High incidence of pacemaker syndrome in patients with sinus node dysfunction treated with ventricular-based pacing in the Mode Selection Trial (MOST)

Patients

The principal eligibility criteria for the MOST study included a clinical diagnosis of sick sinus syndrome requiring implantation of a permanent pacemaker. A clinical diagnosis of sick sinus syndrome was defined as one of the following: 1) a documented symptomatic sinus pause ≥3 s; 2) an asymptomatic sinus pause ≥5 s; 3) chronic sinus bradycardia with a heart rate <50 beats/min, accompanied by symptoms of fatigue and dyspnea caused by chronotropic incompetence; or 4) sinus bradycardia with a heart rate <50 beats/min that restricted the use of long-term medical therapy that was deemed necessary for other medical conditions. Patients were eligible if they were age 21 years or older. The patient had to be eligible for a dual-chamber pacemaker utilizing a transvenous approach and to provide written, informed consent. Exclusion criteria included clinically overt congestive heart failure, chronic atrial fibrillation without documented sinus mechanism within the previous six months, malignancy expected to limit the patient's life-span during the course of the trial, severe psychiatric illness that would limit compliance, or inadequate atrial or ventricular capture or sensing thresholds at the time of pacemaker implantation.

Pacemaker implantation

All patients received dual-chamber pacemakers. After adequate atrial and ventricular sensing and threshold parameters were achieved, patients were randomized to either rate-modulated VVIR or dual-chamber (DDDR) pacing. At the time of pacemaker implantation, the implanting physician recorded the presence or absence of retrograde atrial activation during VVIR pacing at both 70 and 100 beats/min, as well as blood pressure during native sinus rhythm and during VVIR pacing at heart rates <15 beats/min greater than the native sinus rhythm. Baseline quality-of-life measures, including the 36-item Medical Outcome Study Short-Form (SF-36) (8), the specific activity scale (9), health utilities using the time tradeoff method (10), and self-reported quality of life on a scale of 0 to 100 (11)were collected before pacemaker implantation. Initial recommended programming of the pacemaker included a lower rate of 60 beats/min, with rate modulation adjusted to a heart rate of 90 to 110 beats/min after a 1-min brisk walk.

Follow-up

Follow-up was performed at one, three, and six months after pacemaker implantation and then every six months. Follow-up data included the interim clinical history, physical examination, pacemaker interrogation, and quality-of-life status. The primary end point of the MOST study was the first occurrence of stroke or death.

A secondary end point and the focus of the current analysis was the incidence of pacemaker syndrome. In the MOST study, pacemaker syndrome was defined by: 1) new or worsened dyspnea, orthopnea, elevated jugular venous pressure, rales, and edema with ventriculoatrial (VA) conduction during ventricular pacing; or 2) symptoms of dizziness, weakness, presyncope or syncope, and a >20 mm Hg reduction of systolic blood pressure when the patient had VVIR pacing compared with atrial pacing or sinus rhythm (performed during follow-up evaluation).

When pacemaker syndrome was diagnosed at the clinical site, investigators were required to discuss the case with the Clinical Coordinating Center. Before crossover from VVIR to DDDR, investigators were encouraged to pace the ventricle at a lower rate and/or to decrease the aggressiveness of the sensor-driven rate. If symptoms did not resolve, they were required to call the principal investigator of the study, who reviewed the indications for crossover. Before and after crossover, patients were evaluated using the SF-36, the specific activity scale, the time tradeoff approach, and the self-reported 0 to 100 scale.

Statistical methods

The association of baseline and implant variables and discharge medications with the risk of pacemaker syndrome was assessed using a single-predictor Cox proportional hazards model for each variable. The model for heart rate contained a linear spline, which allowed a different risk relationship below and above 55 beats/min. For baseline and implant variables, all 996 VVIR patients were included. For discharge medications, all patients who did not meet the criteria for pacemaker syndrome in the hospital were included (n = 961). The Kaplan-Meier methods were used to estimate six-month event rates for subgroups defined by baseline and implant variables.

Because analysis of variables that are measured repeatedly over time is inherently complex, descriptive statistics and formal hypothesis tests for the association of follow-up variables with pacemaker syndrome risk employed two different approaches. For descriptive statistics, follow-up data from before the onset of pacemaker syndrome were used for patients who developed it; for patients who did not develop pacemaker syndrome, follow-up data from the first 196 days were used, as this was the third quartile of time to pacemaker syndrome and represented the time span during which patients were at greatest risk of developing pacemaker syndrome. Variables were defined for each patient as follows: mean value of reported intrinsic heart rates, total percentage of beats that were VVIR paced, total percentage of beats paced at 90 beats/min or higher, programmed low rate at which the patient spent the most time, and presence of VA conduction reported at any time. The association of these variables with risk of pacemaker syndrome was formally assessed in patients who did not develop pacemaker syndrome in the hospital and whose follow-up included at least one visit while in VVIR pacing, using separate Cox proportional hazards models with the variable of interest entered as a time-dependent co-variate. These models allow patients' values for these co-variates to change over the course of the follow-up period. All follow-up data were used in these models. Approximately 800 patients were included in this analysis, but the number differed by variable, depending on the extent of missing data. Assessments for VA conduction after randomization were performed only in patients who were suspected of having pacemaker syndrome; because of the bias in this sample, summary statistics, but no significance tests, are presented for VA conduction during follow-up.

Baseline and follow-up variables that were significant in univariate tests were entered into a multivariable Cox proportional hazards model. A backward selection procedure was used to identify independent predictors.

Changes in quality of life before and after crossover were tested utilizing the paired nonparametric Wilcoxon signed-rank test. These analyses were restricted to patients who crossed over during the first year of follow-up and who had questionnaires obtained at baseline, before crossover, and after crossover.

Patient characteristics

There were 2,010 patients with sinus node dysfunction randomized, of whom 996 were assigned to single-chamber, rate-modulated VVIR pacing. Of those randomized to VVIR pacing, 477 (48%) were female, and the median age was 74 years. New York Heart Association functional class III or IV heart failure was present at baseline in 145 patients (15%). Of the 996 patients, 209 (21%) had concurrent AV block and 52 (5%) had a history of complete heart block.

Pacemaker syndrome

Among VVIR patients, 182 (18.3%) met criteria for pacemaker syndrome, as defined by the study protocol. Forty-one patients met definition 1 only, 64 met definition 2 only, and 77 met definitions 1 and 2, as discussed earlier. By life-table analysis, the incidence of pacemaker syndrome was 13.8% at six months, 16.0% at one year, 17.7% at two years, 19.0% at three years, and 19.7% at four years (Fig. 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Kaplan-Meier graph showing the time from implantation to time of permanent crossover for pacemaker syndrome. The y-axis is truncated at 0.7.

Baseline and implant data

The incidence of pacemaker syndrome was similar regardless of the criterion by which a patient was defined as having sick sinus syndrome. However, a slower sinus rate at the time of implantation was associated with the development of pacemaker syndrome. The hazard ratio (HR) for developing pacemaker syndrome was 1.11 (95% confidence interval [CI] 1.04 to 1.19, p < 0.01) for each 5-beat/min decrease in patients with baseline heart rates above 55 beats/min. For heart rates below 55 beats/min, further reductions in sinus rates were not predictive of the development of the pacemaker syndrome (p = 0.9). A higher programmed ventricular rate was associated with the development of pacemaker syndrome, with a HR of 1.15 (95% CI 1.04 to 1.26) for each 5-beat/min increase in the programmed rate (p = 0.005). No other patient characteristic before or at implantation, including age, gender, left ventricular ejection fraction, presence or absence of VA conduction, systolic blood pressure drop with VVIR pacing at implantation, and duration of QRS or PR intervals, was associated with the development of pacemaker syndrome (Table 1).

Discharge medications

There was no association between use of beta-blockers, calcium antagonists, or amiodarone at hospital discharge and the risk of subsequent pacemaker syndrome.

Follow-up data

The risk of pacemaker syndrome was significantly higher in patients with a higher percentage of ventricular paced beats (p = 0.0001; HR 1.22 [95% CI 1.16 to 1.28] for each 10% increase), a higher programmed, lower pacemaker rate (p = 0.0001; HR 1.27 [95% CI 1.16 to 1.38] for 5-beat/min increase), and a slower underlying sinus heart rate (p = 0.0005; HR 1.09 [95% CI 1.04 to 1.15] for 5-beat/min decrease) (Table 2). The presence of chronotropic incompetence, as documented by having fewer native beats at a rate >90 beats/min, showed a trend toward increased risk of pacemaker syndrome (p = 0.060, HR 1.07 [95% CI 0.99 to 1.14] for a 5% decrease).

Multivariate model

Of the univariate predictors of pacemaker syndrome (baseline and follow-up sinus rate, programmed low rate, percentage of beats paced, and percentage of beats paced at >90 beats/min were entered into the multivariate model), only a higher percentage beats paced was an independent predictor (p = 0.0001; HR 1.22 [95% CI 1.16 to 1.28] for each 10% increase). A programmed lower rate, the second most important factor, made a marginal and nonsignificant contribution to the predictive model (p = 0.073; HR 1.11 [95% CI 0.99 to 1.23] for a 5-beat/min increase).

Quality of life

At the time of diagnosis of pacemaker syndrome, quality of life was lower than that at the time of implantation for nine of the 10 SF-36 scales and was significantly worse for six of these scales (data based on 80 patients) (Table 3). In addition, specific activity scale class (p < 0.001) and self-reported health on a scale of 0 to 100 (p = 0.03) worsened significantly, compared with their values at the time of implantation. After crossover to physiologic pacing, 9 of the 10 SF-36 scales improved, and 6 improved significantly, as did scores on the specific activity scale (p = 0.02), the time tradeoff score (p = 0.01), and the self-reported 0 to 100 scale (p < 0.001; data based on 153 patients). The most striking improvements were in energy and physical role, both of which improved to become substantially better than they were even before pacemaker implantation.