Published ahead of print on April 3, 2008, doi:10.1164/rccm.200701-031OC
© 2008 American Thoracic Society doi: 10.1164/rccm.200701-031OC
Incidence and Prognosis of Sustained Arrhythmias in Critically Ill Patients1 Service de Réanimation Médicale, Hôpital Raymond Poincaré, Assistance Publique–Hôpitaux de Paris, Faculté de Médecine Paris Ile de France Ouest, Université de Versailles Saint-Quentin en Yvelines, Garches, France; 2 Laboratoire de Biostatistiques, Faculté de Pharmacie, Université de Nantes, Nantes, France; 3 Service de Cardiologie, Hôpital Cochin, Assistance Publique–Hôpitaux de Paris, Université Paris V, Paris, France; 4 Service de Cardiologie, Hôpital Européen Georges Pompidou, Assistance Publique–Hôpitaux de Paris, Université Paris V, Paris, France; 5 Service de Cardiologie, Hôpital Brabois, Centre Hospitalier Universitaire, Université de Nancy I, Vandoeuvre Les Nancy, France; and 6 Centre d'Investigation Clinique INSERM 0203, Unité de Pharmacologie Clinique, Service de Pharmacologie, Hôpital de Pontchaillou, Centre Hospitalier Universitaire, Faculté de Médecine, Université de Rennes 1, Rennes, France Correspondence and requests for reprints should be addressed to Professor Djillali Annane, M.D., Ph.D., Service de Réanimation Médicale, Hôpital Raymond Poincaré, 104 Boulevard Raymond Poincaré, 92380 Garches, France. E-mail: djillali.annane{at}rpc.aphp.fr
Rationale: Sustained arrhythmias are common in postoperative and cardiac intensive care units (ICUs), but their incidence and prognosis in general ICUs have never been reported. Objectives: To estimate the incidence and prognosis of sustained arrhythmias in a general ICU population. Methods: Prospective, multicenter, 1-month inception cohort study. Measurements and Main Results: A total of 1,341 patients were included: 12% (163/1,341) had sustained arrhythmias, including 8% (113/1,341) and 2% (30/1,341) with supraventricular and ventricular arrhythmias, respectively, and 2% (30/1,341) with conduction abnormalities. In-hospital death rates were 17% (205/1,178) in patients without arrhythmia and 29% (33/113) in patients with supraventricular arrhythmias (odds ratio [OR], 1.95; 95% confidence interval [CI], 1.27–3.01), 73% (22/30) in patients with ventricular arrhythmias (OR, 13.20; 95% CI, 5.79–30.10), and 60% (18/30) in patients with conduction abnormalities (OR, 7.46; 95% CI, 3.52–15.82). Neurological sequel rates were 6% (55/973) in arrhythmia-free survivors and 15% (12/80) in survivors with supraventricular arrhythmias (OR, 2.92; 95% CI, 1.45–5.89), 38% (3/8) in survivors with ventricular arrhythmias (OR, 7.53; 95% CI, 1.60–35.50), and 17% (2/12) in survivors with conduction abnormalities (OR, 8.77; 95% CI, 1.65–46.57). After adjusting for prognosis factors and propensity scores, ventricular arrhythmias still increased mortality (OR, 3.53; 95% CI, 1.19–10.42) but supraventricular arrhythmias and conduction abnormalities did not. Conclusions: Sustained arrhythmias are observed in 12% of patients admitted to general ICUs. Ventricular arrhythmias increase the risk of death.
Key Words: arrhythmia • critical illness • hospital mortality • neurological sequel • propensity score
Cardiac arrhythmias are frequently observed after cardiac (1) and noncardiac (2) surgery. Atrial fibrillation is the most common feature, but all types of arrhythmias can be observed. Known risk factors include age, male sex, history of arrhythmias, congestive heart failure or hypertension, bicaval venous cannulation, pulmonary vein venting, prolonged cross-clamp times, intraaortic balloon pump, postoperative atrial pacing, postoperative pneumonia, mechanical ventilation for more than 24 hours, and return to the intensive care unit (ICU) (3, 4). Although transient, postoperative arrhythmias may have a major economic impact with increased hospital length of stay (3, 4). Critically ill patients often have underlying heart disease, are frequently submitted to metabolic, ischemic, or neurohormonal stressors, and are therefore particularly at risk for cardiac arrhythmias (5). In a single-center retrospective study performed on 2,820 critically ill patients over a 12-year period, the prevalence of arrhythmias was 78%, with atrial fibrillation being the prominent clinical problem (6). Arrhythmias were more likely to occur in patients admitted in the ICU for cardiovascular disorders and were associated with an increased risk of death. In a single-center prospective 32-month cohort study performed on 756 critically ill patients with predominantly cardiac diseases, the annual incidence of arrhythmias was approximately 17%, with atrial fibrillation and monomorphic ventricular tachycardia as the most common features (7). However, similar data are lacking for patients admitted to general ICUs. Thus, the purpose of our study was to prospectively assess the incidence and prognosis of sustained arrhythmias in a general ICU population.
This was a prospective, multicenter, 1-month (from December 1 to January 1) inception cohort study performed in 26 European general ICUs. The protocol was approved as a minimal risk protocol by the ethics committee at each participating center and informed consent was waived.
Study Population and Data Collection
Follow-up
Definitions
Sudden cardiac death was defined by abrupt change in clinical status (i.e., arrhythmias, hypotension, chest pain, dyspnea), followed within 1 hour by abrupt coma and sudden cardiovascular collapse resulting in biological death (12).
Endpoints
Statistical Analysis
Study Population Participating ICUs were medical ICUs (n = 14), surgical ICUs (n = 3), or mixed, medical and surgical, ICUs (n = 9) (see Table E1 of the online supplement). A total of 1,341 patients were enrolled (Table 1). Most patients were nonsurgical (1,016/1,341, 76%), and admitted for noncardiac diseases (1,166/1,341, 87%). Only 9 patients were admitted for acute myocardial infarction and 18 patients were admitted for cardiogenic shock. Patients' severity of illness was witnessed by high SAPS II scores (33; IQR, 26) and by the high proportions of patients on ventilator (757/1,341, 56%) or on vasopressors (416/1,341, 31%). All patients in whom anticoagulation was not contraindicated were receiving prophylactic doses of low-molecular-weight heparin or unfractionated heparin for the prevention of venous thromboembolic disease. The crude proportion of patients with new onset of arrhythmia was 163 of 1,341 (12%). Overall, patients with arrhythmias were older, were more likely to have past medical history of cardiovascular, pulmonary, or endocrine diseases, had higher McCabe class, were more likely to present with sepsis/septic shock or other acute cardiovascular disorder, had higher SAPS II score, and were more likely to be on ventilator or on vasopressors. Hospital length of stay was 3 (IQR, 6) days in patients without arrhythmias and 6 (IQR, 12) days in patients with arrhythmias.
Description of Arrhythmias One hundred and forty-six episodes of supraventricular arrhythmias were recorded in 8% (113/1,341) of the patients, with atrial fibrillation as the commonest arrhythmia (Table 2). There were 45 episodes of ventricular arrhythmias in 2% (30/1,341) of the patients, and 39 episodes of cardiac conduction abnormality in 2% (30/1,341) of the patients. The incidence of specific arrhythmias slightly decreased when excluding patients with primary cardiac diagnoses or chronic arrhythmias (see Table E2). Cardiac arrhythmias occurred mainly within the first 48 hours of ICU admission; 0.4% (5/1,341) of patients had sudden cardiac death. Univariate analysis of risk factors yielded roughly similar results for the three different types of arrhythmias (see Table E3). Risk factors included age, past medical history of cardiovascular or endocrine diseases, high SAPS II, need for ventilator, and need for vasopressors.
Among variables recorded at onset of arrhythmias, antiarrhythmic drugs, catecholamines, hyperglycemia, or hypoxia were the main possible precipitating factors (see Table E4). Torsades de pointes occurred in patients with unstable coronary artery disease. Ventricular arrhythmias and conduction abnormalities were more often associated with cardiovascular collapse and persistent coma than supraventricular arrhythmias (see Table E5). Anticoagulants were given in 21% of supraventricular arrhythmias, 4% of ventricular arrhythmias, and 10% of conduction abnormalities (see Table E5).
Outcomes
Among survivors at hospital discharge, neurological sequel rates were 6% (55/973) in patients without arrhythmia, and 15% (12/80) in patients with supraventricular arrhythmias (OR, 2.92; 95% CI, 1.45–5.89), 38% (3/8) in patients with ventricular arrhythmias (OR, 7.53; 95% CI, 1.60–35.50), and 17% (2/12) in patients with conduction abnormalities (OR, 8.77; 95% CI, 1.65–46.57) (Tables 3 and 4). Among the 87 patients with atrial fibrillation, 10 had neurological sequelae including 4 patients with focal neurological deficit and 6 with diffuse anoxic injury. The risk of focal neurological events was associated with supraventricular arrhythmia but not with anticoagulation, and the interaction between supraventricular arrhythmia and anticoagulation was not significant (see Table E9). As compared with survivors without neurological sequelae, survivors with neurological sequelae were more likely to have past medical history of hypertension, trauma or central nervous system disease, higher SAPS II, and to be on ventilator (see Table E7). After adjusting for these covariates and for propensity scores, only ventricular arrhythmias remained associated with increased risk of neurological sequelae (OR, 10.40; 95% CI, 1.84–58.93) (Table 4).
This large multicenter trial established the incidence of sustained arrhythmias and/or conduction abnormalities as approximately 12% of the general ICU population, with atrial fibrillation as the main rhythmic event. Of note, although the current study was performed in nonsurgical, noncardiac ICU patients, the observed incidence of arrhythmias was lower than after coronary artery bypass graft surgery (15, 16), and was close to the incidence of arrhythmias reported in cardiac ICU patients (7) and surgical ICU patients (17–20). There are other points of likeness between cardiac arrhythmias occurring in the general ICU population and those occurring in postoperative patients. Cardiac arrhythmias were more likely to occur within the first 48 hours of ICU admission, and were reported to be more frequent on Postoperative Day 2 (3, 4, 15, 17). Risk factors included age, past medical history of cardiovascular or endocrine diseases, and severity of illness as assessed by SAPS II and the need for mechanical ventilation or vasopressor therapy. All these factors were previously reported in postoperative patients (3, 4). Nevertheless, the incidence of sustained ventricular arrhythmias of 2% was higher than the frequency rate of 0.6 to 1.2% reported after cardiac surgery (21–23). This might be explained by the ICU environment, which is characterized by abundant and active metabolic, ischemic, and neurohormonal stressors, as witnessed, for example, by the high proportion of patients on ventilator or requiring vasopressor therapy. The diagnosis of broad QRS complex tachycardia remains difficult at bedside. The diagnosis of ventricular tachycardia or supraventricular tachycardia was made independently by three board-certified cardiologists who were experts in the field of cardiac arrhythmias. They used established predefined algorithms. There were no discrepancies between them and thus misdiagnosed ventricular tachycardia was very unlikely. Supraventricular arrhythmias did not significantly impact on hospital survival in ICU patients, in contrast to previous reports from single-center studies (6, 7, 19). In fact, in the previous studies, there were no attempts to control for differences in other prognosis factors. Thus, it cannot be ruled out that the observed increased mortality rates in these studies may have been explained by factors other than arrhythmias. Indeed, in our study, adjustment for age, past medical history of cardiovascular diseases, admission category, severity of illness (i.e., SAPS II and need for mechanical ventilation or vasopressors), admission diagnosis, and propensity scores erased the difference in risk of death observed in univariate comparison between patients with or without supraventricular arrhythmias. Thus, it is likely that supraventricular arrhythmias are a marker of critical illness severity rather than a direct cause of death (19). As previously reported in trauma patients, the standardized mortality ratio was similar in patients with or without supraventricular arrhythmias (19). In contrast, ventricular arrhythmias were strongly associated with an increased risk of death and the standardized mortality ratio was significantly greater in patients with than in those without ventricular arrhythmias. Both supraventricular and ventricular arrhythmias significantly increased the risk of neurological sequelae, as reported for postoperative sustained arrhythmias (24). However, the increased risk of neurological sequelae remained constant regardless of covariate and propensity score adjustments only for ventricular arrhythmias. Of note, inference (exact small-sample inference was performed) was based on 923 without neurological sequelae (among which only 5 had ventricular tachycardia [VT]) and on 58 patients with neurological sequelae (among which only 3 had VT) and thus should be taken with caution. In conclusion, this large, prospective, multicenter inception cohort study provides an estimate of the incidence of cardiac arrhythmias in a general ICU population of about 12%. Atrial fibrillation was the most common form of arrhythmia. Ventricular arrhythmias increased the risk of in-hospital mortality and the risk of neurological sequelae. In contrast, supraventricular arrhythmias and conduction abnormalities had no impact on the risk of in-hospital mortality and on the risk of neurological sequelae.
Supported by Assistance Publique–Hôpitaux de Paris. This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org Originally Published in Press as DOI: 10.1164/rccm.200701-031OC on April 3, 2008 Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Participating study investigators: C. Arich, service de réanimation, CHU Nîmes; H. Aube, service de réanimation, CHG Dijon; G. Bleichner, service de réanimation, CH Argenteuil; P.E. Bollaert, service de réanimation, CHU Nancy; G. Bonmarchand, service de réanimation, CHU Rouen; L. Brochard, service de réanimation; CHU Henri Mondor Créteil; G. Capellier, service de réanimation, CHU Besançon; C. Charpentier, service de réanimation chirurgicale, CHU Nancy; J.C. Chevrolet, service de réanimation, Hôpital Cantonal, Genève (Switzerland); C. Chopin, service de réanimation, CHU Lille; Y. Cohen, J.M. Fosse, service de réanimation, CHU Avicenne, Bobigny; J.F. Dhainaut, service de réanimation, Hôpital Cochin, Paris; B. François, service de réanimation, CHU Limoges; B. Guidet, service de réanimation, Hôpital Saint Antoine, Paris; G. Troché, service de réanimation, CH Mignot, Versailles; J.M. Korach, service de réanimation, CHG Chalons/Champagne; M. Leeman, service des soins intensifs, Bruxelles (Belgium); O. Lesieur, service de réanimation, CHG La Rochelle; D. Loriferne, service de réanimation, Hôpital Sainte Camille, Bry/Marne; P. Fouet, service de réanimation, CH Aulnay/Bois; D. Villers, service de réanimation, CHU Nantes; H. Outin, service de réanimation, CHIC Poissy St Germain; J.L. Pinsard, service de réanimation, CHG Elbeuf; A. Parrot, service de réanimation, Hôpital Tenon, Paris; D. Pourrat, service de réanimation, CHU Poitiers; and P. Sauder, service de réanimation, CHU Strasbourg. Received in original form January 5, 2007; accepted in final form March 31, 2008
Related articles in AJRCCM:
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
AT A GLANCE COMMENTARY
0.07 second 
2 or Fisher exact tests (for categorical variables) as appropriate. Propensity scores (
