• diabetes mellitus;
  • diabetic foot;
  • infection;
  • osteomyelitis;
  • antibiotics;
  • surgery;
  • systematic review


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of Interest
  9. References
  10. Appendix A
  11. Appendix B

The International Working Group on the Diabetic Foot expert panel on infection conducted a systematic review of the published evidence relating to treatment of foot infection in diabetes. Our search of the literature published prior to August 2010 identified 7517 articles, 29 of which fulfilled predefined criteria for detailed data extraction. Four additional eligible papers were identified from other sources. Of the total of 33 studies, 29 were randomized controlled trials, and four were cohort studies.

Among 12 studies comparing different antibiotic regimens in the management of skin and soft-tissue infection, none reported a better response with any particular regimen. Of seven studies that compared antibiotic regimens in patients with infection involving both soft tissue and bone, one reported a better clinical outcome in those treated with cefoxitin compared with ampicillin/sulbactam, but the others reported no differences between treatment regimens. In two health economic analyses, there was a small saving using one regimen versus another. No published data support the superiority of any particular route of delivery of systemic antibiotics or clarify the optimal duration of antibiotic therapy in either soft-tissue infection or osteomyelitis. In one non-randomized cohort study, the outcome of treatment of osteomyelitis was better when the antibiotic choice was based on culture of bone specimens as opposed to wound swabs, but this study was not randomized, and the results may have been affected by confounding factors.

Results from two studies suggested that early surgical intervention was associated with a significant reduction in major amputation, but the methodological quality of both was low. In two studies, the use of superoxidized water was associated with a better outcome than soap or povidone iodine, but both had a high risk of bias. Studies using granulocyte-colony stimulating factor reported mixed results. There was no improvement in infection outcomes associated with hyperbaric oxygen therapy. No benefit has been reported with any other intervention, and, overall, there are currently no trial data to justify the adoption of any particular therapeutic approach in diabetic patients with infection of either soft tissue or bone of the foot. Copyright © 2012 John Wiley & Sons, Ltd.


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of Interest
  9. References
  10. Appendix A
  11. Appendix B

Infection of the foot is a common complication in patients with diabetes mellitus, and it can lead to significant morbidity (including lower extremity amputation) and mortality. Several groups have developed guidelines for treating diabetic foot complications on the basis of the limited available published data. The Infectious Diseases Society of America has developed evidence-based guidelines specifically directed at managing diabetic foot infections (DFI), but the recommendations are not based on a formal systematic review of the literature.

Two systematic reviews of some aspects of DFIs have been published. In 2008, the International Working Group on the Diabetic Foot conducted a systematic review of treatment of diabetic foot osteomyelitis [1], and in 2011 The National Institute for Health and Clinical Excellence (NICE, UK) published the results of a systematic review of the management of all aspects of care for inpatients with a diabetic foot complication [2]. The present systematic review includes an update of the 2008 osteomyelitis guideline but is extended to include all types of bacterial DFIs. This review focuses on studies of therapeutic interventions and does not cover definitions for infection, methods for diagnosis (clinical, imaging or microbiological sampling) and the interface between critical colonization and infection.


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of Interest
  9. References
  10. Appendix A
  11. Appendix B

Literature search was conducted using PubMed and Embase, seeking all prospective and retrospective studies in any language that evaluated interventions for the treatment of foot infections in people aged 18 years or older with diabetes mellitus. The search strategy employed is described in Appendix A. Eligible studies included randomized controlled trials (RCTs), case–control studies, prospective and retrospective cohort studies and those of interrupted time series (ITS) or controlled before-and-after design (CBA). Uncontrolled case series, studies in which controls were historical and case reports were excluded. Studies where patients with DFIs formed part of the total population were excluded if the data for the subgroup with diabetes were not separately described.

One author assessed each identified reference by reviewing the title and abstract for potential eligibility. Full copies of potentially eligible publications were independently reviewed by two authors to determine whether they met the criteria for being included. When the two reviewers disagreed, they worked to reach consensus, sometimes with input from another reviewer. Using specially prepared forms, the reviewers noted the study design, characteristics of patient populations, type(s) of interventions, all outcomes and the duration of follow-up of included patients. Investigators scored all studies for methodological quality using scoring lists developed by the Dutch Cochrane Centre [3]. Quality items were rated as ‘done’, ‘not done’, or ‘not reported’, and only those rated as ‘done’ contributed to the methodological quality score. Equal weighting was applied to each validity criterion for every study design.

The methodological quality score was translated into a level of evidence according to the Scottish Intercollegiate Guidelines Network (SIGN) instrument as follows: (1) RCTs and (2) case–control, cohort, CBA or ITS studies. Studies were also rated as: ++ (high quality with low risk of bias), + (well conducted with low risk of bias) and − (low quality with higher risk of bias). Co-reviewers agreed on the findings from the data extraction and the evaluation of methodological quality of each paper. Extracted data were summarized in the evidence table (see Appendix B) and described on a study-by-study narrative basis. Because of the heterogeneity of study designs, interventions, follow-up and outcomes, no attempt was made to pool the results. This evidence table was compiled following collective discussions (by electronic and in-person conferences) by all members of the working party.


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of Interest
  9. References
  10. Appendix A
  11. Appendix B

A total of 7517 papers were identified in the initial search: 4549 in PubMed and 2968 in Embase. After reviewing the titles and abstracts and excluding duplicate citations, we selected a total of 509 papers (460 papers in English, 26 in Russian, six in Ukrainian, six in Spanish, four in German, four in French, two in Chinese and one in Bulgarian) for full paper review. Of these, 29 papers met the criteria for inclusion. All of these papers were in English except for one which was in Chinese. Four additional papers that were not identified by our search strategy were added manually [4-7]. The data of all included papers are summarized in the evidence table (See Appendix B).

Types of study

Of the 33 studies, 29 were RCTs, and four were cohort studies. Of the 29 reported RCTs, one was actually a description of two studies in one article [8]. In some reports, patients with diabetes and a foot infection formed a subgroup of a larger group of, for instance, patients with skin and soft-tissue infections; these studies were excluded if sufficient detail was not specifically provided on the diabetic foot subpopulation. Twelve studies reported on the use of antibiotics in skin and soft-tissue infection, eight on patients with DFIs including osteomyelitis, of which one study was on the use of bone biopsy [9]. Treatment with topical antiseptic agents was reported in three studies. There were two studies of the role of surgery in DFIs and two that reported on the financial costs of antibiotic use. There were four studies that investigated treatment with granulocyte-colony stimulating factor (G-CSF); one additional paper on the use of G-CSF had not been identified in the literature search because it was filed as a letter to the editor rather than as an original study. The data of this study were extracted and added to the evidence table [6]. One study on the intramuscular administration of procaine plus polyvinylpyrrolidine was included, as well as one on the use of hyperbaric oxygen therapy (HBOT).

Individual topics

Early surgical intervention

The two papers that reported on this topic were both single-centre cohort studies of the effect of early surgery and antibiotics versus antibiotics alone in deep foot infections with and without osteomyelitis [10, 11]. Both studies suggested that there was a significant reduction of major amputation, from 27% to 13% in one study [10] and 8% to 0% in the other [11], with early minor surgery. Both studies examined outcomes associated with earlier surgery and not the particular indication for operative intervention. Because of the high risk of selection bias in deciding on which patients underwent early surgery in both studies, we find it hard to draw any conclusions from these data.

Health economics

Two studies explored the cost-effectiveness of different antibiotic regimens. The first was a cost-minimization assessment comparing treatment with ertapenem versus piperacillin/tazobactam [12] and was a subgroup analysis of a larger RCT [13]. Because piperacillin/tazobactam requires a more frequent dosing schedule than ertapenem, total costs for this regimen, including those for drug preparation and administration, were higher. The difference in cost per patient per day was, however, only of the order of $6. The second study, which explored cost-effectiveness in subjects admitted to hospital with skin and soft-tissue infection, reported a total potential cost saving of $61 per subject treated with ceftriaxone and metronidazole as opposed to ticarcillin/clavulanate [14].

Topical treatment with antiseptic agents

Two small single-centre RCTs have compared topical treatment with superoxidized water with either soap or povidone iodine. One of these studies was in patients with infected diabetic foot ulcers and outcomes of interest, i.e. odour reduction, cellulitis and extent of granulation tissue, were significantly better in the group of patients treated with superoxidized water than in controls treated with another topical disinfectant [15]. Of note, there was 81% reduction in periwound cellulitis in the intervention group versus 44% reduction in the controls. The other study was non-blinded and was conducted in patients with post-surgical wounds [16]. The duration of antibiotic treatment was significantly longer in patients treated with povidone iodine compared with those treated with superoxidized water (15.8 days versus 10.1 days; p = 0.016). Both studies included long-term outcomes of wound healing, but neither specifically addressed the potentially negative effect of other topical disinfectants in the comparator groups. One additional study in 30 subjects compared the results of a single application of a topical antiseptic, either iodophor and rivanol, with a control group [17]. There was a significantly reduced growth of bacteria after 24 h in the iodophor group compared with either the rivanol or control group, but the short follow-up and strictly microbiological (rather than clinical) outcome criteria limit the clinical usefulness of this study.

Granulocyte-colony stimulating factor

Four single-centre RCTs examining the adjunctive use of G-CSF in DFIs were identified [18-20]. A fifth study was published as a letter to the editor [6]. Patients had soft-tissue infection in four of the five studies and associated osteomyelitis in one [19]. In two studies, the design was double blinded; in one case the assessor was blinded, and in one only the patient was blinded. Blinding was mentioned in the fifth. In the study by Viswanathan et al. [6], a total of 85 patients treated with 5 µg/kg or a fixed dose (263 µg) of G-CSF were compared with 82 controls not treated with G-CSF; both groups received antibiotics and appropriate surgical wound care. Time to infection resolution was significantly lower for subjects who received G-CSF in the one study [21] but not in the others. This study [21] also reported a shorter duration of intravenous antibiotic use in G-CSF-treated patients, but this was not observed in another study [18]. Hospital length of stay was shorter for the G-CSF group in two studies [6, 21] but not in a third [18]. The need for surgical intervention was not statistically different between the two groups in the three studies that examined it [6, 19, 21] and neither was the time to eliminate pathogens from the wound [19, 21]. The results of these five studies are somewhat inconsistent and provide no clear evidence on which, if any, patients with DFIs might benefit in some clinically important way from the use of G-CSF. A meta-analysis of these five studies also concluded that adding G-CSF did not significantly affect the likelihood of resolution of infection or wound healing or the duration of systemic antibiotic therapy but was associated with a significantly reduced likelihood of lower extremity surgical interventions, including amputation and a reduced the duration of hospital stay [22].

Procaine plus polyvinylpyrrolidone

One study assessed intramuscular injection of 0.15 mL/day of procaine and polyvinylpyrrolidone for 10 days in 118 patients with a DFI affecting an ischaemic limb [23]. This observer-blinded, single-centre RCT found no significant difference between groups.

Hyperbaric oxygen therapy

Although a number of trials that have examined the effect of HBOT in patients with diabetic foot complications, including two double-blind RCTs [24, 25], were found, only one study was found that specifically investigated infection as an outcome [26]. This single-centre, open label study in patients receiving standard antibiotic treatment and wound debridement compared outcomes in 15 patients who received HBOT with 15 control subjects. Although it was not explicitly stated that the subjects had a foot infection, this was implied by the use of antibiotics. There were no significant differences in the numbers of positive wound cultures, major and minor amputations or hospital stay between the intervention and control groups.

Antibiotic choice based on bone biopsy

A single cohort study explored the effect of basing antibiotic selection on the results of culture of a bone biopsy specimen in patients with diabetic foot osteomyelitis [9]. Among 50 subjects, 32 had had previous unsuccessful treatment for osteomyelitis. The rate of remission of infection was significantly higher in the group for whom antibiotic choice was based on bone culture than in those in whom therapy was based on wound swab culture [82% vs 50%, respectively (p = 0.02)]. Nevertheless, it is possible that this difference was the result of confounding variables, especially the fact that patients in one of the highest enrolling centres only received a rifampicin-containing regimen if they had a bone culture.

Comparison of antibiotic regimens – skin and soft-tissue infection alone

Of the available studies comparing different antibiotic treatment regimens for skin and soft-tissue infections, 11 were RCTs, and one was a prospective cohort study [27]. Among the randomized trials, nine were multicentre studies [4, 7, 8, 28-33], whereas two were single-centre trials [14, 34]. Trial design was double blinded in three [4, 8, 32], investigator blinded in two [29, 31] and non-blinded in six [7, 14, 28, 30, 33, 34]. Three studies were subset analyses of larger trials [4, 7, 32]. One reported on two consecutive studies of the topical antibiotic peptide pexiganan [8]. The other studies compared systemic antimicrobial regimens: one compared two oral antibiotic regimens [34], whereas the rest involved parenteral regimens, often with a switch to oral antibiotic therapy.

The following classes of antibiotics were compared in the various studies: first, third and fifth generation cephalosporins (cephalexin, ceftriaxone and ceftobiprole, respectively); fluoroquinolones (ofloxacin, levofloxacin, ciprofloxacin and moxifloxacin); lincosamides (clindamycin); extended-spectrum penicillins plus beta lactamase inhibitors (piperacillin/tazobactam, ticarcillin/clavulanate, amoxicillin/clavulanate); carbapenems (ertapenem); nitroimidazoles (metronidazole); lipopeptides (daptomycin); and glycopeptides (vancomycin). Each of these antibiotic agents (except ceftobiprole) is widely used.

The mean duration of administration of antibiotic in patients with skin and soft-tissue infection in the two studies that reported on this ranged from 6 to 27 days [8, 14]. In the study of oral regimens, the duration of administration was only 2 weeks, although three patients were actually treated for longer [34]. No differences between the regimens were observed in the ten studies with regard to infection outcome, duration of hospital admission or rates of amputation. Clinical cure rates in all studies without osteomyelitis ranged from 48% [29] to 90% [8]. One RCT of mildly infected diabetic foot ulcers reported that a topical antibiotic, pexiganan, was similar in clinical and microbiological effectiveness to an oral fluoroquinolone, ofloxacin, with fewer adverse effects [8]. No study demonstrated a significant benefit for any specific antibiotic agent, route of administration or duration of treatment.

Comparison of antibiotic regimens – studies including patients with osteomyelitis

In addition to the previously mentioned cohort study of the use of bone biopsy in patients with osteomyelitis [9], there were seven studies of antibiotic treatment of DFI in which a proportion the patients had infection of underlying bone [5, 13, 35-39]. The other seven studies were RCTs: three were double blind, one was single blind, three were open label, four were multicentre, and three were single-centre trials. The prevalence of osteomyelitis varied from 6% [8, 13, 29, 36] to 81% [5]. The antibiotic classes compared in these studies were as follows: penicillins plus beta lactamase inhibitors (parenteral ampicillin/sulbactam and oral amoxicillin/clavulanate); extended-spectrum penicillins plus beta lactamase inhibitors (piperacillin/tazobactam); carbapenems (imipenem/cilastatin, ertapenem); second generation cephalosporins (cefoxitin); fluoroquinolones (ofloxacin, moxifloxacin); and oxazolidinones (linezolid).

Outcomes investigated included clinical cure [5, 13, 36-39], adverse drug reactions [5, 13, 37-39] and duration of antibiotic therapy [5, 36]. Only one study, a comparison of ampicillin/sulbactam with cefoxitin, reported a difference in clinical and microbiological outcomes [35]. The clinical cure rates in this study were significantly different (p = 0.03) but were exceptionally low, and there were no significant differences between the groups in bacteriological response (100% vs 73%), amputations (eight in each) or duration of hospitalization (21 vs 12 days). In the other studies in which patients with osteomyelitis were included, clinical cure rates (variously defined) ranged from 61% [38] to 94% [13, 39]. The mean duration of antibiotic treatment in the six studies was short, ranging from 6 days [35] to 28 days [5]. No study demonstrated a significant advantage of any particular antibiotic agent or route of administration in diabetic foot osteomyelitis.


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of Interest
  9. References
  10. Appendix A
  11. Appendix B

In planning this review, a search was made only for studies in which a treatment of DFI was compared with a contemporaneous control group, with studies being included only if at least the outcome data of the (sub)population of subjects with diabetes were reported. This led to the identification of only a very small number of suitable publications. It has to be accepted that trial design can pose problems in attempts to determine the effectiveness of different treatments in this field, and this is especially true for studies intended to evaluate the role of surgical interventions. Early surgery is accepted as essential in some cases of foot infection, and yet the trial evidence to substantiate the benefit is weak and based on just two studies – each of which had a very a high chance of bias. Another caution attaches to the use of the SIGN criteria for documenting study quality. This system ranks work mainly on the quality of study design, rather than study conduct, and this can result in apparent anomalies – with weaker studies occasionally achieving higher scores.

In most of the clinical trials that evaluated the efficacy of various antimicrobial agents, patients with DFIs were either excluded or comprised a small proportion of the study population. The design of some clinical trials allowed a post hoc analysis focusing on the subset of patients with a DFI, but the potential for confounding issues and the small number of subjects limits their usefulness. Not only is the number of reasonably designed studies in this field remarkably small, but most had a low score for study design, were marred by the use of small and heterogeneous populations, were poorly described or had a high risk of bias. Thus, readers should be cautious in interpreting the results of the available published work. Furthermore, circumstances dictating the choice of treatment in different countries and settings will vary according to the behaviours of affected population, nature of the presentation of infection, prevalence of different microorganisms and their antibiotic sensitivities. Selection of treatment is also severely restrained by limitation of resources in many parts of the world and poses particular problems in the management of those who live far from urban centres.

Notwithstanding the limitations, it is possible to draw several important conclusions. The reported data on skin and soft-tissue infection confirmed earlier observations suggesting that Gram-positive microorganisms play the leading role in DFI. There is, however, emerging observational evidence that Gram-negative species (especially Pseudomonas aeruginosa) are frequent pathogens in some populations, especially those in warm climates and developing countries [40-42]. If confirmed, this would have an important impact on the selection of antibiotic regimens. The available published data also suggest that it is possible to treat selected patients with a DFI in an outpatient setting with an oral antibiotic regimen, either initially or after a switch from parenteral therapy. The study of a topical antibiotic, pexiganan, is promising, but this agent will need to undergo further testing before it can be evaluated for approval. We identified few new data on the management of diabetic foot osteomyelitis since our relatively recent systematic review [1]. There is a great need for studies in patients with diabetic foot osteomyelitis to define the need for surgical intervention, the optimal antibiotic agents and the duration of therapy.

In the studies reported here, it was also of note that no great difference was observed in comparisons between antibiotic regimens with a relatively broad compared with a narrower spectrum of activity. It is also noteworthy that the randomized comparisons of antibiotic regimens generally reported a rather short duration of treatment (usually less than 2 weeks) – even in the few patients with bone infection– yet reported good outcomes. These observations, which conflict with most current clinical practice regarding the duration of antibiotic therapy especially in patients with osteomyelitis, need to be formally tested.

This systematic review makes clear the need for more studies on treatment of DFIs. We particularly need prospective studies that are robust, well-designed, comparative trials. These should be aimed at helping clinicians make an optimal choice of both empiric and targeted antibiotic regimens in various situations, including the choice of specific agents, the route and the duration of administration. Such studies should use a validated system for defining and classifying infections [43, 44] and be designed to evaluate all relevant clinical, microbiological, financial and other outcomes for both soft-tissue and bone infections.


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of Interest
  9. References
  10. Appendix A
  11. Appendix B

We thank Dr. Oleg Udovichenko, Russia, and Prof. Zhangrong Xu, China, for their help in translating papers published in languages other than English.

We thank the following, who served as corresponding members of the Expert Panel:

  • Dr Z. G. Abbass, Tanzania
  • Dr F. J. Aragón Sánchez, Spain
  • Dr M. Eneroth, Sweden
  • Dr B. M. Ertugrul, Turkey
  • Dr H. Gawish, Egypt
  • Dr I. Gurieva, Russia
  • Dr A. Jirkovska, Czech Republic
  • Dr F. de Lalla, Italy
  • Dr S. Kono, Japan
  • Dr A. Nather, Singapore
  • Dr J.-L. Richard, France
  • Dr N. Rojas, Chile
  • Dr L. Tudhope, South Africa
  • Dr S. Twigg, Australia
  • Dr V. Viswanathan, India
  • Dr O. Udovichenko, Russia
  • Dr D. Yue, Australia
  • Dr Z. Xu, China

Conflict of Interest

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of Interest
  9. References
  10. Appendix A
  11. Appendix B

BAL has served as a consultant to Merck, Pfizer, Cubist, DiPexium, Johnson & Johnson. ARB was a Pfizer Visiting Professor in 2011 to the University of Washington Department of Allergy and Infectious Diseases, an educationally unrestricted programme for which funding was awarded to DAID on a peer reviewed and competitive basis. ARB received no personal financial benefit from this programme. ES has served as an investigator in the EU-CORE database study (NOVARTIS).


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of Interest
  9. References
  10. Appendix A
  11. Appendix B
  • 1
    Berendt AR, Peters EJ, Bakker K, Embil JM, Eneroth M, Hinchliffe RJ, Jeffcoate WJ, Lipsky BA, Senneville E, Teh J, Valk GD. Diabetic foot osteomyelitis: a progress report on diagnosis and a systematic review of treatment. Diabetes Metab Res Rev 2008; 24(Suppl 1): S145S161.
  • 2
    National Institute for Health and Clinical Excellence. Diabetic foot—inpatient management of people with diabetic foot ulcers and infection. 2011; 2011(6/4/2011): .
  • 3
    Dutch Cochrane Centre. 2011; 2011(8/4/2011): .
  • 4
    Graham DR, Lucasti C, Malafaia O, Nichols RL, Holtom P, Perez NQ, McAdams A, Woods GL, Ceesay TP, Gesser R. Ertapenem once daily versus piperacillin-tazobactam 4 times per day for treatment of complicated skin and skin-structure infections in adults: results of a prospective, randomized, double-blind multicenter study. Clin Infect Dis 2002; 34(11): 14601468.
  • 5
    Saltoglu N, Dalkiran A, Tetiker T, Bayram H, Tasova Y, Dalay C, Sert M. Piperacillin/tazobactam versus imipenem/cilastatin for severe diabetic foot infections: a prospective, randomized clinical trial in a university hospital. Clin Microbiol Infect 2010; 16(8): 12521257.
  • 6
    Viswanathan V, Mahesh U, Jayaraman M, Shina K, Ramachandram A. Beneficial role of granulocyte colony stimulating factor in foot infection in diabetic patients. J Assoc Physicians India 2003; 51: 9091.
  • 7
    Graham DR, Talan DA, Nichols RL, Lucasti C, Corrado M, Morgan N, Fowler CL. Once-daily, high-dose levofloxacin versus ticarcillin-clavulanate alone or followed by amoxicillin-clavulanate for complicated skin and skin-structure infections: a randomized, open-label trial. Clin Infect Dis 2002; 35(4): 381389.
  • 8
    Lipsky BA, Holroyd KJ, Zasloff M. Topical versus systemic antimicrobial therapy for treating mildly infected diabetic foot ulcers: a randomized, controlled, double-blinded, multicenter trial of pexiganan cream. Clin Infect Dis 2008; 47(12): 15371545.
  • 9
    Senneville E, Lombart A, Beltrand E, Valette M, Legout L, Cazaubiel M, Yazdanpanah Y, Fontaine P. Outcome of diabetic foot osteomyelitis treated nonsurgically: a retrospective cohort study. Diabetes Care 2008; 31(4): 637642.
  • 10
    Tan JS, Friedman NM, Hazelton-Miller C, Flanagan JP, File TM Jr. Can aggressive treatment of diabetic foot infections reduce the need for above-ankle amputation? Clin Infect Dis 1996; 23(2): 286291.
  • 11
    Faglia E, Clerici G, Caminiti M, Quarantiello A, Gino M, Morabito A. The role of early surgical debridement and revascularization in patients with diabetes and deep foot space abscess: retrospective review of 106 patients with diabetes. J Foot Ankle Surg 2006; 45(4): 220226.
  • 12
    Tice AD, Turpin RS, Hoey CT, Lipsky BA, Wu J, Abramson MA. Comparative costs of ertapenem and piperacillin-tazobactam in the treatment of diabetic foot infections. Am J Health Syst Pharm 2007; 64(10): 10801086.
  • 13
    Lipsky BA, Armstrong DG, Citron DM, Tice AD, Morgenstern DE, Abramson MA. Ertapenem versus piperacillin/tazobactam for diabetic foot infections (SIDESTEP): prospective, randomised, controlled, double-blinded, multicentre trial. Lancet 2005; 366(9498): 16951703.
  • 14
    Clay PG, Graham MR, Lindsey CC, Lamp KC, Freeman C, Glaros A. Clinical efficacy, tolerability, and cost savings associated with the use of open-label metronidazole plus ceftriaxone once daily compared with ticarcillin/clavulanate every 6 hours as empiric treatment for diabetic lower-extremity infections in older males. Am J Geriatr Pharmacother 2004; 2(3): 181189.
  • 15
    Martínez-De Jesús FR, Ramos-De la Medina A, Remes-Troche JM, Armstrong DG, Wu SC, Lázaro-Martínez JL, Beneit-Montesinos JV. Efficacy and safety of neutral pH superoxidised solution in severe diabetic foot infections. Int Wound J 2007; 4(4): 353362.
  • 16
    Piaggesi A, Goretti C, Mazzurco S, Tascini C, Leonildi A, Rizzo L, Tedeschi A, Gemignani G, Menichetti F, Del PS. A randomized controlled trial to examine the efficacy and safety of a new super-oxidized solution for the management of wide postsurgical lesions of the diabetic foot. Int J Low Extrem Wounds 2010; 9(1): 1015.
  • 17
    Chen W, Xu K, Zhang H, Shang Y, Hao P. A comparative study on effect of bacterial load in diabetic foot ulcers dealing with iodophor and rivanol respectively. Zhongguo Xiu Fu Chong Jian Wai Ke Za Zhi 2008; 22(5): 567570.
  • 18
    Yönem A, Cakir B, Guler S, Azal OO, Corakci A. Effects of granulocyte-colony stimulating factor in the treatment of diabetic foot infection. Diabetes Obes Metab 2001; 3(5): 332337.
  • 19
    de Lalla F, Pellizzer G, Strazzabosco M, Martini Z, Du JG, Lora L, Fabris P, Benedetti P, Erle G. Randomized prospective controlled trial of recombinant granulocyte colony-stimulating factor as adjunctive therapy for limb-threatening diabetic foot infection. Antimicrob Agents Chemother 2001; 45(4): 10941098.
  • 20
    Kästenbauer T, Hornlein B, Sokol G, Irsigler K. Evaluation of granulocyte-colony stimulating factor (Filgrastim) in infected diabetic foot ulcers. Diabetologia 2003; 46(1): 2730.
  • 21
    Gough A, Clapperton M, Rolando N, Foster AV, Philpott-Howard J, Edmonds ME. Randomised placebo-controlled trial of granulocyte-colony stimulating factor in diabetic foot infection. Lancet 1997; 350(9081): 855859.
  • 22
    Cruciani M, Lipsky BA, Mengoli C, de Lalla F. Granulocyte-colony stimulating factors as adjunctive therapy for diabetic foot infections. Cochrane Database Syst Rev 2009; 8(3): CD006810.
  • 23
    Duarte HA, Fernández Montequín JI, Fors López MM, Carretero JH, Vilas MM, Mesa MG. Clinical evaluation of De Marco formula as an adjunctive therapy for infected ischemic diabetic foot: a prospective randomized controlled trial. Can J Clin Pharmacol 2009; 16(2): e381e391.
  • 24
    Abidia A, Laden G, Kuhan G, Johnson BF, Wilkinson AR, Renwick PM, Masson EA, McCollum PT. The role of hyperbaric oxygen therapy in ischaemic diabetic lower extremity ulcers: a double-blind randomised-controlled trial. Eur J Vasc Endovasc Surg 2003; 25(6): 513518.
  • 25
    Löndahl M, Katzman P, Nilsson A, Hammarlund C. Hyperbaric oxygen therapy facilitates healing of chronic foot ulcers in patients with diabetes. Diabetes Care 2010; 33(5): 9981003.
  • 26
    Doctor N, Pandya S, Supe A. Hyperbaric oxygen therapy in diabetic foot. J Postgrad Med 1992; 38(3): 112114.
  • 27
    Lobmann R, Ambrosch A, Seewald M, Dietlein M, Zink K, Kullmann KH, Lehnert H. Antibiotic therapy for diabetic foot infections: comparison of cephalosporines with chinolones. Diabetes Nutr Metab 2004; 17(3): 156162.
  • 28
    Bradsher RW Jr, Snow RM. Ceftriaxone treatment of skin and soft tissue infections in a once daily regimen. Am J Med 1984; 77(4): 6367.
  • 29
    Siami G, Christou N, Eiseman I, Tack KJ. Clinafloxacin versus piperacillin-tazobactam in treatment of patients with severe skin and soft tissue infections. Antimicrob Agents Chemother 2001; 45(2): 525531.
  • 30
    Harkless L, Boghossian J, Pollak R, Caputo W, Dana A, Gray S, Wu D. An open-label, randomized study comparing efficacy and safety of intravenous piperacillin/tazobactam and ampicillin/sulbactam for infected diabetic foot ulcers. Surg Infect (Larchmt ) 2005; 6(1): 2740.
  • 31
    Lipsky BA, Stoutenburgh U. Daptomycin for treating infected diabetic foot ulcers: evidence from a randomized, controlled trial comparing daptomycin with vancomycin or semi-synthetic penicillins for complicated skin and skin-structure infections. J Antimicrob Chemother 2005; 55(2): 240245.
  • 32
    Noel GJ, Bush K, Bagchi P, Ianus J, Strauss RS. A randomized, double-blind trial comparing ceftobiprole medocaril with vancomycin plus ceftazidime for the treatment of patients with complicated skin and skin-structure infections. Clin Infect Dis 2008; 46(5): 647655.
  • 33
    Vick-Fragoso R, Hernández-Oliva G, Cruz-Alcázar J, Amábile-Cuevas CF, Arvis P, Reimnitz P, Bogner JR, STIC Study Group. Efficacy and safety of sequential intravenous/oral moxifloxacin vs intravenous/oral amoxicillin/clavulanate for complicated skin and skin structure infections. Infection 2009; 37(5): 407417.
  • 34
    Lipsky BA, Pecoraro RE, Larson SA, Hanley ME, Ahroni JH. Outpatient management of uncomplicated lower-extremity infections in diabetic patients. Arch Intern Med 1990; 150(4): 790797.
  • 35
    Erstad BL, McIntyre J. Prospective, randomized comparison of ampicillin/sulbactam and cefoxitin for diabetic foot infections. Vasc Surg 1997; 31(4): 419426.
  • 36
    Lipsky BA, Baker PD, Landon GC, Fernau R. Antibiotic therapy for diabetic foot infections: comparison of two parenteral-to-oral regimens. Clin Infect Dis 1997; 24(4): 643648.
  • 37
    Lipsky BA, Itani K, Norden C. Treating foot infections in diabetic patients: a randomized, multicenter, open-label trial of linezolid versus ampicillin-sulbactam/amoxicillin-clavulanate. Clin Infect Dis 2004; 38(1): 1724.
  • 38
    Lipsky BA, Giordano P, Choudhri S, Song J. Treating diabetic foot infections with sequential intravenous to oral moxifloxacin compared with piperacillin-tazobactam/amoxicillin-clavulanate. J Antimicrob Chemother 2007; 60(2): 370376.
  • 39
    Grayson ML, Gibbons GW, Habershaw GM, Freeman DV, Pomposelli FB, Rosenblum BI, Levin E, Karchmer AW. Use of ampicillin/sulbactam versus imipenem/cilastatin in the treatment of limb-threatening foot infections in diabetic patients. Clin Infect Dis 1994; 18(5): 683693.
  • 40
    Bansal E, Garg A, Bhatia S, Attri AK, Chander J. Spectrum of microbial flora in diabetic foot ulcers. Indian J Pathol Microbiol 2008; 51(2): 204208.
  • 41
    El-Tahawy AT. Bacteriology of diabetic foot. Saudi Med J 2000; 21(4): 344347.
  • 42
    Viswanathan V, Jasmine JJ, Snehalatha C, Ramachandran A. Prevalence of pathogens in diabetic foot infection in South Indian type 2 diabetic patients. J Assoc Physicians India 2002; 50: 10131016.
  • 43
    Lipsky BA, Berendt AR, Deery HG, Embil JM, Joseph WS, Karchmer AW, LeFrock JL, Lew DP, Mader JT, Norden C, Tan JS. Diagnosis and treatment of diabetic foot infections. Clin Infect Dis 2004; 39(7): 885910.
  • 44
    International Working Group on the Diabetic Foot. International Consensus on the Diabetic Foot and Supplements. 2007; DVD.

Appendix A

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of Interest
  9. References
  10. Appendix A
  11. Appendix B

Search strings for each of the sections


June 1960 to August 2010

The basic search was combined with searches for specific interventions of interest by adding the search term.

((Diabetes Mellitus OR diabetic))


(((Clinical Trials) OR (comparative study) OR (epidemiologic study characteristics) OR (Clinical Trial*) OR (case-control stud*) OR (case control stud*) OR (cohort stud*) OR (Comparative stud*)))


((Infection OR infected OR cellulitis OR abscess OR necrotizing fasciitis OR osteomyelitis OR gangrene OR erysipelas OR osteitis OR (Bone Diseases, Infectious) OR (Diabetic Foot)) AND (Surgery OR Amputation OR (Surgery, Plastic) OR (Preoperative Care) OR (dead space) OR drain OR hardware OR (bone samples) OR biopsy OR (Vascular Surgical Procedures) OR (Thrombolytic Therapy) OR (Costs and Cost Analysis) OR (Wound Healing) OR (Anti-Bacterial Agents) OR (Anti-Infective Agents) OR (administration and dosage) OR (Drug Administration Routes) OR parenteral OR oral OR topical OR duration OR cement OR (Methylmethacrylate) OR (Calcium Sulfate) OR implant OR collagen OR ceramic OR (Aminoglycosides OR gentamicin OR amikacin OR tobramycin) OR (Glycopeptides OR vancomycin OR Oritavancin OR dalbavancin) OR teicoplanin OR Metronidazole OR Linezolid OR (Fusidic Acid) OR Daptomycin OR Monobactam OR (Carbapenem OR imipenem OR meropenem) OR (beta-Lactams) OR (Cephalosporins) OR cefuroxime OR ceftazidime OR cephalexin OR ceftriaxone OR cefpirome OR (Clavulanic Acids) OR (Clavulanic Acid∗) OR (Moxalactam) OR (Penicillins) OR penicillin OR flucloxacillin OR oxacillin OR Methicillin OR nafcillin OR ampicillin OR penicillin OR piperacillin OR (Tetracyclines) OR tetracycline OR minocycline OR doxycycline OR (Macrolides) OR erythromycin OR azithromycin OR clarithromycin OR (Lincomycin) OR clindamycin OR (Trimethoprim-Sulfamethoxazole Combination) OR cotrimoxazole OR co-trimoxazole OR (Quinolones) OR ciprofloxacin OR ofloxacin OR moxifloxacin OR levofloxacin OR (Anti-Infective Agents, Local) OR (Silver OR Silver Sulfadiazine OR iodine) OR honey OR larvae OR maggots OR larval OR (hyperbaric oxygen therapy OR hyperbaric OR (vacuum assisted wound therapy) OR (VAC therapy) OR (negative pressure therapy) OR (growth factors) OR (G-CSF) OR (granulocyte colony stimulating growth factor)))


June 1960 to August 2010

The basic search was combined with searches for specific interventions of interest by adding the search term.

Map to preferred terminology (with spell check).

Also search as free text.

Include sub-terms/derivatives (explosion search).

(Diabetes Mellitus) OR diabetic


(Clinical Trials) OR (comparative study) OR (epidemiologic study characteristics) OR (Clinical Trial*) OR (case-control stud*) OR (case control stud*) OR (cohort stud*) OR (Comparative stud*) OR (case control study) OR (Comparative study) OR (RCT ) OR (Randomised controlled trial) OR (Costs and Cost Analysis)


Infection OR infected OR cellulitis OR abscess OR (necrotizing fasciitis) OR osteomyelitis OR gangrene OR erysipelas OR osteitis OR (Bone Diseases, Infectious) OR (Diabetic Foot)


(Wound Healing) OR (Anti-Bacterial Agents) OR (Anti-Infective Agents) OR (administration and dosage) OR (Drug Administration Routes) OR parenteral OR oral OR topical OR duration OR cement OR Methylmethacrylate OR (Calcium Sulfate) OR implant OR collagen OR ceramic OR Aminoglycosides OR gentamicin OR amikacin OR tobramycin OR Glycopeptides OR vancomycin OR Oritavancin OR dalbavancin OR teicoplanin OR Metronidazole OR Linezolid OR (Fusidic Acid) OR Daptomycin OR Monobactam OR Carbapenem OR imipenem OR meropenem OR (beta-Lactams) OR Cephalosporins OR cefuroxime OR ceftazidime OR cephalexin OR ceftriaxone OR cefpirome OR (Clavulanic Acids) OR (Clavulanic Acid∗) OR Moxalactam OR Penicillins OR penicillin OR flucloxacillin OR oxacillin OR Methicillin OR nafcillin OR ampicillin OR penicillin OR piperacillin OR Tetracyclines OR tetracycline OR minocycline OR doxycycline OR Macrolides OR erythromycin OR azithromycin OR clarithromycin OR Lincomycin OR clindamycin OR (Trimethoprim-Sulfamethoxazole Combination) OR cotrimoxazole OR (co-trimoxazole) OR Quinolones OR ciprofloxacin OR ofloxacin OR moxifloxacin OR levofloxacin OR (Anti-Infective Agents, Local) OR Silver OR (Silver Sulfadiazine) OR iodine OR honey OR larvae OR maggots OR larval OR (hyperbaric oxygen therapy) OR hyperbaric OR (vacuum assisted wound therapy) OR (VAC therapy) OR (negative pressure therapy) OR (growth factors) OR (G-CSF) OR (granulocyte colony stimulating growth factor)

Appendix B

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of Interest
  9. References
  10. Appendix A
  11. Appendix B

Evidence tables

ReferenceStudy design and scorePopulationIntervention and control managementOutcomesDifferences and statistical resultsLevel of evidence (SIGN)CommentsOpinion
  1. ABPI, ankle : brachial pressure index; A/C, amoxicillin/clavulanate; A/S, ampicillin/sulbactam; BKA, below knee amputation; DFI, diabetic foot infection; DMF, De Marco formula; EMA, European Medicines Agency; EOT, end of therapy; FDA, US Food and Drug Administration; G-CSF, granulocyte-colony stimulating factor; HBOT, hyperbaric oxygen therapy; I/c, imipenem/cilastatin; i.v., intravenously; MRSA, methicillin-resistant Staphylococcus aureus; NS, not significant; P/T, piperacillin/tazobactam; RCT, randomized controlled trial; TOC, test-of-cure; WCC, white cell count.

Early surgery
Tan et al. [10]Cohort

Single centre

Study quality, 3/8

Cohort study of 112 patients with 164 DFIs, hospitalized for treatment of the foot infection. Of these, 76 had a deep infection, 65 had osteomyelitis. No early surgery and antibiotics in 87 subjects versus early surgery and antibiotics in 77 subjects87 infections treated without surgery in the first 3 days versus 77 treated with antibiotics + surgery (of which 46 antibiotics and debridement and 31 antibiotic and early local amputation). Duration of treatment with antibiotics unknownInfection outcome: above ankle amputationAmputation rate 27.6% vs 13.0% antibiotic group and antibiotic and surgical intervention groups, respectively (p < 0.01)2−No information regarding (appropriateness of) antibiotic treatment. High risk of bias as there is no assessment of severity and there is a high chance of indication biasNo sponsor identified. Provides no evidence to confirm the role of surgery, as opposed to timing of intervention
Faglia et al. [11]Cohort

Single centre

Study quality, 5/8

Diabetes and deep foot space abscess

N = 106, group 1: 43 subjects, group 2: 63 subjects

Group 1: immediate surgical debridement; group 2: referred from another hospital after a mean delay of 6.2 ± 7.5 days without debridement

Duration of treatment with antibiotics unknown

Drainage without amputation

One or more ray amputations

Transmetatarsal amputation


Major amputation

Group 1: 9 vs Group 2: 4

Group 1: 21 vs Group 2: 21

Group 1: 12 vs Group 2: 10

Group 1: 1 vs Group 2: 23

Group 1: 0 vs Group 2: 5

(p < 0.001)

Χ2 24.4

2−Poor quality study despite the 5/8 score

Concluded that delay in drainage increases the incidence of amputation, but this is not justified by these data because of the possibility of bias

No sponsor identified
Health economics
Tice et al. [12]RCT

Subset analysis based on multicentre, double-blinded study

Study quality, 6/9

99 patients with DFI, including osteomyelitis provided all infected bone was surgically removed; 56 subjects in ertapenem group vs 43 in P/T groupSubstudy of SIDESTEP [13], cost-minimization assessment of ertapenem versus P/T. No duration of treatment givenInfection outcomes:

Mean days of treatment

Total i.v. drug doses

Total antibiotic dosages

Mean drug preparation and administration cost

7.6 vs 7.4 (p = 0.8) days of treatment, 7.5 vs 25.5 (p < 0.0001) total i.v. doses

8.6 vs 26.8 (p < 0.0001) total i.v. and oral doses, $356 vs $503 (p < 0.001) total cost of treatment for ertapenem and pip/tazo, respectively

1+High dropout rate. Length of stay was a proxy measure. The length of stay might have been prolonged because of the trial designSponsored by Merck
Topical treatment with antiseptic agents
Martínez-De Jesús et al. [15]RCT

Single centre

Patient blinded

Study quality, 4/9

Type 2 diabetes and infected, deep diabetic foot ulcers, 21 subjects in intervention group vs 16 in control groupn = 21 intervention group: neutral pH superoxidized aqueous solution

n = 16 control group: disinfectant such as soap or povidone iodine. Duration of antibiotic treatment more than 10 days

Odour, periwound cellulitis and granulation tissueOdour reduction was achieved in all superoxide patients (100% vs 25%; p < 0.01) and surrounding cellulitis diminished (p < 0.001) in 17 patients (80.9% vs 43.7%)1−Alternate patient group allocation, yet different numbers in each group. Non-standardized wound classification criteriaNo sponsor identified
Chen et al. [17]RCT

Single centre

Patient blinded

Study quality, 6/9

30 patients with diabetic foot ulcers, 10 patients in each subgroup10 diabetic foot ulcers treated with iodophor, 10 with rivanol, 10 controls

One single application of topical treatment after ulcer debridement

Infection outcomes:

bacteria number in wound

Number of colonies after 24 h/number of colonies at t = 0 was 0.961, 0.918 and 0.986 for the control group, iodophor group and rivanolol, respectively. Significantly less growth of bacteria after 24 h in the iodophor group compared with the rivanol and the control group1+Use of systemic antibiotics not mentioned. Study only looked at bacterial growth after 5 min and 24 hNo sponsor identified
Piaggesi et al. [16]RCT

Single centre

Open label

Study quality, 6/9

40 patients with diabetes with post-surgical wounds, who had surgery for a DFI, 20 subjects in each treatment groupDermacyn versus povidone iodine. All patients had systemic antibiotic therapy and surgical debridement if needed. Ischaemia was an exclusion criterion

Duration of treatment with P/T and metronidazol with or without teicoplanin 10.1 ± 6.1 weeks for Dermacyn and 15.8 ± 7.8 weeks for control group (p = 0.016). Duration of antibiotic use was an outcome measure

Infection outcomes: use of antibiotics

Non-infection outcomes: Healing rate at 6 months and healing time

Duration of antibiotic use: 10.1 ± 6.1 weeks Dermacyn group vs 15.8 ± 7.8 weeks in povidone iodine group (p = 0.016). Healing rate at 6 months 90% in Dermacyn vs 55% in iodine group (X2 9.9, p = 0.002). Healing time 10.5 ± 5.9 vs 16.5 ± 7.1, respectively (p = 0.007)1+2 patients lost to follow-up

Details of the interventions and outcomes were suboptimal

Possible adverse effect of iodine on wound healing not taken into account

Very long antibiotic treatment period

Sponsored by Oculus Innovative Sciences
Gough et al. [21]RCT

Single centre

Double blind

Study quality, 9/9

40 patients with diabetes with moderate (International Consensus Guidelines Grade 3) infection of DFU, n = 20 subjects in both treatment armsIntervention: G-CSF 5 µg/kg adjusted on basis of WCC, for 7 days versus saline control

Both groups received 4 antibiotics, mean duration of i.v. antibiotics 8.5 for G-CSF and 14.5 for controls (p = 0.02)

Infection outcome measures:

(1) Time to resolution of infection

(2) Total time of intravenous antibiotics

(3) Hospital length of stay

(4) Need for surgery

(5) Time taken to eliminate pathogens from wound

Non-infection outcome:

(6) Effect of G-CSF on generation of neutrophil superoxide

Intervention: 7 (5–20) days vs control: 12 (5–93)

(p = 0.03)

Intervention: 8.5 (5–30) vs control: 14.5 (8–63) days (p = 0.02)

Intervention: 10.0 (7–31) days vs 17.5 (9–100)

(p = 0.02)

Intervention: 0 vs 4/20 (20%)

(p = 0.114)

Intervention: 4 (2–10) days vs control: 8 (2–75) days (p = 0.02)

Intervention: 16.1 (4.2–24.2) nmol per 106 neutrophils/30 min vs 7.3 (2.1–11.5) (p < 0.0001)

1++Well-designed RCT showing significant benefit in moderate infection

See meta-analysis [22]which concluded that G-CSF did not have a significant benefit with regard to either resolution of infection or healing of wounds, although there was a significant reduction in the need for lower extremity surgery

Sponsored by Amgen
Yönem et al. [18]RCT

Single centre

Blinding unknown

Study quality, 2/9

15 subjects with cellulitis or Wagner 2 or less in each of the two treatment arms15 patients treated with standard treatment (antibiotics and wound care), 15 patients treated with standard treatment + G-CSF 5 µg/kg, duration of antibiotic treatment 22.9 ± 2.0 days in G-CSF 23.3 ± 1.9 days in control group, standard treatment with G-CSF 3 days

Duration of antibiotic treatment 22.9 ± 2.0 days in G-CSF 22.3 ± 1.9 days in control group

Infection outcomes: Time to resolution of infection, duration of hospitalization, duration of parenteral antibiotic administration, need for surgical interventionNo significant differences in time to resolution of infection, duration of hospitalization duration of parenteral antibiotic administration, amputation in the G-CSF treated group compared with the standard group1−Also includes results of respiratory burst, granulocyte count etc. Typing error in abstract (p < 0.05 should be p > 0.05)No sponsor identified
De Lalla et al. [19]RCT

Single centre


Study quality, 6/9

Severe limb-threatening foot infection, all with osteomyelitis in diabetes N = 40

Patients with ABPI < 0.5 or ankle systolic pressure <50 mmHg, and patients with serum creatinine >1.6 mg/100 mL were excluded

20 subjects in each treatment group

Intervention: conventional treatment plus G-CSF 263 µg s.c. daily for 21 days versus conventional treatment (no placebo). Mean duration of antibiotics 68.9 ± 29.2 days for G-CSF patients and 58.7 ± 23.7 days for controls (not significant)(1) Cure (complete closure of the ulcer without signs of bone infection)

(2) Improvement (eradication of pathogens in addition to marked or complete reduction of cellulitis but incomplete ulcer healing, or ulcer healing but persistent osteomyelitis)

(3) Failure (absence of any clinical improvement) or amputation for persistent infection

At 3 weeks: intervention 0 vs controls 0

At 9 weeks: intervention 7 vs controls 7 (p = 1.0)

At 3 weeks intervention 12 vs controls 9 (p = 0.34)

At 9 weeks: intervention 8 vs controls 4 (p = 0.17)

At 3 weeks: intervention 8 vs controls 11 (p = 0.34)

At 9 weeks: intervention 5 vs controls 9 (p = 0.19)

1+No effect of G-CSF on eradication of infection, in contrast to [21]. Differences with [21] study relating to prevalence of osteomyelitis and choice of outcome measuresNo sponsor identified
Kästenbauer et al. [20]RCT

Single centre

Patient blinded

Study quality, 7/9

Soft-tissue infection of DFU, 20 subjects in intervention group, 17 in placebo groupThe patients in the intervention group received daily an initial dose of either 5 µg/kg G-CSF or placebo (0.9% sterile saline solution), s.c. Subjects were treated with i.v. antibiotics (clindamycin and ciprofloxacin) until the inflammation had visibly improved. Oral antibiotics were administered thereafter if necessary. Duration of treatment with G-CSF 10 days. Mean antibiotic treatment duration 5.6 ± 2.5 days in G-CSF group, 5.8 ± 2.3 days in placebo groupInfection outcomes: infection scores pre-treatment versus post-treatment, putrid, erythema, oedemaPatients who received G-CSF did not have an earlier resolution of clinically defined infection than placebo patients1+Infection score non-validatedSponsored by Amgen

No differences, like [18], whereas significant differences in resolution of signs of infection contrast with [21]

Viswanathan et al. [6]RCT

Single centre

Double blind

Study quality, Not scored

N = 20, with extensive cellulitis, Wagner II–III ulcers, 10 subjects in each treatment armThe patients received daily initial dose of either 5 µg/kg body weight G-CSF or placebo (0.9% sterile saline solution), injected subcutaneously

Duration of treatment with G-CSF 10 days

Eradication of infection


Hospital length of stay

Intervention: 9

Control: 3 (NS)

Intervention: 0

Control: 3 (NS)

Intervention: 7.4

Control: 8.8 (p = 0.02)

 ‘Foot ulcers excluded’Sponsored by Amgen
Procaine plus polyvinylpyrrolidone
Duarte et al. [23]RCT,

Single centre

Assessor blinded

Study quality, 7/9

118 patients with ischaemic DFI, of which ischaemic gangrene n = 63, ischaemic ulcer n = 55. 59 subjects in each treatment arm59 patients treated with De Marco formula 0.15 mL/day intramuscular injection (DMF = combination of procaine HCl and polyvinylpyrrolidone), for 10 days, then twice weekly until wound healing or completion of a 6-week period. 59 patients treated with standard careInfection outcomes: AmputationAmputation rate 45.8% vs 25.4% (toes 30.4% vs 28.8%, transmetatarsal amputation 18.6% vs 8.5%) in the control group and DMF group, respectively (NS)1+Unknown risk of bias, unclear criteria for reason of amputation or level of amputation

Little obvious evidence of benefit

Sponsored by Gen Cell
Hyperbaric oxygen therapy
Doctor et al. [26]RCT

Single centre

Open label

Study quality, 3/9

30 patients with chronic diabetic foot ulcers, 15 subjects in both intervention and control groupAll patients treated with systemic antibiotic therapy and wound debridement, 15 patients treated with HBOT, 15 treated conservatively

4 HBOT sessions of 45 min over 2 weeks. Antibiotic duration 3 days

Infection outcome: positive wound cultures

Non-infection outcomes: hospital days, amputation and level

Hospital stay 41 vs 47 days, major amputation 2 vs 7, minor amputations 4 vs 2, pre-procedure positive wound culture 19 vs 16, post-procedure positive wound culture 3 vs 12, in the HBOT group versus the control group, respectively. All differences were not significant1−It was not described how many diabetic foot ulcers were infected, but most received antibiotics

Method of randomization was not described

Antibiotics used were based on sensitivity spectrum and included cephalosporins, aminoglycosides and metronidazole

No sponsor identified

No evidence of benefit

Comparison of antibiotic regimens – skin and soft-tissue infection alone
Bradsher and Snow [28]RCT


Open label

Study quality, 4/9

Subset of RCT in 84 patients with soft-tissue infection. Of these 84, 20 patients had DFI, 10 subjects in each treatment arm10 patients with DFI treated with ceftriaxone, 10 with cefazolin

Duration of antibiotic treatment unknown

Infection outcomes: only microbiological evaluation, elimination, reduction, persistence, relapse, reinfectionElimination of infection: 6 vs 4, reduction: 3 vs 2, persistence: 1 vs 4 in ceftriaxone and cefazolin, respectively1−Insufficient data available for the DFI subgroup

No clinical assessment in DFI subgroup

No sponsor identified
Lipsky et al. [34]RCT

Single centre

Open label

Study quality, 4/9

Outpatient infected diabetic foot ulcers N = 56, 27 vs 29 subjects in each treatment armOral clindamycin hydrochloride (n = 27) or cephalexin (n = 29)

Duration of therapy 2 weeks. Additionally, 3 patients received an additional 2 weeks of antibiotic treatment after their initial course

Infection outcomes: Eradication of bacteria by wound culture, clinical cureNo difference in eradication, clinical response or wound healing response between the two antibiotics. Fifty-one infections (91%) were eradicated, 42 (75%) after 2 weeks of treatment; only 5 (9%) were initially treatment failures, and 3 (5%) were subsequently cured with further outpatient oral antibiotic treatment. After a mean follow-up of 15 months, no further treatment was required in 43 (84%) of the cured patients1−No ITT analysis

No data on blinding of patient/clinician/assessor

Sponsored by the Department of Veterans Affairs and Upjohn Company

Only study on clindamycin monotherapy

First treatment trial of out-patients

Siami et al. [29]RCT Multicentre

Investigator blinded

Study quality, 5/9

409 patients with skin and soft-tissue infection, of which 279 patients clinically evaluable, 54 patients with DFI, 25 subjects in each of the two treatment arms

Patients with osteomyelitis were excluded

29 patients treated with clinafloxacin i.v., 25 with P/T i.v. with or without vancomycin (in case of MRSA). Duration of treatment for whole group (including group with DFIs): at least 3 days of i.v. therapy followed by oral therapy for a maximum total duration of 14 days. Median duration of treatment of patients that completed treatment was 13 days (total patients)Infection outcomes:

Clinical cure

Microbiological eradication

15/29 clinically cured in clinafloxacin group vs 12/25 in pip/tazo group

Microbiological eradication: 32/73 and 15/47 isolates eradicated for clinafloxacin and P/T groups, respectively

1+Approximately one-third of patients not clinically or microbiologically evaluableSponsored by Parke-Davis

Short duration of treatment, also relatively low rate of clinical cure

Clay et al. [14]RCT

Single centre

Open label. (Veterans Admin.)

Study quality, 3/9

DFU N = 70, only men with diabetes and lower extremity infection were included, 36 vs 34 subjects in each treatment armGroup 1 n = 36

Metronidazole 1 g and 1 g ceftriaxone i.v. each day for a mean of 6.7 ± 3.3 days in patients with successful outcome

15 protocol violations

Group 2 n = 34

Ticarcillin/clavulanate 3.1 g i.v. each 6 h for a mean of 6.1 ± 4.3 days in patients with successful outcome

12 protocol violations



Finger stick blood glucose

Improvement of wound stage

Creatinine clearance


No statistically significant differences (NS)



Stage changed ‘minimally’ – details not shown


Cost saving of $61 per hospital admission in metronidazole/ceftriaxone group

1−Men only

27 patients had antibiotics changed and no indication of whether the analysis was strictly per protocol

No stated time of day for blood glucose measurement, and only undertaken in 39/70

Creatinine clearance assessed in only 31/70

‘Treatment success’ achieved in 29 patients in Cef/Met and in 29 in the Tic/clav group

(p = NS)

Inappropriate measures of treatment success besides clinical staging (which is not standardized)

Treatment duration is only assessed in those who achieved treatment success. This causes some doubt on the cost analysis

No conclusions can be drawn from the study

Sponsored by Roche

Pharmacist-led study

Lobmann et al. [27]Cohort. Prospective

Study quality, 4/8

180 diabetic patients with severe limb-threatening foot infection were consecutively enrolled. 300 patients were screened, 90 vs 90 subjects in each treatment arm90 ceftriaxone vs 90 quinolones in addition to standard treatment of foot infection. Mean duration of treatment in ceftriaxone group 18.7 days, in quinolone group 23.8 days. Median duration of treatment in ceftriaxone group 11.5 days, and 16.5 days in the quinolone groupNon-infection outcomes

Wound healing, amputation rate, length of stay

Infectious: clinical (reaching Wagner I or 0) and microbiological cure rate of infection, duration of antibiotic therapy, need to change antibiotic therapy

Treatment with a third generation cephalosporin is as effective as a treatment with quinolones. Clinical response was achieved in 58.0% in the ceftriaxone group and in 51.1% in the quinolone group (NS). Fourteen days after initiation of treatment, the number of patients with microbiological isolates decreased in both groups (52 to 5 in the ceftriaxone group and 60 to 12 in the quinolone group). At hospital discharge, 66.0% of ceftriaxone and 64.4 of quinolone-treated diabetic ulcers were cured or improved. Median duration of antibiotic therapy 11.5 days for ceftriaxone vs 16.5 days for quinolone (p < 0.01). Need to change antibiotic therapy 7.8% ceftriaxone vs 16.7% for quinolones2−Clindamycin could be added in both groups (added in 27%)

Not clear how many patients in each group received clindamycin. Definition of clinical response is unusual (change in Wagner grades)

Sponsored by Hoffmann La Roche
Harkless et al. [30]RCT Multicentre

Open label

Study quality, 2/9

314 Patients with polymicrobial infections involving MRSA also received vancomycin 1 g q12h, n = 155 vs n = 159 subjects in each of two treatment arms155 adult patients with moderate-to-severe infected diabetic foot ulcers received P/T (4 g/0.5 g q8h), and 159 received A/S (2 g/1 g q6h) as a parenteral treatment

Median duration of treatment 8.0 days P/T vs 8.5 days A/S

Clinical success (resolution of ulcer and of symptoms of infection, no additional antibiotics needed)

Bacteriological success (end of cure or end of treatment eradication or presumed eradication)

Clinical efficacy rates (cure or improvement) were statistically equivalent overall (81% for P/T vs 83% for A/S), and median duration of treatment was similar in the clinically evaluable populations (9 days for P/T, 10 days for A/S). Drug-related adverse events for both study drugs were comparable in frequency and type1−Very large number of dropouts (38 and 44% non-evaluable)Sponsored by Wyeth
Lipsky and Stoutenburgh [31]RCT

Subset analysis of multicentre trial

Investigator blinded

Study quality, 4/9

N = 133, all subjects had a DFU with infection, 47 vs 56 subjects in each of two treatment armsPatients with a diabetic ulcer infection were prospectively stratified to ensure they were equally represented in the treatment groups, then randomized to either daptomycin (n = 47) [4 mg/kg every 24 h i.v.] or a pre-selected comparator (n = 56) (vancomycin or a semi-synthetic penicillin) for 7–14 days

Exact duration of treatment not given

Infection outcomes: Success rates microbiological adverse eventsOf 133 subjects, 103 were clinically evaluable. Most infections were monomicrobial, and Staphylococcus aureus was the predominant pathogen. Success rates for patients treated with daptomycin or the comparators were not statistically different for clinical (66% vs 70%, respectively; 95% CI, −14.4, 21.8) or microbiological (overall or by pathogen) outcomes. Both treatments were generally well tolerated, with most adverse events of mild to moderate severity1−Infection presumptively caused by Gram-positive organisms. 30 of 133 subjects were not clinically evaluable. 8 patients had MRSA infections: 1 in daptomycin group, 7 in vancomycin group

Note: No ITT analysis

Sponsored by Cubist
Lipsky et al. [8]2 RCTs consecutive, multicentre

Double blind

Study quality, 8/9

Mildly infected diabetic foot ulcers. N = 835 subjects, of whom 418 in the intervention group, and 417 in the control group2 studies: 303 and 304: 418 subjects received the active topical agent pexiganan plus an oral placebo versus 417 subjects who received oral ofloxacin plus a topical placebo

Mean duration 23 days in study 303 and 25 days in study 304. Median duration 27 days in study 303 and 22 days in study 304

Infection outcomes: clinical cure or improvement of the infection, eradication of wound pathogens, bacterial resistance, adverse events

Non-infection outcomes: wound healing

Although study 303 failed to demonstrate equivalence, study 304 and the combined data for the 2 trials demonstrated equivalent results (within the 95% confidence interval) for topical pexiganan and oral ofloxacin in clinical improvement rates (85–90%), overall microbiological eradication rates (42–47%) and wound healing rates. The incidence of worsening cellulitis (2–4%) and amputation (2–3%) did not differ significantly between treatment arms. Bacterial resistance to ofloxacin emerged in some patients who received ofloxacin, but no significant resistance to pexiganan emerged among patients who received pexiganan. More adverse effects in the ofloxacin group1++Mild infection not adequately defined. Development of resistance in the oral antibiotic group

Only study of oral versus topical treatment

Low incidence of pexiganin resistance

Sponsored by Magainin and SKB
Noel et al. [32]RCT


Double blind

Study quality, 6/9

Subgroup analysis of 257 people with DFI in a larger study of skin and skin-structure infections: total N = 828 (31% patients with DFI)

169 vs 89 subjects in each of the two treatment arms. Group allocation 2:1, only 222/257 were clinically evaluable

Group 1: 169 subjects: Ceftobiprole 500 mg i.v. 8 hourly

Group 2: 89 subjects: Vancomycin 1 g each 12 h and ceftazidime 1 g each 8 h

Both for 7–14 days. Mean duration of total population 9.0 days for ceftobiprole and 9.1 days for control group (per protocol analysis of N = 828)

Clinical outcomes assessed at TOC visit (7–14 days after EOT) and defined as: cure, failure and not evaluable

Microbiological outcomes assessed at TOC visit (7–14 days after end of treatment):

Eradication, presumed eradication, persistence, resumed persistence, colonization, superinfection, not evaluable

Clinically cured:

Group 1 86.2%

Group 2 81.8% (CI of comparison −5.4 to 15.7)

No further details given for the DFI group

1+No baseline details of the DFI patients

Only outcome measure available for the DFI patients is the proportion of clinically cured in the clinically evaluable patients at TOC visit

Efficacy of the two groups seemingly equivalent. Drug is currently not FDA or EMA approved. Centres targeted had high prevalence of MRSA

Sponsored by Johnson & Johnson
Vick-Fragoso et al. [33]RCT


Randomized open label

Study quality, 4/9

Large multinational study of skin and soft-tissue infections, N = 804. Subset with DFI n = 134. Group 1, 63 subjects vs group 2, 71 subjects. Total withdrawal 22/134Group 1: sequential i.v./oral moxifloxacin 400 mg/day

Group 2: sequential i.v./oral A/C

1000/200 mg three times daily

Mean duration of antibiotic treatment 14.1 ± 5.5 days for moxifloxacin and 15.2 ± 5.4 days for A/C (i.v. and oral combined)

Clinical success rate (success = total resolution or marked improvement of all symptoms and signs; no additional or alternative antimicrobial treatment)Group 1: 25/49 (51.0%)

Group 2: 42/63 (66.7%)

(5%CI for difference −34 to 2.7, formal statistical significance not calculated)

1−No apparent difference in the subset with DFI

No difference observed in the total population

Sponsored by Bayer
Graham et al. [4], ertapenem versus P/TRCT


Double blind

Study quality, 5/9

540 adults with complicated skin and skin-structure infections. Of these, 98 had a lower extremity infection with diabetes, of which the data of 66 patients were evaluable

Subjects with osteomyelitis were excluded

53 subjects received 1 g daily ertapenem with TID placebo infusions, compared with 45 subjects who received 3.375 g QID P/T

Mean duration of therapy was 9.1 ± 3.1 days for ertapenem and 9.8 ± 3.3 days for P/T

Clinical cureClinical cure in evaluable patients (modified intention to treat analysis): Ertapenem group: 23/35 (66%) cure

P/T group: 22/31 (71%) cure (no significant difference)

1−Very limited demographical and baseline data on the subjects of the subgroup with diabetes

Only data of 66 of 98 subjects were available for review and analysis

More outcome measures are available for the total studied group but not for the subgroup of patients with diabetes-related lower extremity infection

Sponsored by Merck
Graham et al. [7], levofloxacin versus ticarcillinRCT

Multicentre Open label

Study quality, 1/9

399 adults with complicated skin and skin-structure infections. Of these, 66 had an infected diabetic foot ulcer

Subjects with osteomyelitis or who needed emergency surgery were excluded

Patients were randomized to 1 of the 2 study arms:

Ticarcillin/clavulanate (3.1 g given i.v. every 4–6 h) with a switch to oral A/C (875 mg BID) at the investigator's discretion, or levofloxacin (750 mg given by mouth and/or i.v. QD). Subjects in both groups received 7–14 days of therapy. The randomization schedule was stratified by study centre and by diagnosis of diabetic ulcer

Mean duration of therapy was 12.1 ± 4.9 days in the levofloxacin group and 12.1 ± 4.9 days in the ticarcillin/clavulanate group

Clinical cureClinical cure in evaluable patients in ticarcillin/clavulanate group 18/26 (69%) vs 16/28 (57%) in the levofloxacin group (no significant difference)

Seven patients taking levofloxacin and 2 taking TC/AC had osteomyelitis diagnosed after admission to the study, resulting in 4 amputations

Five of 9 of the osteomyelitis cases were due to diabetic ulcers

1−Very limited demographical and baseline data on the subjects of the subgroup with diabetes

Only data of 54 of 66 subjects were available for review and analysis

More outcome measures are available for the total studied group but not for the subgroup of patients with diabetes-related lower extremity infection

Not reported to which group the subjects with osteomyelitis were randomized

Sponsored by Johnson & Johnson Research and Development
Comparison of antibiotic regimens – studies including patients with osteomyelitis
Grayson et al. [39]RCT

Single centre

Double blind

Study quality, 9/9

Limb-threatening infection of the foot in 93 hospitalized subjects with 96 episodes of DFI, some despite previous antibiotic therapy. Prevalence of osteomyelitis 68% vs 56% episodes in the A/S and I/c groups, respectively. Group 1: 48 episodes in 47 participants, group 2: 48 episodes in 46 participants

One person was randomized in error

Group 1: A/S 2 g/1 g (A/S) i.v. 6-hourly

Group 2: I/c 500 mg i.v. every 6 h

Doses adjusted to renal function

Mean duration of treatment in A/S group 13 ± 6.5 days vs 15 ± 8.6 days in the I/c group, follow-up period 1 year

Eradication of infection at 5 days

Eradication of infection at EOT

Microbiological eradication

Failure at EOT

Adverse reactions

Group 1 28/48 vs Group 2 29/48

Group 1 39/48 vs Group 2 41/48 (p = 0.78)

Group 1 32/48 vs group 2 36/48 (p = 0.5)

Group 1 8/48 vs Group 2 6/48

Group 1 16 vs Group 2 17

1++High quality RCT

No difference between two intravenous regimens in terms of resolution of signs of STI and of systemic signs

There was a very high incidence of amputation 69 vs 58% for A/S and I/c, respectively

Osteomyelitis – cannot be assessed in this way although did have follow-up for 1 year in this study, but they were also treated surgically

Sponsored by Pfizer

High prevalence of osteomyelitis

Osteomyelitis also treated with resection of bone

1 year follow-up

No difference between narrow (G + ve targeted) versus broad-spectrum antibiotics

Research question must be to isolate osteomyelitis. But if included, need 1 year follow-up

Erstad and McIntyre [35]RCT

Single centre

Double blind

Study quality, 6/9

36 patients with DFI, majority superficial infection (56%). 18 patients in each of 2 treatment arms. 44% vs 28% suspected or proven osteomyelitis in the A/S group versus the cefotixin group, respectively18 patients treated with A/S 3 g QID (A/S), 18 treated with cefoxitin, 2 g QID (Cef) for at least 5 days, in both groups combined with surgical intervention

Mean duration of hospitalization 21.1 (range 6–58) days in A/S group, 12.1 (range 4–39) days in Cef group (p = 0.06)

Infection outcomes: Cure (=complete alleviation of signs or symptoms of infection), improvement, bacteriological response, amputation, duration of hospitalizationCure 6% vs 39% (p = 0.03), improvement in 78% vs 50%, cure + improvement 15/17 vs 16/17, bacteriological response in 100% vs 73%, toe/ray amputation n = 7 vs n = 7, BKA n = 1 vs n = 1 and days of hospitalization 21.1 vs 12.1 in the A/S and cefotixin groups, respectively1+Unclear what day of treatment the assessment of clinical outcome was madeSponsored by Pfizer

Note: higher cure rate. Difficult to see why there is a difference in cure as opposed to cure plus improvement

Lipsky et al. [36]RCT

MulticentreOpen label

Study quality, 5/9

N = 108, 55 vs 53 subjects in the treatment arms. Prevalence of osteomyelitis 4/55 vs 1/53 in the ofloxacin and amoxicillin/sulbactam groups, respectively. Subjects with osteomyelitis included if the infected bone was removed55 subjects i.v. then oral ofloxacin vs 53 subjects A/S, then oral A/C

Mean duration of treatment i.v. ofloxacin 7.8 days, oral 13.2 days, A/S i.v. 7.1, oral 12.0 days, duration of treatment in osteomyelitis: ofloxacin i.v. 9.2 days oral 11.5 days, A/S i.v. 7.0, days, 12.9 days oral A/C

Infection outcomes Treatment of infection

Cured or improved, mean duration of therapy

No differences in outcomes between groups. Cured or improved 85% ofloxacin vs 83%

A/S. The mean duration of therapy with the ofloxacin regimen was 7.8 days (range, 1–25 days) intravenously and 13.2 days (range, 3–25 days) orally. The mean duration of therapy with the aminopenicillin regimen was 7.1 days (range, 1–20 days) intravenously and 12.0 days (range, 1–24 days) orally. Patients with osteomyelitis received a somewhat longer course of intravenous therapy (mean duration, 9.2 vs 7.0 days, respectively) but a slightly shorter course of oral therapy (mean duration, 11.5 vs 12.9 days, respectively) than did patients with only soft-tissue infections

1+20 Subjects non-evaluable

Persistence of streptococci in ofloxacin treatment group

Infected bone was supposed to be removed, but in the results it turned out that it was only removed in 71%

Numbers of osteomyelitis do not seem to match in the tables

Sponsored by Johnson Pharmaceuticals
Lipsky et al. [37]RTC


Open label

Study quality, 4/9

371 enrolled, of whom 10 were not treated, 241 vs 120 in each treatment arm

Prevalence of osteomyelitis 24% vs 17% in the linezolid and A/S group, respectively

241 linezolid, 120 A/S and A/C. Mean duration linezolid 17.2 ± 7.9 days, A/S 16.5 ± 7.9 days

Duration of i.v. linezolid therapy 7.8 ± 5.5 days, oral linezolid therapy 15.9 ± 7.4 days, duration of A/S therapy 10.4 ± 5.7 days, oral A/C therapy 15.0 ± 7.8 days

Infection outcomes: clinical cure and safety dataOverall, the clinical cure rates were statistically equivalent (linezolid 81% vs A/S 71%). Subjects with linezolid had a higher cure rate for infected DFU (81% vs 68%; p = 0.018) and in cases without osteomyelitis (87% vs 72%; p = 0.003)

Significantly more anaemia, thrombocytopenia and discontinuation of therapy in the linezolid group. Any event 26.6% vs 10.0% in the linezolid group versus the A/S group, respectively (p < 0.01)

1−Investigators diagnosed osteomyelitis in 77 patients. The analysis of clinical outcome by pathogen is a modified intent-to-treat population which consisted of patients in the intent-to-treat population with a baseline pathogen and evaluable clinical response of success or failure

The clinical cure rate is actually a per protocol analysis instead of an ITT analysis

Sponsored by Pfizer

Only study to show higher incidence of AEs in one group

Lipsky et al. [13]RCT Multicentre

Double blind

Study quality, 7/9

586 subjects with a DFI classified as moderate to severe and requiring intravenous antibiotics, 295 vs 291 subjects in each treatment arm. 12% of subjects had leg ulcers. Prevalence of osteomyelitis 8% vs 6% in the ertapenem and P/T groups, respectively. Osteomyelitis was surgically removed <48 hIntravenous ertapenem (1 g daily; n = 295) or P/T (3.375 g every 6 h; n = 291) given for a minimum of 5 days, after which oral A/C (875/125 mg every 12 h) could be given for up to 23 days

Mean duration of treatment 11.1 days for ertapenem and 11.3 days for P/T. Mean duration of oral follow-up therapy 9.7 days

Infection outcomes

Clinical cure, bacteria eradication, safety data

Of the 576 treated subjects, 445 were available for assessment at the end of intravenous therapy. Both baseline characteristics and favourable clinical response rates were similar for the 226 who received ertapenem and the 219 who received P/T (94% vs 92%, respectively). Rates of favourable microbiological responses (eradication rates and clinical outcomes, by pathogen) and adverse events did not differ between groups. No differences in number of adverse events1+Dropout rate 23% analysed by modified ITT

12% were leg ulcers. Data on site missing in n = 174. Proportion of cure for organisms resistant to ertapenem (pseudomonas and enterococci) was similar to success rates of P/T

Number of patients who received additional antibiotics is not mentioned

Sponsored by Merck

Only 11 days treatment duration

Lipsky et al. [38]RCT, Sub-analysis of multicentre

Double-blind, double dummy study

Study quality, 8/9

617 Subjects, hospitalized for DFI, 78 patients with DFIs available for treatment efficacy. Prevalence of osteitis 11% vs 20% in moxifloxacin versus P/T group, respectively. Bone infection was surgically ‘fully or partially resected’Moxifloxacin (400 mg/day) or P/T (3.0/0.375 g every 6 h) for at least 3 days, followed by moxifloxacin (400 mg/day orally) or A/C (800 mg every 12 h orally)

Duration of treatment: moxifloxacin i.v. 6.7 days, oral 7.4 days, A/C 6.3 days i.v., 7.9 days oral

Infection outcomes: clinical response of the infection at TOC, 10–42 days post-therapy, pathogen eradication, safety dataAmong 617 patients enrolled in the original study, 78 with DFIs were evaluable for treatment efficacy. Clinical cure rates at TOC were similar for moxifloxacin and P/T/A/C (68% vs 61%) for patients with infection (p = 0.54). Overall pathogen eradication rates in the microbiologically valid population were 69% vs 66% for moxifloxacin and comparator, respectively (p = 1.0)

No differences in safety outcomes

1++Patients with osteomyelitis were excluded if the bone could not be fully or partially resectedSponsored by Bayer

Short duration of total treatment

Senneville et al. [9]Cohort


Investigator blinded

Study quality, 4/7

50 patients with diabetic foot osteomyelitis treated in different centres, of whom 16 (32%) had already been treated for osteomyelitis of the footBone culture based antibiotic therapy

Duration of treatment 11.0 ± 4.1 weeks for success group and 12.4 ± 4.2 week for failure group (p = 0.19). In the two groups combined: 11.5 ± 4.2 weeks

Infection outcomes: Failure 18 (36%), 32 remission (64%). 20 predictive criteria were evaluatedPositive association: bone culture based antibiotic therapy 4 (22.2%) in failure group, 18 (56.3%) in remission group (p = 0.02). Multivariate analysis OR 4.78, CI 1.02–22.7 (p = 0.04)2+9 patients lost to follow-up

There was variability in practice and antibiotic therapy among centres

No sponsor identified

et al. [5]


Single centre

Open label

Study quality, 5/9

In patients with diabetes and severe DFI and who were known to have organisms sensitive to study drugs. Total number randomized N = 64, but two withdrawn early and not included in analysis. Group 1 n = 30 (osteomyelitis 73%), group 2 n = 32

(osteomyelitis 81%) (p = 0.05)

Group 1 i.v. P/T 4.5 g 8 hourly, group 2: i.v. I/c 0.5 g 6 hourly, with glycopeptide added if MRSA positive (n = 3), with excision of infected bone and with negative pressure therapy if necessary

Intended duration of treatment: 14 days for soft-tissue infection; 28 days for soft tissue plus bone but only 5 days if all infected bone removed surgically

Clinical success rate (success = total resolution of all symptoms and signs, without amputation)

Relapse within 2 months of hospital discharge

Treatment duration

Microbiological response

Total amputations

Major amputations

Adverse events

Group 1: 14 (46.7%)

Group 2: 9 (28.1%)

(p = 0.13)

0/14 vs 2/9

21 days vs 24 days

Complete in 96% in both groups

18 total amputations vs 22 total (p = 0.739)

22.5% of the whole group had a BKA

No difference between groups (p = 0.55)

1−Total number recruited was 64 and yet analysis conducted on only 62: technically per protocol analysis

Despite inclusion criteria, microbiological data available in only ‘approximately 80%’

57% isolated organisms were Gram-negative, reflecting disease duration and previous treatment

Mean duration of infection was 30 days and 40.5 days in the two groups – and yet all had had no antibiotics for 48 h prior to inclusion – which is surprising in people with severe infection

Not sponsored