Objective: To investigate the prevalence and determinants of unrecognised myocardial infarction (UMI).
Design, setting, patients: In this cross-sectional study in a tertiary centre, a delayed enhancement cardiac MRI (DE-CMR), which identifies both Q-wave and non-Q wave MIs, was performed in 502 subjects with manifest extracardiac atherosclerotic disease or marked risk factors for atherosclerosis without symptomatic coronary artery disease.
Main outcome measures: UMI was defined as the presence of delayed enhancement without corresponding clinical history.
Results: DE-CMR was of sufficient image quality in 480 (95.6%) subjects. A UMI was present in 45 (9.4%) of all subjects; in 13.1% of men and in 3.7% of women. The risk of UMI increased from 6.0% (95% CI 2.2 to 9.8%) in those with two vascular risk factors up to 26.2% (95% CI 15.2 to 37.3%) in those with four or five risk factors. In a multivariable analysis, the risk of UMI was related to male gender (OR 2.3 (95% CI 1.0 to 5.6)), age (OR 1.04 (95% CI 1.00 to 1.07) per year), ever smoking (OR 3.1 (95% CI 1.0 to 9.1), history of stroke (OR 1.9 (95% CI 0.8 to 4.3)) and history of aneurysm of the abdominal aorta (OR 2.6 (95% CI 1.0 to 6.9)).
Conclusions: In cardiac asymptomatic subjects at high vascular risk, UMI is common. The risk of UMI increases with increasing presence of risk factors.
Statistics from Altmetric.com
Several population-based studies indicated that a considerable proportion of the acute myocardial infarctions (MIs) remain clinically unrecognised. Previous studies showed that 22–40% of all MIs occur unrecognised.1–9 Although clinically not recognised, these silent or unrecognised MIs (UMIs) have been associated with an increased mortality risk that is similar to that found in those with a clinically recognised MI (RMI).4 8–10 Most previous studies used Q-waves on the electrocardiogram (ECG) to assess the presence of UMIs. However, delayed enhancement cardiac magnetic resonance imaging (DE-CMR) with Gadolineum contrast allows for visualisation of myocardial scar tissue, and thereby has the capacity to detect previous MI.11 Because DE-CMR detects both Q-wave and non-Q-wave MIs, the sensitivity of DE-CMR for the detection of MI is in principle inherently greater than that of ECG.12 Only a few studies have used the DE-CMR approach to study UMI. The PIVUS study reported a prevalence of 19.8% UMIs in 70-year-old subjects from the general population.13 In contrast, in another study in 298 corporation employees with a mean age of 50 years without known MI, stroke, diabetes mellitus (DM) or malignancy, only one subject with a UMI was found.14 Clearly, the prevalence depends on the characteristics of the populations studied, such as age, gender and risk-factor distribution. DE-CMR studies in subjects with vascular disease in areas other than the coronary arteries and among those with vascular risk factors are lacking. Importantly, as argued previously in this journal, despite high levels of cholesterol-lowering drugs, blood-pressure-lowering drugs and antiplatelet aggregation drugs, prescribing opportunities exist for increasing the benefits of secondary prevention.15–17 In this regard, it is important to select those subjects who would benefit most from secondary prevention measures. Therefore, we set out to study the prevalence and determinants of UMI in a population with longstanding hypertension but without known symptomatic coronary artery or valvular heart disease by means of DE-CMR.
We used data from subjects enrolled in the SMART study (Second Manifestations of ARTerial disease).18 The SMART study is an ongoing prospective single-centre cohort study in subjects with manifest arterial disease or vascular risk factors. Starting in September 1996, consecutive subjects aged 18–80 years, referred to the University Medical Center Utrecht (UMCU) with manifest arterial disease or a vascular risk factor, underwent a vascular screening including a questionnaire and blood chemistry. Written informed consent was obtained from all participants. The study complies with the declaration of Helsinki and was approved by the medical ethics committee of the UMCU. The rationale and design of the SMART study have been described in detail elsewhere.18
All participants underwent a vascular screening on a single day at the UMCU. Blood samples were collected after an overnight fast. Glucose, total cholesterol, triglycerides and high-density lipoprotein-cholesterol were measured. Low-density lipoprotein cholesterol was calculated using the Friedewald formula. Height and weight were measured without shoes and heavy clothing. Blood pressure was measured in the supine position at the right brachial artery every 4 min during the arterial stiffness measurement with a semiautomatic oscillometric device (Omega 1400, Invivo Research Laboratories, Broken Arrow, Oklahoma). Medical history, use of current medication and packyears smoked were derived from a questionnaire described elsewhere.18 In addition, all participants are being followed every 6 months for the occurrence of vascular events, including myocardial infarction, stroke and vascular death.18 Hypertension was defined as systolic blood pressure (BP) ⩾140 mm Hg or diastolic BP ⩾90 mm Hg or antihypertensive treatment based on self report and BP measurements at the baseline of the SMART study. Subjects who currently used alcohol or tobacco, or had done so in the past, were classified as alcohol drinkers or ever smokers, respectively. Albuminuria was defined as a urinary albumin:creatinine ratio greater than 2.5 mg/mmol for men or greater than 3.5 mg/mmol for women.19 A history of aneurysm of the abdominal aorta (AAA) was defined as a self-reported history of AAA or the presence of AAA at ultrasonography screening at the time of entry in SMART. Peripheral artery disease (PAD) was defined as a self-reported history of PAD or the presence of PAD at screening at the time of entry in SMART.
For the present study, we selected 1273 subjects from the SMART population who reported in the questionnaire as having had hypertension for at least 3 years and who were free from previous symptomatic coronary or valvular heart disease according to the questionnaire and follow-up to undergo CMR.20 The symptomatic status was assessed by direct questioning for heart attack, Rose questionnaire,21 bypass surgery or percutaneous transluminal coronary angioplasty. Subjects with severe concomitant illness (n = 8), severe renal insufficiency (serum creatinine >160 μmol/l; n = 56), morbid obesity (BMI >40 kg/m2; n = 3) or contraindications to MRI examination (including claustrophobia; n = 34) were excluded. On the day of CMR resting BP, height and weight were recorded again. Written informed consent was obtained from all subjects for the CMR study, and the substudy was approved by the medical ethics committee of the UMCU. Five-hundred and two subjects agreed to participate.
CMR was performed using 1.5 T clinical MRI scanners (Achieva, Philips Medical Systems, Best, The Netherlands). With a steady-state free precession (SSFP) pulse sequence for optimal blood-myocardium contrast, longitudinal and, subsequently, short-axis views of the left ventricle (LV) were acquired. Scan parameters were TR/TE = 2.8/1.4 ms, RF excitation angle 55°, field of view (FOV) 350×350 mm, matrix 256×192, reconstructed resolution interpolated 1.4×1.4 mm and 8 mm slice thickness (with contiguous slices). Fifteen minutes after the injection of 30 ml Gd-DTPA contrast (Magnevist, Bayer Schering Pharma, Mijdrecht, The Netherlands), DE images were acquired along the short axis and longitudinal axis with an inversion recovery segmented k-space 3D gradient echo sequence. Scan parameters were TR/TE = 4.4/1.4 ms, RF excitation angle 15°, FOV 350×280×100 mm, matrix 256×245×10, interpolated to a resolution of 1.4×1.4×5 mm. The inversion time was individually adjusted for optimal nulling of the myocardium. Figure 1 shows a DE image.
All DE-CMR images were analysed by two experienced observers (MM and BV) independently on a PACS workstation (PACS, Philips Medical Systems, Best, The Netherlands). In case of disagreement, a third reader (EJV) was consulted. The observers were blinded for clinical history. Areas of hyperintense myocardium were classified as myocardial scar tissue. The areas were classified as subendocardial (ie, scar tissue on the endocardial side of the myocardium occupying <50% of the total myocardial thickness), transmural (ie, scar tissue on the endocardial side of the myocardium occupying >50% of the total myocardial thickness) or epicardial (ie, scar tissue on the epicardial side of the myocardium occupying <50% of the total myocardial thickness).11 Subendocardial and transmural scar tissue was considered MI. As all subjects had no known history of coronary artery disease, these MIs were considered UMI. The location of UMI was classified according to the American Heart Association segmentation system.22 As described earlier, LV mass and end-diastolic and end-systolic volumes were measured by a skilled investigator on serial short-axis views by tracing the endocardial and epicardial borders with dedicated software (ViewForum, Philips Medical Systems, Best, The Netherlands).23 Papillary muscles were included in the LV mass. Infarct size was quantified as scar mass relative to LV mass.13
The prevalence was estimated by age and gender, and presented as proportions with 95% CIs. The relation between vascular risk factors and presence of an UMI was assessed with univariable logistic regression models. In addition, we evaluated the prevalence of UMI with increasing presence of vascular risk factors (male gender, age above mean age, that is 53 years, ever smoking, history of stroke and history of AAA). A multivariable logistic regression model was used to assess the independent relationship of factors which showed a relation with p⩽0.15 in the age- (as a continuous variable) and gender-adjusted analysis to UMI.
To examine whether the presence of UMI was related to cardiac dimensions and cardiac function, we used a univariable linear regression model to assess age- and gender-adjusted differences in LV mass and LV ejection fraction (LVEF) between subjects with and without UMI. A p value less than 0.05 was considered statistically significant. A data analysis was performed using SPSS for Windows v.15 (SPSS, Chicago).
The DE-CMR was of sufficient image quality in 480 (95.6%) subjects. The interobserver agreement between MM and BV was excellent (weighted kappa = 95%) for the detection of UMI on a per-patient level. The general characteristics are listed in table 1.
In these 480 subjects, a UMI was present in 45 (9.4%, 95% CI 6.8% to 12.0%) of all subjects, that is 38 (13.1%, 95% CI 9.2% to 16.9%) of 291 men and in seven (3.7%, 95% CI 1.0% to 6.4%) of 189 women. No differences in risk-factor prevalence between subjects with and without DE-CMR of sufficient image quality were observed (data not shown). Overall, subjects who did not agree to participate in the present study had a higher BP, were more frequently diabetics and had less often a vascular history than subjects who did agree to participate (data not shown). On average, two (range 0 to 11) LV segments were affected by DE. Of the UMIs, 54% were present in the lateral segments. Forty per cent of the UMIs were subendocardial, and 60% were transmural. The relative scar size was 4.4 (SD 3.9) g of scar per 100 g of LV myocardium. The prevalence of UMI increased with age, as shown in fig 2.
In the analysis adjusted for age and gender, the risk of UMI was related to male gender, increasing age and ever smoking. Relations were of borderline significance for a history of stroke and a history of AAA. A statistically weaker relation was observed for a higher prevalence of UMI and antiplatelet treatment, and for a higher risk of UMI and diabetes. (table 2). The multivariable analysis showed similar results, except that antiplatelet treatment did not have a relation to the risk of UMI in the multivariable model.
The prevalence of UMI increased with the number of risk factors (table 3) from 6.0% (95% CI 2.2 to 9.8) in subjects with two risk factors to 26.2% (95% CI 15.2 to 37.3) in subjects with four or five risk factors. Infarct size was not related to the number of risk factors.
Adjusted for age and gender, subjects with UMI had a significantly higher LV mass index (59.0 (13.4) g/m2 vs 53.0 (12.0) g/m2, p <0.035) and a significantly lower LVEF (57.1 (11.7)% vs 62.6 (7.7)%, p <0.001) than those without UMI.
This DE-CMR based study shows that in high-risk cardiac asymptomatic subjects, UMI is a common finding. The prevalence of UMI was higher in subjects with more vascular risk factors.
In this cross-sectional study, we report a prevalence of UMI of 9.4% in all subjects combined, and of 13.1% in men and of 3.7% in women. Earlier studies have used ECG to detect UMI. In the Vascular Health Study, a population-based study in subjects over 65 years of age, 3.4% of subjects had a UMI.7 In a study in inhabitants of Reykjavik, Iceland, the prevalence of UMI increased from approximately 0% in men under age 40 to 5% in men aged 70–75 years.8 In the same study, the prevalence of UMI was almost 0% in women aged 35, increasing to 1.9% in women aged 75 years.3 In a study in inhabitants older than 75 years of the Bronx, New York City (USA), 6.4% had UMI.24 In the Rotterdam study in 3272 men and women aged 55 years or older, the overall prevalence of UMI was 4.6% in men and 3.5% in women. In the age range of 55 to 64 years, the prevalence of UMI was 2.6% and 2.0% for men and women, respectively, increasing to 7.5% and 9.8% for men and women, respectively, over 85 years of age.25 As mentioned above, more recent studies have used DE-CMR. In the PIVUS (Prospective Investigation of the Vasculature in Uppsala Seniors) study, a prevalence of UMI 19.8% was reported in 70-year-old subjects from the general population.13 The prevalence reported in the PIVUS study is higher than the prevalence reported for our population. Importantly, in the PIVUS population, 13% of subjects without UMI had a history of chest-pain symptoms, and 3.7% had angina, whereas none of our subjects had a history of symptomatic coronary artery disease. In contrast to the high prevalence reported in the PIVUS study, in a different study in 298 corporation employees with a mean age of 50 years without known MI, stroke, DM or malignancy, only one subject with an UMI was found.14
Others have used echocardiography or single photon emission computed tomography (SPECT) to investigate the prevalence of UMI. In patients undergoing vascular surgery, the prevalence of UMI as detected by dobutamine stress echo was 23%.26 In patients with diabetes mellitus at high cardiovascular risk, the prevalence of UMI, detected as irreversible perfusion abnormalities on SPECT, was 4%.27
Since the prevalence of UMI depends on the age, gender and vascular risk profile of the study subjects, an accurate comparison of overall UMI prevalence rates between studies cannot be made. However, the prevalence of UMI as reported by DE-CMR in our study and the PIVUS study seems higher than in ECG-based UMI prevalence studies. The higher sensitivity of DE-CMR compared with ECG for the detection of UMI might, at least in part, explain this.12
The prognosis of UMIs detected by ECG has been reported to be similar to that of recognised myocardial infarction.4 8–10 Although it seems likely that the prognosis of UMIs detected by DE-CMR will also be similar to that of RMI, to our knowledge, prospective cohort studies to assess the relation between UMI detected by DE-CMR in cardiac asymptomatic subjects and future events are lacking. There is some evidence to support the prognostic value of UMI detected on DE-CMR. Small areas of DE (less than 10 g) were reported to be prognostically significant in subjects presenting to the emergency department with chest pain.28 Also, the presence and extent of DE were strong predictors of major adverse cardiac events and death in subjects suspected of coronary artery disease without a history of MI.29 In addition, indirect evidence from the present study may indicate that UMI is associated with a poor vascular prognosis, that is, cardiac function, as indicated by LVEF, was significantly lower in subjects with UMI, and LV mass was significantly higher in subjects with UMI. Increased LV mass is associated with an increased vascular morbidity and mortality,30 and an increase in LV mass is associated with an increased oxygen demand, making the LV more vulnerable to MI. These findings are in line with results from the PIVUS study.13 In contrast, the PIVUS study also showed that the prevalence of whole-body atherosclerosis measured by magnetic resonance angiography did not differ between subjects without scar and those with UMI on DE-CMR, while in those with recognised myocardial infarction (RMI), more atherosclerosis was present.31 This finding raised the possibility that UMIs detected by DE-CMR may not always be related to atherosclerosis.
If UMI detected by DE-CMR is shown to be of prognostic importance, detection of subjects with UMI may be important, especially when evidence would show that subjects with UMI detected by DE-CMR benefit similarly from secondary prevention measures as patients with a RMI.32 However, evidence on this subject is still insufficient. As mentioned above, opportunities exist for increasing the benefits of secondary prevention drug therapy.15–17 If future evidence would demonstrate that screening for UMI would prognostically benefit the individual patient, the benefits of the various detection methods should be compared. While the ECG can detect Q-wave MIs at low cost, DE-CMR and SPECT additionally detect non-Q wave MIs. As the spatial resolution of DE-CMR is higher than that of SPECT, DE-CMR may prove to have a higher sensitivity for the detection of UMI than SPECT. In addition, stress perfusion DE-CMR and SPECT can detect reversible perfusion abnormalities, which may be additional valuable prognostic information.
We acknowledge the limitations of the study. First, CMR scans were made at different time intervals after inclusion in SMART. Since the development of UMI under the influence of risk factors is time-dependent, these different time intervals between inclusion and CMR might have led to an overestimation of the prevalence of UMIs. Importantly, as part of the follow-up after inclusion in SMART, all participants completed a questionnaire on vascular events, including myocardial infarctions, every 6 months. Second, DE-CMR was of insufficient image quality in 4.4% of subjects. This percentage is comparable with previous reports.13 Importantly, no differences in risk-factor prevalence between subjects with and without DE-CMR of sufficient image quality were observed, therefore suggesting that our prevalence estimate is valid. Third, an ECG was only available for the time at inclusion in SMART. Therefore, we could not compare the sensitivity of ECG and CMR for the detection of UMI. However, the validity of the presented results on the prevalence and determinants of UMI is not affected. Fourth, we observed differences in vascular risk-factor profile and history between subjects who agreed and did not agree to participate in the study. Overall, the former subjects had a higher BP, were more frequently diabetics and had less often a vascular history. Therefore, we might have underestimated the prevalence of UMI in the total study population.
In conclusion, in high-risk cardiac asymptomatic subjects, UMI is common. The risk of UMI increases with increasing presence of risk factors.
The authors acknowledge all MRI technicians, research nurses and medical students involved in SMART Heart for valuable support.
Funding: The SMART study was financially supported by a grant of the University Medical Center Utrecht. MFLM was financially supported by EUGene.
Competing interests: None.
Ethics approval: Ethics approval was provided by the medical ethics committee of the UMCU.
Patient consent: Obtained.
SMART Study Group: A Algra, Julius Center for Health Sciences and Primary Care and Rudolph Magnus Institute of Neuroscience, Department of Neurology; Y van der Graaf, GEHM Rutten, DE Grobbee, Julius Center for Health Sciences and Primary Care; FLJ Visseren, Department of Internal Medicine; PA Doevendans, Department of Cardiology; FL Moll, Department of Vascular Surgery; LJ Kappelle, Department of Neurology; WPThM Mali, Department of Radiology.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.