Osteomyelitis is diagnosed in 50–60% of patients who have been hospitalised for diabetic foot infections (Lipsky et al, 2016). Besides standard-of-care offloading, optimisation of glycaemic control and arterial vascularisation, the treatment of diabetic foot osteomyelitis (DFO) involves long-term systemic antibiotics and surgical debridement (Lipsky et al, 2016). Diabetic foot ulcers with osteomyelitis are difficult to heal, resulting in high amputation rates (6–14%) (Mutluoglu et al, 2013).
Recently, the local antibiotic therapy antibiotic-loaded calcium sulphate/hydroxyapatite (CaS/HA) has been used to treat DFO with promising initial results (Karr, 2011; Whisstock et al, 2015; Drampalos et al, 2017; Niazi et al, 2019) and this was the treatment used in the following case study.
Case study
The patient was a 49-year-old white woman who had hypertension and type 1 diabetes for 20 years with peripheral neuropathy but without peripheral artery disease. She used valsartan (160 mg once daily), hydrochlorothiazide (25 mg once daily) and insulin (glargine, Lantus® and aspart, NovoRapid®). She had no allergies. Her right foot was severely deformed due to Charcot neuroarthropathy-related collapse and transmetatarsal amputations of the first and second rays and cuboid exostectomy due to recurrent infected foot ulcers (Figure 1). Her previous ulcers had not been complicated by DFO. Despite the foot deformities, the patient was able to mobilise and weight-bearing was possible with orthopaedic footwear.
The patient presented at a multidisciplinary diabetic foot unit with a recurrent plantar ulcer which was about 3 cm2 located centrally over the collapsed midfoot. Pre-existent oedema of the entire foot was seen with generalised erythema and a locally elevated temperature on the dorsum of the midfoot. Instability of the foot was clearly present in the tarsometatarsal and midfoot joints.
It was possible to reach multiple dislocated loose bone fragments on the midfoot level using forceps. The patient’s vital signs were normal with a temperature of 37.7°C. PEDIS classification (Schaper, 2004) of the ulcer was P1 (no symptoms or signs of peripheral artery disease in the affected foot and palpable dorsal pedal and posterior tibial arteries), E3 (ulcer size of 3 cm2), D3 (ulcer involvement of all subsequent layers of the foot, including bone and joints), I3 (erythema >2 cm surrounding the ulcer, swelling, warmth, discharge and suspicion of osteomyelitis but no systemic inflammatory response signs), S2 (loss of protective sensation on the affected foot).
The authors hospitalised the patient and prescribed intravenous amoxicillin/clavulanate (1,000/200 mg, four times daily). Laboratory findings showed a normal white blood cell count (7.6 x 10^9/L) and an elevated C-reactive protein level (20 mg/L). Magnetic resonance imaging (MRI) showed inflammation of the soft tissues around the ulcer with an underlying abscess and oedematous bones and bone marrow on the level of the collapsed midfoot (Figure 2). These findings confirmed osteomyelitis of the midfoot but did not rule out secondary Charcot neuropathy, which was suspected due to the instability of the foot, oedema, erythema and warmth.
The authors advised offloading with strict bedrest to prevent further damage to the foot. Urgent surgical debridement was performed with drainage of the abscess and partial resection of the infected remnants of the lateral cuneiform bone underlying the ulcer tract. Gentamicin-impregnated collagen sponges were put in place and the wound was partially closed. Further resection of the infected tarsals was not performed because we expected this could cause damage that could render the foot non-functional. Bedrest was continued postoperatively and intravenous antibiotics flucloxacillin (12 g, once daily) and oral ciprofloxacin (500 mg, twice daily) were prescribed based on cultures of the resected bone, positive for Staphylococcus aureus and Klebsiella pneumoniae (resistant to amoxicillin/clavulanate and piperacillin/tazobactam). No further healing was observed after 2 weeks and the oedema, erythema, elevated temperature and midfoot instability persisted. The authors discussed lower leg amputation with the patient, which she refused.
As a final attempt for limb salvage, an extensive surgical debridement of the ulcer and infected bones of the midfoot was performed, which included complete resection of the cuboid bone, after which the bone void was filled (about 18 cm3) with gentamicin-loaded CaS/HA, and the skin was closed (Figure 3). The aim was to eradicate the infection and to regain osseous stability of the midfoot joints. The operation was performed under general anaesthesia and a pressure tourniquet was used. The authors prescribed 5 days of strict bedrest postoperatively after which total contact cast (TCC) immobilisation was started.
Serous wound leakage persisted for 3 weeks, which required a removable TCC to allow wound care. This delayed hospital discharge by two weeks. Systemic amoxicillin/clavulanate (500/125 mg, three times daily) was continued postoperatively despite resistance of the isolated Klebsiella pneumonia, since clindamycin and ciprofloxacin were not tolerated by the patient. Intra-operative bone cultures were negative.
The authors observed gradual healing of the ulcer and improvement in MRI findings at monthly outpatient visits. At 4 months, ulcer healing was complete and systemic antibiotics were stopped. At 7 months postoperatively, normal foot stability without signs of inflammation was found. MRI showed further reduction of oedema of the tarsals and the underlying soft tissues (Figure 4). The authors then stopped TCC immobilisation and the patient started weight-bearing mobilisation with custom-made orthopaedic footwear. At final follow-up, 1-year postoperatively, there were no recurrent ulcers or signs of Charcot neuroarthropathy (Figure 5) and radiography showed no signs of residual osteomyelitis (Figure 6).
Discussion
Local antibiotic treatment
Various techniques and materials have been developed for local antibiotic treatment since the introduction of joint arthroplasties in the 1970s. Non-biodegradable polymethyl methacrylate (PMMA) in preformed beads has mostly been used as a carrier material for local antibiotic treatment (Gogia et al, 2009). Filling of bone voids with antibiotic-loaded PMMA beads is easy and it improves dead space management after surgical debridement and reduces the need for systemic antibiotics.
Disadvantages are partial retention of antibiotics in the PMMA beads and the requirement of further surgery to remove the beads after treatment (Gogia et al, 2009).
Many biodegradable materials have been developed to improve local antibiotic treatment, ranging from protein-based materials, such as collagen, to bone grafts and bone graft substitutes, such as calcium sulphate and hydroxyapatite, to synthetic polymers (McLaren, 2004). CaS/HA composite (Cerament™, BoneSupport) has recently been introduced as a biodegradable bone graft substitute. This material acts as an osteoconductive scaffold to improve healing of bone defects and can be loaded with antibiotics (gentamicin or vancomycin) for local antibiotic treatment (Nilsson et al, 2013).
Previous studies showed that a gradual release of gentamicin from CaS/HA results in a local concentration above the minimal inhibitory concentration for about 4 weeks (Stravinskas et al, 2016). Complete reabsorption of the material requires about 2 years (Stravinskas et al, 2016).
Treatment outcomes in previous studies
Initial results of the treatment of post-traumatic osteomyelitis with surgical debridement and bone void filling with antibiotic-loaded CaS/HA are promising (McNally et al, 2016; Stravinskas et al, 2016). Specific results of this treatment for DFO were studied by Niazi et al (2019), where 63 out of 70 patients with DFO achieved wound healing by surgical debridement and bone void filling with gentamicin-loaded CaS/HA, in addition to culture-specific systemic antibiotics. Another study by Drampalos et al (2017) described the healing of 12 patients with calcaneal DFO after surgical debridement and holes drilled in the adjacent bone and filled with gentamicin-loaded CaS/HA in combination with systemic antibiotics for 6–12 weeks. Whisstock et al (2015) described the successful treatment of 16 out of 20 patients using surgical debridement and gentamicin-loaded CaS/HA bone void filling without systemic antibiotics. Karr (2011) reported the successful treatment of a patient with methicillin-resistant Staphylococcus aureus DFO using surgical debridement including minor amputations and filling of the bone voids with vancomycin-loaded CaS/HA beads, in addition to oral trimethoprim/sulfamethoxazole for 2 weeks.
Charcot neuroarthropathy
The success of the treatment in this case study supports these previously published successes. The management of the patient was more complicated due to her severely deformed and deformed foot due to chronic Charcot neuroarthropathy, which is a reason for exclusion in most studies on the management of diabetic foot infections. Charcot neuroarthropathy-related foot deformation increases the risk of recurrent plantar ulcers and leads to a higher risk of lower-limb amputation compared with people with diabetic foot ulcers without Charcot neuroarthropathy (Sohn et al, 2010). Suspected secondary Charcot neuroarthropathy activity further complicated this treatment, requiring seven months of TCC immobilisation.
The observed symptoms of inflammation at presentation — oedema, elevated temperature, erythema and instability of the tarsometatarsal and midfoot joints — were suggestive of Charcot neuroarthropathy activity but obviously overlap with symptoms of extensive infection. MRI did not distinguish between Charcot neuroarthropathy activity and osteomyelitis in this case since bone and bone marrow oedema on the midfoot level is present in both pathologies (Ertegrul et al, 2013). These MRI findings in the presence of an underlying ulcer confirm osteomyelitis, but do not rule out secondary Charcot neuroarthropathy activity. Of the Charcot neuroarthropathy-related symptoms, mainly the instability improved slowly, requiring seven months of TCC immobilisation, even though the wound healing was complete after four months without signs of persistent infection. Ignoring secondary Charcot neuroarthropathy activity may have led to early weight-bearing mobilisation of the foot with the risk of further damage.
Antibiotics
In this case study, systemic antibiotics were prescribed perioperatively and postoperatively for a duration of 4 months. The literature suggests that the treatment of osteomyelitis with antibiotic-loaded calcium sulphate-based bone graft substitutes can be successful without systemic antibiotics (Gauland, 2011). However, most previous studies on local antibiotic treatment of osteomyelitis by antibiotic-loaded CaS/HA provided systemic antibiotics for about 3 months (McNally et al, 2016; Drampalos et al, 2017).
The authors prescribed systemic antibiotics since we were aiming for limb salvage in a severely threatened foot with a difficult-to-treat infection and it was decided that it would be irresponsible to stop the antibiotics until all signs of infection were absent and wound healing was complete.
Initial clinical experience
The authors found surgical debridement and gentamicin-loaded CaS/HA bone void filling quick and easy to achieve. No local or systemic complications of the material or the required procedures occurred. Serous wound leakage was witnessed for about 3 weeks, which has been reported in previous studies of antibiotic-loaded CaS/HA and does not indicate persistent or recurrent infection (Drampalos et al, 2017). The authors performed surgical debridement and subsequent gentamicin-loaded CaS/HA bone void filling as a second stage after initial standard-of-care was unsuccessful. Although unintended, multiple stages of surgical treatment are often required in the standard treatment of DFO.
Conclusion
In this case study, the authors presented the successful treatment of a difficult-to-manage diabetic foot ulcer with DFO in a severely deformed foot complicated by Charcot neuroarthropathy, by using surgical debridement and gentamicin-loaded CaS/HA bone void filling, offloading and systemic antibiotics after unsuccessful standard treatment. This result adds credibility to the promising initial results showing how effective surgical debridement and antibiotic-loaded CaS/HA bone void filling can be when treating DFO. This treatment has potential benefits, when compared with the current conventional treatment, by enabling single-stage treatment and making long-term systemic antibiotics unnecessary. More research is required regarding potential procedures, postoperative treatment, such as offloading and systemic antibiotics, and comparisons with conventional treatment of DFO before this treatment modality can be implemented as an addition to standard care. For now, it should be considered for patients who do not heal after standard-of-care treatment for DFO to prevent amputations in this vulnerable group of patients.
