A randomized trial was done to compare single-chamber atrial (AAI) and dual-chamber (DDD) pacing in patients with sick sinus syndrome (SSS). Primary end points were changes in left atrial (LA) size and left ventricular (LV) size and function as measured by M-mode echocardiography. Background In patients with SSS and normal atrioventricular conduction, it is still not clear whether the optimal pacing mode is AAI or DDD pacing.

​

Methods: 

A total of 177 consecutive patients (mean age 74 ± 9 years, 73 men) were randomized to treatment with one of three rate-adaptive (R) pacemakers: AAIR (n = 54), DDDR with a short atrioventricular delay (n = 60) (DDDR-s), or DDDR with a fixed long atrioventricular delay (n = 63) (DDDR-l). Before pacemaker implantation and at each follow-up, M-mode echocardiography was done to measure LA and LV diameters. Left ventricular fractional shortening (LVFS) was calculated. Analysis was on an intention-to-treat basis.

​

Results: 

Mean follow-up was 2.9 ± 1.1 years. In the AAIR group, no significant changes were observed in LA or LV diameters or LVFS from baseline to last follow-up. In both DDDR groups, LA diameter increased significantly (p < 0.05), and in the DDDR-s group, LVFS decreased significantly (p < 0.01). Atrial fibrillation was significantly less common in the AAIR group, 7.4% versus 23.3% in the DDDR-s group versus 17.5% in the DDDR-l group (p = 0.03, log-rank test). Mortality, thromboembolism, and congestive heart failure did not differ between groups.

​

Conclusions ​

During a mean follow-up of 2.9 ± 1.1 years, DDDR pacing causes increased LA diameter, and DDDR pacing with a short atrioventricular delay also causes decreased LVFS. No changes occur in LA or LV diameters or LVFS during AAIR pacing. Atrial fibrillation is significantly less common during AAIR pacing.

Our study is the first randomized trial comparing AAIR and DDDR pacing in patients with SSS and normal AV conduction. The study indicates that long-term DDDR pacing induces LA dilation and, in the case of a high proportion of RV pacing, also reduces LV function. Furthermore, AF is significantly less common during AAIR pacing. These findings support AAIR pacing as the preferred pacing mode in this group of patients. .In the present trial, DDDR pacing with 90% RV pacing induced changes identical to the changes observed in the VVI group in the AAI versus VVI trial (1- 2), with a decrease in LVFS and an increase in LA dilation, and DDDR-l pacing (with a mean 17% of RV pacing) caused only an increase in LA diameter, with no change in LVFS. These findings support that a high proportion of RV pacing causes a decrease in LV function. This is in accordance with recent findings from the randomized MOST (The Mode Selection Trial in Sinus-Node Dysfunction) (11) trial, where increasing proportions of pacing in the RV was associated with increase in the risk of hospitalization for HF (12). .The persistent LA dilation and LVFS decrease in the DDDR-s group after programming to AAI mode (DDDR-s-AAI) indicate that the effects of long-term RV pacing on LV function and LA size also persist after cessation of pacing. The echocardiographic study in AAI mode was, however, done only 5 min after programming to AAI mode, and it is not known whether the LA dilation and LVFS decrease would revert, given a longer period without RV pacing. In dogs, long-term AV synchronous LV pacing has been reported to induce ventricular remodeling with asymmetrical hypertrophy of the LV wall, thinning of the earliest activated free wall, and thickening of the late-activated septum (13- 14). The present results do not answer whether a similar remodeling occurs during long-term RV pacing in humans. .In the present study, LA diameter increased significantly in both DDDR groups but not in the AAIR group. The LA dilation is possibly caused by the abnormal activation sequence and mechanical contraction pattern of the ventricles induced by RV pacing (3- 4,15- 16) associated with a decrease in the LV systolic (3- 4) and diastolic function (15) and an increase in the right atrial pressure and the pulmonary capillary wedge pressure (3,17- 18). The LA dilation was more marked in the DDDR-s group with 90% RV pacing than in the DDDR-l group with 17% RV pacing, supporting an association between these two parameters. The non-significant increase in LA diameter observed in the AAIR group could represent a feature in the natural evolution of the SSS and/or be related to increasing age (19). In the present study, the ventricular lead was implanted in the RV apex. Using RV septal pacing might have influenced the present results (20- 21). .Patients with DDDR pacemakers were programmed with either a short rate-adaptive AV delay or a long fixed AV delay. Optimizing the AV delay individually might have influenced the present results (22). However, the study was designed to evaluate effects of different proportions of ventricular pacing rather than DDDR pacing with an optimized AV delay. Despite programming of a fixed long AV interval, pacing in the RV was reduced to a mean of only 17%. The explanation of this finding probably is RV fusion beats, RV pacing during AF, and RV pacing during different forms of pacemaker tachycardias, as previously documented (22). .When the present study was designed in the early 1990s, two-dimensional echocardiography was expected to yield more precise information of the changes over time in LA and LV dimensions than M-mode echocardiography. The results of our two-dimensional echocardiographic studies do not support the findings done by M-mode echocardiography. The changes observed in left chamber volumes and LVEF during follow-up were all small in size and within the 95% confidence intervals of repeated two-dimensional echocardiographic studies (23- 24). Measuring the LA volume by two-dimensional echocardiography has never found any place in clinical or scientific echocardiography. Furthermore, in contrast with LA diameter measured by M-mode echocardiography (25), LA volume has never been found predictive of later cardiovascular events. .The correlations between mean proportion of RV pacing and the changes in LA and LV diameters during follow-up and the correlations between mean proportion of pacing in the atrium and changes in LA diameter were all non-significant, most likely because of the insufficient accuracy of echocardiography (23,26). .In the present trial, AF was significantly more common in the two DDDR groups, indicating that RV pacing may promote AF, most likely because it causes LA dilation. Similar changes in the echocardiographic parameters were observed after excluding patients who had AF at their echocardiographic examinations, indicating that AF was not the cause of the echocardiographic changes observed. We observed no differences in occurrence of thromboembolism, congestive HF, or death between pacing modes during follow-up, indicating that AAIR and DDDR pacing is similar regarding these outcomes. The limited sample size must, however, be kept in mind when interpreting these data. The incidence of high-degree AV block in the AAIR group was comparable to the annual rate of 1.7% observed in a recent retrospective study of 399 consecutive patients treated with AAI(R) pacing in our institution (27). .At present time, AAI(R) pacing mode seems to be the optimal treatment for isolated SSS. In the future, new mode switching abilities between DDD and AAI pacing may enable AAI pacing the majority of the time and, in addition, protect the patients from severe bradycardia due to AV block. The currently ongoing DANPACE trial (10) is expected to answer whether the risks associated with AAI(R) pacing and AV block are less than the risks associated with RV pacing in DDD(R) mode. .Study limitations .In our previous study, the differences between pacing modes increased markedly during long-term follow-up (2,28). In the present study the mean follow-up was just below three years. The present results cannot be extrapolated beyond this period after pacemaker implantation. Our study was initially designed to include 450 patients, but inclusion was stopped prematurely after randomizing 177 patients, reducing the statistical power. .The echocardiographic measurements were done un-blinded with regard to pacing mode and randomization group. Furthermore, at the ambulatory follow-up visits, echocardiography was done in the AAI mode 5 min after echocardiography in the DDD mode when ventricular pacing had been present in that mode. These factors may have introduced an observer bias. However, results from prior echocardiographic studies of any particular patient in the study were not known at the time of later echocardiographic studies or analyses of echocardiographic data in that patient. .For evaluating the proportions of pacing and sensing in the chambers, we had to rely on telemetered data. Far-field over-sensing of R waves in the atrium and of T waves in the ventricle as well as under-sensing of small atrial electrograms during AF may have influenced these data. To reduce these sense problems, bipolar atrial leads were used in the majority of patients. It is unlikely that the reliability of telemetered data should be different between randomization groups. .Conclusions .Our study is the first randomized trial comparing AAIR and DDDR pacing in patients with SSS and normal AV conduction. The study indicates that long-term DDDR pacing induces LA dilation and, in the case of a high proportion of LV pacing, also reduces LV function. Furthermore, AF is significantly less common during AAIR pacing. These findings support AAIR pacing as the preferred pacing mode in this group of patients.