Noninvasive determination of healing of major lower extremity amputation: The continued role of clinical judgment

Transcription

1 Noninvasive determination of healing of major lower extremity amputation: The continued role of clinical judgment Willis H. Wagner, MD, Blair A. Keagy, MD, Mamdouh M. Kotb, MD, Steven J. Burnham, MD, and George Johnson, Jr., MD, Chapel Hill, N.C. Various tests are used preoperatively to differentiate patients who require an above-knee amputation (AKA) from those whose vascular supply is adequate to heal a below-knee procedure (BKA). This 15-month study of 109 amputations compared four of these methods: segmental Doppler systolic pressure measurements, transcutaneous oxygen measurement (tcpo2), fluorescein angiography, and skin thermometry. There were 66 BKAs (85% healed primarily) and 43 AKAs (93% healed primarily). The actual level of amputation was determined by the operating surgeon without consideration of the preoperative test results, and the incidence of healing was then related to the test parameters. The average skin temperature at the amputation site was higher (93.7 F) in the group that healed primarily compared with those who required operative stump revision (89.9 F) (p < 0.001). The mean midcalf tcpo2 was also higher in the BKA group that healed (Po mm Hg) compared with those who failed (Po mm Hg) (p < 0.001). Qualitative skin fluorescence was less successful in differentiating success from failure. Of the 63 BKAs that fluorescein predicted would heal, eight failed (13%). Doppler pressures at the thigh, popliteal, midcalf, or ankle level were unreliable in predicting healing of a BKA. Formulation of indexes relating absolute pressures to the brachial systolic pressure did not improve the value of this examination. From this review it is concluded that the skin temperature and tcp02 obtained at the site of proposed amputation were the most reliable prognostic noninvasive examinations. Depending on which test was used, routine application of the noninvasive predictors in this study as the sole determinant of amputation level would have resulted in an AKA in 17% to 60% of patients whose amputations otherwise would have healed at a lower level. Objective data can supplement clinical judgment but not replace it. ( J VAsc SURG 1988; 8: ) Amputation of the lower extremity is a frequently performed operation that may be associated with long-term disability. Whenever possible, a belowknee amputation (BKA) should be performed because rehabilitation is easier and more successful if the knee joint is present. 1~ For this reason many surgeons frequently attempt a BKA instead of an above-knee amputation (AKA), even when the clinical impression is that the lower extremity circulation is less than optimal for healing. We used a belowknee operative site in most of the 1028 consecutive major lower extremity amputations (64%) performed at our institution during a 13-year period? The failure rate of these BKAs was 19%. During this From the Department of Surgery, University of North Carolina School of Medicine. Presented at the Symposium on Non-Invasive Diagnostic Techniques in Vascular Disease, San Diego, Calif., Feb , Reprint requests: Blair A. Keagy, MD, 108 Burnett-Womack Bldg., CB 7065, Chapel Hill, NC period the decision regarding site of resection was made predominantly on the basis of clinical parameters, such as the vascularity of the muscle and the condition of the skin at the proposed site of resection. Infrequently, a patient had a primary AKA when a flexion contracxure or the neurologic status prevented ambulation or effective transfer on the involved extremity. In an effort to decrease the number of unsuccessful BKAs, while at the same time offering that procedure to patients who would benefit from it, various noninvasive and semiinvasive tests have been developed to discriminate between healers and nonhealers of a BKA. These tests include segmental Doppler systolic blood pressures, skin thermometry, transcutaneous oxygen measurements, and fluorescein angiography. Doppler-derived arterial pressures provide a simple, noninvasive, and inexpensive evaluation of regional blood flow. Skin thermometry offers an objective means of determining the site of amputation by monitoring the skin temperature at 703

2 Journal of VASCULAR 704 Wagner et al. SURGERY Table I. Relationship of Doppler-derived pressure to amputation healing Healed Failed Level (mra Hg) No. (mm Hg) No. p Value* Above-knee Femoral Popliteal NS Calf t 2 NS Ankle NS Popliteal-arm NS Calf-arm NS Ankle-arm NS Below-knee Femoral NS Popliteal 146 _ NS Calf NS Ankle NS Popliteal-arm NS Calf-arm NS Ankle-arm NS Data expressed as mean _+ standard error. *Student's nonpaired two-tailed t test. Significance determined atp < 0.05 level. t O, No pulsatile Doppler flow detected at this level. various sites on the lower extremity. Assuming that skin temperature correlates with blood flow, proponents of the test suggest that a relatively high regional temperature is associated with an increased likelihood of healing. Transcutaneous oxygen tension (tcpo2) measurements can be considered a reflection of the amount of oxygen available for diffusion to the skin surface under conditions of maximal local hyperthermia. The amount of oxygen detected by the sensor is a balance of oxygen delivery and local physiologic demands. A low tcpo2 suggests little metabolic reserve capable of contributing to wound healing. Fluorescein angiography is a qualitative, yet reproducible estimate of local skin perfusion. When fluorescein is injected intravenously, it rapidly diffuses through the capillary wall of well-vascularized tissue and into the extracellular spaces. Regional distribution of dye is then detected under ultraviolet light. It was the purpose of this study to evaluate each of these tests in a prospective fashion. The examinations were performed in a group of patients before they were to undergo a major lower extremity amputation. The surgeon had no knowledge of the test results when he made his operative decision. These various parameters were then correlated with the healing rate of the amputation. MATERIAL AND METHODS During the 15-month study period, 109 patients had major amputation of a single limb, including 66 below-knee procedures (61%) and 43 above-knee operations (39%). The determination of the level of amputation was made by the operating surgeon solely on clinical grounds (condition of skin, presence of pulses, and vascularity of the muscle) without knowledge of any of the study results. The series consisted of 73 men and 36 women whose average age was 64.8 years. Seventy-four patients were black; 35 were white. Diabetes was present in 49 patients (45%). Amputations were considered to have healed primarily if no significant debridement or revision was required. Primary healing was accomplished in 56 of 66 BKAs (85%) and 40 of 43 AKAs (93%). Segmental Doppler examinations were performed on 95 patients, skin thermometry on 92, transcutaneous oxygen measurements in 103 patients, and fluorescein angiography in all 109. To ensure uniformity, all examinations were performed with one of the authors (M. M. K.) present. Segmental Doppler arterial pressures were obtained from the femoral, above-knee popliteal, mid-calf (anterior or posterior tibial), and dorsalis pedis or posterial tibial (ankle) arteries. Brachial artery pressures were used to calculate ratios of the lower extremity pressures to the arm. Measurements of tcpo2 were made at room temperature with a Novametrix model 818 transcutaneous oxygen and carbon dioxide monitor (Novametrix Medical Systems, Waningford, Conn.) using a sensor heated to 45 C. The patients were supine during the determinations, and no supplemental oxygen was administered. Sensors were first calibrated to air, then the zero point was determined with a zeroing solution (#8231, Novametrix Medical Systems). The skin was cleansed with an alcohol swab

3 Volume 8 Number 6 December 1988 Determination of lower extremi~ amputation healing 705 *(500) (5130) o(5) (5) O ~ 160 ~ ~ ~ x 1.0 _ z z~ (,~m; # ~ 0.8 l r 40 - ~* (rr.~) i < : L-."! 0 - U.4 - p=ns HEAL FAIL Fig. 1. Distribution of Doppler systolic calf pressures for below-knee amputations. and the electrode attached to the skin with a doublesided adhesive ring. Readings were taken on both the medial and lateral surfaces of the mid-calf and lower thigh. A reference reading was also taken approximately 5 cm below the clavicle on the midclavicular line, representing a central tcpo2 baseline for the patient. Skin thermometry was carried out in the patient's room at an ambient temperature of approximately 75 F (23.8 C). The electronic telethermometer (model 43TD, Yellow Springs Instrument Co., Yellow Springs, Colo.) used a single input from a surface temperature probe. During the test the door was closed and a light blanket was used to cover the patient's legs to minimize air current and the effect of the surrounding environment. Readings were taken on the medial and lateral surface of the leg and thigh corresponding to the possible levels of amputation, and the readings were reported as a mean for each site. Fluorescein angiography was carried out in the operating room after induction of anesthesia, before skin preparation. Sodium fluorescein (15 to 20 mg/kg of body weight, depending on the pigmentation of the patient) was injected intravenously. After 10 to 15 minutes the room was darkened and the skin was examined for fluorescence with an 0.2 IHI ~ HEAL p=ns Ih FAIL Fig. 2. Distribution of Doppler systolic calf~arm indexes for below-knee amputations. ultravolet-emitting lamp (Wood's lamp). Healthy tissue reflects a yellow fluorescence, whereas ischernic tissue reflects a blue to black light. A photograph of each limb was taken with a modified Polaroid camera (Polaroid SLR 680, New Orleans, La.) and a fluorescent filter. A prediction regarding at what level the amputation would heal was then made by the examiner on the basis of this qualitative test but that information did not influence the operative decision regarding amputation site. Values for each quantitative examination are reported as the mean + the standard error. Mean values for the transcutaneous oxygen measurement, temperature determinations, and the segmental Doppler pressures were compared by means of Student's nonpaired t test. RESULTS Segmental Doppler arterial pressures were predictive only of AKA healing. Although femoral pressures yielded the only statistically significant results for AKA, a definite trend was evident at each level of the measurements (Table I). Significance was lh'nited by the small number of AKA failures. Doppler

4 706 Wagner et al. Journal of VASCULAR SURGERY OJ o 30 fl_ 1.2 p- 20 I0 0 Q I. k (mean)~ - U ~e _ 0 r 00 (nleon),.o p<.o01 HEAL FAIL Fig. 3. Distribution of transcutaneous Po2 for below-knee amputations. pressures were not predictive of healing at the belowknee level (Table I). The relationship of lower extremity to brachial artery pressure did not improve the value of this examination (Table I). Separate analysis of Doppler pressures and ratios in diabetic and nondiabetic patients showed no difference in their lack of correlation with BKA healing. Analysis of Doppler pressures obtained at the site of a potential BKA shows substantial overlap of values for those who healed and those whose amputation did not heal (Fig. 1). Comparison of calf pressures with arm pressures did not increase discrimination (Fig. 2). Measurements of tcpo2 were determined in 103 amputated extremities. The mean chest tcpo2 (reference value) was not significantly different between those patients with healed and those with nonhealed amputation ( vs 58.8 _ 3.7 mm Hg, p = NS). The mean thigh and calf tcpo2 determinations were higher for both AKA and BKA in those patients whose procedure was successful than in those whose operations failed (Table II). The distribution of tcpo2 measurements for patients undergoing BKA is given in Fig. 3. Table II. Relationship of transcutaneous Po2 to amputation level Healed Failed Level (ram Hg) No. (ram Hg) No. p Value* Above-knee Thigh 46 _ _+ 7 3 NS Calf NS Below-knee Thigh _ Calf 37 _ <0.001 Data expressed as mean _ standard error. * Student's nonpaired two-tailed t test. Significance determined at p < 0.05 level. Skin temperature at the site of the amputation was obtained in 92 patients. The mean temperature at the amputation site was _ 0.2 F in the patients with a successful amputation and _ 1.0 F in those who required revision (p < 0.001). The distribution of values for the subset of patients having a BKA is given in Fig. 4. All 109 patients had fluorescein angiography. Of the 88 patients whom the fluorescein test indicated would heal a BKA, 63 actually underwent BKA and 55 (87%) of these healed. Poor skin fluorescence suggested the need for AKA in 21 patients. Three of these patients actually had a BKA and one healed (33%). Of the remaining 18 extremities that had an AKA in agreement with the fluorescein results, 16 (89%) healed. To compare the accuracy of these noninvasive tests, cutoff points were derived that optimized the prediction of healing a BKA with each study. Sensitivities and specificities were calculated at various cutoff points and priority was given to the sensitivity. The final cutoff values were: calf pressure greater than 40 mm Hg, calf-arm ratio greater than 0.25, skin temperature greater than 90 F, and calftcpo2 greater than 16 mm Hg. Using these thresholds to calculate sensitivity, specificity, and overall accuracy, skin temperature and tcpo2 were the most reliable indicators (Table III). Receiver operating characteristic (ROC) curves were generated with a series of cutoff points for a positive examination (Fig. 5). Comparing the accuracy of various tests, each with a continuous scale of values, requires the arbitrary determination of a cutoff for a positive test. These thresholds are established to coincide with the goals of the examiner. The ROC curve provides analysis of different tests at several thresholds, thereby improving diagnostic decision making. 6,7 With ROC curves, skin thermometry and calf tcpo2 were again the most accurate examinations for BKA healing. The positive predictive value is a more clinically useful figure than sen-

5 Volume 8 Number 6 December 1988 Determination of lower extremity amputation healing 707 i 96 - Im u_ 94 o Ld F-- < 92 ry 90 I,I 0.. Ld m i (mean) :: p<.o01 * ' 80 I I- 60 / > ~-- ~l//,/ -~'a'--- S kin Temperoture 4 F,/_/ -o-tc o2 ILl ~ /.,~ --n-- Colf Pressure (n 2 0 ~ I SPECIFICITY (%) Fig. 5. Receiver operating characteristic curve compares accuracy of tcpo2 skin temperature, calf Doppler pressures, calf-arm Doppler indexes, and skin fluorescence in predicting healing of a below-knee amputation. HEAL FAIL Fig. 4. Distribution of skin temperature for below-knee amputations. sitivity or specificity because it relates how often a positive test result correctly predicts healing. Conversely, the negative predictive value is the rate at which a BKA will fail when the test result suggests failure. By this analysis, skin temperature is the most balanced noninvasive test in the preoperative determination of BKA success (Table IV). The specificity and negative predictive values reported for all the tests are probably artificially low, because many patients who were predicted to fail by the established criteria actually underwent AKA on the basis of clinical assessment. The results of these examinations were therefore withdrawn from the analysis of BKA healing. The percentage of unnecessary AKAs recommended by noninvasive results were: calf pressure, 60%; calf-arm index, 60%; fluorescein, 33%; calf tcpo2, 30%; and skin temperature, 17%. DISCUSSION Preserving the knee joint is important in the rehabilitation of a patient undergoing an amputation. The knee performs a valuable proprioceptive function and provides a more even circumferential support than an AKA, with 30% of the weight borne by each tibial condyle and 40% by the patellar ten- don. In addition, the below-knee prosthesis is easier to use and requires less energy expenditure to ambulate on than does an above-knee prosthesis. 2 However, if this procedure fails, the patient must be subjected to the inherent risks of a second major surgical procedure in addition to the prolonged hospitalization associated with two major operations. Therefore, various preoperative studies have been investigated to improve on the clinical assessment of potential healing. In evaluating these noninvasive examinations, emphasis is given to the prediction of a successful BKA. It is unlikely that we would primarily perform an amputation at a higher level than an AKA on the basis of a noninvasive test. Doppler ultrasound has proved to be a rapid noninvasive means of initially evaluating arterial occlusive disease. Early in the application of this technique to assess amputation healing, Barnes et al.s found an association between ankle pressures and the success of a BKA. In 1980, Pollock and Ernst 9 concluded that a pressure of 55 mm Hg at the knee or an anldebrachial ratio of 0.3 would accurately predict healing of amputations below the knee. Similar parameters have been reported by Nicholas et al.10 Robinson, H and Lep~intalo et a132 Recent studies of segmental Doppler pressures have not been so favorable. Malone et al.,a3 using a series on noninvasive tests, found that neither popliteal artery pressure nor anldebrachial index was reliable in predicting healing. Our experience likewise showed that Doppler pressures and ratios were unable to separate BKA amputation failures from those that healed. The unreliability of

6 708 Wagner et al. Journal of VASCULAR SURGERY Table III. Sensitivity, specificity, and accuracy of noninvasive tests for below-knee amputation Test Sensitivity* Specificityi- Accuracy~ Calf pressure (>40 mm Hg) 94% (49/51) 25% (2/8) 85% (51/60) Calf-arm index (>0.25) 94% (49/51) 25% (2/8) 85% (51/60) Skin temperature (>90 F) 98% (49/50) 71% (5/7) 95% (54/57) CalftcPo2 (>16 mm Hg) 94% (51/54) 70% (7/10) 91% (58/64) Fluorescein test 98% (55/56) 20% (2/10) 86% (57/66) The actual number of tests is given in parentheses. True positive (TP) 100. *Sensitivity = TP + false negative (FN) True negative (TN) 100. fspecificity = TN + false positive (FP) TP +TN :~Accuracy = TP + TN + FP + FN 100 Table IV. Predictive value of clinical judgment and noninvasive tests for below-knee amputation Test Positive test* Negative test~ Clinical judgment 85% (56/66) NA Calf pressure (>40 mm Hg) 89% (49/55) 40% (2/5) Calf/arm index (>0.25) 89% (49/55) 40% (2/5) Skin temperature (>90 F) 96% (49/51) 83% (5/6) Calf tcpo2 (>16 mm Hg) 94% (51/54) 70% (7/10) Fluorescein test 87% (55/63) 67% (2/3) The actual number of tests is given in parentheses. NA, This value is unavailable because all patients who were predicted to have a failed BKA by clinical judgment underwent AKA. True positive (TP) 100. * Positive predictive value = TP + false positive tnegative predictive value = True negative (TN) x 100. TN + false negative Doppler pressures may be related to the arterial medial calcinosis found in diabetic vessels. This significantly reduces vessel compliance and the ability to occlude the artery with an external cuff, thereby spuriously raising the derived arterial pressure. However, in our series as well as in studies by others, ls'14 segmental pressures were no more valuable in nondiabetic than in diabetic patients. Tactile assessment of skin temperature has long been an element of the clinician's armamentarium in predicting amputation healing. Regional skin perfusion is related to thermal gradients, as heat flow to the skin surface is primarily a convective process determined by blood flow. *s Therefore, thermometry or thermography has the potential for providing an inexpensive, noninvasive, quantitative representation of skin blood flow. Using an infrared technique, Henderson and Hacketd 6 concluded that thermography could have improved the prediction of am- putation level in one third of their cases. Spence et al.17 also believed that infrared thermography was highly predictive, although their success rate of 80% is comparable to that achieved by clinical impression alone. Direct skin temperature measurement, thermometry, to predict amputation healing has been shown previously to correlate well with amputation healing. In a series of 30 limbs, all patients with temperature readings above 32 C healed and all those with readings below 30.5 C failed. 17 Seventy-eight percent of limbs with skin temperatures between these values healed. A skin temperature 5 C or more above ambient temperature was also associated with primary healing of an amputation in 25 of 26 cases (96%). TM In our own experience, using a cutoff of 90 F, skin temperature demonstrated the best combination of sensitivity, specificity, accuracy, and predictive value of all the noninvasive techniques evaluated. Comparison of skin to ambient temperatures and exclusion of limbs with local inflammatory changes could further improve results. Fluorescein is an organic dye that, if administered systemically, can be detected in well-vascularized tissue by its green-yellow fluorescence under ultraviolet light. Fluorescein was first used to assess amputation level by Lange and Boyd 19 in Their technique involved bulky electronic equipment and was not widely accepted. In 1982, Tanzer and Home 2 reported a retrospective, comparative analysis of fluorescein angiography in predicting outcome of 27 lower extremity amputations performed on the basis of clinical assessment. Use of fluorescein as the determinant of level would have decreased the failure rate from 33% to 15%. Conversely, 11% of amputations would have been unnecessarily high on the basis of fluorescein perfusion. McFarland and

7 Volume 8 Number 6 December 1988 Determination of lower extremity amputation healing 709 Lawrence 21 found that qualitative skin fluorescence correctly predicted healing in 9 of 11 BKAs (82%). We found that qualitative evaluation of fluorescein angiography is limited by its poor specificity (20%). The ability to correctly predict BKA healing with fluorescein (positive predictive value) was not significantly greater than that with clinical judgment alone (Table W). Silverman et al have popularized quantitative, fiber-optic fluorometry as a means of predicting healing of an amputation. Besides requiring a lower dose of fluorescein than the qualitative method, it is more reliable in black patients 2s and the results are easily reproducible. We have recently completed an evaluation of this technique in a separate group of amputation patients, and this will form the basis for a future report. The use of tcp02 monitoring to assess skin blood flow has found increasing application over the last decade. It provides an easily reproducible, physiologic correlate of the metabolic balance between oxygen delivery and utilization. Transcutaneous P02 electrodes have been used to assess skin flap viability, 26 success of limb reimplantation, 27 and severity of peripheral vascular disease. 283 It is particularly useful as a noninvasive discriminator in diabetic patients where Doppler pressures may be spuriously elevated.sl's2 Many investigators have used tcp02 in the evaluation of patients undergoing amputation. Is's '3s-s7 Recently, Malone et al. ~s observed that a cutoff of 20 mm Hg provided 100% accuracy in the determination of level of lower limb amputation. Previous authors have not reported this degree of accuracy, with each study showing some degree of overlap of tcp02 values for successful and unsuccessful amputations. In addition, the cutoffs for healing varied between 10 ss and 40 mm Hg? In our experience, with a cutoff of 16 mm Hg, tcp02 was superior to clinical assessment in predicting healing at the belowknee level. However, before tcp02 can be recommended for general use, standardization of cutoffs must be accomplished. Some investigators 3s believe that the rate of change in tcpo2 with supplemental oxygen inhalation will prove to be more reliable than an absolute value. An inherent flaw in all studies of preoperative determination of healing involves the performance of AKAs on the basis of some defined criterion, such as clinical assessment. Therefore the possibility exists that unnecessary AKAs were performed that might have been appropriately identified by the tests being investigated as healing at a below-knee site. The ideal method to evaluate the prediction of BKA healing by any noninvasive test would necessitate the performance of an amputation at the below-knee level in all patients requiring major lower extremity amputation. It is unlikely that such a human study will be conducted. Another variable in evaluating noninvasive predictors of amputation is the reality that wound healing is multifactorial, and failure may be related to factors other than blood flow (e.g., nutritional status, operative technique, postoperative care, infection, or stunap trauma). Considering these limitations, we found that measurement of skin temperature and tcpo~ at the midcalf level were the most accurate predictors of BKA healing. However, with the optimal cutoff points that we established for skin temperature and tcpo2, these tests incorrectly predicted the failure of a BKA in 17% and 30% of patients, respectively. Thus a significant number of patients would have received unnecessary AKAs had the level been determined solely on the basis of these studies. Therefore noninvasive tests may improve, but not supplant, clinical judgment in the preoperative evaluation of patients undergoing BKA. REFERENCES 1. Castronuovo JJ Jr, Deane LM, Deterling RA Jr, O'Donnell TF Jr, O'Toole DM, Callow AD. Below-knee amputation: is the effort to preserve the knee joint justified? Arch Surg 1980;115: Couch NP, David JK, Tilney NL, Crane C. Natural history of the leg amputee. Am J Surg 1977;133: Robinson KP. Long posterior flap amputation in geriatric patients with ischaemic disease. Ann R CoIl Surg Engl 1976;58: De Cossart L, Randall P, Turner P, Marcuson RW, The fate of the below knee amputee. Ann R CoIl Surg Engl 1983;65: Keagy BA, Schwartz JA, Kotb M, Burnham SJ, Johnson G Jr. Lower extremiv amputation: the control series. J VASC SURG 1986;4: Sumner DS. Evaluation of noninvasive testing procedures: data analysis and interpretation. In: Bemstein EF, ed. Noninvasive diagnostic techniques in vascular disease. St. Louis: CV Mosby, 1985: Metz CE. Basic principles of ROC analysis. Semin Nucl Med 1978;8: Barnes RW, Shanik GD, Slaymaker EE. An index of healing in below-knee amputation: leg blood pressure by Doppler ultrasound. Surgery 1976;79: Pollock SB, Ernst CB. Use of Doppler pressure measurements in predicting success of amputation of the leg. Am I Surg 1980;139: Nicholas GG, Meycrs JL, Demuth WE. The role of vascular laboratory criteria in the selection of patients for lower extremiw amputation. Ann Surg 1982;195: Robinson K. Amputation in vascular disease. Ann R CoIl Surg Engl 1980;62:87-91.

8 710 Wagner et al. Journal of VASCULAR SURGERY 12. Lep'antalo MJ, Haajanen J, Lindfors O, Paavolainen P, Scheinin TM. Predictive value of preoperative segmental blood pressure measurements in below-knee amputations. Acta Chir Scand 1982;148: Malone JM, Anderson GG, Lalka SG, et al. Prospective comparison of noninvasive techniques for amputation level selection. Am J Surg 1987;154: Cederberg PA, Prithard DJ, Joyce JW. Doppler-derived segmental pressures and wound-healing in amputations for vascular disease. J Bone Joint Surg (Am) 1983;65: Spence VA, Walker WF. The relationship between temperature isotherms and skin blood flow in the ischemic limb. J Surg Res 1984;36: Henderson HP, Hackett MEJ. The value of thermography in peripheral vascular disease, Angiology 1978;29:65: Spence VA, Walker WF, Troup IM, Murdoch G. Amputation of the ischemic limb: selection of the optimum site by thermography. Angiology 1981;32: Golbranson FL, Yu EC, Gelberman RH. The use of skin temperature determinations in lower extremity amputation level selection. Foot Ankle 1982;3: Lange K, Boyd LJ. Use of fluorescein method in establishment of diagnosis and prognosis of peripheral vascular diseases. Arch Intern Med 1944;74: Tanzer TL, Home JG. The assessment of skin viability using fluorescein angiography prior to amputation. J Bone Joint Surg [Am] 1982;64: McFarland DC, Lawrence PF. Skin fluorescence, a method to predict amputation site healing, J Surg Res 1982;32: Silverman DG, Roberts A, ReiUy CA, et al. Fluorometric quantitation of low-dose fluorescein delivery to predict amputation site healing. Surgery 1987;101: Silverman DG, Rubin SM, ReiUy CA, et al. Fluorometric prediction of successful amputation level in the ischemic limb. J Rehabil Res Dev 1985;22: Silverman DG, Wagner FW, Prediction of leg viability and amputation level by ftuorescein uptake. Prosthet Orthot Int 1983;7: Brousseau DA, Klein SG, Weinstock BA, Silverman DG. Fluorometric assessment of perfusion in multicolored skin. Surg Forum 1983;34: Achauer BM, Kirby SB, Litke DK. Transcutaneous Po2 in flaps: a new method of survival prediction. Plast Reconstr Surg 1980;65: Matsen FA, Bach AW, Wyss CR, Simmons CW. Transcutaneous Poz: a potential monitor of the status of replanted limb parts. Plast Reconstr Surg 1980;65: Byme P, Provan JL, Ameli FM, Jones DP. The use oftranscutaneous oxygen tension measurements in the diagnosis of peripheral vascular insufficiency. Ann Surg 1984;200: Hauser CJ, Appel P, Shoemaker WC. Pathophysiologic classification of peripheral vascular disease by positional changes in regional transcutaneous oxygen tension. Surgery 1984;95: White R.A, Nolan L, Harley D, et al. Noninvasive evaluation of peripheral vascular disease using transcutaneous oxygen tension. Am J Surg 1982;144: Hauser CJ, Klein SR, Mehringer M, Appel P, Shoemaker WC. Superiority of transcutaneous oximetry in noninvasive vascular diagnosis in patients with diabetes. Arch Surg 1984;119: Wyss CR, Matsen FA, Simmons CW, Burgess EM. Transcutaneous oxygen tension measurements on limbs of diabetic and nondiabetic patients with peripheral vascular disease. Surgery 1984;95: Burgess EM, Matsen FA, Wyss CR, Simmons CW. Segmental transcutaneous measurements of Po2 in patients requiring below-the-knee amputation for peripheral vascular insufficiency. J Bone Joint Surg (Am) 1982;64: Franzeck UK, Talke P, Bemstein EF, Golbranson FL, Fronek A. Transcutaneous Po2 measurements in health and peripheral arterial occlusive disease. Surgery 1982;91: Karanfilian RG, Lynch TG, Zirul VT, et al. The value of laser Doppler velocimetry and transcutaneous oxygen tension determination in predicting healing of ischemic forefoot ulcerations and amputations in diabetic and nondiabetic patients. J VASC SURG 1986;4: Katsamouris A, Brewster DC, Megerman J, et al. Transcutaneous oxygen tension in selection of amputation level. Am J Surg 1984;147: Mustapha NM, Redhead RG, Jain SK, Wielogorski JWJ. Transcutaneous partial oxygen pressure assessment of the iscbemic lower limb. Surg Gynecol Obstet 1983;156: McCollum PT, Spence VA, Walker WF. Oxygen inhalation induced changes in the skin as measured by transcutaneous oxymetry. Br J Surg 1986;73:882-5.