the effect of increasing clamping forces on endothelial...

Cardiovascular Surgery,Vol. 7, No. 4, pp. 457–463, 1999 1999 The International Society for Cardiovascular Surgery. Published by Elsevier Science Ltd

All rights reserved. Printed in Great Britain0967–2109/99 $20.001 0.00

www.elsevier.com/locate/cardiosur

PII: S0967-2109(98)00154-9

The effect of increasing clamping forces onendothelial and arterial wall damage:an experimental study in the sheep

A. I. Margovsky, A. J. Chambers and R. S. A. LordSurgical Professorial Unit, St Vincent’s Hospital, University of NSW, Sydney, Australia

Purpose: This study aimed to relate the level of physical force applied to the arterial wall byatraumatic clamps to the degree of endothelial and wall damage. Methods: Sixteen sheepcarotid and femoral arteries were each demarcated into four segments 1 cm apart (total 64segments). Each segment was clamped for 15 min with a standard angled DeBakey vascularclamp. Four levels of force were generated by closing the clamp at three, four, five and sixnotches of closure. The extent of endothelial injury was assessed by using a dedicated com-puter assisted image acquisition program to measure the area stained by Evan’s blue dye.The extent of damage to the layers of the arterial wall was analyzed and compared by scanningelectron microscopy and light microscopy. Results: For femoral arteries, the area of endothelialinjury was considerably less for three notch (3.76 6 0.28 newtons) and four notch (5.68 60.29 newtons) closure compared with that for five notch (6.19 6 0.31 newtons) and six notch(6.61 6 0.16 Newtons) closure (p 5 0.01). For carotid arteries, three notch (5.68 60.28 newtons) closure caused less damage than did four notch (7.98 6 0.29 newtons), fivenotch (9.17 6 0.40 newtons) and six notch (9.57 6 0.64 newtons) closure (P 5 0.02). Scan-ning electron microscopy confirmed the extent and depth of arterial injury correspondeddirectly to the forces generated by the vascular clamps. Conclusions: The closing forces gener-ated by arterial clamps correlated positively with the extent of artery wall injury. Vascularclamps should be applied at the minimum level of force that will arrest blood flow. 1999The International Society for Cardiovascular Surgery. Published by Elsevier Science Ltd. Allrights reserved

Keywords: arterial wall, endothelial wall, sheep, vascular clamp

IntroductionClamps are used to arrest flow and to control bleed-ing from arteries [1], but they may also cause moder-ate to severe damage to the vessel wall [2–4], whichcould jeopardize an otherwise successful vascularreconstruction. Some of these complications, such ascreation of a flap, dissection, through-and-throughinjury and arterio-venous fistulae follow relativelygross instrumentation. Other effects, including intra-

Correspondence to: Professor Reginald S. A. Lord, Surgical Prof-essorial Unit, Level 17, O’Brien Building, St Vincent’s Hospital, Vic-toria Street, Darlinghurst, NSW 2010, Australia

CARDIOVASCULAR SURGERY JUNE 1999 VOL 7 NO 4 457

vascular thrombosis and later stricture formation [5,6], reflect more subtle injuries including endothelialdisruption [7].

Previous reports described some of the features ofclamp design and the effects of different clamps onthe vessel wall. These studies concentrated parti-cularly on relating available vascular ‘atraumatic’clamps in terms of their occlusive ability and holdingcapacity to the effect on the vessel wall [1–3, 5, 8–11]. In some of these reports, the arteries studiedwere relatively small, such as rat and rabbit aortas,and femoral arteries [5, 12–14]. In other studies,canine arteries were employed [2, 3, 8, 9, 11].Despite the differences in arterial diameters, wallthickness’, lamellar units and other morphological

Vascular clamps—endothelium and artery wall interactions: an experimental study in the sheep: A. I. Margovsky et al.

characteristics, these studies consistently uncoveredendothelial damage and other cellular injuries. Thepressures generated by different types of clamps wereoccasionally measured [3, 4, 10, 13], but no previousstudies graded the pressures applied to the vesselwall to the subsequent injury, nor have previousstudies investigated clamp-related injuries in relationto differences in artery diameter.

Investigators have generally used histologicalexamination and scanning electron microscopy todetermine vessel wall damage [1, 2, 5, 8, 9, 11, 12,15]. Other methods for measuring the extent ofinjury include the assessment of functional ability[14], vasomotor responses [12] and local plateletuptake [4]. In vivo staining with Evan’s blue has alsoproved useful for defining the extent of endothelialdisruption [3, 8], especially when a computer-assisted program was employed to calculate the dam-aged areas [16].

This study aimed to investigate the correlationbetween the forces generated by commonly used vas-cular clamps and the ensuing endothelial and arterywall damage. The carotid and femoral arteries of thesheep were studied in vivo to achieve a reproducibleanimal model.

Materials and methods

Experimental protocol

Sheep were studied, as the carotid and femoralarteries are of similar calibre to those in man. Thecarotid and femoral arteries of four adult sheepweighing between 40 and 50 kg were used. A totalof 16 arteries were subjected to varying clamp forcesand the extent of endothelial damage was assessedby Evan’s blue staining, histological examination andscanning electron microscopy.

Anesthaesia was induced using intravenous Nem-butal (15 mg/kg) and maintained using oxygen with1.5–2.5% Halothane. An intravenous line providedfluid replacement with normal saline through anexternal jugular vein. Systemic blood pressure wasmonitored through a catheter in the deep femoralartery and mean blood pressure was maintainedbetween 90 and 100 mmHg.

Both carotid arteries were exposed via a midlineneck incision. The femoral arteries were exposedthrough groin incisions. An aliquot of 150 IU/kg ofsodium Heparin was administered intravenously5 min before the first set of clamps was applied tothe carotid arteries.

Four identical angled DeBakey arterial clamps(Downs Surgical, UK) were placed on the middlesection of each exposed artery at 1-cm intervals, avo-iding any branches and surrounding tissues, andcommencing at the distal end to avoid clamping acollapsed vessel (Figure 1). The jaws of each clamp

CARDIOVASCULAR SURGERY JUNE 1999 VOL 7 NO 4458

Figure 1 Diagram showing four angled DeBakey vascular clamps appliedto the artery with forces generated at (a) three notch, (b) four notch, (c)five notch and (d) six notch, respectively. t 5 thickness of clamped arteryis proportional to the force applied

were placed at right angles to the long axis of thevessel, with the vessel held 1 cm from the tip of theclamp. The four clamps were closed to three, four,five and six notches, respectively, and left in placefor 15 min.

After the last clamp was removed, a 0.5% solutionof Evan’s blue dye (up to a total volume of 3 ml/kg)was administered intravenously and allowed to circu-late for 30 min [16]. All carotid and femoral arterieswere then removed, opened longitudinally andpinned flat. The areas of blue staining were exam-ined under a dissecting microscope at 3 7 magnifi-cation. A video camera (KP-M1E, Hitachi–Denshi,Japan) connected to the viewing post of the micro-scope was used to acquire images. The images weretransferred to a computer (Apple Macintosh, usingImage Scion, version 1.51, National Institutes ofHealth, USA), which calculated the area of bluestaining in each vessel using appropriate calibration.At the end of each experiment, the animals werehumanely killed. Half of the samples were then fixedin glutaraldehyde and prepared for scanning electronmicroscopy (Stereoscan S150, Cambridge Instru-ment CO, UK). The rest of the arterial specimenssubjected to clamping were fixed in 10% bufferedformalin. The morphological changes were studiedafter staining the paraffin sections with haematoxylinand eosin.

Clamping force and pressure measurements

Before the clamps were applied, the closing force foreach notch of the DeBakey angled vascular clampsemployed in this experiment was pre-measured by anelectronic loading-cell device with 0.1 N resolution


using the following procedure (Figure 2) where: A 5electronic load-cell measured force to 0.1 N resol-ution; and B 5 electronic digital callipers measureddistance to 0.01 mm resolution.

For a given notch setting D refers to the displace-ment of the application point of the callipers. Theforce from the load cell (A) was correlated with thegap between the jaws of the arterial clamp (C) asmeasured by the gauge (B). The procedure wasrepeated for a range of gaps with the displacementdetermined by the screw thread (E).

At any notch setting, five different loads wereapplied giving a load in newtons (N)/deflection(t mm) ratio for that notch setting. The measure-ments were repeated three times for each notch set-ting giving a family of curves. Using a statisticalcurve fitting software package (Microsoft Excel 4.0)one empirical function was derived to obtain the gapthickness (deflection) in terms of the two variables:No (the notch settings) and F (force). The family ofcurves was then replotted using this derived math-ematical function to ensure agreement with the orig-inal experimental data (Figure 3).

The external diameters of the arteries and thethickness of the clamped arteries were measuredusing callipers. These measurements were used tocalculate the average forces generated by the clamps,using the equation:

F 5t 2 (5.9957 2 1.4688 3 No)

0.4163 2 0.0172 3 No

where: F 5 force applied (newtons); t 5 clampedthickness (mm); No 5 notch number.

The areas subjected to clamp application were cal-culated using the equation:

A 5 p 3 D (mm) 3 3 mm,

where: A 5 area of the clamped artery; D 5 external

Figure 2 Diagram showing setup for electronic loading-cell measure-ments of forces generated by clamps (explanation in text)


Figure 3 Diagram showing computer-analysed correlations between theclamped thickness and forces generated by DeBakey vascular clamp

diameter of clamped artery; 3 mm 5 width of theangled DeBakey vascular clamp.

To calculate the specific compressive pressure (P)immediately under the clamp face, the estimatedclamping force was divided by the compressed area,which is related to the area of the clamp face as fol-lows:

P 5 F/A,

where: P 5 pressure (newtons/mm2); F 5 force(newtons); A 5 area of the clamped artery (mm2).

All the surgical and technical procedures were per-formed at the Department of Biomedical Engineer-ing of the University of NSW, with approval fromthe local Animal Care and Ethics Committee(ACEC). The data were analysed using the Wil-coxon signed-rank non-parametric statistical test.Values are shown as mean 6 s.d.

Results

Clamping force and pressure measurements

The mean external diameters were 7.2 6 0.8 mm forcarotid arteries and 5.3 6 0.7 mm for femoralarteries.

The clamping forces generated by the standardangled DeBakey vascular clamps were calculated forcarotid and femoral arteries in relation to the thick-ness of the vessel clamped (Table 1). The mean area(mean 6 s.d.) subject to clamp application in thecarotid arteries was 67.86 6 7.5 mm2 for carotidarteries and 49.95 6 6.5 mm2 for the femoralarteries.

The results of the specific compressive pressureswere expressed as mean 6 s.d. (Table 2). Other vari-ables, including systemic blood pressure and vesselelasticity, were not included in the calculationsbecause all measurements were performed under


Table 1 Forces generated by angled DeBakey vascular clamp when clamping the sheep femoral and carotid artery (mean 6 s.d.)

Notch closure Femoral artery (newtons) Carotid artery (Newtons)

3 notch 3.76 6 0.28 5.68 6 0.284 notch 5.68 6 0.29 7.98 6 0.295 notch 6.19 6 0.31 9.17 6 0.406 notch 6.61 6 0.16 9.57 6 0.64

Table 2 Clamping pressures applied to the femoral and carotid arteries, calculated in respect to the clamped surface area (mean 6 s.d.)

Notch closure Femoral artery (g/mm2) Carotid artery (g/mm2)

3 notch 75.4 6 5.6 83.8 6 4.14 notch 113.9 6 5.8 117.6 6 4.35 notch 124.1 6 6.2 135.2 6 6.66 notch 132.4 6 3.2 141.1 6 9.5

standard conditions using animals of similar weight,size, and artery wall thickness (0.75 6 0.05 mm).

Evan’s blue staining

The areas of Evan’s blue staining in carotid arteriesseen with the different clamping forces are shown inFigure 4. The areas of endothelial damage in thecarotid arteries showed that significantly less damageoccurred at three notch closure (15.75 6 1.1 mm2,P 5 0.02) compared with damage at four notch(20.92 6 4.2 mm2) and five notch (21.31 62.3 mm2, P 5 0.6) and at six notch closure (23.596 2.5 mm2, P 5 0.1).

The extent of endothelial injury in femoral arterieswith the clamping force at three notch (10.28 61.8 mm2) was not significantly different comparedwith four notch (13.22 6 3.4 mm2, P 5 0.067) butwere less than for five notch (18.31 6 3.1 mm2, P5 0.01), and at six notch (21.23 6 2.5 mm2, P 50.01).

Scanning electron microscopy

All specimens underwent SEM examination, withtypical findings for the clamped segments shown inFigure 5. In the areas subjected to three-notchclamping there were minimal intimal changes, shownmostly as changes in endothelial cell orientation.Clamping at four notches produced mild endothelialcell disruption. Clamping at five and six notches pro-duced consistent desquamation of endothelium, withpartial disturbance and rupture of the internal elasticlamina and the creation of deep fissures.

Light microscopy

The artery wall injuries observed by light microscopywere classified according to our scheme describedelsewhere [4].


Figure 4 Graphs demonstrate a comparative area of endothelial damagefor: (a) femoral arteries and (b) carotid arteries


Figure 5 Scanning electron microscopy images at the sites of clamps application: a, b, c, d 5 femoral artery (original magnification 3 25); e, f, g, h 5carotid artery (original magnification 3 25); a, e 5 three notch; b, f 5 four notch; c, g 5 five notch and d, h 5 six notch

In mild injuries, light microscopy analysis did notidentify any significant endothelial injury and theinternal elastic lamina was intact. However, vacuoliz-ation of some smooth muscle cells in deep medialareas was observed in some specimens (Figure 6A).

In moderate injuries, vacuolization of smoothmuscle cells in the media was consistently detected.The internal elastic lamina was partially disrupted infissures and the adjacent endothelial cells showeddefinite signs of injury within the fissure zone with-out extensive desquamation of the cells around theclamped site (Figure 6B).

In severe injuries, the cell response was differentcompared with moderate injury, including completedesquamation of the endothelial cells in the zone ofclamp application. Vacuolization of smooth musclecells in the superficial and deeper part of media was


apparent. Continuity of the internal elastic laminawas broken in the fissure zone, with the fissure usu-ally extending through the medial layer to form acavity often filled with thrombus (Figure 6C).

DiscussionNumerous reports confirm that vascular clamps candamage the vessel wall [1, 3, 5–7, 12, 17–19]. Mostof these studies evaluated different vascular clampsin a general way with relatively few analysing theeffects of specific features of clamp design. Clampgeometry, closing force, weight and holding abilityare known to influence the extent of trauma. Thearchitecture of the jaw face [3, 8, 10], the type ofvessel [15] and the duration of clamping are alsoimportant. In the present study, the extent of endo-


Figure 6 Light microscopy showing damage of the sheep carotid artery upon clamp application (original magnification 3 100). (A) Mild damage, (B)moderate damage, (C) severe damage

thelial disruption has been directly correlated withincreasing forces of clamp application, with the studydesign limiting the effect of other variables includingthe duration of clamping [16].

A sheep model was used in this investigationbecause ovine carotid and femoral arteries are com-parable in size to those in man. Standard angledDeBakey clamps were chosen because these clampsare commonly used in human vascular procedures.These clamps are similar to other serrated clampsthat, in general, have proved to be most reliable andresistant to slipping. A further reason for using theDeBakey clamp in our study was that the DeBakeyclamp ranked as average among other vascular‘atraumatic’ clamps [2].

The force needed to occlude a vessel is determinedby four variables: vessel diameter, blood pressure,vessel elasticity and blade contact area. The closingpressure of clamps used to occlude small calibre ves-sels has been estimated [13], but the precise relation-ship between the forces generated by the clamp, theactual closing pressures, and the subsequent arterialwall injuries have not been quantified.

Berlin and Berlin [10] established that com-pression forces of 50 to 75 g are needed to arrestflow in a vessel under 300 mm pressure. Most of thewidely used vascular clamps examined, however,required 200 to 1000 g or even greater force toengage the latches [10].

Harvey and Gough [8] demonstrated that when aclamp was applied with sufficient force to occlude avessel, the damage was considerably less than whenthe clamp was fully closed [8] and Pabst et al. [20]confirmed that excessive clamping forces caused sev-ere endothelial damage.

In the present experiment, the forces generated at


three-notch clamping induced a significantly smallerarea of endothelial disruption than forces generatedat four, five and six notches. The area of damagereached a maximum level at five notches, beyondwhich there was no further significant change. Theresults indicate that damage to endothelium dependson the closing pressure of the specific clamp withina limited range, but increases in pressure beyond thishave no additional effect on the endothelium. Thisobservation may relate to blood vessels being viscoel-astic and thus sharing some of the characteristics ofviscous liquids, namely, continuous deformationafter a force is applied.

Scanning electron microscopy demonstrated adirect relationship between vascular compressiveforces and arterial wall damage. Higher magnifi-cation revealed deposition of platelets on the surfaceproportional to the endothelial lesions, but there waslittle evidence of thrombogenesis, perhaps becauseof the brief duration of clamp application.

The minimal endothelial damage produced byforces generated at three-notch clamping suggeststhat effective occlusion of arteries 4.5–8 mm externaldiameter can be performed safely at this level ofclamp closure. However, the amount of damagecaused will be additionally influenced by vessel size,systemic blood pressure, clamp jaw geometry,arterial wall elasticity and the duration of clampapplication.

Evan’s blue dye injected intravenously stains dam-aged endothelial cells, and this property has provedreliable in outlining the areas of clamp application[12]. The authors refined the use of Evan’s bluestaining by a new method of computer-assistedanalysis to achieve precise calculation of the dam-aged surface. Using this and other technology the


authors have previously shown that endothelial andwall injury is proportional to the duration of clamp-ing up to a period of ~30 min [16]. After this thres-hold period, clamping for a longer time at constantpressure did not cause any further damage. Endo-thelial damage from so-called ‘atraumatic’ clampsincrease platelet uptake [4] and may progress toaccumulation of thrombus with the risk of embolismor vessel occlusion. The present studies confirm thatthe pressure exerted by a clamp is also proportionalto the extent of wall injury but, unlike the effect ofduration, no threshold pressure was identified. Inother words, the deleterious effects of increasingpressure continued over the range of pressures stud-ied and corresponded to the maximal range of notchclosure in the clamp. In clinical practice clampsshould be applied at the minimal pressure level thatsafely arrests flow, and if possible, the clamp shouldbe released every ~30 min [16]. Excessively longclamping at high application pressures will damagethe arterial wall and increase the likelihood of clini-cally significant clamp-related complications. At lowocclusion pressures only the endothelium is injured.Higher pressures caused deeper damage with thepossibility of dissection, strictures, rupture andthrombotic occlusion.

AcknowledgementsThis study was supported by St Vincent’s ClinicFoundation, Sydney, Australia.

References1. Henson, G. G. and Robb, C. G., A comparative study of the

effects of different arterial clamps on the vessel wall. British Jour-nal of Surgery, 1956, 43, 561–564.

2. Hickman, G. A. and Mortensen, J. D., A comparative evaluationof vascular clamps. Journal of Thoracic and Cardiovascular Surgery,1962, 44, 561–569.

3. Grant, P. A., Egerton, W. S., Strutton, G. and Balderson, G.,The Martin titanline arterial clamp — a new lightweight alterna-tive. Australian and New Zealand Journal of Surgery, 1988, 58,67–71.

4. Margovsky, A. T., Lord, R. S. A., Meek, A. C. and Bobryshev,Y. V., Artery wall damage and platelet uptake from so-calledatraumatic arterial clamps: an experimental study. CardiovascularSurgery, 1997, 5, 42–47.


5. Slaybeck, J. B., Bowen, W. W. and Hinshaw, D. B., Intimalinjury from arterial clamps. American Journal of Surgery, 1976,132, 183–188.

6. Aukland, A. and Hurlow, R. A., Carotid stenosis due to clampinjury. British Medical Journal, 1981, 282, 2013.

7. Kuo, J., Ramstead, K., Salih, V. et al., Effect of vascular clampon endothelial integrity of the internal mammary artery. Annalsof Thoracic Surgery, 1993, 55, 923–926.

8. Harvey, J. G. and Gough, M. H., A comparison of the traumaticeffects of vascular clamps. British Journal of Surgery, 1981, 68,267–272.

9. Moore, W. M., Manship, L. L. and Bunt, T. J., Differentialendothelial injury caused by vascular clamps and loops. I: normalvessels. American Journal of Surgery, 1985, 51, 392–400.

10. Berlin, R. D. and Berlin, R. D., Vascular clamping: a new con-cept with manometric studies comparing available instrumentswith a novel design. Vascular Surgery, 1978, 12, 108–112.

11. Guidoin, R., Martin, L., Levaillant, R. et al., Endothelial lesionsassociated with vascular clamping-surface micropathology byscanning electron microscopy. Biomaterials and Medical Develop-ments in Artificial Organs, 1978, 6, 179–197.

12. Barone, G. W., Conerly, M., Farly, P. C. et al., Assessing clamprelated vascular injuries by measurement of associated vasculardysfunction. Surgery, 1989, 105, 465–471.

13. Dujovny, M., Wakenhut, N., Kossovsky, N. et al., Minimumvascular occlusive force. Journal of Neurosurgery, 1979, 51,662–668.

14. Risberg, B. and Bylock, A., Vascular trauma induced by clamp-ing — correlation between surface ultrastruction and fibrinolyticactivity. Acta Cir. Scandinavica, 1981, 147, 25–32.

15. Manship, L. L., Moore, W. M., Bynoe, R. and Bunt, T. J., Dif-ferential endothelial injury caused by vascular clamps and vesselloops. II: Artherosclerotic vessel. American Journal of Surgery,1985, 51, 401–406.

16. Margovsky, A. I., Chambers, A. J. and Lord, R. S. A., The effectof arterial clamp duration of on endothelial injury: an experi-mental study. Australian and New Zealand Journal of Surgery,1997, 67, 448–451.

17. Litchford, B., Okies, J. E., Sugimura, S. and Starr, A., Acuteaortic dissection form cross-clamp injury. Journal of Thoracic andCardiovascular Surgery, 1976, 72, 709–713.

18. Coelho, J. C., Sigel, B., Flanigan, D. P. et al., Arteriographicand ultrasonic evaluation of vascular clamp injuries using an invitro human experimental model. Surgical Gynaecology andObstetrics, 1982, 155, 406–412.

19. Archie, J. P. Jr, Early postoperative femoral-distal bypass graftfailure due to vascular clamp injury induced common femoralartery thrombosis. American Journal of Surgery, 1988, 54, 167–168.

20. Pabst, T. S., Flanigan, D. P. and Buchbinder, D., Reducedintimal injury to canine arteries with controlled application ofvessel loops. Journal of Surgical Research, 1989, 47, 235–241.

Paper accepted 2 November 1998

the effect of increasing clamping forces on endothelial...

Documents