Hemodynamic characteristics of failing infrainguinal in situ vein bypass

Hemodynamic characteristics of failing infrainguinal in situ vein bypass Benjamin B. Chang, MD, Robert P. Leather, MD, FACS, Jeffrey L. Kaufinan, biD, FACS, Anna Marie Kupinski, MS, RVT, Peter W. Leopold, MBBCh, and Dhiraj M. Shah, MD, FACS, Albany, N.Y. The successful follow-up of distal arterial reconstructions for the identification of the failing bypass in the postoperative period hinges on a knowledge of the natural history o f flow characteristics in these reconstructions. Over a 4-year period resting and hyperemic bypass flow, fistula flow, conduit diameter, and distal peak systolic velocity of 350 in situ bypasses were measured serially. B-mode ultrasound imaging of the entire bypass was performed to identify specific stenoses. Measurements were performed 5 to 9 days after operation, every 2 months for the first year, and ever), 6 months thereafter. Distal bypass flow <25 ml/min, a ratio of hyperemic/resting distal bypass flow <2.5, and vein size ---3.0 mm inner diameter all correlated with bypass stenosis (>50%) or occlusion (p < 0.01). Contrary to previous studies, a distal peak systolic velocity of <45 cm/sec did not correlate with bypass stenosis or occlusion. A low distal peak systolic velocity did correlate with bypass stenosis or occlusion in bypasses larger than or equal to 3.5 nun inner diameter (p < 0.03). However, no combination of these factors was able to accurately predict preocclusive stenosis or occlusion. Distal bypass flow was highest initially but reached a plateau 6 to 12 months after operation. Fistula flow, although very- high initially, showed marked decrement with time. (J VAsc SugG 1990;12:596-600,)

In spite of improvements in femoropopliteal and femorotibial vein bypass techniques, late failure remains one of the most common complications of these procedures. Identification and correction of preocclusive lesions ofinfrainguinal vein bypasses allow the surgeon to improve bypass patency at least 10% to 15%. 1'2 However, commonly used methods of noninvasive bypass assessment such as physical examination, pulse volume recording, and segmental Doppler pressures are generally insensitive to imminent bypass failure,s'4 Serial postoperative duplex ultrasonography has been purported to be a sensitive means of assessing bypass function, although the exact criteria for identifying the failing bypass remain to be defined. Use of the distal peak systolic velocity (DPSV) has been proposed as a test for imminent failure.S However, other work has questioned the use of this parameter. 6 More recently the combination of an abnormal DPSV and decrease in ankle/brachial index (ABI) has been proposed as a more accurate determinant of imminent bypass failure. 7 The use of any such criteria depends on an accurate knowledge of the natural history of bypass flow volume and velocity characteristics. Therefore in this study we From the Vascular Surgery Service,AlbanyMedical College. Reprint requests: Benjamin B. Chang, MD, Department of Surgery (A61), AlbanyMedical College,Albany, NY 12208. 24/37/23533

596

have sought to identify more accurate criteria for bypass assessment with duplex ultrasonography. It k, just as important to note that this 4-year study seeL, to define the long-term flow volume and velocit~ characteristics of in situ vein bypasses. MATERIAL AND M E T H O D S Over a 4-year period, serial examinations of 350 in situ vein bypasses were conducted. These bypasses were all performed at a single institution; all patients who were able and willing to return for routine follow-up visits were included. The patient population consisted of 193 men and 114 women; 14% had bilateral reconstructions. Mean age was 69 _+ 11 years. Fifty-four percent of patients were smokers, 55% were diabetic, and 52% were hypertensive. The indication for the original surgery was limb salvage in 91% and claudication in 9%. Thirty-six percent were femoropopliteal bypasses, whereas 64% were femorotibial bypasses. The follow-up protocol involved physical examination, pulse volume recording, segmental Doppler' pressures, and duplex scanning of each bypass. These examinations were conducted 5 to 9 days after operation, every 2 months for the first year, and ever), 6 months thereafter. Bypasses felt to be at higher risk for occlusion were often subjected to more frequent surveillance. ',~ Each duplex examination consisted of simulta-

Volume 12 Number 5 November 1990

Hemodynamics offailing infrainguinal bypass 597

Table I. Changes in bypass and estimated fistula flow over time Postgperative interval 5- 9 2- 4 4- 6 6-12 12-18 >18

Total flow (ml / min)

Distal flow (ml / min)

Fistula flow (ml/ min)

345 257 205 119 103 97

83 75 59 54 49 47

257 180 143 64 48 39

No.

days mo mo mo mo mo

76 ~ 282 216 253 193 109

± ± + ± ± ±

37 36 55 16 14 11

+ ± ± ± ± ±

8 7 4 3 5 4

± + ± ± ± ±

31 34 54 14 11 9

~Because of incisional pain and early discharge from the hospital, measurements at this interval could not be done in all patients.

Table II. Bypass outcome as a function of distal bypass flow

Table III. Bypass outcome as a function of low DPSV

Stenosed or

Distal flow < 2 5 ml/min Distal flow 25-49 m l / m i n Distal flow > 4 9 ml/min

Sg¢nosed or

occluded

Not stenosed

15

46

17

83

15

174

neous B-mode ultrasound imaging by use of a 10 MHz sector scanning probe coupled with a 4.5 MHz pulsed Doppler (Diasonics DRF 400 [Diasonics, Inc., South San Francisco, Calif.]); deeper bypasses required the use of a 7.5 MHz probe coupled to a 4.0 MHz pulsed Doppler. Images were obtained of the entire bypass and the inflow and outflow vessels during each examination. However, these data were not prospectively recorded, because the intent of this study was to define hemodynamic criteria. Distal by,pass flow volume was measured by a previously validated technique involving time-averaged flow velocity and ultrasonically determined conduit diameter. s Hyperemic distal bypass flow was measured after 3 minutes of tourniquet occlusion just below the ~mee. In addition, proximal bypass flow, DPSV conduit diameter, and fistula flow were also recorded. Images were obtained of identifiable bypass stenoses, and the degree of stenosis was estimated by a combination of B-mode imaging and measurement of peak systolic velocity within the lesion. 9 A doubling o f peak systolic velocity was consistent with a greater than 50% diameter reduction stenosis. tLESULTS Results of longitudinal measurement of bypass flow volumes are summarized in Table I. Total bypass ~low was highest shortly after the initial operation. ~ y subtracting the distal bypass flow from the proximal bypass flow fistula flow was estimated. As can

DPSV ->45 cm/sec DPSV -<45 cm/sec

occluded

Not stenosed

26 21

201 102

be seen, distal bypass flow progressively decreased over the first few months ~fter operation but reached a plateau within 4 to 6 months after surgery. Average fistula flow demonstrated a more marked progressive decrement in the first postoperative year. During the study period 303 bypasses remained patent without further surgery. Among these patent bypasses, 35 stenoses were detected and placed under observation because they exhibited less than a 50% diameter reduction and the: patients were asymptomatic. None of these 35 patients' bypasses occluded. Thirty-six of the remaining 47 bypasses in the study underwent revision for tiow-limiting lesions (more than 50% diameter reduction stenosis), and nine bypasses occluded. At the time of revision, the cause of the lesions was found to be hyperplastic changes in the vessel wall. Among the group of occluded bypasses, three were denovo occlusions, and six resulted from stenoses that progressed to occlusion. Table II presents the fate of in situ bypasses with distal bypass flows divided into three classes. As can be seen, there were a large number of bypasses with flow volumes measured to be less than 25 ml/min; occlusion/stenosis was seen significantly more often in this group, although most remained patent. In evaluating measured[ duplex-generated parameters in respect to occlusion/stenosis, a DPSV of less than 45 cm/sec recorded at the time of follow-up was not found to correlate with a negative outcome (Table III). In contrast, a ratio ofhyperemic to resting distal bypass flow of less than 2.5 did significantly correlate with occlusion/stenosis (Table IV). In ad-

598

Journal of VASCULAR SURGERY

Chang et al.

Table IV. Bypass outcome as a function of hyperemia/resting bypass flow

Hyperemic/ resting distal flow >2.5 Hyperemic/ resting distal flow -<2.5

Stenosed or occluded

Not stenosed

15

194

17

62

clifton, a measured conduit inner diameter of less than or equal to 3.0 mm was correlated with failure

(t < 0.05). However, in using the parameters of DPSV <45 cm/sec, hyperemic/resting bypass flow <2.5, distal bypass flow < 2 5 ml/min, and vein inner diameter less than or equal to 3.0 mm, no one factor or combination of factors was suitable for use as screening test(s) because sensitivity and specificity in any combination of parameters was no greater than 20%. In examining the reasons for this, it was found that low DPSV correlated especially poorly in the bypasses with diameters of less than 3.0 mm, which formed a large portion o f this series (Table V). In larger bypasses low D P S V did correlate with bypass failure/occlusion, but again was not useful as a screening test for failing bypasses. In 30 patients in the occlusion/stenosis group a significant D P S V decrease was observed between the initial and final examinations before occlusion or revision (p < 0.01). DISCUSSION The surgical treatment of infrainguinal occlusive disease does not end after the femoropopliteal or femorotibial bypass is completed. Systematic postoperative surveillance can potentially identify preocelusive lesions before actual bypass failure, s'7'1° This is especially important in light of reports documenting the relatively poor results of graft revision after occlusion; in contrast, most preocclusive lesions may be corrected quite easily with lasting patency. 2"::a2 However, the most useful tests for identifying preocelusive lesions remain to be defined. In part this is because the long-term hemodynamic behavior of such reconstructions is only now being elucidated with the help of duplex ultrasound scanning. In the early postoperative period, there is maximal vasodilation of the arterial bed in limb salvage patients; as a result, measured resting distal bypass flow was highest at this time. In addition, there is continual diastolic flow as manifested by an end-diastolic flow velocity greater than zero: Pulse volume re-

Table V. Correlation of DPSV and hyperemia to bypass outcome: effect of conduit diameter Condu# diameter (ID) <-3.0 mm <3.5 mm

DPSV <45 cm/sec Hyperemic/resting flow <2.5

NS NS

p <0.03 p <0.004

cordings performed at this time will show an exaggerated increase in amplitude. As limb ischemia resolves and any tissue breaksdown heals, the resistance of the distal arterial bed returns toward normal. This was manifested by a reduction of the distal bypass flow to an average of 49 ml/min 12 months later; resting distal flow remained relatively stable thereafter. In some bypasses resting distal flow was remarkably low. Although bypasses with resting flow rates less than 25 ml/min were somewhat more likely to occlude, most of these very low flow bypasses maintained long-term patency. It is interesting to note that fistula flow through the iatrogenic arteriovenous fistulas unique to the in situ bypass spontaneously decreased over time; thi~ supports the contention that small, cutaneous fistulas need not be ligated during the course of the opera ~tion. 13 It continues to be our practice to ligate deep. fistulas or deep perforating veins that fill the deep venous system on Doppler examination or completion angiography. Cutaneous branches and small fistulas need not be ligated. In spite of the remarkable ability of vein bypasses in this setting to remain open ha the face of extremely low flow rates and velocities, bypass occlusion in the late postoperative period remains a major problem. The most common identifiable reasons for late bypass failure are generally felt to be progression o f o u t f l o ~ or inflow disease or the development o f isolated or diffuse stenoses within the bypass itself. ~4 Idcntkfication of bypasses prone to imminent occlusion ( " t h e failing bypass") often allows the surgeon to perform a simple elective short bypass to bridge the stenosis ("jump graft") or patch angioplasty with good long [ term patency. 2 Unfortunately, commonly used noninvasive techniques of bypass assessment such as pulse palpation or pulse volume recording may be insensitive in this setting. 3~ Many of the limbs in this series never had, palpable distal pulses at any time after operation in spite of having a patent bypass; bypasses to peroneai

Volume 12 Ntunber 5 November 1990

arteries and isolated popliteal arteries are very unlikely to have palpable foot pulses. Feeling of pulse in in situ bypass also may be inaccurate, because with ~listal occlusion pulse may feel sharper proximally. Duplex ultrasonography has been suggested .as a more accurate method of noninvasive bypass assessment. Although most graft stenoses can be identified by direct imaging of the entire bypass looking for areas of increased peak systolic velocity, this proce'~ture is time-consuming. Other proponents of the use o f duplex ultrasonography have tried to establish criteria requiring only assessment of flow parameters over the distal bypass to identify failing bypasses, in a manner analagous to that established for the carotid arteries. The potential benefits in operator time are obvious. Bandyk et al. s have proposed that a low resting D P S V of less than 45 cm/sec identifies vein bypasses that are at risk for future occlusion. H o w ever, our data do not entirely support this assertion, because a large percentage ofnonrevised bypasses in this series had DPSV values of below 45 cm/sec for prolonged periods o f time without identifiable stenoses. In addition, no other value of DPSV could be used to identify the fairing bypasses. As DPSV is measured at rest in a high-resistance arterial bed, most lesions will only become apparent by this method when there is almost complete occlusion of the bypass. This effect will be increased as the resting flow rate decreases; that is, resting DPSV will be most insensitive in the very low flow bypasses that are most prone to occlusion. Furthermore, because D P S V must be lower in larger bypass conduits and higher in smaller conduits to the same outflow, it is possible that DPSV could most accurately be applied as a function of bypass diameter; DPSV was especially inaccurate in smaller bypasses. The use of low D P S V coupled with a decrement in ABI of 0.15 has also been proposed to identify failing bypasses. 7'1s In this group of patients, ABI ,could not be applied to patients with distal anastomoses at the ankle or foot levels, because the perfusion bed for these bypasses is more distal to the area of reference. Bypass flow during tourniquet-induced hyperemia was studied in the hope that it would be a more accurate measure of bypass status. Increasing blood flow by tourniquet, exercise, or papaverine has been widely used as a means of unveiling undetected arteriai disease by pulse volume recording or intraarterial pressure measurements. 16-~s In this setting the decrease in peripheral resistance brought about by temporary ischemia causes a marked increase in flow except in the presence of a critical stenosis. The use

Hemodynamics offailing infrainguinal bypass 599 of a 3-minute period of tourniquet occlusion is critical in that this is a repeatable and reproducible ischemic event that can be compared after serial visits. In addition, many patients in this group would not be able to use a treadmill in a reproducible manner. The data in this study indicate a ratio of hyperemic blood flow/resting blood flow o f less than 2.5 was significantly correlated with bypass compromise. The use of this ratio helped normalize blood flow values obtained with varying amounts of outflow disease. The value of 2.5 was chosen because a value of 3 to 5 is regarded as normal. ~9 However, because there was a high number of noncompromised bypasses with ratios of less than 2.5, the predictive value of this single test was limited. Therefore it is not recommended that this be used as an exclusive criterion for further invasive investigation. It is recommended that bypasses with low blood flow ratios be subjected to duplex scatming of the entire bypass, because images ofstenoses can be obtained and identiffed by this method. 9,~8 In this study there was no single parameter evaluated that adequately identified the failing bypass. Although certain hemodynamic criteria were identiffed for screening failing: bypasses, further studies are needed to determine :more exacting criteria for impending bypass failure. It remains to be seen whether stenoses located anywhere along the length of a bypass can be quickly identified by a simple flow measurement at the distal bypass.

REFERENCES

1. Leather RP, Shah DM, Chang BB, Kaufman JL. Resurrection of the in situ saphenous vein bypass: 1,000 cases later. Ann Surg 1988;208:435-42. 2. Bandyk DF, Cato RF, Towne JB. Durability of the in situ saphenous vein bypass: a comparison of primary and secondary patency. J VASCSUV,G 1.987;5:256-68. 3. Blackshear WM, ThMe BL, Strandness DE. Natural history of above and below-knee femoropopliteal grafts. Am J Surg 1980; 140:234-41. 4. Johnson WC, Logerfo FW, Corson JD, et al. The prognostic value of preoperative and postoperative Doppler ankle systolic pressures in vein bypass grafts of the lower extremity. In: Dietrich ED, ed. Non-invasive cardiovasoalar diagnosis. Littleton, Mass: PSG Publishing Co, 1981:175-84. 5. Bandyk DF, Cato RF, Towne JB. Alow-flow velocity predicts failure of femoropopliteal and femorotibial bypass grafts. Surgery 1985;98:799-809. 6. Nicholls SC, Kohler TR, Martin RL, et al. Diastolic flow as a predictor of arterial stenosis. J VAsc SuRG 1986;3:498501. 7. Green RM, McNamara J, Ouriel K, DeWeese JA. Comparison of infralngninal graft surveillance techniques. J VAsc SURG 1990;11:207-15. 8. Leopold PW, Chang BB, Shandall AA, et al. Transcutaneous

600

9.

10.

11.

12.

13.

Chang et M.

flow measurements in in situ bypasses: an assessment of duplex scanning. Angiology 1986;3:143-8. Leopold PW, Chang BB, Kupinstd AM, et al. Flow/velocity characteristics of arterial bypass stenoses. J Surg Res 1989; 46:23-8. Whittemore AD, Clowes AW, Couch NP, Mannick IA. Secondary femoropopliteal reconstruction. Ann Surg 1981; 193: 35-42. Green RM, Ouriel K, Ricotta JJ, DeWeese IA. Revision of failed infrainguinal bypass graft: principles of management. Surgery 1986;100:646-54. Belldn M, Donaldson MC, Whittemore AD, et al. Observations on the use of thrombolytic agents for thrombotic occlusion of infrainguinal vein grafts. J VASC S~JI~6 1990; 11:289-96. Leather R.P, Karmody AM. In situ saphenous vein arterial bypass for the treatment of limb ischemia. In: Mannick JA~ ed. Advances in surgery. Chicago: Year Book Medical Pubfishers, Inc, 1986:175-220.

~nJof VASCULAR SURGERY

14. Szilagyi DE, Elliott JP, Hagcman JH, et al. Biologic fate of autogenous vein implants as arterial substitutes. Ann Surg 1973;I78:232-46. 15. Bandyk DF. Perioperative use of the duplex scan. In: Bergan~ Jl, Yao IST, eds. Arterial surgery. New diagnostic and operative techniques. Orlando: Grtme and Stratton, I988:46782. 16. Carter SA. Response of anlde systolic pressure to leg exercise in mild or questionable arterial disease. N Engl t Med 1972; 287:578-8i. 17. Fronek A, Johansen KH, Dilley RB, et al. UltrasonographL icaUy monitored postocclusive reactive hyperemia in the diagnosis of peripheral arterial occlusive disease. Circulation 1973 ;48:149-52. 18. Flanigan DP, Williams LR, Schwartz IA, et al. Hemodynamic evaluation of the aortoiliac system based on pharmacologic dilatation. Surgery I983;91:709-I4. I9. Strandness DE, Sumner DS. Hemodynamics for surgeons; New York: Grune and Stratton, 1975.