause mortality was significant (P < 0.0001). Panel C. Survival curves for myocardial infarction or death from coronary heart disease in women without coronary heart disease. The curves were superimposed on each other. The difference between arrhythmia groups for myocardial infarction or coronary heart disease was not significant (P > 0.2). Panel D. Survival curves for all-cause mortality in women without coronary heart disease. The difference between arrhythmia groups for all-cause mortality was not significant (P > 0.2).
in apparently healthy men with complex or frequent ventricular arrhythmia are consistent with results from the Manitoba study (12), in which 3983 Canadian air pilots without overt coronary heart disease were prospectively followed for 10 years. However, that study failed to explore the role of potentially important covariables, such as hypertension and cigarette smoking, factors that may be associated with both arrhythmia and outcome. Kennedy and coworkers (30-32) also investigated the prognostic importance of complex or frequent ventricular arrhythmias in apparently healthy persons. They reported no increase in all-cause mortality or myocardial infarction or death from coronary heart disease in persons without coronary heart disease who had complex or frequent arrhythmia. The differences between our findings and those of Kennedy and colleagues (32) may be due to differences in sample size (659 994
participants with complex or frequent arrhythmia in our study compared with 73 such persons in the study by Kennedy and colleagues) or to the influence of betablocker use (27% of Kennedy and colleagues' study sample were receiving beta-blockers compared with 4% of men without coronary heart disease in our study). Other differences in study samples may have played a role, with selection bias more likely in the earlier clinical series. In persons with coronary heart disease, no statistically significant association was observed between arrhythmia and either outcome (all-cause mortality or myocardial infarction or death from coronary heart disease) (see Table 4). This finding is consistent with the results of two previous studies (33, 34). Nevertheless, several clinical studies of survivors of myocardial infarction have shown an increased risk for cardiac death
15 December 1992 • Annals of Internal Medicine • Volume 117 • Number 12
Downloaded From: http://annals.org/ by a University of Otago User on 12/08/2016
in those with ventricular arrhythmias, particularly when complex or frequent forms are present (1-7). The difference between our findings and those of these other studies may be attributable to the different selection processes, particularly in relation to the interval between onset of coronary heart disease and ECG monitoring. In our study, participants with coronary heart disease survived long enough to attend a routine screening examination. Unstable patients with recent myocardial infarction and those with coronary heart disease of new onset in whom arrhythmias may contribute to early morbidity or mortality may have been under-represented in our study. Complex or frequent arrhythmia may be a marker for occult coronary heart disease or silent ischemia. However, if occult coronary heart disease or silent ischemia was a contributor to both arrhythmia expression and adverse outcome, we would expect to find similar, if not stronger, associations between arrhythmia and outcome in patients with overt coronary heart disease. Because we did not observe a significant association between arrhythmia and outcome in patients with coronary heart disease, it is possible that a nonischemic mechanism underlies our findings in men without clinically apparent coronary heart disease. Previous studies suggested that hypertensive patients treated with diuretics may be at an increased risk for fatal cardiac events, perhaps because of an increased frequency of ventricular premature beats (35-38). The antiarrhythmic properties of beta-blockers might, in part, be responsible for improved survival among patients with myocardial infarction who received these drugs in several secondary prevention trials (39-41). The significance of antiarrhythmic therapy in patients with ventricular arrhythmias remains controversial (42, 43). Because of the potential effect of diuretics, beta-blockers, and other antiarrhythmic drugs on arrhythmia expression and outcome, we included these variables in our multivariate model. Congestive heart failure was used as a covariate because Swerdlow and colleagues (44) reported that the presence and severity of heart failure were the most powerful independent predictors of both cardiac death and sudden death in patients with ventricular tachyarrhythmias. We included left ventricular mass in the multivariate model because previous studies (45-47) showed an increased risk for ventricular arrhythmias in patients with left ventricular hypertrophy and increased mortality and an increased incidence of coronary heart disease in patients with increased left ventricular mass (48-50). Additionally, left ventricular fractional shortening was also entered in multivariate model because a previous study showed that patients with left ventricular systolic dysfunction had increased mortality (51). After adjustments were made for left ventricular mass and fractional shortening, the increased risk for all-cause mortality and myocardial infarction or death from coronary heart disease in men without coronary heart disease who had complex or frequent arrhythmia persisted with only minimal change. This finding would suggest that complex or frequent ventricular arrhythmia is not merely a marker for occult left ventricular dysfunction or hypertrophy.
Because our study was conducted in a large population-based sample, we avoided referral biases inherent in earlier clinical studies. Ambulatory ECG monitoring was successfully done in about 97% of the persons who attended the index examination. In addition, we carried out extensive analyses to explore the roles of numerous potential confounding variables that may influence outcome and the association between arrhythmia and outcome. Our sample consisted predominantly of white persons, and therefore our data should be interpreted with caution when considering other races. In our study, 1-hour ambulatory ECG monitoring was done in 6033 persons during a routine examination. Because of substantial variation in the occurrence of ventricular arrhythmia, the use of 1-hour ambulatory ECG monitoring may have resulted in an underestimation of the number of participants predisposed to arrhythmia. However, the large sample size would minimize the effect of variation. Even though the use of 24-hour monitoring would have resulted in a higher prevalence of arrhythmia, the 1-hour monitoring was sufficient to predict both outcomes in men without coronary heart disease who had complex or frequent arrhythmia. Because stress testing and angiographic studies were not done in our study, the possibility of occult coronary heart disease cannot be excluded in participants without clinically apparent coronary heart disease. Our results in the group with coronary heart disease should be interpreted with caution because this group was small and included patients with angina pectoris, coronary insufficiency, or myocardial infarction who survived their acute coronary event and were able to attend a routine examination, often several years after disease onset. Thus, we studied a relatively healthy group with coronary heart disease. Our results should not be extrapolated to patients with unstable or recent-onset coronary heart disease. Last, the arrhythmias detected in our study participants were asymptomatic and low grade. Our findings may not be applicable to patients with symptomatic or higher-grade ventricular arrhythmias. Currently, no evidence exists that antiarrhythmic therapy in patients with asymptomatic ventricular arrhythmias improves prognosis, regardless of whether or not coronary heart disease is present. The mechanism of the observed increased risk associated with ventricular arrhythmias in men without clinically apparent coronary heart disease is unknown. The extent to which 1-hour ambulatory ECG monitoring has utility and any therapeutic implications in the clinical setting are also unknown. Further prospective studies are needed to identify the mechanisms of increased risk in apparently healthy subjects with ventricular arrhythmias. Grant Support: In part by an educational grant from Parke-Davis (Dr. Bikkina). Requests for Reprints: Daniel Levy, MD, Framingham Heart Study, 5 Thurber Street, Framingham, MA 01701. Current Author Addresses: Dr. Bikkina: University of South Alabama, Department of Cardiology, 2451 Fillingim Street, Tenth Floor, Mobile, AL 36617. Drs. Larson and Levy: Framingham Heart Study, 5 Thurber Street, Framingham, MA 01701.
15 December 1992 • Annals of Internal Medicine • Volume 117 • Number 12
Downloaded From: http://annals.org/ by a University of Otago User on 12/08/2016
995
References 1. Kotler MN, Tabatznik B, Mower MM, Tominaga S. Prognostic significance of ventricular ectopic beats with respect to sudden death in the late post infarction period. Circulation. 1973;47:959-66. 2. Prognostic importance of premature beats following myocardial infarction. Experience in the coronary drug project. JAMA. 1973;223: 1116-24. 3. Ruberman W, Wernblatt M, Goldberg JD, Frank CW, Shapiro S. Ventricular premature beats and mortality after myocardial infarction. N Engl J Med. 1977;297:750-7. 4. Moss AJ, DeCamilla J, Engstrom F, Hoffman W, Odoroff C, Davis H. The posthospital phase of myocardial infarction: identification of patients with increased mortality risk. Circulation 1974;49:460-6. 5. Bigger JT Jr, Dresdale RJ, Heissenbuttel RH, Weld FM, Wit AL. Ventricular arrhythmias in ischemic heart disease: mechanism, prevalence, significance, and management. Prog Cardiovasc Dis. 1977;19:255-300. 6. Lie KI, Wellens HJJ, Downar E, Durrer D. Observations on patients with primary ventricular fibrillation complicating acute myocardial infarction. Circulation. 1975;52:755-9. 7. Conolly SJ, Carins JA. Comparison of one-, six- and 24-hour ambulatory electrocardiographic monitoring for ventricular arrhythmia as a predictor of mortality in survivors of acute myocardial infarction. CAMIAT Pilot Study Group. Canadian Amiodarone Myocardial Arrhythmia Trial. Am J Cardiol. 1992;69:308-13. 8. Moss AJ. Clinical significance of ventricular arrhythmias in patients with and without coronary artery disease. Prog Cardiovasc Dis. 1980;23:33-52. 9. Horan MJ, Kennedy HL. Ventricular ectopy. History, epidemiology, and clinical implications. JAMA. 1984;251:380-6. 10. Chiang RN, Perlman LV, Ostrander LD Jr, Epstein FM. Relationship of premature systoles to coronary heart disease and sudden death in the Tecumseh epidemiologic study. Ann Intern Med. 1969; 70:1159-69. 11. Hinkle LE Jr, Carver ST, Stevens M. The frequency of asymptomatic disturbances of cardiac rhythm and conduction in middle-aged men. Am J Cardiol. 1969;24:629-50. 12. Pell S, D'Alonzo CA. A three year study of myocardial infraction in a large employed population. JAMA. 1961;175:463-70. 13. Rabkin SW, Mathewson FA, Tate RB. Relationship of ventricular ectopy in men without apparent heart disease to occurrence of ischemic heart disease and sudden death. Am Heart J. 1981 ;101: 135-42. 14. Rodstein M, Wolloch L, Gubner RS. Mortality study of the significance of extrasystoles in an insured population. Circulation. 1971; 44:617-25. 15. Fisher FD, Tyroler HA. Relationship between ventricular premature contractions on routine electrocardiography and subsequent sudden death from coronary heart disease. Circulation. 1973;47:712-9. 16. Elkon KB, Swerdlow TA, Myburgh DP. Persistent ventricular ectopic beats: a long-term study. S Afr Med J. 1977;52:564-6. 17. Dawber TR, Meadors GF, Moore FE. Epidemiologic approaches to heart disease: the Framingham study. Am J Public Health. 1951;41: 279-86. 18. Dawber TR, Kannel WB, Lyell LP. An approach to longitudinal studies in a community: the Framingham Study. Ann NY Acad Sci. 1963;107:539-56. 19. Gordon T, Moore TE, Shurtleff D, Dawber TR. Some methodologic problems in the long-term study of cardiovascular disease: observations on the Framingham study. J Chronic Dis. 1959;10:186-206. 20. Kannel WB, Feinleib M, McNamara PM, Garrison RJ, Castelli WP. An investigation of coronary heart disease in families. The Framingham Offspring Study. Am J Epidemiol. 1979;110:281-90. 21. Shurtleff D. Some characteristics related to the incidence of cardiovascular disease and death: Framingham Heart Study 18 follow-up. In: Kannel WB, Gordon T, eds. The Framingham Study. Section 30. Washington, DC: U.S. Government Printing Office; 1974 (DHEW publication No. [NIH] 74-599): 17-25. 22. Lown B, Wolf M. Approaches to sudden death from coronary heart disease. Circulation 1971;44:130-42. 23. Levy D, Savage DD, Garrison RJ, Anderson KM, Kannel WB, Castelli WP. Echocardiographic criteria for left ventricular hypertrophy: The Framingham Heart Study. Am J Cardiol. 1987;59:956-60. 24. Feigenbaum H. Echocardiography. Third edition. Philadelphia: Lea & Febiger, 1981:51-118. 25. Devereux RB, Reichek N. Echocardiographic determination of left ventricular mass in man. Anatomic validation of the method. Circulation. 1977;55:613-8. 26. SAS Institute Inc. SAS/STAT User's Guide, Version 6, Fourth Edition, Volume 1. Cary, North Carolina: SAS Institute Inc.; 1989. 27. SAS Institute Inc.. SAS/STAT User's Guide, Version 6, Fourth Edition, Volume 2. Cary, North Carolina. SAS Institute Inc.; 1989.
996
28. Kalbfleisch JD, Prentice RL. The Statistical Analysis of Failure Time Data. New York: John Wiley & Sons.; 1980. 29. SAS Institute Inc. SAS Technical Report P-217, SAS/STAT Software: The PHREG Procedure, Version 6. Cary, North Carolina: SAS Institute Inc.; 1991. 30. Kennedy HL, Underbill SJ. Frequent and complex ventricular ectopy in apparently healthy subjects: a clinical study of 25 cases Am J Cardiol. 1976;38:141-8. 31. Kennedy HL, Pescarmona JE, Bouchard RJ, Goldberg RJ. Coronary artery status of apparently healthy subjects with frequent or complex ventricular ectopy. Ann Intern Med. 1980;92:179-85. 32. Kennedy HL, Whitlock JA, Sprague MK, Kennedy LJ, Bucklingham TA, Goldberg RJ. Long-term follow-up of asymptomatic healthy subjects with frequent and complex ventricular ectopy. N Engl J Med. 1985;312:193-7. 33. de Soyza N, Bennett FA, Murphy ML, Bissett JK, Kane JJ. The relationship of paroxysmal ventricular tachycardia complicating the acute phase of ventricular arrhythmia during the late hospital phase of myocardial infarction to long-term survival. Am J Med. 1978; 64:377-81. 34. Cats VM, Lie KI, Van Capelle FJ, Durrer O. Limitations of 24 hour ambulatory electrocardiographic recording in predicting coronary events after acute myocardial infarction. Am J Cardiol. 1979;44: 1257-62. 35. Multiple Risk Factor Intervention Trial: risk factor changes and mortality results. Multiple Risk Factor Intervention Trial Research Group. JAMA. 1982;248:1465-77. 36. MRC trial of treatment of mild hypertension: principal results. Br Med J (Clin Res Ed). Medical Research Council Working Party. 1985;291:97-104. 37. Levy D, Anderson KM, Christiansen JC, Campanile G, Stokes J 3d. Antihypertensive drug therapy and arrhythmia risk. Am J Cardiol. 1988;62:147-9. 38. Ventricular extrasystoles during thiazide treatment: sub-study of MRC mild hypertension trial. The Medical Research Council Working Party on Mild to Moderate Hypertension. Br Med J. 1983;287: 1249-53. 39. Koppes GM, Beckmann CH, Jones FG. Propranolol therapy for ventricular arrhythmias two months after acute myocardial infarction. Am J Cardiol. 1980;46:322-8. 40. Uchstein E, Morganroth J, Harrist Hubble E, for the BHAT study group. Effect of propranolol on ventricular arrhythmia. The betablocker heart attack trial experience. Circulation. 1983;67(Suppl 1): 5-10. 41. Olsson G, Rehnquist N. Ventricular arrhythmias during the first year after myocardial infraction: influence of long-term treatment with metaprolol. Circulation. 1984;69:1129-34. 42. Preliminary report: Effect of encanide and flecanide on mortality in a randomized trial of arrhythmia suppression after myocardial infarction. The Cardiac Arrhythmia Suppression Trial (CAST) Investigators. N Engl J Med. 1989; 321:406-17. 43. Talajic M. Long term management of ventricular arrhythmias-to treat or not to treat? Can J Cardiol. 1991;7:96-7. 44. Swerdlow CD, Winkle RA, Mason WJ. Determinants of survival in patients with ventricular tachyarrhythmias. N Engl J Med. 1983; 308:1436-42. 45. Levy D, Anderson KM, Savage DD, Balkus SA, Kannel WB, Castelli WP. Risk of ventricular arrhythmias in left ventricular hypertrophy: The Framingham Heart Study. Am J Cardiol. 1987;60:560-5. 46. Levy D, Anderson KM, Plehn J, Savage DD, Christiansen JC, Castelli WP. Echocardiographically determined left ventricular structural and functional correlates of complex or frequent ventricular arrhythmias on one-hour ambulatory electrocardiographic monitoring. Am J Cardiol. 1987;59:836-40. 47. Messerili FH, Ventura HD, Elizardi DJ, Dunn FG, Frohlich ED. Hypertension and sudden death. Increased ventricular ectopic activity in left ventricular hypertrophy. Am J Med. 1984;77:18-22. 48. Kannel WB, Gordon T, Castelli WP, Margolis Jr. Electrocardiographic left ventricular hypertrophy and risk of coronary heart disease. The Framingham Study. Ann Intern Med. 1970; 72:813-22. 49. Levy D, Garrison RJ, Savage DD, Kannel WB, Castelli WP. Left ventricular mass and incidence of coronary heart disease in elderly cohort. The Framingham Heart Study. Ann Intern Med. 1989;110: 101-7. 50. Casale PN, Devereux RB, Milner M, Zullo G, Harshfield GA, Pickering TG. Value of echocardiographic left ventricular mass in predicting cardiovascular morbidity events in hypertensive men. Ann Intern Med. 1986;105:173-8. 51. Risk stratification and survival after myocardial infarction. N Engl J Med. 1983;309:331-6.
15 December 1992 • Annals of Internal Medicine • Volume 117 • Number 12
Downloaded From: http://annals.org/ by a University of Otago User on 12/08/2016
