To test the hypothesis that supraventricular arrhythmia remains a frequent event that is associated with major adverse outcomes after surgery, we studied a consecutive sample of 4181 patients who underwent major, nonemergency, noncardiac procedures and were in sinus rhythm at the preoperative evaluation. Our findings provide insight into the contemporary epidemiology and natural history of perioperative supraventricular arrhythmia.

Methods

Top Methods Results Discussion Author & Article Info References

Patients

All patients 50 years of age or older who underwent major, nonemergency, noncardiac procedures at Brigham and Women's Hospital, Boston, Massachusetts, from 18 July 1989 to 28 February 1994 were eligible for the study. Major noncardiac procedures were defined as those associated with an expected length of stay of 2 or more days. Procedures were electively scheduled or were done during inpatient admissions. The Brigham and Women's Hospital Human Research Committee approved the enrollment and clinical data collection protocols. Patients were included in this analysis if they were in sinus rhythm at the preoperative evaluation. Of the 4315 consecutive patients enrolled in the overall study, 4181 (97%) met this criterion and constituted the study sample for this analysis.

Data Collection

The data collection protocol is described elsewhere ([4]; Lee TH, Marcantonio ER, Mangione CM, Thomas EJ, Cook EF, Sugarbaker DJ, et al. Derivation and prospective validation of a simple index for prediction of cardiac risk of major noncardiac surgery. In preparation). In brief, patients who provided informed consent for the full study protocol had preoperative evaluations done by study personnel, who used a structured data form. These evaluations included detailed medical histories, physical examinations, and laboratory testing. Intraoperative data were obtained from anesthesia flow sheets and surgeons' notes. For patients who did not have this evaluation because they could not be approached or because they refused to participate in the interview portion of the study, clinical data were obtained from the structured evaluation performed by the anesthesiologist noted in the medical record. This data source was also used to obtain American Society of Anesthesiologists (ASA) classifications for all patients. Consenting patients agreed to postoperative sampling of creatine kinase. If total creatine kinase levels were elevated, creatine kinase-MB levels were measured immediately after surgery, at 8 p.m. on the evening of surgery, and on the next two mornings. For other enrolled patients, creatine kinase-MB testing was done according to the physician's orders. Electrocardiography was performed in the recovery room and 1, 3, and 5 days after surgery if the patient remained hospitalized. Study personnel reviewed charts daily for the occurrence of complications.

Classification of Outcome

The perioperative period was divided into two parts: the intraoperative period, which ended at the patient's departure from the operating room, and the postoperative period. A perioperative supraventricular arrhythmia was diagnosed whenever a patient was noted by his or her physicians or by the study's research personnel to have new onset of atrial fibrillation, atrial flutter, paroxysmal atrial tachycardia, multifocal atrial tachycardia, or regular supraventricular tachyarrhythmia without discernible P waves. For the purpose of this study, only episodes of supraventricular arrhythmia that were reported to be persistent or were treated with specific therapies were considered in the analysis. Review of records for a randomly selected subset of 60 patients with supraventricular arrhythmia revealed that 87% received specific treatment (cardioversion or antiarrhythmic therapy).

The postoperative occurrence of major cardiac events was classified by a single reviewer [who was blinded to preoperative clinical data] on the basis of postoperative clinical information, including cardiac enzyme measurements, electrocardiograms, and clinical events. Criteria for perioperative myocardial infarction were 1) a new Q wave lasting 0.04 seconds or longer, 2) creatine kinase-MB levels greater than 5% of an elevated total creatine kinase level with a typical increase and decrease in creatine kinase-MB levels, and 3) creatine kinase-MB levels greater than 3% of an elevated total creatine kinase level in the setting of electrocardiographic changes consistent with ischemia [4].

Major cardiac events were myocardial infarction, cardiogenic pulmonary edema, ventricular fibrillation or primary cardiac arrest, and sustained complete heart block. Intermediate cardiac events, recorded by study research personnel on the basis of record review, were congestive heart failure without pulmonary edema, myocardial ischemia with documented electrocardiographic changes, and documented ventricular tachycardia. Noncardiac events were pulmonary embolism, cerebrovascular accident, noncardiac pulmonary edema, respiratory failure requiring intubation, renal failure requiring dialysis, gastrointestinal bleeding, bacteremia, pneumonia, urinary tract infection, and wound infection.

Statistical Analysis

Univariate analyses of the correlation between clinical characteristics and the occurrence of perioperative supraventricular arrhythmia were performed by using a chi-square test and the Fisher exact test for categorical variables and a t-test or Wilcoxon test for continuous variables. Variables with a P value less than 0.10 were then entered into the multiple regression analysis. Stepwise logistic regression analysis was used to determine independent correlates of supraventricular arrhythmia. A two-sided P value less than 0.05 was considered statistically significant. Clinically relevant variables from the regression model were analyzed for potential interactions. Potentially significant interaction terms were considered in the regression models.

In the analysis of the association between intraoperative events and the occurrence of supraventricular arrhythmia, only postoperative episodes of supraventricular arrhythmia were included (n = 256). A clinically significant hypotensive episode during surgery was defined as a decrease of one third in the original systolic blood pressure or a systolic blood pressure less than 90 mm Hg.

Linear regression models were used to estimate the independent variation in length of stay attributable to supraventricular arrhythmia; these models controlled for age, sex, preoperative clinical characteristics, ASA classification, type of procedure, and postoperative events (major and intermediate cardiac events and noncardiac events). The log transformation of length of stay was used because of the non-normal distribution of this variable. The adjusted hospital days attributable to supraventricular arrhythmia were estimated from the percentage change in length of stay by using the geometric mean of length of stay from the final model.

Results

Top Methods Results Discussion Author & Article Info References

The patients' mean age (±SD) was 66 ± 9 years. A total of 2244 patients (54%) were male, 563 (14%) were diabetic, 609 (15%) had a history of myocardial infarction, 244 (6%) had congestive heart failure, and 1366 (33%) had evidence of ischemic heart disease. The types of procedures performed were orthopedic (35%), thoracic (12%), abdominal (12%), abdominal aortic aneurysm (5%), other vascular (16%), and other general surgical (20%) procedures.

A new perioperative supraventricular arrhythmia developed in 317 patients (7.6%) (Table 1). Less than 25% of patients who had supraventricular arrhythmia had more than one episode of arrhythmia. Atrial fibrillation and supraventricular tachycardia without discernible P waves were the most frequent types of supraventricular arrhythmia and usually occurred in the postoperative period. The time of occurrence of postoperative supraventricular arrhythmia peaked during the first 3 days after surgery (Figure 1). This pattern of occurrence after surgery was similar for all types of supraventricular arrhythmia.

View larger version (15K): [in this window] [in a new window]   Figure 1. Distribution of supraventricular arrhythmias (SVA) according to the time of first occurrence among the 256 patients who developed arrhythmias after surgery.

Correlates of Supraventricular Arrhythmia

Compared with patients who did not develop supraventricular arrhythmia, patients who developed this arrhythmia were significantly older and were more likely to be male; to have a history of myocardial infarction, congestive heart failure, diabetes, chronic lung disease, cancer, or arrhythmia; to have evidence of clinically significant valvular disease (such as systolic murmurs of grade III or higher) on physical examination; and to have premature atrial complexes on preoperative electrocardiography (Table 2).

Preoperative use of calcium-channel blockers, digoxin, procainamide, and quinidine was more common in patients who developed supraventricular arrhythmia. Use of ß-blockers did not differ between the two groups. Patients with ASA classification III or IV were more likely to develop supraventricular arrhythmia. The most significant univariate correlate of supraventricular arrhythmia was the type of procedure performed. Patients who had abdominal aortic aneurysm repair or an intrathoracic, abdominal, or other vascular procedure had an increased risk for perioperative supraventricular arrhythmia compared with patients who did not undergo these procedures (Table 2). The type of supraventricular arrhythmia was not significantly associated with the type of procedure. Patients who developed atrial fibrillation underwent procedures similar to those of patients who had other types of arrhythmia.

In multiple regression analysis, independent correlates of supraventricular arrhythmia were age 70 years or older, male sex, congestive heart failure, evidence of clinically significant valvular disease on physical examination, history of supraventricular arrhythmia, history of asthma, presence of premature atrial complexes on preoperative electrocardiography, ASA class III or IV, and type of procedure (Table 3). Intrathoracic, abdominal aortic aneurysm, abdominal, and other vascular procedures were the operations independently associated with perioperative supraventricular arrhythmia.

Although calcium-channel blockers and ß-blockers were not correlated with increases or decreases in risk for supraventricular arrhythmia, current digoxin therapy had complex associations with supraventricular arrhythmia. Among patients with a history of supraventricular arrhythmia, those who were currently receiving digoxin had a higher risk for supraventricular arrhythmia than those not receiving digoxin. In contrast, among patients who had intrathoracic surgery, those receiving digoxin had a lower risk for supraventricular arrhythmia (odds ratio, 0.2 [95% CI, 0.04 to 0.7]) than those not receiving digoxin. In fact, the risk for supraventricular arrhythmia in patients who had intrathoracic surgery and were receiving digoxin did not significantly differ from the risk in patients who had other procedures (odds ratio, 1.8 [CI, 0.4 to 7.7]).

Associated Operative and Postoperative Events

In an analysis of the association between intraoperative events and postoperative supraventricular arrhythmia, the development of hypotensive episodes during surgery was significantly associated with postoperative supraventricular arrhythmia (relative risk, 1.3 [CI, 1.1 to 1.7]). If these episodes lasted more than 10 minutes, the association was even stronger (relative risk, 1.5 [CI, 1.2 to 2.0]). After adjustment for preoperative characteristics, the occurrence of an intraoperative hypotensive episode lasting more than 10 minutes was independently associated with postoperative supraventricular arrhythmia (Table 4).

Table 5 shows the rates of cardiac and noncardiac postoperative events that occurred in the patients who developed perioperative supraventricular arrhythmia and in those who did not develop supraventricular arrhythmia. One hundred sixty-two patients (51%) had at least one other major or intermediate event associated with the occurrence of supraventricular arrhythmia. Patients who had acute cardiac events (congestive heart failure, myocardial infarction, ischemia without myocardial infarction, ventricular tachycardia, or cardiac arrest) had particularly high relative risks for perioperative supraventricular arrhythmia. Noncardiac events associated with supraventricular arrhythmia were bacterial pneumonia, bacteremia, wound infection, urinary tract infection, cerebrovascular accident, pulmonary embolism, and gastrointestinal bleeding. Patients who had hypotensive episodes at any time during the perioperative period also had an increased risk for supraventricular arrhythmia (relative risk, 2.0 [CI, 1.6 to 2.6]).

Analysis of the timing of supraventricular arrhythmia in relation to these other complications revealed that for most major complications, the supraventricular arrhythmia occurred on the same day as the complication or on subsequent days. For example, supraventricular arrhythmia occurred at least 1 day before the diagnosis of acute myocardial infarction in 3 (21%) of 14 patients with this complication and at least 1 day after the diagnosis in 16 (44%) of 36 patients with this complication. One exception to this trend was the occurrence of cerebrovascular accident among patients with supraventricular arrhythmia: In all 5 of these patients, the supraventricular arrhythmia occurred before the cerebrovascular accident was diagnosed; in 2 of these patients, the supraventricular arrhythmia identified was atrial fibrillation.

Effect on Length of Hospital Stay

The overall mean hospital length of stay was 8.1 ± 7.8 days (median, 7 days) (Table 6). Stay was significantly longer in patients who developed supraventricular arrhythmia. In the multivariate analysis, after adjustment for clinical factors, type of procedure, and associated events, supraventricular arrhythmia was independently associated with a 33% (CI, 25% to 41%) increase in length of stay. This increase was equivalent to 2.5 days (CI, 1.9 to 3.1 days).

Discussion

Top Methods Results Discussion Author & Article Info References

Our findings suggest that supraventricular arrhythmia remains a common event that is strongly associated with increased rates of complications and greater length of stay after noncardiac surgery. Increasing use of ß-blockers and calcium-channel blockers in the general population seems to have no effect on the development of postoperative supraventricular arrhythmia. Our data indicate that digoxin may help protect against supraventricular arrhythmia in patients having thoracic surgery, but the association of digoxin with the occurrence of this complication is complex.

Our results are consistent with and extend findings from previous research on supraventricular arrhythmia. The independent predictors of postoperative supraventricular arrhythmia identified in our study-age, male sex, presence of congestive heart failure, valvular disease, history of lung disease, and history of supraventricular arrhythmia-are reportedly associated with the development of supraventricular arrhythmia in the general population [5-7]. Frequent premature atrial complexes have been extensively linked with supraventricular arrhythmia after cardiac surgery, and our study also suggested an association after noncardiac surgery [8-10].

The correlates that we identified are consistent with previous findings. In the 1970s, one large study identified three correlates for the development of postoperative supraventricular arrhythmia: age greater than 70 years, major procedures (intrathoracic, intraabdominal, or major vascular), and presence of pulmonary rales on physical examination [3]. Other studies, most done after thoracic procedures, reinforced these findings and also described associations of supraventricular arrhythmia with preoperative pulmonary function and electrocardiographic findings [1, 2, 11].

In our study and in previous studies, a strong correlate for the development of supraventricular arrhythmia was the type of procedure performed. Several studies in patients undergoing thoracic procedures have reported a high incidence of supraventricular arrhythmia [12-14]. Because of the high risk for postoperative supraventricular arrhythmia, some researchers have suggested that patients undergoing thoracic procedures should receive medical prophylaxis before surgery [15, 16].

What such therapy might be is unclear. Some research suggests that in patients undergoing cardiac surgery, ß-blockers and calcium-channel blockers may help protect against supraventricular arrhythmia. Other data indicate that perioperatively administered metoprolol can reduce the frequency of atrial fibrillation after thoracotomy for lung resection [17]. In our study, however, calcium-channel blockers and ß-blockers were not associated with increases or decreases in the occurrence of supraventricular arrhythmia. It should be emphasized that our analysis was based on preoperative outpatient use of these medications. Serum concentrations of these drugs may have been subtherapeutic at the time of surgery or after surgery. Therefore, our data do not exclude the possibility that perioperative therapy with these agents might reduce rates of supraventricular arrhythmia.

In contrast, preoperative use of digoxin showed important associations with an increased risk for postoperative supraventricular arrhythmia in patients with a history of supraventricular arrhythmia and a decreased risk for supraventricular arrhythmia in patients undergoing thoracic surgery. These interactions must be considered with caution. Among patients with a history of supraventricular arrhythmia, those receiving digoxin were at increased risk for postoperative supraventricular arrhythmia. In this context, long-term digoxin therapy might have been prescribed because of frequent recurrences of supraventricular arrhythmia. Another possibility is that digoxin increases the threshold for the development of arrhythmias [18, 19].

Preoperative factors were not the only correlates of supraventricular arrhythmia in our sample; we also noted that intraoperative hypotension was significantly associated with an increased risk for postoperative supraventricular arrhythmia. This finding has previously been described in patients undergoing thoracotomy and in experimental studies [2, 19, 20]. The responsible mechanisms in this setting are not clear but may involve surgical manipulation around the pericardium and lung hilum. In the overall population of patients having noncardiac surgery, postoperative supraventricular arrhythmia may be a consequence of ischemia induced by hypotension. The occurrence of intraoperative hypotension might also just be a marker for patients who have more severe cardiovascular disease and are therefore at greater risk for supraventricular arrhythmia.

An analogous problem characterizes the interpretation of the association of supraventricular arrhythmia with an increased risk for postoperative complications seen in our study. Whether supraventricular arrhythmia caused these complications, was a consequence of them, or was a marker for increased risk for complications is not clear. In most cases, supraventricular arrhythmia occurred on the same day as the complications or after them. Among these patients, supraventricular arrhythmia was probably not the cause of the complication but resulted from effects associated with these complications, such as hypoxia, hypovolemia, and electrolyte imbalances.

In our patients, perioperative supraventricular arrhythmia was strongly associated with increased duration of hospitalization, even after adjustment for clinical characteristics and the occurrence of postoperative complications. Although part of this increase in length of stay may be due to incomplete adjustment for the severity of complications in patients with supraventricular arrhythmia, it is reasonable to hypothesize that some prolongation of hospitalization was due to the measures used to treat the arrhythmias. This finding raises the question of whether increased efficiency might result from use of different treatment strategies for patients with supraventricular arrhythmia, such as deferring cardioversion to the outpatient setting.

Our findings also suggest that care might be improved by identifying patients at increased risk for supraventricular arrhythmia and targeting these patients with prophylactic measures. However, what such measures might be remains unclear. Our data emphasize the need for further research to test strategies to reduce perioperative supraventricular arrhythmia and to minimize its consequences.

Dr. Goldman: University of California, San Francisco, 505 Parnassus Avenue, M-994, San Francisco, CA 94143.

Dr. Lee: Partners Community HealthCare, Inc., Prudential Tower Suite 1150, 800 Boylston Street, Boston, MA 02199.