Summary
Osteomyelitis is probably the oldest known infection in the history of life. It can develop secondary to local tissue disruption, ischemia and associated chronic wounds or via hematogenous infection. Although it has been known to the medical community for a long time, treatment remains challenging. Detection of the microbial agent remains crucial for the associated antibiotic therapy. Also, tissue specimens for culture and histology must be obtained. Several factors such as biofilm formation, resistance development and special virulence factors can impede the efficiency of the antibiotic treatment. In the last two decades, developments of antibiotic agents with available data in the field of osteomyelitis primarily include brilacidin, ceftaroline, ceftobiprole, dalbavancin, daptomycin, tedizolid, telavancin, tigecycline. Many of them are not on the market, or under study, or only found in selected countries. However, they are expected to become more accessible in coming years.
Introduction
Several classification schemes have been developed in the past to guide treatment. Following its etiology, osteomyelitis can be divided into three clinical entities[10]. Hematogenous osteomyelitis can also occur in adults and especially elderly patients, but affects mainly pediatric patients[10-12]. The most common form in adults is associated with local tissue disruption that can follow bone surgery or trauma[10, 12]. A third type is associated with vascular insufficiency and consecutive wounds of the lower extremities[3]. Foot ulcers often develop in diabetic patients, where neuropathy and metabolic changes add up to the infection susceptibility due to ischemia[3, 10]. Treatment of osteomyelitis can be purely medicamentous in some acute settings with absence of necrotic areas or abscess formation, but is mostly combined with surgical debridement in chronic cases[13, 14]. Effectiveness of antibiotic therapy depends not only on its availability and concentration at the infection site, but also on susceptibility of the infectious agent. Overuse, misuse and easy availability of antibiotics in combination with the property of bacteria to evolve when subjected to selective pressure has lead to resistance. The latter is increasing worldwide and also affects Europe and certainly the countries around the Mediterranean sea[15-17]. Unfortunately, only few new substances have been approved for osteomyelitis treatment in the last years, but several are currently being investigated[13].
Epidemiology
Staphylococcus aureus is the most common bacterium isolated from bone samples in pediatric osteomyelitis, and the incidence of Kingella kingae in little children is rising, probably due to better detection and diagnostic performances[11]. In adults, roughly half of osteomyelitis cases might be implant-associated[18]. Therefore, the antimicrobial spectrum differs from pediatric osteomyelitis[18], in adults being 33% related to S. aureus and 32% to coagulase-negative staphylococci. More rarely, other germs can be found depending on endemic bacteria and fungi can be associated with immune suppression[2]. In bedridden patients, pressure sores have a high incidence and can lead to osteomyelitis via direct infection of neighboring tissues[6]. They have a high risk of complications and recurrence. Interestingly, pathogens isolated in recurrence of osteomyelitis in the same bone are different from the initial microbiology in 86% of cases[19]. While hematogenous spread mostly leads to mono-bacterial infection, ulcer-related infection is usually polymicrobial. Table 1 gives an overview of possible pathogens of different population of osteomyelitis patients.
Diagnosis of Osteomyelitis
Laboratory Tests
As a general principle, the detection of pathogenic bacteria within bone samples remains the gold standard for the confirmation of clinical diagnosis of osteomyelitis. Histology is supplementary. Of note, the microbiological samples have to be done within clinically or radiologically infected bone. There are numerous studies advocating an acceptable concordance of repeated consecutive fistula samples with bone samples, indicating that if several fistula samples are the same, the underlying bone is likely to be infected with the same pathogen[20]. We validated this widespread attitude in France also for the subset of patients with (diabetic) toe osteomyelitis[21]. However, these specific attitudes are not accepted in the international community. While serum inflammation markers such as C-reactive protein (CRP) and erythrocyte sedimentation rate can be useful especially during follow-up of osteomyelitis, values within the normal does not exclude the latter. Especially, in the presence of fistula, diabetic foot and absence of systemic signs of infection such as fever, redness, or heat, the CRP values might be negative. We are not aware of any cut-off level or negative predicting values to exclude underlying osteomyelitis. Procalcitonin levels have not proven useful for diabetic foot osteomyelitis[22] or skeletal infection in children[23]. The microbial culture and resistance testing of the responsible infectious agent remains crucial. Histological samples as well as up to 5 soft-tissue samples for culture should be obtained[10]. Antibiotic prophylaxis given at the induction of anesthesia does not interfere with intra-operative sampling cultures[1]. If image-guided needle biopsy is obtained, at least 2 mL of fluid should be aspirated to improve sensitivity[24]. Depending on endemic and patient-specific factors, rarer pathogens like mycobacteria and fungi should be searched for. It is important to incubate implant-related samples for up to three weeks, as slow-growing bacteria like Propionibacterium acnes can otherwise be missed[25]. Lastly, cultures and histological examination are adjuncts to a clinical suspicion, but, when negative, cannot exclude osteomyelitis[26].
Radiologic Imaging
X-rays of bone lesions remain a minimum standard, but the sensitivity is low especially in early osteomyelitis[27]. Magnetic resonance imaging (MRI) is an excellent tool and provides additional information about soft tissues[26-29]. The sensitivity and specificity of FDG positron emission tomography-computed tomography (PET-CT) is superior to all other imaging methods, but its availability is limited in most parts of the world (Figure 1)[27]. On standard X-rays, the earliest visible changes include swelling of soft tissue, periostal thickening or elevation, and focal osteopenia. Before the radiographs show lytic changes, probably 50% to 75% of the bone matrix must be destroyed, which takes at least two weeks[27]. MRI is very sensitive and can show tissue edema and increased regional perfusion. However, these changes can last for a long time after surgery and distinction between fibrovascular scarring, "overuse syndromes", gout, neuropathic osteoarthropathy and reactive infection is often difficult. Thus, MRI lacks specificity, especially in the post-surgery setting or in diabetic foot alterations. Kaim et al.[30] reported a sensitivity, specificity, and accuracy of 100%, 69%, and 78% for MRI in chronic post-traumatic osteomyelitis. Today, CT scans are better for the visualization of sequestra and are less expensive than MRI.
Figure 1: Chronic osteomyelitis in the right femur
In the future, the most specific and accurate radiological exam for osteomyelitis could become PET[31, 32]. This performance was recently confirmed in another study with corresponding results of 100%, 76%, and 90%, respectively[33]. It is the most expensive radiological exam and lacks its established place in daily clinical life. To our best knowledge, there are no studies investigating the evolution of the metabolic signal post-surgery or during the antibiotic treatment of osteomyelitis. Scientific proof that PET permits to distinguish between physiologic remodeling and infection after a cut-off of 6 weeks post-surgery, has been studied only in the rabbit model[34]. Studying its quantitative signal during long lasting therapy in humans might become interesting, since it could help identifying patients for which a prolongation of therapy would not be necessary or it could detect zones of early collection for which a surgical re-intervention might become warranted. Scintigraphy has become less important harboring a low specificity for implant-associated infections[35]. Moreover, bone scintigraphy alone cannot distinguish between aseptic loosening and infection, and needs combination with a leukocyte-scintigraphy. Sensitivity, specificity and accuracy for a leucocyte-labeled scintigraphy are 63%, 97%, and 77% for implant-related osteomyelitis[36].
Treatment of Osteomyelitis
Surgical Treatment
Surgical treatment consists of radical debridement and lavage to diminish bacterial load and remove non-vital tissue. A relative wide resection with 5 mm clinical margins thereby diminishes recurrence[37]. Nevertheless, it should be as atraumatic as possible for adjacent soft-tissue covering[38]. Pulsed lavage irrigation has been shown to clear off bacteria more effectively than simple irrigation in animal studies[39]. To fill the remaining dead space, antibiotic-loaded polymethyl-methacrylate (PMMA) has been introduced almost 40 years ago[40]. The topical application allows very high local concentration without systemic side effects. Antibiotic-impregnated collagen fleece is widely used in clinical practice and shows higher release rates than PMMA, but shorter elution time[41]. Multiple other biodegradable substances have been developed during the last decades and are currently under investigation[42, 43]. Unfortunately, we are unaware of new developments regarding surgical techniques. As in the antiquity, the cornerstone of surgical management implicate amputation or at least the removal of all foci, especially sequestrate and fistulae, accompanied by intramedullar reaming or other techniques of intramedullar lavage[2].
Antibiotic Treatment
The efficiency of antibiotic treatment is dependent on the complex interaction of the drug, the host and the microbial agent[44]. Theoretically, the concentration on the target site is crucial for antibiotic action. Not only type of administration and bioavailability, but also pharmacodynamic (PD) parameters have been taken into account when modeling treatment efficiency[44]. In a recent review published in this journal, our research group summarized the latest insights in the pharmacokinetic (PK) PK/PD aspects of antibiotics in the bone, which we would like to reference our previous review in this journal for the more interested reader[2]. Furthermore, microorganisms have developed mechanisms to inactivate antibiotic action, including resistance development and biofilm formation. In nature, 60-80% of bacteria are believed to exist in biofilms, where they are embedded in amorphous biomatrix[45, 46]. The altered metabolism reduces susceptibility to antibiotics within hours[46]. Recent reports have shown that resistance developed during treatment in initially susceptible S. aureus can also be related to dynamic small colony variants. Some antibiotics like clindamycin, moxifloxacin and gentamicin can even induce the latter[47, 48].
In S. aureus infection, the expression of Pantone-Valentine leukocidin (PVL) is a virulence factor. Most community-acquired methicillin-resistant S. aureus (MRSA) strains produce PVL[49]. Its production can be induced by certain microbial agents, penicillin-binding protein 1 has to been shown to trigger the latter. While oxacillin increased the release 2.5-fold, combination with clindamycin, rifampicin and linezolid prevented from this effect[49].
Necessity of Parenteral Antibiotic Therapy?
In former times, experts usually recommended an intravenous (IV) therapy for 4 to 6 weeks[50] followed by an oral course of additional weeks or months. The belief for long periods of supplementary oral treatment evolved from cases of relapsing osteomyelitis in the 1970s which may be less frequent today due to improved surgical and antibiotic therapy. The rationale for a prolonged IV course was elevated serum concentrations. Today, the opinion has rather switched for IV treatment during the initial 2 weeks[51]. This initial two to six weeks of IV medication bases on experts" opinion rather than on clinical trials. Without doubt, bone penetration of antibiotic agents in parenteral administration is good and bioavailability per definition 100%[52]. At the same time, IV medication should be limited as far as possible to save unnecessary costs, prevent catheter-related complications and to increase patient and nursing comfort. The estimated proportion of complications attributed to prolonged IV course ranges around 15%[50]. Some antibiotics, such as ertapenem and ceftriaxone, could be administered via the subcutaneous route, but this route of administration is currently unlabelled. Prospective studies evaluating these points are urgently required.
Local Antibiotic-releasing Delivery Systems
The ideal local antibiotic delivery system is lacking[53, 54]. Antibiotic-containing cement is used for the treatment[53] and prophylaxis[53] of bone and prosthetic joint infections[54], but remains controversial in terms of additional benefit. Spacers for knee joint surgery may equally contain antibiotics[54]. All of these systems release antibiotics locally at concentrations exceeding up to one thousand times those of the minimum inhibitory concentrations (MICs) for the most common pathogens without releasing in the systemic circulation and without producing adverse effects[53]. However, the duration of time over which these antibiotics continue to be active and released is less certain. Moreover, the advantage appears minimal in two-stage procedures for arthroplasty infections[55]. Currently, there are few antibiotic-laden bone cement composites that have been approved by the Food and Drug Administration for clinical use: tobramycin, gentamicin, vancomycin, quinolones[53], cephalosporins[56], amphotericin B, and fluconazole[54]. Rifampin should not not be mixed with cement, since it may prolong the time to cement hardness by several hours. Gentamicin, the most frequently used antibiotic compound[57], may lead to development of small colony variants. Hand-mixing into cement is feasible to increase antibiotic dosage. The cement should be mixed first, and the antibiotics should then be added[54]. However, the addition of high doses of antibiotics (>4.5 g of powder) substantially weakens bone cement[54].
It is unknown if local antibiotic delivery could be equivalent to systemic administration of antibiotics. Few available data suggest an equivalent remission rates up to 78% in osteomyelitis patients treated with beads alone[57]. The major disadvantage of the PMMA beads is the presumed need for surgical removal, which usually takes place 3-4 weeks after their implantation[53]. Biodegradable implants are preferable to antibiotic-laden bone cement, because they do not require surgical removal. The PMMA is used in osteomyelitis to fill a bone gap, and facilitate the induction of a membrane (Masquelet) before bone grafting. Please note that some PMMA combine two antibiotics that could be synergic against the pathogen (gentamicin + vancomycin or gentamicin + clindamycin). Prospective studies are required to demonstrate that these new cements are better from a curative and a preventive point of view (prevention of superinfection). Finally, new local agents are absorbable. For instance, colleagues in Oxford developed a gentamicin-loaded, calcium sulphate/ hydroxyl-apatite biocomposite that is absorbable and proved to be effective in the treatment of chronic osteomyelitis[58].
Duration of Treatment
The duration of antibiotic treatment for osteomyelitis has traditionally been several months with initial IV treatment. No concluding evidence regarding these questions exists so far. While data are very sparse that a regimen shorter than 4 weeks shows higher risk of failure[59], there is accumulating evidence that therapy longer than 6 weeks does not improve the outcome[60-63]. Even in periprosthetic joint infection with retained material, 8 weeks has been shown to be non-inferior to longer treatments[64, 65]. When surgery cannot be performed (e.g. large infected area of the pelvis), antibiotics are only occasionally prescribed during one to several weeks upon clinical indication. This might be fever or increased purulent discharge. Importantly, the sinus tract should not be closed, because it represents the spontaneous drainage of a chronic infection.
Recent Antibiotics in the Market and Scientific Evaluation
Ceftaroline
This novel broad-spectrum cephalosporine is potent against MRSA[66]. It has been tested and approved safe and efficient in phase 3 studies for complicated skin and soft tissue infection and community-acquired pneumonia[66]. In a rabbit osteomyelitis model with MRSA it was superior to vancomycin[66]. Data for human osteomyelitis are numerous but anecdotal[67]. In these few reports and a recent review of roughly 180 osteomyelitis cases in the USA[67], ceftaroline reveals similar success rates to comparator drugs[67-69], but studies are clearly needed[69].
Daptomycin
One of the recently studied agents for osteomyelitis is daptomycin, a cyclic lipopeptide. Its dosage regimen of q24h during 2 minutes makes it favorable for outpatient IV treatment. The agent is currently on the market for treatment of Grampositive infections, including complicated soft tissue infections, S. aureus bacteremia and right-sided infectious endocarditis. Clinical effectiveness for osteomyelitis with MRSA might be non-inferior[70-72], or even better than vancomycin[73] or other new agents like tigecycline, dalbavancin, linezolid and telavancin. High-dose daptomycin (up to 8 or 10 mg/kg/day) in combination with or without rifampicin was most effective in treatment of implant-associated MRSA infections[74]. In a large retrospective report including 638 cases, it was shown to be safe and effective in patients with osteomyelitis or orthopaedic implant-related infections[75].
Telavancin
This lipoglycopeptide is a semisynthetic derivative of vancomycin[76]. It is approved in the USA for MRSA soft tissue infection[77]. In a rabbit osteomyelitis model with MRSA, telavancin shows lower MICs than linezolid and vancomycin, but similar efficacy[78]. In a case series of 4 patients with osteomyelitis all were treated successfully with telavancin and surgical intervention[79].
Tigecycline
This glycylcycline, a semi-synthetic tetracycline, is administered IV and is approved for skin infections, abdominal infections and community-acquired pneumoinia[80].
Sometimes it is employed against resistant pathogens and resistant pathogens in osteoarticular infections[80]. In a rat osteomyelitis model due to MRSA, tigecycline revealed at least as efficacious as teicoplanin[81]. Side-effects are mainly gastro-intestinal. In one phase 3 trial for osteomyelitis and chronically infected diabetic ulcers, it was inferior to ertapenem with or without vancomycin[82] but a binational multicenter retrospective cohort study in France and Turkey underlined its potential as a salvage therapy with prolonged administration (mean follow up 54 weeks) in 36 patients with various multiresistant bone and joint infections[83]. Tigecycline might be probably used in combination therapy, especially in case of Gram-negative multi-resistant infections or combined infection with Enterobacteriacae and staphylococci[83].
Promising Antibiotics in Development
Although the development of microbial resistance is not yet covered by respective effective antibiotic treatments, several new molecules are currently been studied in clinical trials for multi-resistant bacteria.
Brilacidin
This defensin-mimetic non-peptidic molecule has bactericidal efficacy for Gram-positive and Gram-negative bacteria, even in non-replicative state[84]. Phase 2-studies have been completed and phase 3 studies are underway for acute bacterial skin and skin structure infections[85]. Topical administration for ophthalmologic infection has been studied[86]. No data for osteomyelitis is available nor are studies currently planned to our knowledge. Due to its potential action in biofilms, this drug could be interesting for the treatment of implant-relatedb infection.
Ceftobiprole
It is the first broad-spectrum cephalosporin with bactericidal activity against MRSA and broad Gram-positive and -negative range. Its safety and non-inferiority to vancomycin has been shown for soft-tissue infections[87]. It was effective in vivo against PVL producing community-acquired MRSA osteomyelitis in rabbits[88]. Effectiveness for humans has only been shown in case-reports[89].
Dalbavancin
The application interval of dalbavancin is even more advantageous as the above-mentioned daptomycin. It is usually administered 1/ week IV, but a single-dose regimen is sufficient for complicated soft-tissue infection including MRSA[90]. Its action is bactericidal, the synthetic lipoglycopeptide blocking enzymes involved with polymerization and cross-linking of peptidoglycan[91]. It is approved and available on the market in the USA since 2014 and in Europe since 2015 for skin infections[91, 92]. Minimum inhibitory concentrations for S. aureus isolated from diabetic foot ulcers and vancomycin-intermediate and heteroresistant types were measured in vitro, dalbavancin showed excellent activity and was superior to vancomycin[93]. Its distribution to bone and synovial fluid was measured in healthy subjects and shows concentrations superior to the MIC of S. aureus over 50 days[94, 95].
Debio 1450
This FabI inhibitor has been developed for soft-tissue infection and osteomyelitis. It is derived from the crystal structure of the active site of the enzyme and is staphylococcus-specific[96]. It is currently being tested in a phase 2 study for acute bacterial skin and skin structure infections [NCT02426918]; no data for osteomyelitis is available.
Lefamulin
The semi-synthetic compound, a pleuromutilin-derivate inhibits bacterial protein synthesis by binding to the 50S ribosome. It was tested in a phase 2 trial for acute bacterial skin and skin-structure infections against vancomycin. It showed good efficiency against multi-resistant Gram-positive bacteria, including MRSA. It is expected to be available both parenterally and peroral. Studies are underway[97]; no large data for osteomyelitis is available.
Nemonoxacin
Non-fluorinated quinolone with a broad spectrum against atypical pathogens, Gram-positive and Gram-negative microbes. It has been released in Taiwan for community-acquired pneumonia, but not yet for its other field of development, diabetic foot infection[98]. Clinical phase 2 and 3 trials have been registered and are underway, intermediate results are promising, but definitive results missing and the substance is not yet approved for the latter[99].
Tedizolid
This oxazolidinone prodrug is converted to tedizolid in vivo. The molecule inhibits translation by binding to the bacterial 23S ribosome initiation complex[100]. Its spectrum covers Grampositive pathogens, including linezolid-resistant S. aureus[101]. It has been shown to be efficient, safe and well-tolerated in phase 3 studies and a post-hoc analysis for Latino patients[101-103]. Its efficiency has been studied in rat foreign-body osteomyelitis, but no data for human osteomyelitis exists[104].
Finally, there are many other new agents already on the market in some countries or not, but which are powerful candidates for the future: delafloxacin, finafloxacin, zabofloxacin, eravacycline, omadacycline, ceftazidime/avibactam, ceftolozane/tazobactam, and others that still have to prove their non-inferiority to current regimens and which might become more interesting and valuable candidates in future reviews.
Hyperbaric Oxygen Therapy
Hyperbaric oxygen therapy consumes very substantial resources[105]. It provides oxygen to promote collagen production, angiogenesis, osteogenesis, and healing in the ischemic or infected wound[106]. Animals receiving hyperbaric oxygen showed an acceleration in all phases of fracture repair[106]. Several authors have suggested that adjunctive hyperbaric oxygen therapy might be useful in the treatment of human chronic osteomyelitis, even if the results are not consistent though. The adjunctive role of hyperbaric oxygen in osteomyelitis is difficult to assess because of the multiple confounding variables of patient, surgery, organism, bone, and antibiotics. Today, although recognized for reimbursement by some insurers, the evidence base for hyperbaric oxygen therapy for diabetic foot care still remains weak[105].
Conclusion
For future developments, the results of the attended studies have to be included in clinical practice. Other interesting areas include the evaluation of absorbable antibiotic carriers for local therapy. Host-defense protein imitating drugs are ideal candidates for further research due to their low risk of development of resistance[107]. Raising economic pressure on medical treatment establishes the need for prolonged antibiotic therapy on an outpatient base. Even if modern treatment algorithms may cure the majority of acute and chronic bone infections, the risk of relapse or persisting disability is high. Finally, given that the prevalence of glucose-intolerance in Europe has risen to 22%[108] and treatment of post-traumatic osteomyelitis remains challenging, treatment of osteomyelitis will remain an important clinical issue.
Acknowledgements
We are indebted to our colleagues from the Orthopedic Service, the Service of Infectious Diseases and the Research facilities of Geneva University Hospitals for their help.
Ethics
Peer-review: Externally and internally peer-reviewed.
Authorship Contributions
Surgical and Medical Practices: S.H., İ.I., İ.U., Concept: S.H., İ.I., İ.U., Data Collection or Processing: S.H., İ.I., İ.U., Analysis or Interpretation: S.H., İ.I., İ.U., Literature Search: S.H., İ.I., İ.U., Writing: S.H., İ.I., İ.U.