Local Antibiotic Delivery Systems in Diabetic Foot Osteomyelitis: A Brief Review

Diabetic foot osteomyelitis (DFO) is a severe, difficult to treat infection. Local antibiotic delivery has been studied as a potential therapeutic adjunct following surgery for DFO. This review aims to summarize the evidence on local antibiotic delivery systems in DFO. PubMed database was searched up to March 2020. Overall, 16 studies were identified and included: 3 randomized controlled trials (RCTs), 3 retrospective studies (RSs), and 10 case series. In the RCTs, gentamicin-impregnated collagen sponges significantly improved clinical healing rates and slightly improved duration of hospitalization. In the RSs, antibiotic-impregnated calcium sulfate beads non-significantly improved all healing parameters, but did not reduce post-operative amputation rates or time of healing. The majority of case series used calcium sulfate beads, achieving adequate rates of healing and eradication of infection. In conclusion, evidence for add-on local antibiotic delivery in DFO is still limited; more data are needed to assess this therapeutic measure.


Introduction
iabetes mellitus increases the risk of foot infections, some cases of which progress to diabetic foot osteomyelitis (DFO) [1,2]. Foot deformity, peripheral neuropathy, peripheral arterial disease, and minor injury increase the risk of diabetic foot lesions [2][3][4][5]. The development of biofilms in chronic wounds represents an additional challenge, since they protect pathogens from removal by host immunity and systemically administered antibiotics [6].
Management of DFO may be surgical or medical, depending on patient characteristics [7]. Surgery is especially useful in the event of pus, sequestrum, gangrene, or antibiotic-resistant bacteria [8]. Instead of amputation, debridement offers the advantage of removing necrotic while preserving healthy bones and tissues [9]. This approach is sometimes accompanied by local antibiotic delivery [3].
Local antibiotics offer the following advantages: higher local antibiotic concentration, longer duration, and fewer side effects [3]. At the same time, they act as a bone substitute that fills the dead space caused by bone resection [10]. Polymethylmethacrylate (PMMA) beads are the major representative of nonbiodegradable carriers [11]. Antibiotic release from PMMA beads is initially high during the first 48-72 hours, but quickly falls to lower levels, and may elute for weeks or even years [11]. Disadvantages include the high temperature it produces and the surgical removal of the beads required upon completion of drug release [12,13].
During the last 2 decades, biodegradable carriers have been developed: proteins (collagen, gelatin, thrombin etc.), synthetic polymers, grafts, and substitutes (calcium sulfate or phosphate) [14]. These act as a matrix for new bo+ne growth. During their degradation, additional release of antibiotics occurs, prolonging their action and preventing biofilm formation on their surface [15].
The aim of this brief review is to summarize the evidence on add-on local antibiotic delivery in the surgical management of DFO.

Search strategy
We performed a search in the PubMed database for studies published up to December 2020 on the management of patients with DFO using an implantable antibiotic delivery system. We excluded case reports, case series with fewer than 5 patients, in vitro studies, reviews, comments, letters, and studies on other locations of osteomyelitis. Studies in which >10% of patients did not have diabetes were excluded unless the results for these patients were presented separately. Only publications in the English language were included.
Parameters evaluated included healing rates, time, and complications, such as further surgical interventions, amputation rates, and mortality. Clinical presentation, laboratory investigation, radiological evaluation, antibiotic therapy, duration of symptoms, previous surgical procedures, and comorbidities were also recorded.

Randomized controlled trials
Lipsky et al. randomized 56 patients with moderately infected diabetic foot ulcers in a 2:1 ratio into 2 groups, one with and the other without the use of a gentamicincollagen sponge in addition to standard care for up to 28 days [15]. Significantly higher rates of clinical cure and eradication of pathogens were achieved in the gentamicin-collagen sponge group [15]. Safety data were similar between the 2 groups.
Varga et al. investigated the efficacy of a gentamicincollagen sponge application into wounds after minor amputation for non-healing ulcers with DFO [16]. Fifty diabetes patients were randomized to the add-on gentamicin sponge or usual care group. All patients received systemic antibiotics according to their antibiogram profile. In the gentamicin sponge group, wound healing duration was significantly shortened by almost 2 weeks. No differences were observed between the groups in length of hospitalization or number of revisions for wound breakdown or subsequent amputations.
Uçkay et al. continued their initial trial re-examining the potential benefits of gentamicin-collagen sponges in a larger RTC of 88 patients with DFO and prolonged follow-up [17]. There was no difference in clinical cure rates in favor of the gentamicin-sponge. However, a small trend towards faster healing was noted in the gentamicin-collagen sponge group. Similar to the other studies, local antibiotic delivery was not associated with safety concerns.

Retrospective studies
Krause et al. assessed the effect of local tobramycinimpregnated calcium sulfate beads in addition to standard treatment after transmetatarsal amputation in diabetes patients with non-healing forefoot fullthickness ulceration with DFO or skin necrosis [18]. In total, data from 65 amputations were reviewed, including 49 cases in the beads group and 16 cases in the group without beads. Wound breakdown rates were significantly lower in the beads group. In this group, there was also a non-significant reduction in time of healing. There were no differences between the groups in length of hospitalization and need for ipsilateral second transtibial amputation.
Qin et al. compared infected bone resection combined with adjuvant antibiotic-impregnated calcium sulfate vs. infected bone resection alone for the treatment of DFO in 46 patients [19]. Antibiotic-impregnated calcium sulfate prevented the recurrence of DFO, but showed no significant improvement in healing rates, post-operative amputation rate, and time of healing.
In the most recent retrospective study, Chatzipapas et al. recruited 25 patients with forefoot and calcaneal DFO who were divided into 3 groups [20]. Healing rates were 100% in the PMMA group (surgical debridement in combination with the local application of antibioticloaded PMMA beads), 87.5% in the calcium sulfate group (surgical debridement in combination with the local application of hydroxyapatite and calcium sulfate beads), and 87.5% in the control group (surgical debridement only) [20].

Case series
Roukis et al. studied 16 patients (15 diabetes) with medical comorbidities and offered two-stage reconstruction with a surgical skin flap of fullthickness foot ulcers using a V-Y technique for patients with high risk of wound breakdown [21]. Antibioticimpregnated PMMA spacers were used in all first-stage procedures. After 3 days, a second-stage procedure was performed, involving further debridement, V-Y flap cover, and filling of bone defects with either bone graft combined with platelet-rich plasma or with antibioticimpregnated PMMA. This treatment resulted in primary healing in 9 of 15 patients and secondary healing with dressings in a further four. Two patients required transmetatarsal amputation at a mean followup of 15 months. The large case series by Gauland evaluated vancomycin-and gentamicin-loaded calcium sulfate tablets for lower-extremity osteomyelitis in 337 patients [22] . Damaged bone and soft tissues were resected and calcium sulfate tablets were inserted in the dead space. Overall, 86.4% of patients were treated without intravenous antibiotics and 7.4% with intravenous antibiotic administration. The remaining 6.2% was treated with amputation. Furthermore, 70% of patients healed even without oral antibiotic administration.
Melamed and Peled investigated the use of an antibiotic-impregnated cement spacer in 23 cases of forefoot DFO [23]. Of the 23 cases treated by meticulous debridement and antibiotic-impregnated cement spacer implantation, 21 healed successfully, while the spacer was left permanently in 10 patients. Transfer lesions occurred in one patient only. In two patients, it was necessary to amputate the affected part that did not heal [23].
Walsh and Yates reported that 5 out of 7 Wagner grade 3 ulcers healed with calcanectomy plus gentamicin-impregnated collagen sponge or calcium sulfate with tobramycin [24]. Mean healing time was 64 days (range not stated).
Dalla Paola et al. used vancomycin-and gentamycin-loaded bone cement after surgical debridement with removal of the infected bone for first metatarsophalangeal DFO [25]. They reported healing in 24 of 28 patients without new ulceration, shoe fit problems, or gait abnormalities. Four patients developed a relapse of the ulceration and one of them underwent percutaneous revascularization and transmetatarsal amputation.
Jogia et al. reported 100% cure in 20 DFO patients who had failed to respond to routine wound debridement, systemic antibiotics, and offloading [26]. Their approach included excision of bone sequestrate and application of biodegradable highly purified synthetic calcium sulfate pellets containing vancomycin and gentamicin. Postoperative systemic antibiotic treatment was decided on an individual basis.
Panagopoulos et al. included 8 patients with chronic metatarsal or calcaneal DFO [3]. These patients were successfully treated with gentamicin delivery either with PMMA cement beads or bone graft substitutes. Local antibiotics were applied after minor surgery in combination with systemic antibiotics. Gentamycin beads were absorbed in ≤2 months without surgical removal. Wound healing was seen in 6 patients.
Elmarsafi et al. evaluated 27 DFO patients treated by PMMA and gentamicin/vancomycin-eluting cement spacers after surgical excisional debridement [27]. Among these patients, 20 spacers were successfully retained or exchanged. Of the 10 patients requiring spacer removal, 4 underwent removal with subsequent arthrodesis and 6 with subsequent pseudoarthrosis, while 8 required ipsilateral partial foot amputation not related to spacer use or removal.
Drampalos et al. successfully treated 12 consecutive patients with calcaneal DFO using bone debridement and local delivery in drilled tunnels of a gentamycinloaded absorbable calcium sulfate/hydroxyapatite biocomposite [28]. One patient required a subsequent flap operation and 6 needed vacuum-assisted closure. There was also one case of prolonged wound leakage. No major amputation was required.
Niazi et al. evaluated 70 DFO patients treated by debridement, local antibiotic-loaded calcium sulfate/ hydroxyapatite bio-composite, and systemic antibiotics based on intra-operative cultures [29]. This treatment resulted in a healing rate of 90%, and there was no recurrence of infection.

Discussion and conclusions
DFO remains difficult to treat. Therefore, add-on local antibiotics have been attempted post-operatively to improve outcomes [30]. However, definitive supportive evidence is still rare. In RCTs, gentamicinimpregnated collagen sponge significantly improved clinical cure rates and slightly improved duration of hospitalization. In RSs, antibiotic-impregnated calcium sulfate beads non-significantly improved all healing parameters, but did not reduce post-operative amputation rates or time to healing. The majority of case series used calcium sulfate beads, achieving adequate rates of healing and eradication of infection.
Importantly, there are limitations in available evidence. In fact, only 3 RCTs were identified [15][16][17]. Further limitations include small patient numbers, wide range of inclusion criteria, heterogeneity of DFO, and the variety of surgical techniques. Accordingly, large rigorously designed RCTs with clear inclusion criteria and procedures are required to shed more light on this issue. Finally, it is still unclear how patients who would most benefit from this add-on therapy can be identified or selected.
In conclusion, add-on local antibiotic delivery following surgery for DFO has achieved some favorable results, mainly healing and eradication rates. Nonetheless, evidence is still limited, while methods and criteria used in the studies have been heterogeneous. Certainly, local antibiotic delivery represents an important step towards improved wound treatment, but more robust evidence is needed, especially on its efficacy in DFO. If its efficacy is finally confirmed, this therapeutic adjunct will certainly enrich our armamentarium for one of the most dangerous diabetic complications [31,32].