The predictability of the success of arterial reconstruction by means of transcutaneous oxygen tension measurements

Transcription

1 The predictability of the success of arterial reconstruction by means of transcutaneous oxygen tension measurements Paul I. T. Oh, John L. Provan, M.B., M.S., F.R.C.S., F.R.C.S.(C), F.A.C.S., and F. M. Ameli, M.B., F.R.C.S.(Ed), F.R.C.S.(C), F.A.C.S., Toronto, Ontario, Canada The cases of 100 patients with severe peripheral vascular disease were reviewed to determine whether the success or feasibility of arterial reconstruction could be correlated with the increase in transcutaneous tissue oxygen tension (tcpo2) that occurs in all persons when assuming an erect posture from the supine position. Resting supine tcpo2 levels of 20 mm Hg or less at the foot were noted in all patients in this study (normal values mm Hg). Group I patients were defined by an increase of less than 15 mm Hg on standing, whereas group II patients showed an increase of 15 ram Hg or more. Group I patients had a supine tcpo2 value of mm Hg, with an increase of mm Hg on standing. The group II subjects also had a low supine tcpo2 level, mm Hg, but the increase on standing, mm Hg, was significantly higher (p < 0.001). When these levels increased by less than 15 mm Hg on standing in group I (31 patients), only 29% of limbs (10 of 34) were saved or had patent grafts at 3 months and 55% of attempted vascular reconstructions (11 of 20) failed. Amputations were performed in 50% of the limbs (17 of 34). This contrasted in group II (69 patients) with a significantly higher limb salvage and graft patency rate (81%, 57 of 70 limbs), success of reconstruction (79%, 37 of 47 procedures), and a significantly lower rate of amputation (11%, 8 of 70 limbs) (p < 0.001). We conclude that a failure to increase the tcpo2 reading at the foot by at least 15 man Hg on standing indicates a poor prognosis for arterial reconstruction in patients suffering from gangrenous or pregangrenous lesions of that foot. (J VAsc SURG 1987;5: ) Many patients suffering from severe arterial ischemia are subjected to multiple operative procedures in the expectation that their rest pain or gangrene may be ameliorated. Such procedures are expensive, ~ confine the patient to the hospital for long periods, and ultimately may not succeed, 2-s requiring after some time the performance of a below- or aboveknee amputation. 6 If early success or failure of the reconstruction could be predictcd, unnecessary surgical treatment could be avoided, thereby shortening hospital stay and facilitating rehabilitation. 7 Attempts have bcen made to identify factors that may affect the outcome of arterial reconstructions s'8'9 and various physiologic measurements, such as seg- From the Division of Vascular Surge U, The Wellesley Hospital, University of Toronto. Presented at the Thirty-fourth Scientific Meeting of the North American Chapter, International Society for Cardiovascular Surgery, New Orleans, La., June 10-11, Supported by The Wellesley Hospital Research Institute. Reprint requests: J. L. Provan, M.B., Ste. 217 E. K. Jones Bldg., 160 Wellesley St. East, Toronto, Ontario, Canada, M4Y 1J3. mental blood pressures, Doppler ultrasonic velocity analyses, postocclusive reactive hyperemia, and pulse volume recordings, have all been evaluated for their predictive value in peripheral vascular disease.l '11 Recently, transcutaneous measurements of arterial ox~y gen tension (tcp02) have been discovered as an exciting method of assessing the peripheral circulation. 12-1s Preliminary observations by Byrne and Provan had indicated that in patients with gangrene or pregangrene, tcp02 levels of 20 mm Hg or less were frequently observed at the foot with the patient in the supine position (unpublished data). When these levels failed to increase by more than 15 mm Hg on standing, reconstruction was either not feasible, as shown arteriographically, or frequently resuited in thrombosis and amputation. Accordingly, we have retrospectively examined 100 cases to assess whether a failure to increase the tcp02 reading at the foot by at least 15 mm Hg on standing indicates a poor prognosis for arterial reconstruction in patients suffering from gangrenous or pregangrenous lesions of that foot. 356

2 Volume 5 Number 2 February 1987 Measurement of tcpo2 to predict arterial reconstruction success 357 Table I. Patient populations and tcpo2 values Preoperative tcpo2 (ram Hg) No. of patients Age (yr) Male Female Supine Increase Control _ (47%) 19 (53%) (50-77) (0-125) Group I (58%) 13 (42%) _ (50-85) (0.,15) (0-13) Group II _ (55%) 31 (45%) (36-83) (0-18) (15-60) ~Measured in healthy adults without clinical evidence of arterial disease. Table II. Clinical presentation No. of patients Rest pain Gangrene Ulceration Claudication Group I (48%) 13 (42%) 6 (19%) 12 (39%) Group II (46%) 8 (12%) 12 (17%) 47 (68%) NOTE: More than one symptom often present. MATERL&L AND METHODS Measurements of tcpo2. Measurement of tcpo2 was donc by mcans of a tcpo2 monitor (Kontron Instruments, Everett, Mass.) equipped with a miniature Clark electrode, which was placed on the dorsum of the foot. This electrode contained a heating thcrmistor that raised the skin temperature to 45 C, thereby causing dilatation of capillaries, increased blood flow, and diffusion of oxygen to the skin surface where it was electrochemically reduced and quantified. The patient remained supine for 10 to 15 minutes to allow for equilibration once the dectrode was in place. Readings wcre then continued for 3 to 5 minutes while the patient remained standing. The ~acrease (r) in tcpo2 levels was equal to the difference between standing and supine measurements. Postoperative in addition to preoperative measurements of tcpo2 were taken in some cases. A more &tailed instrumental methodology has been reported by Byrne ct al. is Patients. One hundred patients with preoperative supine tcpo2 values of 20 mm Hg or less were reviewed under the following criteria: (1) indications for surgery--rest pain, gangrene, nonhealing ulceration, and/or claudication; (2) associated risk factors and conditions--diabetes mcllitus, smoking more than one package per day, coronary arte~ disease (myocardial infarction, angina, or congestive heart failurc), cerebrovascular accident, hypertension (greater than 150/95 mm Hg), pulmonary insufficiency, and prcvious reconstructive surgery; (3) arteriography; (4) type of procedure--arterial recon- struction, sympathectomy, transluminal dilatation, and amputation; and (5) outcome of treatmentm patency of graft or failure because of thrombosis, amputation, or no relief of rest pain. Two study populations were delineated on the basis of the increase (r) in tcpo2: group I, r < 15 mm Hg; and group II, r > 15 mm Hg. Patients were investigated and underwent treatment between January 1984 and May Data betwecn and within groups were analyzed by means of paired and unpaired t tests. RESULTS A total of 100 patients underwent 112 procedures after a preoperative tcpo2 value of 20 mm Hg or less at the foot in the supine position had been measured. Of these 100 patients, 31 showed an increase in tcpo2 of less than 15 mm Hg on standing (group I) and 69 patients had an increase of 15 mm Hg or greater (group II) (Table I). Persons in group I had a mean age of 68.7 years with a standard deviation of 10.9 years and consisted of 18 men and 13 women. Group II consisted of 38 men and 31 women with a mean age of 66.2 years with a standard deviation of 10.9 years. The two groups were well matched with respect to age and sex. The clinical presentations of the populations are shown in Tables II and III. Rest pain and gangrene were the most frequent indications for operation in group I, whereas severe claudication was the most common complaint in group II. The finding of a significantly higher proportion of patients with gan-

3 Journal of 358 Oh, Provan, and Ameli VASCULAR SURGERY Table IIL Associated conditions No. of Diabetes Coronary artery Cerebrovaseular Pulmonary Previous vascular patients mellitus Smoking disease accident Hypertension insufficiency surgery Group I (35%) 18 (58%) 16 (52%) 9 (29%) 16 (52%) 4 (13%) I0 (32%) Group II (22%) 53 (77%) 31 (45%) 15 (22%) 31 (45%) 9 (I3%) 23 (33%) Table IV. Postoperative course Group I Group II Procedure No. Success Fail No. Success Fail Aortobifemoral bypass 5 4 AKA THR, BKA Femoropopliteal bypass THR, 2 BKA 19 I7 THR, BKA Femorotibial bypass THR 2 0 THR, BKA Fcmorofemoral cross- 3 I AKA, THR over graft Axillofemoral bypass 3 i 2 THR THR, BKA Profimdaplasty Lumbar sympathectomy BKA Transluminal dilatation Primary amputation BKA, 1 AKA Secondary amputation BKA, 2 AKA BKA None 1 0 THR Total BKA = below-knee amputation; AKA = above-knee amputation; THR = thrombosis with no amputation. grene in group I (p < 0.001) and claudicants in group II (p < 0.005) may reflect the relative severity of the disease process in the two groups. The greater incidence of diabetes mellitus in the group I population very likely contributed to this observation. Surprisingly, a greater number of smokers (more than one package per day) was found in group II patients (p < 0.05). The occurrence of the other conditions corresponded fairly well between the two groups. The preoperative tcpo2 values are shown in Table I. Before all procedures in the group I population, the supine tcpq was 4.24 ± 5.31 mm Hg and an increase of 3.91 ± 4.59 mm Hg on standing was seen. Group II patients had a supine tcpo2 of 5.73 ± 4.98 mm Hg with an increase of _ mm Hg. The increase in tcpo2 was significantly higher in the group II patients (p < 0.001). The surgical treatments and outcomes for both groups are presented in Table IV. Generally, operation was performed for limb salvage and most claudicants in group II were managed conservatively. In group I, 31 people underwent 37 procedures. Ten patients had had reconstructive surgery before this date. One patient with rest pain was not operated on as she was considered too great a cardiac risk. She later died with gangrene of the foot. Two other deaths occurred in this group, both from cardiac causes, one after an axillofemoral bypass and the other aftcr a below-knee amputation. No significant difference was found between thc initial tcpo2 values of the successes vs. the failures. Graft patency and limb survival were achieved in 10 of 34 limbs (29%) at the time of follow-up with 11 patients having unilateral amputations, three with bilateral amputations, and one revision from below-knee to aboveknee; this gave a 50% amputation ratc among the treatment failures. Five persons who had advanced gangrene required primary amputation without a re: constructive procedure. The 12 sccondary amputations were performed after a total of 17 failed attempts at rcvascularization. Of the 21 reconstructive arterial procedures carried out, 11 (55%) resulted in early (less than 3 months) graft failure. All patients in group II were evaluated for pcriphcral vascular disease, resulting in 55 operative procedures and 19 cases in which conservative treatment was deemed appropriate for intermittent clandication only. The latter course was followed either if the risk of surgery was considered too great (three patients) or if the disease was not seriously affecting the patient's lifestyle or limb viability (16 patients). Twenty-four persons had had a previous arterial operation and one patient was excluded from the study because of other conditions that precluded a reconstructive proccdurc. One death occurred in this group, caused by a myocardial infarction after an

4 Volume 5 Number 2 FebruaDT 1987 Measurement of tcpo2 to predict arterial reconstruction success 359 Table V. Preoperative vs. postoperative tcpo2 Preop (ram Hg) Postap (ram Hg) No. Supine Erect Supine Erect Group I Success 7 Failure 7 Group II Success 27 Failure 3 *p < tno statistical significance. ~:p < ? 5.2 _ _ _ t 42.6 ± _ ~ 19.1" 17.7 z ± 16.5" 34.7 ± 31.1~ * 24.7 _+ 28.8:~ 54.6 _ _+ 21.1~ Table VI. Angiographic findings Group I Group II Total Runoff" vessels No. Success No. Success No. Success None 14 7 One 11 2 Two 4 2 Three 4 3 Total axillobifemoral procedure. Graft paten W and limb survnal were apparent in 57 of 70 limbs examined (81%). The failures included six unilateral belowknee amputations and one bilateral, giving an 11% rate of limb loss. All amputations were performed after a failed bypass. Thirty-seven of 47 reconstructions (79%) were patent at 3 months. Limb survival and graft patency at 3 months were both significantly higher in group II patients (p < 0.001) whereas the proportion of amputations performed was significantly greater in group I (p < 0.001). In the comparison of pre- vs. postoperative tcp02 measurements (Table V), an increase in both supine and standing levels was seen after reconstructive surgery in both groups (p < 0.1). However, a significant elevation was evident in the procedures that succeeded (p < 0.001), reflecting quantitative improvement in tissue perfusion. There were five cases in which the tcp02 level failed to increase at all after operation; all five required amputation. Forty-eight angiograms were also examined. Runoffwas assessed either for the single leg that was operated on or for both limbs in aortobifemoral procedures (Table VI). A significantly greater proportion of lbnbs had what can be described as "good" runoff (two- or three-vessel) in group II patients (p < 0.001) and a significantly greater proportion of successes was noted when runoff was good (76%) rather than poor (39%) (p < 0.001). In comparing successes with failures within groups, no difference was found in the group I population because most patients had basically poor runoff. However, in group II patients, the limbs in which operations succeeded had significantly better runoffthan those that failed (p < 0.01). This seems to indicate that twoor three-vessel runoff, as visualized angiographically, holds a fairly good prognosis for reconstruction. Similarly, when the runoff is poor, so is the outlook. However, procedures in this study did succeed with no or one-vessel runoff and the angiogram alone did not differentiate the successes from the failures. Unfortunately, an assessment of the completeness of the plantar arch was not possible in most instances because the plantar arch was not usually seen on the routine angiograms. DISCUSSION A successful reconstructive procedure for limb salvage is gratifying to both patient and surgeon. However, if operation fails, the status of the patient may be worsened, with an increase in suffering, time, and expense for all concerned. There is clearly a need to assess the chances of success in all patients and, if possible, to identify, those in whom reconstructive surgery is least likely to succeed. In general, if a graft fails, the limb returns to its preoperative state, 2 but it may be made more isch-

5 360 Oh, Provan, and Ameli Journal of VASCULAR SURGERY emic because of distal thrombosis and interruption of collateral channels.l Unsuccessful limb salvage attempts may require further reconstructions, causing increased morbidity, and can adversely affect the ultimate amputation level in patients initially considered to be candidates for a below-knee amputation. 6'2 Kazmers, Satiani, and Evans 21 reported a study in which 19 of 40 limbs (48%) with failed rcvascularizations ultimately required above-knee amputations, whereas only 13% did when 51 primary amputations were performed. Others have also reported a lower incidence of successful healing of below-knee amputations after failed bypasses. 2'I9 The importance of preserving the knee joint to allow more effective rehabilitation and mobilization, decreased energy demands and pain, and lower mortality rates from the surgery has been outlined by Berardi and Keonin 22 and this must be considered before any surgery is attempted. Adequate oxygenation is one of the main factors involved in tissue viability and wound healing. White et al.16 showed that tcpo2 follows oxygen delivery, the product of arterial oxygen content and blood flow, in the patient with peripheral vascular disease. therefore, tcpo2 carl provide some objective measure of the viability of an ischemic limb. Burgess ct al. 2s have used tcp02 to quantitate the healing potential of possible sites of amputation in dysvascular limbs, whereas Achaver, Kirby, and Litke 24 have reported a method of predicting the survival of skin graft flaps. The low resting supine tcp02 levels (0 to 20 mm Hg) seen in the patients in our study reflect the severity of the underlying arterial disease, which limits the blood flow and oxygen supply to the cutaneous capillary beds. It was pointed out by Tonncson 2s that a tcp02 of zero did not indicate that there was no oxygen reaching the tissue but rather that there was none in excess of the metabolic requirements of that tissue. When blood flow is further compromised as in exercise or prolonged recumbency, tissues become more severely ischemic and gangrene may develop. Our earlier studies have related the reduction in resting tcpo2 values to the clinical severity of the arterial insufficiency. When a person arises from a supine to an erect position, a redistribution of peripheral blood flow occurs. Rushmcr 26 described a significant increase in blood volume in the dependent extremities with a concurrent decrease in blood flow through the legs. The oxygen content of the femoral venous blood was reduced to about half normal. This increased arteriovenous oxygen gradient reflects increased oxygen extraction by the tissues in the erect position. The long vertical column of blood in this position also creates an elevated hydrostatic pressure on the arterial side, which leads to an increased perfusion of cutaneous capillary beds. is This heightened extraction and perfusion leads to improved oxygen delivery to the skin and thereby an increase in tcpo2 levels. This change can be quite marked, as in the group II patients, and is an explanation for the decrease in rest pain on dependency. The situation just described can only occur if ar, terial channels, both primary runoff vessels, and secondary collateral vessels exist in the first place. When the arteries to accommodate the increased blood supply on standing are occludcd or so severely diseased as to be nonfunctional, there can be no enhancement of perfusion or oxygen delivery nor will there be significant increase in tcp02. The group I patients display such characteristics. The situation is even worse in the absence of pedal arches 9 or with distal thrombosis of the small arteries of the foot, as seen in diabetic patients. 27'28 The placement of a graft to bypass a proximal stenosis or occlusion is an attempt to improve oxygen delivery to the distal extremity by increasing blood flow through the leg. Again, distal arterial channels must be patent to distribute the augmented blood supply. This situation is similar to the distribution of peripheral blood flow on standing. Therefore, it would seem that a greater clinical improvement after reconstructive surgery might be achieved in patients in whom a large increase (/>15 mm Hg) in tcp02 on standing could be demonstrated preoperatively than in those with a small increase (<15 mm Hg). The greater graft patcncy and limb salvage rates in the group II patients provide some support for this hyt' pothesis. Other noninvasive studies have also recently been correlated with success of reconstructive arterial operations.10 Anklc/brachial pressure indices measured with Doppler ultrasound have been studied by several groups, 811"29"3 but there has been variation in the findings reported. Quantitative velocity and reactive hyperemia analyses, pulse volume recordings, and pulse reappearance time after occlusion have shown some promise? Transcutaneous oxygen measurements provide certain advantages over other available methods, as outlined by Burgess et al. 23 and Byrne et al.i8 The measurement of tcpo2 can be made simply at any location along the limb and it appears to be superior to other methods in demonstrating the capacity for oxygen delivery to the skin. The electrodes respond rapidly to any fluxes in the local circulation, as seen

6 Volume 5 Number 2 February 1987 Measurement of tcpo2 to predict arterial reconstruction success 361 in postural changes. Earlier work in our laborato U has shown that tcp02 is more sensitive and specific than ankle pressure and readings can be taken in the absence of distal pulses or Doppler signals? ~ Measurements of tcp02 directly assess the adequacy of nutritional factors reaching the tissues. Other methods, including angiography, evaluate mainly the pathways that the blood follows, not whether an area is viable or whether a wound or ulcer will heal. Finally, this technique is strictly noninvasive and objective, and patients do not experience discomfort during the testing. Therefore, tcp02 is a highly effective method to monitor peripheral vascular disease and to identify patients in whom the disease is severely advanced. A prospective study is ideally required to test the predictive value of the criteria reported herein, but the present study does indicate the value of tcp02 in the surgical management of the severely ischcmic limb and also places another objective tool at the surgeon's disposal to help him in assessing the suitability of a patient for arterial reconstruction. REFERENCES 1. Gupta SK, Veith FJ, Samson RH, Scher LA, Weiser R, White-Flores SA. Cost analysis of operations for infrainguinal arteriosclerosis. Circulation 1982;66(Suppl II):9. 2. Naji A, Chu J, McCombs PR, Barker CF, Berkowitz HD, Roberts B. Results of 100 consecutive femoropopliteal vein grafts for limb salvage. Ann Surg 1978;188: Couch NP, Wheeler HB, Hyatt DF, Crane C, Edwards EA, Warren R. Factors influencing limb survival after femoropopliteal reconstruction. Arch Surg 1967;95: Myers KA, King RB, Scott DF, Johnson N, Morris pj. Surgical treatment of the severely ischaemic leg. II: Salvage rates. Br J Surg 1978;65: O'Donnel ]A, Brener BJ, Brief DK, Alpert J, Parsonnet V. Realistic expectations for patients having lower extremity bypass surgery for limb salvage. Arch Surg 1977; 112: Wooster DL, Provan JL. Fate of the limb after failed femoropopliteal reconstruction. Can J Surg 1982;25: Bell PRF. Are distal vascular procedures worthwhile? Br J Surg 1985;72: Dean RH, Yao JST, Stanton PE, Bergan JJ. Prognostic indicators in femoropopliteal reconstruction. Arch Surg 1975; 110: Imparato Advl, Kim GE, Madayag M, Haveson S. Angiographic criteria for successful ribial arterial reconstructions. Surgery 1973;74: Bernstein EF, Stuart SH, Fronek A. The predictive value of noninvasive testing in peripheral vascular disease. In: Bernstein EF, ed. Noninvasive diagnostic techniques in vascular disease. St. Louis: The CV Mosby Co, 1982: Samson RH, Gupta SK, Veith FJ, Ascer E, Scher L. Perioperative noninvasive hemodynamic ankle indices as predictors of infrainguinal graft patency. J VASC SURG 1985;2: Shoemaker WC, Vidyasagar D. Physiological and clinical significance of Ptco2 and Ptco2 measurements. Crit Care Med 1981;9: Lubbers DW. History of transcutaneous PQ measurement. Crit Care Med 1981;9: Matsen FA, Wyss CR, Pedegana LR, et al. Transcutaneous oxygen tension measurement in peripheral vascular disease. Surg Gynecol Obstet 1980;150: Franzeck UK, Talke P, Bemstein EF, Golbranson FL, Fronek A. Transcutaneous Po2 measurements in health and peripheral arterial occlusive disease. Surgery 1982;91: White RA, Nolan L, Harley D. Noninvasive evaluation of peripheral vascular disease using transcutaneous oxygen tension. Am J Surg 1982;144: Cina C, Katsamouris A, Megerman J, et al, Utility of transcutaneous oxygen tension measurements in peripheral arterial occlusive disease. J Vase SURG 1984; 1: i8. Byrne P, Provan JL, Ameli FM, Jones DP. The use oftranscutaneous oxygen tension measurements in the diagnosis of peripheral vascular insufficiency. Ann Surg 1984;200: Dardik H~ Kahn M, Dardik I, Sussman B, Ibrahim IM. Influence of failed vascular bypass procedures on conversion of below-knee to above-knee amputation levels. Surgery 1982; 91: Raviola CA, Nichter L, Baker JD, et al. Femoropopliteal final bypass: what price failure? Am J Surg 1982;I44: Kazmers M, Satiani B, Evans WE. Amputation level following unsuccessful distal limb salvage operations. Surgery 1980;87: Berardi RS, Keonin Y. Amputations in peripheral vascular occlusive disease. Am J Surg 1978;135: Burgess EM, Matsen FA, Wyss CR, Skmmons CW. Segmental transcutaneous measurements of Po2 in patients requiring below-the-knee amputation for peripheral vascular insufficiency. J Bone Joint Surg 1982;64: Achaver BM, Kirby SB, Litke DK. Transcutaneous Po2 in flaps: a new method of survival prediction. Hast Reconstr Surg t980;65: Tormeson KH. Transcutaneous oxygen tension in imminent foot gangrene. Acta Anaesth Scand 1978;68: Rushmer RF. Cardiovascular dynamics, 4th ed. Philadelphia: WB Saunders Co, 1976: Steer HW, Cuckle HS, Franklin PM, Morris PJ. The influence of diabetes mellitus upon peripheral vascular disease. Surg Gynecol Obstet 1983;157: Edwards EA. Postamputation radiographic evidence for small artery obstruction in arteriosclerosis. Ann Surg 1959; Sumner DS, Strandiaess DE. Hemodynamic studies before and after extended bypass grafts to the tibial and peroneal arteries. Surgery i979;86: Corson JD, Johnson WC, LoGerfo FW, et al. Doppler ankle systolic blood pressure: prognosticvalue in vein bypass grafts of the lower extremity. Arch Surg 1978;i13: Provan JL, Byrne P, Ameli FM. Transcutaneous O2 tension as a noninvasive index of successfial vascular reconstruction. (Abstr) San Diego Symposium on Noninvasive Diagnostic Techniques in Vascular Surgery, October 1982.

7 362 Oh, Provan, and Ameli Journal of VASCULAR SURGERY DISCUSSION Dr. Willard C. Johnson (Boston, Mass.). I certainly enjoyed this presentation and its technique of measuring the increase in local tissue oxygen tension by dependency. I wish to ask whether that type of observation can be used to aid in healing of amputations of the forefoot. For example, in a patient with digital gangrene, if the tissue oxygen level increases when you put him in arterial position and possibly give him supplementary nasal oxygen and an increase from 20 up to 35 mm Hg is seen, is it appropriate to do a digital amputation on the basis of these augmentations of tissue oxygen? Dr. Provan. Dr. Johnson, we have not used this technique for the assessment of digital amputations in a way that I can describe to you. We have used it for an assessment of the determination whether a below-knee amputation will heal or not, and we found that a level of transcutaneous oxygen tension below the knee of 25 mm Hg predicted with almost 100% accuracy whether a below-knee amputation would heal; this was the subject of a prospective study that we did some time ago. Certainly we do not believe that if you find a tcpo2 level in the foot on the order of 2, 3, 4, or 5 mm Hg, that a primary amputation of the toe is likely to succeed, but we have not related that to an increase in tcpo2 on standing. Dr. T. J. Bunt (Columbia, S.C.). Have the authors had the opportunity to measure the oxygen tensions after revascularization and can they correlate the success or the lack of success with a change in these oxygen tensions after a revascularization? Dr. Provan. Yes, we have indeed done that. The incidence is in fact much lower in those patients whosc reconstruction does not succeed, although for some reason the levels were actually higher than before operation, which we didn't entirely understand. The trouble with postoperative measurement, particularly in the early postoperative period, is that it is subject to considerable variation, depending on the patient's cardiac output and hemodynamic status, so that it is not so useful in the postoperative period as one would like it to be. Dr. J. Dennis Baker (Los Angeles, Calif.). I have two questions. First, it seemed from your data that you attempted fewer bypasses in group I patients, the bad-risk group. What other criteria did you use to reject the patients for bypass operation? Second, did you find any consistent correlation be- tween the lack of increase in oxygen measurement in group I patients and the physical examination? It would seem that these patients would have either a very pale or a cyanotic forefoot. Dr. Provan. We used the conventional criteria for the performance of arterial reconstruction. I think there were 11 patients altogether in group I who did not undergo revascularization either because their general cardiac status did not permit it or because we did not think it was feasible from an arteriographic standpoint. Regarding your second question, there was no difference between the two groups on clinical examination, apart from the higher incidence of gangrenous lesions in the group I patients. Many of these patients did in fact show dependent rubor and it was similar in the two groups. Clinically, both groups of parents were severely ischemi,. Dr. Charles O. Brantigan (Denver, Colo.). Your observations are interesting, but I would like to question their practical significance. If I understand the data, you were able to achieve successful revascularization in a third of the poor-risk group of patients. Therefore, it seems to me that irrespective of transcutaneous oxygen measurements, you are still going to attempt limb salvage in what appear to be otherwise favorable cases. What do you perceive as the practical application of this technique? Dr. Provan. That is a very good question and it is one that we have been concerned with ourselves. I think this is one more factor to consider before advising revascularization. This technique is the most useful form of noninvasive assessment, we believe, and probably superior to other noninvasive examinations. I think if a patient has angiographically poor runoff, which makes us somewhat concerned as to whether revascularization can be done at all, and no increase in the oxygen tension is seen on standing, I think the surgeon might suggest the possibility tha revascularization will not succeed. Without this test the suggestion might be to explore. Dr. Richard J. Gusberg (New Haven, Conn.). Did you notice any difference in the predictive value of tcpo2 in the diabetic patients compared with the nondiabetics? Dr. Provan. The tcpo2 levels are on the whole always worse in the diabetic patient; again, I think this reflects the severity of their distal vascular disease. We did not specifically investigate the diabetic patients as a subgrou p in terms of this predictive value, but I think it is worth doing.