Viscoelastic Measurement of Clot Formation: A New Test of Platelet Function

Transcription

1 ANNALS OF CLINICAL AND LABORATORY SCIENCE, Vol. 13, No. 2 Copyright 1983, Institute for Clinical Science, Inc. Viscoelastic Measurement of Clot Formation: A New Test of Platelet Function ABDUS SALEEM, M.D.,* CINDY B LIFELD, M.D., SALWA A. SALEH, M.D., DAVID H. YAWN, M.D., MYLES L. MACE, Ph.D, MARY SCHWARTZ, M.D., and E. STANLEY CRAWFORD, M.D. Departments of Pathology, Pediatrics, Anesthesiology and Surgery, Baylor College of Medicine and The Methodist Hospital, Houston, TX ABSTRACT Assessment of platelet function presents a challenge to the investigators in the clinical field. An instrum ent, Sonoclot, is described which measures changes in the viscoelastic properties (clot im pedance on a vibrating probe) of plasma as it is recalcified. Platelet poor plasma (PPP) and platelet rich plasma (PRP) show distinct and characteristic Sonoclot tracings. Tracings of PPP show a lag period and a primary wave. T hese param eters correspond to recalcification tim e and fibrin polymerization. Tracings of PRP show, in addition, a secondary wave and a downward wave. These parameters correspond to the incorporation of platelets in the clot and retraction of the clot. The PRP param eters are influenced by the num ber and quality of the platelets. This instrum ent was utilized to assess the platelet function in patients after coronary artery bypass surgery (CAB). In 69 patients studied, 20 patients did not have excessive bleeding. The clot im pedance tracings in this group were normal. Twenty-four patients had excessive bleeding, normal clot impedance, and coagulation tests indicating the possibility of surgical bleeding. This was confirmed in 22 (92 percent) patients upon exploration to control bleeding. Twenty-five patients had excessive bleeding, normal coagulation tests and abnormal clot im pedance suggesting platelet dysfunction. In twenty-one (84 percent) of the patients, transfusion of platelet concentrate controlled the bleeding with corresponding correction of clot impedance. Sonoclot studies are quick and easy to perform, and appear to be a valuable and reliable adjunct in the diagnosis of hemostatic problems after CAB surgery. Introduction ated on for coronary artery bypass surgery (CAB), not only in large institutions In the last decade there has been an increase in the number of patients oper- but also in smaller community hospitals. Address reprint requests to Abdus Saleem, M.D., Im provem ent in surgical techniques have M.S. #205, The Methodist Hospital, 6565 Fannin, significantly reduced the m orbidity and Houston, TX mortality of this procedure. However, /83/ $01.50 Institute for Clinical Science, Inc.

2 1 16 SALEEM, BLIFELD, SALEH, YAWN, MACE, SCHWARTZ, AND CRAWFORD uncontrollable bleeding rem ains one of the serious com plications, and the incidence of life threatening hemorrhage has been reported to be betw een five to 20 percent. The cause of excessive bleeding may be surgical, hemostatic failure, or both. Several recent studies have em phasized platelet dysfunction as a major cause of hemostatic defect.4,5,6,8 9 Although recent advances in coagulation technology provide valuable inform ation on clotting factor defects, the evaluation of platelet dysfunction remains insensitive and time consuming. The platelet count provides a num erical b u t not qualitative index. Template bleeding time, although superior to older techniques of m easuring bleeding time, is subject to w ide variation in precision.7 Platelet aggregation tests are tim e consum ing and the interp retatio n of resu lts is g enerally subjective.2 The Thromboelastograph has been reported to measure platelet function.10 Its utility is lim ited by the requirem ent of large blood volumes and difficulty with data collection and interpretation.11 The system* provides an insight into platelet function. Basically, the instrument measures changes in the viscoelastic properties (clot im pedance upon vibrating probe) of plasma or whole blood as it is recalcified. The instrum ent is a simple viscoelastometer in series with an analog recorder. It has been used previously by von Kaulla to examine overall coagulability11 and, more recently, others have used it to m onitor heparin adm inistration.3 However, its use has not been generalized and it has not been used specifically to assess platelet function. In this paper, technical data are presented as well as theoretical considerations generated by clot impedance tracings and its application in m onitoring * Sonoclot, manufactured by Sienco Inc., Morrison, CO. p la te le t fu n ctio n follow ing coronary artery bypass surgery. Our specific objective was to determine: (1) if the instrument s tracings correlate with platele t function, and (2) if th e abnorm al tracings correlate w ith excessive bleeding in patients with platelet dysfunction, and if it normalizes after platelet transfusion. Materials and Methods T h e I n s t r u m e n t As previously described11 the instrum ent has a continuously vibrating probe (am plitude less than one micron, frequency less than 200 Hz) partially imm ersed in a m edium of static or changing viscoelastic properties. A mechanical impedance is im posed upon the probe by the medium. Any change in viscoelastic properties of the m edium is reflected in a change in mechanical im pedance. This change is converted to analog output and recorded on th e adjacent chart. A fter calibration against a set of standards, the instrum ent has the capacity to measure relative viscosity. T h e T e s t The blood was collected in a Vacutainer containing 3.8 p ercen t sodium citrate, maintaining a ratio of blood: anticoagulant of 9:1. T he specim en was placed in a centrifuge and spun at 100 x g for 10 m inutes to obtain p la te le t rich plasma (PRP). Further centrifugation of the rem aining contents at 1000 x g for 15 m inutes y ield ed p latelet poor plasm a (PPP). Unless specified, all tests w ere performed as follows: F o u r h u n d re d /u.1 o f sam p le w ere placed in a cuvet containing a stir bar and were equilibrated to 37 C for three minutes. Tw enty /u.1 of 0.5 M C ac l2 w ere added to the plasma. The mixture was stirred for 5 seconds and the probe was

3 then carefully low ered into the cuvet. The recorder was activated. N o r m a l R a n g e Tw enty healthy donors w ith no history of anticoagulant drug use or of a bleeding disorder were used to define the normal Sonoclot tracing and the effect of clotting factor and platelet concentration on the viscoelasticity of platelet poor plasma. E f f e c t o f C a l c i u m C o n c e n t r a t i o n A stock solution of 0.5 M CaCl2 was diluted in distilled water such that 20 /u.1 were added to each of test plasma to yield final concentrations of 0.006, 0.010, 0.020, and M CaCl2. E f f e c t o f T h r o m b i n C o n c e n t r a t i o n A stock solution of throm bin at 100 U p e r m l was m ade by re c o n stitu tin g throm bin in distilled water. Further dilutions were made such that 20 x\ were added to each 400 u.1 of test plasma to yield final throm bin concentrations of 0.16, 0.31, 0.63, and 1.25 U per ml. No CaCl2 was added. E f f e c t o f F a c t o r D e f i c i e n t P l a s m a Factor II, V, VII, VIII, IX, X, XI, XII, and XIII deficient plasm as w ere obtained. Four hundred /xl of each of these plasmas were recalcified with 20 1 of 0.5 M CaCl2. A 25 percent factor concentration of each of the previously m entioned factors was prepared by mixing three parts of factor deficient plasma w ith one part of pooled normal PPP. Each of these substrates was run as previously described. E f f e c t o f P l a t e l e t C o n c e n t r a t i o n A platelet count was obtained on a fresh platelet concentrate from the blood bank. This was appropriately d ilu ted VISCOELASTIC MEASUREMENT OF CLOT 1 17 with analogous plasma to obtain platelet counts of 0, 15, 30, 60, 88, or 250 K per ju.1. Each of these samples was run as previously stated. S o n i c a t e d P l a t e l e t s Platelet rich plasma adjusted to 250 K platelets per / I was sonicated in a Bonwill sonicator with a red probe for two intervals of 15 seconds each. Absence of intact platelets was confirmed by phase contrast microscopy. O u t d a t e d P l a t e l e t s Three and eight day old platelet concentrates obtained from the blood bank, adjusted to a platelet count of 250 K per / I with pooled PPP, were run as previously stated. P a t i e n t s Excessive bleeding after coronary artery bypass surgery is defined as blood loss of greater than 100 ml p er hour. Sixty-nine p atients, adm itted to T he M ethodist H ospital for coronary artery bypass surgery, w ere studied. All patients had a pre-operative screening test consistin g of b le e d in g tim e, p la te le t count, Prothrom bin tim e, (PT), partial throm boplastin time, (PTT), fibrinogen and clot im pedance tracing. Postoperatively, all patients were studied with an identical screen w ithin one hour, or at the tim e of excessive hemorrhage. Patients with abnormal pre-operative screens, history of oral anticoagulants, or antiplatelet drugs w ithin seven days prior to surgery were excluded. Patients who had excessive post-operative bleeding attributable to coagulation factor defects (coagulation factor deficiency, prolonged PT, PTT, and inadequate heparin neutralization) w ere also excluded from this study. In this study, 20 patients did not have excessive bleeding as defined previously

4 118 SALEEM, BLIFELD, SALEH, YAWN, MACE, SCHWARTZ, AND CRAWFORD (controls), while 49 patients showed excessive post-operative bleeding. Results C l o t I m p e d a n c e o f P l a t e l e t P o o r P l a s m a Following recalcification of PPP there was a lag period (A), followed by an upward deflection (primary wave, B). The primary wave reached a plateau at C as shown in figure 1. C l o t I m p e d a n c e o f P l a t e l e t R i c h P l a s m a In addition to lag period (A) and primary wave (B), the recalcification of PRP produced a shoulder (C). This was followed by a secondary wave (D) which, after reaching a peak (E), was followed by a downward deflection (F), (figure 1). N o r m a l R a n g e In table I are shown m ean ± 1 SD of various parameters of the clot im pedence data obtained from PPP and PRP of 20 normal donors. It is dem onstrated that for PRP the lag p erio d was significantly shorter and the primary slope was significantly steeper than those of PPP (p < 0.01). As shown earlier, the instrum ent tracing of PPP did not produce a secondary wave and, therefore, no shoulder peak interval. E f f e c t o f P l a t e l e t C o n c e n t r a t i o n o n C l o t I m p e d a n c e In figure 2 are shown a series of tracings at different platelet concentrations. It will be noted that (1) the lag phase decreased with increase in platelet count; (2) the slope of the primary wave increased with increase in platelet count; (3) the secondary wave made its appearance at 30 K p er fil p latelets; (4) the downward wave appeared at 60 K per / I platelets; and (5) with increasing platelet count, the slope of the secondary wave increased and the shoulder to peak interval decreased. TABLE I Normal Range of Parameters* Lag P e r io d S e c o n d s 1 Wave S lo p e c m /m in 2 Wave S lo p e cm/m in S h o u ld e r t o P ea k I n t e r v a l S e c o n d s Minutes F i g u r e 1. Comparison of clot im pedance of platelet rich plasma (PRP) and platelet poor plasma (PPP). A = lag period; B = primary wave; C = shoulder; D = secondary wave; E = peak; and F = downward wave. Platelet poor plasma Platelet rich plasma ± ± 45t 2.3 ± ± ± 30 Note: All values are expressed as mean ± 1 S.D. *Sonoclot (n = 20) tcompared with platelet poor plasma, p < 0.01

5 VISCOELASTIC MEASUREMENT OF CLOT 1 19 wave was apparent. No primary wave was obtained from PPP w hen the throm bin concentration was less than 0.16 U per ml. With higher concentrations of thrombin, i.e., 0.3, 0.6, or 1.2 U per ml, tracings of PPP showed both a lag period and a primary wave. Minutes F ig u r e 2. Effect of increasing concentration of platelets on Sonoclot tracings. A = platelet poor plasma; B = 15 K per / l; C = 30 K per /xl; D = 60 K per /il; E = 8 8 K per / l; and F = 250 K per /til. E f f e c t o f S t o r a g e o n P l a t e l e t s The clot im pedance from three day old blood bank PRP, adjusted to a platelet count of 250 K per /xl, was similar to a tracing obtained from fresh PRP. However, eight day old blood bank PRP produced a tracing resem bling PPP, showing only a lag phase and primary wave. E f f e c t o f S o n i c a t i o n Clot im pedance of plasma containing sonicated platelets showed a lag period and a prim ary wave. No secondary or downward wave was seen. E f f e c t o f C a l c i u m a n d T h r o m b i n C o n c e n t r a t i o n No primary wave was obtained from PPP when the calcium concentration was less than 0.01 M. With concentrations of 0.02 and M C ac l2, th e prim ary E f f e c t o f C l o t t i n g F a c t o r D e f ic ie n c y Clot im pedance from II, V, and VIII- XII deficient plasma did not produce a primary wave. A 25 percent of concentration of each of these clotting factors did produce a lag period and primary wave. However, in each case, the lag period was longer and primary wave slope less steep than normal. The tracings from VII and XIII deficient plasma produced a lag period and a primary wave which were not significantly different from the corresponding parameters obtained from normal PPP. P a t ie n t s All patients had normal pre-operative bleed in g tim es, PT, PTT, fibrinogen, p la te le t counts, and clot im p ed an ce tracings. In table II are shown the postoperative clinical and laboratory parameters. All patients in the Control group (Group A) had normal clot im pedance. In the group w ith in c re a se d p o st-o p erativ e blood loss, 24/49 patients had normal clot im pedance (Group B), while 25/49 had abnormal clot im pedance (Group C). The cardiopulmonary bypass time in all three groups did not show any significant differences. C o m p a r i s o n o f G r o u p B w i t h C o n t r o l s ( G r o u p A) Maximum post-operative bleeding per hour was significantly higher (p < 0.01) in group B patients (table II). The pum p time, bleeding time, platelet count, and

6 120 SALEEM, BLIFELD, SALEH, YAWN, MACE, SCHWARTZ, AND CRAWFORD TABLE II Post-operative Clinical and Laboratory Data on 69 Patients After Coronary Artery Bypass Surgery Maximum Tem plate Sonoclot P aram eters pump P o s t- o p. P l a t e l e t B le e d in g 1 Wave 2 Wave S h o u ld e r- Time B lood L oss Count Time L ag S lo p e S lo p e p e a k Time M in u te s m l/h o u r k / \ i l M in u tes Seconds cm /m inute cm /m inute Seconds Normal range Controls Group A (n = 20) Group B (n = 24) Group C (n = 25) N/A N/A 60 ± ±130* 275 ± ± ± ± :± t ± ± ± ± ± ± ± ± 25*+ 8.0 ± 4.2* 450 ± 60* * 1.7± 0.4* 325 ±40* Note: All values are expressed as mean ± 1 S.D. *p < 0.01 compared with controls. +p < 0.01 compared with normal. the lag, primary, and secondary wave, and shoulder to peak interval, did not show any statistical difference (p > 0.05). C l i n i c a l F o l l o w - u p o f G r o u p B P a t i e n t s B ecause o f excessiv e h em o rrh ag e, normal bleeding tim e, PT and PTT, and normal clot im pedance, surgical bleeding was suspected in all Group B patients. All patients in this group (24/24) were explored surgically. In 22/24 patients, one or more bleeding vessels w ere detected. These vessels w ere ligated, resulting in satisfactory hemostasis. In 2/24 patients, no bleeding vessels were found. However, the b leed in g tap ered off w ithin two hours of exploration. All patients in this group had blood loss replaced by packed red cells and fresh frozen plasma. C o m p a r i s o n o f G r o u p C w i t h C o n t r o l s ( G r o u p A) Maximum post-operative bleeding was significantly higher (p < 0.01) in Group C patients than in the Control Group. The platelet count was lower and bleeding time was greater than the Control Group. Clot im pedance data show ed abnormalities of all the parameters (table I). The lag phase was longer, the slopes of p rim ary an d seco n d ary w aves w ere lower, and shoulder to peak interval was longer than the Controls (p < 0.01). The clot im pedance abnormalities were more severe than would be expected for the level of platelet count. F o l l o w - u p o f G r o u p C P a t i e n t s Because o f prolonged bleeding tim e and clot impedance abnormalities, platelet dysfunction was suspected to be the cause of excessive bleeding. Ten units of platelet concentrate were adm inistered to all patients (25/25) in this group; the blood loss was replaced by packed red cells and fresh frozen plasma. In 18/25 patients, the b leeding was controlled within 30 minutes of the platelet transfusion. In figure 3 are shown clot impedance tracings of one of the patients before and after platelet transfusion. In 3/25 patients, the response was less dramatic; however, after platelet transfusion, th e b le e d in g ta p e re d off w ith in tw o hours. In 4/25 patients, th e b leed in g continued, and surgical exploration did not reveal any localized bleeding. The bleeding ultimately tapered off in four to

7 Minutes F i g u r e 3. C lot im pedance of p la te le t rich plasma from patient after coronary artery bypass surgery. A = after surgery; B = after transfusion with platelet concentrate. See text for details. 12 hours after exploration. The clot impedance parameters returned to normal in all patients (25/25) after platelet transfusion. Discussion As th e se d ata show, th e com plex coagulation process can be followed from thrombin generation through clot retraction. Clot im pedance tracing curves of PPP and PRP are distinct and reproducible. While the exact determ inants of all phases of the tracing are not known, the basis can be postulated of many of the im p ed an ce changes and som e o f the findings can be correlated w ith a clinical setting. The generation of the clot im pedance tracing is dependent upon multiple factors. Tracings of PPP demonstrate that VISCOELASTIC MEASUREMENT OF CLOT 121 the process begins with fibrin monomer formation, whether it be from the addition of exogenous throm bin or utilization of the throm bin generating capacity of calcium. Our experim ents indicate that the concentration of factors II, V, VIII, IX, X, XI, and XII affects the lag period and the prim ary wave param eters. In contrast, factor VII and XIII have no effect. The Sonoclot distinctly differentiates the PPP from PRP. In fact, the instrum ent has been utilized primarily for assessing platelet function. Q uantitative dependence is dem onstrated in the series of curves g e n e ra te d in fig u re 2. T hus, platelet count is a necessary param eter if the tracings are able to be correlated with implications about platelet function. The lag period of PRP is significantly shorter and the prim ary wave slope is significantly steep er than that of PPP (figure 1 and table I). In addition, PPP does not show the secondary and downward waves present in PRP. From the experim ent presented, it was noted by the present authors that secondary and downward waves made their appearance at a platelet concentration of 30 K and 60 K per / I, respectively. At higher platelet counts, the complex, consisting of the secondary and downward waves, became more compact. Electron microscopic studies of the clot at the tim e of production of the secondary wave revealed that the platelets became an integral part of the clot, and fibrin threads became attached to the platelet surface (figure 4). At about the same time, a visual inspection revealed the clot retracting from the sides of the cuvet. This phenomenon appeared to correspond to the stimulation of contractile protein of the platelets and the contraction of the cytoplasm toward the center of the platelet. It is believed that these events increased the clot density w hich is reflected in the production of the secondary wave. About the time the secondary wave peaked, the

8 SALEEM, BLIFELD, SALEH, YAWN, MACE, SCHWARTZ, AND CRAWFORD 122 F 4. Scanning electron micrograph of a recalcified normal platelet rich plasma showing attach ment of fibrin threads to clumps of activated platelets (x22,000). F 5. Scanning electron micrograph of a recalcified platelet rich plasma from a patient with platelet dysfunction showing poor attachment of fibrin threads to inactive platelets ( x22,000). ig u r e ig u r e

9 clot appeared to have com pletely retracted onto the probe. Thus, the system then consisted of a new probe, the clot plus th e original p robe, v ib ratin g in serum. Serum was a medium low of viscoelasticity. This resulted in production of the downward wave of the Sonoclot tracing. Absence of the secondary and downward waves in the tracing of eight day old blood bank and sonicated platelets indicated th at biologically active and intact platelets w ere necessary for this part of the tracing. The absence of the secondary wave below 30 K platelets per /a1 and absence of the dow nw ard wave below 60 K platelets per ul, roughly correlated with the clinical observation of increased risk of bleeding and prolonged bleeding time at low platelet counts. In figure 2 it is shown that the concentration of platelets influenced the lag period, the slopes of primary and secondary wave, and the duration of the secondary wave (shoulder to peak interval). Since the clotting factor deficiencies also affected the lag period and primary wave slope, the secondary wave slope and shoulder peak interval appeared to be specific for platelet function. The shoulder to peak interval was easily m easured on the chart recorder and may be the m ore conven ien t p aram eter to define platelet function. In our study, 20 patients did not bleed excessively (less than 100 ml per hour) following CAB, and had normal PT, PTT, bleeding tim e and Sonoclot parameters. The platelet count was decreased (200 ± 38 K per il). In 24 patients, surgical bleeding was suspected because of excessive bleeding in face of normal coagulation values, b leed in g tim e, and clot impedance. In 22 (92 percent) patients, surgical bleeding was confirm ed upon re-exploration. In two patients (8 p ercent), the tracings w ere normal, and no surgical bleeding was found. This could represent either failure to predict correctly or difficulty in locating a bleeding VISCOELASTIC MEASUREMENT OF CLOT 123 vessel. In 25 patients w ith excessive bleeding following CAB, p latelet dysfunction was suspected because of abnormal bleeding tim e and/or clot impedance and normal coagulation values. Electron microscopic study of the patients w ith severe platelet dysfunction showed that platelets participated poorly in clot formation and interaction betw een fibrin and platelets was suboptimal (figure 5). Tw enty -one (84 percent) responded clinically to platelet transfusion, and the tracings returned to normal. In 4 (16 percent) patients, the response to platelet transfusion was either delayed or inconclusive despite im provem ent in the viscoelastic abnorm alities of the clot. No surgical bleeding was dem onstrated in these cases. These patients represented a diagnostic dilemma. It is probable that a com bination of surgical defect, platelet dysfunction, and/or vascular dysfunction, was responsible for excessive bleeding. Several investigations have used bleeding tim e as a prim ary test for platelet dysfunction.4,5,8 In our study, although the mean bleeding time was significantly prolonged in patients with platelet dysfunction, several patients had bleeding tim e w ithin normal limits. The shoulder to peak interval dem onstrated a m uch better separation of the groups with normal or abnormal platelet function. Moreover, in peri-operative patients with several intravenous lines, it is often difficult to find a site to do a b leed in g tim e. Often the intravenous infusion had to be discontinued temporarily or the shoulder area had to be used w ithout tourniquet to do bleeding times. The study was relatively easy to perform, and th e resu lts w ere gen erally available w ithin 20 m inutes after the specim en was received in the laboratory. Since PT, PTT, and platelet count results w ere also available within a short tim e, it was possible to indicate whether the patient had a surgical or hemostatic defect. A quick answer to this difficult question

10 1 2 4 SALEEM, BLIFELD, SALEH, YAWN, MACE, SCHWARTZ, AND CRAWFORD a le rte d th e e n tire team of surgeons, anesthesiologists, blood bankers, and coagulationists to institute specific therapy for life threatening hemorrhage. References 1. B ic k, R. L.: Alteration of hemostasis associated with cardiopulmonary bypass. Pathophysiology, prevention, diagnosis and management. Sem. Thromb. Haemost. 3:59 82, C r o n b er g, S.: Evaluation of platelet aggregation. Coagulation 3: , E n s, G. and H e m s t r a, R.: ReAct test. Recalcification of activated whole blood using the Sonoclot. Thromb. Haemost. 38:282, F r i e d e n b e r g, W. R., M y e r s, W. O., P l o t k a, E. D., B e a t h a r d, T. N., K u m m e r, D. J., G a t l in, P. F., S t o ib e r, D. L., R a y, J. F., Ill, and S a u t t e r, R. D.: Platelet dysfunction associated with cardiopulmonary bypass. Ann. Thorac. Surg. 25: , H a r k e r, L. A., M a l p a s s, T. W., B r a n s o n, H. E., H a s s e l, E. A., II, and S l ic h t e r, S. L.: Mechanism of abnormal bleeding in patients undergoing cardiopulmonary bypass: Acquired transient p la te let dysfunction associated w ith selective granule release. Blood 56: , H e n n e s s y, V. L., H ic k s, R. E., N ie w ia r o w s k i, S., E d m u n d s, L. H., Jr., and C o l e m a n, R. W.: Function of human platelets during extracorporeal circulation. Amer. J. Physiol. 232: , H i r s h, J., B l a jc h m a n, M., and K a e g i, A.: The bleeding time. Platelet Function Testing. Day, H. J., Holmsen, H., and Zucker, M. B., eds. Washington, U.S. Department of Health, Education and Welfare, 1978, pp M c K e n n a R., B a c h m a n n, F., W h i t t a k e r, B., G i l s o n, J. R., and W e i n b e r g, M., Jr.: The hem ostatic m echanism after open heart surgery. II. Frequency of abnormal platelet function during and after extracorporeal circulation. J. Thorac. Surg. 70:76-85, U m l a s, J.: In vivo platelet function following cardiopulmonary bypass. Transfusion i5 : , von Kau lla, K. N.: Quantitative methods for recording blood coagulation: theoretical and practical aspects. Progress in Hematology III, New York, G rune and Stratton, 1962, pp v o n K a u l l a, K. N., O s t e n d o r f, P., and v o n K a u l l a, E.: The impedence machine: A new bedside coagulation recording device. J. Med. 6:73-87, 1975.