Statistical reliability of bone biopsy for the diagnosis of diabetic foot osteomyelitis.

ABSTRACT

Bone biopsy is often referred to as the reference standard for the diagnosis of diabetic foot osteomyelitis (OM), and it also serves as an important interventional tool with respect to diabetic foot infections and limb salvage. However, the phrase bone biopsy lacks a standardized definition, and the statistical reliability of the pathologic diagnosis has not been previously examined. The objective of the present study was to quantify the reliability of the histopathologic analysis of bone with respect to the diagnosis of diabetic foot OM. Four pathologists, kept unaware of the previous pathology reports and specific patient clinical characteristics, retrospectively reviewed 39 consecutive tissue specimens and were informed only that it was "a specimen of bone taken from a diabetic foot to evaluate for OM." As a primary outcome measure, the pathologists were asked to make 1 of 3 possible diagnoses (1) no evidence of OM, (2) no definitive findings of OM, but cannot rule it out, or (3) findings consistent with OM. There was complete agreement among all 4 pathologists with respect to the primary diagnosis in 13 (33.33%) of the 39 specimens, with a corresponding kappa coefficient of 0.31. A situation of clinically significant disagreement, or in which at least 1 pathologist diagnosed "no evidence of OM," but at least 1 other pathologist diagnosed "findings consistent with OM," occurred in 16 (41.03%) of the specimens. These results indicate agreement below the level of a "reference standard" and emphasize the need for a more comprehensive diagnostic protocol for diabetic foot OM.