FDA’s newest draft guidance lays out considerations for developing treatments for diabetic foot infections
In a new draft guidance document released within weeks of a related clinical practice guideline, the FDA provides considerations for developing therapies to treat diabetic foot infections, focusing on Phase 3 efficacy trials. The FDA’s recommendations for noninferiority trials are built on sparse data, a fact acknowledged in an appendix to the draft.
Background: Diabetic foot infections
- Diabetes mellitus, a disease that occurs when blood glucose levels are too high, is highly prevalent. The most common form, type 2 diabetes (T2DM), occurs when the body does not make or use insulin well. The Centers for Disease Control and Prevention’s National Diabetes Statistics Report estimates that over 11% of the U.S. population has diabetes. Of these, T2DM “accounts for 90% to 95% of all diabetes cases” in the country.
- Of the multiple complications that come with diabetes, neuropathy is one of the most common. Neuropathy is a result of nerve damage. In patients with diabetes, this can take two forms – autonomic or peripheral. Autonomic neuropathy involves damage to the nerves that control automatic functions of the body and can cause various cardiovascular, gastrointestinal, and genitourinary complications. Peripheral neuropathy, on the other hand, is a sensory neuropathy, resulting in burning pain, tingling, numbness, and/or the loss of various sensations. In diabetic peripheral neuropathy, the legs and feet are most commonly affected, although the hands can sometimes be affected as well. This form of neuropathy is estimated to be present in at least 20% of adult patients with T2DM.
- One of the major complications associated with diabetic peripheral neuropathy is diabetic foot infections (DFIs). As the name implies, DFIs are soft tissue or bone infections of the foot that occur in people with diabetes. These infections often, but not always, occur after a diabetic foot ulceration, which is a break in the protective skin barrier. Although the sensational changes from peripheral neuropathy contribute directly to the increased risk of an ulcer and subsequent infection in people with diabetes, there are other contributing factors, such as a higher rate of dry cracking skin and reduced blood flow to the foot. Additionally, people with diabetes often have a reduced ability to fight infection and heal from wounds.
- Preventive measures are a vital component of battling DFIs; however, once a DFI is present, it can be challenging to treat and resolve. If a DFI is not caught and treated early on, it can progress rapidly, killing the infected tissue and even spreading the infection to nearby tissues, including fascia, tendons, muscles, joints, and bones. Although these infections are most often caused by staphylococcus or streptococci initially, as they progress and deepen, they may become polymicrobial, with the involvement of pseudomonas and anaerobic bacteria. Many of the issues that predispose patients to DFIs, including reduced blood flow, reduced immune function and inhibited healing, can also act as barriers to treatment and resolution.
- DFIs are associated with significant morbidity and mortality. The three-year mortality rate for patients who develo a diabetic foot ulcer increases from 13% to 28%. In patients who have an infected ulcer, or DFI, the five-year mortality rate has been estimated to be as high as 45%. Approximately 15% of patients who develop a bone infection as part of a DFI will require amputation, and the five-year mortality rate in these patients jumps to 60%. Research has found that polymicrobial growth is associated with a higher risk of mortality in patients with DFI.
- DFIs are often treated with multiple interventions, including various methods of wound cleaning and drainage, in conjunction with antibiotics. Currently, these infections are treated with oral or intravenous antibiotics that are already known to provide appropriate coverage of the pathogen that was isolated from that patient’s infection. In their joint 2023 guideline on the treatment of DFIs, the Infectious Diseases Society of America (IDSA) and the International Working Group of the Diabetic Foot (IWGDF) state that, “Based on many studies (most limited by methodological flaws) that compared various oral or parenteral antibiotic agents in patients with DFI, treatment with any appropriately selected agent of most classes of antibiotics by either route is effective in the great majority of cases.”
- However, there is limited evidence on the use of antibiotics specifically for DFIs, resulting in minimal knowledge on comparative effectiveness, antimicrobial resistance, or the appropriate dose and duration of treatment. The IDSA notes that although “no antimicrobial class or agent has been found to be superior to others for treating DFIs,” this evidence is limited to two retrospective studies and a systematic review. Also, two studies have suggested that one antibiotic, tigecycline, may be inferior to other antibiotics. Additionally, the IDSA acknowledges a “paucity of data on the resolution of infection, recurrence of infection, and the acquisition of antimicrobial resistance” in DFIs, as well as a lack of DFI-specific dosing guidance. Finally, the guidelines notes that newer antimicrobial agents that can improve adherence through administration as a one-time infusion (i.e., Dalvance and Orbactiv) have not been evaluated for this indication, resulting in insufficient evidence to recommend use.
Now, the FDA has issued a draft guidance document specifically addressing the treatment of DFIs
- The DFI-specific draft guidance document came about, writes the FDA, because the existing industry guidance on Acute Bacterial Skin and Skin Structure Infections [ABSSSI]: Developing Drugs for Treatment “does not address DFI due to the differences between DFI and other ABSSSI related to definitions, clinical manifestations, microbiology, management, and measurement of clinical outcomes,” necessitating a separate guidance.
- In terms of general drug development considerations during the early phase, sponsors should consider obtaining Phase 1 data that demonstrate that the candidate drug has adequate penetration to the outer skin layers, after consultation with the agency about drug penetration study technique.
- The population studied should be people with diabetes who have a bacterial foot infection that began at a point not above the malleoli. Although DFIs can range from cellulitis to deep wounds, products to treat bone and joint infections are excluded from the scope of the guidance. Sponsors can consider using a classification system to define the study population; the FDA provides citations for several such systems.
- Noninferiority trials are “interpretable and acceptable” to support a drug to treat DFI if the sponsor can justify a noninferiority margin that the FDA deems acceptable. An appendix dedicated to the question of justification of noninferiority margins provides extensive further context – see below. Sponsors can also choose to conduct trials showing superiority of the investigational product over an active comparator; these are “also readily interpretable and provide direct evidence of treatment benefit,” according to the draft guidance.
Nuts and bolts of trial design
- To show effectiveness of a product for DFI, the FDA is looking for two adequate and well-controlled trials, though evidence may also come from one adequate and well-controlled trial supported by, for example, confirmatory evidence from a related indication such as acute bacterial skin and skin structure infections. If a single trial with supporting evidence is being considered, sponsors should first discuss this strategy with the FDA.
- For safety, the FDA recommends gathering data from at least 500 participants, though some of the safety data could also be drawn from clinical trials for other indications, as long as the dosage and duration of treatment in those trials was at least as great as that proposed for DFI. Novel treatments with promising benefit may be able to use a smaller preapproval safety database, but sponsors should consult with the agency on this question.
- Phase 3 efficacy trials should be prospective, randomized, and double-blinded, comparing the investigational drug with a standard-of-care active control to assess superiority or noninferiority. Enrollees should have a DFI with a range of degrees of involvement. Any surgical intervention such as abscess incision and drainage or debridement should occur within the first 48 hours after study enrollment, and patients who need procedures performed after the 48-hour window should be considered clinical failures. However, “minor prespecified procedures” like “superficial debridement of devitalized tissue” are permitted throughout the study. The FDA advises avoiding the use of topical antibacterials.
- Patients with either type 1 or type 2 diabetes and moderate to severe foot infections without bone involvement may be enrolled, so long as the infection does not currently extend above the knee and began at or below the malleoli. The guidance document refers sponsors to the IWGDF DFI criteria for defining moderate and severe DFI. These criteria look at signs and symptoms such as degree of erythema, depth of tissue involvement, and presence of any systemic manifestations such as fever, tachypnea, and leukocytosis. Presence of two or more systemic manifestations indicates severe infection. The draft guidance provides suggestions for methods to measure lesions size and recommends using uniform methodology at all trial sites. Sponsors should include patients with neuropathy and vascular insufficiency, to build a representative patient population.
- Patients with osteomyelitis, septic arthritis, gangrene requiring amputation, and critical limb ischemia should be excluded, as should those with primarily non-bacterial infections, among other exclusion criteria. The FDA also recommends excluding those who have received more than 24 hours of “effective antibacterial drug therapy” to treat DFI, and capping enrollment at 25% for those who have received any “potentially active antibacterial therapy for the current DFI episode.”
Microbiologic, statistical, and clinical considerations
- The draft guidance includes a brief section on clinical microbiology considerations, including acceptable sources for obtaining specimens before initiating therapy; wound swabs and sinus tract aspirates are “generally not acceptable” culture sources. Sponsors should perform Gram stains and culture the specimens, with in vitro susceptibility testing, with final confirmation of culture isolates performed at a central laboratory. Investigators should obtain blood cultures before initiating any treatment.
- An important provision in this section is that study investigators should determine whether bacteria isolated are “pathogens, colonizers, or contaminants;” only pathogens should be used to determine whether a patient is microbiologically evaluable – that is, whether their data can be included in the efficacy determination. To this point, sponsors should assemble a list of accepted pathogens and review it with the FDA.
- Before enrollment, participants should be assessed for osteomyelitis, which can either be diagnosed by imaging (modality unspecified), or with a positive probe to bone through a wound, with magnetic resonance imaging (MRI) considered for patients with sepsis or an open and infected ulcer that doesn’t probe to bone.
- Stratification of participants should account for various factors, including patients with and without foot ulcers, “to account for the differences in the natural history of the disease entities.” Other covariates that sponsors can consider including as prespecified adjustments include infection severity and degree of vascular insufficiency, according to the draft guidance. Studies should not have an upper age limit for participation, and comorbidities shouldn’t be used as exclusion criteria, barring safety concerns. Obese patients should be included, and pharmacokinetic (PK) parameters for study drugs should be determined prior to phase 3 in both normal-weight and obese patients. Patients with renal insufficiency – a common complication of diabetes – should be included, with appropriate early-stage PK investigations and dose adjustment. The guidance offers similar considerations for patients with hepatic impairment, although early-stage PK investigations are not considered an expectation per the guidance.
- The draft guidance offers additional considerations for dose selection, advising sponsors to avoid concurrent antibacterial therapy except in the context of a planned add-on trial. Rescue therapy “will generally be interpreted as failure of the study drug.” However, concomitant antibacterial treatment for bacteria not targeted by the study drug may be considered in consultation with the FDA.
- Sponsors may have a hard time parsing out the relative contributions of adjunctive measures, which form a core part of treatment of DFI. The FDA advises that sponsors prespecify and track which of these measures, such as specialized wound care products and techniques and various off-loading mechanisms, are permitted during the study period.
- Duration of treatment for a DFI is at least seven to 14 days, according to the draft guidance. Plans for prolonged treatment should be well defined and reviewed with the FDA before the study begins. The FDA also provides advice about how to handle an intravenous-to-oral switch if an oral formulation of the investigational product is available.
Efficacy and safety endpoints
- The FDA is clear about the primary efficacy endpoint, which should be “resolution or improvement of all signs and symptoms of DFI to the extent that no further antibacterial drug therapy is needed and none of the following events have occurred: receipt of rescue therapy, unplanned surgical debridement, amputation, or death.” These criteria should be predefined for the sake of uniform clinical assessments.
- Importantly, all assessments occurring after baseline should be timed according to time from randomization, rather than from the end of therapy. The FDA recommends that the test-of-cure visit, which occurs at about 21 days after randomization and seven to 14 days after end of treatment, represent the timepoint for assessment of the primary outcome measure. Treatment effect should be tracked at end of therapy and subsequent follow-ups, “to evaluate for durability of the treatment effect.”
- Sponsors should consider a secondary endpoint measure of “the composite of death, unplanned amputation, and infectious complications at 21 days post-randomization,” according to the FDA, because such an endpoint “objectively measures key patient benefits.” Clinical response at end of therapy and all-cause mortality at a fixed post-randomization time point could be used as additional secondary endpoints.
- The FDA reiterated that osteomyelitis, unplanned surgical debridement, and amputation all constitute treatment failures. Here, the agency also makes brief mention of patient-reported outcomes (PROs), noting they “can be considered,” and sponsors should discuss such plans with the FDA.
- The draft guidance also sets out fairly prescriptive recommendations for assessments to be taken during the entry visit and thereafter. In addition to the entry visit, the FDA recommends on-therapy visits beginning 48-72 hours after initiating the study drug, as well as an end of therapy visit, which should follow patients who have discontinued the study drug per protocol. A test-of-cure visit and a Day 28 post-randomization visit are also called for; the FDA specifies which assessments are recommended at each visit.
Statistical considerations and the noninferiority appendix
- The FDA recommends subgroup analyses to look at the primary endpoint in patients who did and did not receive any prior antibacterial therapy, with an additional recommendation to perform further sensitivity analyses and exploratory analysis “for factors that could modify the primary analysis findings.” This portion of the draft guidance also defines the safety population, the intent-to-treat population, and the “microbiological intent-to-treat (micro-ITT)” population. This latter group includes patients who were randomized, and who have “a baseline bacterial pathogen known to cause DFI.” The FDA provides further considerations for patients who can be included in the trial’s per-protocol, clinically evaluable, and microbiologically evaluable populations.
- Regardless of whether a trial is being run as a noninferiority or superiority trial, the primary efficacy analysis should be based on the ITT population rather than the micro-ITT population. This acknowledges the fact that patients can have DFIs even though microbiologic identification of a pathogen might not be possible, as when a patient has cellulitis but no open wound.
- Additional statistical considerations are provided for both noninferiority and superiority trials, including how to calculate the sample size to account for prespecified type I and type II statistical error rates, the clinical response expected, and the noninferiority margin or magnitude of superiority expected.
- The guidance document itself and an appendix provide considerations specific to the condition of DFI. The FDA acknowledges that a noninferiority trial should be based on determining the degree to which the active comparator can be differentiated from placebo, but such trials are lacking in DFI. “Because no historical, randomized, placebo-controlled trials for patients with diabetic foot infection (DFI) could be identified, direct estimation of the treatment effect was not possible. Therefore, we considered retrospective case series comparing the pre- with the post-antibacterial drug era.”
- The appendix explains that the FDA reached back to studies conducted in the 1940s to look at amputation and mortality rates before and after penicillin became standard of care for DFIs. Major limitations of these studies are discussed by the FDA, with limitations of case studies and other potential data sources also called out. The FDA’s best estimate from available data supports a noninferiority margin of 10%, though sponsors could discuss other approaches with the FDA.
- Recent reviews of DFI treatment comment on the fact that patients with DFI are often explicitly excluded from antibiotic clinical trials. Multiple novel drugs that might be expected to address unmet need for DFI, such as Teflaro (ceftaroline), Fetroja (cefiderocol), Zerbaxa (Ceftolozane/tazobactam) and Avycaz (ceftazidime-avibactam), have not been evaluated in this population. And as noted previously by the IDSA, antibiotics with novel methods of administration have also excluded these patients from studies. Interestingly, one antibiotic, Tygacil (tigecycline), contains labeling language recommending against its use in this population due to a clinical trial failing to demonstrate non-inferiority.
- Currently, there are three drugs available on the market with specific indications for the treatment of DFI – Zosyn (piperacillin/tazobactam), Zyvox (linezolid), and Invanz (ertapenem). All three of these drugs are approved for the treatment of DFIs without bone involvement (osteomyelitis) and include a list of pathogens against which they are expected to be active. In this new draft guidance, the FDA recommends a similar approach to the labeling language: “Drug X is indicated for the treatment of adults with diabetic foot infections (without concomitant osteomyelitis or septic arthritis) caused by … [list genus and species of bacteria].”
- This new draft guidance document did not appear on the FDA’s roster of guidance priorities until just three months ago. Twice per year, the FDA’s Center for Drug Evaluation and Research (CDER) typically publishes a guidance agenda that provides a list of possible topics for future guidance documents. This DFI draft guidance did not appear on CDER’s guidance agenda until this July, suggesting that it was conceived and developed relatively recently.
- The timing of this newly conceived guidance document, which closely follows the release of the IDSA/IWGDF guidelines this month, suggests that the FDA guidance has been, in some way, shaped by the infectious disease guidelines. The IDSA/IWGDF guidelines note the limitations of the available evidence for antibiotic use in DFI and also provide step-wise recommendations for staging DFIs and diagnosing bone involvement.
- The FDA document is vague in some areas, as when recommending osteomyelitis diagnosis by the probe-to-bone test or “imaging.” Similarly, the document refers both to inclusion criteria that would have the infection at or below the level of the malleoli, and to having the source of the infection at or below the malleoli, with involvement not extending above the knee – a markedly larger area of involvement. However, when read in conjunction with the much more thorough body of literature on diabetic foot promulgated by IDSA and IWGDF, investigators could have more clarity on diagnostic criteria and inclusion/exclusion criteria for trial participation. Sponsors could consider asking the FDA to be clearer in referencing or deferring to these guidelines in the final guidance document.
- Diagnosis of mild to moderate DFIs is largely clinical, and the draft guidance recommends inclusion of patients without microbiologically confirmed infections. This means that sponsors will need to rely heavily on investigators’ clinical acumen for selection of the trial population and, in many cases, for assessing endpoints. Especially for noninferiority trials where a margin of 10% will be the cutoff between meeting and missing a primary endpoint, sponsors will benefit from having measures in place to ensure uniform and accurate assessment and recording.
- The FDA has a separate guidance for the conduct of noninferiority studies to support approval of antibacterial drug products, but much of the content of that brief 2010 guidance is mirrored and contextualized in this draft guidance. The 2013 guidance document on developing drugs for ABSSSI also has a discussion of noninferiority trials that parallels the recommendations of the DFI guidance document. The FDA’s appendix acknowledges the difficulties in constructing noninferiority trials for a clinical area with a paucity of recent data. Stakeholders may wish to reinforce the need for good communication and flexibility while designing trials for drugs to treat DFI.
- What’s next? The docket for comments on this guidance will be open until December 18, 2023.
To contact the authors of this item, please email Kari Oakes ( [email protected]) or Chelsey McIntyre ( [email protected]).
To contact the editor of this item, please email Chelsey McIntyre ( [email protected]).
Key Documents and Dates
- Diabetic Foot Infections: Developing Drugs for Treatment
- Docket no. FDA-2023-D-4067
- Comments due: December 18, 2023