Differential diagnosis of osteomyelitis in the diabetic foot using combined radionuclide methods

Aim: The aim of this prospective study was to evaluate the contribution of combined radionuclide methods, mainly based on ⁹⁹ᵐTc-HMPAO-labelled leucocyte scan (HMPAO-LS), to the differential diagnosis of osteomyelitis (OM) in the diabetic foot. Patients and methods: Forty five patients with diabetes mellitus (20 men, 8 with type 1), aged 59.8±9.9 years, with mean duration of diabetes 15.9+9.4 years and clinical suspicion of OM in 54 pedal sites (36 in the forefoot and 18 in the mid/hidfoot) were enrolled in the study. All patients were presented with pre-existing pedal abnormalities: 36 patients with a total of 46 pedal ulcers, 27 with Charcot arthropathy (4 of them bilateral) and 21 with previous toe or foot amputation. After physical examination of the foot, C-reactive protein (CRP) levels, erythrocyte sedimentation rate (ESR), total (WBC) and differential white blood cell count and blood glucose and glycohemoglobin (HbAlc) levels were measured in all patients. Plain film radiographs of the foot and ⁹⁹ᵐTc-MDP-bone scan (MDP) were obtained in all patients followed by ⁹⁹ᵐTc-HMPAO-labelled leucocyte scan (HMPAO-LS) in 36 of them (45 foot sites). In order to interpret abnormal leucocyte uptake in sites of Charcot arthropathy or amputation, ⁹⁹ᵐTc-tin colloid bone marrow scan (BMS) was additionally performed in 6 patients. Scintigraphic images were interpreted according to the following criteria: MDP scan showing focal arterial hyperperfusion, hyperemia and focally increased bony uptake on delayed images was considered compatible with OM. HMPAO-LS demonstrating abnormal leucocyte accumulation in a site concordant with the area of uptake on bone scan (spatially congruent images) was considered positive for OM. Leucocyte activity without corresponding BMS activity (incongruent HMPAO-LS/BMS images) indicated OM, while congruent HMPAO-LS/BMS images suggested the presence of active bone marrow due to Charcot arthropathy or amputation. Lesion to contralateral foot ratios were calculated in all radionuclide studies. Diagnosis was confirmed by long-term (at least 6 months) clinical and radiological follow-up or bone biopsy. Follow-up HMPAO-LS was additionally performed in 7 patients to evaluate response to antibiotic treatment. Results: Among the 54 pedal sites investigated, 26 foci of OM (48.1%), 6 cases with acute Charcot arthropathy (ACA) and 18 sites with isolated soft tissue infection (STI) were diagnosed. Stress fracture was diagnosed in 2 cases, acute arthritis of metatarsophalangeal joint in 1 case and diabetic gangrene in 1 case. Underlying OM was present in 50% of ulcers in this study, while 88.5% of the OM cases were pedal ulcer related. No association was found between the presence of OM and the duration, the surface area of the ulcer or the presence of pus discharge. Conversely, there was a correlation between the depth or Wagner’s stage of the ulcer and the frequency of OM. OM patients showed higher CRP levels compared to those with ACA or STI (24.6 vs 10.3 and 7.5 mg/L respectively, p<0.05). Conversely, ESR, WBC count and neutrophil percentage were not significantly different among the main three groups of patients. ESR was high in most patients, while severe leucocytosis or elevated neutrophil percentage were not observed in any group. Blood glucose and HbAlc levels were equally high in all three groups of patients. Fever was absent in the 78% of patients with OM. In the present study, plain radiography was proven insensitive (34.6%), but highly specific (96.4%) in the diagnosis of diabetic foot OM. The accuracy, positive predictive value (PPV) and negative predictive value (NPV) of radiographs were 66.6%, 90.0% and 61.4% respectively. MDP scan although extremely sensitive (100.0%), was not specific (10.7%) in the diagnosis of OM in the complicated diabetic foot. The accuracy, PPV and NPV of bone scan were 53.7%, 51.0% and 100.0% respectively. Three phase MDP scan was not useful in differentiating OM from ACA because both conditions were characterized by equally increased blood flow (higher than STI, p<0.0001 and p<0.001 respectively). PPV and NP of the combination of the studies were also high (86.4% και 95.7% respectively). The majority of the OM cases (95%) demonstrated concordant MDP/HMPAO-LS findings, while incongruent MDP/HMPAO-LS images were present in most (94%) cases of STI. Interestingly, both congruent and incongruent MDP/HMPAO-LS images were observed in cases of ACA. There were three false positive results, 2 of them in sites of Charcot arthropathy. Lesion to contralateral foot ratio was higher in sites of OM compared with those with ACA or STI (p<0.005). Although semiquantitative analysis was not found diagnostically useful, it could be helpful in the evaluation of disease activity in repeated studies. The additional performance of BMS helped to diagnose OM superimposed on Charcot arthropathy or on site of previous amputation. The addition of BMS improved the specificity and accuracy of combined MDP/HMPAO-LS from 88.0% and 91.1% to 100.0% and 97.8% respectively. Conclusions: OM is a commonly encountered complication in non-healing diabetic foot ulcers. Clinical presentation is usually atypical and inflammatory blood marker levels are not helpful in the differential diagnosis of OM from ACA or STI in the diabetic foot. Among laboratory tests, elevated CRP seems to be better marker of OM in the diabetic foot compared with leucocytocis or elevated ESR, but still not specific. Plain radiography is the least sensitive imaging method for early diagnosis of OM in the complicated diabetic foot. Its main drawbacks are delayed appearance of the characteristic OM changes and difficulty in diagnosing OM superimposed on Charcot arthropathy. However, radiography should be the initial screening study performed on diabetic foot, because it provides anatomic informations and is useful as a baseline study. Among radionuclide imaging modalities, MDP scan alone, either single- or multiphase, is of limited value in OM diagnosis in the diabetic foot. MDP scan is useful in OM diagnosis in the forefoot in patients without preexisting bony abnormalities. It may also helpful for better localization and interpretation of HMPAO-LS findings in patients with pedal ulcers. Conversely, MDP scan is not useful for detecting OM in sites of Charcot joints and it could be omitted in patients with established Charcot arthropathy. HMPAO-LS is the most accurate radionuclide method for the diagnosis of OM complicated pedal ulcers or Charcot arthropathy in the diabetic foot. The study has higher diagnostic accuracy compared with the other imaging methods (91.1%) and a negative study excludes infection with a high degree of certainty. Moreover, it is also useful in follow-up response assessment to antibiotic treatment. From a practical point of view, both ⁹⁹ᵐTc labelling and multiple views acquisition improve lesion localization. Consequently, we suggest that HMPAO-LS should be the first radionuclide imaging performed for OM diagnosis on the diabetic foot. MDP scan could improve anatomical localization of focal HMPAO-LS findings in the forefoot. Finally, a complementary BMS is required for correct interpretation of a positive HMPAO-LS in the mid or hind foot, on sites of Charcot arthropathy or amputation.Σκοπός της διδακτορικής διατριβής είναι η αξιολόγηση της συμβολής συνδυασμού ραδιοϊσοτοπικών μεθόδων, βασιζόμενου κυρίως στο σπινθηρογράφημα με επισημασμένα in vitro με ⁹⁹ᵐTc-HMPAO αυτόλογα λευκά αιμοσφαίρια, στη διαφορική διάγνωση της οστεομυελίτιδας στο διαβητικό πόδι. Ασθενείς και μέθοδοι: Μελετήθηκαν 45 ενήλικες διαβητικοί ασθενείς με κλινική υπόνοια οστεομυελίτιδας σε 54 θέσεις των ποδών. Οι ασθενείς υποβλήθηκαν σε μέτρηση βιοχημικών δεικτών φλεγμονής, απλές ακτινογραφίες άκρων ποδών, σπινθηρογράφημα οστών (3 ή 4 φάσεων) με ⁹⁹ᵐTc-MDP (MDP), σπινθηρογράφημα με επισημασμένα με ⁹⁹ᵐTc-HMPAO αυτόλογα λευκά αιμοσφαίρια (HMPAO-LS) (36 ασθενείς) και σπινθηρογράφημα μυελού οστών με ⁹⁹ᵐTc-κολλοειδές κασσιτέρου (BΜS) (6 ασθενείς). Τα ευρήματα των ραδιονουκλιδικών εξετάσεων ερμηνεύτηκαν σύμφωνα με τα τρέχοντα κριτήρια. Η διάγνωση επιβεβαιώθηκε κλινικώς και ακτινολογικώς ή με βιοψία οστού. Σε 7 ασθενείς έγινε επανάληψη του HMPAO-LS για την αξιολόγηση της απόκρισης στην θεραπεία. Αποτελέσματα: Από τις 54 κλινικά ύποπτες θέσεις, οστεομυελίτιδα (ΟΜ) διαγνώστηκε σε 26, οξεία αρθροπάθεια Charcot (ΟΑC) σε 6, απλή φλεγμονή των μαλακών μορίων (ΦΜΜ) σε 18, κάταγμα εκ καταπονήσεως σε 2, οξεία μονοαρθρίτιδα σε 1 και διαβητική γάγγραινα σε 1 θέση. Συνυπάρχουσα ΟΜ διαγνώστηκε στο 50% των ελκών. Η CRP ήταν υψηλότερη στους ασθενείς με ΟΜ συγκριτικά με εκείνους με ΟΑC ή ΦΜΜ (24.6 έναντι 10.3 και 7.5 mg/L αντίστοιχα, p