Author + information
- Received August 12, 2002
- Revision received October 17, 2002
- Accepted October 31, 2002
- Published online March 5, 2003.
- Bruce A Koplan, MD*,* (, )
- William G Stevenson, MD, FACC*,
- Laurence M Epstein, MD, FACC*,
- Sary F Aranki, MD† and
- William H Maisel, MD, MPH*
- ↵*Reprint requests and correspondence:
Dr. Bruce A. Koplan, Cardiac Arrhythmia Service/Division of Cardiology/Department of Medicine, Brigham and Women’s Hospital, 75 Francis Street, Boston, Massachusetts 02115, USA.
Objectives The study objective was to develop and validate a simple risk score to predict postoperative permanent pacing (PPM) after valve surgery.
Background Our ability to identify patients preoperatively that will require PPM is poor. A simple preoperative risk score to predict PPM after valve surgery could assist both clinical practice and research.
Methods All valve surgery patients at our institution from 1992 to 2002 were included (n = 4,694). Two-thirds of the patients were randomly selected to form a risk score prediction group (PG), and the score was then applied to the remaining patients (validation group [VG]).
Results Preoperative right bundle branch block (odds ratio [OR], 3.6; 95% confidence interval [CI], 2.3 to 5.7) and multivalve surgery that included the tricuspid valve (OR, 3.7; 95% CI, 2.3 to 6.1) were the strongest independent predictors of PPM, while multivalve surgery that did not include the tricuspid valve (OR, 2.1; 95% CI, 1.3 to 3.3), preoperative left bundle branch block (OR, 2.0; 95% CI, 1.3 to 2.9), preoperative PR interval >200 ms (OR, 1.9; 95% CI, 1.3 to 3.0), prior valve surgery (OR, 1.8, 95% CI, 1.2 to 2.7), and age >70 years (OR, 1.4; 95% CI, 1.04 to 2.0) also predicted PPM. A risk score from 0 to 6 identified patients in the VG with incidences of PPM of 1.9%, 5.2%, 8.7%, 11.5%, 21%, 36%, and 50%, respectively.
Conclusions A simple risk score incorporating preoperative conduction, age, prior valve surgery, and surgery type predicts PPM after valve surgery. This score may be useful in the perioperative management of valve surgery patients.
Approximately 5% of the 100,000 patients who undergo cardiac valve surgery each year in the U.S. will require postoperative permanent pacemaker (PPM) implantation before hospital discharge (1,2). Advanced age, previous valve surgery, and multivalve surgery have been associated with the need for PPM, but the ability to preoperatively identify those patients who will require PPM is poor (3–6). Furthermore, the independent predictive value of the preoperative electrocardiogram (ECG) to predict patients who will need permanent pacing is unknown (2). In addition to more accurately informing patients of the potential need of permanent pacing after surgery, preoperative identification of high-risk patients could lead to different approaches to temporary postoperative pacing and altered surgical techniques.
This study was designed to develop and validate a simple preoperative risk score, utilizing demographic, ECG, and surgical characteristics to predict the need for permanent pacing after cardiac valve surgery.
All patients who underwent cardiac valve surgery at the Brigham and Women’s Hospital from March 1992 to February 2002 were included in the study. Patient records were analyzed for baseline demographics, preoperative ECG characteristics, surgery indication and type, and the need for PPM before discharge. Patients who had an indication for PPM or an implanted cardioverter-defibrillator (ICD) preoperatively were excluded from analysis. Patients who died within six days of surgery were also excluded. Patients who underwent postoperative ICD implantation were considered to have a PPM only if they also had an indication for permanent pacing independent of their need for an ICD; otherwise they were classified in the nonpacemaker group. After these exclusions, 4,694 patients met these inclusion criteria and were available for analysis. Preoperative ECG intervals were measured by an automated system (1991, Marquette Electronics, Milwaukee, Wisconsin), which has been previously validated (7,8). The final ECG report generated by the automated system was validated by a cardiologist at our institution.
Two-thirds of the patients were selected at random to form the risk score prediction group (PG). The primary outcome was “postoperative implantation of a permanent pacemaker.” Multivariate logistic regression was performed on the PG to identify independent predictors of the need for PPM. The relative values of the beta coefficients of the independent predictors were then used to devise the risk score. This score was then applied to the remaining one-third of the cohort (the validation group [VG]).
All statistical analysis was performed using SAS Version 8.0 software (Cary, North Carolina). Random selection was performed by random number generation using SAS. Continuous outcomes were analyzed using ttests, and dichotomous variables were analyzed using chi-square tests. Continuous variables are expressed as mean ± SD, and dichotomous variables are expressed as a frequency percentage. A p value of <0.05 was considered to be statistically significant. Multivariate analysis was performed on the PG using stepwise logistic regression with forward selection and backward elimination using a p value of 0.2 for entry, and 0.05 for elimination until the last variable entered was subsequently eliminated. Results are expressed as an odds ratio (OR) with 95% confidence interval (CI).
Baseline patient characteristics are displayed in Table 1. There were no significant differences between the prediction group and the validation group. For the two groups, mean age was 65 years, and nearly half (42%) were female. Approximately 11% of patients had prior valve surgery, 13% of patients underwent multivalve surgery, 56% underwent some form of aortic valve surgery, and 51% of patients underwent some form of mitral valve surgery. Preoperative ECG demonstrated first degree atrioventricular block in 12% of patients, left bundle branch block (LBBB) or right bundle branch block (RBBB) in 8% of patients, and bifasicular block in nearly 2% of patients. Some form of preoperative conduction abnormality was noted in 20% of patients. The PPMs were implanted in 256 of 4,694 patients (5.5%) with equal likelihood of PPM implantation in the PG and VG (5.4% vs. 5.6%, p = NS). For the entire cohort, the indication for permanent pacing was atrioventricular block in 63%, and sick sinus syndrome (including asystole) in the remaining 37%, which is consistent with patient cohorts reported from other institutions (9).
Univariate analysis of PG
Age ≥70 years and prior valve surgery were found to be univariate predictors of permanent pacing after valve surgery, while gender and concomitant coronary artery bypass (CABG) were not predictive of permanent pacing (Table 2). Preoperative ECG predictors of permanent pacing are displayed in Figure 1. Permanent pacing was more common in patients with preoperative first degree atrioventricular block (8.9% vs. 4.9% with a normal PR interval, p = 0.001), LBBB (10% vs. 4.5% in patients with no bundle branch block, p = 0.004), RBBB (not including bifasicular block) (18% vs. 4.7%, p < 0.0001), and bifasicular block (16% vs. 4.5%, p < 0.0001). Rhythm on admission (sinus vs. atrial fibrillation), QRS widening in the absence of bundle branch block, and gender were not predictive of the need for PPM. The type of valve surgery greatly impacted the likelihood of PPM (Fig. 2). Patients undergoing aortic valve surgery had a higher risk of postoperative PPM than those undergoing mitral valve surgery (5.1% vs. 3.5%, p = 0.04). Surgery involving the tricuspid valve appeared to confer a higher risk compared with surgery not involving the tricuspid valve. Multivalve surgery conferred an almost threefold increased risk of PPM, while valve repair was associated with a significantly lower univariate risk of permanent pacing compared with valve replacement.
Multivariate analysis of the PG
Multivariate logistic regression was performed utilizing a model that included age, gender, baseline rhythm, concomitant CABG, and each of the univariate predictors of PPM. Significant independent predictors of permanent pacing after valve surgery in the PG included age ≥70 years, prior valve surgery, PR interval >200 ms, RBBB, LBBB, multivalve surgery with tricuspid valve included, and multivalve surgery not including the tricuspid valve. The beta coefficients and ORs with 95% CI derived from the multivariate regression model are displayed in Table 3. Multivalve surgery involving the tricuspid valve (OR, 3.7; 95% CI, 2.3 to 6.1) and RBBB (OR, 3.6; 95% CI, 2.3 to 5.7) were the strongest independent predictors of the need for PPM. Multivalve surgery not involving the tricuspid valve and baseline LBBB or first degree atrioventricular block each conferred a twofold increased risk. Although mitral valve surgery was a negative univariate predictor of PPM, it was not an independent predictor.
The beta coefficients from the multivariate analysis performed on the PG group were utilized to devise a simple risk score (Table 4). A score of 2 points each was assigned for RBBB or multivalve surgery including the tricuspid valve, and a score of 1 each was assigned for the remainder of the significant independent predictors. The VG (the one-third of the original cohort not randomly selected to be in the risk score group) was then used to assess the accuracy of the risk score in predicting the need for PPM. For patients in the VG with a score of 0, 1, 2, 3, 4, 5, and 6, the incidence of PPM was 1.9%, 5.2%, 8.7%, 12%, 21%, 36%, and 50%, respectively (Fig. 3). No patient had the maximum possible score of 7 points. This scoring system allowed for patients to be divided into low- (score of 0 to 1), moderate- (score of 2 to 3), and high-risk (score ≥4) groups (Fig. 4). The incidence of PPM in these three groups was 3.6%, 9.8%, and 25%, respectively.
The need for PPM after cardiac valve surgery is common. Nevertheless, to our knowledge, this is the first study to demonstrate that a simple risk score incorporating preoperative clinical characteristics, preoperative ECG characteristics, and the type of surgery effectively predicts the need for a PPM. Equally important, the score accurately identifies a large, low-risk subset of valve surgery patients that are unlikely to need a PPM.
The preoperative identification of a high-risk subset of valve surgery patients who may require permanent pacing has important clinical implications. The number and location (ventricular vs. dual-chamber) of temporary epicardial pacing wires implanted at the time of surgery are clinical decisions that have typically been determined by surgeon preference based on clinical experience and the perceived risk of postoperative bradyarrhythmia. Temporary epicardial pacing wires are not completely without risk. Their postoperative placement and removal can cause bleeding, cardiac tamponade, bypass graft injury, or infection (10–13). These risks may be higher with temporary atrial epicardial pacing electrodes. This risk score would allow for a clinical decision to be made in the operating room based on an accurate estimate of the patient’s risk of pacing. In many patients, the absence of temporary epicardial pacing electrodes imposes a risk associated with delayed treatment of the bradyarrhythmia. For example, a patient with a markedly reduced left ventricular ejection fraction at high risk for atrioventricular block may benefit from dual-chamber temporary epicardial electrodes to preserve AV synchrony. Other high-risk patients may benefit from more than one set of ventricular epicardial electrodes to ensure reliable pacing and to avoid the urgent placement of temporary transvenous pacing. In some cases epicardial placement of a permanent system may be appropriate. This is particularly relevant for patients with reduced ejection fraction who may benefit from epicardial left ventricular pacing for hemodynamic reasons and who have detrimental effects of right ventricular apical pacing alone.
On the other hand, the risk score can be utilized to identify a large proportion of valve surgery patients that are low risk. Patients with a risk score of 0 (37% of the cohort) had less than a 2% risk of the need for permanent pacing. A single set of epicardial ventricular electrodes may be a reasonable option for these patients.
The risk score has other important clinical benefits. The score allows patients to be accurately advised preoperatively regarding the likelihood of a PPM. The importance of this information to the patient should not be underestimated, as patients would be expected to have increased confidence in their surgeons and increased satisfaction with their overall care when potential complications are discussed thoroughly and accurately.
A number of specific risk factors were identified as independent predictors of the need for PPM. Among these factors were several indicators of preoperative conduction system delay on the surface ECG, including a PR interval >200 ms, RBBB, and LBBB, which each independently predicted the need for PPM. While prior studies have demonstrated the importance of multivalve surgery and previous valve surgery, this study is the first to utilize multivariable analysis to demonstrate the magnitude of preoperative conduction abnormality as an independent predictor of PPM (1,3). Furthermore, the somewhat surprising finding that RBBB is a more powerful predictor of PPM than LBBB is likely due to a higher rate of injury to the left bundle than the right bundle during aortic or mitral valve surgery, resulting in complete atrioventricular block when RBBB is already present.
The decision to implant a permanent pacemaker after valve surgery in this study was at the physician’s discretion in agreement with the current American College of Cardiology/American Heart Association guidelines for permanent pacing (14). Although potential biases such as a lower threshold for permanent pacing in the elderly and perceived cost-benefit decisions related to length of stay may have influenced this decision, biases such as these are probably not unique to our institution, and they should not make our results any less valid. The mean postoperative day of pacemaker implantation was 8.4 ± 5.8 days (median 7 days) in this study. Recovery of conduction after valve surgery usually occurs by the seventh postoperative day (15), although some studies suggest that up to one-third of patients who receive permanent pacing after cardiac surgery will ultimately not be pacemaker-dependent (9). Nevertheless, prolonged immobilization from temporary pacing impedes patient recuperation and may increase the risk of pneumonia, deep venous thrombosis, and pulmonary embolism. Given these issues of risk and benefit, the timing of pacemaker implantation in this study is probably similar to what other physicians advocate in their practice and in the literature (16).
The risk score could be a useful tool for developing and evaluating new surgical techniques to reduce the incidence of permanent pacing after valve surgery. By identifying high-risk patients before surgery, surgeons may explore alternative approaches to valve surgery.
Finally, previous authors have established methodologic standards for clinical predication rules (17,18). These standards, which include a clearly defined and clinically important outcome, absence of ascertainment bias, and precise definition of possible predictors, are fulfilled by our model. The large cohort size and the large number of patients requiring permanent pacing in this study allowed both the PG and the VG to be drawn from the same cohort. In addition, the rate of permanent pacemaker implantation in our cohort is similar to postvalve surgery permanent pacemaker implantation rates reported elsewhere, and the demographics of the patients in this study are typical for people undergoing cardiac valve surgery. These findings suggest that our results could be extrapolated to patients undergoing cardiac valve surgery at other institutions.
In conclusion, a simple risk score based on preoperative clinical characteristics, preoperative ECG characteristics, and type of surgery accurately predicts the need for permanent pacing after cardiac valve surgery. This score should prove valuable for the perioperative management of valve surgery patients and hopefully will lead to the exploration of alternative approaches to valve surgery in patients at high risk for permanent pacing.
- aortic valve
- coronary artery bypass grafting
- confidence interval
- implanted cardioverter-defibrillator
- left bundle branch block
- odds ratio
- risk score prediction group
- postoperative permanent pacemaker/pacing
- right bundle branch block
- risk score validation group
- Received August 12, 2002.
- Revision received October 17, 2002.
- Accepted October 31, 2002.
- American College of Cardiology Foundation
- ↵Willems JL, Arnaus P, van Bemmel JH, et al. A reference data base for multilead electrocardiographic computer measurement programs. J Am Coll Cardiol 1987;10:1313–21
- Kudenchuk P.J.,
- Ho M.T.,
- Weaver W.D.,
- et al.
- Price C.,
- Keenan D.
- Gregoratos G.,
- Cheitlin M.D.,
- Conill A.,
- et al.