Patency of heparin-WA coated tubes at low flowrates W.F. ID and M.V. Sefton Department of Chemical Engineering and Applied Chemistry and Centre for Biomaterials, University of Toronto, Toronto, Ontario M5S lA4, Canada (Received 8 September 1988; accepted 3 January 1989)
The patency of heparin-PVA coated 1.14 mm i.d. polyethylene tubes was greater than that of control tubes without heparin in a novel test section that enabled ex viva evaluation in dogs at low flow and shear rates (2-10 ml/min). low shear rates were achieved by diverting a small portion of a chronic A-V shunt flow through Y-connectors into the small diameter test tubing. For shear rates in the range of 200-650 s-’ half the heparin-PVA tubes remained patent for longer than 72 min whereas half the control tubes (PVA only) only remained patent for longer than 30 min. At higher shear rates (500-l 000 s-l), the 50% patency times for heparin-PVA and PVA were 225 and 123 min respectively. The increased patency at the higher shear rates was attributed more to the effect of the connector than just shear rate. The higher shear rates were achieved by changing from a Y-connector in which the small diameter side-branch was cut flush to the larger diameter main line to one in which the small diameter test tubing protruded into the lumen of the main line; the latter ‘protruding’ connector was presumed to result in less platelet pre-activation than the other. This method has been found useful for assessing the thrombogenicity of heparinized tubes at low flow rates. Keywords: Thrombogenicity, heparin. tubing
Heparinization continues to be a particularly popular technique for reducing the thrombogenicity of cardiovascular materials. Such an approach is expected to be more useful under ‘low flow’ conditions where fibrin formation is the dominant host response. Unfortunately, there are few methods of evaluating the thrombogenicity of biomaterials at low flow rates, ex orin viva. Acute testing is complicated by the presence of surgical artifacts, while conventional A-V shunts (acute or chronic) usually involve testing at high flow rates and shear rates. The A-A shunt used by Okano’ appears to be a useful alternative. The Dudley test* is an artifact free, low flow, ex viva test. However, it cannot be used to examine materials with low thrombogenicities (and hence long ‘clotting’ times) because of the limited volume of blood that can be drawn from an animal. We have reported previously on the use of a parallel flow test section3 in a chronic A-V shunt in the dog as a means of assessing low flow rate thrombogenicity (Figure 7). A length of small diameter test tubing (1 mm i.d.) is inserted in parallel with a larger diameter (Silastic main branch (3 mm i.d.) between the arterial and venous portions (limbs) of a chronic shunt. The difference in diameters results in a fraction (nominally 2%) of the flow being diverted into the test tubing, which patency can be monitored. Occlusion of the test tubing does not compromise the main shunt, enabling it (and the animal) to be reused several times. Correspondence 0 1989
to Dr M.V. Sefton
Butterworth
8 Co (Pubkhers)
Initially3 we were unable to measure directly the flow rate in the small diameter branch. Hence we were unable to explore directly the effect of (low) flow rate on thrombogenicity. Also it was noted that substantial activation of thrombosis occurred in the Y-connector used to divert flow from the main branch of the shunt into the smaller diameter test tubing. Hence we have modified the Y-connector to assess its impact on the results from this procedure. These studies were conducted with heparin-PVA coated tubing to assess further the thrombogenicity of this material. The effect of shear rate on platelet deposition and
Hepannized _~ From artery
Figure 1 Parallel flow test section used in chronic A-V shunt to evaluate thrombogenicity at low flow rates. First published in reference 3 and reproduced by permission of the publishers, John Wiley 8 Sons, Inc.
Ltd. 0142-9612/89/050313-05$03.00 B/omaterials
1989, Vol 10 July
313
Heparin-PVA
coated tubes: W. F. lp and M. V. Sefton
fibrin formation is well known4. For shear rates < 1000 s-’ (the diffusion controlled regime) Turitto et a/.4 found that platelet deposition on subendothelium increased with increasing shear rate reflecting the increased transportofthe platelets to the surface. Simultaneously the amount of fibrin in the adhering thrombi decreased with increasing shear rate, presumably because of the increased dispersion or dilution of activated clotting factors. Also at higher shear rates, embolization becomes more important and the amount of thrombi deposit found on a surface will decrease at higher shear rates. In practice, then, the patency of tubes or catheters is presumed to increase at higher flow rates, because of the reduced amount of fibrin formation and increased embolization. The relationship among shear rate, tube size and patency has recently been modelled theoretically (D. Basmadjian, unpublished).
MATERIALS
AND METHODS
Tubing and parallel flow test section Heparin-PVAand PVAcoated polyethylenetubing (1 .I 4 mm Intramedic, Clay-Adams, i.d. X 1.57 mm o.d., PE 160, Parsippany NJ) were prepared as before3 with and without heparin (Canada Packers intestinal mucosal, 170 USP units/ mg). Heparin was added to the poly(vinyl alcohol) coating solution at levels of 1% and 2%. The 1% heparin coated tubing was found to contain =: 15 pg/cm’ of immobilized heparin, with a release rate of 1 X 10m4 yg/cm’ min. The 2% heparin coated tubings were presumed to contain twice as much immobilized heparin. Coating thicknesses (dry) were ~6 pm, based on differences in coated and uncoated (dry) weights. The PE tubing was etched in chromic acid and glow discharge cleaned to prepare it for coating. No bare spots were expected to be present based on prior observations with similarly coated tubes, nor were any detected (by SEM) during these studies. The parallel flow test section was prepared as before3. Holes were punched in the side of the Silastic main branch (3.18 mm i.d. X 6.35 mm o.d.) at an approximate angle of 30”, using specially prepared jigs and tools, at both ends of the main branch to accommodate the test tubing. The coated polyethylene tubing was inserted as before, sealed in place, and cut flush with the inside wall of the Silastic branch. This type of connection (Figure 2a) was termed the ‘flush connector’. A ‘protruding connector’ (Figure 26) was prepared in the same way, except that the coated polyethylene tube was inserted l-l .5 cm into the lumen of the Silastic branch and not cut flush.
Patency assessment The ethylene oxide sterilized parallel flow test section was inserted between the arterial and venous portions of a chronic A-V shunt that had been previously established (at least 3 wk before and typically several months) in a dog. The surgical protocol has been described elsewhere5. The dog was restrained in a sling, and the shunt clamped to permit insertion of the test section. The animal was not anaesthetized. Periodically during connection of the test section, the flow in the small diameter test tubing was monitored using a Bi-directional Doppler flowmeter (Model BVM 30, BachSimpson Ltd. London, Canada). This unit replaced our earlier flowmeter and enabled us to measure velocities in the range of O-50 cm/s (flow rates O-240 cm3/min). Flow measurement was obtained by placing the 8 MHz cuff probe flat on
314
Biomaterials
1989, Vol 10 July
SILASTIC tubing
~ e tubing
+
+ I
t+--
2.5cm 4 SILASTIC tubing
a
e tubing
w
2.5cm 4
\
SILASTIC tubing
b Figure 2 Y-connectors used to divert flow from larger diameter Silastic main branch into small diameter coated polyethylene branch: (a) ‘flush’ connector, (b) ‘protruding’ connector. Figure 2a was first published in reference 3 and is reproduced by permission of the publishers. John Wiley & Sons, Inc.
the tubing; the flow signal was displayed on a chart recorder. The tube was considered to be patent as long as the flow was not zero. Since flow reduction was generally fairly rapid (see Results), this was a reasonable endpoint. Previously3 we were only able to measure the flow rate in the main branch and used Poiseuille’s law to estimate the flow rate in the small diameter branch. In that case patency of the small diameter was monitored by touch and tube occlusion was detected as a relative coolness due to lack of flow. The flowmeter was calibrated with the timed collection of blood.
RESULTS Figure 3 shows typical plots of mean blood velocity in the coated polyethylene small diameter branch against time for both flush and protruding connectors and both 1% heparinPVA and PVA coated tubes. The mean velocity was calculated as the average of maximum systolic and minimum diastolic velocities. Time-averaged velocity was determined from approximately 10 velocity cycles. A variety of patterns were found (Figure 3) reflecting variations in animal cardiac output (all animalswere awake, running or resting during the test) and the process of thrombosis in the various test tubings and around the connectors. It is possible that some of this variation reflected changes in the degree of tube occlusion as thrombus developed and then embolized but alternative explanations are equally probable. The dog needed to be restrained for each velocity measurement, so that measurements were done every 15 min initially and then every half hour. Since the dog was restrained, the tube velocity might have been lower than that during the intervening period when he was running or walking freely around the animal laboratory. This artifact was unavoidable, but was presumably similar for all tests. With obvious exceptions, most patterns involved the relatively sharp termination of flow, occurring between the penultimate and last velocity monitorings. Hence, it was reasonable to use the time for zero flow as the patency time, although it tended to overestimate the patency time of the tubes. There does not appear to be a characteristic time
Heparin-PVA
constant for thrombus growth reflected in the onset and drop in velocity in these plots. From Figure 3, it is seen that the patency of heparinPVA coated tubes was greater than that for PVA tubes using both connectors and that the protruding connector, with a
PVA I
0 10
120
60
I
160
heparln
0
60
- PVA
120
Time
160
(minutes)
coated
tubes:
W.F.
lp and
M.V
Sefton
higher velocity gave higher patencies (for either material) than the flush connector with its lower velocity. The same conclusions are derived from the cumulative patency (fraction of tubes that remained patent at particular time point) curves seen in Figure 4, with results taken from a larger number of animals and tubes than shown in Figure 3. There was unfortunately a wide variability in the patencies of individual tubes, so that in many cases individual heparinPVA tubes may have had a patency no longer than that of individual PVA tubes without heparin. There was no correlation in these results with animal weight, shunt age or other standard haematological parameters (e.g. haematocrit, fibrinogen level); blood viscosity was not measured. The indicated mean wall shear rates (i) were calculated from the mean of the measured average velocities using i = 4v/r, where f = radius and v = velocity) assuming a parabolic velocity profile; pulsatility was not considered. Increasing the heparin content from 1% to 2% (in the original gel solution) had a small effect on patency (Figure 5). although very few tests were done at the higher heparin content. Here results are reported as the time for which half the tubes remained patent. In Figure 3, both small and large diameter branches were the same length (20 cm), while in Figure 4, the small diameter branch was kept in the range 20-25 cm, depending on how the coated tube was trimmed while the Silastic branch was kept at 20 cm. The effect of changing the length of tube branch to 30 cm or the Silastic branch to 35 cm while keeping the other one at 20 cm is shown in Figure 6, with the Figure 4 data reproduced for comparison (see also Table 7). Increasing the length of the small diameter branch and reducing the corresponding velocity/flow rate/shear rate, resulted in a slight reduction in patency (flush connector: i = 200-450 s-’ instead of 200-650 so’; protruding connector: i = 350-600 so’ instead of 5001000 s-l). On the other hand, increasing the length of the larger diameter branch and increasing the velocity in the small diameter branch (flush connector: ?; = 650-850 s-l) resulted in a small increase in patency. With the flush connector the effect of changing PE length was mainly seen at the high end of the shear rate range. The effect for PVA with the longer Silastic branch was similar to that shown in Figure 6 for heparin-PVA. 1
0
60
120
180 heparin
:g-----
=
- PVA
-
0.2 0.1
i
0 0
10
100
200
5 0
Time 60
120
160
240
300
360
(mmutes)
Figure
4
(minutes)
3
Averaged
coatedpolyethylene either
(a) ‘flush’
blood tubes
velocities
in parallel
connector
m 20 with
cm
long
PVA
20 cm long Silastic
or (b) ‘protruding’
connector.
or heparin-PVA main
branches
Cumulativepatencyofheparin-PVA
polyethylene
tubes
Silastic
either
connector
with
400
I
0
Time
Figure
300
with (i
experiment
‘flush’ 1000
times.
long)
(0,o)andPVA in parallel
connector s-‘J.
was discontinued
than indicated number
= 500-
(20-25cm
Numbers
Points
(f
= 200-650
marked
at the times in brackets
with
with
indicated
(O.m)coated 20cm
s--‘) arrows
or
lengths
indicate
andpatencywas
give number
of tubes,
of
;Orotruding’ that the greater followed
by
of animals.
E~omaterials
1989,
Vol
10 July
315
Heperin-PVA
coared tubes: W.F. ip end M.V. Sefton
Table 1
Fifty per cent pafency values* in different flow configurations
Flow configuration
30 cm PV20 Silastic
”
0
1
0
2
1
Heparin content (%)
It is interesting to note that the effect of changing tube length and hence shear rate was less dramatic than that of changing the connector. Although the shear rates with the longer polyethylene branch in the protruding connector, were similar to that seen with the flush connector and shorter PE tube, the patency of the former were greater. Similarly increasing the shear rate with the flush connector to that seen with the protruding connector, by increasing the length of the Silastic branch did not result in as great an increase in patency as changing connectors. Of course there was a wide range in shear rates for each group of tests, let alone the variation during the test, so more controlled shear rate conditions would be needed to refine these observations.
DISCUSSION Under all (low) flow conditions examined, the average or cumulative patency of heparinized PVA coated tubes was greater than that of PVA coated tubes without heparin. This was attributed to the biological activity of the immobilized heparin, which is presumed to accelerate the inactivation of thrombin by antithrombin Ill, similar to that observed before \\\
\.
1
\
heparin - PVA
0
100
200
360
400
Time (minutes) Cumulative patency of heparin-PVA tubes of different lengths Figure 6 relarive fo that of the Silastic branch: 0 30 cm coared polyethylene/20 cm Silastic (flush connector n = 11); l 30 cm coated polyethylene/20 cm Silastic (protruding connector, n = 16); A 20 cm coated polyethylene/ 35 cm Silastic (flush connector n = 7J; ----- 20 cm polyethylene/20 cm Silastic (data from Figure 4).
316
Biomaterials
1989, Vol 10 July
cm
48(4)
20-25 cm PV20 cm Silastic
30(44)
20 cm PV35 cm Silastic
45(7)
Heparin-PVA flush connector 69(11) (i = 200-450
protruding connector
s-’
72(54) (i = 200-650
s-l)
113(7) (i = 650-850
s-l)
)
lOO(15) (; = 350-600 225(15) (i = 500-l
s-’
)
000 s-‘)
nd.
‘Timeforwhich halfthetubes remained patent; number in brackets is number of tubes. n.d. = not determined.
2
Figure 5 Fifty per cent patency values (i.e. rime for which half the tubes remainedpatent) for heparin-PVA coatedpolyethylene tubes, with different heperin contents. Number in brackets is number of tubes examined.
1
PVA flush connector
with this test system3 and in other in vitro assays?*. The patency, however, was not constant, so that cumulative patency curves were preferable to average patencies as a means of presenting the results. This presumably reflects tube-to-tube variations interanimal/shunt variability, the difficulty of preparing identical connectors by hand, the natural variability in the process of thrombosis itself and the wide variation in test tubing velocities. A reason for using the parallel flow test section with the chronic shunt was the desire to avoid the surgical artifacts associated with acute biomaterial testing. Unfortunately the parallel flow test section introduces another artifact -the initiation of thrombosis and pre-activation of platelets/clotting factors in the Y-connector. The flow is disturbed as a portion of the flow is diverted into the coated test branch. Also the uncoated end of the polyethylene tube is exposed to the blood at the Y-connector/main line interface and the Silastic portion of the test section may have a different thrombogenicity than the Silastic of the previously implanted, and perhaps conditioned, chronic shunt. Consequently, the observed patency is not strictly that associated only with the test tubing. Hence, the observation that the patency of the heparin PVA tubes although greater than that of the tubes without heparin, was still limited (the maximum of -6 h is too low for practical application) may be explained by this contribution of the Y-connector to the development of the occluding thrombus. Similarly the absence of a significant dose-dependent effect of immobilized heparin on patency (Figure 5) is explained in the same way. The occluding thrombus may have begun in the inlet Y-connector before exposure to much of the heparinized surface, or at the outlet Y-connector where the two flows recombine. The occluding thrombus was generally in the tubing at the inlet or occasionally at the outlet Y-connectors and rarely in the middleof the test tubing. No thrombi werefound in the body of the connector, presumably because of the high flows in the connector, although thrombi loosely adhering here could have been washed away when the connector was rinsed. Alternatively the immobilized heparin may not have been adequate either in amount of antithrombin activity to handle the load of activated clotting factors or platelets generated in the Y-connector. The absence of an effect of heparin-PVA on platelets has been documented elsewhereg,“. TheYconnector effect is seen most clearly in Figure 4, where changing the Y-connector to the protruding type resulted in a significant increase in patency (for both heparin-PVA and WA). We presumed this to be the consequence of more gentle connections to the test tubing as the
Heparin-PVA
flow (at the inlet) is able to reduce speed first before changing direction. The velocity (and shear rate) in the test tubing also increased, because of the reduced drag in the Y-connectors. In an attempt to distinguish the direct effect of shear rate from the change in Y-connector associated pre-activation, the mean shear rate in the test tubing wasvaried by changing the relative lengths of the test tubing and main line. Changing the relative lengths was expected to change the test tubing shear rate in direct proportion (i.e. f a LJL,, where LJL, are silastic, polyethylene branch lengths respectively) assuming Poiseuille’s law. Unfortunately the range of shear rates examined was too small to distinguish much of an effect. Certainly small changes in shear rate had a small effect which is not inconsistent with various relationships which relate shear rate to transport processes4,“. These theoretical expressions involve a maximum dependency of ‘thrombosis’ on shear rate to the l/3 power (the Leveque flow regime). If these expressions were to apply to the experiments conducted here, then the observed change in shear rate would result in only a 15% change in patency, which is obviously less than that seen on changing Y-connectors. Unfortunately, it is impossible to relate directly these theoretical calculations with the patency experiments, so that the relative importance of shear rate and Y-connector pre-activation must remain unclear. Nevertheless we assume that the direct effect of shear rate is minor. However, at much higher flow rates/shear rates, as in the 3 mm i.d. Silastic portions, we presume that the expected effect of shear rate on patency is important. Elsewhere” we report patencies of greater than a week for heparin-PVA, PVA and uncoated polyethylene tubes when placed in the chronic shunt in a series mode, where all the flow is directed through these tubes. Here the flow is so high that occlusive thrombi do not have the opportunity to develop regardless of material (provided it is not too highly thrombogenic).
coated tubes: W.F. lp and M.V. Sefton
the Y-connector from one in which the side-branch was cut flush with the lumen of the main line to a more gentle one in which the test tubing protruded into the main line resulted in an increase in patency for both heparin-PVA and PVA coated tubes. Nevertheless the parallel flow test section has been found to be a useful means of assessing the biological activity of heparinized tubes at low flow rates. However, further refinements are required to distinguish the effects of the test material from those of the Y-connectors.
ACKNOWLEDGEMENTS We gratefully acknowledge the support of the National Institutes of Health undergrant HL24020 and theassistance of Dr W. Zingg and his staff of the Division of Surgical Research, Hospital for Sick Children.
REFERENCES 1
2
3
4
5
6
CONCLUSIONS
7
Immobilized heparin has been shown again to be biologically active in delaying thrombosis. Patencies of heparin-PVA coated tubes were greater than that of PVA coated tubes without heparin at low flow rates (2-10 ml/min) and moderate shear rates (=200-1000 s-l). However, the parallel flow test section used in conjunction with a chronic A-V shunt in dogs is not perfect. Thrombosis appears to be initiated in the Y-connectors used to divert a fraction of the chronic shunt flow into the 1 mm i.d. test tubing. Changing
8 9
10
11
Park, K.D., Okano,T., Nojiri, C. and Kim, S.W., Hepann immobolization onto segmented polyurethane-urea surface - effect of hydrophilic spacer, Trans. 3rd World Biomaterials Congress, Kyoto, Japan, 1988, 287 Dudley, 6.. Williams, J.L., Able, K. and Muller, B.. Syntheses and charactenzation of blood compatible surfaces, Trans. ASAIO 1976, 22, 538 Ip, W.F., Zingg, W. and Sefton, M.V., Parallel flow arteriovenous shunt for the ex viva evaluation of heparinized materials,, J. Biomed. Mater. Res. 1985, 19, 161-178 Turitto, V.T., Muggli, Ft. and Baumgartner, H.R.. Physical factors influencing platelet deposition on subendothelium: importance of blood shear rate. Ann. NYAcad. Sci. 1977, 283, 284-292 Zingg. W., Ip, W.F., S&on, M.V. and Mancer, K., A chronic arteriovenous shunt for the testing of biomaterials and devices in dogs, Life Suppoti Systems 1986, 4. 221-229 Goosen, M.F.A. and Sefton, M.V.. Properties of a hepann-polyvinyl/ alcohol hydrogel c0ating.J. Biomed. Mater. Res. 1983,17,359-373 Goosen, M.F.A., Sefton, M.V. and Hatton, M.W.C., Inactivation of thrombin by antithrombin Ill on a heparinized biomatenal. Thromb. Res. 1980. 20, 543-554 Rollason, G. and Sefton, M.V., Factor Xa inactivation by a heparinized hydrogel, Thromb. Res. 1986,44, 517-525 Cholakis, C.H., Zingg, W. and Sehon. M.V., ln vitro platelet Interaction with a heparin-polyvmyl alcohol hydrogel. J. Biomed. Mater. Res. 1989, 23,417-442. Cholakis, C.H.. Zingg, W. and Sehon. M.V., Effect of hepann-PVA hydrogel on platelets in a chronic canine arteno-venous shunt, J. Biomed. Mater. Res 1989, 23, 399-4 16 Basmadjian, D. and Sefton, M.V., A model of thrombin inactrvation in heparinized and nonheparinized tubes with consequences for thrombus formation, J. Biomed. Mater. Res. 1986, 20, 633-651
Biomatenals
1989, Vol 10 July
317