Author + information
- Andrea Messori, PharmD∗ ( )( and )
- Sabrina Trippoli, PharmD
- ↵∗ESTAV Centro, Area Vasta Centro Toscana, Regional Health System, Via Guimaraes 9-11, 59100 Prato, Italy
In evaluating the effectiveness of the Watchman appendage occlusion device (AOD) in patients with atrial fibrillation, the main strength of the study by Holmes et al. (1) is represented by the large population of patients included in the analysis (which tends to approximate the totality of evidence). The authors necessarily designed their study of effectiveness (endpoint: stroke or systemic embolism) by comparing AOD with warfarin because this comparison reflected the available clinical material or, at least, that resulting from the randomized trials. Consequently, the results expressing the comparison between AOD and warfarin is the key message conveyed by their study.
In discussing their results, however, Holmes et al. (1) pointed out that novel oral anticoagulants (NOACs) also currently represent a treatment option indicated for these patients, but no data of (direct) clinical comparison are available for AOD versus NOACs, and so this question remains open.
Despite their well-known limitations, indirect comparisons can be useful to explore a therapeutic question on which no direct comparative data are available (2,3). In 2014, we carried out one such indirect comparison between AOD and NOACs (3), at a time when the published population of patients treated with Watchman was relatively small (data from 1,107 patients published by Bajaj et al. ).
The study by Holmes et al. (1) now offers a much larger population (n = 2,406) for performing this indirect comparison. Hence, in the analysis presented herein, we have updated our previous results published in 2014 (3), and we have redetermined this indirect comparison between AOD and NOACs by including the data of the patient-level meta-analysis by Holmes et al. (1).
Figure 1 shows the results of this re-analysis (the figure includes a noninferiority margin that reflects the same margin adopted in the randomized trials ). Firstly, it appears that the effectiveness data published in 2014 for Watchman (event rate = 0.99% ) were more favorable to the device than those published by Holmes et al. (1) (event rate = 2.72%). Despite this, the overall message resulting from this noninferiority analysis is that for both datasets, the effectiveness of Watchman fully satisfies the noninferiority criterion. In fact, the upper limit of the confidence interval (CI) of Watchman (see Figure 1) remains well within the “right” noninferiority margin (i.e., remains “on the left” of the margin without intercepting the margin itself), thus satisfying the noninferiority criterion (i.e., AOD is noninferior to NOACs). This finding represents the main result of our reanalysis.
However, it should also be noted that in Figure 1, the right margin of the CI for AOD also remained on the left of the line of identity (although to a minimal extent); this suggests that the superiority of AOD versus NOACs could also be the conclusion from these data.
It should be pointed out that these indirect analyses are largely speculative, and so caution should be exercised in drawing any conclusion from these data. Despite this, the (indirect) evidence resulting from the present reanalysis does support the hypothesis that AOD is noninferior to NOACs.
Please note: The authors have carried out this study in the context of their activity at the above mentioned institution; ESTAV Centro belongs to the Italian national health system. Both authors have reported that they have no relationships relevant to the contents of this paper to disclose. Noel Boyle, MD, PhD, served as Guest Editor for this paper.
- 2015 American College of Cardiology Foundation
- Holmes D.R. Jr..,
- Doshi S.K.,
- Kar S.,
- et al.
- Bajaj N.S.,
- Parashar A.,
- Agarwal S.,
- et al.
- Messori A.,
- Fadda V.,
- Maratea D.,
- Trippoli S.
- ↵Fadda V. Meta-analysis on novel oral anticoagulants vs warfarin for stroke prevention: re-expressing the results as risk difference (Comment). PubMed Commons. Available at: http://www.ncbi.nlm.nih.gov/pubmed/24315724#cm24315724_3448. Accessed September 1, 2015.