OUP user menu

Prevalence and prognostic significance of daily-life silent myocardial ischaemia in middle-aged and elderly subjects with no apparent heart disease

Ahmad Sajadieh , Olav Wendelboe Nielsen , Verner Rasmussen , Hans Ole Hein , Jørgen Fischer Hansen
DOI: http://dx.doi.org/10.1093/eurheartj/ehi169 1402-1409 First published online: 17 March 2005


Aims We aimed to determine the prevalence and prognostic significance of daily-life silent myocardial ischaemia (SMI) in healthy middle-aged and elderly subjects with no previous heart disease.

Methods and results Six hundred and seventy-eight healthy men and women between 55 and 75 years of age and with no history of cardiovascular disease or stroke were included. Baseline examinations included physical examination, fasting laboratory testing, and 48 h ambulatory electrocardiogram monitoring. An episode of ischaemia was defined by a down-sloped or horizontal ST depression of at least 1 mm at a duration of at least 1 min. Seventy-seven subjects (11.4%) had SMI. All participants were followed for up to 5 years. In 77 subjects with SMI, 16 (20.7%) had an event (death or myocardial infarction). In 601 subjects without SMI, 50 (8.3%) had an event. The hazard ratios for SMI in relation to cardiac and combined events after correction for conventional risk factors were 3.1 [(1.24–7.97), P=0.016] and 1.97 [(1.06–3.69), P=0.033], respectively.

Conclusion SMI as detected by Holter monitoring was detected in 11.4% of these subjects and was associated with more than three-fold increase in the cardiac event rate after correction for risk factors, implying that this test could be used to identify high-risk individuals among these subjects.

  • Silent myocardial ischaemia
  • Holter recording
  • Coronary heart disease
  • Risk factors


In patients with stable or unstable coronary artery disease (CAD), silent myocardial ischaemia (SMI) is quite prevalent and has about the same grave prognostic value as symptomatic ischaemia.13 The prognostic significance of SMI in asymptomatic subjects without known CAD has been controversial. In apparently healthy and relatively young subjects with a low a priori probability of CAD, exercise testing usually reveals a low prevalence of SMI and a high false-positive rate.4,5 Therefore, the use of exercise test as a screening modality in the general population has not been recommended.6,7 However, in a recent large study of apparently healthy subjects, SMI detected during and after exercise testing was associated with a two- to six-fold increase in the coronary event rate in a 5 year period, especially in the high-risk subgroups, and the authors suggest exercise testing to identify subjects at very high risk.8 These results are in agreement with other studies.6 Daily SMI as detected by Holter monitoring may detect subgroups other than those identified by exercise testing and may thus have a different prognostic value.

The value of Holter monitoring to identify SMI and its prognostic significance in apparently healthy subjects is not well established. Thus, we planned to study the prevalence and prognostic significance of daily SMI in a middle-aged and elderly population without apparent or previous heart disease.


This study is a part of the Copenhagen Holter study, which aimed to address the value of 48 h Holter recording in risk assessment of middle-aged and elderly men and women with no apparent heart disease, especially in relation to other risk factors. The study was performed from 1998 to 1999. Every person in Denmark can be identified by a unique number in the ‘central personal register’, Ministry of the Interior. This enabled us to perform a representative epidemiological survey of subjects living within two well-defined postal regions in Copenhagen. Every man aged 55 and all men and women aged 60, 65, 70, and 75 received a questionnaire (n=2969) asking about cardiovascular risk factors, use of medication, and history: 2041 returned the questionnaire and 1743 agreed to participate in further investigations. We immediately excluded 348 for the following reasons: history of myocardial infarction, other cardiac diseases, stroke, cancer, and other significant or life threatening conditions. Thus, the remaining 1395 apparently healthy subjects were ranked according to the number of the following self-reported risk factors: hypertension, diabetes mellitus, smoking habits, familial pre-disposition to cardiac disease (sudden death or AMI in a parent or sibling before the age of 60), obesity [body mass index (BMI)>30], or hypercholesterolaemia. This study invited all subjects with two or more of these risk factors (n=576), and a completely random sample consisting of 60% (490) subjects with one or no apparent risk factor. In total, 775 subjects attended the Holter study (Figure 1). The rationale for including participants with multiple risk factors was that these would be expected to have the highest prevalence of SMI. The rationale for including a random sample of participants with none or one risk factor was to obtain a large reference population to study the prevalence of silent ischaemia in apparent low-risk subjects.

Figure 1 Flow diagram demonstrating the study population at different exclusion steps.

Of the 775 subjects who were enrolled in the Holter study, 33 were excluded from this study because of detection of one or more exclusion criteria at the start of the study (seven with cancers, nine with apparent cardiac diseases, six with other significant or life threatening diseases, two with baseline ST-segment depression of ≥1 mm, four with left bundle branch block, and five with significant Q-waves), and 64 subjects had technically unacceptable or incomplete recordings. In total, 678 subjects participated in the study and had acceptable Holter recordings. All participants were subject to an interview by a physician, physical examination including fasting laboratory testing, and 48 h continuous Holter monitoring. All subjects were asked about chest pain or equivalent symptoms, according to a modified ‘Rose questionnaire’. Participants were subdivided into two groups on the basis of the level of physical activity: group 1 with an almost sedentary life-style or a light physical activity level of 2–4 h per week, group 2 with a moderate to hard physical activity level of at least 4 h weekly physical activity or regular training. Diabetes mellitus was defined as known diabetes or a fasting plasma glucose of ≥7 mmol/L. Laboratory testing was performed in the morning after overnight fasting.

Continuous 48 h Holter recording was made by two-channel SpaceLabs tape recorders (9025, SpaceLabs, Inc., Redwood, WA, USA). Lead II and V5 were selected in all participants. Analysis of ST-segment depression was performed semi-manually by trained personal at the Holter laboratory. An episode of ischaemia was defined by a down-sloped or horizontal ST depression of at least 1 mm and duration of at least 1 min, separated from another episode by at least 1 min of no ST depression (the usual definition). In addition to this usual definition and to reduce the possible false-positive number, another group with more pronounced ST depression (at least 2 mm at maximum) was defined and studied (strict definition). The episodes of ischaemia were detected by the computer program and were evaluated by the technicians, who decided to accept or discard the episode, and additionally accepted or changed the suggested level of ST depression by the machine. Evaluation of the Holter recordings were completed before follow-up studies and performed by subjects blinded to and without any access to participants' files.


Before inclusion, all participants gave their written informed consent. The study was approved by the regional ethical committee of Copenhagen and Frederiksberg. The Helsinki Declaration was complied with.


Data on death and acute myocardial infarction (AMI) were obtained through the national central patient registry. All deaths and hospital admissions and discharges in Denmark should be reported to this registry within 2 weeks. Hospital admissions were additionally studied from hospital discharge letters. The diagnosis of AMI was based on history, typical ECG changes, and cardiac enzyme elevation. Combined endpoint was based on combined all-cause mortality or first AMI. A cardiac event was defined as a first AMI or death from coronary heart disease. Deaths from coronary heart disease were based on information on death certificates and hospital files.

Subgroup analyses

The following subgroups were targeted for separate evaluation: men, women, smokers, non-smokers, systolic blood pressure above or below median value (155 mmHg), total S-cholesterol above or below median value (6 mmol/L), BMI above or below median value (27 kg/m2), subjects with low vs. high level of physical activity, subjects with and without glucose intolerance (diabetes mellitus or fasting blood sugar >6 mmol/L), and subjects with a Framingham risk-score above vs. below the calculated average. Cut-off values for blood pressure, BMI, and S-cholesterol were based on the median values. To calculate the Framingham risk-score, the method described by Wilson et al.9 was used, on the basis of information on age, sex, total cholesterol, HDL cholesterol, blood pressure, diabetes mellitus, and smoking habits.

Statistical analyses

Statistical analyses were performed using SAS statistical software program (version 8.2). The univariate association between SMI and other parameters were evaluated by Student's t-test, Mann–Whitney, Kruscal–Wallis, or χ2 test as appropriate. Two-tailed tests of significance are reported and P-values less than 0.05 are considered statistically significant. Event-free survival in patients with and without SMI was compared by the method of Kaplan–Meier, and statistical differences were evaluated by means of the log-rank test. Cox proportional hazard models were used to evaluate the risk factor adjusted associations of SMI with death or AMI. The assumption of proportional hazards was assessed by visual judgement of the log-minus-log survival plots. Covariates were entered as continuous variables when possible. We checked the assumption of linearity for a continuous variable by entering the transformed variable in addition to the variable of interest. The natural logarithm and square transformations were used. A significant change in the −2 log-likelihood for any model was considered as a sign of non-linearity, otherwise the linearity assumption was accepted. All variables met the linearity assumption except cholesterol, which showed a J-form association with combined events because of increased mortality in a subgroup of 18 subjects, indicating worse prognosis if cholesterol was <4 mmol/L. Cholesterol was nonetheless maintained as a continuous untransformed variable to cohere with other studies, because linearity was demonstrated when a pure cardiac event was used and because a dichotomized variable (above or below 4 mmol/L) did not change any of the main results.

The following variables of potential prognostic importance were evaluated in the Cox models: age, sex, smoking, serum cholesterol, systolic arterial blood pressure, diabetes mellitus, BMI, physical activity level, and sign of hypertrophy on electrocardiogram (ECG). No model building procedures were used to obtain the presenting models. In backward elimination procedure, P>0.1 was used as an elimination criteria. Both full models and backward elimination models are presented.

Comparison of the heart rate at the beginning of the episodes of ST depression with the mean heart rate was based on the assumption that these values are independent. In each subjects, the mean heart rate is an average of more than 60 000 numbers, of which the heart rate at the beginning of the episodes are only a few, and thus negligible.


The basic characteristics of 678 participants of this study are shown in Table 1. SMI was more prevalent in older subjects and in subjects with higher systolic blood pressure and lower BMI, other variables were not significantly different between the groups (Table 1).

View this table:
Table 1

Baseline characteristics of study population and subjects with and without SMI

CovariatesAll (n=678)Silent myocardial ischaemia
Yes (n=77)No (n=601)P-value
Age (year)64.5±6.866.6±6.764.2±6.80.004
Women, n (%)281 (41.2%)39 (50.6%)242 (40.3%)0.08
Current smoker, n (%)314 (46.3%)40 (51.9%)274 (45.6%)0.29
Diabetes, n (%)75 (11.1%)6 (7.8%)69 (11.5%)0.33
Low level of physical activity174 (25.9%)21 (27.3%)153 (25.7%)0.76
Body mass index (kg/m2)26.1±4.224.7±4.226.3±4.20.001
Systolic blood pressure (mmHg)156.4±24.2165.8±28.3155.1±23.40.003
Diastolic blood pressure (mmHg)90.9±10.991.9±12890.7±10.70.40
Resting pulse (beat/min)72±12.371.8±12.872.7±12.30.40
Cholesterol (mmol/L)6.1±1.16.1±1.06.0±1.10.73
High density lipoprotein (mmol/L)1.5±0.461.6±0.521.5±0.450.04
Triglycerides (mmol/L)1.5±1.21.4±0.741.5±1.30.50
Blood glucose (mmol/L)5.82±1.75.84±1.75.82±1.70.90
Aspirin usage103 (15.2%)13 (16.9%)90 (15.0%)0.66
β-Blocker usage34 (5.0%)6 (7.8%)28 (4.7%)0.24
Calcium-antagonist usage56 (8.3%)6 (7.8%)50 (8.3%)0.87
ACE-inhibitor usage32 (4.7%)5 (6.5%)27 (4.5%)0.44
Diuretic usage121 (17.8%)14 (18.2%)107 (17.8%)0.94
Fibrinogen (µmol/L)11.2±2.110.9±2.011.2±2.10.29
C-Reactive protein (µg/mL)4.7±10.84.5±8.04.8±11.10.87
0–1 self-reported risk factors (%)316 (46.8%)35 (45.4%)281 (46.8%)0.90
≥2 self-reported risk factors (%)362 (53.2%)42 (54.6%)320 (53.2%)0.90

Episodes of SMI

In total, 77 participants (11.4%) had at least one episode of ST depression. In 53 (7.8%) subjects, ST depression reached a maximum of at least 2 mm. In total, 624 episodes of SMI were detected. Figure 2 (see Supplementary material online) shows the circadian variations of the episodes of SMI. The Mean heart rate at the onset of episodes was 92±19, which is significantly higher than mean heart rate 75±7 (P<0.001) for the whole observation. Heart rate during SMI episodes reached a maximum of 101±21, which is significantly higher than heart rate at onset (P<0.001) and mean heart rate (P<0.001). Number of episodes in each subjects varied from 1 to 40, with 60% having one to five episodes. Duration of SMI episodes varied from 1 to 1170 min, with a median value of 58 min.

Figure 2 Circadian variation in the episodes of SMI in apparently healthy middle-aged and elderly subjects. Number of episodes (solid line) and number of subjects with at least one episode (dashed line) of this figure.

The ‘Rose questionnaire’

None of the participants had any spontaneous complaints of chest pain or equivalent symptoms. However, when questioned according to a modified ‘Rose questionnaire’, 97 subjects (14.3%) had at some point experienced chest pain or discomfort that, to the evaluation of physician at the time of interview, had resemblance to anginal symptoms. The frequency of chest pain in subjects with and without SMI was 14.3 (11/77) and 14.3% (86/601), respectively (P=0.99). The prevalence of SMI in subjects with and without chest pain was 11.34 and 11.36%, respectively (P=0.99). In subjects with and without chest pain, the combined event percentages were 8.3 and 9.9%, respectively (P=0.56).

Baseline ECG

All subjects had sinus rhythm at baseline. Significant Q-waves (I 1–3), left bundle branch block (VII-1), and ST depression ≥1 mm (IV-1) were excluded. Non-specific T-wave changes (V 1–4) were observed in 51, and non-significant ST changes (IV 2–4) were detected in 20 subjects. First-degree atrioventricular (AV) block (VI-3) was detected in 16 subjects and short PQ (VI-5) in 1 subject. No other AV conduction abnormality was detected. ECG indications of left ventricular hypertrophy (III 1,3) were detected in 22 subjects: 7 (9.0%) in those with SMI and 15 (2.5%) in those without SMI (P=0.002).

Follow-up and event rates

The median follow-up time was 53 months (interquartile range 51–55). During this time, 66 events were recorded: 23 among 281 women (8.2%) and 43 among 397 men (10.8%), P=0.31. Events included 54 deaths, including 14 cases of documented cardiac death and 12 cases of AMI.

In 316 subjects with no or one self-reported risk factor, 8.8% had events, and in 362 subjects with two or more self-reported risk factors, 9.7% had events (P=0.75).

SMI and survival

Figure 3 (see Supplementary material online) shows the Kaplan–Meier survival plots in men and women stratified, on the basis of presence of SMI.

Figure 3 Event-free (death, myocardial infarction) survival in apparently healthy middle-aged and elderly subjects with (dashed line) and without (solid line) SMI: (A) women and (B) men.

Usual definition

In 77 subjects with the usual definition of SMI, 16 (20.7%) total and 8 (10.4%) cardiac events were registered during the follow up. In 601 subjects without SMI, 50 (8.3%) combined and 18 (3.0%) cardiac events were registered. Hazard ratio in relation to SMI for combined events was 2.7 [(1.5–4.6), P<0.001]; for cardiac events, it was 3.7 [(1.6–8.6), P=0.002]. These remained significant after correction for all selected covariates (Table 2). If subjects with sign of left ventricular hypertrophy were excluded, SMI would still be a significant predictor of both combined events. [HR: 2.2 (1.2–4.1), P=0.013] and cardiac events [HR: 3.7 (1.4–9.6), P=0.009] after correction for covariates.

View this table:
Table 2

Cox regression analysis: risk of combined and cardiac events in relation to daily-life SMI, according to the usual definition (≥1 mm of horizontal or down-sloped ST depression for 1 min) and other conventional risk factor (full model)

Combined eventsCardiac events
CovariateHazard ratio (CI)P-valueHazard ratio (CI)P-value
Age (year)1.13 (1.08–1.18)0.00011.14 (1.06–1.22)0.0002
Smoking (yes/no)2.51 (1.45–4.33)0.0012.57 (1.08–6.11)0.033
Sex2.0 (1.14–3.51)0.0153.63 (1.38–9.55)0.009
Diabetes mellitus (yes/no)1.84 (0.94–3.62)0.0763.17 (1.25–8.05)0.015
Systolic blood pressure (mmHg)1.0 (0.99–1.02)0.331.02 (0.99–1.03)0.06
Cholesterol (mmol/L)0.93 (0.73–1.19)0.580.88 (0.60–1.31)0.54
Body mass index (kg/m2)1.0 (0.94–1.07)0.941.06 (0.96–1.17)0.22
Left ventricular hypertrophy0.71 (0.17–3.07)0.650.99 (0.12–8.23)0.99
Physical activity level (high/low)0.76 (0.44–1.29)0.301.38 (0.54–3.51)0.49
SMI 1.97 (1.06–3.69) 0.033 3.1 (1.24–7.97) 0.016

Strict definition

In 53 subjects with the strict definition of SMI, 11 combined and 5 cardiac events were observed. In comparison, in 625 subjects without SMI, according to this definition, 55 total and 21 cardiac events were observed. Hazard ratio in relation to this kind of SMI for combined events was 2.7 [(1.4–5.1), P=0.003]; for cardiac events, it was 3.1 [(1.2–8.3), P=0.02]. Both of these remained significant after correction for all covariates: 2.6 (1.26–5.32), P=0.01 for combined events and 5.54 (1.82–16.83), P=0.003 for cardiac events.

In a backward elimination Cox-model with all of the mentioned covariates, the following variables were identified as predictors: SMI [HR: 1.9 (1.1–3.6), P=0.03], male sex [HR: 2.1 (1.2–3.6), P=0.008], smokers [HR: 2.4 (1.4–4.1), P=0.001], and age [HR: 1.13 (1.09–1.18), P<0.001].

There was no association between the duration of SMI (sum of the duration of all episodes of SMI in each person) and event rates: 47% of the events occurred in the subjects with SMI with a duration below median and 53% of the events in the subjects with longer duration SMI (P=0.82).

SMI and event rates in subgroups

Table 3 shows frequencies of SMI and hazard ratios for SMI in pre-defined subgroups. SMI in women was associated with significant increased event rate in univariate analysis, but not in multivariate analysis (Figure 3 and Table 3). In subjects without any history of chest pain (581), SMI was associated with increased event rates: HR: 2.42 (1.33–4.40), P=0.004 after correction for age and sex.

View this table:
Table 3

SMI and risk of death or myocardial infarction in different subgroups of middle-aged and elderly healthy subjects (corrected for age and sex)

Silent myocardial ischaemia (event rates)Hazard ratios
Yes (%)No (%)HR (CI)P-value
Women (281) (0.95–5.75)0.064
Men (397) (1.14–5.0)0.021
Smokers (314) (0.89–3.78)0.10
Non-smokers (364) (1.11–7.32)0.029
Systolic blood pressure <155 (363) (0.68–4.68)0.24
Systolic blood pressure ≥155 (315) (1.39–6.09)0.005
Total cholesterol >6 mmol/L (358) (0.70–3.87)0.26
Total cholesterol <6 mmol/L (313) (1.54–8.46)0.003
BMI≥26.4 (333) (0.91–6.6)0.08
BMI<26.4 (345) (1.14–4.58)0.024
Physically active (498) (1.02–4.64)0.045
Physically inactive (174) (1.0–6.37)0.05
Glucose intolerant (182)36.811.72.15 (0.88–5.28)0.09
Not glucose intolerant (496) (1.08–4.85)0.03
Frammingham risk-score below average (311) (1.36–8.31)0.009
Frammingham risk-score above average (359)18.410.31.55 (0.68–3.56)0.30
0–1 self-reported risk factors (316) (0.66–4.60)0.26
≥2 self-reported risk factors (362) (1.37–5.71)0.005
All (678) (1.35–4.21)0.003

Number in parenthesis indicates the total number of subjects.


Major findings

In this study, we found daily-life SMI in 11.4% of middle-aged and elderly subjects with no apparent heart disease and it was associated with a three- to five-fold increase in cardiac event rate in a mean observation period of 53 months. The increased risk remained significant after correction for all conventional risk factors. Strikingly, 30% of all cardiac events occurred in the group with SMI. Subjects with SMI and otherwise a lower Framingham risk-scorer had a two-fold increase in event rates compared with those at higher end of Framingham risk-score and without SMI.

Comparison with other studies

The prevalence of SMI in this study (11.4%) is in agreement with other studies showing a prevalence of 10–23% in comparable materials.6,10,11 Fleg6 found a 3.6-fold increase in relative risk of coronary events in apparently healthy subjects with SMI, detected with both exercise testing and thallium scintigraphy. Laukkanen et al.8 reported a two- to six-fold increase in relative risk of coronary events in men with exercise-induced SMI, in agreement with our study. As far as daily-life SMI is concerned, Hedblad et al.10 found a 4.4-fold increase in cardiac event rate in subjects with SMI in a study of 394 middle-aged men without known CAD. Otherwise, the prognostic significance of daily-life SMI has not been addressed in larger studies. To our knowledge, this is the largest Holter study in apparently healthy middle-aged and elderly men and women.

SMI and event rate in subgroups

Even though this study may not have the appropriate size to be conclusive about all subgroups, we looked at different pre-defined subgroups to evaluate the tendencies. The results of subgroup analyses, thus, should be interpreted with caution owing to the risk of both type I and type II errors.

The prognostic significance of daily-life SMI in different risk groups has not been addressed before. In a recent study, exercise-induced SMI was shown to be more strongly associated with cardiac event rate in healthy subjects with hypercholestrolaemia, hypertension, and smoking.8 As far as hypertension is concerned, our results are consistent with other studies, both regarding the high rate and significance of SMI.8,1214 In smokers, the absolute risk associated with SMI was higher, but the relative risk was lower owing to high risk in smoker without SMI (Table 3). In subjects with hypercholestrolaemia, our findings are in contrast to the earlier mentioned study. This may partially be due to use of different methods of detecting SMI (exercise testing vs. Holter monitoring) and differences in populations studied; Laukkanen et al. studied only men and with a wider age spectrum. It is known that different modalities, i.e. exercise testing, Holter monitoring, or isotope scintigraphy can detect different groups of patients with SMI.15 The prognostic significance of SMI in other subgroups showed the same tendency as for the whole population (Table 3).

Even though the absolute risk associated with SMI was high in subjects with Framingham risk-score both above and below average, the relative risk was much higher in the low-risk group. This may be due to the lower event rate in subjects with Framingham risk-score below average and no SMI. Thus, ECG monitoring can identify high risk subjects among apparently healthy subjects and also among those normally considered as belonging to the low-risk group according to Framingham scoring. Taking the number of risk factors into consideration, we may emphasize that most subjects in this study (569 subjects, 84%) had at least one risk factor. This is in agreement with the study by Laukkanen et al.,8 suggesting that SMI has significance in asymptomatic subjects with at least one risk factor. Many subjects with one risk factor may, however, have a Framingham risk-score below average.

SMI in relation to other baseline variables

The prevalence of SMI was higher in older subject, and subjects with higher blood pressure, which is in agreement with previous studies.1,8,1417 Interestingly, SMI was somewhat more prevalent in women and in subjects with lower BMI. Laukkannen et al. also found SMI associated with lower BMI, even though their method of detection was exercise testing and they studied only men. A higher HDL level in subjects with SMI may reflect the lower BMI. Otherwise, SMI was not associated with any other risk factors studied (Table 1). In a logistic multivariate analysis, high systolic blood pressure (P=0.002) and low BMI (P=0.003) were the only factors associated independently with SMI.


ST depression in asymptomatic subjects may be an indicator of coronary artery stenosis or dysfunction or a false-positive finding due to other reasons for ST-segment depression, such as hypertrophy or hypokalaemia. Although false-positive test results are high in subjects with a low pre-test probability of CAD, in middle-aged and elderly subjects, like the population of this study, the rate of false positive may be lower.18,19 About 30% of all cardiac events in this population occurred in subjects with SMI at baseline, indicating that in populations like this more than one-fourth of the cases of the manifest ischaemic heart disease start as silent ischaemia before manifestation. The poor prognosis in these subjects as demonstrated in this study and other studies may indicate that most of these subjects may have CAD. Taking the strict definition, the probability of CAD would be even higher.


To our knowledge, this is the first study to show the prognostic significance of daily-life SMI in middle-aged and elderly men and women in different risk strata. Additionally, this is the largest study of ECG monitoring in apparently healthy subjects. The prognostic burden of SMI showed the same tendency in most pre-defined subgroups inclusive subjects with a Framingham risk-score below average. From a preventive cardiology point of view, this may have potential clinical relevance. In subjects at the lower end of the Framingham risk-score, the threshold of primary preventive measures is usually high owing to relative low cost-effectiveness. Identification of high-risk subjects in this group may give them the potential benefits of primary prevention, such as aggressive lipid lowering and other risk factor modification. However, screening for SMI in the general population has not been advised mostly due to the relatively low prevalence and probability of high false-positive results.20 If screening for SMI is to be considered, it might be important to identify groups of subjects with higher prevalence of SMI and/or lower probability of false positive. The relatively high event rate associated with SMI as shown in this and other studies6,8,10,11 makes it justified that these subjects, if identified, should receive primary prevention and undergo further investigation to reduce the high risk of coronary events.


This study was performed in middle-aged and elderly caucasians. SMI in younger ages may not have the same significance. Because not all eligible subjects were able to or willing to participate, selection bias of some extent cannot be excluded. In addition, application of these data to other ethnic groups should be done with caution.


SMI as detected by ambulatory ECG monitoring is found in –11% of middle-aged and elderly subjects with no apparent heart disease and is associated with a three- to five-fold increase in the risk of coronary event rate over a 5-year follow-up period, independent of other conventional risk factors, implying that this test can be used to identify high-risk individuals among these subjects.

Supplementary material

Supplementary material is available at European Heart Journal online.


This study was supported by grants from ‘The Danish Heart Foundation’.


View Abstract