Efficacy of the dorsal pedal bypass for limb salvage in diabetic patients: Short-term observations Frank B. Pomposelli, Jr., MD, Stephen J. Jepsen, MD, Gary W. Gibbons, MD, David R. Campbell, MD, D o r o t h y V. Freeman, MD, Arnold Miller, MD, and Frank W. LoGerfo, MD, Boston, Mass. Limbs of diabetic patients with distal tibial disease are frequently considered unreconstructible; however, when studied with intraarterial digital subtraction angiography, the dorsal pedal artery is frequently found to be patent. We have reviewed our recent experience with 96 patients, 94% of whom had diabetes and had 97 bypasses placed to the dorsal pedal artery. All procedures were for limb salvage. Superimposed infection was present in 42.3%. In 92 instances where intraarterial digital subtraction angiography successfiaUyvisualized the dorsal pedal artery, 91 bypasses were placed. In 12 other cases where the dorsal pedal artery was not visualized by intraarterial digital subtraction angiography but audible with the continuous-wave Doppler, bypasses were completed successfiflly in six. All procedures were performed with vein. Inflow was taken from the femoral artery in 48, popliteal artery in 45, tibial artery in 2, and from a femoral tibial graft in 2. Perioperative mortality was 1.92%. Actuarial graft patency, limb salvage, and patient survival were 82%, 87%, and 80%, respectively at 18 months. We conclude that bypass grafting to the dorsal pedal artery can be reliably performed with acceptable shortterm results. An attempt should always be made to visualize the foot vessels angiographically, especially in diabetic patients, so that this valuable option in arterial reconstruction will not be overlooked. (J VAse StrRG 1990;11:745-52.) Advances in angiographic technology, surgical techniques, 1,2 and anesthesia/critical care management have all contributed to the improved results with infrapopliteal reconstruction for limbthreatening ischemia. This has proved especially important in diabetic patients whose pattern of occlusive disease frequently involves the tibial and perk,deal vessels? Our practice is somewhat unique because of the high incidence of diabetic patients (greater than 90%) referred with ischemic limb-threatening lesions. The use of routine intraarterial digital subtraction angiography (IADSA) improved our ability to visualize the distal arterial tree in these patients. As a result of our improved imaging of the distal circulation, we have identified a subset of individuals where the tibial vessels were extensively diseased but the dorsal pedal artery was spared. From the Division of Vascaflar Surgery, New England Deaconess Hospital, Harvard Medical School. Presented at the Sixteenth Annual Meeting of the New England Society for Vascular Surgery, Bretton Woods, N.H., Sept. 2122, 1989. Reprint requests: Frank Pomposelli, MD, 110 Francis Street, Suite 9C, Boston, MA 02215. ~/6/19884
Because our early occasional experience with vein bypass grafting to the dorsal pedal artery proved quite favorable, we standardized our technique and broadened our indications. We present our initial experience with the dorsal pedal artery bypass graft (DP bypass) including indications, technique, and early results. MATERIAL AND METHODS Patients The charts of 96 patients who underwent 104 attempted DP bypasses between August 1986 and May 1989 were reviewed. The clinical characteristics of this group and the indications for surgery are depicted in Table I. All procedures were done for limb salvage, and no patient was operated on for claudication alone. The 44 patients (42.3%) with associated secondary infection in addition to ischemia were initially treated with broad spectrum intravenous antibiotics and underwent bedside or intraoperative debridement and incision and drainage as required. Seven of these patients required partial forefoot open amputation for control of sepsis before vascular reconstruction. Fifty-eight patients underwent preoperative noninvasive testing in our diagnostic vascular laboratory 745
746
Journal of VASCULAR SURGERY
Pomposelli et al.
Table I. Clinical data and indications
Table II. Angiographic findings
for surgery Clinical data Males Females Diabetes Myocardial infarction/angina Previous peripheral vascular surgery Previous amputation Previous transmetatarsal amputation Indications ~ Ischemic ulcer Rest pain Failed amputation Gangrene Secondary infection
Findings
No. of patients (n = 96)
%
64 32 90 33 28 41 6
67 33 94 34 29 43 6
51 36 6 31 44
49 35 6 30 42
~Some patients had more than one indication present.
including segmental limb pressures, Doppler waveform analyses, and pulse volume recordings.4 The mean preoperative ankle/brachial index was 0.64. All patients studied had flat or nearly flat (less than 5 mm amplitude) forefoot pulse volume tracings. Angiographic criteria All 104 extremities were studied with preoperative IADSA. The arterial tree was visualized from aortic bifurcation to the base of the toes. All levels of occlusion or stenosis were determined, and the best vessel in continuity with the pedal arch was identiffed when present. Anteroposterior and lateral projections of the foot with magnification views as needed were obtained to visualize the pedal vessels. The usual pattern of occlusive disease suitable for DP bypass grafting was occlusion of all three tibial vessels either proximally or distally and no demonstrable continuity with the reconstituted dorsal pedal bxtery below the ankle. The angiographic findings are depicted in Table II. Fourteen of the 92 patients where the dorsal pedal artery was visualized had a segment of diseased or small distal peroneal artery in continuity with the dorsal pedal artery. In all 12 cases where the artery was not visualized, an audible Doppler signal was present at its anatomic location on the dorsal foot.
Technique The technical principles for successful performance of this procedure are similar to those for other distal reconstructions including use of magnification, patience, gentleness, and attention to detail s In dealing with heavily calcified vessels we prefer interrupting blood flow with intralurninal coronary occluders
Inflow SFA severely diseased or occluded SFA patent or patent fem-pop Isolated popliteal segment Patent fem-tib into isolated tibial segment Outflow Dorsal pedal artery visualized Dorsal pedal artery not visualized
No. of extremities (n = 104) 46 53 3 2 92 12
SFA, Superficial femoral artery; fem-pop, femoral-popliteal bypass; fern-rib, femoral tibial bypass.
(Flo-Restor, Bio-dash Vascular Inc., St. Paul, Minn.) to vascular clamps. Fine monofilament polypropylene sutures (7-0 or 8-0) on small cardiovascular cutting needles are used for the distal anastomoses. By using these techniques, we have not had to fracture the calcified tunica media to make these vessels ~:~ft enough for hemostatic control and s u t u r i n g . 6 We prefer to expose the dorsal pedal artery as the initial step in the operative procedure. The dorsal pedal artery can usually be identified by a vertical incision on the dorsum of the foot starting 1 cm below an imaginary transverse line between the distal ends of the malleoli and 1 cm lateral to the long extensor muscle of the great toe. The artery is fotmd below a dense fascial layer, the inferior extensor retinaculum. The sterile intraoperative continuous-wave Doppler probe is helpful in localizing the artery. Proximally, the dorsal pedal artery gives off its lateral tarsal branch. Distally, at the level of the proximal first metatarsal, it bifurcates into the deep plantar artery and the first dorsal metatarsal artery (Fig. 1). In most cases the artery between these branches is suitatS_~ for bypass grafting. In some circumstances it may be preferable to use one of the branches when the main trunk of the dorsal pedal artery is severely diseased or occluded. The only appropriate conduit to the dorsal pedal artery is vein. We prefer ipsilateral saphenous vein when available but have successfully used contralateral saphenous vein, arm vein, and composites of the above when necessary. Bypass grafts arising from the femoral level are best placed with the in situ technique z when possible. Our preference is to completely expose the vein and cut the valves under direct vision using the modified Mills valvulotome. 8 Mobilizing the distal vein for in situ bypass grafting requires certain precautions. Because of the inframalleolar location of the dorsal pedal artery, it is generally necessary to expose the medial marginal
Volume 11 Number 6 June 1990
Efficacy of the dorsalpedal bypass in diabetic patients 747
t ii Correct position of graft i
!
Perforating branch of
//Peroneal
a. ~t a.
Incorrec' position of graft
medis a.
~te a Great
Saphen(
.,ep antar a.
lnci for
1st Met
Fig. 1. Anatomy of the dorsal pedal artery and its principal branches.
Fig. 2. Sketch of the correct and incorrect pathways for tunneling the saphenous vein for in situ DP bypass. Tunneling through the skin bridge may result in necrosis.
vein, which is the extension of the greater saphenous vein on the foot. This usually results in a skin bridge between the two required parallel incisions to expose the distal vein and the artery. It is important to keep the distal forefoot vein incision as medial as possible t° avoid a narrow skin bridge that may be devitalized. ~, addition, undermining of the skin bridge to tunnel the graft should be avoided by mobilizing the vein far enough proximally to tunnel it proximal to the skin bridge (Fig. 2). To preserve vein and because of theoretic advantages in long-term patency ~,2we routinely take inflow from the most distal vessel possible. When inflow is taken more distally, most commonly from the belowknee popliteal artery, we have used reversed, nonreversed, and in situ techniques for handling the vein. The in situ technique has not been preferred in our more recent experience since the segment left in situ is quite short, and it necessitates the formation of the two parallel incisions on the foot, which can be avoided with the other techniques. When the caliber of the vein at the ankle is larger than at the knee (a frequent occurrence in our experience), we favor a reversed vein bypass. If the caliber is larger at the
knee, we will cut the valves and use the vein in a nonreversed fashion. For the occasional bypass, which takes its inflow from a tibial vessel, the vein is used in a reversed fashion. Occasionally, when neither a tibial vessel nor the dorsal pedal artery has been visualized on the arteriogram and amputation is inevitable, bypass grafting to the dorsal pedal artery may still be possible if a Doppler signal is audible over it. In this circumstance we will make an incision directly over the Doppler signal and expose the vessel in the usual fashion. If no dorsal pedal artery can be identified, the procedure is terminated. If the artery is visualized, the patient is heparinized, and the artery is carefully opened longitudinally. An extended arteriotomy may be necessary to find an adequate lumen of at least 1 mm internal diameter. Our experience with on-table arteriography to assess "outflow" before bypass grafting has been poor. Instead we will rely on intubation of the distal lumen with a small 22-gauge angiocatheter and gently inject heparinized balanced salt solution. If flow is easy we will proceed. High resistance to injection or inability to inject contraindicates bypass placement. Routine injection of the dorsal pedal
Journal of VASCULAR SURGERY
748 Pomposelli et al.
100%
>,..
80
o
Z LLI I-,< 13,.
il 6O
40 SECONDARY PATENCY
o
.............
20
0
N = 97
PRIMARY PATENCY
6
12
18
MONTHS
Fig. 3. Kaplan-Meier life-table plots of primary and secondary patency.
Table III. Types of conduit and source of arterial inflow Conduit~source of arterial inflow
No. of bypasses (n = 97)
Type of conduit ISV RSV NRSV AV COM
48 21 16 6 6
I n f l o w source
FEM POP TIB FEM-TIB
48 45 2 2
ISV, In situ saphenous vein; RSV, reversed saphenous vein; NRSV, nonreversed saphenous vein; AV, arm vein; COM, composite vein; POP, popliteal; FEM, femoral artery; POP, popateal artery; TIB, tibia/artery; FEM-TIB, femoral-tibial bypass graft.
artery in other cases where the artery is clearly patent on the preoperative arteriogram gives the surgeon a baseline of comparison for this maneuver.
Follow-up All patients discharged with patent grafts were seen in 2 weeks and every 3 months during the first year. Patients with open foot wounds at discharge were seen more frequently as needed. Patients were seen in the subsequent years at 6-month intervals. Graft patency was determined by the presence of a palpable pulse in the graft and on the dorsum of the foot. Graft pulsations are easy to detect because of the superficial location of the distal anastomosis, and all patients were instructed to palpate their foot for
Table IV. Additional procedures required to achieve a healed foot in patients with successful DP bypasses Procedure
No. ofpatients
Digit/ray amp TMA Debride/STSG Total
17 1 15 33
Amp, Amputation; T/vIA, transmetatarsal amputation; STSG, spat thickness skin graft.
the presence of a pulse on a daily basis. Nine patients living a great distance from our institution were seen less frequently after the first year of follow-up. ~ calculating graft patency for this study, the interval used for these nine patients was the time of the last clinic visit if less than 6 months. If the interval was greater than 6 months (four patients), the patient was queried by telephone about the status of their foot and the presence of a palpable foot pulse. All patients were able to confidently determine that a foot pulse was present. Ankle pressures and pulse volume recordings were not routinely used to document graft patency. Moreover, the efficacy of these parameters in discovering failing grafts has not been determined. Nevertheless, recurrence of symptoms, deterioration of forefoot pulse volume tracings, ankle pressures, graft pulsations, or Doppler flow signals all mandate repeat arteriography. Patient survival, graft patency, and limb salvage rates were calculated by use of the actuarial life-table method.
Volume 11 Number 6 June 1990
Efficacy of the dorsalpedal bypass in diabetic patients
749
100 1
90
55
32
_
_
18
80
LU
~~
60
133 v
40 J
N = 97
20
A
0
5
1'2
1'8
MONTHS 100
~
45
.n, 67
[
8O
-
~
N = 97
2O
0 B
0
6
12 MONTHS
18
24
Big. 4. A, Kaplan-Meier life-table plot of limb salvage; B, Kaplan-Meier life-table plot of patient survival.
RESULTS Ninety-six patients underwent 104 attempted bypass procedures to the dorsal pedal artery. In the 92 (88.5%) cases where the angiogram successfully visualized the artery, 91 bypass grafts were successfully placed. In the 12 instances where the dorsal pedal artery could not be visualized by IADSA, in spite of an audible Doppler signal, all patients were explored, and six were found to have suitable arteries for bypass gra_~ing and underwent successful operations. Five of the other six patients underwent below-knee amputation. This left a total of 97 extremities for further analysis. The types of vein grafts used and the sources Jf arterial inflow are depicted in Table III.
Two patients died within 30 days of surgery (1.92%). Six bypasses failed within this period (6.1%). The remaining 89 patients with patent grafts at discharge had their feet saved. Thirty-three patients required additional surgical procedures to achieve a healed foot (Table IV). Actuarial life-table analysis of graft patency, limb salvage, and patient survival were 81.8%, 86.9%, and 80%, respectively at 18 months (Figs. 3 and 4). DISCUSSION
This experience further emphasizes the importance of visualizing foot vessels in diabetic patients with limb-threatening ischemia even when all tibial vessels are occluded and a popliteal pulse is pres-
750
Journal of VASCULAR SURGERY
Pomposelli et al.
entfl'2'9 We have found digital subtraction angiography to be helpful in dais group, since it gives good visualization with a minimum of contrast and patient discomfort. Other authors 1°,~ have reported reasonable resuits in tibial reconstructions to vessels where continuity to the pedal circulation had not been determined. However, our previous experience supports the concept that disappointing results may occur with bypasses into tibial segments not in continuity with the pedal vessels when ischemic tissue loss or ulceration is present in the foot.~2 This has stimulated us to place a bypass to the dorsal pedal artery itself in the absence of a connected, more proximal tibial vessel. Our initial concern about the adverse affect the limited outflow bed of the dorsal pedal artery might have on immediate graft patency has proved to be unfounded. Moreover, bypasses were successfully placed in several patients with prior transmetatarsal amputations who subsequently developed ischemic lesions at the end of their stumps. In spite of the exposed location of the distal anastomosis, no graft failed from compression by foot wear. The high rate of limb loss in patients with unsuccessful bypasses or early graft thrombosis or both underscores the severity of ischemia in this group of patients. The advantages of bypass grafting to the dorsal pedal artery seem most evident in situations with forefoot gangrene and secondary sepsis where pulsatile flow is required to promote healing after radical debridement or open amputation or both. It is in these circumstances of extensive tissue loss and where IADSA imaging fails to visualize the dorsal artery of the foot that we will attempt "blind" exploration of the artery when a Doppler signal is present. The 50% success rate we have observed with this last-ditch effort at limb salvage seems appropriate in this situation and in many individuals can be attempted before amputation on the same anesthetic. In patients with lesser degrees of ischemia involving uncomplicated, nonhealing ischemic ulcers, and/or ischemic rest pain, the need to perform a DP bypass is less compelling. This is especially true when a segment of the popliteal artery is visualized distal to a superficial femoral artery occlusion (blind popliteal segment). However, femorall-popliteal bypass grafting in these circumstances may prove technically successful but not relieve symptoms, especially if extensive tibial disease is present. When this occurs, secondary bypass grafting to the dorsal pedal artery usually relieves symptoms. It is difficult to determine which of these patients would best be served by a
primary DP bypass. In general, in the absence ~,f gangrene, extensive tissue loss, or failed amputation, we will choose the more proximal procedure first and only perform a secondary bypass to the pedal level if symptoms persist. In situations where a distal peroneal artery (distal third of the tibia) is in continuity with the dorsal pedal artery, especially when the former is of small diameter or diseased, we will preferentially bypass into the artery since it is a less technically demanding procedure. We continue to favor bypass grafting into more proximal tibial vessels in all other circumstances where continuity to the foot has been demonstrated. If the durability of the DP bypass holds up to longer follow-up, however, old rules may hve to be reappraised. Although the efficacy of bypass grafting into blind popliteal segments and tibial vessels without established continuity to the foot has been demonstrated, ~°,11,1aour own experience with these procedures in diabetic patients with limb-threatening ischemia has proved tm(::edictable and occasionally disappointing. In circumstances of limb-threatening ischemia as a result of tibial occlusive disease, especially in diabetic patients, every effort should be made to visualize the complete distal arterial circulation. When no tibial vessel can be found in continuity with a patent dorsal pedal artery, bypass grafting into the artery itself affords an excellent likelihood of durable limb salvage. REFERENCES
1. Sidawy AN, Menzoian JO, Cantelmo NL, LoGerfo FW. Effect of inflow and outflow on the results of tibioperoneal vein grafts. Am j" Surg 1986;152:211-4. 2. Veith FJ, Ascer E, Gupta SK, et al. Tibiotibial vein bypass grafts: a new operation for limb salvage. J Vasc SUWG 1985;2:552-7. 3. LoGerfo FW, Coffman JD. Vascular and microvascular d/sease of the foot in diabetes: implications for foot care. N Engl J Med 1984;311:1615-9. 4. Gibbons GW, Wheelock FC. Problems in the non-invasive evaluation of the peripheral circulation in the diabetic. Practical Cardiol 1982;8:115-26. 5. Ascer E, Veith FJ, Gupta SK. Bypasses to plantar arteries and other tibial branches: an extended approach to limb salvage. J"VASC SURG 1988;8:434-41. 6. Ascer E, Veith FJ, White Flores SA. Infrapopliteal bypasses to heavily calcified rock-like arteries. Am J Surg 1986;152: 220-3. 7. Leather RP, Powers SR, Karmody AM. A reappraisal of the in situ saphenous vein arterial bypass: its use in limb salvage. Surgery 1979;86:453-61. 8. Leather RP, Shah DM, Corson JD, et al. Instrumental evolution of the valve incision method of in situ saphenous vein bypass. J VAsc SURG 1984;1:113-23. 9. Cantelrno NL, Snow JR, Menzoian JO, LoGerfo FW. Successful vein bypass in patients with an ischemic limb and a palpable popliteal pulse. Arch Surg 1986;121:217-20.
Volume 11 Number 6 June 1990
10. Reichle FA, Rankin KP, Tyson RR, et al. Long-term results of 474 arterial reconstructions for severely ischemic limbs: a fourteen year follow up. Surgery 1978;85:93-100. 11. Buchbinder D, Rollins DL, Sernrow CM, et al. In sire tibial reconstruction. State of the art or passing fancy. Ann Surg 1988;207:184-8,
Efficacy of the dorsalpedal bypass in diabetic patients 751
12. Imparato AM, Kim GE, Madayag M, et al. Angiographic criteria for successful tibial arterial reconstructions. Surgery 1973;74:830-8. 13. Mannick JA, Jackson BT, Coffman JD, et al. Successof bypass vein grafts in patients with isolated popliteal artery segments. Surgery 1967;6 I: 17-25.
DISCUSSION
Dr. A n t h o n y D. Whittemore (Boston, Mass.). The authors are to be commended for this truly remarkable series, in a unique patient population of primarily diabetic patients. Ninety-seven vein grafts to the dorsal pedal artery in the space of a relatively short 3-year period of time is most definitely remarkable, and the resulting 86% limb salvage rate is equally notable, partioalarly since most of these patients had either osteomyelitis or gangrene. During the past 5 years there have been many groups who have published very encouraging results with respect to distal 'bi "'~revascularization. We have had a very favorable experience with the in situ technique. The femorotibial patency rates in nearly 200 femorotibial bypass operations is now over 80% with the in situ methodology. Now whether this is unique to the in situ method or not is open debate, since as you know John Porter has published very similar statistics with autogenous, but reversed vein grafts. Enrico Ascer and Frank Veith have published some very encouraging resuks of grafting to not only the dorsal pedal artery but also to the branches of the dorsal pedal artery. Nevertheless, the bypasses and the patency rates expressed in the current series, although limited to the 18-month mark, are extraordinarily commendable, and we certainly hope that it remains sustained over the next 31/2 years. We have had only 18 of these truly inframalleolar bypasses over the last 5 years, but our results show tht more than 80% of them do in fact remain patent. Several important points tb~3gh, deserve emphasis. The first is the importance of the extensive digital subtraction angiography that can reveal a very suitable vessel, even if there is extensive proximal tibial disease. Certainly the distal tibial or dorsal pedal artery is more attractive than an isolated tibial segment. If, however, such vessels are not visualized even with extensive digital subtraction angiography, exploration is still dearly warranted. They were able to find suitable vessels in more than 50% of the vessels that were not visualized anglographically. Second, as Enrico Ascer has advocated, if the main pedal artery proves to be too diseased, there is nothing wrong with some of the branches distally. Acceptable results have clearly been achieved. Third, I think the issue of temporary distal occlusion deserves special mention. The authors make a point of avoiding the use of external clamps in these often circumferentially calcified vessels and make a point to use intraluminal balloon occlusion control. I suspect that is very important. The authors have used a variety of inflow sources and autologous and alternatives to saphenous vein. Have you noticed any difference in the
various subgroups with respect to the long-term results? That is, do you expect the same results with an arm vein or a lesser saphenous vein from the popliteal down to a dorsal pedal branch? Second, would you address the issue of timing? If you are faced with a patient with osteomyelitis or gangrene, do you stage the procedures and carry out simultaneous definitive procedure for the tissue necrosis at the same time yon place the bypass graft or do you advocate a staged approach? Dr. David Brewster (Boston, Mass.). In our experience with in situ grafts to the dorsal pedal artery, the positioning of two parallel incisions in front of the medial malleolus and on the dorsum of the foot has led to some problems with cutaneous necrosis. Have you experienced this? Second, in the 10% or so of patients without visualization on preoperative angiography, would you recommend direct exploration? What about so-called prebypass intraoperative arteriography to minimize wound complications? Finally, do you recommend any long-term pharmacologic measures to enhance long-term patency? Dr. F. B. Pomposelli (dosing). In answer to Dr. Brewster's questions, we have also been very concerned about the two parallel incisions on the foot that are necessary when placing an in situ bypass graft to obtain sufficient vein length to reach the dorsal pedal artery. It is for this reason that we have recently been avoiding the in situ technique in bypasses arising from more distal inflow sources, principally the popliteal artery, where the extra vein length can be obtained by mobilizing the saphenous vein into the distal thigh and terminating the vein incision at the ankle. In bypasses arising from the femoral level, however, the dorsal foot vein must be exposed which results in an island of skin between this incision and the separate incision to expose the dorsal pedal artery. We have avoided problems with this skin bridge by mobilizing the vein far enough proximally that it can be tunneled subcutaneously proximal to the skin bridge itself. Tunneling the vein through the skin bridge may result in necrosis. Second, we will try to space these incisions far enough apart that the resulting skin bridge is as wide as possible. This may also decrease the chances for necrosis. By using these principles we have not had a case of necrosis of the skin bridge when the graft was patent. However, necrosis developed in the skin bridge of one of my own patients when the graft failed and the foot became ischemic again. We have occasionally tried "on table" arteriography and have been disappointed with the information we obtain in
752
Pomposelli et al.
these very distal and small vessels. What we will usually do in the situation where we are uncertain about patency o f the dorsal pedal artery is to explore the artery first, usually with a small incision directly over the Doppler signal. The patient is then heparinized, and the artery is carefully opened. If the internal luminal diameter is 1 mm as determined by a small coronary dilator, we will take a small polyethylene 22-gauge catheter and inject into the distal artery with heparinized saline solution. If flow is easy we will proceed with bypass grafting. If there is excessive resistance or no flow possible we will not proceed with bypass grafting. In some patients who are facing certain amputation this last attempt at limb salvage can occasionally be done on the same anesthetic as amputation if the patient is willing to accept that. The only pharmacologic measure we will use in the attempt to enhance long-term patency on a routine basis is aspirin as long as the patient has no contraindication. However, whether or not this truly improves long-term results is uncertain at the present time.
Journal of VASCULAR SURGERY
Dr. Whittemore asked questions about the variety of inflow sources and the variety of vein sites and whether we have been able to detect any differences in durability over the long term. At this juncture it is impossible to tell. Moreover, we have not detected any trends, although we are in the process o f evaluating that issue. It is not unusual for these patients to have had previous peripheral vascular or coronary artery procedures, and for that reason we feel a flexible approach is necessary with regard to sources of inflow and venous conduit to maximize the likelihood that you will be able to perform a successful bypass. Finally, in the patient with osteomyelitis or gangrene, we advocate a staged approach. Generally, we will control infection first with broad spectrum intravenous antibiotics and incision and drainage, debridement, or open amputation as needed. Once infection is controlled, we will proceed to arteriography and vascular reconstruction. We then will delay definitive reconstructive procedures on the foot until after revascularization.
Because our early occasional experience with vein bypass grafting to the dorsal pedal artery proved quite favorable, we standardized our technique and broadened our indications. We present our initial experience with the dorsal pedal artery bypass graft (DP bypass) including indications, technique, and early results. MATERIAL AND METHODS Patients The charts of 96 patients who underwent 104 attempted DP bypasses between August 1986 and May 1989 were reviewed. The clinical characteristics of this group and the indications for surgery are depicted in Table I. All procedures were done for limb salvage, and no patient was operated on for claudication alone. The 44 patients (42.3%) with associated secondary infection in addition to ischemia were initially treated with broad spectrum intravenous antibiotics and underwent bedside or intraoperative debridement and incision and drainage as required. Seven of these patients required partial forefoot open amputation for control of sepsis before vascular reconstruction. Fifty-eight patients underwent preoperative noninvasive testing in our diagnostic vascular laboratory 745
746
Journal of VASCULAR SURGERY
Pomposelli et al.
Table I. Clinical data and indications
Table II. Angiographic findings
for surgery Clinical data Males Females Diabetes Myocardial infarction/angina Previous peripheral vascular surgery Previous amputation Previous transmetatarsal amputation Indications ~ Ischemic ulcer Rest pain Failed amputation Gangrene Secondary infection
Findings
No. of patients (n = 96)
%
64 32 90 33 28 41 6
67 33 94 34 29 43 6
51 36 6 31 44
49 35 6 30 42
~Some patients had more than one indication present.
including segmental limb pressures, Doppler waveform analyses, and pulse volume recordings.4 The mean preoperative ankle/brachial index was 0.64. All patients studied had flat or nearly flat (less than 5 mm amplitude) forefoot pulse volume tracings. Angiographic criteria All 104 extremities were studied with preoperative IADSA. The arterial tree was visualized from aortic bifurcation to the base of the toes. All levels of occlusion or stenosis were determined, and the best vessel in continuity with the pedal arch was identiffed when present. Anteroposterior and lateral projections of the foot with magnification views as needed were obtained to visualize the pedal vessels. The usual pattern of occlusive disease suitable for DP bypass grafting was occlusion of all three tibial vessels either proximally or distally and no demonstrable continuity with the reconstituted dorsal pedal bxtery below the ankle. The angiographic findings are depicted in Table II. Fourteen of the 92 patients where the dorsal pedal artery was visualized had a segment of diseased or small distal peroneal artery in continuity with the dorsal pedal artery. In all 12 cases where the artery was not visualized, an audible Doppler signal was present at its anatomic location on the dorsal foot.
Technique The technical principles for successful performance of this procedure are similar to those for other distal reconstructions including use of magnification, patience, gentleness, and attention to detail s In dealing with heavily calcified vessels we prefer interrupting blood flow with intralurninal coronary occluders
Inflow SFA severely diseased or occluded SFA patent or patent fem-pop Isolated popliteal segment Patent fem-tib into isolated tibial segment Outflow Dorsal pedal artery visualized Dorsal pedal artery not visualized
No. of extremities (n = 104) 46 53 3 2 92 12
SFA, Superficial femoral artery; fem-pop, femoral-popliteal bypass; fern-rib, femoral tibial bypass.
(Flo-Restor, Bio-dash Vascular Inc., St. Paul, Minn.) to vascular clamps. Fine monofilament polypropylene sutures (7-0 or 8-0) on small cardiovascular cutting needles are used for the distal anastomoses. By using these techniques, we have not had to fracture the calcified tunica media to make these vessels ~:~ft enough for hemostatic control and s u t u r i n g . 6 We prefer to expose the dorsal pedal artery as the initial step in the operative procedure. The dorsal pedal artery can usually be identified by a vertical incision on the dorsum of the foot starting 1 cm below an imaginary transverse line between the distal ends of the malleoli and 1 cm lateral to the long extensor muscle of the great toe. The artery is fotmd below a dense fascial layer, the inferior extensor retinaculum. The sterile intraoperative continuous-wave Doppler probe is helpful in localizing the artery. Proximally, the dorsal pedal artery gives off its lateral tarsal branch. Distally, at the level of the proximal first metatarsal, it bifurcates into the deep plantar artery and the first dorsal metatarsal artery (Fig. 1). In most cases the artery between these branches is suitatS_~ for bypass grafting. In some circumstances it may be preferable to use one of the branches when the main trunk of the dorsal pedal artery is severely diseased or occluded. The only appropriate conduit to the dorsal pedal artery is vein. We prefer ipsilateral saphenous vein when available but have successfully used contralateral saphenous vein, arm vein, and composites of the above when necessary. Bypass grafts arising from the femoral level are best placed with the in situ technique z when possible. Our preference is to completely expose the vein and cut the valves under direct vision using the modified Mills valvulotome. 8 Mobilizing the distal vein for in situ bypass grafting requires certain precautions. Because of the inframalleolar location of the dorsal pedal artery, it is generally necessary to expose the medial marginal
Volume 11 Number 6 June 1990
Efficacy of the dorsalpedal bypass in diabetic patients 747
t ii Correct position of graft i
!
Perforating branch of
//Peroneal
a. ~t a.
Incorrec' position of graft
medis a.
~te a Great
Saphen(
.,ep antar a.
lnci for
1st Met
Fig. 1. Anatomy of the dorsal pedal artery and its principal branches.
Fig. 2. Sketch of the correct and incorrect pathways for tunneling the saphenous vein for in situ DP bypass. Tunneling through the skin bridge may result in necrosis.
vein, which is the extension of the greater saphenous vein on the foot. This usually results in a skin bridge between the two required parallel incisions to expose the distal vein and the artery. It is important to keep the distal forefoot vein incision as medial as possible t° avoid a narrow skin bridge that may be devitalized. ~, addition, undermining of the skin bridge to tunnel the graft should be avoided by mobilizing the vein far enough proximally to tunnel it proximal to the skin bridge (Fig. 2). To preserve vein and because of theoretic advantages in long-term patency ~,2we routinely take inflow from the most distal vessel possible. When inflow is taken more distally, most commonly from the belowknee popliteal artery, we have used reversed, nonreversed, and in situ techniques for handling the vein. The in situ technique has not been preferred in our more recent experience since the segment left in situ is quite short, and it necessitates the formation of the two parallel incisions on the foot, which can be avoided with the other techniques. When the caliber of the vein at the ankle is larger than at the knee (a frequent occurrence in our experience), we favor a reversed vein bypass. If the caliber is larger at the
knee, we will cut the valves and use the vein in a nonreversed fashion. For the occasional bypass, which takes its inflow from a tibial vessel, the vein is used in a reversed fashion. Occasionally, when neither a tibial vessel nor the dorsal pedal artery has been visualized on the arteriogram and amputation is inevitable, bypass grafting to the dorsal pedal artery may still be possible if a Doppler signal is audible over it. In this circumstance we will make an incision directly over the Doppler signal and expose the vessel in the usual fashion. If no dorsal pedal artery can be identified, the procedure is terminated. If the artery is visualized, the patient is heparinized, and the artery is carefully opened longitudinally. An extended arteriotomy may be necessary to find an adequate lumen of at least 1 mm internal diameter. Our experience with on-table arteriography to assess "outflow" before bypass grafting has been poor. Instead we will rely on intubation of the distal lumen with a small 22-gauge angiocatheter and gently inject heparinized balanced salt solution. If flow is easy we will proceed. High resistance to injection or inability to inject contraindicates bypass placement. Routine injection of the dorsal pedal
Journal of VASCULAR SURGERY
748 Pomposelli et al.
100%
>,..
80
o
Z LLI I-,< 13,.
il 6O
40 SECONDARY PATENCY
o
.............
20
0
N = 97
PRIMARY PATENCY
6
12
18
MONTHS
Fig. 3. Kaplan-Meier life-table plots of primary and secondary patency.
Table III. Types of conduit and source of arterial inflow Conduit~source of arterial inflow
No. of bypasses (n = 97)
Type of conduit ISV RSV NRSV AV COM
48 21 16 6 6
I n f l o w source
FEM POP TIB FEM-TIB
48 45 2 2
ISV, In situ saphenous vein; RSV, reversed saphenous vein; NRSV, nonreversed saphenous vein; AV, arm vein; COM, composite vein; POP, popliteal; FEM, femoral artery; POP, popateal artery; TIB, tibia/artery; FEM-TIB, femoral-tibial bypass graft.
artery in other cases where the artery is clearly patent on the preoperative arteriogram gives the surgeon a baseline of comparison for this maneuver.
Follow-up All patients discharged with patent grafts were seen in 2 weeks and every 3 months during the first year. Patients with open foot wounds at discharge were seen more frequently as needed. Patients were seen in the subsequent years at 6-month intervals. Graft patency was determined by the presence of a palpable pulse in the graft and on the dorsum of the foot. Graft pulsations are easy to detect because of the superficial location of the distal anastomosis, and all patients were instructed to palpate their foot for
Table IV. Additional procedures required to achieve a healed foot in patients with successful DP bypasses Procedure
No. ofpatients
Digit/ray amp TMA Debride/STSG Total
17 1 15 33
Amp, Amputation; T/vIA, transmetatarsal amputation; STSG, spat thickness skin graft.
the presence of a pulse on a daily basis. Nine patients living a great distance from our institution were seen less frequently after the first year of follow-up. ~ calculating graft patency for this study, the interval used for these nine patients was the time of the last clinic visit if less than 6 months. If the interval was greater than 6 months (four patients), the patient was queried by telephone about the status of their foot and the presence of a palpable foot pulse. All patients were able to confidently determine that a foot pulse was present. Ankle pressures and pulse volume recordings were not routinely used to document graft patency. Moreover, the efficacy of these parameters in discovering failing grafts has not been determined. Nevertheless, recurrence of symptoms, deterioration of forefoot pulse volume tracings, ankle pressures, graft pulsations, or Doppler flow signals all mandate repeat arteriography. Patient survival, graft patency, and limb salvage rates were calculated by use of the actuarial life-table method.
Volume 11 Number 6 June 1990
Efficacy of the dorsalpedal bypass in diabetic patients
749
100 1
90
55
32
_
_
18
80
LU
~~
60
133 v
40 J
N = 97
20
A
0
5
1'2
1'8
MONTHS 100
~
45
.n, 67
[
8O
-
~
N = 97
2O
0 B
0
6
12 MONTHS
18
24
Big. 4. A, Kaplan-Meier life-table plot of limb salvage; B, Kaplan-Meier life-table plot of patient survival.
RESULTS Ninety-six patients underwent 104 attempted bypass procedures to the dorsal pedal artery. In the 92 (88.5%) cases where the angiogram successfully visualized the artery, 91 bypass grafts were successfully placed. In the 12 instances where the dorsal pedal artery could not be visualized by IADSA, in spite of an audible Doppler signal, all patients were explored, and six were found to have suitable arteries for bypass gra_~ing and underwent successful operations. Five of the other six patients underwent below-knee amputation. This left a total of 97 extremities for further analysis. The types of vein grafts used and the sources Jf arterial inflow are depicted in Table III.
Two patients died within 30 days of surgery (1.92%). Six bypasses failed within this period (6.1%). The remaining 89 patients with patent grafts at discharge had their feet saved. Thirty-three patients required additional surgical procedures to achieve a healed foot (Table IV). Actuarial life-table analysis of graft patency, limb salvage, and patient survival were 81.8%, 86.9%, and 80%, respectively at 18 months (Figs. 3 and 4). DISCUSSION
This experience further emphasizes the importance of visualizing foot vessels in diabetic patients with limb-threatening ischemia even when all tibial vessels are occluded and a popliteal pulse is pres-
750
Journal of VASCULAR SURGERY
Pomposelli et al.
entfl'2'9 We have found digital subtraction angiography to be helpful in dais group, since it gives good visualization with a minimum of contrast and patient discomfort. Other authors 1°,~ have reported reasonable resuits in tibial reconstructions to vessels where continuity to the pedal circulation had not been determined. However, our previous experience supports the concept that disappointing results may occur with bypasses into tibial segments not in continuity with the pedal vessels when ischemic tissue loss or ulceration is present in the foot.~2 This has stimulated us to place a bypass to the dorsal pedal artery itself in the absence of a connected, more proximal tibial vessel. Our initial concern about the adverse affect the limited outflow bed of the dorsal pedal artery might have on immediate graft patency has proved to be unfounded. Moreover, bypasses were successfully placed in several patients with prior transmetatarsal amputations who subsequently developed ischemic lesions at the end of their stumps. In spite of the exposed location of the distal anastomosis, no graft failed from compression by foot wear. The high rate of limb loss in patients with unsuccessful bypasses or early graft thrombosis or both underscores the severity of ischemia in this group of patients. The advantages of bypass grafting to the dorsal pedal artery seem most evident in situations with forefoot gangrene and secondary sepsis where pulsatile flow is required to promote healing after radical debridement or open amputation or both. It is in these circumstances of extensive tissue loss and where IADSA imaging fails to visualize the dorsal artery of the foot that we will attempt "blind" exploration of the artery when a Doppler signal is present. The 50% success rate we have observed with this last-ditch effort at limb salvage seems appropriate in this situation and in many individuals can be attempted before amputation on the same anesthetic. In patients with lesser degrees of ischemia involving uncomplicated, nonhealing ischemic ulcers, and/or ischemic rest pain, the need to perform a DP bypass is less compelling. This is especially true when a segment of the popliteal artery is visualized distal to a superficial femoral artery occlusion (blind popliteal segment). However, femorall-popliteal bypass grafting in these circumstances may prove technically successful but not relieve symptoms, especially if extensive tibial disease is present. When this occurs, secondary bypass grafting to the dorsal pedal artery usually relieves symptoms. It is difficult to determine which of these patients would best be served by a
primary DP bypass. In general, in the absence ~,f gangrene, extensive tissue loss, or failed amputation, we will choose the more proximal procedure first and only perform a secondary bypass to the pedal level if symptoms persist. In situations where a distal peroneal artery (distal third of the tibia) is in continuity with the dorsal pedal artery, especially when the former is of small diameter or diseased, we will preferentially bypass into the artery since it is a less technically demanding procedure. We continue to favor bypass grafting into more proximal tibial vessels in all other circumstances where continuity to the foot has been demonstrated. If the durability of the DP bypass holds up to longer follow-up, however, old rules may hve to be reappraised. Although the efficacy of bypass grafting into blind popliteal segments and tibial vessels without established continuity to the foot has been demonstrated, ~°,11,1aour own experience with these procedures in diabetic patients with limb-threatening ischemia has proved tm(::edictable and occasionally disappointing. In circumstances of limb-threatening ischemia as a result of tibial occlusive disease, especially in diabetic patients, every effort should be made to visualize the complete distal arterial circulation. When no tibial vessel can be found in continuity with a patent dorsal pedal artery, bypass grafting into the artery itself affords an excellent likelihood of durable limb salvage. REFERENCES
1. Sidawy AN, Menzoian JO, Cantelmo NL, LoGerfo FW. Effect of inflow and outflow on the results of tibioperoneal vein grafts. Am j" Surg 1986;152:211-4. 2. Veith FJ, Ascer E, Gupta SK, et al. Tibiotibial vein bypass grafts: a new operation for limb salvage. J Vasc SUWG 1985;2:552-7. 3. LoGerfo FW, Coffman JD. Vascular and microvascular d/sease of the foot in diabetes: implications for foot care. N Engl J Med 1984;311:1615-9. 4. Gibbons GW, Wheelock FC. Problems in the non-invasive evaluation of the peripheral circulation in the diabetic. Practical Cardiol 1982;8:115-26. 5. Ascer E, Veith FJ, Gupta SK. Bypasses to plantar arteries and other tibial branches: an extended approach to limb salvage. J"VASC SURG 1988;8:434-41. 6. Ascer E, Veith FJ, White Flores SA. Infrapopliteal bypasses to heavily calcified rock-like arteries. Am J Surg 1986;152: 220-3. 7. Leather RP, Powers SR, Karmody AM. A reappraisal of the in situ saphenous vein arterial bypass: its use in limb salvage. Surgery 1979;86:453-61. 8. Leather RP, Shah DM, Corson JD, et al. Instrumental evolution of the valve incision method of in situ saphenous vein bypass. J VAsc SURG 1984;1:113-23. 9. Cantelrno NL, Snow JR, Menzoian JO, LoGerfo FW. Successful vein bypass in patients with an ischemic limb and a palpable popliteal pulse. Arch Surg 1986;121:217-20.
Volume 11 Number 6 June 1990
10. Reichle FA, Rankin KP, Tyson RR, et al. Long-term results of 474 arterial reconstructions for severely ischemic limbs: a fourteen year follow up. Surgery 1978;85:93-100. 11. Buchbinder D, Rollins DL, Sernrow CM, et al. In sire tibial reconstruction. State of the art or passing fancy. Ann Surg 1988;207:184-8,
Efficacy of the dorsalpedal bypass in diabetic patients 751
12. Imparato AM, Kim GE, Madayag M, et al. Angiographic criteria for successful tibial arterial reconstructions. Surgery 1973;74:830-8. 13. Mannick JA, Jackson BT, Coffman JD, et al. Successof bypass vein grafts in patients with isolated popliteal artery segments. Surgery 1967;6 I: 17-25.
DISCUSSION
Dr. A n t h o n y D. Whittemore (Boston, Mass.). The authors are to be commended for this truly remarkable series, in a unique patient population of primarily diabetic patients. Ninety-seven vein grafts to the dorsal pedal artery in the space of a relatively short 3-year period of time is most definitely remarkable, and the resulting 86% limb salvage rate is equally notable, partioalarly since most of these patients had either osteomyelitis or gangrene. During the past 5 years there have been many groups who have published very encouraging results with respect to distal 'bi "'~revascularization. We have had a very favorable experience with the in situ technique. The femorotibial patency rates in nearly 200 femorotibial bypass operations is now over 80% with the in situ methodology. Now whether this is unique to the in situ method or not is open debate, since as you know John Porter has published very similar statistics with autogenous, but reversed vein grafts. Enrico Ascer and Frank Veith have published some very encouraging resuks of grafting to not only the dorsal pedal artery but also to the branches of the dorsal pedal artery. Nevertheless, the bypasses and the patency rates expressed in the current series, although limited to the 18-month mark, are extraordinarily commendable, and we certainly hope that it remains sustained over the next 31/2 years. We have had only 18 of these truly inframalleolar bypasses over the last 5 years, but our results show tht more than 80% of them do in fact remain patent. Several important points tb~3gh, deserve emphasis. The first is the importance of the extensive digital subtraction angiography that can reveal a very suitable vessel, even if there is extensive proximal tibial disease. Certainly the distal tibial or dorsal pedal artery is more attractive than an isolated tibial segment. If, however, such vessels are not visualized even with extensive digital subtraction angiography, exploration is still dearly warranted. They were able to find suitable vessels in more than 50% of the vessels that were not visualized anglographically. Second, as Enrico Ascer has advocated, if the main pedal artery proves to be too diseased, there is nothing wrong with some of the branches distally. Acceptable results have clearly been achieved. Third, I think the issue of temporary distal occlusion deserves special mention. The authors make a point of avoiding the use of external clamps in these often circumferentially calcified vessels and make a point to use intraluminal balloon occlusion control. I suspect that is very important. The authors have used a variety of inflow sources and autologous and alternatives to saphenous vein. Have you noticed any difference in the
various subgroups with respect to the long-term results? That is, do you expect the same results with an arm vein or a lesser saphenous vein from the popliteal down to a dorsal pedal branch? Second, would you address the issue of timing? If you are faced with a patient with osteomyelitis or gangrene, do you stage the procedures and carry out simultaneous definitive procedure for the tissue necrosis at the same time yon place the bypass graft or do you advocate a staged approach? Dr. David Brewster (Boston, Mass.). In our experience with in situ grafts to the dorsal pedal artery, the positioning of two parallel incisions in front of the medial malleolus and on the dorsum of the foot has led to some problems with cutaneous necrosis. Have you experienced this? Second, in the 10% or so of patients without visualization on preoperative angiography, would you recommend direct exploration? What about so-called prebypass intraoperative arteriography to minimize wound complications? Finally, do you recommend any long-term pharmacologic measures to enhance long-term patency? Dr. F. B. Pomposelli (dosing). In answer to Dr. Brewster's questions, we have also been very concerned about the two parallel incisions on the foot that are necessary when placing an in situ bypass graft to obtain sufficient vein length to reach the dorsal pedal artery. It is for this reason that we have recently been avoiding the in situ technique in bypasses arising from more distal inflow sources, principally the popliteal artery, where the extra vein length can be obtained by mobilizing the saphenous vein into the distal thigh and terminating the vein incision at the ankle. In bypasses arising from the femoral level, however, the dorsal foot vein must be exposed which results in an island of skin between this incision and the separate incision to expose the dorsal pedal artery. We have avoided problems with this skin bridge by mobilizing the vein far enough proximally that it can be tunneled subcutaneously proximal to the skin bridge itself. Tunneling the vein through the skin bridge may result in necrosis. Second, we will try to space these incisions far enough apart that the resulting skin bridge is as wide as possible. This may also decrease the chances for necrosis. By using these principles we have not had a case of necrosis of the skin bridge when the graft was patent. However, necrosis developed in the skin bridge of one of my own patients when the graft failed and the foot became ischemic again. We have occasionally tried "on table" arteriography and have been disappointed with the information we obtain in
752
Pomposelli et al.
these very distal and small vessels. What we will usually do in the situation where we are uncertain about patency o f the dorsal pedal artery is to explore the artery first, usually with a small incision directly over the Doppler signal. The patient is then heparinized, and the artery is carefully opened. If the internal luminal diameter is 1 mm as determined by a small coronary dilator, we will take a small polyethylene 22-gauge catheter and inject into the distal artery with heparinized saline solution. If flow is easy we will proceed with bypass grafting. If there is excessive resistance or no flow possible we will not proceed with bypass grafting. In some patients who are facing certain amputation this last attempt at limb salvage can occasionally be done on the same anesthetic as amputation if the patient is willing to accept that. The only pharmacologic measure we will use in the attempt to enhance long-term patency on a routine basis is aspirin as long as the patient has no contraindication. However, whether or not this truly improves long-term results is uncertain at the present time.
Journal of VASCULAR SURGERY
Dr. Whittemore asked questions about the variety of inflow sources and the variety of vein sites and whether we have been able to detect any differences in durability over the long term. At this juncture it is impossible to tell. Moreover, we have not detected any trends, although we are in the process o f evaluating that issue. It is not unusual for these patients to have had previous peripheral vascular or coronary artery procedures, and for that reason we feel a flexible approach is necessary with regard to sources of inflow and venous conduit to maximize the likelihood that you will be able to perform a successful bypass. Finally, in the patient with osteomyelitis or gangrene, we advocate a staged approach. Generally, we will control infection first with broad spectrum intravenous antibiotics and incision and drainage, debridement, or open amputation as needed. Once infection is controlled, we will proceed to arteriography and vascular reconstruction. We then will delay definitive reconstructive procedures on the foot until after revascularization.
