Assessing The Findings
From The SIDESTEP Study
Dr. Armstrong: The SIDESTEP study was published last year and has added significantly to the information we have now about diabetic feet in the literature.6 The SIDESTEP study was the study of ertapenem versus piperacillin/tazobactam. Dr. Lipsky, can you briefly summarize this study? How was it unique from other studies of diabetic foot infections?
Dr. Lipsky: After some initial discussions in 2000, we began the study in 2001. It took three years before we were able to meet our target enrollment of 600 patients, which would make this the largest study of diabetic foot infections. We called the study SIDESTEP, which stands for the Study of Infections in Diabetic feet comparing Efficacy, Safety and Tolerability of Ertapenem versus Piperacillin/tazobactam.
The idea of the trial was to compare these two parenteral antibiotic agents, which were given for at least five days in all patients. However, patients in either group could be switched, when appropriate, to oral therapy with amoxicillin/clavulanate for a maximum total of 28 days of antibiotic therapy (i.e., a minimum of five days of IV therapy and a maximum of 23 days of oral therapy).
For this randomized, double-blinded, multicenter trial, we enrolled 586 patients who had diabetic foot infections that were classified as moderate to severe infections, and required intravenous antibiotic therapy. Among these patients, 576 received some study therapy and ultimately 445 patients could be fully assessed at the end of intravenous therapy, which was the primary endpoint. Among these patients, 226 received ertapenem and 219 were treated with piperacillin/tazobactam. Patients enrolled in the two arms were similar by all parameters at baseline.6
When examining various groups at different endpoints, we found that the clinical, microbiological and safety outcomes were all statistically comparable in patients who received ertapenem as in those who got piperacillin/tazobactam. These outcomes demonstrated that ertapenem is a useful antibiotic for treating diabetic foot infections as the outcomes were similar to one of the most widely used agents for this indication.6 This study was the primary evidence for the efficacy of ertapenem for treating diabetic foot infections and it was subsequently approved by the Food and Drug Administration (FDA) for this specific indication. This was in addition to the drug’s previous indication for complicated skin-skin structure infections.
Key Factors That Contributed To Successful Outcomes
Dr. Armstrong: Dr. Sheehan, you were intimately involved in this project from the early stages. One of the things that many of us noticed about this study was the very favorable outcome rate in both treatment arms, whether the patients received the once-a-day parenteral agent ertapenem or the four-times-a-day agent piperacillin/tazobactam. The outcome rate was outstanding and probably superior to any previous study of diabetic foot infections if one were to look at it. Why do you think this was the case in over 80 centers that were involved in this study and what can we make of it? How will these findings help our patients when we go back to the clinic?
Peter Sheehan, MD: First of all, the study looked at two very efficacious antibiotics. Ertapenem and piperacillin/tazobactam are excellent antimicrobials for diabetic foot infections based on the historical assessment of what we know is in the microbiological flora. We would expect a good response rate with two excellent antibiotics. There was also the patient selection based upon the inclusion/exclusion criteria. These were largely moderate to severe diabetic foot infections that were being treated. There was a paucity of ischemic patients, whom are the most challenging of all in terms of managing their infections, as well as those with osteomyelitis.
Another factor was the center selection. The centers involved in the study are excellent centers. The investigators were skilled clinicians, who understood how to manage these infections and, perhaps more importantly, understood how to manage ulcerations according to the standards of care that have been published.
Lawrence A. Lavery, DPM: Following up on what Dr. Sheehan said, there is often a “study effect” that occurs with randomized clinical studies. Certainly, the people who participate in these studies have some expertise. Generally, clinicians in a study are giving a high level, detailed evaluation on a regular basis. This is certainly driven by the outcomes for the clinical measures that we are evaluating in the study. Accordingly, there is probably information that you are driven to look at when you are a clinician participating in a randomized clinical study that may not be in front of you for decision-making in the average clinic or office setting.
Dr. Sheehan also mentioned that limb-threatening infections were excluded from the study. People that were going to have revascularizations were not included in the study. Patients with osteomyelitis were not included in the study.
Dr. Armstrong: People with osteomyelitis could be included in the study if one resected the osteomyelitis. I think you would agree that resection of osteomyelitis is probably the standard of care in many places.
Dr. Joseph, in addition to the factors discussed by Drs. Lavery and Sheehan, how would you explain the favorable clinical outcomes in the 90 percent range for the SIDESTEP study when the clinical outcomes in many other studies are in the 70 and 80 percent range for complicated skin-skin structure infections and diabetic foot infections?2,6
Dr. Joseph: I think what Dr. Lavery was talking about is certainly a factor. I also agree with Dr. Sheehan’s points. There were 89 centers in this study with people who were really interested in diabetic foot infections and are known to be excellent clinicians in that area.
The Value Of Improved Study Protocols
Dr. Joseph: However, I think another advantage is the protocols have been getting better over the years when it comes to studying diabetic foot infections. This whole issue of looking at diabetic foot infections and diabetic foot ulcerations, and how to standardize them for the protocols was not that easy to do as recently as maybe 10 or 15 years ago. Some of the studies on becaplermin gel, which were very sophisticated for the time, are about 15 years old now but were the first studies that standardized some of these issues.7 The studies that Dr. Lipsky did with pexiganan also standardized some of these issues.8,9 I do think the increasing sophistication of these protocols over the last 10 or 15 years might play a role.
This also ties into the notion of a specific indication for diabetic foot infection through the FDA that Dr. Lipsky alluded to earlier with ertapenem. Ten or 15 years ago, no drug received a diabetic foot indication. The indications were more along the lines of complicated skin-skin structure. The first drug to actually have the diabetic foot infection indication was trovafloxacin.
Dr. Lipsky: In addition to the site selection and the involvement of clinicians who are interested in and skilled at caring for people with diabetic foot infections, we trained these people rigorously when they became investigators for the study. Dr. Armstrong and I met with them in Arizona and had a full day session teaching them about the importance of debridement, obtaining optimal specimens for culture, how to use the study equipment, etc. They were very well prepared to go into the study.
Dr. Armstrong: Dr. Lipsky, did you want to discuss the wound scoring system that was utilized for the SIDESTEP study?
Dr. Lipsky: We originally used a wound scoring system for a study that compared the topical antimicrobial peptide pexiganan against the oral drug ofloxacin.8,9 We modified that wound scoring system for use in the linezolid diabetic foot infection study and further modified it for use in the ertapenem study.6,10
This wound scoring system allows us to take the size and depth of the wound into consideration as well as all signs and symptoms of inflammation and infection. We came up with a quantitative score that allows us to assess wounds both at baseline and during therapy to compare wound scores over time.
In both the pexiganan and ertapenem studies, we found a good correlation of the wound score with clinical outcome.6,8,9 This may be something that could be standardized and used to allow various studies to compare their results as well as to allow investigators to know the severity of wound infections when the patients are first seen.
Dr. Lavery: Dr. Lipsky, can you briefly summarize the factors involved in coming up with the wound score?
The wound scoring system has two basic components. The first part looks at size, comprised of measurements of the length, width and depth by standard methods. The second part examines all the components of inflammation. We looked at redness, induration, warmth, pain, tenderness, etc. Then we developed a quantitative system by which the investigators could determine a score for each of those factors. In some instances, you could simply measure a component such as the amount of redness. In other instances, you could grade secretions or measure depths.
A Closer Look At The Clinical Outcomes
Dr. Lipsky: It is important to be aware that the primary outcome we selected for the SIDESTEP study was the clinical response at the discontinuation of IV therapy. As we discussed, this could be as early as five days into the course of their treatment. The outcome was a 94 percent response rate for ertapenem and 92 percent for piperacillin/tazobactam.6 If you looked at the clinical outcome 10 days after the completion of antibiotic therapy, the so-called “follow-up assessment,” you get a somewhat lower response rate that is a little more similar to what has been reported in other studies.6,10-12
For patients who had a moderate infection, 89 percent who received ertapenem and 88 percent who got piperacillin/tazobactam had a favorable clinical outcome. For the one-third of patients who had a severe infection, the clinical response rate was 83 percent for ertapenem and 71 percent for piperacillin/tazobactam. Part of the excellent response rate was the fact that we were studying a parenteral drug and the primary time of response we were looking at was fairly early on in the duration of treatment and many patients did not yet have a chance to fail therapy.6
In fairness, I should also point out that when the FDA obtained and evaluated the primary data and made its own decisions about which patients should and should not be included and which ones were cures or clinically favorable outcomes, they came up with a clinical success rate of 75 percent for ertapenem and 70.8 percent for pipericillin/tazobactam at the 10 day follow-up assessment.6
While we start out with 90-plus percent clinical response rates, in fact, depending upon how one looks at and interprets the data, the response rates are much more similar to those in other recent studies, like the linezolid study, the daptomycin study and the piperacillin/tazobactam study. 6,10-12
Making The Switch From Parenteral Treatment To Oral Therapy
Dr. Armstrong: This leads to a potentially intriguing question. Your theory is that these patients just did not have time to fail after, say, between five and 10 days of parenteral therapy. Dr. Joseph, can you think of any other reasons why that might be the case? Why are the clinical outcomes lower after that period of time? Were those endpoints, the post-oral and parenteral therapy endpoints, similar to outcome measures in previous studies?
Dr. Joseph: It is hard to say. Were they reasonable endpoints? Yes, I think they were reasonable endpoints because clinical response is what we are looking for. It is a very clinically relevant parameter when you are able to say, “Okay, I have given this patient this parenteral drug for five days and look how much better the foot looks.” Some studies get bogged down with the microbiological responses which, in my opinion, do not mean as much to clinicians.
Dr. Armstrong: Exactly. The shift from parenteral therapy to oral therapy is a key decision point, isn’t it? An effective agent and an effective response rate are critical to giving the clinician the confidence to move in that direction. Let me pose the question to the panel. What are the clinical landmarks that you are looking for that will help you decide when to change the treatment therapy or when to discharge the patient?
Dr. Lipsky: In the SIDESTEP study, when the patient was discontinued from IV (DCIV) therapy, two things could have happened. He or she could have been switched over to oral therapy to continue being treated or the investigator could have decided that the patient did not need any further antibiotic therapy. About two-thirds of the time, patients were switched to oral therapy so the great majority of patients at the DCIV endpoint were still getting antibiotic therapy. We defined failure as a patient who was getting worse or needed continued antibiotic therapy beyond the maximum allowable duration of study therapy (i.e., 28 days). A patient who needed an operation or whose condition was clearly becoming worse would be a patient who was not clinically responding.
In essence, what would happen for most of these patients is that after five days or more of IV therapy, a decision was made that they were now stable enough to switch to oral therapy but they were not yet cured of their infection. These patients would have been considered to have a favorable clinical response at the DCIV. That is why I say they did not have time yet to fail. Then if we look at the outcome at 10 days, after the patient has completed his or her antibiotic therapy (the follow-up assessment visit), the clinical outcome was in the mid-80 percent range.6 Therefore, approximately 5 to 7 percent of these patients did go on to be considered clinical failures of therapy despite looking like they were doing well at the end of their IV therapy.6
Dr. Sheehan: In my experience, that is unusual. After I see a favorable response and switch to oral therapy based on the organisms that were isolated, it is unusual that the patient would then get worse.
Dr. Lipsky: Right. That is in fact what we found. It happened no more than 10 percent of the time and usually less.
