Testimony of Peter Lurie, M.D., M.P.H., and Eunice Yu
Public Citizen’s Health Research Group
Public Citizen’s Health Research Group opposes the approval of REPEL-CV, a resorbable polymer adhesion barrier, for the proposed indication of reduction in surgical adhesions. Minimum criteria for the approval of an adhesion barrier should be the demonstration of a clinically significant improvement and a reasonable assurance of safety. Neither is present in this case.
Opening the sternum (sternotomy) during cardiac surgery can result in the formation of dense scar tissue called adhesions between the outside of the heart (epicardium) and the inner portion of the chest wall. REPEL-CV is implanted between the epicardium and the chest wall and, in theory, resorbed within 28 days. During this period, it is supposed to reduce the incidence and severity of adhesions, making subsequent surgery less difficult. It is noteworthy that the sole proposed indication (“reducing the incidence, severity and extent of post-operative adhesion formation in patients undergoing cardiac surgery via sternotomy”) makes no claim of any clinical benefit to the patient.
In its pre-market application (PMA) for REPEL-CV, SyntheMed Inc. submitted a single randomized, evaluator-blinded pivotal trial of 142 pediatric patients (73 treatment, 69 control) who were expected to require at least two sternotomy procedures for repair of congenital heart malformations. The device was implanted in the first surgery and adhesions were measured in the second. In the treatment group, REPEL-CV was sutured to the margins of the open pericardium (a sac surrounding the heart) below the sternum in the first operation, while in the control group the pericardium was left open. Following various patient withdrawals and discontinuations, the trial yielded an intent-to-treat population of 56 patients in the treatment arm and 54 in the control arm, all of whom underwent the second procedure. Most study endpoints were based upon the following four-point scale developed in the feasibility trials.
Grade 0 = No adhesions
Grade 1 = Mild adhesions (filmy, non-cohesive adhesions requiring blunt dissection to separate the space between the epicardium and sternum)
Grade 2 = Moderate adhesions (filmy, non-cohesive adhesions requiring a combination of blunt and selective sharp dissection to separate the space between the epicardium and the sternum)
Grade 3 = Severe adhesions (dense, cohesive adhesions requiring extensive sharp dissection to separate the space between the epicardium and the sternum)
The primary endpoint was the percentage of the surgical site with Grade 3 adhesions detected during the second surgery. Three of the four secondary endpoints also drew from this scale by measuring the prevalence of Grade 0-2 adhesions, or by characterizing the patients by their most severe adhesion grade. The final secondary endpoint measured the time required to dissect adhesions at the second sternotomy.
REPEL-CV Does Not Reduce the Incidence of Adhesions
The trial showed a significant reduction in the prevalence at second sternotomy of Grade 3 adhesions in patients with the REPEL-CV implant (21.3%) compared with those whose pericardiums were left open (47.3%; p=0.0008). However, the product failed to actually prevent adhesions overall, whether measured by the percentage of the operative surface area with Grade 0 adhesions (mean 2.9% vs. 0.9%; p=0.32) or by the percentage of patients who were completely adhesion-free (1.8% vs. 0%). The product did seem to reduce the severity of adhesions. For the primary efficacy variable, there was an overall redistribution from Grade 3 adhesions into Grade 2 and Grade 1 adhesions. The percentage of patients characterized by their worst adhesion showed a similar trend: REPEL-CV resulted in a “one-grade shift downwards,” from Grade 3 into Grade 2. This is very different from “reducing the incidence” of adhesions, part of the indication sought by the company.
Lack of Clinical Endpoints
The above adhesion scale has never been validated with clinical outcomes such as mortality, infection, or complications of adhesions. Indeed, we are aware of no circumstance in which it has been used except for in the development of REPEL-CV. One observer suggests that pericardial adhesions “may be beneficial” for patients because adhesions prevent “excessive movement of the heart devoid of its normal pericardial support.” Lack of clarity over the significance of adhesions makes the demonstration of an actual clinical benefit all the more important.
Given that the apparent purpose of adhesion prevention is the facilitation of follow-up surgeries, the only (secondary) endpoint with a resemblance to clinical significance is adhesion dissection time. Yet this endpoint was not influenced by REPEL-CV (25.9 minutes for the treatment group vs. 25.0 minutes for the control group; p=0.84). The lack of a positive result for the only clinical endpoint calls into question the usefulness of the device.
The major limitations of using adhesions as the primary outcome are underscored by the history of FDA’s guidance on clinical trials for resorbable adhesion barrier devices. In its original 1999 Draft Guidance for such devices in abdominal and/or pelvic surgery, the agency made clear the importance of clinical endpoints: “Optimally, endpoints should directly address clinical outcome measures … The most direct method of providing valid scientific evidence of effectiveness is to select an appropriate clinical endpoint(s) and design a study that may demonstrate a statistically significant and clinically meaningful effect on recognized adhesion-related morbidity.” After discussing particular endpoints for abdominal and pelvic surgery and acknowledging possible impediments to the use of clinical endpoints, the section concludes, “sponsors are encouraged to directly assess clinical endpoints whenever possible.”
This did not sit well with the Adhesion Barrier Task Force, which represented the manufacturers of adhesion barriers including SyntheMed’s predecessor company, Life Medical Sciences. In comments on the Draft Guidance submitted to the FDA, the Task Force declared that, “Until there is more information and standards established for conducting these studies with highly specific clinical endpoints, it would be overly burdensome to suggest that measuring specific clinical outcomes might be the means of assessing product effectiveness as this has never been accomplished to date.”
The objections evidently had the desired effect. In the final Guidance, after stating that clinical outcomes are “the most direct method of providing valid scientific evidence of effectiveness,” the FDA concludes, “The clinical outcomes associated with adhesions may be reasonably assessed by parameters which are more immediately measurable and potentially less confounded.” At least three of five examples of appropriate outcomes given in the final Guidance mention only adhesions; the other two are unclear.
While the study design rightly emphasized blinding of the evaluating surgeon at second sternotomy, this design was undermined. Although REPEL-CV should be resorbed within 28 days, “implanted test material or a fibrous capsule, or other abnormal tissue” was observed in 30.4% of patients in the REPEL-CV group and 1.9% of patients in the control group at second sternotomy (p<0.0001). Finding this material during the second sternotomy would likely unblind the evaluator, potentially leading to downgrading of adhesion severity in the REPEL-CV group.
In addition to these efficacy concerns, there were troubling signs of possible dangers associated with the device. There were trends toward higher risk of death (16.4% vs. 13.0%), mediastinial infection (5.5% vs. 1.4%, or 4 vs. 1 patient), and adverse events possibly, probably or definitely related to the study (8.2% vs. 1.4%). These trends were not statistically significant, but with a total population of 142 patients who underwent the first sternotomy, the study was only, by the sponsor’s own admission, “adequate to rule out a 18% disadvantage (15% [mortality rate in the control group] vs. 33% [mortality rate in the treatment group], 2.8 odds ratio) with 80% power and one-sided 5% Type I error.” Thus, REPEL-CV-treated patients would have had to have died at almost three times the control rate for this study to have reached statistical significance. The consistent direction of the adverse effects observed, even if non-significant, is concerning.
Parallels with Intergel
The data on REPEL-CV bring to mind the case of Intergel, a product made from sodium hyaluronate and intended to reduce pelvic adhesions. In that case, adhesions were also shown to be reduced by the product. However, despite the use of an adhesion scale with arguably more validation than that used in the REPEL-CV study, FDA was concerned that, “There is little experience in the clinical literature correlating the [Modified American Fertility Society] score with clinical outcomes.” The pivotal Intergel trial demonstrated, as here, a consistent but non-statistically significant increase in infection rates. Initially, the FDA rejected the sponsor’s application, but the company appealed to an external Dispute Resolution Panel, which recommended approval. The FDA then reversed itself and approved the device. On April 16, 2003, less than two years after the device was approved, the company removed the device from the market due to dozens of reports of post-operative pain requiring repeat surgery, foreign body reactions and tissue adherence, including three deaths. This history should give one pause before approving an adhesion barrier with only surrogate endpoints and a questionable safety record.
SyntheMed has simply failed to demonstrate that its product will have any important impact upon the public health. To do so, the following conditions would have to be met:
- The patients receiving the device would have to undergo resternotomy; in fact, only a minority of patients will undergo resternotomy and all implanted patients face the potential dangers of the device.
- The product would have to reduce adhesions; in fact, the product reduces the severity but not the incidence of adhesions.
- The adhesions would have to have clear clinical significance; in fact, their significance remains unclear and the product had no impact upon the only clinical outcome.
- The product would have to have an appropriate safety profile; in fact, there are trends in the direction of increased infection and even increased mortality.
For these reasons, Public Citizen’s Health Research Group opposes the approval of this device.
 SyntheMed. Proposed Package Insert for REPEL-CV. September 17th, 2007, p. 3.
 SyntheMed. REPEL-CV P07005: Summary of Safety and Effectiveness. September 17th, 2007, p. 34.
 SyntheMed. REPEL-CV P07005: Summary of Safety and Effectiveness. September 17th, 2007, Table 17.
 ibid., Table 18.
 ibid., Table 17.
 ibid., p. 53.
 Nkere UU. Postoperative adhesion formation and the use of adhesion preventing techniques in cardiac and general surgery. ASAIO Journal 2000;46:654-6.
 SyntheMed. REPEL-CV P07005: Summary of Safety and Effectiveness. September 17th, 2007, p. 40.
 Center for Devices and Radiological Health. Guidance for Resorbable Adhesion Barrier Devices for Use in Abdominal and/or Pelvic Surgery; Draft Guidance. Food and Drug Administration, December 16, 1999.
 Burns JW. Letter to Dockets Management Branch (Docket 99D-5199). Adhesion Barrier Task Force, March 13, 2000.
 Center for Devices and Radiological Health. Guidance for Resorbable Adhesion Barrier Devices for Use in Abdominal and/or Pelvic Surgery; Guidance for Industry. Food and Drug Administration, June 18, 2002.
 SyntheMed. REPEL-CV P07005: Summary of Safety and Effectiveness. September 17th, 2007, p. 51.
 ibid., Table 21.
 ibid., p. 47.
 Richter KC. Letter to Lifecore Biomedical, Inc. Office of Device Evaluation, Center for Devices and Radiological Health, November 15, 2000.
 Sullivan MG. Intergel sales halted pending investigation of deaths, pain. Ob/Gyn News, May 15, 2003.