496 Acta Cardiol 2014; 69(5): 496-502 doi: 10.2143/AC.69.5.3044876

Address for correspondence :

Johan Bennett, MB, BCh, MD,Dept. of Cardiovascular Medicine, University Hospitals Leuven,Herestraat 49, B-3000 Leuven, Belgium.E-mail: johan.bennett@uzleuven.be

Received 5 May 2014; accepted for publication 27 May 2014.

INTRODUCTION

Transient left ventricular ballooning syndrome (TLVBS), also known as takotsubo or stress-induced car-diomyopathy, is a distinctive reversible condition charac-terized by acute transient left ventricular dysfunction, typically affecting postmenopausal women and frequently

Left ventricular function and clinical outcome in transient left ventricular ballooning syndrome

Johan BENNETT1, MD; Bert FERDINANDE1, MD; Peter KAYAERT1, MD; Stefanus WIYONO1, MD; Dries DE COCK1, MD; Christophe DUBOIS1,2, MD PhD; Peter SINNAEVE1,2, MD, PhD; Tom ADRIAENSSENS1, MD, PhD; Kaatje GOETSCHALCKX1, MD; Walter DESMET1,2, MD, PhD1Dept. of Cardiovascular Medicine, University Hospitals Leuven, Leuven, Belgium; 2Dept. of Cardiovascular Sciences, Catholic University of Leuven, Leuven, Belgium.

Background It is unknown if the severity of left ventricular dysfunction in patients with transient left ventricular ballooning syndrome (TLVBS) adversely aff ects clinical outcome. Furthermore, it remains unclear if the patterns of ventricular involvement are distinct patterns or if they represent varying stages of ventricular involvement.

Methods and results All patients with TLVBS who presented to our hospital from August 1998 to August 2012 were prospectively identifi ed and entered into a clinical database. Available ventriculograms were reviewed, the ejection fraction (EF) calculated and a new severity score of left ven-tricular (LV) involvement was developed to determine the degree of LV dysfunction. The incidence of in-hospital mortality, cardiogenic shock and major cardiac events (heart failure/pulmonary oedema or major cardiac arrhythmia) was recorded. In total, 145 TLVBS episodes were identifi ed in 139 patients. Age at presentation was 67 ± 12 years and 89% (n = 123) of patients were female. Patients who developed cardiogenic shock or other acute cardiac events had a worse LVEF compared to those who did not (P < 0.01 and P = 0.05, respectively). In-hospital mortality was not related to worse EF (P = 0.58). In-hospital and 1-year mortality rates were 6.9% and 12.6%, respectively. Median time from symptom onset to clinical diagnosis was similar between the apical ballooning (n = 104; 12 [3-30] hours) and the mid-ventricular ballooning group (n = 25; 11 [4-35] hours, P = 0.97).

Conclusions In TLVBS patients the severity of LV dysfunction determines the incidence of cardiogenic shock and early cardiac events. Apical and mid-ventricular forms of TLVBS appear to be distinct patterns.

Keywords Takotsubo cardiomyopathy – clinical outcome – systolic heart failure – mortality.

occurring after significant emotional or physical stress. First described in Japan nearly two decades ago1, the first series of TLVBS patients outside of Japan was reported by our group in 20032. The prevalence is between 0.7 and 2.5% in patients with suspected acute coronary syn-drome3. Many aspects of its pathogenesis still remain poorly understood, and prognosis is generally considered good after the treatment of acute phase complications such as heart failure with cardiogenic shock, pulmonary congestion and ventricular arrhythmias. While the orig-inal description of takotsubo cardiomyopathy referred exclusively to apical ballooning, more recently mid-ven-tricular3,4 and basal5 ballooning have been described. It is not clear if these patterns are distinct variants of ven-tricular involvement, or if they represent varying stages

97682.indb 49697682.indb 496 3/10/14 10:523/10/14 10:52

LV function and outcome in TLVBS 497

6 segments (figure 1); the basal anterior wall, the mid-anterior wall, the apical anterior wall, the apical inferior wall, the mid-inferior wall and the basal inferior wall. All segments were given a score of 1 to 4; 1 = normal contrac-tion, 2 = hypokinesia, 3 = akinesia, and 4 = dyskinesia. The scores from all segments were added up to give an overall LV score (range 6-24). All available ventriculo-grams were reviewed and a score was given.

Statistical analysis

All data were expressed as mean ± SD or median (interquartile range) as appropriate, or as percentage (number). The Student’s t-test and the chi-square test were used to compare groups. Pearson’s correlation was used to compare the severity of LV involvement with ejection fraction and degree of QTc prolongation. Sta-tistical analyses were performed using SPSS (version 20, SPSS Inc. Chicago, Illinois). A P value of < 0.05 was considered significant.

RESULTS

A total of 145 TLVBS episodes were identified in 139 patients; 4 patients had one recurrent episode, with one patient suffering 2 recurrent episodes of TLVBS. The average age at presentation was 67 ± 12 years and 89% (n = 123) of patients were female. Hypertension (54%) and hyperlipidaemia (41%) were the most com-mon cardiovascular risk factors (table  1). Chest pain (71%), dyspnoea (44%) and nausea/vomiting (9%) were

of ventricular involvement depending on the timing of ventricular imaging during the disease process. Further-more it is unknown if the extent of ventricular involve-ment correlates with the incidence of in-hospital mortal-ity, adverse cardiovascular events, or with the degree of QTc interval prolongation.

We conducted a retrospective analysis of data, pro-spectively entered into a clinical database of patients presenting with TLVBS to our department, to assess if the severity of ventricular dysfunction adversely affects clinical outcome, to determine precisely the in-hospital and 1-year mortality rates, and to characterize the dif-ferent patterns of ventricular involvement in TLVBS.

METHODS

Between January 1998 and August 2012 all consecutive patients presenting with TLVBS to our university hospi-tal department of cardiology were prospectively identified using pre-specified criteria, and entered into a clinical database. Patients were diagnosed with TLVBS according to the Modified Mayo Clinic criteria6. The apical balloon-ing form was defined as akinesia of the mid to apical segments of the left ventricle, with normo- to hyperkine-sia of the basal segments, whilst the mid-ventricular form was defined as akinesia of the mid-ventricular segments, and normokinesia of the basal and apical segments. The basal ballooning form was defined as basal left ventricu-lar akinesia with normal mid and apical wall motion. Clinical information including demographic data, time of onset of symptoms, time of diagnosis, echocardio-graphic and angiographic data, were retrospectively ana-lysed. The incidence of in-hospital mortality and other major cardiac events (cardiogenic shock, heart failure/pulmonary oedema or major cardiac arrhythmia) was recorded and the peak QTc interval was calculated for all patients. Cardiogenic shock was clinically defined based on the following criteria: (a) systolic blood pressure < 90 mmHg for 30 minutes before inotropes/vasopressors, or vasopressors/IABP were required to maintain systolic blood pressure ≥ 90 mmHg, (b) evidence of decreased organ perfusion, and (c) heart rate ≥ 60 beats per minute. A major cardiac arrhythmia consisted of any ventricular arrhythmia or atrial tachyarrhythmia with haemody-namic compromise.

Only patients who had coronary angiography and ventriculography performed and available for retrospec-tive analysis, were included in the ventricular function-clinical outcome analysis group. A severity score of left ventricular (LV) involvement was developed to investigate the hypothesis that the extent of LV wall involvement may be a better prognostic indicator than the overall LV ejec-tion fraction (EF). The left ventricle was divided into

Fig. 1 Left ventricular wall segments: (1) basal anterior wall, (2) mid-anterior wall, (3) apical anterior wall, (4) apical inferior wall, (5) mid-inferior wall, (6) basal inferior wall.

97682.indb 49797682.indb 497 3/10/14 10:523/10/14 10:52

J. Bennett et al.498

In terms of seasonal occurrence, there was a numer-ical trend to higher rates of TLVBS in summer months though this did not reach statistical significance (fig-ure 2A, P = 0.28).

In 87 patients (60%), a clear preceding emotional stressful event (48%) or preceding physical stressor (52%) was present. The admission cardiac troponin I (1.5 ± 2.2 µg/ml ) and CK-MB (8.2 ± 10.3 µg/ml) were raised in 88.3% and 76.5% of TLVBS episodes, respec-tively. The peak troponin I and peak CK-MB were 4.1 ± 11.0 µg/ml and 13.8 ± 17.7 µg/ml, respectively.

Left ventricular dysfunction and clinical outcome

In 119 TLVBS episodes, coronary angiography and left ventriculography were available for review and ret-rospective analysis. Seventy-four percent (n = 88) of patients had classical apical ballooning and 26% (n = 31) had mid-ventricular ballooning. No forms of basal TLVBS were identified. On ventriculography, the mean LVEF was moderately reduced, at 43 ± 12%, and the mean left ventricular end-diastolic pressure (LVEDP) was raised at 24 ± 7 mmHg. The mean LV score was 12.2 ± 1.9. There was a significant but weak correlation between the LVEF and novel LV score (R = 0.374, P = 0.01).

In patients who died in-hospital (n = 8, 7.1%), the mean EF was 40 ± 12% and the mean LV score was 12.9 ± 2.0, compared to a mean EF of 43 ± 12% and a mean LV score of 12.2 ± 1.9 for in-hospital survivors (P = 0.58 and P = 0.40, respectively; table 3). Seventeen patients (14%) developed cardiogenic shock, of whom 4 patients (24%) required temporary IABP support and 4 patients (24%) died in the hospital. There was a sig-nificant difference in severity of LV impairment (assessed by EF) between patients who did develop cardiogenic shock and the patients who did not (34 ± 8% vs 44 ± 12%, P < 0.01) whilst, although there was a trend, there was no significant difference in LV score between these

the most common presenting symptoms. The presenting ECG patterns of this cohort of TLVBS patients present-ing within 3 hours of symptom onset have recently been described7.

Serial ECGs showed QTc interval prolongation at some point in time in 86% (125) of 145 TLVBS episodes. The overall mean peak QTc interval was 516 ± 61 ms. Eight patients (5.5%) presented following successful resuscitation of an out-of-hospital cardiac arrest, of whom 5 had suffered a ventricular fibrillation or ven-tricular tachycardia cardiac arrest. These 5 patients had prolonged QTc intervals (mean 520 ± 37 ms) on presen-tation. One other patient had ventricular fibrillation on day 5 after admission during emergency vascular surgery for acute limb ischaemia as a result of a displaced vas-cular closure device. This was successfully treated with urgent electrical cardioversion. The patient’s QTc was significantly prolonged (560 ms) when the arrhythmia occurred. In the 6 patients above with documented ven-tricular arrhythmias the QTc intervals were all prolonged but returned to normal limits within 14 days, suggesting no baseline QTc prolongation (table 2).

Table 1 Patient characteristics

Patient characteristics n = 145

Age (years) 67.0 ± 12.2

Female 88.5% (123)

Hypertension 53.8% (78)

Hyperlipidaemia 40.7% (59)

Family history CAD 24.1% (35)

Current smoker 13.8% (20)

Diabetes mellitus 12.4% (18)

Obesity 9.0% (13)

PVD 4.8% (7)

Values expressed as mean ± SD, or as percentage (number). CAD: coronary artery disease, PVD: peripheral vascular disease.

Table 2 Patients with ventricular arrhythmias.

Patient Age, sex Presentation QTc (ms) LV EF (%) TLVBS type Clinical outcome

1 38, female OOH VF cardiac arrest 497 30% Apical Died day 8, severe hypoxic brain injury

2 30, female OOH VF cardiac arrest 519 43% Apical Survived

3 49, female OOH VF cardiac arrest 484 50% Apical Survived

4 71, female OOH VF cardiac arrest 580 33% Apical Died day 5, Severe hypoxic brain injury

5 70, female OOH VT cardiac arrest 522 27% Apical Survived

6 61, female VF day 5 post presentation 560 35% Mid-ventricular Survived

QTc: corrected QT interval, LV EF: left ventricular ejection fraction, OOH: out of hospital, VF: ventricular fibrillation, VT: ventricular tachycardia.

97682.indb 49897682.indb 498 3/10/14 10:523/10/14 10:52

LV function and outcome in TLVBS 499

Fig. 2 Seasonal occurrence of TLVBS (panel A). Correlation between the severity of LV involvement (LV score) and the peak QTc interval (panel B).

Table 3 Comparisons of left ventricular (LV) scores and ejection fraction (EF) in patients with and without in-hospital mortality, cardiogenic shock or major cardiac events.

Death prior to discharge Survived to discharge P value

LV score 12.9 ± 2.0 12.2 ± 1.9 0.40

EF (%) 40.3 ± 11.8 42.7 ± 12.0 0.58

Cardiogenic shock No cardiogenic shock P value

LV score 13.1 ± 1.7 12.2 ± 1.9 0.08

EF (%) 33.7 ± 8.4 44.4 ± 11.8 < 0.01

Major cardiac event* No major cardiac event P value

LV score 12.6 ± 1.9 12.5 ± 1.9 0.25

EF (%) 39.4 ± 11.6 44.1 ± 11.9 0.05

Values expressed as mean ± SD. *Congestive heart failure/pulmonary oedema, or major cardiac arrhythmia.

97682.indb 49997682.indb 499 3/10/14 10:523/10/14 10:52

J. Bennett et al.500

and haemodynamic resolution of TLVBS, with 1 elderly patient suffering severe septic shock and subsequent cardiac arrest, and the other 2 patients succumbing to post-gastrointestinal surgical complications and vascu-lar surgery complicated by sepsis and bowel ischaemia, respectively. The median time from admission to in-hospital death was 11 [6-45] days.

During the year following discharge from hospital a further 7 patients died (1-year follow-up data available for 135 TLVBS episodes); 2 patients as a result of cardiac failure, 1 patient following cardiac perforation during implantation of a pacemaker, and 4 patients due to non-cardiac illnesses. The overall 1-year mortality was 12.6% (n = 17).

Type of TLVBS and time from onset of symptoms

In 129 episodes of TLVBS there was a clear time of onset of symptoms and a clear time of clinical diagnosis by echocardiogram or ventriculography, whichever hap-pened first. The median time from symptom onset to clinical diagnosis was 12 [4-30] hours. In 81% (n = 104) of patients, there was apical ballooning and in 19% (n = 25) mid-ventricular ballooning was present. There was no significant difference in time from onset of symp-toms to clinical diagnosis between the apical ballooning group (12 [3-30] hours) and the mid-ventricular bal-looning group (11 [4-35] hours, P = 0.97) suggesting that in these TLVBS patients the 2 forms are distinct patterns of TLVBS.

2 groups (13.1 ± 1.7 vs 12.2 ± 1.9, P = 0.08). Patients who developed congestive heart failure/pulmonary oedema or suffered a major arrhythmia (n = 33, 28%) had a mean EF of 39 ± 12% which was (borderline) significantly lower than the mean EF of 44 ± 12% in those patients who did not develop these acute complications (P = 0.05). In terms of LV score, there was no significant difference between these 2 groups (12.6 ± 1.9 vs 12.5 ± 1.9, P = 0.25, table 3).

In this group of 119 TLVBS episodes, the mean peak QTc was significantly prolonged at 524 ± 62 ms. There was no significant correlation between the degree of peak QTc prolongation and the severity of LV invol-vement on LV score (R = 0.089, figure  2B) or EF (R = 0.187).

Overall in-hospital and 1-year mortality rates

Of the 145 episodes of TLVBS, 10 patients (6.9%, male = 2) died in the hospital (table 4). There was no significant difference in the age of patients who survived or died before discharge, 67 ± 12 years vs 69 ± 15 years, respectively (P = 0.6). Seven succumbed as a direct result of TLVBS-related complications: 4 patients due to severe hypoxic brain injury following out-of-hospital (OOH) cardiac arrest, 2 due to multi-organ failure resulting from progressive heart failure, and 1 patient due to combined cardiac and respiratory failure requiring prolonged and ultimately futile ventilatory support. The non-directly TLVBS attributable deaths occurred after full clinical

Table 4 In-hospital mortality data.

Patient Age, sex Presentation Mean peak QTc (ms)

LV EF (%) Cardiogenic shock

IABP TLVBS type Timing death (days)

Cause of death

1 81, female OOH PEA cardiac arrest 536 41 Yes Yes Apical Day 4 MOF due to progressive heart failure

2 82, female Dyspnoea 408 30 No No Apical Day 14 Septic shock with acute cardiac PEA arrest

3 78, male Chest pain 455 30 Yes No Apical Day 78 MOF due to progressive heart failure

4 71, female OOH VF cardiac arrest 580 33 Yes No Apical Day 5 Severe hypoxic brain injury

5 77, female Chest pain day 2 post gastrectomy

404 45 No No Apical Day 50 Renal failure, sepsis (surgical complications)

6 74, female OOH PEA cardiac arrest 527 35 No No Mid-vent Day 7 Severe hypoxic brain injury

7 69, female Chest pain, dyspnoea 583 28 Yes No Apical Day 57 Ventilatory weaning problems, underlying chest disease

8 48, female OOH PEA cardiac arrest 542 43 No No Apical Day 28 Severe hypoxic brain injury

9 38, female OOH VF cardiac arrest 497 30 Yes No Apical Day 8 Severe hypoxic brain injury

10 73, male Pulmonary oedema 6 hours post AKA

508 40 No No Apical Day 4 Sepsis, ischaemic bowel

In-hospital mortality data. QTc, corrected QT interval; LV EF, left ventricular ejection fraction; IABP, intra-aortic balloon pump; OOH, out of hospital; PEA, pulseless electrical activity; VF, ventricular fibrillation; MOF, multi-organ failure; AKA, above knee amputation.

97682.indb 50097682.indb 500 3/10/14 10:523/10/14 10:52

LV function and outcome in TLVBS 501

(mean EF 36%), whilst 2 had only mild LV impairment. The finding that the degree of LV dysfunction, as assessed by the LV scoring system and calculated EF, does not correspond with the degree of QTc prolonga-tion is interesting as it would suggest that monitoring TLVBS patients until their QTc interval has normalized, even if their LV function is not severely impaired, should be advocated.

Mid-ventricular ballooning: a distinct isolated pattern or phase of classical apical ballooning?

Several forms of TLVBS, including apical, mid-ven-tricular and, more recently, basal TLVBS, have been recognized. The classical apical form of TLVBS appears to be the most common (74% of cases in our cohort) although it is possible that cases of basal TLVBS were not recognized as it has only recently been described. The observation that the time elapsed between the onset of symptoms and assessment of LV function was not different between apical and mid-ventricular TLVBS, suggests that apical and mid-ventricular forms are dis-tinct forms of TLVBS and not the same variant captured at different time intervals. The pathogenesis and under-lying mechanism by which in selected cases only the mid-ventricular region and in other cases the apex and mid-ventricular segments are affected remain unclear.

LIMITATIONS

The relatively small sample size is due to the uncom-mon nature of this condition although this series repre-sents one of the largest cohorts of patients with TLVBS described. Nevertheless, the small numbers in the vari-ous subgroups analysed are acknowledged as a limitation and need to be considered when interpreting the results. Left ventricular function is dynamic and therefore it is not possible to ensure capture and analysis of the vary-ing patterns and severities of LV dysfunction. It would have been interesting to assess the evolving morpho-logical and functional ventricular changes on echocar-diography, however, ventricular assessment was per-formed at diagnosis and to confirm normalization of ventricular function, and unfortunately more frequent ventricular assessment during the acute phase was not part of the clinical protocol.

CONCLUSIONS

In our large cohort of patients presenting with TLVBS, the degree of LV dysfunction on angiography appears to correspond with the incidence of cardiogenic

DISCUSSION

In keeping with previous publications, the classical TLVBS patient in our cohort is a 67-year old female who presents with chest pain and dyspnoea following a stress-ful event. In the majority of cases, the patient makes a relatively quick recovery with resolution of symptoms, and normalization of ventricular function within 1-2 weeks, although ECG changes may take significantly longer to normalize7. The overall in-hospital mortality rate of 6.9% is consistent with previous studies, which demonstrate mortality rates of up to 8%,1,8 and is there-fore similar to that of myocardial infarction9,10. These in-hospital and 1-year mortality rates of 6.9% and 12.6%, respectively, add to the increasing evidence that TLVBS is not a benign cardiac condition.

The new LV scoring system that was developed with as aim to better characterize the extent of ventricular wall involvement, sometimes underestimated by EF measurement, did not prove to be of any help in pre-dicting an increased risk of in-hospital mortality, car-diogenic shock (although a definite trend was there) or other acute cardiac complications. There was how-ever a significant, if weak, correlation between the LV score and the EF, suggesting that perhaps in a big-ger cohort, or where differences in the magnitude of LV score are more pronounced, it may have a role to play.

In this large cohort of patients with TLVBS, the angiographic severity of ventricular ballooning, as assessed by EF, did appear to correspond with the inci-dence of cardiogenic shock and, although with border-line significance (P = 0.05), with the incidence of in-hospital major cardiac events. Interestingly, there was no significant correlation (P = 0.58) between the sever-ity of LV dysfunction and in-hospital mortality. This may be because the number of patients who died in this group analysis was too small (n = 8) to detect any significant differences. However, it may also reflect the fact that there are other important contributing factors such as underlying co-morbidities, sepsis-related com-plications and, perhaps undervalued, significantly pro-longed QTc intervals predisposing to cardiac arrhyth-mias.

We observed a prolonged QTc interval at some point during the acute episode in 86% of patients. Although QTc prolongation has been previously recognized as a typical electrocardiographic feature of TLVBS, the clin-ical importance of this phenomenon has not been clearly described and therefore has probably been underesti-mated. In our cohort, six patients (4.1%) with severely prolonged QTc intervals (mean QTc 527 ms) suffered a ventricular fibrillation cardiac arrest. The ventricular function in 4 of the 6 patients was severely reduced

97682.indb 50197682.indb 501 3/10/14 10:523/10/14 10:52

J. Bennett et al.502

QTc prolongation was a frequent (86% of patients) finding and advocates the need for careful cardiac mon-itoring.

CONFLICT OF INTEREST: none.

shock and, to a lesser extent, with the incidence of early cardiac complications such a clinical heart failure, pulmonary oedema and major cardiac arrhythmias. In-hospital mortality did not appear to be related to the severity of LV dysfunction. Apical and mid-ventricular forms of TLVBS appear to be distinct patterns. Significant

1. Tsuchihashi K, Ueshima K, Uchida T, Oh-mura N, Kimura K, Owa M, Yoshiyama M, Miyazaki S, Haze K, Ogawa H, Honda T, Hase M, Kai R, Morii I; Angina Pectoris-Myocardial Infarction Investigations in Japan.Transient left ventricular apical ballooning without coronary artery stenosis: a novel heart syndrome mimicking acute myocardial infarction. Angina Pectoris-Myocardial Infarction Investigations in Japan. J Am Coll Cardiol 2001; 38: 11-8.

2. Desmet WJ, Adriaenssens BF, Dens JA. Apical ballooning of the left ventricle: first series in white patients. Heart 2003; 89: 1027-31.

3. Pilgrim TM, Wyss TR. Takotsubo cardiomyopathy or transient left ventricular apical ballooning syndrome: A systematic review. Int J Cardiol 2008; 124: 283-92.

4. Bybee KA, Kara T, Prasad A, Lerman A, Barsness GW, Wright RS, Rihal CS. Systematic review: transient left ventricular apical ballooning: a syndrome that mimics ST-segment elevation myocardial infarction. Ann Intern Med 2004; 141: 858-65.

5. Muller O, Roguelov C, Pascale P. A basal variant form of the transient ‘midventricular’ and ‘apical’ ballooning syndrome. QJM 2007; 100: 738-39.

6. Prasad A, Lerman A, Rihal CS. Apical ballooning syndrome (Tako-Tsubo or stress cardiomyopathy): A mimic of acute myocardial infarction. Am Heart J 2008; 155: 408-17.

7. Bennett J, Ferdinande B, Kayaert P, Wiyono S, Goetschalkx K, Dubois C, Sinnaeve P, Adriaenssens T, Coosemans M, Desmet W. Time course of electrocardiographic changes in transient left ventricular

ballooning syndrome. Int J Cardiol 2013; 169: 276-80.

8. Madhavan M, Rihal CS, Lerman A, Prasad A. Acute heart failure in apical ballooning syndrome (TakoTsubo/stress cardiomyopathy): clinical correlates and Mayo Clinic risk score. J Am Coll Cardiol 2011; 57: 1400-01.

9. Brinjikji W, El-Sayed AM, Salka S. In-hospital mortality among patients with takotsubo cardiomyopathy: a study of the National Inpatient Sample 2008 to 2009. Am Heart J 2012; 164: 215-21.

10. Roe MT, Messenger JC, Weintraub WS, Cannon CP, Fonarow GC, Dai D, Chen AY, Klein LW, Masoudi FA, McKay C, Hewitt K, Brindis RG, Peterson ED, Rumsfeld JS. Treatments, trends, and outcomes of acute myocardial infarction and percutaneous coronary intervention. J Am Coll Cardiol 2010; 56: 254-63.

REFERENCES

97682.indb 50297682.indb 502 3/10/14 10:523/10/14 10:52