Ventricular fibrillation (VF) during reperfusion (rVF) in ST-segment elevation myocardial infarction (STEMI) is an infrequent but serious event that complicates coronary interventions. The aim of this study was to analyze clinical predictors of rVF in an unselected population of patients with STEMI treated with percutaneous coronary intervention (PCI). Consecutive patients with STEMI admitted to a tertiary care hospital for primary PCI from 2007 to 2012 were retrospectively assessed for the presence of rVF. Admission electrocardiograms, stored in a digital format, were analyzed for a maximal ST-segment elevation in a single lead and the sum of ST-segment deviations in all leads. Clinical, electrocardiographic, and angiographic characteristics were tested for associations with rVF using logistic regression analysis. Among 3,724 patients with STEMI admitted from 2007 to 2012, 71 (1.9%) had rVF. In univariate analysis, history of myocardial infarction, aspirin and β-blocker use, VF before PCI, left main coronary artery disease, inferior myocardial infarction localization, symptom-to-balloon time <360 minutes, maximal ST-segment elevation in a single lead >300 μV, and sum of ST-segment deviations in all leads >1,500 μV were associated with increased risk for rVF. In a multivariate analysis, sum of ST-segment deviations in all leads >1500 μV (odds ratio 3.7, 95% confidence interval 1.45 to 9.41, p = 0.006) before PCI remained an independent predictor of rVF. In-hospital mortality was 18.3% in the rVF group and 3.3% in the group without VF (p <0.001), but rVF was not an independent predictor of in-hospital death. In conclusion, the magnitude of ST-segment elevation before PCI for STEMI independently predicts rVF and should be considered in periprocedural arrhythmic risk assessment. Despite higher in-hospital mortality in patients with rVF, rVF itself has no independent prognostic value for prognosis.
Ventricular fibrillation (VF) during reperfusion for ST-segment elevation myocardial infarction (STEMI) is an infrequent event, but it complicates coronary interventions and subsequent hospital stays. Because of its relatively low incidence, the predictors and prognostic value of VF during reperfusion are usually analyzed together with other VF episodes at any time of acute STEMI. Some studies have divided VF into early and late timing, and some have dealt with prereperfusion, periprocedural, or postprocedural VF. Although experimental studies have been conducted to search for specific underlying mechanisms of reperfusional arrhythmias at the cellular level, clinical studies focused on VF during reperfusion for STEMI in unselected populations, to the best of our knowledge, are lacking. Most clinical studies have analyzed clinical and angiographic predictors of VF, whereas the data on dynamic electrocardiographic changes that can predict VF, especially VF during reperfusion, are scarce. Our aim was to analyze electrocardiographic characteristics associated with VF during reperfusion in unselected patients with STEMI treated with percutaneous coronary intervention (PCI).
Methods
Consecutive patients with STEMI admitted to Lund University Hospital for primary PCI from 2007 to 2012 were retrospectively assessed for the presence of VF during reperfusion. Information about cardiopulmonary resuscitation or defibrillation for VF was retrieved from the Swedish National Register of Information and Knowledge About Swedish Heart Intensive Care Admissions, and information about reperfusion arrhythmias was obtained from the Swedish Coronary Angiography and Angioplasty Register. Then medical histories of patients were reviewed to crosscheck VF occurrence and its timing in relation to infarct-related artery (IRA) opening. Relevant clinical information was taken from the Swedish National Register of Information and Knowledge About Swedish Heart Intensive Care Admissions, and angiographic characteristics were determined from the Swedish Coronary Angiography and Angioplasty Register.
Electrocardiograms (ECGs) stored in digital format in either GE Marquette MUSE system (GE Medical Systems, Milwaukee, Wisconsin) or Infinity MegaCare ECG Management System (Dräger, Lübeck, Germany) databases were analyzed for predictors of VF during reperfusion. We looked for admission ECGs and historical ECGs recorded before coronary events.
A previously recorded standard 12-lead ECG unrelated to STEMI available for interpretation was defined as a historical ECG. The most recent ECG was used for analysis if several historical ECGs were available. Apart from the standard criteria (P, PR, QRS, QT, and corrected QT intervals and the presence of right bundle branch block [RBBB] or left bundle branch block [LBBB]), the presence of J-point elevation ≥1 mm higher than baseline in 2 contiguous inferior or lateral leads was analyzed.
An ECG recorded after the onset of STEMI but before coronary intervention was defined as an admission ECG. If several ECGs were recorded before PCI, the latest ECG was considered the admission ECG, either in-hospital or prehospital ECG (in those cases in which in-hospital ECGs before PCI were not taken or not saved in the database). ECGs with paced rhythm were excluded. ECGs with complete RBBB or LBBB were excluded from the analysis of parameters characterizing ventricular repolarization (i.e., ST level, QT interval, corrected QT interval). On the basis of the admission ECG, the maximal ST-segment elevation in a single lead with most prominent elevation, the sum of ST-segment deviations in all 12 leads, including ST-segment elevation and reciprocal depression, as well as the Anderson-Wilkins and Sclarowsky-Birnbaum scores were calculated.
In brief, the Anderson-Wilkins acuteness score takes into consideration the presence of abnormal Q waves to the Selvester QRS scoring system and the morphology of the T wave, classified as tall, positive, flat or negative. The acuteness of each standard lead with ST-segment elevation is classified from 0 to 4 points, and the total sum in all leads is then divided by the number of leads involved to correct for the overall extent of the myocardial involvement.
The Sclarowsky-Birnbaum score assess depolarization changes during ischemia progression. Tall peaked T waves are classified as grade I ischemia, ST-segment elevation is present at grade II, and changes in the terminal part of the QRS complex appear with grade III ischemia. Criteria for grade III include disappearance of S waves in leads with Rs configuration and J point/R ratio ≥0.5 in leads with qR configuration.
To identify clinical factors associated with VF during reperfusion, relevant clinical, angiographic, and electrocardiographic factors were compared across groups using chi-square or Fisher’s exact tests for categorical variables and Student’s t tests for continuous variables with approximately normal distributions, or the Mann-Whitney U test as appropriate. Significantly associated covariates were further evaluated in univariate logistic regression models with estimation of odds ratios and likelihood ratio tests. To determine independent factors of risk, factors significantly associated with reperfusional VF in univariate models were included in a stepwise regression analysis with backward elimination. A p value <0.05 was considered significant. All analyses were performed using SPSS version 22.0 (SPSS, Inc., Chicago, Illinois).
Results
Among 3,274 patients with STEMI admitted for primary PCI from 2007 to 2012, 71 (1.9%) had VF during reperfusion. The incidence of reperfusion VF did not differ in different years: 13 of 627 (2.1%) in 2007, 11 of 538 (2.0%) in 2008, 11 of 553 (2.0%) in 2009, 12 of 633 (1.9%) in 2010, 10 of 678 (1.5%) in 2011, and 14 of 735 (1.9%) in 2012.
All patients who had VF during reperfusion from 2007 to 2012 constituted the rVF group, and 614 consecutive patients without arrhythmias admitted during 2007 were taken as controls (no rVF group). Clinical characteristics were analyzed for all 685 patients (71 in the rVF group and 614 in the no rVF group), but admission ECGs were not available for 17%, thus leaving 567 patients included in the analysis of electrocardiographic characteristics (55 patients in the rVF group and 512 in the no rVF group). Among them, 108 were ambulance ECGs and 459 in-hospital pre-PCI ECGs. Two patients in the no rVF group were excluded from electrocardiographic analysis because of paced rhythm. Assessment of repolarization, including the level of ST segment and QT interval, was performed in patients without LBBB or RBBB, a total of 489 patients (42 in the rVF group and 447 in the no rVF group).
Historical ECGs were available for 447 patients: 40 from the rVF group and 407 from the no rVF group. The time from historical ECG to STEMI did not differ between groups: 27 ± 32 months in the rVF group and 40 ± 50 in the no rVF group (p = 0.12).
Patients with VF during reperfusion were more likely to have histories of myocardial infarction and more often used β blockers and aspirin than those without VF ( Table 1 ). Patients with VF during reperfusion more often had VF before PCI, either out of hospital, during ambulance transport, or in hospital before balloon inflation. Eleven patients of 71 in the rVF group had VF before and during reperfusion, whereas 60 patients had no VF before reperfusion and had reperfusion VF only. There were no differences between the groups with regard to age, gender, smoking, presence of hypertension, diabetes, anamnesis of stroke, history of congestive heart failure, using digitalis at admission, proportion of patients with Killip class >I, and serum creatinine at admission. The level of potassium at admission was significantly lower in the rVF group, though within the normal range. Patients with VF during reperfusion had shorter symptom-to-balloon time and more often had inferior localization of myocardial infarction and left main stenosis on angiography. The percentage of multivessel disease and IRA distribution did not differ between groups.
| Variable | VF at reperfusion | p-value | |
|---|---|---|---|
| Yes (n=71) | No (n=614) | ||
| Age (years) | 68±12 | 66±12 | 0.36 |
| Man | 52 (73%) | 430 (70%) | 0.68 |
| Previous myocardial infarction | 15 (22%) | 80 (13%) | 0.04 |
| Hypertension | 31 (46%) | 244 (40%) | 0.36 |
| Stroke | 6 (9.0%) | 46 (7.5%) | 0.63 |
| Chronic heart failure | 4 (6.0%) | 17 (2.8%) | 0.14 |
| Diabetes mellitus | 5 (8%) | 67 (11%) | 0.53 |
| B-blockers | 25 (40%) | 159 (27%) | 0.02 |
| Aspirin | 24 (38%) | 149 (25%) | 0.03 |
| Digitalis | 2 (3.2%) | 8 (1.3%) | 0.24 |
| Past or present smoker | 24 (73%) | 392 (67%) | 0.57 |
| HF at admission Killip>1 | 7 (13%) | 47 (10%) | 0.045 |
| Creatinin at admission (mmol/l) | 97±51 | 87±30 | 0.15 |
| K + at admission (mmol/l) | 3.8±0.5 | 3.9±0.4 | 0.002 |
| Hypokalemia at admission | 4 (5.8%) | 14 (2.7%) | 0.14 |
| VF before reperfusion | 11 (16%) | 26 (4%) | 0.001 |
| Symptom-to-balloon time | 185 (164) | 227 (254) | 0,006 |
| Symptom-to-balloon time <360 min | 51 (84%) | 359 (70%) | 0.025 |
| Multivessel disease | 42 (64%) | 303 (55%) | 0.19 |
| Left main stenosis | 10 (15%) | 34 (6%) | 0.02 |
| RCA as IRA | 34 (49%) | 226 (41%) | 0.2 |
The patients who had VF during reperfusion were more likely to have longer QRS intervals on historical ECGs before their events ( Table 2 ). Only 1 patient from the rVF group had J-point elevation in the lateral leads meeting criteria for early repolarization on the historical ECG. Patients with VF during reperfusion had higher ST-segment elevation before PCI but did not differ in either Anderson-Wilkins acuteness score or Sclarovsky-Birnbaum score compared with the no rVF group. The percentage of LBBB or RBBB and conventional electrocardiographic criteria did not differ between groups.
| VF at reperfusion | p-value | ||
|---|---|---|---|
| Yes | No | ||
| Historical ECG: | |||
| P duration (ms) | 107±28 | 110±19 | 0.63 |
| PR duration (ms) | 164±30 | 164±30 | 0.97 |
| QRS duration (ms) | 102±20 | 96±16 | 0.02 |
| QTc duration (ms) | 421±28 | 415±26 | 0.34 |
| J-point elevation in lateral leads | 1 (2.1%) | 0 (0%) | 0.08 |
| Admission ECG: | |||
| P duration (ms) | 105±29 | 110±21 | 0.3 |
| PR duration (ms) | 176±39 | 171±34 | 0.45 |
| RBBB | 7 (13%) | 41 (8%) | 0.60 |
| LBBB | 6 (11%) | 22 (4%) | 0.15 |
| Inferior MI localization | 44 (67%) | 188 (51%) | 0.02 |
| ST max (μV) | 498 [330] | 300 [261] | <0.001 |
| ST max >300 μV | 47 (84%) | 253 (52%) | <0.001 |
| Sum ST (μV) | 2289 [1933] | 1518 [1205] | <0.001 |
| Sum ST >1500 μV | 48 (87%) | 249 (52%) | <0.001 |
| Birnbaum grade 3 | 13 (45%) | 138 (40%) | 0.72 |
| Anderson-Wilkins score | 2.5±0.9 | 2.5±1.1 | 0.94 |
In a univariate regression analysis, the following factors were associated with increased risk for VF during reperfusion: history of myocardial infarction, aspirin and βa-blocker use, VF before PCI, potassium level at admission, left main coronary artery disease, inferior localization of myocardial infarction, duration of QRS on historical ECG, symptom-to-balloon time <360 minutes, ST-segment elevation in a single lead >300 μV, and sum of ST-segment deviations in all leads >1,500 μV ( Table 3 ). In a multivariate analysis, sum of ST-segment deviations in all leads >1,500 μV before PCI and left main stenosis by angiography remained independent predictor of VF during reperfusion.
| Charasteristics at admission | Univariate analysis | Multivariate analysis | ||||
|---|---|---|---|---|---|---|
| OR | 95% CI | p-value | OR | 95% CI | p-value | |
| Previous MI | 1.93 | 1.04-3.58 | 0.039 | – | – | – |
| QRS duration at historical ECG | 1.02 | 1.003-1.04 | 0.020 | – | – | – |
| K + at admission | 0.40 | 0.22-0.73 | 0.003 | – | – | – |
| VF before PCI | 4.15 | 1.95-8.81 | <0.001 | |||
| Medications: | ||||||
| Aspirin | 1.88 | 1.09-3.22 | 0.023 | – | – | – |
| B-blockers | 1.86 | 1.09-3.19 | 0.024 | – | – | – |
| Symptom-to-baloon time <360 min | 2.19 | 1.08-4.42 | 0.029 | – | – | – |
| Left main stenosis | 2.60 | 1.22-5.54 | 0.013 | 4.47 | 1.19-18.80 | 0.027 |
| Inferior MI | 1.89 | 1.09-3.29 | 0.023 | |||
| ST max >300 μV | 4.87 | 2.34-10.16 | <0.001 | – | – | – |
| Sum ST >1500 μV | 6.44 | 2.86-14.53 | <0.001 | 4.00 | 1.52-10.54 | 0.005 |
Because prereperfusional VF appeared to be a strong predictor of rVF, we also performed a separate analysis of the predictors of lone VF during reperfusion in patients who did not have VF before IRA opening ( Table 4 ). In the univariate regression analysis, the following factors were associated with increased risk for lone VF during reperfusion: history of myocardial infarction, aspirin and β-blocker use, low potassium level at admission, left main coronary artery disease, duration of QRS on historical ECG, symptom-to-balloon time <360 minutes, ST-segment elevation in single lead >300 μV, and sum of ST-segment deviations in all leads >1,500 μV. In a multivariate analysis, sum of ST-segment deviations in all leads >1,500 μV, aspirin use, and QRS duration on historical ECG remained independent predictor of lone VF during reperfusion.
| Charasteristics at admission | Univariate analysis | Multivariate analysis | ||||
|---|---|---|---|---|---|---|
| OR | 95% CI | p-value | OR | 95% CI | p-value | |
| Previous MI | 2.33 | 1.22-4.48 | 0.011 | – | – | – |
| QRS duration at historical ECG | 1.02 | 1.00-1.04 | 0.014 | 1.02 | 1.00-1.05 | 0.042 |
| K + admission | 0.40 | 0.21-0.77 | 0.006 | – | – | – |
| Medications: | ||||||
| Aspirin | 1.99 | 1.12-3.54 | 0.020 | 2.97 | 1.29-6.80 | 0.010 |
| B-blockers | 2.12 | 1.20-3.74 | 0.009 | – | – | – |
| Symptom-to-baloon time <360 min | 2.21 | 1.05-4.62 | 0.036 | – | – | – |
| Left main stenosis | 2.29 | 1.01-5.22 | 0.048 | – | – | – |
| ST max >300μV | 4.99 | 2.29-10.85 | <0.001 | – | – | – |
| Sum ST >1500μV | 6.80 | 2.84-16.30 | <0.001 | 4.40 | 1.57-12.28 | 0.005 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
