This patient with type 2 diabetes mellitus most likely has osteomyelitis of the foot and should be treated with vancomycin and piperacillin-tazobactam. Surgical debridement is also important to ensure removal of all nonviable and purulent tissue and allows for collecting more representative tissue specimens for microbiologic studies. Infections in the feet of patients with diabetes may come on suddenly, present with sepsis, and rapidly become limb-threatening. Most often these infections result from ulcerations created by prior surgical procedures, trauma, or local pressure. Diminished sensation, decreased blood flow, and poor glycemic control are frequent contributing factors. Empiric broad-spectrum antimicrobial therapy aimed at gram-positive cocci (β-hemolytic streptococci and staphylococci) along with aerobic and anaerobic gram-negative bacilli should be initiated as soon as possible in patients with suspected osteomyelitis and systemic findings. Methicillin-resistant strains of Staphylococcus aureus should be considered potential pathogens in these clinical situations and included in the antimicrobial coverage until proven absent. Vancomycin and a broad-spectrum β-lactam/β-lactamase inhibitor, such as piperacillin-tazobactam, would be an appropriate empiric regimen; it is appropriate to narrow the antibiotic spectrum once bone culture results are known. Surgical debridement is essential in the treatment of this patient's obviously gangrenous foot.

Antimicrobial therapy with the combination of cefazolin and metronidazole has too narrow a spectrum under these clinical circumstances. Coverage against a greater variety of aerobic, gram-negative bacilli is required.

Clindamycin has activity against most of the gram-positive cocci and anaerobic gram-negative bacilli associated with diabetic foot infections. However, gentamicin to provide coverage against aerobic gram-negative bacilli in complex diabetic foot infections is not recommended because of its narrow toxicity-to-benefit ratio.

Grossly visible bone or the ability to probe to bone is highly predictive of osteomyelitis, making the need for a bone biopsy and delay in treatment inadvisable.