Safety of vein bypass grafting to the pedal artery in diabetic patients with foot infections

Transcription

1 Safety of vein bypass grafting to the pedal artery in diabetic patients with foot infections dorsal Gary A. Tannenbaum, MD, Frank B. Pomposelli, Jr., MD, Edward J. Marcaccio, MD, Gary W. Gibbons, MD, David R. Campbell, MD, Dorothy V. Freeman, MD, Arnold Miller, MD, and Frank W. LoGerfo, MD, Boston, Mass. The results of 56 vein bypasses to the dorsal pedal artery performed in 53 diabetic patients who were admitted with ischemic foot lesions complicated by infection were reviewed. All patients had one or more of the following: infected ulcers (73%), cellulids (45%), osteomyelitis (29%), gangrene (20%), or abscess (2%). Organisms were cultured from 84% of patients (average 2.6, range i to 9 organisms per infection). Elevated temperature (> 37.7 C) or leukocytosis (> 9.0 x 103/ml) were seen in 13% and 50% of patients, respectively. All patients were treated with broad-spectrtlm antibiotics, local debridement, wound care, and bed rest. Operative debridement or open partial forefoot amputation were required to control sepsis in 11 patients (20%). Treatment of infection delayed revascularization by an average of 10.7 days. All patients underwent autogenous vein bypasses to the dorsal pedal artery. Two grafts failed within 30 days (3.6%), and one patient died (1.8%). Wound infections developed in seven patients (12.5). One wound infection resulted in graft disruption and patient death at 2 months. Average length of stay of the initial hospitalization was 29.8 days. Fifty-two patients were discharged with patent grafts and salvaged limbs; however, 31 subsequent foot procedures and 35 rehospitalizadons were required to ultimately achieve foot healing. Actuarial graft patency and limb salvage were 92% and 98%, respectively at 36 months. Pedal bypass to the ischemic infected foot is efficacious and safe as long as infection is adequately controlled first. The complexity of these situations often requires multiple surgical procedures and extensive wound care, resulting in prolonged or multiple hospitalizations. (J VAsc SURG 1992;15: ) The refinements in technique and improved results with distal bypass have led vascular surgeons to extend arterial reconstruction for limb salvage progressively more distally in the lower extremity. 13 The influence on our own practice, mostly made up of diabetic patients, has been to attempt to bypass to the best vessel in continuity with the foot in preference to isolated popliteal or tibial segments whenever possible. 4 Because of the frequent occurrence of extensive From the Division of Vascular Surgery, Harvard-Deaconess Surgical Service, New England Deaconess Hospital and Harvard Medical School, Boston. Presented at the Eighteenth Annual Meeting of the New England Society for Vascular Surgery, Quebec, Canada, Sept , Reprint requests: Frank B. Pomposelli, Jr., MD, Department of Surgery, New England Deaconess Hospital, 110 Francis St., Suite 9C, Boston, MA /6/36096 infrapopliteal occlusive disease in diabetic patients, s the best vessel in continuity with foot as determined by intraarterial digital subtraction angiogram is the dorsal pedal artery itself in 15% to 20% of our patients. 6 We have been performing vein bypass grafts to the dorsal pedal artery routinely in this situation for the last 6 years and have found it to be effective in healing ischemic foot lesions with excellent durability at our most recent follow-up of 36 months. 7 In spite of these results, we have been concerned about use of this procedure in those patients admitted with active foot infection in addition to ischemia. Although it would seem advantageous to maximize arterial flow by bypassing all occlusive lesions in these patients who frequently have significant tissue loss, it is tmclear flit is safe to place the distal anastomosis in such close proximity to infection. The purposes of this study are to review our experience with pedal 982

2 Volume 15 Number 6 June 1992 Dorsal pedal bypass in diabetics with foot infoctions 983 bypass in the subset of patients admitted with active foot infection and to determine if the presence of foot infection adversely affects outcome or results in higher morbidity and mortality rates. PATIENTS IN THE STUDY AND METHODS During the period March 1986 to August 1989, 56 dorsal pedal bypasses were performed in 53 diabetic patients who were admitted with ischemic foot lesions complicated by infection. These operations comprise 36% (56 of 156) of all dorsal pedal bypasses performed in patients during this time period. 7 The hospital records and outpatient charts of all patients were reviewed, with particular emphasis ~laced on determining the severity of foot infection on admission. Patients were retrospectively classified (by two of the authors, G.A.T. and F.B.P.) into three groups on the basis of the severity of infection, according to previously established criteria. 8 Patients in group 1 had shallow ulcers involving only skin or subcutaneous fat with localized cellulitis or purulence and no involvement of deep structures or systemic signs of infection. Group 2 comprised infections with more extensive cellulitis or with involvement of adjacent bone or tendon, but still localited and without extensive necrosis. Patients in group 3 had severe overtly limb-threatening infections with extensive ceuulitis, gangrene, osteomyelitis, and/or deep space abscess. Systemic signs of infection were often present. The clinical characteristics of patients in the three groups are summarized in Table I. All patients were admitted to the vascular surgery unit of the New England Deaconess Hospital where a standard management protocol was followed. After a history was taken and physical examination was done, aerobic and anaerobic cultures were obtained with use of standard cotton swab culturettes. Febrile or septic appearing patients also had blood cultures drawn. Plain-film foot radiographs were obtained to detect the presence of foreign bodies, soft tissue gas, neuropathic foot deformities or fracture, and bone changes suggestive of osteomyelitis. Patients with undrained pus or extensive necrosis underwent incision and drainage, debridement or open amputation as needed. All amputations, extensive debridements, or procedures in sensate patients were performed in the operating room. Densely neuropathic patients with limited necrosis or abscess underwent incision and drainage or debridement at the bedside. Bedside procedures were performed with sterile gloves and disposable instruments on a sterile drape. Incisions were placed in such a manner as to promote dependent drainage and not compromise possible future partial forefoot amputations. Wounds were packed with gauze impregnated with dilute one quarter strength povidone-iodine (Betadine) solution or isotonic saline. Medical consultation was routinely obtained. Patients were placed on strict bedrest with elevation of the foot and leg initially and then no weight placed on the involved extremity until healed. Dressings were changed two or three times daily by a specially trained nursing staffwho also educated the patients about diabetic foot disease. All wounds were examined daily by an attending surgeon and a member of the house staff. Follow-up bedside sharp debridement was performed as needed to remove residual infected or necrotic tissue. Bedside debridement was performed in 25% of patients. Eleven extremities required formal surgical procedures (Table II). Bacterial cultures were obtained from all patients. Table III lists the bacterial isolates retrieved in the study groups. Multiple organism infections were common (mean, 2.2 isolates per case) with grampositive infections most prevalent. No organisms were retrieved in nine patients, despite clinical assessment consistent with acute or chronic infection. The initial antibiotic regimen was influenced by the severity of the infection and other factors, including preexisting renal or hepatic disease, allergies, and prior antibiotic therapy. In general, broadspectrum agents active against gram-positive, gramnegative, and anaerobic organisms were used initiauy, with modification to more specific agents when initial culture and sensitivity results became available. Infection was considered adequately controlled when cellulitis, lymphangitis, and edema had resolved and wounds were free of gross purulence or "wet" gangrene. Systemic signs of improvement included resolution of fever, leukocytosis, and return of glycemic control, although the presence of these findings (with the exception of hyperglycemia) varied greatly in patients in this study. After local control of infection had been achieved, all patients underwent intraarterial digital subtraction arteriography with visualization of the pedal vasculature. Bypass was delayed until there was complete resolution of any cellulitis or lymphangitis involving areas of proposed incisions to expose the dorsal pedal artery or saphenous vein. Once this had occurred, dorsal pedal bypass was performed, as previously described. 6 No special technical maneuvers were used other than isolating the surgical field from open wounds with sterile plastic adhesive barrie r drapes. 6

3 984 Tannenbaum et al. Journal of VASCULAR SURGERY Table I. Clinical characteristics and distribution of foot lesions Group I Group U Group 111 Total No. patients No. DP bypasses Male: female 6 : 6 16 : : 3 34 : 19 Diabetes mellitus (I00%) Admission temperature ( C) > 37.7 C (33%) 5 (9%) Admission mean WBC ( x 103/ml) 4 (33%) 11 (38%) 12 (80%) 27 (48%) Admission blood sugar (mg/dl) Foot lesions* Ulceration 11 (92%) 23 (79%) 7 (47%) 41 (73%) CeUulitis 4 (33%) 11 (40%) 10 (67%) 25 (45%) Osteomyelitis 1 (8%) 10 (34%) 5 (33%) 16 (29%) Gangrene 0 8 (28%) 3 (20%) 11 (20%) Abscess (7%) 1 (2%) Lymphangitis 0 1 (3%) 3 (20%) 4 (7%) Sepsis (7%) 1 (2%) DP, Dorsal pedal; WBC, white blood count. *Some patients had more than one finding present. Table II. Procedures required to control infection before dorsal pedal bypass Group I Group 17 Group 1ZI Total Procedure (N = 12) (N = 29) (N = 15) (N = 56) Bedside debridement* (25%) Operative debridementt (7%) Open toe amputationt (11%) Open transmetatarsal amputationt (2%) *All bedside procedures performed without anesthesia in neuropathic patients and were limited to the removal or drainage of small amounts of infected or necrotic material. Gloves and sterile instruments were used. tall performed under anesthesia in the operating room. Table III. Bacterial isolates Group I Group 17 Group i~ (N = 12) (N = 29) (N = 12) Gram-positive aerobes Gram-negative aerobes Pseudomonas Others Anaerobes/fungi Negative cultures* Average isolates/cases with positive cultures *All had local signs of active infection including purulence, ceuulitis, or lymphangitis.,all had been treated with oral or intravenous antibiotics before this admission. The time from admission to bypass operation averaged 10.7 days (range, 1 to 35 days) for the entire study group. Operative delay varied with the severity of~infection: means of 8.2, 10.2, and 13.7 days from groups 1, 2, and 3, respectively~, After bypass surgery, ~wound care was Continued to allow granulation of open wounds. Bedside debridement was performed as needed, on the basis of the appearance of the wound. Concurrent examination by the staff of the podiatry department was obtained when reconstructive procedures were necessary to achieve a functional foot or to correct underlying neuropathic foot deformities that may have caused ulceration. Additional operative procedures, including debridement, skin grafts, and minor (toe, ray, or transmetatarsal) amputation were performed after sufficient healing had occurred. Primary wound closure was strived for in all revascularized limbs requiring additional operative procedures. Status of the foot lesion on discharge from the hospital determined posthospitalization care. In general, patients were continued on parenteral antibiotics until signs of healing were dearly present. Those patients discharged from the hospital with partially healed wounds were continued on oral agents until healing was complete. Wound care after discharge was provided by visiting nurses or family members.

4 Volume 15 Number 6 June 1992 Dorsal pedal bypass m diabetics with foot infections 985 Table IV. Cases requiring bypass procedures after dorsal pedal bypass for foot healing ~ Group I Group II Group 12I Total Procedure (N = 12) (N = 29) (N = 15) (N = 56) None 7 (58%) 12 (41%) 1 (6%) 20 (36%) Debridement Minor amputation Toe Ray TMA Skin graft Podiatric proceduret Total no. of procedures *Some patients had more than one procedure. texcision of a metatarsal-phalangeal joint (1) and an Achilles tendon lengthening (1). Table V. Readmission procedures Group I Group II Group 1ZI (N = 12) (N = 28*) (N = 15) Total No. cases requiring readmission 4 (33%) 9 (31%) 7 (58%) 20 (36%) No. readmissions No. cases requiring procedure Readmission procedures Debridement Toe amputation TMA BKA Skin graft Podiatric procedure** Total no. procedures TMA, Transmetatarsal amputation; BKA, below-knee amputation. *One patient died before discharge from hospital. **Exostectomy (1), transmetatarsal amputation revision (1). All patients discharged with patent grafts were monitored frequently (approximately once weekly) until full healing occurred. Additional visits every 2 to 3 months continued for the first year. Patients were seen at 6-month intervals in subsequent years. Graft patency was determined by the presence of a palpable pulse in the graft and on the dorsum of the foot. Follow-up was obtained from outpatient records and telephone interviews. Average length of follow-up is months (median, 26 months), with a maximum of 52 months. Ten patients were lost to follow-up, seven during the first year after bypass. Patient survival, primary graft patency, and limb salvage rates were calculated by the use of the actuarial life-table method. 9 RESULTS There was one postoperative death (1.8%). Thirty-day patency was 96.4%: two grafts occluded within 3 days of bypass. Neither resulted in major amputation. Early wound infections developed in seven cases (12.5%), including one infected wound hematoma in the mid tibial region that resulted in a graft disruption and patient death at 2 months. Postbypass foot procedures were performe d in 36 cases (64%) and were required more frequently in those patients with worse infection (42%, 62%, and 93% of the cases in groups I, II, III, respectively, Table IV). Primary closure was accomplished in 20 of 26 minor amputations (77%). All survivors were discharged with a healed or healing foot. Total hospitalization time averaged 29.8 days (range, 10 to 94; median, 25 days). Full healing was achieved in 26 (47%), limbs and partial healing was achieved in the remaining 29 (53%). Two from the latter group never healed completely: one patient with an early graft failure died 2 months after operation (different from the patient with a disrupted graft), and one patient required below-knee amputation I year after

5 986 Tannenbaum et al. Journal of VASG~ULAR SURGERY Full Healing n=26 Partial Healing n=29 Not Readmitte n=22 (85%) Readmitted n=l 6 (55%) Readmitted n=4 (15%) Readmitted n=l 3 (45%) Fig. 1. Breakdown of the need for readmission on the basis of the status of the foot at the time of discharge, after the first hospitalization when dorsal pedal bypass was performed. Only 15% of patients with fully healed limbs required admission ~... IT... ~... IT... IT... --[... Z... Z... _T J. J. ' " ' ± i I l I _ Primary Patency... Limb Salvage 20 I I I I I I I I f I I I Months Fig. 2. Cumulative primary graft patency and limb salvage in 56 vein grafts to the dorsal pedal artery. bypass. Final healing in the remainder occurred an average of 5.5 months after operation (range, 1 to 20; median, 4 months). Five grafts thrombosed over the course of the study. Two failed within the first 30 days, and one failed at 3, 20, and 37 months, respectively. Two of these were salvaged with revision procedures (one vein patch angioplasty and one interposition vein graft, both for midgraft stenoses). Two of the three limbs with thrombosed, unsal-

6 Volume 15 Number 6 June 1992 Dorsal pedal bypass in diabetics with foot infections O T T T T T T T T 80 m > > O9 60 >. m 40 E " I I I I I I I I I I I I Months Fig. 3. Cumulative patient survival in 53 patients undergoing 56 dorsal pedal artery bypass grafts. vaged grafts required below-knee amputation. These were the only major amputations throughout the study period. Thirty-five readmissions for recurrent or new ipsilateral foot or wound infections were required in 20 patients (35%) during the follow-up period. Thirty-one surgical procedures were necessary in 15 cases (27%), a breakdown by group is listed in Table V. Fig. 1 demonstrates the difference in frequency of subsequent readmissions in patients discharged with fully versus partially healed limbs. Fifteen percent of fully healed limbs were readmitted during follow-up, whereas 55% of partially healed limbs required one or more readmissions. Primary graft patency, limb salvage, and patient survival were 91.8%, 97.8%, and 83.8%, respectively, at 36 months (Figs. 2 and 3). DISCUSSION This report confirms the observation that diabetic foot infections are often multimicrobial, deeply invasive, and frequently require aggressive measures, including debridement and drainage procedures or partial open forefoot amputation, in addition to administration of broad-spectrum antibiotics and non weight bearing to bring them under control, s,1 12 In ischemic patients, the need to control sepsis delays revascularization in a situation where tissue loss has already occurred or an open wound is present. In this study the average duration between admission and bypass was 10 days and rose in direct relation to the severity of infection. This delay, which was required to stop spreading infection, proved successful, since no patient had further progression of necrosis to the point of loss of the foot before bypass was performed. No patient lost his/her limb with a patent graft in this report, whereas two of three patients with thrombosed grafts ultimately lost their foot. These findings underscore the importance of prompt treatment of underlying ischemia, once infection has been adequately drained. Restoration of arterial circulation is necessary for healing, because of the frequent loss of tissue from both necrotizing infection and ablative procedures performed, such as debridements and partial forefoot amputations to bring it under control. If circulation is not restored, ongoing necrosis and foot loss may result, even if the sepsis has been adequately drained. The increased metabolic demands of active infection and extensive tissue loss may overcome the healing capabilities of an ischemic limb, with otherwise adequate circulation. 13 Although control of sepsis alone, followed by a limited amputation may suffice in some of these cases, 14 prediction of a successful outcome by this treatment course is difficult, and an erroneous judgment usually results in loss of the foot. As a result, we will "err" on the side of foot bypass in these difficult situations, given the low morbidity and

7 988 Tannenbaum et al. Journal of VASCULAR SURGERY mortality rates and high likelihood of limb salvage reported in this study. The high percentage of patients initially discharged with an incompletely healed foot (55%) reflects our practice of shifting the care of stable, slowly healing wounds to the outpatient setting, in an attempt to reduce the high cost of treating them in the hospital. Whether or not this practice is truly appropriate, given the 55% readmission rate in these patients for continued problems with foot healing, is difficuk to determine from this study. The decision of when to discharge a patient with an incompletely healed foot lesion from the hospital is complex and is based on psychosocial as well as medical issues. Many patients are often depressed and discouraged by a lengthy hospitalization, and others have traveled a distance from their homes to be treated in a tertiary care institution such as our own, where they are isolated from their families. Often these patients can have the care of their foot wound managed as an outpatient by family members or visiting nurses. Moreover, 15% of the patients with a fully healed foot at the time of discharge still required at least one readmission for a recurrent lesion or infection in the ipsilateral foot, which was thought to be directly related to the initial problem. These findings reflect the difficult and chronic nature of diabetic foot problems and the need for patient education and regular follow-up, even when circulation has been restored.15 Further study is needed to determine if the additional use of vulnerary agents, such as plateletderived growth factor can accelerate wound healing to alleviate some of these problems. 16 The present study confirms that bypass to the dorsal pedal artery can be safely performed in those patients admitted with ischemia and an active infection. The 12% wound infection rate in this study (including one death as a result of infection) although high, seems reasonable given the 92% graft patency and 98% limb salvage seen at 36 months' follow-up. The costs of limb salvage in this patient population are great, however, because of the need for multiple surgical procedures and lengthy or repeated hospitalizations, especially in patients who initially are admitted with severe infections, lz As distal arterial reconstructive techniques have improved, vascular surgeons have successfully extended vein bypasses progressively more distally in the leg. In diabetic patients, salvaging the greatest number of limbs frequently requires bypass to the foot itself, z in preference to isolated popliteal or tibia] segments. 4,~s Although this may often place the distal anastomosis in close proximity to active foot sepsis (35% in our experience), the presence of infection need not diminish expectations for achieving ultimate limb salvage. REFERENCES 1. Reichle FA, Shuman CR, Tyson RR. Femorotibial bypass in the diabetic patient for salvage of the ischemic lower extremity. Am I Surg 1975;129: Auer AI, Hurley JJ, Binnington HB, Nunnelee JD, Hershey FB. Distal tibial vein grafts for limb salvage. Arch Surg 1983;118: Rosenblatt MS, Quist WC, Sidawy AN, Paniszyn CC, LoGerfo FW. Results of vein graft reconstruction of the lower extremity in diabetic and nondiabetic patients. Surg Gynecol Obstet 1990;171: Mannick JA, Jackson BT, Coffman JD. Success of bypass vein grafts in patients with isolated popliteal artery segments. Surgery 1967;61: LoGerfo FW, Coffman JD. Vascular and microvascula disease of the foot in diabetes. N Engl J Med 1985;311: Pomposelli FB, Jepsen SJ, Gibbons GW, et al. Efficacy of the dorsal pedal bypass for limb salvage in diabetic patients: short term observations. J VASC SURG 1990;11: Pomposelli FB, Jepsen SJ, Gibbons GW, et al. A flexible approach to infrapophteal vein grafts in patients with diabetes mellims. Arch Surg 1991;126: Gibbons GW, Eliopoulos GM. Infection of the diabetic foot. In: Kozak GP, et al. Management of diabetic foot problems. Philadelphia: WB Saunders, 1984: Rutherford RB, Flanigan DP, Gupta SK, et al. Suggested standards for reports dealing with lower extremity ischemia. J VAsc St3RG 1986;4: Penn I. Infections in the diabetic foot. In: Sammarco GJ. The foot in diabetes. Malvern: Lea & Febiger, 1991: Levin ME. The diabetic foot: pathophysiology, evaluation, and treatment. In: Levin ME, O'Neal LW, eds. The diabetic foot. 4th ed. St. Louis: Mosby, 1987: Gibbons GW. The diabetic foot: amputations and drainage of infection. J VASC SURG 1987;5: Anderson CB, Munn JS. Cutaneous ulcers in the diabetic foot. In: Ernst CB, Stanley JS, ed. Current therapy in vascular surgery. 2nd ed. Philadelphia: BC Decker, 1991: Mclntyre KE. Control of infection in the diabetic foot: the role of microbiology, immtmopathology, antibiotics, and guillotine amputation. J VAsc Si3R~ 1987;5: Bild DE, Selby JV, Sinnock P, Browner WS, Braveman P, Showstack JA. Lower extremity amputation in people with diabetes: epidemiology and prevention. Diabetes Care 1989; 12: Knighton DR, Fylling CP, Fiegel VD, Cerra F. Amputation prevention in an independently reviewed at-risk diabetic population: using a comprehensive wound care protocol. Am J Surg 1990;160: Mackey WC, McCuUough JL, Conlon TP, et al. The costs of surgery for limb-threatening ischemia. Surgery 1986;99: Hurley JJ, Auer AI, Hershey FB, et al. Distal arterial reconstruction: patency and hmb salvage in diabetics. J Vase SURG 1987;5: Submitted Oct. 4, 1991, accepted Jan. 4, 1992.

8 Volume 15 Number 6 June 1992 Dorsal pedal bypass m diabetics with feot infections 989 DISCUSSION Dr. George Meier: (New Haven, Conn.). Dr. Pomposelli and his colleagues are to be commended for their thoughtful management of limbs in a group of patients that until recently were thought to be unsalvageable. In the current report, diabetic foot infections presented only minor obstacles to limb salvage. Close careful management of infection by appropriate surgical debridement, both before and after operation, provided excellent results in this difficult group of patients. A number of points deserve reemphasis. First, in spite of severe infections in more than 30% of their patients, primary graft patency was achieved in 92% at 36 months. Perhaps even more indicative of these exemplary results was the wound infection rate of 12.5%. ]'his in patients with active foot infections and, in many cases, with preoperative cellulitis or lymphangitis. These results compare favorably to the overall infection rates seen after infrainguinal bypass without diabetic foot infection. We at Yale share the enthusiasm of Dr. Pomposelli and his colleagues for the dorsal pedal bypass. Nonetheless, diabetics present a difficult management problem in differentiating the more common neuropathic ulcerations from the truly ischemic ulcerations. If the origin of the foot lesion is ischemic, then bypass after effective infection control is paramount. However, if infection results from neuropathic injury, then bypass may be unnecessary. We have found that in our last 2 years of experience in approximately 40 patients, about 40% of our dorsal pedal bypasses are in patients with ulceration in the setting of moderate but not severe ischemia, where the ischemia combines with neuropathy to a variable degree. My first question to Dr. Pomposelli addresses this issue specifically. What adjunctive modalities do you use to select patients for bypass? Is a certain duration of nonhealing sufficient or are more criteria required before a bypass is offered? Our belief is that truly ischemic lesions in these patients result in many more complications from wound healing, as well as vein bypass conduit complications, than are seen in primarily neuropathic ones. Have you seen differences in the wound complications related to the degree of distal ischemia? Since these data comprise a subgroup of the series presented at the Society meeting in Newport last year, some minor demographic issues should be settled. We see more than 90% of our dorsal pedal bypasses in patients with infected foot lesions, and reported series confirm this experience. Therefore, what were the indications in the remaining 100 bypasses that were reported last year but not included in this subgroup? Were these limbs also ulcerated yet failed to have certain infectious criteria present and thus were excluded from the current analysis? What exactly were the differences between the groups reported today and the remainder? The efficacy of dorsal pedal bypass has also led us at Yale to explore the dorsal pedal artery in patients with only a Doppler signal in the absence of angiographic confirmation of a dorsal pedal artery. Approximately 20% of our dorsal pedal bypasses are aborted after examination of the dorsal pedal artery during operation and completion of intraoperative arteriography. We view this incidence of negative exploration as necessary to ensure bypass for all patients possible. For us, negative exploration was particularly likely in a small group of patients with end-stage diabetic nephropathy on hemodialysis where, in our experience, 75% of the explorations were abandoned. Otherwise, exploration of the dorsalis pedis, without an adequate vessel, by angiography resuked in a successfifl bypass about 60% of the time, although these data certainly depend on the preoperative definition of an adequate vessel. Dr. Pomposelli, what was your negative exploration rate over this time period and what was the success of blind exploration at your institution? Finally, I am curious as I am sure many of my colleagues here today are, as to the current surgical technique for this bypass at the Deaconess. Previously, you have emphasized the open technique with either reversed, nonreversed, or in situ vein. At Yale, we feel that one of the great benefits of angioscopy may ultimately be its use in guiding in situ bypass resulting in limited incisions and ideally in lower complication rates. Since the Deaconess is well known as a leader in the field of angioscopy, has angioscopically guided in situ bypass replaced the open technique reported here? In summary, this is a well-analyzed series that shows the excellent results that can be obtained with dorsal pedal bypass. Dr. Frank Pomposelli. We emphasize that the treatment of dysvascular diabetic patient with a foot lesion complicated by infection often requires complex, lengthy, and costly treatment. One of the greatest obstacles in the future care of these patients may well be whether third party payors are willing to cover the costs incurred during their lengthy hospitalizations. At the present time the allocated length of stay under the diagnosis-related group compensation scheme for a diabetic patient with a foot ulcer and ischemia is 12 days. Yet, the average length of stay of these patients at our institution has been nearly 30 days. Our criteria for bypass in the neuropathic patient relate to the degree of ischemia that we feel is present at the time of initial presentation. At times, this may be difficult to determine. In general, we continue to rely quite heavily on the use of clinical judgement in making this determination. The most important criterion continues to be the absence of a palpable pulse in the foot. Symptoms of rest pain or dandication are not often helpful, since many of these patients are asymptomatic as a result of the presence of their neuropathy and inactivity. If an infected foot requires debridement and/or open partial forefoot amputation, observing the wound on a daily basis is also important. Once infection is eradicated, there should be prompt signs of healing, including the development of wound granulations within several days. If wounds are not showing signs of prompt healing, arteriography is mandatory. There

9 990 Tannenbaum et al. Journa/of VASCULAR SURGERY appears to be a group of patients with significant occlusive disease who are asymptomatic until they develop a foot lesion and infection, at which time their ischemia becomes more critical, perhaps because of the increased metabolic demands placed on the tissues with the presence of infection. In all such patients and even in those patients where healing is marginally acceptable, we will perform dorsal pedal bypasses since we have had good results both in terms of patency and more important, limb salvage. In general, anlde-brachial indexes have been unreliable in helping us determine how ischemic these patients are and whether vascular reconstruction is necessary. Pulse volume recordings have proved more helpful. If the amplitude of the forefoot pulse volume deflection is less than 5 mm, healing is unlikely, especially when there is a large soft tissue defect present in the foot. This does not substitute for clinical judgment, however, since we have seen several patients with better forefoot tracings who would not heal until dorsal pedal bypass was performed. Is there a difference in wound complications between the more ischemic and less ischemic patients? The present data do not answer that question; however, we do not think that it does. With regard to our indications in the 100 patients who underwent dorsal pedal bypass but did not have active infection, these data have been reported previously. Noninfected, nonhealing ischemic ulceration was, by far, our most common indication. Ischemic rest pain, dry gangrene, and failed healing of minor me or forefoot amputations made up the rest. We do not perform this procedure for clandication. Approximately 15 patients in this study underwent exploration of the dorsal pedal artery without a vessel visualized on the preoperative arteriogram, and 50% were found to be reconstructible and underwent successful bypass. We have also previously described our method of determining reconstructibility of these arteries. The artery is opened, and a small angiocatheter is placed into the distal outflow tract, and heparin saline solution is gently injected. If injection meets little or no resistance, we will go ahead and perform the bypass. Another criterion include an internal diameter of the distal "artery of at least 1 mm. Calcification is not, in and of itself, a contraindication of bypass, although it makes the performance of the procedure more difficult and this is more likely in patients on hemodialysis. Finally, we continue to completely expose the vein in these procedures. Although angioscopy has proved invaluable in preparation of the conduit, including valve cutting, assessment for strictures, webs, and other defects not externally visible and in assessing the adequacy of the distal anastomosis at the completion of the proce, dure, it has not been usefial for occluding side branches and arteriovenous fistulas. Technology is under development to occlude the side branches endoscopically; however, at its present stage of development, the time involved and the fluid required does not offset the potential wound problems with completely exposing the vein. Careful placement of incisions in exposing the vein and avoidance of parallel incisions on the foot whenever possible, should keep the wound complication rate low from these procedures.