Advertisement

Regarding “Inherent problems with randomized clinical trials with observational/no treatment arms”

      The unusual conclusions in the article by Buckley et al
      • Buckley C.J.
      • Rutherford R.B.
      • Diethrich E.B.
      • Buckley S.D.
      Inherent problems with randomized clinical trials with observational/no treatment arms.
      that life-threatening diseases like abdominal aortic aneurysm (AAA) “should not be evaluated using RCTs containing an observation/no treatment arm” and that “intent-to-treat analysis … should not be used in this situation” were based on the authors' belief that high crossover rates are an “inherent weakness” of these studies. This belief stems from a misunderstanding of the term “crossover.” A crossover occurs when a patient receives the alternative intervention rather than the randomly assigned intervention, and represents a serious protocol violation which, if sufficiently frequent, can obscure the trial results. The Aneurysm Detection and Management (ADAM) trial
      • Lederle F.A.
      • Wilson S.E.
      • Johnson G.R.
      • Reinke D.B.
      • Littooy F.N.
      • Acher C.W.
      • et al.
      Aneurysm Detection and Management Veterans Affairs Cooperative Study Group Immediate repair compared with surveillance of small abdominal aortic aneurysms.
      compared two strategies in patients with an AAA 4.0 to 5.5 cm: immediate open repair vs surveillance with repair of AAA that enlarged to 5.5 cm or greater. Patients in the surveillance group who had repair when the AAA enlarged to 5.5 cm or greater were not crossovers, they were treated according to protocol and, therefore, contribute to the valid comparison of the outcomes of the two strategies. The statement by Buckley et al regarding the ADAM trial that “61.6% of those randomized to ultrasound surveillance crossed over to open repair” is incorrect. As reported in the text and Figure 1 of the ADAM manuscript, less than 10% of the surveillance group had repair off-protocol and represent crossovers.
      Apart from Buckley et al, there is near universal agreement that randomized trials remain “the gold standard in evaluating healthcare interventions” and that intent-to-treat analysis remains an essential aspect of their validity.
      • Schulz K.F.
      • Altman D.G.
      • Moher D.
      CONSORT Group
      CONSORT 2010 statement: updated guidelines for reporting parallel group randomised trials.
      Randomized trials established the benefit of coronary artery bypass surgery
      Eighteen-year follow-up in the Veterans Affairs Cooperative Study of Coronary Artery Bypass Surgery for stable angina The VA Coronary Artery Bypass Surgery Cooperative Study Group.
      and carotid endarterectomy,
      Beneficial effect of carotid endarterectomy in symptomatic patients with high-grade carotid stenosis North American Symptomatic Carotid Endarterectomy Trial Collaborators.
      and there is no “inherent” reason why AAA repair should be studied any differently.

      References

        • Buckley C.J.
        • Rutherford R.B.
        • Diethrich E.B.
        • Buckley S.D.
        Inherent problems with randomized clinical trials with observational/no treatment arms.
        J Vasc Surg. 2010; 52: 237-241
        • Lederle F.A.
        • Wilson S.E.
        • Johnson G.R.
        • Reinke D.B.
        • Littooy F.N.
        • Acher C.W.
        • et al.
        Aneurysm Detection and Management Veterans Affairs Cooperative Study Group.
        N Engl J Med. 2002; 346: 1437-1444
        • Schulz K.F.
        • Altman D.G.
        • Moher D.
        • CONSORT Group
        CONSORT 2010 statement: updated guidelines for reporting parallel group randomised trials.
        BMJ. 2010; 340: c332
      1. Eighteen-year follow-up in the Veterans Affairs Cooperative Study of Coronary Artery Bypass Surgery for stable angina.
        Circulation. 1992; 86 ([No authors listed]): 121-130
      2. Beneficial effect of carotid endarterectomy in symptomatic patients with high-grade carotid stenosis.
        N Engl J Med. 1991; 325 ([No authors listed]): 445-453

      Linked Article

      • Inherent problems with randomized clinical trials with observational/no treatment arms
        Journal of Vascular SurgeryVol. 52Issue 1
        • Preview
          Randomized clinical trials (RCTs) offering an observation/no treatment (OBS/NoRx) arm as control and which are focused on the management of a condition with potentially life-threatening consequences, however small the risk, often experience a significant rate of crossover to treatment by those randomized to the OBS/NoRx arm. Results of these trials when analyzed on intent-to-treat basis often fail to resolve the issue at which they were directed. The authors have observed this in trials of abdominal aortic aneurysms with this design and use these to exemplify the dilemmas RCTs of such design create, with crossovers ranging from 27% to over 60% (EVAR II, UKSAT, ADAM, PIVOTAL).
        • Full-Text
        • PDF
        Open Archive
      • Reply
        Journal of Vascular SurgeryVol. 53Issue 2
        • Preview
          The primary purpose of the authors' article was to identify potential problems, which can occur when large numbers of patients in randomized abdominal aortic aneurysm (AAA) clinical trials cross over from no treatment to treatment. We feel these cross overs obfuscate some of the trial's conclusions. Despite comments to the contrary, the predominant reasons for trial participants to cross over from no treatment to treatment was either aneurysm becoming symptomatic, a very subjective event, or the patients' request for treatment because of concerns related to their aneurysm.
        • Full-Text
        • PDF
        Open Archive