Review

Platelet functions in patients with aortic ball valves

Jon Dale, M.D. Erik Myhre, M.D. Oslo, Norway

Patients with aortic ball valves have a high incidence of arterial thromboembolic complica­tions,' and the emboli are usually derived from thrombi located on the prosthetic valve itself.'"^ Anticoagulant treatment cannot fully prevent these complications, although intense therapy has been found to offer some protection." ' The tendency to arterial thrombus formation in such patients focused our attention on the function of their blood platelets.

Another characteristic finding in patients with aortic ball valves is intravascular hemolysis, which is almost invariably increased to a greater or lesser degree.* The possibility that hemolysis may modify the function of blood platelets, and thereby the hemostatic and thrombotic mech­anisms, is indicated by Hellem's discovery that red cells contain a factor necessary for platelet retention in glass bead columns,** later identified to be adenosine diphosphate (ADP).'"

The purpose of this investigation was to study platelet functions in a larger series of patients with aortic ball valve prostheses, and to relate the findings to the degree of intravascular hemolysis as well as to the occurrence of arterial throm­boembolic complications. Results from a prelimi­nary study have been published previously." '̂

Materials and methods

Patients with single Starr-Edwards aortic ball valves were studied. Such valves were implanted in 253 patients at our center from 1967 until 1970, and 169 of the 175 patients still alive reported for an extensive follow-up examination in 1972.' Two series of Starr-Edwards aortic valves were used.

From the Institute for Thrombosis Research and Medical Department B, Rikshospitalet, University Hospital, Oslo, Norway. Received for publication Dec. 3, 1975. Reprint requests: Dr. J. Dale, Institute for Thrombosis Research, Rikshospitalet, Oslo 1, Norway.

type 1200 with silicone rubber ball and metal cage." Most of the platelet function studies were done before or at the follow-up examination. After this the patients entered a controlled clin­ical trial and were given either acetylsalicylic acid or placebo in addition to their anticoagulant therapy. Platelet function tests could therefore not be performed later, since the results could disclose the type of drug given.

The platelet counts were determined in all subjects, whereas the other tests were done in a varying number of unselected patients. Platelet function was also evaluated in healthy individ­uals, mainly hospital personnel, selected to obtain the best possible similarity to the patients with regard to sex and age distribution. The number of patients and healthy persons subjected to the different tests is specified in the Results section. In addition, the bleeding time was measured in 23 patients receiving anticoagulants for other rea­sons than valve implantation, including postoper­ative prophylaxis after orthopedic surgery, coro­nary heart disease, peripheral arterial disease, and atrial fibrillation. Platelet survival was studied in seven patients with aortic valve disease unoper-ated upon.

The degree of intravascular hemolysis was esti­mated from the serum lactate dehydrogenase (LDH) activity at 25° C. as earlier described." The following platelet function tests were done: The bleeding time was measured from two inci­sions by Borchgrevink and Waaler's'^ modifica­tion of Ivy's method by one of us (J. D.). The cuts were approximately 1 cm. long, but the depth was adjusted so that, instead of a 1 mm. depth, a blood drop the size of a matchhead emerged in 30 seconds. Platelets were counted in a hemocytom-eter by a modification of Nygaard's method.** Platelet adhesiveness was determined in native blood by the modified method of Hellem."

September, 1977, Vol. 94, No. 3, pp. 359-366 American Heart Journal 359

Dale and Myhre

Table I. Serum LDH levels in healthy individuals and patients with Starr-Edwards aortic ball valve prostheses*

Table 111. Platelet counts in healthy individuals and patients with Starr-Edwards aortic ball valves.*

Healthy persons

Patients, valve 1200

Patients, valve 2300

No. of subjects

60

21

48

Serum LDH (U./L.)

Mean S.EM.

120.5 2.7

269.9 28.3

580.7 54.7

Compared with

normal

p < 0.001

p < 0.001

Healthy indi­viduals

Patients, valve 1200

Patients, valve 2300

No. of subjects

28

46

111

Platelets (per fil)

Mean S.E.M.

224,200 8,000

183,000 7,200

189,100 5,600

Compared with

normal

p < 0.001

p < 0.001

*LeveI of significance of differences between mean values in normal subjects and in valve patients.

Table II. Bleeding time in healthy individuals, patients with Starr-Edwards aortic ball valves and other patients on anticoagulants*

•Number of subjects, mean values, and standard error of the mean. Level of significance of the difference between normal subjects and patients.

Table IV. Platelet adhesiveness in healthy indi­viduals and patients with Starr-Edwards aortic ball valves

Healthy indi­viduals

Patients, valve 1200

Patients, valve 2300

Others on anti­coagulants

No. of subjects

25

10

22

17

Bleeding time (min.)

Mean S.E.M.

4.8 0.24

5.5 0.51

7.2 0.38

5.6 0.23

Compared with

normal

N.S.

p < 0.001

p < 0.01

Healthy indi­viduals

Patients, valve 1200

Patients, valve 2300

No. of subjects

22

21

48

Adhesiveness (%)

Mean S.E.M.

71.8 1.8

50.9 3.7

27.2 2.3

Compared with

normal

p < 0.001

p < 0.001

•The number of subjects, mean value in minutes, and standard error of the mean are listed, and the level of significance of the differences between the mean values in normal subjects and patients is calculated (N.S. = not significant).

Platelet aggregation was estimated in platelet-rich plasma (PRP) by the turbidimetric method described by Born," with the Unigalvo EEL-titrator (Evans Electroselenium, Ltd., Essex, England). P R P was obtained by centrifugation of citrated blood at 300 X g for 15 minutes. Aggre­gation was initiated with collagen, ADP, and epinephrine in final concentrations of 2.1 jug per milliliter, 0.7 juM, and 3.64 juM, respectively. The curves were always recorded in this sequence, adding collagen suspension to P R P 20 minutes after blood withdrawal. The maximal rate of collagen- and ADP-induced aggregation was esti­mated. This response has been found to be proportional to the platelet count in normal subjects (Dale and Lund-Riise: To be published). The rate in each patient was therefore calculated

platelet count. Whether epinephrine induced irreversible secondary aggregation was evaluated from the shape of the curves.

Platelet survival was determined according to Abrahamsen's'* method by labeling autologous platelets with "CrOi " and reinjecting them. The platelet half-life was determined from 16 hours after the injections and until the radioactivity had reached half of the 16 hour value.

Results

Intravascular hemolysis, as reflected by the serum LDH values, was increased in nearly all patients with prosthetic valves, and significantly higher enzyme levels (p < 0.001) were found in patients with valve type 2300 than in those with valves of series 1200 (Table I), which is in accord with our earlier results.* Only LDH values from patients in whom platelet adhesiveness was also determined are presented here.

360 September, 1977, Vol. 94, No. 3

Platelet functions with aortic ball valves

The mean bleeding time in healthy individuals was 4.8 minutes and the standard deviation (S.D.) 1.2, giving a normal range (mean ± 2 S.D.) of 2.4 to 7.2 minutes. The precision of the method could be estimated since duplicate incisions were made, and the coefficient of variation was calculated to be 17.5 in normal subjects and 15.3 in valve patients.

The bleeding time was prolonged on average by 2.4 minutes in patients with valves of type 2300, which is statistically highly significant, whereas a moderate, nonsignificant prolongation was found in those with valve 1200 (Table II). In the heterogeneous group of patients on anticoagu­lants, a slight but significant increase was found. Compared with this group, the increase of the mean bleeding time in individuals with valve type 2300 was, however, also highly significant (p < 0.001). The tests were repeated in nine patients with bleeding times in the upper range, on average 8.8 minutes, after 2 to 12 months, in order to study the consistency of the prolonga­tion. The values were well reproducible, since the coefficient of variation of the difference was as low as 10.1.

The mean platelet count was moderately but significantly lower in both groups as compared with that of normal subjects (Table III), the range in healthy individuals being 140,000 to 309,000.

The mean platelet adhesiveness in normal subjects was 71.8 per cent and the S.D. 8.6, the normal range therefore being 55 to 89 per cent. Platelet adhesiveness was markedly reduced in valve patients and most pronounced in those with valves of type 2300 (Table IV). The difference between the mean values in the two patient groups was also highly significant (p < 0.001). The variability of the adhesiveness values was estimated by repeated testing after 1 week to 2 years in 17 patients. The mean adhesiveness in these was 28 per cent and the coefficient of variation of the differences was 28.2.

The approximate mean numbers of adhesive and nonadhesive platelets could be calculated, since the mean total platelet counts and the mean platelet adhesiveness were known in normal subjects and valve patients. Not only were the numbers of adhesive platelets reduced in the patients, but nonadhesive platelets were consider­ably increased.

An inverse linear relationship was found

Table V. The maximal rate of reversible platelet aggregation induced by 0.7 juM of ADP, and of irreversible aggregation initiated with 2.1 /xg per milligram of collagen in per cent of the normal response

No. of subjects

ADP-induced aggregation: Healthy

individuals Patients,

valve 1200 Patients,

valve 2300

52

13

32

Collagen-induced aggregation Healthy

individuals Patients,

valve 1200 Patients,

valve 2300

51

12

32

Aggregation <%of

Mean

100.1

96.5

90.2

99.9

98.9

86.4

normal)

S.E.M.

2.7

6.7

4.4

2.9

6.2

3.6

Compared with

normal

N. S.

N. S.

N. S.

p = 0.005

Table VI.Platelet half-life in patients with unop-erated aortic valvular disease and Starr-Edwards aortic ball valves.*

Healthy individuals" Unoperated valve disease Starr-Edwards aortic

valves

No. of subjects

20 7

17

Platelet half-life (days)

Mean

3.78 3.78 3.42

S.E.M.

0.06 0.24 0.13

*For comparison, the half-life in normal men between 40 and 70 years of age is calculated from Abrahamsen's" study.

between adhesiveness and bleeding time in 17 valve patients (Fig. 1), the correlation coefficient of the regression line being —0.57 (0.05 > p > 0.01). The bleeding time and actual number of adhesive platelets correlated equally well. A comparison between serum LDH levels and platelet adhesiveness was made in 46 patients with ball valves and a significant inverse correla­tion (p > 0.01) was found (Fig. 2).

The studies on platelet aggregation did not reveal marked deviations from the normal response. The slightly reduced rate of primary aggregation initiated with ADP in patients with valve type 2300 did not reach statistical signifi-

American Heart Journal 361

Dale and Myhre

80-1

70

60

w 50 0)

I ̂ ° I 20

10 T 1 1 1 1 1 1 1 1

5 5 7 8 9 10 11 12 13 1 2 3 4 B leed ing - t ime ,min.

Fig. 1. Relation between platelet adhesiveness and bleeding time in 17 patients with Starr-Edwards aortic ball valve prostheses. The formula for the regression line is: Adhesiveness = —4.28 x bleeding time +60.1. Coefficient of correlation: r = —0.57.

90

U) CO LLl Z LU > (/) LU I

n <

80

70

60

50

40

JU-

20-

10-

/ • •

100 200 3^0 4^0 500 sSo 700 800 900 lO'oO il'oo I2i'0 1300

L D H . U / L

Fig. 2. Relation between platelet adhesiveness and serum LDH levels.

cance. However, the mean maximal rate of irre­versible aggregation induced by collagen was moderately—but significantly—lower in these patients (Table V). Epinephrine elicited second­ary, irreversible aggregation in 64 per cent of the patients with prosthetic valves as compared to 89 per cent of the healthy individuals, a statistically significant difference (0.02 > p > 0.01).

The platelet half-life was insignificantly short­er in patients with ball valves than in those with unoperated-upon aortic valve disease (Table VI).

Lacking a normal material of our own, we compared our values with those of individuals between 40 and 70 years of age from Abraham-sen's'^ study. The mean value in these was exactly the same as in our own unoperated-upon patients, whereas a significant difference (p = 0.02) was found when a comparison was made with our valve patients. Since only three patients had valves of type 1200, a separate evaluation of the survival in these could not be made. The inverse correlation between platelet survival and degree

362 September, 1977, Vol. 94, No. 3

Platelet functions with aortic ball valves

Table VII. Comparison of platelet functions in ball-valve patients with and without arterial thromboembolic complications, with level of significance of the differences

Patients with thromboembolism

No. Mean S.E.M.

Patients without thromboembolism

No. Mean S.E.M. Significance

Platelets (per fil) Bleeding time (min) Adhesiveness (%) ADP-induced aggregation (%

of normal) Collagen-induced aggregation

(% of normal) Platelet half-life (days)

39 6

17 11

11

4

179,200 6.4

29.4 89.8

89.8

3.63

8,100 0.46 4.3 5.3

7.4

0.22

113 25 40 34

32

13

193,000 6.8

37.3 92.7

87.3

3.36

5,800 0.41 3.3 4.5

3.2

0.17

N. S N. S N. S N. S

N.S.

N. S.

Table VIII. Comparison of platelet functions in ball-valve patients with and without bleeding episodes, with level of significance of the differences

Platelets (per fA) Adhesiveness (%) Bleeding time (min.)

Patients with bleeding

No.

19 6 8

Mean

176,400 25.3 6.7

S.E.M.

Patients without bleeding

No.

13,900 133 4.0 67 0.55 24

Mean

191,300 35.4

6.7

S.E.M.

5,000 2.5 0.4

Significance

N . S . p < 0.05

of hemolysis as expressed by serum LDH was poor (r = 0.28, p > 0.10).

The incidence of arterial thromboembolic complications in the patients before' and after^" this study has been recorded. Platelet functions were compared in patients who suffered such complications and those who did not (Table VII). None of the differences observed was statistically significant. The number of observations was, however, so small in the group with thromboem­bolism with regard to some of the tests that a statistical evaluation was of limited value, since the investigation was performed in unselected patients.

Some patients developed bleeding episodes during the subsequent controlled clinical trial, mostly gastrointestinal bleeding in those receiv­ing acetylsalicylic acid in addition to anticoagu­lants.^" Platelet counts, adhesiveness, and bleed­ing time were compared in patients with and without bleeding episodes (Table VIII), and significantly lower mean platelet adhesiveness was found in those who later suffered from bleed­ing.

Discussion

The study has demonstrated abnormal platelet function in patients with aortic ball valve pros­

theses, with most marked disturbances of platelet adhesiveness and bleeding time, whereas platelet counts, irreversible platelet aggregation, and platelet survival deviated slightly from normal.

The mean value and distribution of platelet counts in healthy individuals are in good agree­ment with the findings of others.^' The slightly reduced mean platelet number in the ball-valve patients may be related to the moderately short­ened platelet survival. Decreased amounts of circulating platelets after localized thrombosis have been found by some^^ but not by others.^^

The mean normal bleeding time was approxi­mately 2 minutes shorter than that originally reported,''' but the method was slightly modified. Moreover, our normal values correspond well with those of others using comparable tech­niques,^' "-^^ and the precision as well as the reproducibility were acceptable."' " The prolon­gation of the bleeding time in patients with valves of type 2300 indicates a defect in the hemostatic mechanism that may render them prone to bleed­ing. Acetylsalicylic acid causes a marked addi­tional prolongation,^* and might therefore induce a particularly strong bleeding tendency in such patients.

Platelet adhesiveness was markedly reduced in our patients, whereas others^" have found normal

American Heart Journal 363

Dale and Myhre

values with a modified Salzman's method after mitral valve implantation. The inverse relation­ship between number of adhesive platelets and bleeding time is in accord with the original discovery of Hellem and associates," and supports the theory that adhesive platelets are important for primary hemostasis. This is further indicated by the relation between in vivo and in vitro adhesiveness," and by some similarities between the structure of the primary platelet plug™ and the platelet aggregates retained by glass bead columns.^^ Thus, the prolongation of the bleeding time observed is most probably due to the reduced number of adhesive platelets.

The primary aggregation induced by ADP probably partly reflects a similar platelet reac­tion as does the adhesiveness test, since platelet retention is also dependent upon available ADP," '" and since the structure of retained aggregates closely resembles that of platelets aggregated by ADP." However, the primary ADP-induced aggregation was normal in our patients whereas the retention was reduced. A likely explanation is that the adhesiveness test to a considerable extent reflects the in vivo reac­tivity of the platelets, whereas the primary aggre­gation does not. For adhesiveness measurements, native blood is used directly after careful with­drawal," and rough handling of the blood reduces retention considerably, probably through libera­tion of ADP.^^ Before testing aggregation, however, the blood is exposed to addition of citrate, to centrifugation, and finally to rather violent mechanical stirring. These procedures, in addition to the time from blood withdrawal until aggregation is recorded, most probably influence the platelets to such an extent that the original reactivity is seriously altered. The secondary, irreversible platelet aggregation is possibly less influenced by the procedure than the more sensi­tive primary aggregation. Irreversible aggregation has been reported to be normal in patients with mitral valve prostheses.^" The substance 2,3-diphosphoglycerate, which is present in red cells and liberated during intravascular hemoly­sis, inhibits the platelet release reaction and irre­versible aggregation,''^ and might influence aggre­gation in patients with prosthetic valves.

Platelet survival has been found to be normal in patients with unoperated-upon valvular disease.^^ Since our mean platelet half-life in such platelets was equal to the mean normal value in

Abrahamsen's'" study for the same age group, a comparison between the latter material and ours seemed justified. The moderate shortening of platelet survival in our valve patients is in good agreement with the results from other studies,'*^- ^̂ whereas the more pronounced reduction reported by one group" may be due to methodological differences. The exponential slopes of the survival curves from several patients indicate random disappearance of the platelets.

Three mechanisms for disturbed platelet func­tion in patients with prosthetic valves have been suggested'^: Refractoriness toward ADP liberated from red cells, consumption of adhesive platelets during thrombus formation, and mechanical trauma of platelets by passage through the valve. After exposure of platelets to ADP, further addi­tion of ADP combined with mechanical stirring results in a lowered aggregating response, the platelets having become refractory.^'''" Platelet retention in glass bead columns is dependent upon ADP from disrupted red cells," '" from platelets adhering to the beads and undergoing release,^" or from both sources. Retention is inhib­ited by addition of ADP prior to the testing,^^ by rapid transfer of blood between syringes or slow centrifugation,'^ or by intravenous infusion of water causing intravascular hemolysis." The destruction of red cells by the valves could possibly provide ADP in concentrations suffi­ciently high near the valves to induce platelet refractoriness. However, the rate of continuous hemolysis is low as compared to that reported after water infusions,** " ADP is immediately diluted by the blood stream, and it is rapidly cleared from plasma."^' " I t therefore seems un­likely that ADP liberation from red cells can explain the disturbances in platelet function.

Consumption of adhesive platelets by throm­bus formation could possibly induce a certain reduction in the number of these platelets, but not to the extent found in patients with valves of type 2300. Moreover, the incidence of arterial thromboembolic complications was lower in these patients than in those with valve type 1200.^ Finally, such consumption would not induce the actual increase of nonadhesive platelets seen in the patients.

The most important mechanism is therefore probably mechanical damage of the platelets by the prosthetic valves, which could produce all the changes observed in platelet behavior. A slight

364 September, 1977, Vol. 94, No. 3

Platelet functions with aortic ball valves

damage could reduce platelet adhesiveness, possi­bly the most delicate and vulnerable reactivity measured, in a large proportion of the platelets. A stronger trauma could cause reduced irreversible aggregation or removal of platelets from the circulation, giving slightly reduced platelet numbers and shortened survival. According to this, intravascular hemolysis and disturbed platelet function are parallel phenomena, in­duced by mechanical trauma of red cells and platelets, respectively.

Platelet functions were the same in patients who developed arterial thromboembolism and those who did not, but our material was too small with regard to platelet survival to allow conclu­sions. In a larger series of patients with Starr-Edwards aortic ball valves a slightly—but signifi­cantly—shorter survival was found in those who had thromboembolic episodes.̂ *

Our findings indicate that the occurrence of arterial thrombosis is not chiefly determined by individual differences in platelet reactivity, the important factors triggering thrombus formation probably being the foreign material represented by the valve itself,"- " and the turbulence of the blood stream caused by the valve.*" However, although the patients with valves of type 2300 have a higher degree of hemolysis and lower platelet adhesiveness than those with valve type 1200, they have a much lower incidence of arterial thromboembolic complications.^ Although the difference in thrombogenicity has been attributed to the cloth covering of the cage in the newer valves,' the low number of adhesive platelets could offer some protection against thrombus formation. Thus, the reduced platelet reactivity in patients with aortic ball valves may be regarded as a defense against arterial thrombu's formation.

Summary

Platelet functions were studied in normal subjects and patients with single Starr-Edwards aortic ball valves of series 1200 and 2300. The most pronounced changes were found in platelet adhesiveness, measured with Hellem's modified method. The mean percentage of adhesive platelets was reduced from 71.8 in normal subjects to 50.9 in patients with valve type 1200 and to 27.2 in those with type 2300. An inverse correlation was found between platelet adhesive­ness and the degree of intravascular hemolysis, as

reflected by serum LDH levels. The mean bleeding time was significantly prolonged in patients with valve 2300, and the individual values correlated inversely to the adhesiveness. The mean values of platelet counts, of irreversible aggregation induced by collagen or epinephrine, and of platelet survival were all moderately—but significantly—reduced as compared to normal. The most important mechanism behind the disturbed platelet reactivity is probably mechan­ical damage of the platelets by the valve, whereas refractoriness of platelets toward ADP liberated from red cells as well as consumption of adhesive platelets by thrombus formation is thought to have limited influence on platelet behavior. Platelet function was altered to the same extent in patients with a history of arterial thromboem­bolic complications as in those without. The disturbed platelet reactivity may predispose to bleeding, but may also offer some protection against arterial thromboembolism.

We are indebted to Dr. K. Rootwelt, Section of Nuclear Medicine, for the survival studies, and to Mrs. A.-L. Almaas and Mrs. Aa. Lund-Riise for skilled technical assistance.

REFERENCES

1. Dale, J.: Arterial thromboembolic compUcations in patients with Starr-Edwards aortic ball valve prostheses, AM. HEART J. 91:653, 1976.

2. Fleming, J., Hamer, J., Hayward, G., Tubbs, O. S., and Hill, I.: Long-term results of aortic valve replacement with the Starr-Edwards valve, Br. Med. J. 1:139, 1969.

3. Cleland, J., and Molloy, P. J.: Thrombo-emboUc compli­cations of the cloth-covered Starr-Edwards prostheses No. 2300 aortic and No. 6300 mitral, Thorax 28:41, 1973.

4. Arrigoni, M. G., Danielson, G. K., Mankin, H. T., and Pluth, J. R.: Aortic valve replacement with cloth-covered composite-seat Starr-Edwards prosthesis, J. Thorac. Cardiovasc. Surg. 65:376, 1973.

5. Spencer, F. C, Reed, G. E., Clauss, R. H., Tice, D. A., and Reppert, E. H. Cloth-covered aortic and mitral valve prostheses, J. Thorac. Cardiovasc. Surg. 59:92, 1970.

6. Friedh, B., Acrichide, N., Grondin, P., and Campeau, L.: Thromboembolic complications of heart valve pros­theses, AM. HEART J. 81:702, 1971.

7. Matloif, J. M., Collins, J. J., Sullivan, J. M., Gorlin, R., and Harken, D.: Control of thromboembolism from prosthetic heart valves, Ann. Thorac. Surg. 8:133, 1969.

8. Myhre, B., Dale, J., and Rasmussen, K.: Erythrocyte destruction in different types of Starr-Edwards aortic ball valves. Circulation 42:515, 1970.

9. Hellem, A. J.: The adhesiveness of human blood platelets in vitro, Scand. J. Clin. Lab. Invest. 12 (Suppl. 51):20, 1960.

10. Gaarder, A. M., Jonsen, J., Laland, S., Hellem, A. J., and Owren, P. A.: Adenosine diphosphate in red cells as a factor in the adhesiveness of human blood platelets, Nature 192:531, 1961.

11. Myhre, E., Hellem, A. J., Stormorken, H., and Rootwelt, K.: Intravascular haemolysis, platelet consumption and

American Heart Journal 365

Dale and Myhre

platelet adhesiveness in patients with prosthetic heart valves, Scand. J. Thorac. Cardiovasc. Surg. 5:13, 1971.

12. Stormorken, H.: Platelets, thrombosis and hemolysis. Fed. Proc. 30:1551, 1971.

13. Herr, R. H., Starr, A., Pierie, W. R., Wood, J. A., and Bigelow, J. C : Aortic valve replacement. A review of six years' experience with the ball-valve prosthesis, Ann. Thorac. Surg. 6:199, 1968.

14. Dale, J., Myhre, E., and Rasmussen, K.: Elimination of hemoglobin and lactate dehydrogenase from plasma in normal subjects and patients with intravascular hemoly­sis, Scand. J. Clin. Lab. Invest. 29:389, 1972.

15. Borchgrevink, C. F., and Waaler, B.: The secondary bleeding-time. A new method for the differentiation of hemorrhagic diseases. Acta Med. Scand. 162:361, 1958.

16. Hellem, A. J.: Platelet adhesiveness in von Willebrand's disease. A study with a new modification of the glass-bead filter method, Scand. J. Haematol. 7:374, 1970.

17. Bom, G. V. R.: Quantitative investigation into aggrega­tion of blood platelets, J. Physiol. 162:67, 1962.

18. Abrahamsen, A. F.: A modification of the technique for "Cr-labelling of blood platelets, Scand. J. Haematol. 5:53, 1968.

19. Abrahamsen, A. F.: Platelet survival studies in man, Scand. J. Haematol. Suppl. 3:14, 1968.

20. Dale, J., Myhre, E., Storstein, O., Stormorken, H., and Bfskind, L.: Prevention of arterial thromboembolism with acetylsalicylic acid. A controlled clinical study in patients with aortic ball valves. AM. HEART J. In press.

21. ten Gate, J. W., de Vries, S. I., Sixma, J. J., and van Berkel, W.: Reproducibility and normal values of various platelet function tests, Foha Med. Neerl. 15:10, 1972.

22. Moschos, C. B., Lahiri, K., Manskopf, G., Oldewurtel, H. A., and Regan, T. J.: Effect of experimental coronary thrombosis upon platelet kinetics, Thromb. Diath. Haemorrh. 30:339, 1973.

23. Baumgartner, H. R., Grimm, L., and Zbinden, G.: Effect of generation of mural platelet thrombi on circulating blood in rabbit aorta, Experientia 29:442, 1973.

24. Mielke, C. H., Ramos, J. C., and Britten, A. F. H.: Aspirin as an antiplatelet agent. Template bleeding time as a monitor of therapy. Am. J. CHn. Pathol. 59:236, 1973.

25. Harker, L. A., and Slichter, S. J.: Studies on platelet and fibrinogen kinetics in patients with prosthetic heart valves, N. Engl. J. Med. 283:1302, 1970.

26. Harker, L. A., and Slichter, S. J.: The bleeding time as a screening test for evaluation of platelet function, N. Engl. J. Med. 287:155, 1972.

27. Hellem, A. J., Borchgrevink, C. F., and Ames, S. B.: The role of red cells in haemostasis: The relation between hematocrit, bleeding time and platelet adhesiveness, Br. J. Haematol. 7:42, 1961.

28. Dale, J., Myhre, E., and Rootwelt, K.: Effects of dipyrid­amole and acetylsalicylic acid on platelet functions in patients with aortic ball valve prostheses. AM. HEART J. 89:613, 1975.

29. Weily, H. S., and Genton, E.: Altered platelet function in patients with prosthetic mitral valves. Effects of sulfin­pyrazone therapy. Circulation 42:967, 1970.

30. Hovig, T., and Stormorken, H,: Ultrastructural studies on the platelet plug formation in bleeding time wounds from normal individuals and patients with von Wille­brand's disease. Acta Pathol. Microbiol. Scand. (Sect. A, Suppl.) 248:105, 1974.

31. Hattori, A., Tsukada, T., Ito, S., Koike, K., and Matsuoka, M.: Scanning electron microscope study of platelet "adhesiveness" to glass beads (Hellem II meth­od). Normal subjects, the effect of aspirin and patients with von Willebrand's disease, Thromb. Diath. Hae­morrh, Suppl. 60:447, 1974.

32. Friedberg, N. M., and Zucker, M. B.: ADP as the cause of reversible inhibition of platelet retention in glass-bead columns, J. Lab. Chn. Med. 80:603, 1972.

33. latridis, S. G., latridis, P. G., Markidou, S. G., and Ragatz, B. H.; 2,3-Diphosphoglycerate: A physiological inhibitor of platelet aggregation. Science 187:259, 1975.

34. Steele, P., Weily, H., Davies, H. Pappas, G., and Genton, G.: Platelet survival time following aortic valve replace­ment, Circulation 51:358, 1975.

35. Stuart, R. K., McDonald, J. W., Ahuja, S. P., and Coles, J. C : Platelet survival in patients with prosthetic heart valves, Am. J. Cardiol. 33:840, 1974.

36. Born, G. V. R., and Cross, M. J.: The aggregation of blood platelets, J. Physiol. 168:178, 1963.

37. Simard-Duquesne, N.: Enhancement by ADP of throm-bin-induced aggregation in ADP-refractory platelets, Thromb. Diath. Haemorrh. 29:445, 1973.

38. Rozenberg, M. C , and Holmsen, H.: Adenosine nucleo­tide metabolism of blood platelets. IV. Platelet aggrega­tion response to exogenous ATP and ADP, Biochim. Biophys. Acta 157:280, 1968.

39. O'Brien, J. R.: Changes in platelet membranes possibly associated with platelet stickiness. Nature 212:1057, 1966.

40. McPherson, J., and Zucker, M. B.: Platelet retention in glass bead columns: Adhesion to glass and subsequent platelet-platelet interactions. Blood. 47:55, 1976.

41. Busfleld, D., and Tomich, E. G.: Acute intravascular hemolysis and platelet behaviour. Nature 218:696, 1968.

42. Brashear, R. E., and Ross, J. C : Disappearance of adenosine diphosphate in vivo, J. Lab. Clin. Med. 73:54, 1969.

43. Holmsen, H., and Stormorken, H.: Kinetic studies on the breakdown of adenosine diphosphate in human plasma, in Stormorken, H., editor: Studies in haematology, Oslo, 1965, Universitetsforlaget, p. 138.

44. Dragojevic, D., Hetzer, R., and Corterier, H.: Throm-boembolieprophylaxe nach Hertzklappenersatz in einem experimentellem Modell, Thoraxchir. Vask. Chir. 20:27, 1972.

45. Salzman, E. W., Berger, S., Merrill, E. W., and Wong, P. S. L. L. Thrombosis induced by artificial surfaces, Thromb. Diath. Haemorrh. Suppl. 59:107, 1974.

46. Stein, P. D., and Sabbah, H. N.: Measured turbulence and its effect on thrombus formation, Circ. Res. 35:608, 1974.

366 September, 1977, Vol. 94, No. 3