Diagnosing Diabetic Foot Osteomyelitis

Early, accurate diagnosis of osteomyelitis is vital to inform timely decisions about managing or ruling out diabetic foot ulcer-related osteomyelitis (DFO) to avoid unnecessary antibiotic treatment. Techniques to diagnose or screen DFO include histological evaluation of a bone biopsy as the standard, probe-to-bone (PTB) test, clinical evaluation of signs and symptoms of infection, and culture of the ulcer specimen or radiographic signs of DFO. Each technique has its own criteria for a positive or negative diagnosis of osteomyelitis. Validity of a diagnostic technique is measured as sensitivity and specificity. For diagnosing DFO, a technique’s diagnostic sensitivity is the percent of all patients completing a study or interval of care with DFO confirmed by a bone biopsy, who were correctly diagnosed using the technique on enrollment. Specificity is the percent of patients completing the study without DFO, who were correctly diagnosed as DFO-free initially using this technique. Diabetic foot osteomyelitis screening measures estimate predictive validity of the technique on admission to care. Positive predictive value and negative predictive value are respectively the percent of all patients admitted that were correctly predicted to develop or not develop DFO. The percent of a technique’s correct positive and negative overall diagnosis is called diagnostic efficiency of the technique. False positives of either diagnostic or predictive validity can result in overuse of antibiotics with resulting development of resistant organisms. False negatives can delay treatment essential to save legs or lives. The 2 studies reviewed herein explore the validity of clinical signs, the PTB test, or plain or enhanced radiographic tests in diagnosing or screening DFO.^1,2

Reference: Morales Lozano R, González Fernández ML, Martinez Hernández D, Beneit Montesinos JV, Guisado Jiménez S, Gonzalez Jurado MA. Validating the probe-to-bone test and other tests for diagnosing chronic osteomyelitis in the diabetic foot [published online ahead of print July 9, 2010]. Diabetes Care. 2010;33(10):2140–2145.

Rationale: Histopathologic study of an involved bone specimen is the criterion standard for documenting diabetic foot osteomyelitis (DFO). However, surgically obtaining the bone specimen is used cautiously, because it is viewed as an invasive procedure that puts patients at risk. One alternative, PTB tests, routinely performed using a blunt instrument to detect palpable bone through the ulcer, merits stronger validation in diagnosing or predicting DFO.  

Objective: Provide data on the validity of the PTB test when used to detect DFO. 

Methods: A prospective, observational, descriptive study of patients with type 1 or 2 diabetes attending the University Podology Diabetic Foot Clinic of the Universidad Complutense of Madrid, Spain, from May 2006 through November 2008, examined diagnostic and predictive validity of the PTB test. Patients were enrolled if they had 1 full-thickness neuropathic or neuroischemic diabetic foot ulcer (DFU) clinically suspected to be infected or have DFO. This was defined using standardized tests for clinical infection based on (a) clinical signs and symptoms (ie, local heat, erythema, edema, foul odor, pain, purulent exudate) confirmed by a positive ulcer specimen soft tissue culture; (b) if they had undergone surgery for acute osteomyelitis; or (c) if after adequate local or antibiotic treatment and rest, the infection signs did not resolve. Presumptive DFO was diagnosed using clinical signs, plain radiography, soft tissue ulcer biopsy, and a PTB test. The recognized standard diagnostic test for DFO against which these results were validated was a standardized sterile bone biopsy exhibiting histological inflammatory cell infiltrate near spongy, cortical, or necrotic bone, or reactive bone formation or periosteal cell proliferation. Validity and interrater reliability statistical significance (P ≤ .05) of any diagnostic or screening metrics were calculated.

Results: Among the 132 patients enrolled with clinically suspected infections, 105 (79.5%) were diagnosed with DFO. Patients had a mean hemoglobin A1c of 7.9 and an average diabetes duration of 15 years, with a 44-week mean DFU duration. Most ulcers (59%) were neuropathic, with 41% classified as neuroischemic. Most commonly, the ulcers were Wagner grade III (93.9%) and Texas classification IIIA or IIIB (86.4%), with loss of protective sensation per vibration (91.7%) or monofilament (75.8%) tests used. The pedal pulse was absent in 25% of patients. The PTB test was the most reliable (P = .001), with the highest percentage of patients correctly diagnosed as with or without DFO (93.9%). This exceeded efficiency of clinical signs and symptoms of infection (59.1%), plain radiograph (75.8%), or ulcer specimen culture (72.0%) in diagnosing and/or predicting biopsy-confirmed DFO. Accuracy of DFO diagnosis using the PTB test was 98% for neuropathic ulcers and 88% for neuroischemic ulcers. 

Authors’ Conclusions: In this study sample with high (79%) prevalence of osteomyelitis, the PTB test was an easy-to-use, low-cost, reliable, and valid test for prompt diagnosis and screening of DFO, exceeding the diagnostic efficiency of clinical infection signs, plain radiograph, or ulcer specimen culture.  

Reference: La Fontaine J, Bhavan K, Lam K, Van Asten S, Erdman W, Lavery LA, Öz OK. Comparison between Tc-99m WBC SPECT/CT and MRI for the diagnosis of biopsy-proven diabetic foot osteomyelitis. Wounds. 2016;28(8):271–278.

Rationale: Technetium-99m (Tc-99m) white blood cell (WBC) single-photon emission computed tomography with computed tomography (SPECT/CT) combines radiolabelled WBCs with high-resolution X-ray CT, which offers a potential tool for improving diagnostic and prognostic value for DFO. 

Objective: Use a retrospective chart review to compare the diagnostic accuracy of Tc-99m SPECT/CT to that of magnetic resonance imaging (MRI), using bone biopsy as the gold standard for diagnosing DFO. 

Methods: Charts were reviewed for patients admitted to a tertiary care hospital in Dallas, Texas, from 2010 to 2013 with clinically suspected DFO confirmed by a bone biopsy. Included patients had received either a Tc-99m SPECT/CT or an MRI or both within 8 weeks of the biopsy and had standardized patient and foot ulcer data recorded as well as a positive culture result and/or histopathological confirmation of osteomyelitis in the bone sample obtained using clean needle biopsy. Diagnostic criteria for DFO using Tc-99m SPECT/CT were determined as WBC abutting the bone cortex or extending into the marrow space evaluated with dual-head, low-energy, high-resolution SPECT/CT collimators 2 hours after injecting 20–25 mCi of radiolabeled autologous Tc-99m WBC. Standardized T1, T2, and fat-suppressed MRI scans were made at a 3-mm slice thickness of the ankle and involved the forefoot or hindfoot using a specialized extremity coil.  Affected bone marrow was compared to adjacent normal fatty marrow, with low-intensity signals on T1-weighted images and high-intensity signals on fat-suppressed T2-weighted images being attributed to DFO. Incomplete or hazy signals or reticulated patterns were recorded as reactive bone marrow edema. Indeterminate or normal bone marrow signals indicated an absence of DFO. Descriptive data were presented as means and standard deviations of continuous variables or percentages or frequencies of categorical variables. Sensitivity, specificity, and positive predictive value (PPV) and negative predictive value (NPV) were calculated respectively to measure diagnostic and screening validity. 

Results: All 110 subjects enrolled had neuropathic DFUs and were comparable on enrollment in the Tc-99m SPECT/CT  (n = 52) and MRI (n  = 58) groups except for a lower percentage of subjects with type 1 diabetes in the MRI group. The 2 groups had similar DFO prevalence (67%). Both techniques had similar diagnostic validity: 87%–89% sensitivity; 35%–37% specificity; 74% PPV; and 58%–60% NPV, with similar (71%) accuracy in diagnosing biopsy-confirmed DFO. 

Authors’ Conclusions: Both techniques had similar diagnostic and screening efficiency, with high sensitivity, but only modest PPV compared to the gold standard bone biopsy in this study population. 

Diabetic foot osteomyelitis is a strong independent predictor of overall and major amputation.³ Early diagnosis and treatment can improve outcomes. Alternative diagnostic techniques studied by Morales Lozano et al¹ and by La Fontaine et al² are less invasive than the recognized standard DFO diagnostic tool: histologically confirmed bone biopsy. The PTB test appeared more reliable and efficient in diagnosing or screening DFO confirmed by bone biopsy than clinical signs of infection, plain radiography, or ulcer tissue bioburden¹ and approaches the diagnostic efficiency of combining the clinical criterion “ulcer depth > 3 mm,” with laboratory values of C-reactive protein > 3.2 mg/dL or erythrocyte sedimentation rate > 60 mm/h.⁴ Although patient samples diagnosed by these techniques were comparable on some important baseline variables, a retrospective observational study design cannot rule out subtle differences in subjects, environment, or clinicians that may have influenced the choice of diagnostic technique and validity results. To rule out all aspects of bias, the same 3 tests would need to be run simultaneously on all subjects, or a large cohort of subjects would need to be randomly assigned to each diagnostic technique. Additional bone-contact swabs would help identify the infecting organism.⁵ La Fontaine et al² showed MRI or Tc-99m SPECT/CT did not differ from each other in diagnostic and screening efficiency, but their results did not match the reliability or diagnostic or screening efficiency of the PTB test, which reported higher reliability and validity and was easier to do and less costly than other diagnostic tests.^1,2,6 The reported low specificity of both MRI and Tc-99m SPECT/CT (35%–37%) in diagnosing DFO is a special concern, because it means more false positive diagnoses of DFO, leading to unnecessary antibiotic use, which increases patient likelihood of developing antibiotic-resistant microorganisms. The reliability and diagnostic and screening efficiency of the PTB test in identifying DFO suggest the need for future research exploring the value of this simple, low-cost test in diagnosing and screening osteomyelitis in the bone underlying other chronic wounds such as venous⁷ or pressure ulcers.⁸ 

Laura Bolton, PhD
Department of Surgery
Rutgers Medical University
New Brunswick, NJ

This article was not subject to the WOUNDS peer-review process.