Reply:
We appreciate the comments of Dr. Albert Lowenfels regarding our possible selection bias in only reporting on patients who were hospitalized for osteomyelitis. Dr. Lowenfels contends that we may have missed a group of patients who were successfully treated as outpatients and were never hospitalized.
First, we acknowledge that the index study population was identified using ICD-9 CM coding for osteomyelitis of the digit and forefoot. We did try to eliminate as much bias as possible by selecting all hospitalized patients, whether they ended up being admitted to surgery or a nonsurgical service. We identified 61% of these patients had been treated for 6–12 weeks in the outpatient setting prior to admission with antibiotics, prior to hospitalization (and thus failed). Therefore, the population we studied probably reflects more severe osteomyelitis, of which 75% had tissue ulcer or digit gangrene. Patients with cellulitis and no associated tissue lesions may certainly be treated as outpatients, and these patients were 2.8-fold more likely to go on to wound healing and had no increased risk of limb loss.
Secondly, as antibiotic usage prior to admission was independently associated with decreased wound healing, this suggests that true osteomyelitis (eg, where ulcer exists or bone can be probed) should be aggressively treated and not be treated as an outpatient. The best way to resolve any of these issues is, as we stated in the discussion, a prospective randomized trial comparing those patients with osteomyelitis (and not just cellulitis) for early aggressive surgical and antibiotic therapy compared with antibiotic therapy alone.
Lastly, there is scant, if any, literature on patients with confirmed osteomyelitis who have been treated successfully as outpatients.
Peter Henker, MD
University of Michigan
Section of Vascular Surgery, Michigan