The following interview with Medpace physicians Dr. Jon Bruss (JB) and Dr.
Brian Murphy (BM) is based on an accepted manuscript that was published in Clinical Infectious Diseases, by Oxford University Press. The manuscript titled “Recommendations for Source Control Review in Clinical Trials of Anti-infective Agents in Complicated Intra abdominal Infections” was published in March 2013. As contributing authors to the article
(lead author is Dr. Joesph Solomkin, Department of Surgery, University of
Cincinnati College of Medicine), Medpace Infectious Disease experts shared their insights.
Q. Why is source control and the surgeon’s assessment of the adequacy of the initial surgical approach important in determining therapeutic outcome and antibiotic efficacy in complicated intra-abdominal infection (cIAI) clinical trials?
JB: Unlike many other infections, cIAIs require some operative or percutaneous intervention to adequately address the source of infection. When there is a significant infected fluid collection or necrotic tissue, antibiotics alone would not be able to completely address an infection. At a minimum, intervention is needed to drain fluid collections and decrease the bacterial burden. If this intervention is inadequate and the source of infection remains, even the best antibiotic is likely to fail.
BM: And it gets even more complicated when you consider additional elements of source control that may be present in cIAIs. Managing the underlying pathological process—surgical correction of bowel perforation, debridement of devitalized tissue, and appropriate surgical wound management are examples.
Even skilled surgeons, under certain circumstances, may not be able to completely control the source of infection after the first surgery. For example, at the time of surgery the extent of the infection may not be completely appreciated, anatomic conditions like a tumor or adheshions may not allow the preferred procedure to be performed, the patient may be hemodynamically unstable and the surgery cannot be completed. And, unfortunately in some cases, there may be judgment or technical error such as bowel perforation that goes unrecognized or a viscus that is inadvertently perforated during the procedure.
JB: When we took this concept to prior clinical trials in cIAIs we found that inadequate intervention for any of the reasons we mentioned often distorted therapeutic outcomes. So really inclusion of cases where there was inadequate source control in the efficacy analysis of a particular therapeutic antibiotic confounded the results. In other words, the effect of the antibiotic was partially masked by the inadequate procedure and the lack of reduction in bacterial burden.
Obviously, these issues combine to make review of source control in cIAI trials an important step in trial design.
Q. Randomized clinical trials are routinely used to assess various
therapeutic interventions for cIAIs. How do inadequate procedures affect drug
JB: Clinical outcome is dependent on both the procedure and intervention. The clinical assessment of outcome is a composite of the adequacy of the procedure and the efficacy of the antibiotic. How can one determine the efficacy of the antibiotic when the other half of the equation is not optimal? Inclusion of a case where there was documented inadequate source control in the efficacy analysis of an antibiotic in a non inferiority study design distorts the assessment.
While randomization should in theory evenly distribute patients with inadequate procedures between the treatment groups, the studies we reviewed demonstrated that this is not always the case.
Q: Given the wide variations in patient co-morbidities, pre-operative diagnostic studies, anatomy, and pathology encountered, are there firm inclusion criteria that clinical studies should require to address source control?
BM: In our opinion, no, as that may introduce selection bias and enroll populations of patients that do not necessarily represent ‘real life’ cases that physicians manage. A less biased approach would the addition of a blinded consensus or adjudication review process. This would necessitate that the reviewers are trained to the same broad anatomic and physiologic principles that the treating surgeons or interventional radiologists are and could identify care that does not adhere to those principles.
Q: What are your recommendations for assessment of adequacy of source control?
JB: There are two key areas to consider—operative management and percutaneous management. The definition of an adequate operative management is generally agreed involves drainage of all fluid collections, closure or resection of any openings into the gastrointestinal tract, and resection of inflamed tissue. Other elements that must be considered to determine adequacy are timing of the procedure, incision used, the type of closure or diversion, drain placement, and surgical wound management.
Source control for cIAIs using surgery or percutaneous catheter drainage should achieve the same endpoint. However, the procedures and time courses for these two approaches will differ considerably. By the end of a surgical operation, source control is typically achieved and the surgeon is aware of the likelihood of success. This is not always the case with percutaneous catheter drainage as it may take three to four days to define a successful source control for cIAIs.
Q: What are your recommendations for a Source Control Review Process and Panel?
BM: Plan who the members will be and what they will review up front. A detailed plan for source control review should be worked out before the start of the study.
This starts with choosing who the members of the source control review panel will be. Qualified members for the panel should be identified prior to start of the trial and involved in the review and finalization of the specific review plan as well as creation of study documents. We recommend that the panel consist of three qualified surgeons to review cases involving operative interventions and three qualified interventional radiologists to review cases involving percutaneous interventions. To decrease the burden for the panel, two reviewers can perform the initial review, with the third reserved to resolve any discordance.
The members should be active in developing the case report forms. These forms should be thorough enough to accurately record medical history, clinical events, procedures and imaging to inform the source control review panel but not so cumbersome as to be an impediment for the site to record the information or the panel to expeditiously review it.
So what do they review? The panel stays blinded with respect to the investigative site, investigator, patient identifiers, and study drug therapy received by the patient and is provided with the patient’s operative/procedure notes, case report forms, and any radiologic reports. We recommend that the panel stays blinded review all cases judged as failures at Test of Cure (TOC); all cures at TOC with a second procedure, if the second procedure is unexplained or was performed specifically to control the infection; and all cures that received a non-study systemic antibiotic between randomization and the TOC visit or any therapy to address the infection between TOC and the long-term follow-up visit.
The panel then adjudicates on whether the initial procedure was adequate or inadequate to control the source of infection with their decision documented in the clinical trial database.
JB: We also recommend that approximately 10-20% of all cases be reviewed by the panel for adequacy of source control, regardless of investigator assessment of outcome. This information will help to validate the accuracy of the review process.
Keep in mind that the panel does not alter the classification of clinical outcome, regardless of their decision on adequacy of source control. Their charter is to solely determine the adequacy of the procedure.
Q: Should there be any considerations for confounding associated with source control intervention?
BM: Yes. We recommend that the variable of the interventions to achieve source
control be considered in the sample size determination. By including this a
priori in the design of the study and how source control may affect outcome, we
may have more certainty that the planned study has a better chance of achieving
its clinical and statistical endpoints and determining the true effect of the