Author + information
- Yu Jin, MD, PhD,
- Lutgarde Thijs, MSc,
- Alexandre Persu, PhD,
- Sverre Kjeldsen, MD and
- Jan A. Staessen, MD, PhD∗ ( )()
- ↵∗Studies Coordinating Centre, Research Unit Hypertension and Cardiovascular Epidemiology, KU Leuven Department of Cardiovascular Sciences, University of Leuven, Campus Sint Rafaël, Kapucijnenvoer 35, Block D, Box 7001, BE 3000 Leuven, Belgium
We read with interest the recent meta-analysis by Davis et al. (1) on the blood pressure (BP)–lowering effects of renal denervation (RDN) in treatment-resistant hypertensive patients. The authors computed pooled statistics for 3 so-called controlled studies and 9 observational studies. They concluded that the 6-month reductions in systolic/diastolic BP in response to RDN averaged 28.9/11.0 mm Hg and 25.0/10.0 mm Hg in the controlled and observational studies, respectively. A number of issues need clarification.
First, Davis et al. (1) used the term “controlled” to group 2 randomized trials (2,3) and 1 observational study (4). It is questionable whether truly randomized trials and a nonrandomized study can be combined into a single group. Moreover, 1 of the randomized studies was of low quality (2). Sample size and primary and secondary endpoints changed across successive design protocols posted at http://www.clinicaltrials.gov up to 1 year after the paper was published (2). We believe that this study (2) should have been excluded based on CONSORT quality standards (5). Furthermore, the spread of the BP responses was not reported for the control group in SYMPLICITY HTN-1 (4). In their Figure 2 (1), the numerical data for the 6-month systolic responses are switched between 2 studies. Moreover, there is discordance between the SDs reported in their Figure 2 (systolic/diastolic responses at 6 months, SD 9.8/6.0 mm Hg) and the spread of the BP responses in Figure 5 (4.8/3.0 mm Hg at 6 months) and in the text (5.0/2.0 mm Hg at 12 months) of the paper by Pokushalov et al. (2). Without clarification, these observations invalidate the pooled estimates for the controlled studies in the meta-analysis by Davis et al. (1) because they were weighted for the inverse of the variance in individual studies.
Second, turning to the 8 uncontrolled studies with 6-month data reported in Figure 3 of their meta-analysis (1), 6 had a very small sample size, with follow-up data available in <20 patients (range: 8 to 20 patients). Small studies with positive results are more readily reported than larger studies with negative results. Davis et al. (1) set a follow-up rate of <70%, not >70% as stated in their paper, as the limit to determine a high risk of bias. In fact, the 6-month follow-up rate in 4 small studies with 10 to 20 participants was 100%, but in all other studies, it ranged from 53% to 66%. The pooled estimates of the 6-month BP responses in the uncontrolled group were based on office BP in 6 studies and 24-h ambulatory monitoring in 2 studies (6,7). Results based on different techniques of BP measurement cannot be pooled in a single summary statistic.
One wonders how a meta-analysis including only a single randomized trial of sufficient quality (3) adds to current knowledge. Davis et al. (1) concluded that RDN resulted in a substantial BP reduction at 6 months. In our view, this conclusion does not hold in view of the above issues. We believe that the main conclusion of this meta-analysis should have been that currently the evidence in favor of RDN in treatment-resistant hypertension is of very low quality, that a meta-analysis cannot replace properly-sized randomized clinical trials, and that RDN therefore should not be applied in routine clinical practice.
Please note: Dr. Kjeldsen has received honoraria for lecturing from AstraZeneca, Bayer, Medtronic, Merck, Sharp & Dohme, and Takeda; consulting honoraria from Bayer, Medtronic, Takeda, and Serodus; and grant support from AstraZeneca. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
Special Presentation of Letter and Response:
A Letter to the Editor has been submitted to the Journal by Dr. Jan A. Staessen and colleagues regarding the paper: Pokushalov E, Romanov A, Corbucci G, et al. A randomized comparison of pulmonary vein isolation with versus without concomitant renal artery denervation in patients with refractory symptomatic atrial fibrillation and resistant hypertension. J Am Coll Cardiol 2012;60:1163–70. Due to the nature, detail, and length of the letter and the corresponding reply, the editors have made a decision to waive the usual word limit and timelines for letters and replies so as to provide a full airing of the issues for our readers. However, given the length of the material, we have elected to place the letter and reply online (onlinejacc.org). We encourage readers to access this material and trust that it contributes to the editorial process.
- 2013 American College of Cardiology Foundation
- Davis M.I.,
- Filion K.B.,
- Zhang D.,
- et al.
- Pokushalov E.,
- Romanov A.,
- Corbucci G.,
- et al.
- Ahmed H.,
- Neuzil P.,
- Skoda J.,
- et al.