Author + information
- Received November 29, 2012
- Revision received January 24, 2013
- Accepted February 5, 2013
- Published online May 14, 2013.
- Yasushi Matsuzawa, MD⁎,†,
- Masaaki Konishi, MD, PhD⁎,†,
- Eiichi Akiyama, MD⁎,†,
- Hiroyuki Suzuki, MD⁎,
- Naoki Nakayama, MD, PhD⁎,†,
- Masayoshi Kiyokuni, MD, PhD⁎,
- Shinichi Sumita, MD, PhD⁎,
- Toshiaki Ebina, MD, PhD⁎,
- Masami Kosuge, MD, PhD⁎,
- Kiyoshi Hibi, MD, PhD⁎,
- Kengo Tsukahara, MD, PhD⁎,
- Noriaki Iwahashi, MD, PhD⁎,
- Mitsuaki Endo, MD, PhD⁎,
- Nobuhiko Maejima, MD⁎,
- Kenichiro Saka, MD⁎,
- Katsutaka Hashiba, MD⁎,
- Kozo Okada, MD⁎,
- Masataka Taguri, PhD‡,
- Satoshi Morita, PhD‡,
- Seigo Sugiyama, MD, PhD†,
- Hisao Ogawa, MD, PhD†,
- Hironobu Sashika, MD, PhD§,
- Satoshi Umemura, MD, PhD∥ and
- Kazuo Kimura, MD, PhD⁎,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Kazuo Kimura, Division of Cardiology, Yokohama City University Medical Center, 4-57 Urafune-cho, Minami-ku, Yokohama, 232-0024, Japan
Objectives This study sought to determine the additional clinical value of gait speed to Framingham risk score (FRS), cardiac function, and comorbid conditions in predicting cardiovascular events in patients with ST-segment elevation myocardial infarction.
Background There is growing evidence that gait speed is inversely associated with all-cause mortality, particularly cardiovascular mortality, among the elderly.
Methods We undertook a single-center prospective observational study of gait speed in 472 patients with ST-segment elevation myocardial infarction in Japan, between 2001 and 2008. Gait speeds were measured using a 200-m course before discharge in all patients, and we followed up cardiovascular events, which consist of cardiovascular deaths, nonfatal myocardial infarctions, and nonfatal ischemic strokes.
Results During the 2,596 person-years of follow-up, 83 patients (17.6%) experienced cardiovascular events. Cardiovascular events increased across decreasing tertiles of gait speed (fastest tertile: n = 5, 3.2%; middle tertile: n = 20, 12.6%; slowest tertile, n = 58, 36.7%). By multiple adjusted Cox proportional hazards analysis, gait speed was a significant and independent predictor of cardiovascular events (hazard ratio for increasing 0.1 m/s of gait speed: 0.71, 95% confidence interval [CI]: 0.63 to 0.81, p < 0.001). The addition of gait speed to the model incorporating FRS, B-type natriuretic peptide levels, and comorbidity index improved reclassification (net reclassification index: 32.8%, 95% CI: 17.4 to 48.3, p < 0.001) and the C-statistics with a reasonable global fit and calibration (C-statistics: from 0.703 [95% CI: 0.636 to 0.763] to 0.786 [95% CI: 0.738 to 0.829]).
Conclusions Among patients with ST-segment elevation myocardial infarction, slow gait speed was significantly associated with an increased risk of cardiovascular events. (Gait Speed for Predicting Cardiovascular Events After Myocardial Infarction; NCT01484158)
Acute myocardial infarction is one of the leading causes of mortality, and the risk of further cardiovascular complications, including recurrent myocardial infarction, sudden cardiac death, heart failure, and stroke, for those who survive acute myocardial infarction is substantial even in the era of optimal reperfusion strategies (1,2). Proper risk stratification for second cardiovascular events at early stage after acute myocardial infarction is the foundation of current cardiovascular practices.
There is growing interest in using gait speed to assess the functional status and motor performance of older people. Because gait speed is easily determined, it may replace more complex physical performance tests. Recently, gait speed has been proposed as a new useful “vital sign” (3). Decreased gait speed is associated with increased all-cause mortality (4,5), especially cardiovascular mortality in the elderly (6). Atherosclerotic risk factors may also be correlated with decreased motor performances (7,8). However, it is unknown whether gait speed is associated with the long-term prognosis in ST-segment elevation myocardial infarction (STEMI) patients independent of traditional risk factors, cardiac function, and comorbid conditions. Identifying the subset of patients after STEMI who have poor physiological reserves, as assessed by gait speed, may help cardiologists better predict outcomes and provide more accurate risk assessments to their patients.
The purpose of this study is to evaluate the prognostic value of gait speed on long-term outcomes and to determine the degree to which gait speed affects the variability in survival after accounting for traditional cardiovascular risk factors, cardiac function, and comorbid conditions in STEMI patients successfully treated with percutaneous coronary intervention (PCI).
Between October 2001 and October 2008, 513 consecutive patients from the Yokohama City University Medical Center were recruited for this study; these patients had undergone successful primary PCI for their first STEMI within 12 h after symptom onset. The diagnosis of STEMI included continuous chest pain lasting >20 min, the presence of ST-segment elevation ≥0.1 mV in ≥2 contiguous leads on the electrocardiogram, and increased creatine kinase (more than twofold the upper limit of the hospital's reference range). We excluded 32 patients who did not complete the 200-m walk test due to fatigue or symptoms. The survival status and follow-up could not be obtained for 9 patients. Thus, 472 patients were included in the final analysis.
All STEMI patients who were capable of walking were routinely included to participate in the cardiac rehabilitation program during hospitalization according to the Japanese Circulation Society guidelines for rehabilitation in patients with cardiovascular disease. Exercise sessions during hospitalization were performed under the supervision of a physical therapist with the standardized protocol, and the intensity was determined by the Borg scale (11 to 13). Each session lasted approximately 1 h, beginning with a warm-up phase of stretching for 15 to 20 min, followed by 20 to 30 min of aerobic activity using either walking in a hallway or walking on a treadmill, and 10 min of cool down. Instruction about the exercise therapy was done by an attending physician at an individual, 30-min session during hospitalization, and exercise therapy was the patient's responsibility after discharge. Exercise intensity was determined by Karvonen methods (Karvonen method: target heart rate = [(220 − age) − rest heart rate] · 0.5 + rest heart rate]) or Borg scale (11 to 13). Exercise frequency was recommended at least 30 min per day, 4 days per week.
The present study was approved by the Yokohama City University institutional review board and was conducted in accordance with the guidelines of the ethics committee of our institution. Written informed consent was obtained from each patient before participation.
Blood samples were obtained on admission at 3-h intervals during the first 24 h, daily for the next 5 days, and then every other day until discharge. Peak levels of creatine kinase and creatine kinase-myocardial band were measured. The estimated glomerular filtration rate (eGFR) was determined from the creatinine level at the time of discharge using the prediction equation proposed by the Japanese Society of Nephrology (9). We calculated the Framingham risk score (FRS) for secondary prevention in these study patients (10). We used B-type natriuretic peptide (BNP) levels at discharge (11).
We assessed comorbid conditions using the Charlson comorbidity index, which covers 19 major disease categories including diabetes mellitus, congestive heart failure, cerebrovascular diseases, and cancer (12). The Charlson comorbidity index has been reported to be a strong prognostic factor for short-term and long-term mortality in patients with a first-time hospitalization for acute myocardial infarction (13).
Before discharge, patients took a 200-m walk, and walking time was measured using a standard digital stopwatch. All 200-m walk tests were performed by physical therapists with the standardized instructions in a wide hallway on a 50-m course in the daytime (10:00 am to 12:00 am and 3:00 pm to 5:00 pm), and patients were asked to walk at their usual pace without overexerting themselves. Patients did not eat a meal or exercise vigorously within 2 h of beginning the test. Patients were permitted to use an aid such as a cane or a walker as recommended by American Thoracic Society guideline for a 6-min walk test (14).
Follow-up and cardiovascular events
Cardiovascular-disease outcomes were evaluated for an average follow-up of 5.5 years. The primary endpoint was the incidence of cardiovascular events during the follow-up period: cardiovascular death, nonfatal myocardial infarction, or ischemic stroke. Two telephone interviewers called the patients or their families to inquire about all hospital admissions, cardiovascular outpatient diagnosis, and death. To verify diagnosis, 3 physicians on the events committee independently reviewed all medical records and death certificates for blinded endpoint classification and assignment of incidence dates using pre-specified criteria. If the reviewing physicians disagreed on the event classification, they adjudicated differences. Myocardial infarction was defined by combinations of symptoms, abnormalities on the electrocardiogram, and cardiac biomarker elevation. Ischemic stroke required documented focal neurologic deficit and clinically relevant brain imaging of infarction. Cardiovascular death was defined as a death due to myocardial infarction, congestive heart failure, or documented sudden death without apparent noncardiovascular causes. For patients experiencing >1 acute event, only the first event was considered in the analysis. To avoid observer bias, the physician ascertaining outcomes from the medical records was blinded to the gait speed data. The secondary endpoint was all-cause death.
Descriptive statistics were analyzed to describe the baseline characteristics according to cardiovascular events and gait speed tertiles. Continuous variables with normal distributions were expressed as mean (SD), and data with skewed distributions were expressed as median (interquartile range). The differences among continuous variables with normal distributions were analyzed using unpaired Student t tests or 1-way analysis of variance. Mann-Whitney U test and Kruskal-Wallis test were used for continuous variables with skewed distributions. We compared the groups using the chi-square test. Because gait speed was strongly related to sex in previous reports (5,6), we decided to categorize patients into sex-dependent groups by gait speed to maintain balanced sample sizes of men and women before data collection and examination. We used the following cut-off values (tertiles) derived from patients in this study: ≤0.926 m/s, 0.926 to 1.070 m/s, and ≥1.070 m/s for men; and ≤0.689 m/s, 0.689 to 0.916 m/s, and ≥0.916 m/s for women. We used sex-dependent groups by gait speed in Tables 1 and 2 and Figure 1, and we treated gait speed as a sex-independent continuous variable for other analyses.
We calculated the cumulative incidence of cardiovascular events according to thirds of gait speed using the Kaplan-Meier method and compared it with the log-rank test. We used Cox proportional hazards models to estimate hazard ratios (HRs) for cardiovascular events and 95% confidence intervals (CIs) using univariate model and multivariate analyses with the forward entry algorithm. We tested interactions between variables by including interaction terms. Models were initially adjusted for age and sex. We then included additional covariates (height, weight, current smoking status, hypertension, diabetes mellitus, dyslipidemia, BNP, FRS, comorbidity index, eGFR, Killip class, left ventricular ejection fraction, angiotensin-converting enzyme-inhibitors [ACE-I]/angiotensin II receptor blockers [ARB], hydroxymethylglutaryl-CoA reductase inhibitors, use of cane or walker, days in bed, days from admission to measurement of gait speed, and number of times of rehabilitation) in the models. We conducted post-hoc analyses on subgroups stratified by age and sex to assess hazard ratio of gait speed for cardiovascular events.
We confirmed the proportional hazards assumption using Schoenfeld's test and calculated estimates of the C-statistic for Cox proportional hazards regression models (15–17). We also examined whether the addition of various combinations (FRS, BNP, comorbidity index, and gait speed) improved the discriminatory power of the model. Using the bootstrap percentile method (18), 95% CIs of the C-statistics and the increment of the C-statistics were estimated.
The calibration of the Cox regression models was assessed by the Grønnesby and Borgan calibration test (19), which compares the actual and expected number of events within 5 groups that have been developed by partitioning the data based on the estimated risk score (x'β). We performed likelihood ratio tests to evaluate whether the global model fit improved after the addition of gait speed. We also evaluated the incremental effect of adding gait speed to the FRS, BNP, and comorbidity index for predicting cardiovascular events with the use of the net reclassification index (20). For the assessment of reclassification improvement, we defined 3 risk categories on the basis of the FRS plus BNP plus comorbidity index: low risk <10%, intermediate risk 10% to 20%, and high risk >20%.
Statistical significance was defined as p < 0.05 with 2-tailed tests. All analyses were performed using PASW 18 for Windows (SPSS Inc., Tokyo, Japan) and SAS program for Windows, release 9.2 (SAS Institute Inc., Cary, North Carolina).
During a mean follow-up of 5.5 ± 2.2 years, corresponding to 2,596 person-years, 83 (17.6%) patients experienced cardiovascular events (n = 34, cardiovascular death; n = 24, nonfatal myocardial infarction; n = 25, nonfatal ischemic stroke). Noncardiovascular death was observed in 30 patients. Among 32 patients who were excluded because of an incomplete 200-m walk test, 18 (56.3%) had cardiovascular events and 19 (59.4%) had all-cause deaths (n = 16, cardiovascular death; n = 3, noncardiovascular death; n = 1, nonfatal myocardial infarction, and n = 1, nonfatal ischemic stroke).
Table 1 describes the baseline clinical characteristics by cardiovascular events status and gait speed tertiles. Patients who had cardiovascular events during follow-up were significantly older and had a lower height, a lower body mass index, lower eGFR, higher BNP, lower left ventricular ejection fraction, and higher comorbidity index and were more likely to have diabetes mellitus, FRS ≥12, and Killip classification ≥2. The percentage of smokers was lower among patients with cardiovascular events than among patients without events. There were no significant differences in peak creatine kinase level, anterior myocardial infarction, cardiogenic shock, stent implantation, multivessel coronary artery disease, and medications between the 2 groups according to cardiovascular events. Gait speed was significantly slower in patients with cardiovascular events, and slow gait speed was correlated with old age, lower height, lower body mass index, lower eGFR, higher BNP, Killip classification ≥2, lower left ventricular ejection fraction, multivessel coronary artery disease, higher FRS, and higher comorbidity index. Delayed timing and less frequency of rehabilitation were associated with both higher rate of cardiovascular events and slower gait speed.
Cardiovascular events increased across decreasing tertiles of gait speed (fastest, n = 5, 3.2% [95% CI: 0.4% to 6.0%]; middle, n = 20, 12.6% [95% CI: 7.4% to 17.7%]; slowest, n = 58, 36.7% [95% CI: 29.2% to 44.2%]; p < 0.001) (Table 2). In separate analyses for each component of cardiovascular events, the increased risk of cardiovascular death, nonfatal myocardial infarction, and nonfatal ischemic stroke was observed with slower gait speed in patients with STEMI (Table 2). Figure 1 depicts the cumulative risk of cardiovascular events by gait speed tertiles. There was a significant difference in cardiovascular events across stratified gait speeds (log rank, p < 0.001) (Fig. 1).
Cox proportional hazards analysis for the prediction of cardiovascular events
Multivariate Cox proportional hazards analysis with forward entry algorithm revealed that a higher level of BNP, discharge without ACE-I/ARB, a higher comorbidity index, a longer stay in bed, less frequency of rehabilitation, and a slower gait speed were independent predictors of future cardiovascular events in STEMI patients (Table 3). In post-hoc analyses, we investigated the prognostic value of gait speed by further stratifying groups by tertiles of age and sex. In patients of the youngest age group, slow gait speed had no significant prognostic value (HR for increasing 0.1 m/s of gait speed: 0.79, 95% CI: 0.61 to 1.04, p = 0.09) (Fig. 2A). In the 2 older groups, gait speed provided significant prognostic information for cardiovascular events (middle age group HR for increasing 0.1 m/s of gait speed: 0.71, 95% CI: 0.59 to 0.84, p < 0.001; oldest age group HR for increasing 0.1 m/s of gait speed: 0.73, 95% CI: 0.64 to 0.83, p < 0.001). The prognostic value of gait speed was significant both in men (HR for increasing 0.1 m/s of gait speed: 0.68, 95% CI: 0.62 to 0.76, p < 0.001) and in women (HR for increasing 0.1 m/s of gait speed: 0.65, 95% CI: 0.52 to 0.81, p < 0.001) (Fig. 2B).
Multiple adjustment models for the prediction of cardiovascular events, cardiovascular death, and all-cause death
In our multiple adjustment model (see Methods), the association between gait speed and cardiovascular events remained significant (HR for increasing 0.1 m/s of gait speed: 0.71, 95% CI: 0.63 to 0.81, p < 0.001) (Table 4). Similarly, gait speed was significantly associated with cardiovascular deaths (HR for increasing 0.1 m/s of gait speed: 0.67, 95% CI: 0.55 to 0.81, p < 0.001) and all-cause deaths (HR for increasing 0.1 m/s of gait speed: 0.71, 95% CI: 0.62 to 0.82, p < 0.001) in the multiple adjustment models (Table 4).
C-statistics for Cox proportional hazards models and net re-classification index to predict cardiovascular events
The FRS was a significant predictor of cardiovascular events (HR for FRS ≥12: 1.66, 95% CI: 1.07 to 2.57, p = 0.02). First, we estimated the C-statistic of the FRS alone (C-statistic: 0.595, 95% CI: 0.528 to 0.669). The incorporation of BNP, comorbidity index, or gait speed into the FRS produced substantial gains in the prognostic information for cardiovascular events. The addition of gait speed to the FRS, BNP, and comorbidity index resulted in a significant increase in the C-statistic from 0.703 (95% CI: 0.636 to 0.763) to 0.786 (95% CI: 0.738 to 0.829) (Table 5). We confirmed the proportional hazards assumptions using the Schoenfeld test (p = 0.61) and confirmed the calibration for the model including all predictors using Grønnesby and Borgan statistics (p = 0.14) (19). The addition of gait speed to the model with FRS, BNP, and the comorbidity index showed a better global fit compared with the model without gait speed, as evaluated by the likelihood ratio test (p < 0.001). We examined the effect modification of the interactions and found that only the comorbidity index had significant interaction with gait speed (p = 0.04).
We treated BNP, comorbidity index, and gait speed as continuous and reclassified risk scores for the study patients. The resultant net reclassification index was significant by adding gait speed to the FRS, BNP, and comorbidity index (index, 4.8% for patients with cardiovascular events, 28% for patients without cardiovascular events, and 32.8% overall [95% CI: 17.4 to 48.3, p < 0.001]) (Table 6).
Slow gait speed was strongly associated with future cardiovascular events in STEMI patients who underwent successful primary PCI. The present study also revealed that the addition of gait speed to FRS, BNP, and comorbidity index, improved the risk stratification as evidenced by a substantial increase in C-statistics and net reclassification index. The model with gait speed showed good global model fit and calibration. These findings indicated that the clinical assessment of gait speed could identify the subsets of patients at a higher risk for cardiovascular events after STEMI.
Although cardiovascular medicine has advanced, the available therapies do not sufficiently minimize the risk of cardiovascular events. Patients who have a history of myocardial infarction are at high risk for future cardiovascular events despite optimal reperfusion therapy (2). Many factors can influence the prognosis of STEMI patients, including age, sex, cardiac function, comorbidity, atherosclerotic burden, and inflammation. There are currently no well-established approaches to predicting life expectancy that incorporate health and function; therefore, the introduction of a proper risk stratification strategy for long-term mortality is necessary. Past reports and our results demonstrated an increased cardiovascular mortality in slow walkers (6). Gait speed, a simple and informative parameter, is recommended as a potentially useful clinical indicator of well-being to refine survival estimates in clinical practice and research.
There are several possible reasons for the association of gait speed with cardiovascular events. Low cardiovascular fitness is a possible mechanism of the association between increased cardiovascular events and slow gait speed. Walking places demands on multiple organ systems, including the heart, lungs, blood vessels, and nervous and musculoskeletal systems, and requires energy. A slowing gait may reflect both a high-energy cost of walking and decreased organ system functions. The presence of comorbidities and frailty could also cause a slowing gait. The assessment of gait speed could be considered a simple and comprehensive parameter because it integrates known and unknown disturbances in multiple organ systems.
Smokers have been shown to have lower mortality after acute coronary syndrome than nonsmokers (21). That has been attributed to the younger age, lower comorbidity, more aggressive treatment, and lower risk profile of the smoker (21). In this study, smokers were significantly younger (age 60.2 ± 10.8 years vs. 68.9 ± 11.7 years, p < 0.001), and had lower comorbidity (2.2 ± 1.8 vs. 2.7 ± 1.2, p < 0.001) and faster gait speed (0.982 ± 0.193 m/s vs. 0.864 ± 0.227 m/s, p < 0.001) than nonsmokers. Although smokers had fewer cardiovascular events than nonsmokers in this study (14.3% vs. 24.2%, p = 0.001, by log rank), smoking was not significant associated factor of cardiovascular events after adjustment of age, risk factors, and comorbidity index (HR: 0.76, 95% CI: 0.46 to 1.24, p = 0.27).
Regarding medication at discharge, patients in the slowest gait speed tertile were less frequently prescribed hydroxymethylglutaryl-CoA reductase inhibitors and ACE-I/ARB. Patients without hydroxymethylglutaryl-CoA reductase inhibitors at discharge were significantly older (65.5 ± 12.0 years vs. 61.9 ± 11.6 years, p = 0.002) and had lower low-density lipoprotein cholesterol (124.8 ± 36.5 mg/dl vs. 140.6 ± 36.4 mg/dl, p < 0.001) than patients with hydroxymethylglutaryl-CoA reductase inhibitors. Patients without ACE-I/ARB at discharge were significantly older (68.5 ± 12.9 years vs. 62.2 ± 11.4 years, p < 0.001) and had lower eGFR (55.5 ± 19.9 ml/min/1.73m2 vs. 61.1 ± 15.0 ml/min/1.73m2, p = 0.007) than patients with ACE-I/ARB. These imbalances of medications could be caused by higher age, lower cholesterol, and lower renal function and could confound the relationship between gait speeds and cardiovascular outcomes. However, in the multiple adjusted model, including hydroxymethylglutaryl-CoA reductase inhibitors and ACE-I/ARB, the gait speed was still a significant predictor of cardiovascular events (Table 4).
In patients of the youngest age group, slow gait speed did not have significant prognostic value by subgroup analysis. However, a borderline protective effect of higher gait speed in the youngest tertile (HR: 0.79, p = 0.09) may not have reached significance because of low statistical power rather than a true lack of effect.
The strengths of this study include the low proportion of patients who were lost to follow-up (9 of 481). Nine patients were lost because of change of address (n = 3), a quarrel with the hospital (n = 2), and no information about whereabouts (n = 4). Of these 9 patients, 4 patients were categorized as fastest gait speed group, 2 patients were categorized as middle gait speed group, and 3 patients were categorized as slowest gait speed group. There was neither cardiovascular event nor noncardiac death until loss of follow-up (mean follow-up duration 734 days) in these 9 patients. Even if cardiovascular events occurred in the 4 patients of the fastest gait speed group, the HR of gait speed for cardiovascular events after adjustment of all covariates was still significant (HR per 0.1 m/s of gait speed: 0.73, 95% CI: 0.65 to 0.84, p < 0.001).
Our study also has several limitations that may limit generalizability of the results in this study. This is a longitudinal study and cannot assess causality and change of walking speed. This study was conducted with a relatively small number of patients at a single center. Moreover, multivariate analysis may mitigate bias after adjustment of confounding factors, but unmeasured factors (such as dose of rehabilitation), which may have an impact on gait speed, leave room for residual bias. Long-distance walking tests may provide information beyond short-distance walking tests. However, in many clinical settings, the longer distance to perform the test may limit feasibility (22,23). The most common distances used in clinical research fall within the 4-m to 10-m range (24). Further studies are needed to confirm the importance of gait speed, measured with short walking tests, in the cardiovascular prognosis of STEMI patients. Because we assessed exercise capacity in only a small number of patients in this study, we could not compare gait speed with exercise capacity in predicting cardiovascular events. Last, since neither routine magnetic resonance imaging nor ankle brachial index was performed in this study, we probably under-estimated subclinical cerebrovascular diseases, cognitive impairment, cerebral white matter hyperintensity, and peripheral arterial diseases, which are associated with atherosclerosis and could cause slower gait speed (25–28).
Our study indicates that slow gait speed strongly and independently associates with the long-term risk of cardiovascular events in patients after STEMI. The measurement of gait speed is a simple and effective risk assessment in secondary prevention strategies.
The authors have reported they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- angiotensin-converting enzyme inhibitor
- angiotensin-II receptor blocker
- B-type natriuretic peptide
- confidence interval
- estimated glomerular filtration rates
- Framingham risk score
- hazard ratio
- percutaneous coronary intervention
- ST-segment elevation myocardial infarction
- Received November 29, 2012.
- Revision received January 24, 2013.
- Accepted February 5, 2013.
- American College of Cardiology Foundation
- Roger V.L.,
- Go A.S.,
- Lloyd-Jones D.M.,
- et al.
- Ostir G.V.,
- Kuo Y.F.,
- Berges I.M.,
- Markides K.S.,
- Ottenbacher K.J.
- Dumurgier J.,
- Elbaz A.,
- Ducimetiere P.,
- Tavernier B.,
- Alperovitch A.,
- Tzourio C.
- Liston R.,
- Mickelborough J.,
- Bene J.,
- Tallis R.
- Elbaz A.,
- Ripert M.,
- Tavernier B.,
- et al.
- Tapanainen J.M.,
- Lindgren K.S.,
- Makikallio T.H.,
- Vuolteenaho O.,
- Leppaluoto J.,
- Huikuri H.V.
- Schmidt M.,
- Jacobsen J.B.,
- Lash T.L.,
- Botker H.E.,
- Sorensen H.T.
- Nozaki T.,
- Sugiyama S.,
- Koga H.,
- et al.
- Cook N.R.
- Longstreth W.T. Jr..,
- Manolio T.A.,
- Arnold A.,
- et al.