Introduction Foot ulcers in patients with diabetes constitute a growing and costly public health concern. Diabetes is increasing in prevalence, especially in developed nations. In the United States, the prevalence is estimated to be 7.3 percent of adults.1 Foot ulcers in these individuals arise at an incidence of 2.5 percent each year and, unfortunately, may initiate a pathway to amputation and limb loss.2 One adverse consequence of foot ulceration is infection, which becomes limb threatening when there is bone involvement. Repetitive pressures on an ulcer often are rapidly converted to tissue-disruptive inflammation close to an underlying bony prominence; the consequence is that approximately 15 percent of foot ulcers are complicated by osteomyelitis.3 The customary management of osteomyelitis from contiguous foot ulceration is thorough surgical removal of all infected bone and often involves resection of relatively noninfected adjacent soft tissue and bone. Along with concurrent use of appropriate antimicrobials, this aggressive approach may lead to successful eradication of the infection; however, it may also result in the untoward outcome of altered biomechanics, foot instability, and increased residual plantar pressures—placing the patient at greater risk of future ulcerations. Indeed, one of the risk factors for amputation is prior amputation.4 In the past decade, there has been a growing clinical interest in an approach to osteomyelitis in the diabetic foot that is more foot sparing with less aggressive surgical ablation, favoring instead more reliance on conservative surgery and long-term antimicrobials. Besides reducing the number of major amputations as first-line treatment, the advantage to such an approach would be in the maintenance of some biomechanical stability to allow easier accommodation of the foot and prevention of new ulcer events. In order to evaluate the efficacy of a conservative approach, we retrospectively evaluated a series of patients with diabetes and foot ulcers complicated by osteomyelitis that were managed by an interdisciplinary team. Patients were evaluable only if they had at least 12-months follow up for the assessment of healing and recurrence. Methods Patients with diabetes and osteomyelitis of the foot who received inpatient care at Staten Island University Hospital from 1994 to 1996 were identified in the hospital discharge database using codes from the International Classification of Diseases, Version 9 (ICD-9 codes). Using medical record numbers, 160 patients with osteomyelitis were identified as having been registered in the outpatient interdisciplinary foot clinic at the Diabetes Treatment Center. Clinic charts for these patients were reviewed, and 58 patients were found to have an adequately documented follow-up observation period of at least 12 months. A long observation period was chosen to address the concern of persistence and/or recurrence in patients with chronic osteomyelitis. These 58 patients comprised the study cohort. Both inpatient and outpatient records were analyzed for the evaluation and management of the osteomyelitis. The diagnosis of osteomyelitis was made on clinical grounds, defined by grossly infected or exposed bone; an ulcer that probed to bone;5 or a diagnostic radiograph and HMPAO-WBC (Ceretec) or similar nuclear imaging scan. All patients were evaluated surgically. The type and extent of the initial surgical procedure was determined according to clinical judgment of the treatment team, often with a conservative treatment philosophy of minimizing bone resection whenever possible. Patients were all treated with intravenous antibiotics, which were initiated empirically based on expected microbial flora and changed if the patient did not respond clinically or if the bone culture grew clearly resistant pathogens. For patients who had more than one episode or hospitalization for osteomyelitis over 12 months, the first episode during the observation period was used as the index case. After discharge from the hospital, the patients were managed by one or more members of an interdisciplinary foot clinic, which was comprised of specialists in podiatry, endocrinology, physical therapy, pedorthics, vascular surgery, and infectious diseases. As a general treatment protocol, patients received outpatient care with frequent wound inspection and debridement, home antibiotic treatments and home dressing changes by nursing agencies, and appropriate offloading of the foot to protect the involved area from recurrent injury. Intravenous antibiotics were given for an expected course of four to six weeks. Osteomyelitis was considered to have clinical cure when there was complete ulcer healing. Treatment failure was defined as persistence of the ulcer at 12 months or further surgical resection of bone after the initial hospitalization. Patients who had healed were evaluated at 12 months for recurrent ulcer at the same location or new ulcer episode at a different site on either foot. Clinical data collected included location of the osteomyelitis, type of initial surgical procedure, ultimate surgical procedure, antibiotic type and duration of treatment, time to ulcer healing, and clinical outcomes after 12 months of observation (Figure 1). Results A total of 58 discharged patients were identified as having diabetes and osteomyelitis of the foot or lower extremity and, in addition, had post-hospitalization outpatient follow up for at least 12 months. There were 10 women and 48 men, with a mean age of 60 years. Location of ulceration and osteomyelitis was toe in 20 patients (34.5%), metatarsal in 30 patients (51.7%), midfoot in 2 patients (3.4%), heel in 3 patients (5.2%), and ankle in 3 patients (5.2%) (Table 1). Only three patients (5.2%) were surgically treated with a major amputation. Thirteen patients (22.4%) had excision of the involved bone, while eight patients (13.8%) had an amputation of a toe or ray. The majority of the patients underwent simple surgical debridement of soft tissue and/or bone (34 patients, 58.6%) (Table 1). Forty-seven patients (81%) received empiric intravenous antibiotic therapy, which was maintained through the course of treatment for generally four to six weeks. The selection of antibiotics was ceftizoxime in 25 patients (43.1%), ampicillin/sulbactam in 12 patients (20.7%), cefoxitin in 6 patients (10.3%), and, in 4 patients with end-stage renal disease, vancomycin and gentamicin (6.9%). The remaining 11 patients (19%) received culture-specific individualized antibiotic regimens. The mean duration of antibiotic therapy was 40.3 days (range 19–90 days). During the follow-up period of 12 months, 10 patients (17.2%) required readmission for further treatment of their ulcer or infection. At the end of the follow-up period, 46 patients (79.3%) were healed, with a mean healing time of 15.4 weeks (range 2–32 weeks). The remaining 12 patients (20.7%) failed to heal: Three (5.2%) underwent amputation of the involved area, and nine (15.5%) had persistence of the ulceration after a year. Of the patients who healed, 3 out of 46 (6.5%) had a recurrent ulcer at the same site, and 21 out of 46 (45.7%) had a new ulcer episode at a site different from the original. Discussion The management of osteomyelitis in the patient with diabetes and foot ulceration has been increasingly modified to a more conservative foot-sparing approach. Traditionally, it has been felt that cure was only possible with thorough resection of all involved bone along with antimicrobial therapy. With the advent of newer broad-spectrum antibiotics and the understanding of the basics of diabetic foot wound care, clinicians around the world have been attempting a treatment based on local care and potent antibiotic regimens to allow conservation of bony architecture and better long-term outcomes. Bamberger, et al., in 1987 reported success in eradicating osteomyelitis in 27 out of 51 patients (53%), with failures being those with swelling, necrosis, or gangrene.6 The use of intravenous antibiotics against isolated pathogens for at least four weeks improved the likelihood of cure. Subsequently, Peterson, et al., showed that oral antibiotic therapy in chronic osteomyelitis may be useful, successfully treating 19 out of 29 patients (66%) with oral ciprofloxacin.7 Since then, other retrospective studies primarily from Europe have demonstrated reasonable efficacy to a conservative approach. Eneroth, et al., demonstrated healing deep foot infections, including osteomyelitis, without amputation in 39 percent of their large series of 223 patients. The authors emphasize aggressive initial soft tissue surgical debridement.8 Conservative therapy by another multidisciplinary diabetic foot clinic and long-term oral antibiotics resulted in resolution of bone infection in 17 out of 21 patients without bone resection.9 In another larger retrospective study, Pittet, et al., reported successful conservative treatment of diabetic osteomyelitis. In this series, 35 out of 50 (70%) patients were cured of osteomyelitis with long-term antibiotics without any significant surgical intervention.10 The analysis was made, however, after the exclusion of 14 patients who initially underwent amputation. The studies do suggest, however, that cure of osteomyelitis without significant bone resection is possible with long-term antibiotics. A more defined analysis on the value of surgical intervention was a retrospective study by Ha Van, et al., which showed that patients who had a conservative surgical procedure faired better than those who were managed only medically with antibiotics.11 Similarly, Tan and colleagues showed the value of early surgical intervention in a retrospective post-hoc analysis of pooled data from several clinical trials for antibiotic therapy of osteomyelitis. It was found that those that had prompt surgical intervention had significantly less amputations than those treated by medical therapy alone.12 Thus, there is some validation to the clinical approach to osteomyelitis in the diabetic foot that advocates long-term systemic antibiotics in concert with early conservative surgical intervention. Our findings support this conservative management, especially as an initial intervention. Only three of our series of fifty-eight patients required major amputation on their first admission. Patients collectively fared well with either debridement, excision of involved bone, or partial amputation. The outcomes were comparable in all subgroups, but the size of the subgroups was not large enough for analysis to detect significant differences. Furthermore, the choice of the initial procedure was by the treating practitioners’ clinical judgment, interjecting selection bias into the analysis. We can only conclude that achievement of nearly 80-percent cure overall makes several treatment options viable. This contrasts with a prevailing surgical view that osteomyelitis should be treated with surgical resection.13 Another striking observation of this series is the success of empiric antibiotic therapy for ulcers complicated by osteomyelitis. Historically, these infections are often polymicrobial with gram-positive (most commonly Staphylococcal), gram-negative, and anaerobic species represented.14 The treatment approach of empirically treating all three bacterial groups—unless the infection did not respond or the culture grew obvious pathogens that were resistant—was utilized in over 80 percent of the cases. This also avoids the pitfall of too narrow an antibiotic spectrum when antibiotics are targeted against a single isolate, rather than the polymicrobial flora that is historically known to be present in such complicated ulcers.15 The high rates of healing osteomyelitis could also reflect the impact of having a comprehensive center providing interdisciplinary care, which has been held up as a model for management of complex diabetic foot disorders.16 The high rate of new ulcerations, in face of a low rate of persistent or recurrent ulcers, probably reflects the morbidity of the patient population rather than the quality of care. The interpretive difficulties of this series reside largely in the limitations of retrospective analysis. In addition, many eligible patients were not captured or were lost to follow up. We only assessed patients with whom we had continued contact, selecting out some treatment failures and noncompliant individuals. This may also explain why there may have been a higher success rate than prior published series. Conclusion In conclusion, our series supports an approach to diabetic foot ulcers complicated by osteomyelitis that is based on early conservative surgery and long-term empiric antibiotic therapy. When this is complimented by good foot ulcer care and suitable offloading, satisfactory healing rates may be achieved. Although early surgical intervention seems beneficial, the extent of the necessary debridement and/or bone excision is not yet clear. A conservative, foot-sparing approach, however, appears safe in selected patients and may be an alternative to early amputation, especially as an initial intervention. A well-defined prospective study with intent-to-treat analysis is needed to guide clinicians with more certainty.