Hemodynamic and metabolic responses to graded microvascular occlusion

Hemodynamic and metabolic responses to graded microvascular occlusion Edward M. Kwasnik, MD, Samer Y. Siouffi, MD, Philip T. Lavin, PhD, and

Shukri F. Khuri, MD, Boston, Mass. The hemodynamic and metabolic consequences of microvascular occlusion, often present in the runoff bed of distal arterial reconstructions, have been difficult to quantitate clinically. To investigate these pathophysiologlc relationships, a porcine hindlimb model was developed in which arteriolar patency, which we term outflow capacity, may be quantitatively defined and reduced by serial distal microembolization with 70 ~m flow-directed glass bubbles. In 10 anesthetized adult pigs, hindlimb perfusion was limited to femoral artery flow (FAF) by collateral ligation. Serial measurements of outflow resistance (OK), femoral artery flow, and resting muscle pH (mpH), a metabolic index of tissue perfusion, were made as relative outflow capacity (KOC) underwent graded reduction from 1.0 (baseline) to 0 (complete occlusion). Femoral artery flow decreased linearly (FAF = 87 KOC - 3), and outflow resistance increased in hyperbolic fashion (OR = 1.66/KOC) in response to graded peripheral microembolization, whereas resting muscle pH followed a more complex relationship (In mpH = 0.055 ROC + 1.95). An integrated analysis of these results suggests that a 50% to 60% reduction in arteriolar patency represents a critical point beyond which outflow resistance rises rapidly and hindlimb flow decreases to levels that are inadequate to support the metabolic demands of resting tissues. (J VAsc SURG 1991;13:867-75.)

Three decades ago angiographic and histologic studies reported by Edwards ~ as well as Strandness et al. 2 revealed extensive occlusion of small arteries and arterioles in limbs amputated from diabetic and nondiabetic patients with severe peripheral atherosclerosis. Using injection cast techniques in amputation specimens, Conrad 3 subsequently demonstrated that up to 84% of the arterioles in ulcerated digits were occluded, and the greatest number of these occlusions occurred in vessels that measured 31 to 80 ~m in diameter. Advances in arterial reconstructive surgery have now made revascularization technically feasible in many such limbs with decreased arteriolar patency, a parameter we have termed the outflow capacity of the arterial runoff bed. These observations have led us to postulate that variations in microvascular patency may increase hemodynamic outflow resistance (OR) and limit nutrient flow to the peripheral tissues. From the Departments of Surgery, Brockton-West Roxbury Veterans Administration MedicalCenter, Brigham and Women's Hospital and Harvard MedicalSchool, Boston. Supported in part by the Richard Warren SurgicalResearchand Education Fund. Presentedat the SeventeenthAnnualMeetingof the New England Societyfor VascularSurgery,Newport,R.I. Sept. 13-14, 1990. Reprint requests: EdwardM. Kwasnik,MD, 134 GrandviewAve., Waterbury, CT 06708. 24/6/28989

Quantitative relationships between outflow capacity, resistance, and tissue perfusion have been difficult to establish in a clinical setting, since arteriolar patency cannot be objectively determined in human subjects undergoing lower extremity revascularization. Furthermore, experimental models of limb ischemia, which have been useful in studying the effects of inflow reduction or reperfusion injury, have provided little information regarding the hemodynamic or metabolic responses to outflow reduction. 4 To test our hypothesis, and to investigate these relationships, a porcine hindlimb model was developed in which outflow capacity was quantitatively reduced by sequential peripheral microembolization, hemodynamic parameters were assessed, and on-line measurement of resting muscle pH was used as a metabolic index of the adequacy of tissue perfusion. This report elucidates the role of microvascular occlusion in the pathophysiology of lower extremity ischemia by defining in a new experimental model relationships between outflow capacity, resistance, and tissue perfusion. MATERIAL A N D M E T H O D S Experimental preparation Ten adult pigs (30 to 40 kg) were premedicated with ketamine (10 mg/kg), and anesthesia was induced with sodium pentobarbitol (15 mg/kg). 867

868

Journal of VASCULAR SURGERY

Kwasnik et al.

Anesthesia was maintained with continuous drip nembutal (45 mg/l 0.9% sodium chloride [NaCI]) initially administered at 125 ml/hr and adjusted as necessary to produce adequate anesthesia. Mechanical ventilation was instituted via tracheostomy. Respirator settings were modified on the basis of arterial blood gas samples to keep arterial Po 2 < 100 mm Hg, Pco 2 at 35 to 45 mm Hg, and arterial p H between 7.35 and 7.45. Cannulation of the carotid artery and jugular vein was performed for arterial and central venous pressure monitoring. Normal saline was administered intravenously at a rate sufficient to keep mean systemic arterial and central venous pressures near baseline levels. Core temperature was monitored by rectal thermometer and maintained above 36 ° C with a heating blanket. The essential features of the experimental preparation are illustrated in Fig. 1. The femoral vessels were exposed, and fine plastic catheters were inserted proximally into the lateral circumflex branches of the femoral artery and vein so as not to obstruct flow in the main vessels. Mean femoral artery pressure was measured by a Statham P-23D (Viggo Spectramed Inc., Critical Care Div. Oxnard, Calif.) strain gauge. The deep femoral artery was ligated distally, and a plastic cannula was advanced proximally toward the common femoral artery for perfusion purposes. A 2.5 mm Doppler flow probe placed around the superficial femoral artery measured mean femoral blood flow and was calibrated by heparinized blood perfusion at known flow rates through the deep femoral cannula while temporarily occluding the proximal common femoral artery. After completion of the dissection, systemic anticoagulation was produced by intravenous injection of 3000 units ofheparin, and a single 30 mg dose of papaverine was injected distally to dilate the peripheral vascular bed. Pressure and flow were continuously recorded on a Gould (Gould Inc., Test & Measurement Recording Systems Division, Cleveland, Ohio) multichannel recorder.

Resting muscle pH measurement Tissue p H electrodes originally developed for the study of myocardial metabolism as well as temperature probes (Monotherm, Inc., St. Louis, Mo.) were placed into the musculi vastus medialis for continuous monitoring of muscle p H and temperature. 5'6 The i mm sensing tip of the electrode consisted of an Ag/AgCI (silver/silver chloride) wire endosed in lead glass. Reference electrodes were placed in a 3 mol/L KCI (potassium chloride) solution, and connected to the subcutaneous tissue in the abdomen via an agar bridge. The electrodes, which have a stabilization period of 3 to 5 minutes, and a 95% in vitro response

time of 3 to 4 seconds, were connected by a coaxial cable to a Coming model 610A (Coming Inc., Coming, N.Y.) voltmeter set in the millivolt mode. Voltmeter output was recorded continuously on a two-channel recorder (Soltec, Inc., Sun Valley, Calif.). The millivolt recordings and the temperatures were fed into a computer programmed to calculate p H using the Nernst equation.7 Each electrode was calibrated in p H 7.0 and 4.0 buffers before and at the end of each experiment. Data from seven experiments in which a drift of less than 0.01 p H unit was observed between the two calibrations were processed for further analysis. Experimental protocol After allowing the preparation to stabilize, absence of significant collateral flow into the hindlimb was demonstrated by observing an immediate fall in muscle p H in response to temporary occlusion of the common femoral artery, The preparation was allowed to return to baseline values of all parameters before beginning graded microvascular occlusion, which was carried out by the following method of reduction and quantification of outflow capacity: Glass bubbles (3M Co., St. Paul, Minn.) with average diameter of 70 ~m were used to produce graded outflow occlusion. Bubbles were serially injected in 0.05 cc aliquots through the deep femoral cannula while temporarily occluding the proximal common femoral artery. The time interval between microsphere injections varied between 10 and 15 minutes depending on the time required for the measured parameters to stabilize after injection. Delivery of the bubbles into the runoff bed by this method was documented in preliminary studies by serial arteriograms. When mean flow measured in the femoral artery reached zero, total occlusion of all arterioles in the outflow tract was assumed. The total volume of bubbles required to produce complete occlusion varied between 0.8 and 1.0 cc. During the mieroembolization process, it was assumed that each aliquot contained approximately the same number of glass bubbles and therefore occluded the same number of arterioles. If the baseline outflow capacity before microembolization is defined as unity, the fractional relative outflow capacity (ROC) remaining patent after each injection may be calculated as follows: Cum. vol. of bubbles to that point ROC = 1 - Vol. required to produce tot. occl. (1) Cum. = Cumulative; occl. = occlusion.

vol. = volume;

tot. = total;

Volume 13 Number 6 June 1991

Response to microvascular occlusion

869

Anwilll I"

.1

Doppler

~km,~m,, PX orotm

Injectio

L.i

Fig. 1. Schematic representation of experimental preparation. L.F.C.A., Lateral femoral circumflex artery; P.F.A., profunda femoris artery (deep femoral); S.F.A., superficial femoral artery.

Mean femoral pressure and flow were recorded after each aliquot and used to calculate resistance, expressed in peripheral resistance u n i t s mmHg/ml/min), by the equation Resistance =

Mean femoral artery pressure (FAP) Mean femoral artery flow (FAF) (2)

Tissue p H as well as these hemodynamic parameters were recorded and allowed to stabilize before proceeding with further microembolization. These investigations were reviewed and approved by the Animal Studies Committee of the Brockton-West Roxbury Veterans Administration Medical Center. They were carried out over a 3-month period of time in an AAALAC approved laboratory in accordance with the "Principles of Laboratory Animal Care" and the "Guide for the Care and Use of Laboratory Animals" (NIH Publication N. 80-23, revised 1985). All animals were killed in a humane manner with bolus injection of KCI at the completion of the experiment.

Data analysis Values of R O C were rounded to the nearest tenth to allow grouping for correlation with FAP, FAF, OR, and muscle pH. Preliminary analysis of data was performed by plotting mean (_+ SEM) FAP, FAF,

OR, and muscle pH data versus ROC. For each experiment, a linear regression equation was constructed for the responses of FAP and FAF beyond baseline values to changes in ROC. The mean values of slopes and intercepts were then calculated to compute linear regression relationships for FAP versus R O C and FAF versus ROC. Confidence limits for the mean slope were calculated, and to demonstrate significance, means were compared by t test to the null hypothesis (mean slope = 0). Paired t tests were used to test the significance of differences in values of each parameter over several intervals of R O C and were accepted at the p < 0.05 level. A similar analysis was carried out for muscle p H after examination of the preliminary relationship between muscle p H and R O C suggested a linear regression of In p H versus ROC. The hyperbolic relationship suggested between OR, a derived variable defined as the quotient of pressure divided by flow, and R O C was explored and tested by use of a logarithmic transformation that was also used to estimate the hyperbolic shape parameter via maximum likelihood. RESULTS Pressure versus outflow capacity Changes in mean FAP as outflow capacity was reduced are demonstrated in Fig. 2. At baseline

Journal of VASCULAR SURGERY

870 Kwasnik et al.

150 1 40 130 120 110 100 90

~ 80

I

uJ lO

I

p<.01

~ 60 ~ 50 40 30 20

10. 0

.]

.i3

.] RELATIVE

OUTFLOW

.]

6

CAPACITY

Fig. 2. Relationship of femoral artery pressure (FAP) to relative outflow capacity (ROC) (FAP = 139 - 27 ROC).

( R O C = 1), mean FAP was 124 + 4 m m H g and decreased to 105 + 9 m m H g as outflow capacity was reduced to 0.9. Pressure then rose in response to further reduction in R O C as described by the linear regression model equation FAP = 139 - 27 ROC, in which 95% confidence limits for the mean slope were 15.7 to 38.2 (p < 0.001), thereby refuting the null hypothesis that slope = 0. The only significant rise within the intervals tested was between R O C = 0.5 and R O C = 0.2 (p < 0.01). F l o w versus o u t f l o w capacity Mean FAF was 103 _+ 6 ml/min at R O C = 1 and decreased to 80 + 6 ml/min at R O C = 0.9. After this point, FAF decreased linearly in response to further reductions in R O C as depicted in Fig. 3. This relationship is described by the equation FAF = 87 R O C - 3 , in which 95% confidence limits for the m e a n slope were 73.9 to 99.6 (p < 0.0001),. once again refuting the null hypothesis. Decreases in flow were significant (p < 0.001) for all intervals tested (Fig. 3). Resistance versus o u t f l o w capacity Outflow resistance was calculated by use o f equation 2 and plotted versus R O C in Fig. 4. From a value o f 1.2 _+ 0.07 peripheral resistance unit at baseline, increases in O R that occurred in response to reductions in R O C were best described by the

relationship O R = 1.66/ROC, in which 95% confidence limits for the mean constant were 1.34 to 1.98 (p < 0.001). Resistance increases for each interval tested (Fig. 4) were significant (p < 0.001). p H versus o u t f l o w capacity The relationship between p H and R O C is presented in Fig. 5 and was best described by In p H -- 0.055 R O C + 1.95. The 95% confidence limits for this slope were 0.036 to 0.074 (p < 0.001), once again refuting the null hypothesis. The fall in p H in the interval R O C = 0.5 to R O C = 0.2 was significant (p < 0.01). DISCUSSION

Atherosclerosis, microembolization, or thrombosis distal to occlusive lesions may, to a varying degree, reduce the microvascular patency or outflow capacity, o f the arterial runoff bed o f patients undergoing lower extremity revascularization. Since objective assessment o f the physiologic significance o f reduced microvascular patency has been difficult to achieve in a clinical setting, mechanisms that may potentially influence the pathophysiologic role o f the arteriolar vessels in limb loss and graft failure have not been established. Using a new experimental model o f graded outflow reduction, we have demonstrated that beyond a critical degree o f microvascular occlusion, O R rapidly increases and tissue perfusion falls to

Volume 13 Number 6

Response to microvascular occlusion 871

June 1991

120 ¢K.001 110

t~O01

p<.OOt

if

'

ii

i

10o t

9O 80

~7o ~ 6O

~ so 4O 30 20

10 RELATIVE

OUTFLOW

CAPACITY

Fig. 3. Relationship of femoral artery flow (FAF) to relative outflow capacity (ROC) (FAF = 87 ROC - 3). 20 19 18

p<.OOl

p<.OOl

i1

•

17

16 15 14 13

7, i10 .g 9

~8

l 6 5. 4 3 2, 1 0 RELATIVE

OUTFLOW

CAPACITY

Fig. 4. Relationship of outflow resistance (OR) to relative outflow capacity (ROC) (OR = 1.66/ROC). levels that are inadequate to meet the demands of resting tissues. Several features of the experimental preparation deserve further comment. First, the arterial blood supply of the porcine hind/imb was limited to inflow by way of the femoral artery by ligation of major collateral branches. In this way the influence of

potential collateral pathways was minimized as evidenced by an observed fall in tissue p H during temporary femoral artery occlusion. Second, perfusion of the femoral artery by way of the deep femoral branch at known flow rates during temporary proximal occlusion allowed accurate calibration of the flow probe and compensated for any discrepancies

872

Journal of VASCULAR SURGERY

Kwasnik et al.

pH r.5

7.45. 7.4. t 7.35' 7.3'

7.25, 7.2. 7.15 /.1.

7.05' 7"

6.95' 6.9" 6.856.8,

.h RELATIVE

.] OUTFLOW

.'1

6

CAPACITY

Fig. 5. Relationship of muscle pH (mpH) to relative outflow capacity (ROC) (In mpH = 0.055 ROC + 1.95). The logarithmic relationship between pH and relative outflow capacity was plotted on a linear scale as the range of pH response was sufficientlynarrow. between probe and vessel size. Finally, the tissue pH electrode used provided continuous on-line measurement of sequential variations in muscle pH as microvascular occlusion progressed, s'6 Although microembolization techniques have been used in the investigation of the myocardial and pulmonary microcirculations,s'9 they have not been applied in the study of cutaneous, subcutaneous, and muscular vascular beds of the extremity. To maximize the number of vessels available for microembolization, a state of relative vasodilation was achieved by the administration of papaverine before microembolization and by the use of continuous pentobarbitol anesthesia, which has also been shown to produce peripheral vasodilation. ~0 Furthermore, in this preparation microembolization was produced by glass bubbles rather than solid beads. These particles were therefore flow directed to seek patent arteriolar channels, thereby reducing the chance of multiple embolization of a single arteriole. Experience with microembolization in the pulmonary vascular bed also suggests that occlusion of vessels with a single embolic particle is greater with injection of small volumes of spheres. 9 In addition, the 70 ~m size would select fourth order arterioles and yet be too large to pass into or through the capillary bed. Certain limitations o f the preparation must also be noted. First, the influence of autoregulatory mechanisms was eliminated by pharmacologic vasodilation. Second, in this acute preparation, the time

course of outflow occlusion was rapid compared to the gradual occlusion that occurs as a consequence of atherosclerosis. Consequently, adaptive changes that may occur clinically at the microcirculatory level may not have had sufficient time to occur. Furthermore, the process is embolic, rather than occlusive or thrombotic, and is produced by foreign material. Nevertheless, from an anatomic point of view, the outcome is the same with loss of microcirculatory patency. Finally, microembolization studies in other tissues have demonstrated that emboli may produce partial occlusion of an arteriole while allowing partial flow to continue, s However, we have assumed that the occurrence of partial embolization as well as occlusion of a single channel by multiple glass bubbles to be similar for each aliquot injected, thereby maintaining the quantitative aspects of the microembolization process. The results of these studies describe the hemodynamic responses to graded outflow reduction. Mean perfusion pressure was maintained within a physiologic range during the course of outflow reduction, although femoral pressure was observed to rise slightly with progressive arteriolar occlusion. This increase may be attributed to the dissipation of the papaverine effect, but also suggests that microembolization may induce some degree of peripheral vasoconstriction. In pulmonary microembolization studies Shirai et al. 9 noted that vasoconstriction occurred upstream to plugs in small pulmonary

Volume 13 Number 6 June 1991

arteries. As a result, blood flow was redistributed from embolized to nonembolized arteries at branch sites. Proximal vasoconstriction of occluded arterioles therefore may function as a local autoregulatory mechanism that operates to maximize perfusion of patent channels and, by increasing local perfusion pressure, to recruit additional areas of the distal vascular bed. Despite the maintenance of adequate perfusion pressure, flow decreased in direct proportion to reduction of arteriolar patency. This observation confirmed the validity of the experimental hypothesis that sequential microembolization produced equivalent reduction in relative microvascular patency. At an R O C of 0.5 therefore, flow was reduced to less than one half of baseline levels. Since the experiment was performed in the resting state with pharmacologic ablation of autoregulatory mechanisms, tissue pH, which reflects the balance between oxygen supply and demand, was used as an index of tissue perfusion. After reduction of outflow capacity by approximately 50% to 60%, mean tissue p H fell to 7.31, reflecting a local increase in tissue hydrogen ion concentration and hence a metabolic imbalance between oxygen supply and demand. Therefore this degree of reduction in outflow capacity in the experimental model appeared to be a critical point beyond which further loss of patent arteriolar channels led to rapid decreases in blood flow and further alteration of normal tissue metabolism. Resistance, or the ratio of perfusion pressure to flow, is a particularly useful parameter for describing the dimensions of a biologic tube system such as the distal vascular bed whose individual members are inaccessible for direct measurement. The relationship between resistance and outflow capacity demonstrated in these studies forms a conceptual framework that best defines the physiologic consequences of arteriolar occlusion. The basic form of this hyperbolic relationship y = c/x, where y is resistance, x is ROC, and c is the proportionality constant of a given vascular bed, may be derived from a consideration of the peripheral vascular bed as a system of a fixed number of parallel arteriolar channels of approximately equal length and diameter. Since the distribution of resistance over the vasculature is unknown, such a parallel model is an acceptable alternative to the more realistic dichotomously branching tree model and has been used in microembolization studies of the coronary vascular bed. 8 In this model the contribution of the conducting vessels proximal to the site of embolization to O R is assumed to be low, and the maximal resistance is located at or distal

Response to microvascular occlusion 873

to the site of embolization at the arteriolar level. Based on this arrangement, the contribution of each arteriole to the total resistance of the vascular bed may be calculated by the familiar formula: 1

1

1

KTot~ = G + ...... +K1

(3)

where n is equal to the total number of channels in the system. If at this point it is assumed that the resistance of each arteriole is approximately equal, since each is exposed to the same metabolic and pharmacologic milieu, then the above relationship reduces to the following: 1

1

RTotal

=

n x P'~ndivla~aunit

(4)

which rearranges to: R T o t a l _ Rindividual unit

(5)

n If at this point it is further assumed that the resistance of each unit remains constant, resistance of the entire bed (RTot~) becomes a dependent variable that may be related to changes in the independent variable of the number of units of which it is comprised. Since the concept of ROC allows the number of patent outflow channels after graded arteriolar occlusion to be related to the baseline state in which all channels are patent (ROC = 1), equation 5 may be modified and used to predict resistance changes from baseline resistance of the system in response to relative reductions in outflow capacity. This new relationship therefore becomes the following: Resistance constant RTot~, = Relative outflow capacity

(6)

Since the relationship derived experimentally is virtually identical to the analytic solution presented above, the assumptions made in this analysis as well as the quantitative aspects of the microembolization process appear to be validated. The significance of arteriolar occlusive changes as a factor in the pathogenesis and treatment of lower extremity ischemia has not been demonstrated clinically. However, the hemodynamic and metabolic derangements produced by arteriolar occlusion in this preparation have potential clinical implications, particularly in the setting of acute ischemia due to microembolization or macroembolization in which loss of arteriolar patency may influence the degree of tissue ischemia that persists after proximal flow is

874

Journal of VASCULAR SURGERY

Kwasnik et al.

restored. A better understanding of changes in the microcirculatory environment near the point of critical reductions in outflow capacity may ultimately suggest hemorheologic or pharmacologic methods of modifying the consequences of OR on arterial reconstructive procedures. Application of the experimental results to reconstructions performed for chronic atherosclerosis are most appropriately limited to distal tibial, peroneal, and isolated popliteal bypass grafts that most closely resemble the experimental preparation in that their collateral supply is relatively fixed and limited, ll As the experimental results illustrate, once a critical proportion, perhaps 50% to 60%, of the arteriolar channels of what is essentially an end arterial vascular bed are occluded, further occlusion produces a relatively rapid rise in resistance. Consequently, flow is decreased to levels that may be inadequate to meet the metabolic demands of resting tissues despite adequate perfusion pressures. Areas of impending or established tissue necrosis fail to heal and may necessitate amputation in the setting of a patent graft. Alternatively, flow velocity may decrease to a level that correlates with graft occlusion. 12 The hemodynamic resistance produced by reduced outflow capacity must be added to the fixed resistance of stenotic lesions beyond the distal anastomosis of arterial bypass grafts in determining the total OR. Since these macrovascular lesions may be visualized angiographically, their contribution may be minimized by careful placement of the distal anastomosis. Arteriolar occlusion, however, occurs at a level of tile microcirculation that is beyond the resolution of standard angiographic methods and constitutes a major limitation of angiography in assessing the adequacy of runoff. This factor contributes to the inconsistent accuracy of correlations obtained between angiographic grading of runoff and graft patency or limb salvage. 13-~s Physiologic methods of outflow assessment such as properly performed intraoperative resistance measurements, ~+18 perhaps in combination with on-line tissue pH measurements, may more accurately reflect reductions in outflow capacity and thus better predict the outcome of arterial reconstructive procedures. REFERENCES 1. Edwards EA. Postamputation radiographic evidence for small artery obstruction in arteriosclerosis. Ann Surg 1959;150: 177-87. 2. Strandness DE, Nothstein DL, Alexander JA, Bell JW.

3.

4. 5.

6.

7. 8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

Observations on arteriolar disease in arteriosclerosis obliterans. Surgery 1960;47:953-8. Conrad MC. Large and small vessel occlusion in diabetics and nondiabetics with severe vascular disease. Circulation 1967; 36:83-91. Barie PS, Mullins RJ. Experimental methods in the pathogenesis of limb ischemia. J Surg Res 1988;44:284-307. Khuri SF, Marston W, Josa M, et al. First report of intramyocardialpH in man: I. methodology and initial results. Med Instrumen 1984;18:167-71. Khuri SF, Marston WA. On-line metabolic monitoring of the heart during cardiac surgery. Surg Clin North Am 1985;65: 439-53. Westcott CW. pH measurements. New York: Academic Press, Inc, 1978. Wierenga PA, Stassen HG, Laird JD, Spaan JAE. Quantification of arteriolar density and embolization by microspheres in rat myocardium. Am J Physiol 1988;254(Heart Circ Physiol 23):H636-50. Shirai M, Sada K, Ninomiya I. Diameter and flow velocity changes in small pulmonary vessels due to microembolization. J Appl Physiol 1988;65:288-96. Kazmers A, Whitehouse WM, Zelenock GB, Cronenwett JL, Lindenauer SM, Stanley JC. Early derangements of arteriovenous anastomotic and capillary blood flow in the canine hindlimb induced by supplemental pentobarbitol anesthesia. J Surg Res 1984;36:102-7. Corson JD, Karmody AM, Shah DM, Naraynsingh V, Young HL, Leather RP. In situ vein bypasses to distal tibial and limited outflow tracts for limb salvage. Surgery 1984;96:75663. Bandyk DF, Cato RF, Towne JB. A low flow velocity predicts failure of femoropopliteal and femorotibeal bypass grafts. Surgery 1985;98:799-809. Mundth ED, Darling RC, Moran JM, Buckley MJ, Linton RR, Austen WG. Quantitative correlation of distal arterial outflow and patency of femoropopliteal reversed saphenous vein grafts with intraoperative flow and pressure measurements. Surgery 1969;65:197-206. Menzoian JO, LaMorte WW, Cantelmo NL, Doyle J, Sidaway AN, Savenor A. The preoperative angiogram as a predictor of peripheral vascular runoff. Am J Surg 1985;150: 346-52. Peterkin GA, Manabe S, LaMorte WW, Menzoian. Evaluation of proposed standard reporting system for preoperative angiograms in infrainguinal bypass procedures: angiographic correlates of measured runoff resistance. J VAsc SURG 1988;7:379-85. Ascer E, Veith FJ, Morin L, et al. Components of outflow resistance and their correlation with graft patency in lower extremity arterial reconstructions. J VASCSURG 1984;1:81728. Parvin SD, Evans DH, Bell PRF. Peripheral resistance measurement in the assessment of severe peripheral vascular disease. Br J Surg 1985;72:751-3. Ascer E, White SA, Veith FJ, Morin L, Freeman K, Gupta SK. Outflow resistance measurement during infrainguinal arterial reconstructions: a reliable predictor of limb salvage. Am J Surg 1987;154:185-8.

Submitted Oct. 11, 1990; accepted Feb. 26, 1991.

Volume 13 Number 6 June 1991

DISCUSSION Dr. James O. Menz~ian (Boston, Mass.). Anytime I hear someone talk about ROC and runoff I get excited. It seems to me that three things keep bypass grafts open. Inflow, outflow, and the tube in between. We spend a lot of time talking about that tube in between and about the subsequent inflow procedures, and I am delighted to hear someone talk about runoff. The hypothesis for this very interesting paper was that variations in microvascular patency may influence the outcome of arterial reconstruction procedures, both in terms of graft patency and limb salvage. This model is well controlled with excellent hemodynamic parameters measuring FAP, FAF, on-line measurement of muscle pH as an index of perfusion, and then quantitative reduction in outflow capacity using these very interesting 70 txm glass bubbles. I have a couple of questions that I would like to ask about your findings. In your Fig. 2 in the paper you plot the FAP against a flow capacity and you note generally that there were not very many significant pressure changes except at two points between the relative outflow capacity of 0.5 and 0.2. I am really not too sure that there really is a very different change in pressure. I wonder whether that may be just the papaverine wearing off, because I do not really understand why the FAP would change as you progressively reduce the outflow capacity. I would like to discuss the concept of resistance. We have measurements in patients undergoing femoral popliteal and femoral tibial bypass grafting, and although we like to talk about resistance, I think we must point out that you and I are not measuring resistance, but really just estimating resistance. We measure flow, we measure pressure, and we calculate the resistance mathematically. Our statisticians tell me that doing these types of measurements using multiple paired t tests is dangerous, and they suggest that we use analysis of variance (ANOVA). I challenge one comment that you made that generally poor correlations have been found between angiographic assessment of runoff and measurements of resistance and patency. I think the reason for that is that the traditional assessment of runoff is bad. This concept of 1, 2, 3 vessel runoff is very poor because it limits the understanding of what the runoff is all about. For example, it wrongly assumes that each of the three tibial vessels is equally important, and I am convinced that that is not true. Traditional assessment of runoff ignores the arch in the foot, which is very critical to the true measurement of outflow capacity. We first made these kinds of measurements of flow and changes in resistance in animals, and we

Response to microvascular occlusion 875

did a lot of manipulation with papaverine and neosynephrine and mechanical occlusion, but when we tried to repeat these studies in approximately 110 patients undergoing femoral popliteal bypass grafting, we could not duplicate the resuks. For example, papaverine had absolutely no effect on resistance in the first 25 patients we measured. I think that these types of models that you use beg to be done in animals in whom you can raise the serum cholesterol and induce atherosclerosis. There have been some very interesting findings now that many of the agents that we use to manipulate runoff bed or that have an effect on smooth muscle like nitroglycerine, nitroprusside, and papaverine have no effect with increasing age and serum cholesterol. I think that your types of measurements should be repeated in those animals in whom you could raise the serum cholesterol. Dr. Edward M. Kwasnik. I thank Dr. Menzoian for his thoughtful comments, particularly in light of his many contributions to the investigation of the role of runoff in lower extremity revascularization. With regard to perfusion pressure, we would agree that the purpose of presenting those data was to demonstrate that we maintained perfusion within a physiologic range, as an alternative to a constant perfusion pressure or flow. We believe this was a more physiologic method and would agree that papaverine effect was probably the most significant factor in producing any variations that were observed. With regard to statistics, we had discussed with our statistician the possibility of using an analysis of variance. His recommendation, as we brought out in our section on data analysis, was to use linear regression techniques for each experiment, and then calculate mean regression slopes for the whole series that could be tested against the null hypothesis of slope = 0. We were able to quantitate and mathematically define our results, thereby validating our hypothesis as well as demonstrating the quantitative relationships we sought to achieve as the goal of these studies. We agree that a chronic version of this model perhaps in animals in which atherosclerosis rather than microembolism is the cause of reduced microvascular patency would be useful, particularly in the evaluation of arterial prostheses. Furthermore, a better understanding of changes in the microcirculatory environment near the point of critical reductions and outflow capacity may ultimately suggest hemorheologic or pharmacologic methods, as presented in previous studies today, which may modify the consequences of outflow reduction for arterial reconstructive procedures.