Efficacy of Magnetic Resonance Imaging in Diagnosing Diabetic Foot Osteomyelitis in the Presence of Ischemia

The Journal of Foot & Ankle Surgery 52 (2013) 717–723

Contents lists available at ScienceDirect

The Journal of Foot & Ankle Surgery journal homepage: www.jfas.org

Efficacy of Magnetic Resonance Imaging in Diagnosing Diabetic Foot Osteomyelitis in the Presence of Ischemia Miki Fujii, MD, PhD 1, David G. Armsrong, DPM, MD, PhD, FACFAS 2, Hiroto Terashi, MD, PhD 3 1

Department of Plastic and Reconstructive Surgery, Ono Municipal Hospital, Ono, Japan Southern Arizona Limb Salvage Alliance, University of Arizona College of Medicine, Tucson, AZ 3 Department of Plastic and Reconstructive Surgery, Kobe University Hospital, Kobe, Japan 2

a r t i c l e i n f o

a b s t r a c t

Level of Clinical Evidence: 2

Magnetic resonance imaging (MRI) has been recognized as the most accurate imaging modality for the detection of diabetic foot osteomyelitis. However, how accurately MRI displays the extent of diabetic foot osteomyelitis in the presence of ischemia is still unclear. We retrospectively compared the preoperative MRI findings with the results of histopathologic examinations of resected bones and studied the efficacy of MRI in the diagnosis of diabetic foot osteomyelitis of different etiologies. A total 104 bones from 18 foot ulcers in 16 diabetic patients (10 men and 6 women; age range 42 to 84 years) treated by surgical intervention from 2008 to 2012 was examined. In 8 neuropathic ulcers, 29 bones were accurately diagnosed in detail using MRI, even those with severe soft tissue infection. Of 75 bones in 10 ischemic ulcers, only 7 bones evaluated by MRI after revascularization were diagnosed accurately; the other 68 could not be diagnosed because of unclear or equivocal MRI findings. On histopathologic examination, all the bones were found to be infected through the bone cortex by the surrounding infected soft tissue, not directly by articulation. Overall, preoperative MRI is effective in the diagnosis of neuropathic ulcers, but less so of ischemic ones. Ó 2013 by the American College of Foot and Ankle Surgeons. All rights reserved.

Keywords: biopsy bone infection critical limb ischemia culture and sensitivity diabetes mellitus histopathology neuropathic ulcer peripheral neuropathy

The presence of diabetic foot osteomyelitis (DFO) affects the selection of treatment and the prognosis (1). Magnetic resonance imaging (MRI) has been recognized as the most accurate imaging technique for detecting DFO (2,3) and for evaluating soft tissue involvement. However, its efficacy in the evaluation of DFO in the presence of ischemia is unclear. The efficacy of limited resection combined with antibiotic therapy has been demonstrated (4,5), and guidelines (2) have recommend the duration of antibiotic therapy should be determined by the presence of residual osteomyelitis (OM). Consequently, additional details on the extent of DFO are needed. In the present study, we retrospectively compared the preoperative MRI findings with the results from histopathologic examination of the resected bones and evaluated the efficacy of MRI in the diagnosis of DFO in different etiologic types of ulcers (6).

according to the World Health Organization classification. Six patients had undergone hemodialysis for diabetic nephropathy. All the ulcers were classified into 4 types according to the main etiologic factors: type I, neuropathic ulcers; type II, ischemic ulcers; type III, neuropathic ulcers with soft tissue infection; and type IV, ischemic or neuroischemic ulcers with soft tissue infection (7) (Table 1). Type I ulcers result mainly from the peripheral neuropathy (PN). They typically occur under callus formation at high pressure points of the deformed foot, where no soft tissue infection is apparent and the bacteria manifest as wound colonization (8,9). Type II ulcers result mainly from peripheral arterial disease (PAD), are the same as those of Fontaine score IV (the so-called critical limb ischemia), are a complication often observed in hemodialysis patients, and appear cold and hairless, with no sign of infection. Type III ulcers are neuropathic, with soft tissue infection, often severe, resemble necrotizing fasciitis, and demonstrate gas gangrene. Type IV ulcers are caused by a combination of PAD and soft tissue infection, with or without PN. The pathway of most infections can be divided into 3 routes: from an ulcerated bunion of the first or fifth toe, from mal perforans ulcers after callus formation at high pressure points, and from an abscess in the central plantar space

Patients and Methods

Table 1 Classification of ulcers according to main etiologic findings (7)

The records of 16 consecutive patients with diabetic foot ulcers (10 men and 6 women, mean age 66.7 years, range 42 to 84), who had undergone surgery from 2008 to 2012, were examined. All the patients had been diagnosed with type 2 diabetes

Financial Disclosure: None reported. Conflict of Interest: None reported. Address correspondence to: Miki Fujii, MD, PhD, Department of Plastic Surgery, Ono Municipal Hospital, 323 Naka-cho, Ono 675-1332, Japan. E-mail address: [email protected] (M. Fujii).

Main etiologic factors

Classification

Ulcers (n)

Neuropathic ulcers Ischemic ulcers (CLI) Neuropathic ulcers with soft tissue infection CLI with soft tissue infection Total

Type Type Type Type d

3 3 5 7 18*

Abbreviation: CLI, critical limb ischemia. * Two patients, each with 2 different types of ulcers.

1067-2516/$ - see front matter Ó 2013 by the American College of Foot and Ankle Surgeons. All rights reserved. http://dx.doi.org/10.1053/j.jfas.2013.07.009

I II III IV

718

M. Fujii et al. / The Journal of Foot & Ankle Surgery 52 (2013) 717–723

Fig. 1. Diagnosis using magnetic resonance imaging of patient 1, right fifth toe. (A) T1-weighted image. (B) Fat-suppressed T2-weighted image. a, low intensity (A) and high intensity (B); b, incomplete or hazy signals or reticulated patterns; c, normal bone marrow signal. Magnetic resonance imaging diagnosis: a, osteomyelitis; b, bone marrow edema; c, normal bone.

through a web infection (10,11). The feet will demonstrate critical limb ischemia resulting from these infections. Type IV ulcers often do not show any sign of soft tissue infection after such critical limb ischemia because of the severe ischemia; however, bacteria are present as a critical colonization. Therefore, the infection is sometimes exacerbated after revascularization. PN was assessed by touch sensation using the Ipswich touch test (12). PAD was suspected at first because of no palpation and/or no sound using a Doppler stethoscope of the dorsalis pedis artery and posterior tibial artery pulses, an ankle-brachial index less than 0.9, and skin perfusion pressure less than 40 mm Hg. Patients with suspected PAD were evaluated using computed tomographic angiography, duplex ultrasonography, and angiography. Soft tissue infection was assessed by general conditions such as high fever, chills, or malaise, clinical findings such as redness, warmth, swelling, or purulent secretions, or laboratory examination results. OM was suspected in patients with a positive probe-to-bone test, swollen foot, sausage toe, an unexplained high leukocyte count or inflammatory markers, and plain foot radiographic findings (1). Preoperative MRI was performed with 1.5 T MR scanners (slice thickness 3 mm, T1-weighted images, conventional spin echo, repetition time/echo time 460 ms/13 ms; fat-suppressed T2-weighted images, short-tau inversion recovery images, short T1 inversion recovery, and repetition time/echo time/inversion time 3,401 ms/80 ms/150 ms). All MRI findings were checked by 2 of us (M.F. and H.T.) with consensus interpretation. The affected bone marrow was compared with the adjacent normal fatty marrow: low-intensity signals on T1-weighted images and high-intensity signals on fat-suppressed T2-weighted images were attributed to OM. Incomplete or hazy signals or reticulated patterns were attributed to reactive bone marrow edema (BME). Normal bone marrow signals were considered indicative of areas clear of disease (13). All bones were marked according to the presence of OM, BME, or normal bone (Fig. 1). Surgery after the MRI diagnosis included resection of the infected bones and gangrenous tissue, amputation, or disarticulation. The area of resection was estimated preoperatively from the MRI findings, blood flow, and area of soft tissue infection and confirmed intraoperatively. A definitive diagnosis was then made from the histopathologic examination of 5-mm-thick sections of formalin-fixed and paraffin-embedded bones stained with hematoxylin and eosin. Only BME or infiltration of inflammatory cells, or both, was considered indicative of reactive BME; however, these 2 conditions, together with the presence of osteonecrosis, granulation tissue, and/or fibrosis, were considered indicative of OM (Fig. 2). The diagnosis from the preoperative MRI findings was compared with the definitive diagnosis from the histopathologic examination findings.

Results A total of 18 ulcers, from 17 feet of 16 patients, including 2 patients with 2 different types of ulcers each, was classified into 4 types according to the main etiologic factors: type I, neuropathic ulcers (n ¼ 3); type II, ischemic ulcers (n ¼ 3); type III, neuropathic ulcers with soft tissue infection (n ¼ 5); and type IV, ischemic or neuroischemic ulcers with soft tissue infection (n ¼ 7; Table 1). The 16 patients underwent a total of 19 MRI examinations before each of 19 surgeries (Table 2). The average interval between the MRI examinations and surgery was 16.7 (range 2 to 48) days. All ischemic ulcers under the type II and type IV classification were revascularized by bypass or endovascular treatment before surgery. The MRI examinations were performed as follows: 3 examinations for 3 type I ulcers; 3 examinations before revascularization of 3 type II ulcers and 1 after revascularization of 1 type II ulcer; 5 for 5 type III ulcers; and 3 before revascularization of 3 type IV ulcers and 4 after revascularization of 4 type IV ulcers. Pre- and postrevascularization MRI was done for 1 patient who underwent 2 surgeries for 1 type II ulcer (patient 3, see Fig. 5): 1 for resection of the third, fourth, and fifth toes and the other for resection of the first and second toes. A total of 104 bone specimens (distal phalanx, 27; middle phalanx, 20; proximal phalanx, 32; metatarsal bone, 24; and cuboid bone, 1) was obtained from 39 toes: type I (11 bones from 3 ulcers), type II (40 bones from 3 ulcers), type III (18 bones from 5 ulcers), and type IV (35 bones from 7 ulcers; Table 3). Histologic analysis of all bone specimens revealed the presence of OM, BME, normal bone, or gangrene. The histopathologic features of the bone marrow in the resected bones corresponded to the MRI findings for types I and III in every detail (Table 4). OM was detected in 23 bones, with a sensitivity and specificity of 100%. In type II, however, none of the 40 bones was accurately diagnosed using MRI because of unclear or equivocal images (Table 5). Only 7 bones from type IV ulcers examined by postrevascularization MRI were accurately

Table 2 Details of magnetic resonance imaging (MRI) examinations

Fig. 2. Histopathologic diagnosis (hematoxylin and eosin stain, original magnification 40). (A) Osteomyelitis with bone marrow edema, inflammatory cells, osteonecrosis, granulation, and fibrosis. (B) Bone marrow edema with inflammatory cells.

Type

Ulcers (n)

MRI (n)

MRI before and after Revascularization

I II

3 3

3 4*

III IV

5 7

5 7

d Before revascularization: 3 After revascularization: 1 d Before revascularization: 3 After revascularization: 4

Average interval between MRI and surgery, 16.7 days. * One type II ulcer (patient 3) underwent both pre- and postrevascularization MRI.

M. Fujii et al. / The Journal of Foot & Ankle Surgery 52 (2013) 717–723

719

Table 3 Bone specimens of each ulcer type Type

Ulcers (n)

Toes (n)

Bones (n)

Distal Phalanx (n)

Middle Phalanx (n)

Proximal Phalanx (n)

Metatarsal (n)

Cuboid (n)

I II III IV Total

3 3 5 7 18

4* 11 7 16 39

11 40 18 35 104

3 12 5 7 27

3 9 2 6 20

3 11 7 11 32

1 8 4 11 24

1

*

1

Four toes and cuboid bones.

diagnosed; the other 28 bones still could not be diagnosed because of unclear or equivocal images. Histopathologic analysis revealed OM in 4 bones of type II, none of which were diagnosed using MRI (Table 6). OM was detected in 22 bones of type IV, but only 6 bones were correctly diagnosed by postrevascularization MRI, with a sensitivity of 27.3% (Table 7). All bones were infected through the bone cortex by the surrounding infected soft tissue, not directly by articulation (types I, III, and IV; Fig. 3). All the wounds healed.

infected soft tissue, not directly by articulation (Fig. 4E). The wound healed without any complications. (Fig. 4F). Patient 3

Case Report

In a 70-year-old man who required hemodialysis, with type II multiple ulcers in the left foot, no bones could be diagnosed before revascularization (ankle-brachial index 0.5) because of unclear or equivocal MRI findings (Fig. 5A). After revascularization (anklebrachial index 1.05), however, the images had improved slightly (Fig. 5B).

Patient 2

Patient 4

A 64-year-old man with a type III ulcer on the right fifth toe, where the infection had spread from the bunion, demonstrated a high leukocyte count and inflammatory markers, and complained of general malaise (Fig. 4A). A radiograph showed destruction of the fifth metatarsal head (Fig. 4B). From the MRI findings, the fifth metatarsal bone was divided into 3 parts: (1) OM, (2) reactive BME, and (3) normal bone to OM, BME, and normal bone (Fig. 1). Similarly, the distal and middle phalanges were diagnosed as reactive BME. In the proximal phalanx, the distal part was diagnosed as reactive bone marrow, and the proximal part as OM (BME/OM). The area of resection was estimated in the area diagnosed as BME in the metatarsal bone (Fig. 4C). At surgery, the devitalized or severely infected skin and soft tissue were resected. The area of bone resection was almost the same as that estimated preoperatively (Fig. 4D). The histopathologic analysis confirmed our MRI diagnosis (Fig. 4E). The bones were infected through the bone cortex by the surrounding

In a 54-year-old woman who required hemodialysis, with a type IV ulcer, all the toes of the right foot were gangrenous (Fig. 6A). The MRI findings of the gangrenous site were unclear or equivocal (Fig. 6B), and the histopathologic examination showed acellularity at the site of gangrene (Fig. 6C).

Table 4 Magnetic resonance imaging (MRI) and histopathologic diagnosis of types I and III Type Ulcers (n) Bones (n) MRI Diagnosis (n) Histopathologic Diagnosis (n) Correct Incorrect OM OM/BME OM/BME/N BME N I III

3 5

11 18

11 18

0 0

6 6

4 5

0 2

1 2

0 3

Abbreviations: BME, bone marrow edema; MRI, magnetic resonance imaging; N, normal bone; OM, osteomyelitis.

Before revascularization (group A) Type II Type IV After revascularization (group B) Type II Type IV * y

The presence of DFO increases the likelihood of surgical intervention, including amputation, the duration of antibiotic therapy, and the recurrence of infection (1). The efficacy of limited resection, combined with antibiotic administration, has been shown to reduce the changes in the biomechanics of the foot and to shorten the duration of antibiotic therapy (3,4). The guidelines have recommended that the duration of antibiotic therapy be determined by presence and amount of residual dead or infected bone and the status of the soft tissue (2). At present, therefore, recognition of the extent, and additional details, of DFO are needed. MRI has been recognized as the most accurate imaging modality for detecting DFO, with high sensitivity (90% to 100%) and specificity (80% to 100%) (13–20). However, the efficacy of MRI in accurately displaying the extent of DFO in the presence of ischemia is still unclear. The MRI diagnosis of OM is made according to the concordance of primary signs and supportive secondary signs (13). The primary signs on MRI are abnormal bone marrow signals, marrow replacement by

Table 6 Histopathologic diagnosis of type II

Table 5 Accuracy of magnetic resonance imaging (MRI) diagnosis for types II and IV MRI Diagnosis

Discussion

Ulcers

Bones

Correct

Incorrect*

Type II

Bones (n)

Histopathologic Diagnosis (n) OM (n)

BME

BME/N

Gangrene

N

3 3

33 8

0 0

33 8

Group A* Group Bz

33y 7y

4 0

1 4

1 2

27 0

0 1

1y 4

7 27

0 7

7 20

Because of unclear or equivocal MRI findings. Pre- and post-vascularization MRI performed for 1 ulcer, type II (patient 3).

Abbreviations: BME, bone marrow edema; MRI, magnetic resonance imaging; N, normal bone; OM, osteomyelitis. * Group A, Prerevascularization MRI diagnosis. y Because of unclear or equivocal MRI findings. z Group B, postrevascularization MRI diagnosis.

720

M. Fujii et al. / The Journal of Foot & Ankle Surgery 52 (2013) 717–723

Table 7 Histopathologic diagnosis of type IV Type IV

Group A Group By Correct Incorrect

Bones (n)

8* 7 20

Histopathologic Diagnosis (n) OM (n)

OM/BME (n)

OM/BME/N

OM/Gangrene

BME

BME/N

Gangrene

0

2

0

1

2

2

1

1 6

4 4

1 1

0 2

1 0

0 0

0 7

Abbreviations: BME, bone marrow edema; MRI, magnetic resonance imaging; N, normal bone; OM, osteomyelitis. * Incorrect because of unclear or equivocal MRI findings. y Correct indicates that the bones were diagnosed correctly using MRI and incorrect indicates that the bones could not be diagnosed using MRI because of equivocal or unclear images.

Fig. 3. Patient 1. A 42-year-old woman with a type III ulcer in the left first distal phalanx (A and B). The infection went through the bone cortex from the surrounding infected soft tissue (arrow), but the articulation was intact (C).

low-signal intensity on T1-weighted images, correspondingly highsignal intensity on fat-suppressed T2-weighted images, and enhancement after the use of contrast medium. The secondary signs are surrounding soft tissue changes, soft tissue infection, abscess, fistula, ulcer, and sinus tract. Generally, OM can be excluded when the MRI findings are negative or when the altered bone marrow signal is remote from the ulcer (21). However, positive findings from an MRI study are not always sufficient to establish the diagnosis of OM (13–19). The reason for the limited specificity is the presence of reactive BME, including acute neuropathic arthropathy or recent surgery. After surgery, the abnormal signal can persist for 3 to 6 months (17). The use of contrast-enhanced images to distinguish OM from reactive BME has been controversial (16,18). Because of the severe renal insufficiency in our diabetic patients, the use of contrast medium was avoided in the present study. It is now well established from published studies (13,15,19) that the extent of OM tends to be overestimated on T2-weighted images or short T1 inversion recovery images; that the extent of OM correlates best when approximated using T1-weighted images; and that fat-suppressed T2-weighted images tend to be of high-signal intensity in the presence of reactive BME. In the present study, therefore, T1-weighted images were conducive to correctly distinguishing OM from reactive BME. Although diabetic foot ulcers develop from multifactorial conditions, their principal etiologies are PN, ischemia, and infection. Such ulcers can be divided into 3 types according to the principal cause: neuropathic ulcers, neuroischemic ulcers, and ischemic ulcers, each with characteristics differing from the others. The characteristics will show greater variation when complicated by soft tissue infection. No comparison of the effectiveness of MRI in the diagnosis of DOF under these different characteristics has been previously published. We categorized the ulcers into 4 types and investigated how effective MRI would be in evaluating the extent of DFO in each ulcer type (6). In our study, MRI of neuropathic type I and III ulcers clearly distinguished OM from reactive BME, even in the presence of severe soft tissue infection. The microscopic changes observed in the bone marrow closely corresponded to the MRI findings. MRI was not useful in the diagnosis of ischemic type II and IV ulcers, because MRI depends on the presence of fluid. Thus, before revascularization, the images were unclear or equivocal because of insufficient interstitial fluid. After revascularization, however, the images improved slightly for type IV ulcers but remained unclear or equivocal for type II ulcers, regardless of whether the scans were taken before or after revascularization. The type II ulcers were gangrenous owing to ischemia; not only was the interstitial fluid insufficient, but also the OM cells, granulation tissue, and infected soft tissue were absent, resulting in unclear or equivocal MRI findings. Therefore, preoperative MRI is not necessary for type II ulcers but should be used for type IV, at least after revascularization.

M. Fujii et al. / The Journal of Foot & Ankle Surgery 52 (2013) 717–723

721

Fig. 4. Patient 2. A 64-year-old man with a type III ulcer on the right fifth toe. (A) The infection had spread over the right fifth toe from the bunion. (B) Radiograph showing destruction of the fifth metatarsal head (arrow), which accorded with the ulcer. (C) The area of resection was estimated in the area diagnosed as bone marrow edema (asterisk and arrow indicate site of cut). (D) The area of resection was almost correspondent to the preoperative estimate. (E) Histopathologic diagnosis (hematoxylin and eosin stain, original magnification 40). The histopathologic analysis confirmed our magnetic resonance imaging diagnosis. The bones were infected through the bone cortex by the surrounding infected soft tissue, not directly by articulation (asterisk). (F) The patient was able to walk with footwear 2 years after surgery.

The histopathologic examinations revealed that the bones were infected through the bone cortex by the surrounding soft tissue in types I, III, and IV (Figs. 3 and 4E). When soft tissue infection develops or bone is exposed to organisms colonizing the ulcer, bacteria penetrate the cortical bone and gain access to the marrow cavity. Although this pathogenesis of bacterial OM has been previously described (14,22), our study has actually demonstrated it by histopathologic analysis. Soft tissue infection was not apparent in the type I ulcers and was sometimes present in the type IV ulcers. Nonetheless, histopathologic analysis revealed bacterial contamination or colonization in the surrounding soft tissue in both types. Because ischemia arises from soft tissue infection in type IV, the infection itself decreases and is latent; however, bacteria are still present as a critical colonization and

sometimes exacerbate the infection again after revascularization. In our study, histopathologic analysis disclosed the presence of critical colonization. The pathogenesis showed the importance of preventing soft tissue infection and of preventing its spread to the tendons or ligaments; immobilization with a cast is inevitable, even in the presence of a critical colonization. The present study had several limitations. First, the definitive diagnosis of OM requires both histopathologic findings and isolation of bacteria from an aseptically obtained bone sample (3). Our definitive diagnosis was made finally from the histopathologic findings. We could not always obtain the cultured samples aseptically from the bone in all patients. Second, our bone samples were from the toes, metatarsals, and 1 cuboid bone. Although the forefoot is not the only

722

M. Fujii et al. / The Journal of Foot & Ankle Surgery 52 (2013) 717–723

Fig. 5. Patient 3 was a 70-year-old man who required hemodialysis with type II multiple ulcers of the left foot. (A) Clinical appearance and magnetic resonance imaging scan before revascularization. (B) Clinical appearance and magnetic resonance imaging scan after revascularization. T1WI, T1-weighted; T2IR, T2-weighted, inversion recovery.

site of diabetic foot ulcers, it is the most common (15,23). Third, the pre- and postrevascularization MRI scans were compared only for patient 3, with both images unclear or equivocal because of an insufficient blood supply. Also, repeated MRI of the same patient is economically impractical. Fourth, the classification (7) chosen for categorizing ulcers is not universal. Generally, the well-known Wagner classification (24) and the University of Texas wound classification system (23) have focused on the wound depth, ischemia, PN and infection. Our choice was made with the purpose of the present study in minddto explore further the efficacy of MRI in diagnosing DFO in the presence of ischemia. Our Kobe classification (7) is a hybrid of the University of Texas and Wagner classifications and has recently been used in Japan (Table 8).

In Asia, type 2 diabetes is becoming epidemic, characterized by rapid rates of increase within a short period, onset at a relatively young age, and a low body mass index (25). A distinct discipline of podiatric medicine such as that in United States or Europe is not available in Asia; therefore, the establishment of a classification system that is more conducive to Asian populations is requisite. In conclusion, although the present study included a limited number of subjects, it has demonstrated that in neuropathic ulcers, DFO can be reliably distinguished from reactive BME in every detail, even in the presence of severe soft tissue infection. MRI was not useful, however, in the diagnosis of DFO in ischemic ulcers because of the insufficient interstitial fluid. Our onward studies will focus on salvaging as much of the diabetic foot as possible.

M. Fujii et al. / The Journal of Foot & Ankle Surgery 52 (2013) 717–723

723

Fig. 6. Patient 4 was a 54-year-old woman who required hemodialysis with a type IV ulcer of the right foot. (A) All the toes were gangrenous. (B) Magnetic resonance imaging of the gangrenous part was unclear. (C) Histopathologic examination showed acellularity at the site of the gangrene (hematoxylin and eosin stain, original magnification 100).

Table 8 Kobe classification and treatment of each wound type Type

Pathophysiology

Main Treatment

I II III

Neuropathic ulcers Ischemic ulcers (CLI) Neuropathic ulcers with soft tissue infection CLI with soft tissue infection

Pressure relief Revascularization Early debridement

IV

Revascularization and debridement on a case-by-case basis

Abbreviation: CLI, critical limb ischemia. Data from Terashi H, Kitano I, Tsuji Y. Total management of diabetic foot ulcerations–Kobe classification as a new classification of diabetic foot wounds. Keio J Med 60:17–21, 2011.

References 1. Lipsky BA, Berendt AR, Deery HG, Embil JM, Joseph WS, Karchmer AW, LeFrock JL, Lew DP, Mader JT, Norden C, Tan JS; Infectious Diseases Society of America. Diagnosis and treatment of diabetic foot infections. Plast Reconstr Surg 117:212S– 238S, 2006. 2. Lipsky BA, Berendt AR, Cornia PB, Pile JC, Peters EJ, Armstrong DG, Deery HG, Embil JM, Joseph WS, Karchmer AW, Pinzur MS, Senneville E; Infectious Diseases Society of America. 2012 Infectious Diseases Society of America clinical practice guideline for the diagnosis and treatment of diabetic foot infections. Clin Infect Dis 54:e132–e173, 2012. 3. Lipsky BA, Peters EJG, Senneville E, Berendt AR, Embil JM, Lavery LA, Urban ci cRovan V, Jeffcoate WJ. Expert opinion on the management of infections in the diabetic foot. Diabetes Metab Res Rev 28(suppl 1):163–178, 2012. n-Sa nchez J. Treatment of diabetic foot osteomyelitis: a surgical critique. Int J 4. Arago Low Extrem Wounds 9:37–59, 2010. 5. Ha Van G, Siney H, Danan JP, Sachon C, Grimaldi A. Treatment of osteomyelitis in the diabetic foot: contribution of conservative surgery. Diabetes Care 19:1257– 1260, 1996. 6. Fujii M, Terashi H. Efficacy of magnetic resonance imaging in diagnosing diabetic foot osteomyelitis. J Am Podiatr Med Assoc (in press). 7. Terashi H, Kitano I, Tsuji Y. Total management of diabetic foot ulcerationsdKobe classification as a new classification of diabetic foot wounds. Keio J Med 60:17–21, 2011. 8. Smith AJ, Daniels T, Bohnen JM. Soft tissue infections and the diabetic foot. Am J Surg 172:7S–12S, 1996.

9. Schultz GS, Sibbald RG, Falanga V, Ayello EA, Dowsett C, Harding K, Romanelli M, Stacey MC, Teot L, Vanscheidt W. Wound bed preparation: a systematic approach to wound management. Wound Repair Regen 11(suppl 1)):S1–28, 2003. 10. Meade JW, Mueller CB. Major infections of the foot. Med Times 96:154–165, 1968. 11. O’Neal LW. Surgical pathology of the foot and clinicopathologic correlations. In: The Diabetic Foot, pp. 483–512, edited by ME Levin, LW O’Neal, Mosby, St. Louis, 1983. 12. Rayman G, Vas PR, Baker N, Taylor CG Jr, Gooday C, Alder AI, Donohoe M. The Ipswich touch test: a simple and novel method to identify inpatients with diabetes at risk of foot ulceration. Diabetes Care 34:1517–1518, 2011. 13. Morrison WB, Schweitzer ME, Batte WG, Radack DP, Russel KM. Osteomyelitis of the foot: relative importance of primary and secondary MR imaging signs. Radiology 207:625–632, 1998. 14. Shank CF, Feibel JB. Osteomyelitis in the diabetic foot: diagnosis and management. Foot Ankle Clin 11:775–789, 2006. 15. Donovan A, Schweitzer ME. Current concepts in imaging diabetic pedal osteomyelitis. Radiol Clin North Am 46:1105–1124, 2008. 16. Rozzanigo U, Tagliani A, Vittorini E, Pacchioni R, Brivio LR, Caudana R. Role of magnetic resonance imaging in the evaluation of diabetic foot with suspected osteomyelitis. Radiol Med 114:121–132, 2008. 17. Gnanasegaran G, Chicklore S, Vijayanathan S, O’Doherty MJ, Fogelman I. Diabetes and bone: advantages and limitations of radiological, radionuclide and hybrid techniques in the assessment of diabetic foot. Minerva Endocrinol 34:237–254, 2009. 18. Collins MS, Schaar MM, Wenger DE, Mandrekar JN. T1-weighted MRI characteristics of pedal osteomyelitis. AJR Am J Roentgenol 185:386–393, 2005. 19. Russell JM, Peterson JJ, Bancroft LW. MR imaging of the diabetic foot. Magn Reson Imaging Clin North Am 16:59–70, 2008. 20. Ahmadi ME, Morrison WB, Carrino JA, Schweitzer ME, Raikin SM, Ledermann HP. Neuropathic arthropathy of the foot with and without superimposed osteomyelitis: MR imaging characteristics. Radiology 238:622–631, 2006. 21. Thorning C, Gedroyc WM, Tyler PA, Dick EA, Hui E, Valabhji J. Midfoot and hindfoot bone marrow edema identified by magnetic resonance imaging in feet of subjects with diabetes and neuropathic ulceration is common but of unknown clinical significance. Diabetes Care 33:1602–1603, 2010. 22. Ciampolini J, Harding KG. Pathophysiology of chronic bacterial osteomyelitis: why do antibiotics fail so often? Postgrad Med J 76:479–483, 2000. 23. Armstrong DG, Lavery LA, Harkless LB. Validation of a diabetic wound classification system: the contribution of depth, infection, and ischemia to risk of amputation. Diabetes Care 21:855–859, 1998. 24. Wagner FW. The dysvascular foot: a system for diagnosis and treatment. Foot Ankle 2:64–122, 1981. 25. Chan JC, Malik V, Jia W, Kadowaki T, Yajnik CS, Yoon KH, Hu FB. Diabetes in Asia: epidemiology, risk factors, and pathophysiology. JAMA 301:2129–2140, 2009.