Elaine Sunderlin
EBM 6/25/2010
“Probing to Bone in Infected Pedal Ulcers”
M. Lindsay Grayson, MD, et.al.
JAMA;Vol 273, No.9;March 1, 1995;721-723
Background:
- Pedal ulcers occur in up to 25% of diabetic patients
- Important to determine difference between soft tissue infections alone and soft tissue infections with underlying osteomyelitis
- What’s the best way to diagnose osteomyelitis?
Hypothesis: The presence of bone, via gentle probing, in the depths of infected pedal ulcers in patients with diabetes is indicative of osteomyelitis
Type of Study: Prospective cohort study of hospitalized diabetic patients receiving antibiotic treatment for severe limb-threatening foot infection.
Methods: During a 2 year period beginning in December 1998 , 92 patients with 97 infections were enrolled in a randomized, double-blinded, single-center prospective clinical trial to study the efficacy of Ampicillin/Sulbactam versus Imipenem/Cilastatin in the treatment of limb-threatening foot infections in diabetic patients. This data pool was used in this follow-up study. Patients in the original study without pedal ulceration, with nonhealed recent surgical wounds, or with pedal infection that had been debrided in a manner to likely expose the adjacent bone were excluded.
- In patients with open ulcers, probing to detect bone was performed prior to debridement
- When ulcers were covered by an eschar, probing was undertaken after debridement that was limited to removal of overlying eschar
- Probing was done using a sterile, blunt, 14cm stainless steel eye probe held like a pencil. One of the two authors assessed the ulcer at bedside
- Bone was considered palpable (positive probe test) when on gentle probing, the evaluator detected a rock-hard, often gritty structure at the ulcer base without the apparent presence of any intervening soft tissue.
- Diagnosis of osteomyelitis
o Histologic (92%): presence of inflammatory cells within the bone, fibrosis of intertravecular soft tissue, and destruction or necrosis of bone and reactive new bone formation
§ + radiographic (35%)
§ + clinical criteria (46%)
o Radiographic alone (2%): bone destruction on imaging studies
o Clinical criteria alone (4%): identification of purulent friable nonviable bone by the surgeon performing debridement
o Radiographic and clinical (2%)
Results:
Osteomyelitis present / Osteomyelitis absent / TotalsPositive test / 33 / 4 / 37
Negative test / 17 / 22 / 39
Totals / 50 / 26 / 76
Specificity = ______number of true negatives______22 = 0.66 (66%)
# of true negatives + # false positives 26
Sensitivity = ______number of true positives______33 = 0.85 (85%)
# of true positives + # false negatives 50
Positive predictive ______True positives______33__ = 0.89 (89%)
Value (PPV) True positives + false positives 33 + 4
Negative predictive ______True negatives______22__ = 0.56 (56%)
Value (NPV) True negatives + false negatives 22 + 17
Positive Likelihood ______Sensitivity______0.66__ = 4.4
Ratio 1 – specificity 1 – 0.85
Negative Likelihood ___1 – sensitivity______1 – 0.66 = 0.4
Ratio Specificity 0.85
Is the study valid? I think so. It is a cohort study with a small number of patients. As they were testing a physical exam, there is always examiner difference. They tried to limit that with only 2 observers performing the exam. The histology specimens were only examined by 1 pathologist who was blinded to the probe test results.
Would I use this in my clinical practice? Unlikely. The positive predictive value is not robust. In general, the current diagnostic tools for osteomyelitis are limited, at best.