To the Editor: Ryan et al. report on the efficacy of Dual Chamber
pacemaker implantation in patients with cardio-inhibitory carotid sinus
hypersensitivity (CICSH) and falls in SAFEPACE-2 (Syncope And Falls in the
Elderly- Pacing And Carotid sinus Evaluation)[1]. This study had the
potential to expand on the findings of SAFEPACE-1 [2] and address some of
the weaknesses of the latter with a blinded, multicenter, and less select
cohort [3]. However, despite improved external validity, SAFEPACE-2
reports findings at odds with the prior results of SAFEPACE-1 by
demonstrating that the null hypothesis was correct (pacemaker implantation
did not reduce falls in CICSH)[1].
The authors of SAFEPACE-2 are to be applauded for their efforts to
increase the generalizability of their results and for their attention to
detail with regards to data acquisition and analysis. This begs the
question, why the marked difference in findings between the two SAFEPACE
trials [1-2]? While it is true that the cohort in SAFEPACE-2 had more
comorbidity, one wonders whether further review of the trial methodology
will yield another explanation.
The ‘double-blind’ design of SAFEPACE-2 seems difficult to achieve
given the nature of the intervention. Pacemaker insertion requires the
patient to carry a specific pacemaker card and to follow-up at specialized
pacemaker clinics. In contrast, a loop recorder requires specific
education as the patient self-activates the recorder with a hand-held
device. Failure to blind the patients could significantly affect the
primary outcome of self-reported falls, increasing the potential for
recall bias in favor of the intervention. Therefore, this methodological
issue may actually strengthen the author’s findings in SAFEPACE 2 and does
not explain the difference in results between the trials.
At first glance, the difference may be explained by the lack of
statistical power in SAFEPACE-2. It is true that SAFEPACE 2 is
underpowered to detect a difference of 20% in the number of patients who
fell, however it is almost fully powered to detect a difference of 40%.
This was the power criteria used in SAFEPACE 1 [2]. Therefore, lack of
statistical power also does not fully explain the difference in results
between the trials.
One methodological issue which may account for the different trial
results is the ‘randomization to implantation’ time. The overall relative
risk (RR) for falls with pacemaker implantation in SAFEPACE-2 was 0.79
(95% Confidence Interval [CI], 0.4-1.5), but was 0.23 (95% CI, 0.15-0.32)
when comparing patients before and after device implantation [1]. The
latter RR suggests that device implantation made a difference to fall
rates and raises the possibility that the overall result was obscured by a
prolonged ‘randomization to implantation’ time. Given the multicenter
nature of this trial it seems likely that this time was longer in SAFEPACE
-2 than the single centre SAFEPACE-1. Therefore, it seems possible that
inclusion of fall data prior to device implantation (the intervention)
obscured the result of SAFEPACE-2 and may partially explain the difference
in results between the two SAFEPACE trials.
Reporting the mean and median time from ‘randomization to
implantation’, and exclusion of the pre-implantation fall data from the
analysis may help shed more light on this important issue.
John W. McEvoy M.B., M.R.C.P.I.
The Johns Hopkins Hospital
600 N. Wolfe Street
Baltimore, Maryland 21287
Jmcevoy1@jhmi.edu
1. Ryan D, Steen N, Seifer C, Kenny RA. Carotid Sinus Syndrome and
falls, should we pace? A multi-centre, randomised control trial (Safepace
2). Heart 2009.
2. Kenny RA, Richardson DA, Steen N, Bexton RS, Shaw FE, Bond J. Carotid
sinus syndrome: a modifiable risk factor for nonaccidental falls in older
adults (SAFE PACE). J Am Coll Cardiol 2001;38(5):1491-6.
3. McAnulty JH. Carotid sinus massage in patients who fall: will it define
the role of pacing? J Am Coll Cardiol 2001;38(5):1497.
To the Editor: Ryan et al. report on the efficacy of Dual Chamber pacemaker implantation in patients with cardio-inhibitory carotid sinus hypersensitivity (CICSH) and falls in SAFEPACE-2 (Syncope And Falls in the Elderly- Pacing And Carotid sinus Evaluation)[1]. This study had the potential to expand on the findings of SAFEPACE-1 [2] and address some of the weaknesses of the latter with a blinded, multicenter, and less select cohort [3]. However, despite improved external validity, SAFEPACE-2 reports findings at odds with the prior results of SAFEPACE-1 by demonstrating that the null hypothesis was correct (pacemaker implantation did not reduce falls in CICSH)[1].
The authors of SAFEPACE-2 are to be applauded for their efforts to increase the generalizability of their results and for their attention to detail with regards to data acquisition and analysis. This begs the question, why the marked difference in findings between the two SAFEPACE trials [1-2]? While it is true that the cohort in SAFEPACE-2 had more comorbidity, one wonders whether further review of the trial methodology will yield another explanation.
The ‘double-blind’ design of SAFEPACE-2 seems difficult to achieve given the nature of the intervention. Pacemaker insertion requires the patient to carry a specific pacemaker card and to follow-up at specialized pacemaker clinics. In contrast, a loop recorder requires specific education as the patient self-activates the recorder with a hand-held device. Failure to blind the patients could significantly affect the primary outcome of self-reported falls, increasing the potential for recall bias in favor of the intervention. Therefore, this methodological issue may actually strengthen the author’s findings in SAFEPACE 2 and does not explain the difference in results between the trials.
At first glance, the difference may be explained by the lack of statistical power in SAFEPACE-2. It is true that SAFEPACE 2 is underpowered to detect a difference of 20% in the number of patients who fell, however it is almost fully powered to detect a difference of 40%. This was the power criteria used in SAFEPACE 1 [2]. Therefore, lack of statistical power also does not fully explain the difference in results between the trials.
One methodological issue which may account for the different trial results is the ‘randomization to implantation’ time. The overall relative risk (RR) for falls with pacemaker implantation in SAFEPACE-2 was 0.79 (95% Confidence Interval [CI], 0.4-1.5), but was 0.23 (95% CI, 0.15-0.32) when comparing patients before and after device implantation [1]. The latter RR suggests that device implantation made a difference to fall rates and raises the possibility that the overall result was obscured by a prolonged ‘randomization to implantation’ time. Given the multicenter nature of this trial it seems likely that this time was longer in SAFEPACE -2 than the single centre SAFEPACE-1. Therefore, it seems possible that inclusion of fall data prior to device implantation (the intervention) obscured the result of SAFEPACE-2 and may partially explain the difference in results between the two SAFEPACE trials.
Reporting the mean and median time from ‘randomization to implantation’, and exclusion of the pre-implantation fall data from the analysis may help shed more light on this important issue.
John W. McEvoy M.B., M.R.C.P.I. The Johns Hopkins Hospital 600 N. Wolfe Street Baltimore, Maryland 21287 Jmcevoy1@jhmi.edu
1. Ryan D, Steen N, Seifer C, Kenny RA. Carotid Sinus Syndrome and falls, should we pace? A multi-centre, randomised control trial (Safepace 2). Heart 2009. 2. Kenny RA, Richardson DA, Steen N, Bexton RS, Shaw FE, Bond J. Carotid sinus syndrome: a modifiable risk factor for nonaccidental falls in older adults (SAFE PACE). J Am Coll Cardiol 2001;38(5):1491-6. 3. McAnulty JH. Carotid sinus massage in patients who fall: will it define the role of pacing? J Am Coll Cardiol 2001;38(5):1497.