Can Aggressive Treatment of Diabetic Foot Infections Reduce the Need for Above-Ankle Amputation?

Transcription

1 286 Can Aggressive Treatment of Diabetic Foot Infections Reduce the Need for Above-Ankle Amputation? James S. Tan, Norman M. Friedman, Carolyn Hazelton-Miller, J. Patrick Flanagan, and Thomas M. File, Jr. From Northeastern Ohio Universities College ofmedicine, Rootstown; and the Departments ofmedicine and Orthopedics, Summa Health System, Akron, Ohio We retrospectively evaluated the charts of 112 patients with diabetic foot infection to determine if early aggressive surgical intervention improves outcome. All patients were classified into two groups on the basis of the timing of surgical intervention and appropriate antimicrobial therapy. Group I included patients who underwent no surgical intervention during the first 3 days of hospitalization but received intravenous antimicrobial therapy, and group II included patients who underwent surgical intervention promptly and received intravenous antimicrobial therapy. Group II was further divided; group IIA included patients who underwent debridement, and group lib included patients who underwent local limited amputation. A higher rate of patients in group I than in group II (27.6% vs. 13%, respectively; P <.01) required above-ankle amputation during the same hospitalization or subsequent admission. Overall, an aggressive surgical approach against foot infection in hospitalized diabetic patients reduced the need for above-ankle amputation and the length of hospital stay by at least 6 days. Treatment of diabetic foot infection requires the combination of early surgical treatment and antimicrobial therapy. The most common cause of hospitalization of diabetic patients is soft-tissue and bone infections involving the foot [1]. Mild or superficial infections can be treated in the outpatient setting by oral antimicrobial agents and by avoidance offurther irritation [2]. Although many patients with severe infections are hospitalized and treated with intravenous antimicrobial therapy, the final outcome is frequently prolonged immobilization or loss of the limb. Diabetes mellitus is the leading cause of lower-extremity amputation and accounts for about one-half of all non-traumarelated amputations in the United States [3, 4]. The rate of lower-extremity amputation among diabetics is times greater than that among nondiabetics [4, 5]. In a retrospective study from Rochester, Minnesota [4],90% of alllower-extremity amputations were performed on patients with non-insulindependent diabetes. Reported risk factors for lower-extremity amputation include nonhealing ulcers, infection, advanced age, male sex, black race, and a history of smoking [6-8]. Reiber et al. [9] listed additional risk factors of amputation for patients with diabetic foot infection such as insufficient lower-extremity circulation, decreased ankle-ann blood pressure index, low levels of highdensity lipoprotein subfraction, and lack of previous education on outpatient care for diabetes. Foot infection and ulceration Received 17 October 1995; revised 26 March This work was presented in part at the 32nd annual meeting of the Infectious Diseases Society of America held in October 1994 in Orlando. Reprints or correspondence: Dr. James S. Tan, 75 Arch Street, Suite 105, Akron, Ohio Clinical Infectious Diseases 1996; 23: by The University of Chicago. All rights reserved /96/ $02.00 significantly predispose diabetic patients to future amputation [5, 7, 9-13]. Bamberger and colleagues [14] have suggested that early diagnosis and treatment of nongangrenous limb infections in diabetic patients could alter the need for amputation. Other investigators [15] have reported that early aggressive digit amputation may prevent eventual limb loss. Apelqvist et al. [16] have suggested that early aggressive intervention may reduce the need for above-ankle amputation and may also decrease the total cost. The primary objectives in the management of diabetic foot infection are to preserve and restore foot function. Many physicians who are not experienced in the care of hospitalized patients with diabetic foot infections tend to take the nonsurgical approach by prescribing antibiotic therapy alone. We believe that this approach is not cost-effective because these patients are subjected to prolonged antimicrobial therapy without removal of the infected focus and to minimal ambu1ation. Early surgical intervention can reduce the duration of antimicrobial therapy and restore full ambulation earlier. In many comparative drug studies [17-20], any amputation- regardless of the level-was considered as part of the exclusion criteria or as treatment failure. Because limited amputation may restore the ability to function, we consider debridement and limited forefoot amputation as a part of therapy and not as treatment failure. We use the endpoint of above-ankle amputation as treatment failure. Therefore, we reviewed our experience with hospitalized patients with diabetic foot infections and analyzed the influence of early aggressive surgical intervention on the final need for above-ankle amputation. Materials and Methods Patients. One hundred twelve patients (57 males and 55 females) had 164 limb infections. Limb infection was defined

2 em 1996;23 (August) Treatment of Diabetic Foot Infections 287 as infection of one lower extremity regardless oflocation. Subsequent hospitalization because of complications of the same limb was considered as part ofthe original event. These patients were admitted between 1982 and 1990 to Akron City Hospital (Akron, OH; presently part ofsumma Health System) and were participants in study protocols ofantibiotic efficacy that were conducted by the infectious disease service as part ofmulticenter studies with several pharmaceutical companies. Survey. The records of the patients were retrospectively reviewed. Information on clinical illness, microbiological data, antibiotic treatment, surgical intervention, and follow-up for subsequent hospitalizations was recorded. Clinical presentation. Only patients with serious foot infections requiring hospitalization were included in the study. Wounds were classified on the basis ofthe depth ofinfection as cellulitis and bullous lesions (dermis not breached), subcutaneous lesions (dermis breached, but bone tendons or joints not involved), deep infections without bone involvement, and deep infections with bone involvement. Osteomyelitis was diagnosed when clinical evidence was accompanied by positive results offoot radiography or positive histopathologic findings. The site of infection was defined as follows: forefoot, which included the metatarsal heads and phalanges; midfoot, which included the tarsal bones and the proximal aspect of the metatarsals; hindfoot (calcaneus); ankle; and above ankle. In this retrospective study, we found that the clinical findings at admission were not consistently and fully documented. Therefore, the classification of the depth of infection was not based on the admitting diagnosis but rather on the postsurgical or final discharge diagnosis. In all cases, the vascular perfusion was judged to be adequate on the basis of the clinical assessment of the investigator to allow appropriate evaluation of the antimicrobial efficacy. Microbiology. Specimens for bacterial cultures were obtained by local swabbing (accomplished by removing the superficial debris before obtaining a deeper sample), needle aspiration, and/or biopsy. Specimens for bacterial cultures were classified as aseptic if they were obtained by biopsy or needle aspiration through noncontaminated skin or septic if they were obtained by either swabbing or curettage through an open ulcer. Specimens were sent to the microbiology laboratory for aerobic and anaerobic cultures whenever possible. Blood for cultures was obtained only when systemic disease was suspected. Treatment. The patients were divided into two groups on the basis oftreatment. Group I included patients with 87 infections who underwent no surgical intervention within the first 3 days ofhospitalization but received intravenous antimicrobial therapy, and group II included patients with 77 infections who underwent surgical intervention within the first 3 days and received intravenous antimicrobial therapy. Group II was further divided; group IIA included patients who underwent surgical debridement, and group lib included patients who underwent local amputation within the first 3 days ofhospitalization. Because of the presence of neuropathy, many patients underwent debridement at their bedside. Antimicrobial agents. Antimicrobial therapy included agents such as ampicillin/sulbactam, amoxicillin/clavulanate, ticarcillin/clavulanate, moxalactam, and cefoxitin. When bacterial resistance was suspected, therapy was switched on the basis of the investigator's clinical judgment. Outcome study. The efficacy oftreatment was determined by the need for above-ankle amputation during the same hospital stay or a subsequent admission within 1 year. Follow-up of the outcome, including above-ankle amputations, was performed by reviewing the records of the hospital and the physicians' offices. In addition, inquiries were made through letters and telephone calls to the patients, primary care physicians, and surgeons. Defects ofthe study. This study was retrospective and had to rely on review of charts at admission, data from private offices and clinics, admission information from other hospitals, and subsequent follow-up information from the patients and their physicians. Details of initial presentation were not reliably documented; therefore, classification of the depth of infection was based on the postsurgical or final discharge diagnosis. Results The patients' ages ranged from 32 to 91 years (mean age, 58.7 years). There were 112 patients with 164 limb infections (total number of hospitalizations, 205). Table 1 shows the relationship between the site and the depth of the 164 limb infections. Eighty-eight infections presented as cellulitis or subcutaneously. The were 76 deep infections with or without osteomyelitis. The forefoot was the most common site of infection (112 [68.3%] of 164), followed by the midfoot, ankle, and hindfoot. There were 65 deep forefoot infections (10 without osteomyelitis and 55 with osteomyelitis). There were 11 deep infections in sites other than the forefoot. Of the 76 deep infections, 65 had concomitant osteomyelitis. The rate of concomitant osteomyelitis was greatest for forefoot infections (55 [49.1%] of 112). Infections in sites other than the forefoot had a lower rate (10 [19.2%] of52) ofconcomitant osteomyelitis. Table 2 shows the results of 147 bacterial cultures. Thirtyeight aseptic specimens (including 19 bone aspirates and 19 bone biopsy specimens) and 109 septic specimens (all from open ulcers) were obtained. Many patients had received either outpatient or inpatient antimicrobial therapy before the samples were taken. Cultures of the 38 aseptic specimens yielded no growth in 13 instances (6 biopsy specimens and 7 aspirates), a single isolate in 8 instances (3 biopsy specimens and 5 aspirates), and multiple organisms in 17 instances (10 biopsy specimens and 7 aspirates). Cultures of the 109 septic specimens yielded no organisms in 4 instances, single isolates in 23 instances, and multiple organisms in 82 instances. Staphylococcus epidermidis, Staphylococcus aureus, and Streptococcus were the most common single organisms isolated from bone biopsy specimens. Corynebacterium isolates were usually found as one of multiple organisms. Corynebacterium

3 288 Tan et al. em 1996;23 (August) Table 1. Sites and depth of 164 limb infections in 112 patients with diabetes. No. of infections Depth of infection Forefoot Midfoot Hindfoot Ankle Cellulitis Subcutaneous Deep without osteomyelitis Deep with osteomyelitis Total Total and Streptococcus were recovered more frequently from the wound and bone than were aerobic gram-negative rods. Nine patients had bacteremia. Cultures of wound specimens from these patients yielded bacteria with comparable susceptibilities to the antimicrobial agents tested. Coagulase-negative Staphylococcus alone was isolated from blood specimens from five patients. Coagulase-negative Staphylococcus plus Enterococcus was recovered from one patient. S. aureus, Streptococcus agalactiae, and Proteus mirabilis were each isolated from one of the remaining three patients. The effects of initial conservative therapy vs. aggressive treatment on the subsequent outcome are shown in table 3. All patients received therapy with antimicrobial agents considered appropriate for the pathogens isolated. There was no correlation between the microorganism isolated, the susceptibility pattern, the antimicrobial agent used, and the outcome. The lowest rate of clinically diagnosed deep infections was among group I Table 2. Organisms isolated from wound specimens and bone biopsy or aspiration specimens from patients with diabetic foot infections. Microorganism(s) Staphylococcus aureus Coagulase-negative Staphylococcus Streptococci Streptococcus pyogenes Streptococcus agalactiae Group G streptococcus Viridans streptococcus Enterococcus species Corynebacterium species Pseudomonas aeruginosa Escherichia coli Proteus mirabilis Enterobacter cloacae Other aerobic gram-negative rods Anaerobes Other Total Wound specimens (n = 109) Bone biopsy or aspiration specimens (n = 38)* o o * Nineteen biopsy specimens and 19 aspirates. Thirteen cultures yielded no growth. patients (29.9% vs. 64.9% among group II patients), but the rate of above-ankle amputations was significantly higher among group I patients than among group II patients (27.6% vs. 13.0%, respectively). Thirteen group I patients underwent amputation during the same admission, and 11 had to be readmitted for amputation. Even though group II had a higher number of patients with osteomyelitis, only five underwent above-ankle amputation during the same hospital stay, and five were readmitted for amputation. When group II was subdivided into patients who underwent debridement and those who underwent localized amputation, the rates of above-ankle amputation were still lower among these patients than among group I patients (no statistical significance). When the frequency of above-ankle amputation for patients with deep infections was compared, the rate among group I patients was 92.3% (24 of 26), whereas the rate among group II patients was 20% (10 of 50). The amputation rates among group IIA and group lib patients were 16.7% (4 of 24) and 23.1% (6 of26), respectively. When group I was compared with either group IIA or lib, the difference was statistically significant (P <.001). More group I patients (11 of 87) than group II patients (5 of 77) had to be readmitted for amputation. The effect of aggressive therapy on the length of hospital stay is shown in table 4. Patients who underwent surgical intervention were compared with those who did not undergo surgical intervention on the same hospital day. The substantial reduction of hospital stay can be appreciated in both groups; those patients who underwent local amputation had a shorter hospital stay, and the reduction of the length of stay was '"'-'9 days. Regardless of the length of hospitalization before surgery, the length of stay after surgical intervention for group I and group II patients averaged 6 days. Discussion Similar to other investigators [14, 21], we found that the forefoot was the most common site of infection in diabetic patients. The forefoot has a predilection for infection because diabetic neuropathy alters the structure of the foot, thus causing the formation ofnew pressure points on the toes and metatarsal heads that are especially susceptible to mechanical trauma. Neuropathy, not vascular insufficiency, is the major predisposing factor of foot infection. Patients with sensory neuropathy

4 em 1996;23 (August) Treatment of Diabetic Foot Infections 289 Table 3. Comparison of initial conservative treatmentwith aggressive treatment of patients with 164 diabetic foot infections. No. (%) of patients Group I Group II Group IIA Group IIB Variable (n = 87) (n = 77) (n = 46) (n = 31) Cellulitis 25 (28.7) 6 (7.8) 6 (13.0) 0 Subcutaneous infection 36 (41.4) 21 (27.3) 16 (34.8) 5 (16.1) Deep infection without osteomyelitis 5 (5.7) 6 (7.8) I (2.2) 5 (16.1) Deep infection with osteomyelitis 21 (24.1) 44(57.1) 23 (50.0) 21 (67.7) Total deep infections 26 (29.9) 50 (64.9) 24 (52.2) 26 (83.9) Above-ankle amputations 24 (92.3) 10 (20.0)* 4(16.7)* 6(23.1)* Deaths I Total above-ankle amputations 24 (27.6) 10 (13.0)t 4 (8.7)t 6 (19.4)t Same admission 13 (54.2) 5 (50.0) 2 (50.0) 3 (50.0) Subsequent admission II (45.8) 5 (50.0) 2 (50.0) 3 (50.0) Above-ankle amputation and death due to uncontrollable infection 25 (28.7) 10 (13.0)t 4 (8.7)t 6 (19.4)t NOTE. Group I = patients who received antibiotic therapy alone with no surgery within the first 3 days of hospitalization; group II = patients who received antibiotic therapy and underwent surgery within the first 3 days of hospitalization; group IIA = patients who received antibiotic therapy and underwent debridement; group IIB = patients who received antibiotic therapy and underwent localized amputation. * p <.001, X 2 analysis (comparison with group I). t P <.01, X 2 analysis (comparison with group I); no statistical significance (P >.05), X 2 analysis (comparison of groups IIA and IIB). t P >.05, X 2 analysis (comparison with group I). are not aware of repeated irritation and trauma to the same area. Blister formation is followed by ulceration. Classically, a neuropathy ulcer, or mal perforans, is a chronic ulcer with a surrounding callus. Callus formation is a sign ofadequate blood supply. Other contributory factors for foot infection include autonomic neuropathy, motor neuropathy, and vascular insufficiency. Autonomic neuropathy results in decreased sweating and, hence, dry and easily traumatized skin. Motor neuropathy-as mentioned above-remolds the architecture of the foot, thus increasing pressure points in certain areas (particularly the sole under the metatarsal heads and tips as well as the dorsal aspects ofthe joints ofthe toes). These areas become especially vulnerable to trauma and subsequent infection. Vascular insufficiency plays a lesser role. Decreased blood supply may be due either to large-vessel disease or, more com- Table4. The effect of aggressive surgicalintervention on the length of hospitalization of patients with diabetic foot infections. Time of initial surgical intervention >3 hospital days (group 1)* ~3 hospital days (group II)* > 2 hospital days ~2 hospital days * See table 3 for explanation of groups. Mean length of hospital stay (d) Localized amputation (group IIB)* Debridement (group IIA)* monly, to microangiopathy. Reduced blood flow can lead to poor tissue perfusion and repair. Autonomic neuropathy increases arteriovenous shunting that increases larger-vessel blood flow but reduces blood flow to capillary beds [22, 23]. As in other studies [14, 23, 24], osteomyelitis, especially in the forefoot, occurred in more than one-half of our patients. The predilection for osteomyelitis at this site may be attributed to the small amount ofintervening soft tissue between the bone and the skin surface and the higher frequency of forefoot injury. Lipsky and co-workers [2] have shown that S. aureus and Streptococcus are the most common pathogens of mild foot infections in diabetic patients. Deep infections are frequently polymicrobial [18,22,25-30]. Since contaminants from cultures of open-wound specimens cannot be easily excluded as possible pathogens and since most studies have shown that the frequency of isolation of mixed aerobic and anaerobic bacteria is higher, empirical use of therapy with antibiotics active against aerobic and anaerobic bacteria for patients with an infected ulcer or deep foot infection is suggested [2, 22, 26, 27]. In nine of our cases where blood cultures were positive, organisms with comparable susceptibility patterns were obtained from wound cultures as well. In six of these cases, coagulase-negative Staphylococcus was isolated from blood, wound, and bone aspiration or biopsy specimens. Although no typing of the isolated organisms was done, we believe that recovery of coagulase-negative staphylococci requires more serious consideration of antibiotic therapy despite the rarity of bacteremia [27] and that obtaining blood for cultures may be justified.

5 290 Tan et al. em 1996;23 (August) Sapico et al. [31] reported a good correlation between blood cultures and wound cultures when aerobes were isolated. In five cases in which blood cultures were positive for anaerobic organisms, only cultures of two specimens from wound sites yielded similar anaerobic organisms, thus indicating that anaerobes can be missed in wound cultures even in competenthands. Our study shows that compared with broad-spectrum antibiotic therapy alone, early surgical debridement or limited amputation with antibiotic therapy results in a shorter length of hospital stay and reduction in the need for subsequent aboveankle amputations. In many comparative drug studies [17 20], amputation, regardless of the level, was considered as a treatment failure. Therefore, in many publishedstudies, patients who underwent any form ofamputation during the comparative study were excluded from the analysis or their treatment was indicated as a failure. Our study considers debridement and limited amputation of the forefoot as a form of surgical intervention and not as a treatment failure. Using the above-ankle amputation as a negative outcome, we have shown that early aggressive surgical intervention reduces significantly the number of cases with this negative outcome. In this study, we found that the highest rate of subsequent aboveankle amputation was among group I patients, despite the lowest rate of deep infections in these patients. The ratio of cases of deep infection to the total number of infections was 26:87 (29.9%) in group I and 50:77 (64.9%) in group II. The ratios in groups IIA and lib were 24:46 (52.2%) and 26:31 (83.9%), respectively. The rates of above-ankle amputation among patients with deep infections were 92.3% (24 of 26) in group I and 20% (10 of 50) in group II (P <.001); the rates were 16.7% (4 of 24) and 23.1% (6 of 26) in groups IIA and lib, respectively. The higher rate of adverse outcome in group I reinforces the need for aggressive surgical therapy. Our study shows that since some patients' presentations may appear less serious or more superficial, the true extent of the disease can be missed without aggressive management. Because our study is retrospective, the initial clinical findings were not consistently documented. We suspect that deep infection was present in some group I patients who were treated conservatively, which may explain the high rate of aboveankle amputation among those patients with deep infection. We believe that when a patient is sick enough to be hospitalized, every effort should be made to rule out a serious, deep infection or undrained necrotic material. Aggressive surgical therapy may indeed provide the best chance of saving the limb and be the most cost-effective [16, 21, 29, 32]. Both debridement and early limited amputation within the first 3 days of hospitalization markedly reduced the length of hospital stay (between 4 and 8 days) and prevented further amputation. Patients who did not undergo aggressive surgical intervention within 3 days had a mean length of stay that was > 18 days (table 4). Assuming that the severity of infection was comparable, a reduction in hospitalization of,..., 7 days could be achieved (length of stay for group II, 11 days) when the treatment was aggressively pursued. In a Swedish study [16], a retrospective economic analysis was performed on the cost of care for 274 patients who did or did not undergo amputation. The investigatorsconcluded that estimating the cost of foot ulcer care should not be limited to only shortterm care but that the long-term cost can be high because of the need for home care and future recurrent foot problems. Although this Swedish study showed an increase in short-term expense when surgery was performed, short-term savings were realized in our study by shortening the days of hospitalization, and the cost within 6 months was also reduced because of higher rates of above-ankle amputation and readmissions. Examining only the short-term gain (reducing the hospital stay by at least 6 days), assuming a per diem cost to the hospital of between $600 and $1,000, a cost saving of at least $3,600 to $6,000 can be realized. In addition, the reduction in the rate of poor outcome from 28.7% in group I to 13% in group II should also be included in the calculation of cost savings. In an editorial on the Swedish study, Lutter [33] stated, "You will find that monies applied early in the treatment process of diabetes mellitus are well spent, since the costs for an individual without an amputation are approximately $16,000 for the study period oftime and soar to $63,000 ifan above-ankle amputation is performed." In a recent report by Eckman et al. [34], decision and costeffectiveness analyses ofthe treatment ofdiabetic foot infection were performed. These researchers recommended the empirical use of antibiotic therapy as the most cost-effective treatment. We agree with this recommendation but also believe that aggressive therapy for foot infection in every hospitalizeddiabetic patient should be pursued. This approach will reduce the length of hospitalization and rehospitalization, the need for aboveankle amputation, and the cost. Patients will be able to ambulate earlier and hopefully return to work sooner. This approach is possible only through the primary care physician's awareness that early aggressive surgical intervention means a better outcome and that a surgeon must be available for prompt and timely treatment. Conclusions and Recommendations The increasing longevity and expanding population of diabetic patients have resulted in a greater number of diabetic foot infections that continue to be the major cause of hospital admissions among these patients [29]. Prevention of these infections by patient education on common sense foot care [25, 29] and scrupulous attention to control of glucose levels [12, 25, 35] are the best ways to reduce morbidity. When the foot infection is serious enough to require hospitalization, our data suggest that a treatment regimen that combines early aggressive surgical intervention and broad-spectrum antibiotic therapy dramatically reduces the length of hospital stay and improves the outcome of the infection even with concomitant osteomyelitis. Although our data are convincing, they are based on a retrospective study. We plan to confirm these observations by a future prospective study.

6 em 1996;23 (August) Treatment of Diabetic Foot Infections 291 References 1. Gibbons GW. The diabetic foot: amputations and drainage of infection. J Vase Surg 1987;5: Lipsky BA, Pecoraro RE, Larson SA, Hanley ME, Ahroni JH. Outpatient management of uncomplicated lower-extremity infections in diabetic patients. Arch Intern Med 1990; 150: Most RS, Sinnock P. The epidemiology of lower extremity amputations in diabetic individuals. Diabetes Care 1983;6: Centers for Disease Control. Lower extremity amputations among persons with diabetes mellitus-washington, MMWR Morb Mortal Wkly Rep 1991;40: Humphrey LL, Palumbo PJ, Butters MA, et al. The contribution of noninsulin-dependent diabetes to lower-extremity amputation in the community. Arch Intern Med 1994; 154: Ogbuawa 0, Williams JT, Henry WL Jr. Diabetic gangrene in black patients. South Med J 1982;75:285~8. 7. Pecoraro RE, Reiber GE, Burgess EM. Pathways to diabetic limb amputation. Basis for prevention. Diabetes Care 1990; 13: Liedberg E, Persson BM. Age, diabetes and smoking in lower limb amputation for arterial occlusive disease. Acta Orthop Scand 1983;54:383~8. 9. Reiber GE, Pecoraro RE, Koepsell TD. Risk factors for amputation in patients with diabetes mellitus: a case-control study. Ann Intern Med 1992; 117: Newman LG, Stagnaro-Green A. Risk factors for amputation in diabetics [letter]. Ann Intern Med 1993; 118: Bi1d DE, Selby N, Sinnock P, Browner WS, Braveman P, Showstack JA. Lower-extremity amputation in people with diabetes: epidemiology and prevention. Diabetes Care 1989; 12:24~ Moss SE, Klein R, Klein BEK. The prevalence and incidence of lower extremity amputation in a diabetic population. Arch Intern Med 1992; 152: Apelqvist J, Larsson J, Agardh C-D. Long-term prognosis for diabetic patients with foot ulcers. J Intern Med 1993;233: Bamberger DM, Daus GP, Gerding DN. Osteomyelitis in the feet of diabetic patients: long-term results, prognostic factors, and the role of antimicrobial and surgical therapy. Am J Med 1987; 83: Benton GS, Kerstein MD. Cost-effectiveness of early digit amputation in the patient with diabetes. Surg Gynecol Obstet 1985; 161: Apelqvist J, Ragnarson-Tennvall G, Larsson J, Persson U. Long-term costs for foot ulcers in diabetic patients in a multidisciplinary setting. Foot Ankle Int 1995; 16: Tan JS, Wishnow RM, Talan DA, Duncanson FP, Norden CW, the Piperacillin/Tazobactam Skin and Skin Structure Study Group. Treatment of hospitalized patients with complicated skin and skin structure infections: double-blind, randomized, multicenter study of piperacillin-tazobactam versus ticarcillin-clavulanate. Antimicrob Agents Chemother 1993; 37: Grayson ML, Gibbons GW, Habershaw GM, et al. Use of ampicillin! sulbactam versus imipenernlcilastatin in the treatment of limb-threatening foot infections in diabetic patients. Clin Infect Dis 1994; 18: File TM Jr, Tan JS. Ampicillin/sulbactam vs cefoxitin for skin infections or osteomyelitis including diabetic foot infections [abstract no 657]. In: Proceedings of the Seventh Mediterranean Congress of Chemotherapy (Barcelona, Spain). Barcelona, Spain: Mediterranean Society ofchemotherapy, 1990: File TM Jr, Tan JS. Ticarcillin-clavulanate therapy for bacterial skin and soft tissue infections. Rev Infect Dis 1991; 13(suppl 9):S Lipsky BA, Pecoraro RE, Wheat LJ. The diabetic foot: soft tissue and bone infection. Infect Dis Clin North Am 1990;4: Mills JL, Beckett WC, Taylor SM. The diabetic foot: consequences of delayed treatment and referral. South Med J 1991; 84: Boulton AJM. The diabetic foot. Med Clin North Am 1988; 72: Calhoun J, Mader 1. Osteomyelitis of the diabetic foot. In: Frykberg RG, ed. The high risk foot in diabetes mellitus. New York: Churchill Livingstone, 1991: Murphy DP, Tan JS, File TM Jr. Infectious complications in diabetic patients. Prim Care 1981;8:695~ Sapico FL, Witte JL, Canawati HN, Montgomerie JZ, Bessman AN. The infected foot of the diabetic patient: quantitative microbiology and analysis of clinical features. Rev Infect Dis 1984; 6(suppl 1):S Wheat LJ, Allen SD, Henry M, et al. Diabetic foot infections: bacteriologic analysis. Arch Intern Med 1986; 146: Brodsky JW, Schneidler C. Diabetic foot infections. Orthop Clin North Am 1991;22: Calhoun JH, Cantrell J, Cobos J, et al. Treatment of diabetic foot infections: Wagner classification, therapy and outcome. Foot Ankle Int 1988; 9: Elkeles RS, Wolfe lhn. The diabetic foot. 8MJ 1991; 303: Sapieo FL, Bessman AN, Canawati HN. Bacteremia in diabetic patients with infected lower extremities. Diabetes Care 1982;5: Taylor LM Jr, Porter JM. The clinical course of diabetics who require emergent foot surgery because of infection or ischemia. 1 Vase Surg 1987; 6: Lutter LD. Sooner beats the cost of later [editorial]. Foot Ankle Int 1995; 16: Eckman MR, Greenfield S, Mackey WC, et al. Foot infections in diabetic patients: decision and cost-effectiveness analyses. lama 1995;273: Scher KS, Steele F1. The septic foot in patients with diabetes. Surgery 1988; 104:661-6.