Effect of Optimization of Hemodynamics onFibrinolytic Activity and Antithrombotic Efficacyof External Pneumatic Calf Compression

EDWIN W. SALZMAN, M.D., GERALD P. McMANAMA, M.D., ASCHER H. SHAPIRO, Sc.D., LINDA K. ROBERTSON, R.N.,ANN S. DONOVAN, R.N., HOWARD W. BLUME, M.D., JUDY SWEENEY, R.N., ROGER D. KAMM, PH.D.,MARK C. JOHNSON, S.M., and PETER McL. BLACK, M.D.

External pneumatic calf compression is effective but imperfectfor antithrombotic prophylaxis in surgical patients. In prelimi-nary studies, sequential filling of multisegmented leggings withgraded pressure decreasing from ankle to knee increasedvenous flow velocity and wall shear stress, decreased residualvenous volume, and enhanced postoperative fibrinolysis morethan uniform compression. To determine if improved hemody-namics also increased antithrombotic activity, we performed aprospective randomized trial in neurosurgical patients com-paring sequential application of graded pressure with uniformpressure applied to either a segmented bladder or to a singlebladder. Deep vein thrombosis was diagnosed by leg scanningand impedance plethysmography and confirmed by phlebo-graphy. Venous thrombosis developed in 3 of 45 patients withgraded-sequential filling, 6 of 50 with uniform compression-multiple compartments, and 3 of 41 with uniform pressuresingle bladder (differences not significant). These results sug-gest either that uniform compression offers all that can beexpected of external pneumatic calf compression in preventionof venous thrombosis, or that even if a study with greater sta-tistical power showed graded-sequential filling to be superior,the benefit/cost ratio of the more complex latter system is notlikely to be large.

A LTHOUGH THE INCIDENCE of deep vein throm-bosis in postoperative patients can be reducedby prophylactic administration of antithrom-

botic drugs, such preventive measures have not beenuniversally adopted into practice. Several surveys'13have reported that many practitioners use prophylacticantithrombotic drugs only in their highest risk patients,if they use them at all. The efficacy of prophylactic ad-ministration of anticoagulants and antiplatelet agents isdisputed, and is certainly not absolute; failures do occur.

Supported by Grant HL13754 from the National Heart, Lung, andBlood Institute.

Reprint requests: Edwin W. Salzman, M.D., Beth Israel Hospital,330 Brookline Avenue, Boston, MA 02215.Submitted for publication: March 9, 1987.

From the Department of Surgery, Beth Israel Hospital, andthe Charles A. Dana Research Laboratories, Harvard

Medical School, Boston, Massachusetts, and theDepartment of Mechanical Engineering, Massachusetts

Institute of Technology, Cambridge, Massachusetts

The requirement for laboratory control, at least withwarfarin, is also a deterrent. The most prominent obsta-cle to wider adoption of antithrombotic prophylaxis,however, has been the fear of bleeding complications,which inevitably to some degree attend the use of anti-thrombotic drugs in surgical patients.An alternative approach, attractive because of its

freedom from hemorrhagic side effects, is the use ofphysical methods to reduce pooling of venous blood inthe lower extremities and to increase the velocity ofvenous blood flow. Such physical techniques includeelastic stockings, passive leg exercises, electrical calfstimulation, and external pneumatic compression ofthelegs. The latter has been most extensively studied andappears to be most effective among the physicalmethods available.A substantial literature has developed regarding the

use ofexternal pneumatic compression boots or leggingsin patients at risk of venous thromboembolism, includ-ing patients undergoing neurologic operations,4-6 uro-logic surgery,7'8 general surgical procedures for malig-nant or benign disease,9"10 operations for gynecologicmalignancy,"I orthopedic procedures,'2"13 and nonoper-ative conditions. 14 The mechanism of action of externalpneumatic compression in reducing the rate of venousthrombosis in the lower limbs is conjectured to be two-fold: a fluid mechanical effect expelling blood from thelower extremities to substitute for the action of the in-operative muscle pump in the recumbent patient, andenhancement of fibrinolytic activity.'5 The relative in-

636

EXTERNAL PNEUMATIC CALF COMPRESSION

fluence of these two factors is not certain. Externalpneumatic compression appears to be associated withno bleeding complications even in patients undergoingextensive surgery.Although external pneumatic compression of the legs

has proven efficacious in reducing the rate of venous

thrombosis, there are occasional failures associated withits use, particularly in high risk patients. In our trial ofcalf compression in neurosurgical patients,4 for exam,

ple, 9% of patients who were treated with external pneu-matic compression developed deep vein thrombosis(compared with 25% in the control group). As conven-

tionally applied, external pneumatic calf compressionmay not produce an optimal hemodynamic effect. It ispossible that an improvement in the fluid mechanicalaspects of the procedure, brought about by modifica-tions in the apparatus, would be reflected in an increasein its antithrombotic effectiveness.

Shapiro and associates have conducted an extensivestudy of the hydrodynamics of external pneumaticcompression, first in theoretical and experimental cir-culation models'6 and later in human volunteers usinggated blood pool scans of the lower extremities.'7 Theycompared the amount and rate of blood expelled whendifferent modes of intermittent external pneumaticcompression were applied to the lower extremities.Scintillation camera imaging ofthe labelled red cell poolsynchronized to the pressurization cycle provided dataon the time course ofchanges in regional blood volumesin the leg. They calculated regional values ofthe fractionof blood ejected as well as comparative indices propor-tional to flow rate, blood velocity, and wall shear stress,all of which might reasonably be regarded as figures ofmerit to be maximized in an attempt to obtain optimalprophylaxis against deep vein thrombosis. Using a four-compartment cuff, they studied compression modeswith different intercompartmental pressure gradationsand different intercompartmental intervals to the onsetof pressurization. The most effective system hemody-namically was found to be a four-compartment leggingin which the bladders filled in sequence beginning at theankle with graded pressures decreasing from ankle toknee. The optimal values for a four-compartment sys-

tem were approximately 60, 50, 40, and 30 mmHg, de-creasing from ankle to knee, with about 0.25 sec be-tween pressurizations of successive segments, a pressurerise time of less than 0.5 sec, each segment held at fullpressure for about 10 sec, and a period of about 50 sec

with pressure relieved before the next cycle. No com-

mercially available system is hemodynamically optimalor conforms to these features.We tested the hypothesis that an external pneumatic

compression device with these characteristics would im-prove antithrombotic efficacy in patients undergoing

neurosurgical operations, a high-nrsk group that cannotsafely be given antithrombotic drugs because ofthe haz-ard of intracranial bleeding complications. A corollaryof this hypothesis would be that graded-sequential fillingof such a specially designed apparatus would have suffi-ciently improved antithrombotic properties to justify itsincreased cost and complexity. We also studied the ef-fect of different modes of calf compression on wholeblood fibrinolytic activity to explore further the mecha-nism ofthe antithrombotic action ofexternal pneumaticcompression.

Methods

The trial was initiated as a two-armed clinical studywith randomization afforded by a series of sealed enve-lopes. Patients were assigned either to receive uniformcompression delivered by a four-compartment systemapplied to the calf (Group II), with all compartmentsfilled simultaneously to the same pressure, or to havesequential filling of the same system from ankle to kneewith graduated pressure decreasing from ankle to knee(Group I) according to specifications derived from thestudies ofKamm et al.'7 with pressure and timing valuesas stated above. The apparatus was furnished by Gay-mar Industries, Inc., Orchard Park, NY. The four-seg-ment cuffs had "wavy" boundaries between segments, a

configuration that effectively eliminated the pooling ofblood that, in the scintillation studies of Kamm et al.,had been observed to occur between segments of moreconventional four-segment cuffs.The study was conducted in patients undergoing neu-

rosurgical operations at the Beth Israel Hospital. A de-scription ofthe patient population is presented in Table1. External pneumatic compression was applied duringthe induction of anesthesia, except for two patients (1 inGroup I and 1 in Group II) in whom confinement to bedbefore operation suggested the need for earlier prophy-laxis. In these cases the external pneumatic compressiondevice was applied as soon as the patient came to theattention of the investigators.

Diagnosis of deep vein thrombosis was made by 1251I-fibrinogen scanning," which was begun on the first post-operative day, impedence plethysmography (IPG)'9every second day, and ascending phlebography by thetechnique of Rabinov and Paulin20 in patients who hadpositive fibrinogen scans or IPGs. In addition, blood wasdrawn before operation and at intervals in the postoper-ative period for determination of fibrinolytic activity inheparinized whole blood by a radioactive clot lysis tech-nique modified by Coppe et al.2' from the method ofMoroz.22 General surgical patients undergoing opera-tions for nonthrombotic conditions were also studied as

controls for the fibrinolysis assay. Patients were main-

Vol. 206 * No. 5 637

SALZMAN AND OTHERS

TABLE 1. Patient Groups

No. of patientsSex (M/F)Age 10-20

2 1-3031-4041-5051-606 1-7071-80>80

Surgical procedureCraniotomy fortumor

ShuntsOther craniotomiesCervical laminectomyLumbar laminectomySpinal fusion

Risk factors for DVTPrevious DVT or PEObesityVaricose veinsEACA-

Group IGraded-Sequential

4524/21

0

91088622

4697

136

41031

Group II

UniformPressure,Four

Compartment

5023/27

0

481685

81

6599147

820

tained on external pneumatic compression until theybecame ambulatory or ror a maximum of the periodspanned by three injections of labelled fibrinogen,roughly 3 weeks.

After 43 patients had been entered into the trial,Group III receiving uniform compression with a singlebladder system (Venodyne, Lyne-Nicholson, Inc.), wasadded to provide a realistic comparison of graded-se-quential filling versus a commercially available systemhaving uniform pressurization. The randomization was

adjusted to balance the three patient groups. It should benoted that the Venodyne system has a substantiallylonger pressure rise time than do the other systemsstudied.For ethical reasons no unprotected control group was

included in this study, which instead compared a tech-nique ofestablished effectiveness versus another form oftreatment that was potentially superior. This protocolwas approved by the Institutional Review Board at theBeth Israel Hospital.

Results

One hundred fifty-eight subjects were enrolled in thestudy, but 22 were dropped prior to its completion be-cause of nonadherence to the study protocol in 13 pa-tients, including two who refused the procedure becauseof discomfort, nonoperative therapy (5 patients), or fail-

ure of the apparatus (4 patients). Forty-five of the re-maining 137 patients were in Group I, 50 in Group II,and 41 in Group III. The surgical procedures, age, sex,and other relevant clinical details are shown in Table 1.The groups appeared to be comparable in all importantrespects, except that patients in Group I had a somewhathigher prevalence of risk factors for venous thrombosis.Compression of the limbs is known to enhance sys-

temic fibrinolytic activity, which might augment the an-tithrombotic action ofexternal pneumatic compression.We studied whole blood fibrinolytic activity in the first75 patients. A fall in fibnrnolytic activity is characteristicofpostoperative patients, beginning in the first 24 hoursafter operation and reaching its nadir by the third post-operative day.23'24 Such a reduction in fibrinolytic activ-ity was exhibited by the group of general surgical pa.tients studied as controls (Fig. 1; p < 0.01 prdbp vs.postop day 1, Friedman's rank sum test). Patients whoreceived external pneumatic compression of the limbswith a uniform compression device, either in Group II

or Group III, failed to display this expected fall in fibri-nolytic activity in the postoperative period and, in fact, 9of 16 patients (56%) with the single bladder system and21 of 35 (60%) with the four-compartment uniformcompression system showed an increase in the fibrino-lytic index on the first postoperative day (Figs. 1 and 2).

Patients in Group I had an average increase in fibrino-lytic activity that was significantly greater in magnitudethan for those in Groups II and III (mean fibrinolyticindex on first postop day 1.45 for graded sequentialgroup vs. 1.06 for single bladder system and 1.12 formultiple bladder system with uniform compression; p< 0.02 by Dunnett's test or Kruskal-Wallis test), and 23of 25 patients (92%) with graded sequential filling had arise in activity. Thus the effect on fibrinolytic activitywas correlated with the hemodynamic effectiveness ofthe external pneumatic compression systems. Patientswho failed to be protected by external pneumatic com-pression, and who developed venous thrombosis despiteits use, typically also failed to develop enhanced fibrino-lytic activity as a result of the application of externalpneumatic compression after operation. Six ofthe eightsuch patients whose fibrinolytic index was assessed onthe first postop day had values below their group mean(Fig. 2). Patients whose level of fibrinolytic activity wasincreased did not exhibit any clinically obvious bleedingtendency as a result, but blood loss was not measuredand this issue was not rigorously appraised.The effect of the treatments on the incidence of deep

venous thrombosis is shown in Table 2. All patients whodeveloped a positive fibrinogen scan or IPG were re-quested to have a phlebogram to confirm the diagnosisof venous thrombosis, but five refused. Fifteen patientsdeveloped a positive fibrinogen scan; ten of these had a

638 Adin. Surg. * November 1987

Vol. 206 * No. 5 EXTERNAL PNEUMA'

phlebogram, three ofwhich proved to be negative. Thesethree patients are regarded as "false positives," althoughthe positive scan may have been related to venousthrombosis in a muscular tributary of the deep venoussystem that was not visible on phlebography. For pur-poses of analysis, "positive" patients are considered tobe those who have a positive scan with a positive (con-firmatory) phlebogram plus those who had a positivescan but did not have a phlebogram.4 The results ofsuchan analysis suggest a trend toward a greater antithrom-botic effect for graded sequential filling than for the uni-form pressure four-compartment system, but no differ-ence from the single bladder uniform compression sys-tem. The differences are not statistically significant. The95% confidence interval for the difference betweengraded sequential compression and uniform compres-sion with multiple segments is from -16.9% to 6.3%,and for the difference between graded sequential com-pression and uniform compression with a single bladderit is from -1 1.4% to 10.2%. The fact that these intervalsinclude zero is consistent with the lack of statistical sig-nificance.25

FibrinolyticIndex

(Post- Op)

POD1 POD4 POD7FIG. 1. Fibrinolytic activity in the three patient groups on the first,fourth, and seventh postoperative days. Analogous values are alsoshown for general surgical patients undergoing operations ofcompara-ble severity ("no boots"). Results are expressed as a "fibrinolyticindex" (cpm released into supernatant divided by cpm originally con-tained in radioactive clot) and are normalized to preoperative valuesfor each patient.

TIC CALF COMPRESSION

3.00

2.50FibrinolyticIndex

2.00(Post -op,day I )

1.00

0.50

639

No Boots SingleBladderUniformCompression

MultipleBladderUniformCompression

GradedSequentialCompression

FIG. 2. Fibrinolytic index on the first postoperative day in the fourpatient groups. Horizontal bars indicate mean values. Closed circlesindicate patients who subsequently developed deep vein thrombosis.

The same conclusion will be reached if one considersas "positive" only those patients who had a positive scanwith a positive (confirmatory) phlebogram or if one in-cludes as positive all patients with a positive scan regard-less of the phlebogram.We calculated that if the present rate of venous

thrombotic complications were maintained, approxi-mately 500 patients in each group would be required toshow a statistically significant difference between GroupI and the other groups with an alpha error <0.05 and abeta error of 0.20.26 Because this was considered im-practical the study was terminated at this time.

Discussion

External pneumatic compression is an establishedform of prophylaxis for prevention of venous thrombo-sis in postoperative patients. The relative importance ofenhanced fibrinolysis and of physical dispersion of thecomponents of a thrombus by hemodynamic factors incalfcompression is not known. In this study they tendedto vary together. The failure of fibrinolytic activity to beenhanced by pneumatic compression in patients whosubsequently developed venous thrombosis was predict-able. The association oflong whole blood clot lysis timeswith subsequent venous thrombosis in surgical patients

o No DVT* DVT

c

0

8

0~~~~~o~~~~

0

o 8 8

___g -- - o- - - - - C-_ __ _ _I _ _

0 0

§CSc o

1.50

SALZMAN AND OTHERS

TABLE 2. Deep Vein Thrombosis

Group I Group II Group IIIGraded Sequential Uniform Pressure, Uniform Pressure,

Compression Multiple Compartment Single Bladder

N 45 50 41A. + scan 2 3 2

+ venogramB. + scan 1 3 1

no venogramC. + scan 0 2 1

- venogramDVT: A + B 3/45 (6.7%) 6/50 (12%) 3/41 (7.3%)

A + B + C 3/45 (6.7%) 8/50 (16%) 4/41 (9.8%)A only 2/45 (4.4%) 3/50 (6%) 2/41 (4.9%)

Development of DVT, postop days 2, 4, 7 1, 2, 5, 5, 12, 13 2, 2, 3DVT, (A + B), operation Craniotomy for tumor Craniotomy for tumor (2) Craniotomy for tumor

Laminectomy (2) Craniotomy for trauma LaminectomyOther craniotomy (2) Hanington rodLaminectomy

was observed by Comp and associates,27 and Clayton etal.28 found low levels of fibrinolytic activity to be a sen-

sitive predictor of venous thrombosis in patients under-going gynecologic operations.The evidence for the efficacy of pneumatic compres-

sion is persuasive, but calf compression might be even

more effective if its hemodynamic effects were opti-mized. Among the four fluid mechanical parameters in-fluenced by external pneumatic compression (flow velo-city, volume flow rate, shear stress, and residual volumeof blood in the limbs), it is not yet possible to decidewhich is most important for antithrombotic effective-ness of the system; however, since under normal cir-cumstances they all tend to vary together, this point maybe moot. It is clear that commercially available systemsfor applying external pneumatic compression to thelower limbs are not optimized in these respects.We found that a system for external pneumatic com-

pression in which the pressure was applied in a gradedfashion, milking the blood from ankle to knee, was moreeffective hemodynamically, and in the clinical trial itproved to be more effective in enhancement of fibrino-lytic activity than uniform compression applied to theleg either in a multiple bladder system or in a singlebladder. Nicolaides et al.29 have previously reported thata multiple bladder system involving sequential fillingfrom ankle to thigh but with uniform pressure was more

effective in accelerating femoral venous flow and was

superior to a single bladder system in preventing proxi-mal venous thrombosis in patients undergoing abdomi-nal operations; however, there was no significant differ-ence in the rate of calf thrombi. Since the sequentiallyfilled device in their trial had thigh cuffs as well as calfcuffs and the single bladder system had only calf com-

pression, it is not certain that the difference in the fre-quency of proximal (thigh) thrombi was attributable tothe sequential filling. In a subsequent study, Nicolaideset al.30 compared the sequentially filled device appliedfor at least 3 days and followed by elastic stockingsversus electrical calf stimulation versus low-dose hepa-rin, and they found the results of external pneumaticcompression and heparin to be roughly equivalent. Inthe system employed by Nicolaides et al., the fillingpressures, rise time, duration of filling, and interval be-teen successively filled bladders were different from thecorresponding values employed in the optimized devicewe have studied, which is not duplicated by any com-

mercial unit now available.Despite the improved hemodynamic effectiveness of

the graded compression sequentially filled system andthe superior biologic effects reflected by enhanced fibri-nolytic activity, we were unable to show greater anti-thrombotic effectiveness for this device. There were no

significant differences in the frequency ofvenous throm-bosis in the three patient groups. Two possible explana-tions are: (1) regardless of the difference in hemody-namic effects, the difference in antithrombotic effective-ness among the three modes of external pneumaticcompression is negligible, and the uniform compressionsystem accomplishes all that can be expected of externalpneumatic compression in respect to prevention ofvenous thrombosis, or (2) a real difference exists, withgraded sequential compression being superior to uni-form compression, but because ofthe limited number ofpatients or for other reasons, the study was insufficientin statistical power to show the difference (i.e., there is apossibility of a beta error).31 Even if the latter is true,however, the difference in efficacy of the three forms of

640 Ann. Surg. - November 1987

Vol.206 * No.5 EXTERNAL PNEUMATIC CALF COMPRESSION 641calf compression is probably not large, and one mustquestion whether the increment in clinical benefitwould be worth the additional cost and complexity ofengineering and producing the graded sequential fillingsystem.

It remains possible that, in a population at evenhigher risk than the neurosurgical patients (such as thosewith reconstructive operations on the hip), a hemody-namically superior system applying graded pressurewith sequential filling might have demonstrably betterantithrombotic efficacy than systems in which fluid me-chanical features are not optimized.

AcknowledgmentsThe cooperation of Gaymar Industries, Inc., is gratefully acknowl-

edged. Dr. McManama was a Fellow of the American Society forArtificial Internal Organs.

References1. Simon TL, Stengle JM. Antithrombotic practice in orthopaedic

surgery. Clin Orthop 1974; 102:181-187.2. Morris GK. Prevention ofvenous thromboembolism. A survey of

methods used by orthopaedic and general surgeons. Lancet1980; 572-574.

3. Conti S, Daschbach M. Venous thromboembolism prophylaxis.Arch Surg 1982; 117:1036-1040.

4. Skillman JJ, Collins REC, Coe NP, et al. Prevention ofdeep veinthrombosis in neurological patients: a controlled, randomizedtrial of external pneumatic compression boots. Surgery 1978;83(3):354-358.

5. Turpie AGG, Gallus AS, Beattie WS, Hirsh J. Prevention ofvenous thrombosis in patients with intracranial disease by in-termittent pneumatic compression ofthe calf. Neurology 1977;27(5):435-438.

6. Turpie AGG, Delmore T, Hirsh J, et al. Prevention of venousthrombosis by intermittent sequential calf compression in pa-tients with intracranial disease. Thromb Res 1979; 15:611-616.

7. Coe NP, Collins REC, Klein LA, et al. Prevention of deep veinthrombosis in urological patients: a controlled, randomizedtrial of low-dose heparin and external pneumatic boots. Sur-gery 1978; 83:230.

8. Salzman EW, Ploetz J, Bettman M, et al. Intraoperative externalpneumatic calf compression to afford long-term prophylaxisagainst deep vein thrombosis in urological patients. Surgery1980; 87(3):239-242.

9. Hills NH, Pflug JJ, Jeyasingh K, et al. Prevention of deep-veinthrombosis by intermittent pneumatic compression of calf. BrMedJ 1972; 1:131.

10. Roberts VC, Cotton LT. Prevention of postoperative deep veinthrombosis in patients with malignant disease. Br Med J 1974;1:358.

11. Clarke-Pearson DL, Synan IS, Hinshaw WM, et al. Prevention ofvenous thromboembolism by external pneumatic calf com-pression in patients with gynecologic malignancy. Obstet Gy-necol 1984; 63:92-98.

12. Harris WH, Athanasoulis CA, Waltman AC, Salzman EW. Pro-phylaxis of deep-vein thrombosis after total hip replacement. JBone Joint Surg 1985; 67-A(l):57-62.

13. Hull R, Delmore TJ, Hirsh J, et al. Effectiveness of intermittentpulsatile elastic stockings for the prevention of calf and thighvein thrombosis in patients undergoing elective knee surgery.Throm Res 1979; 16:37-45.

14. Salzman EW, Sobel M, Lewis J, et al. Prevention of venousthromboembolism in unstable angina pectoria. N Engl J Med1982; 306:991.

15. Knight MTN, Dawson R. Effect of intermittent compression ofthe arms on deep venous thrombosis in the legs. Lancet 1976;2:1265.

16. Olson DA, Kamm RD, Shapiro AH. Bioengineering studies ofperiodic external compression as preophylaxis against deepvein thrombosis. Part II: experimental studies on a simulatedleg. ASME, J Biomech Eng 1982; 104.

17. Kamm R, Butcher R, Froelich J, et al. Optimization ofparametersof external pneumatic compression for prophylaxis againstdeep vein thrombosis: radionuclide gated imaging studies. Car-diovasc Res 1986; 20:588-596.

18. Flanc C, Kakkar VV, Clarke MB. The detection ofvenous throm-bosis of the legs using 125I-labelled fibrinogen. Br J Surg 1968;55(10):742-747.

19. Wheeler HB, Pearson D, O'Connell D, et al. Impedance phlebo-graphy. Technique, interpretation and results. Arch Surg 1972;104:164.

20. Rabinov K, Paulin S. Roentgen diagnosis ofvenous thrombosis ofthe leg. Arch Surg 1972; 104:134.

21. Coppe D, Wessinger SJ, Ransil BJ, et al. Sex differences in theplatelet response to aspirin. Thromb Res 1981; 23:1-21.

22. Moroz LA. Increased blood fibrinolytic activity after aspirin in-gection. N Engl J Med 1977; 296(10):525-529.

23. Mansfield AO. Alteration in fibrinolysis associated with surgeryand venous thrombosis. Br J Surg 1972; 59:754.

24. Knight MTN, Dawson R, Melrose DG. Fibrinolytic response tosurgery. Labile and stable patients and their relevance to post-operative deep vein thrombosis. Lancet 1977; 2:370.

25. Bulpitt CJ. Confidence intervals. Lancet 1987; 1:494-497.26. Fleiss JL. Statistical Methods for Rates and Proportions. New

York: J. Wiley & Sons, 1973; 23-33.27. Comp PC, Jacocks RM, Taylor FB. The dilute whole blood clot

lysis assay: a screening method for identifying postoperativepatients with a high incidence of deep venous thrombosis. JLab Clin Med 1979; 93(1):120-127.

28. Clayton JK, Anderson JA, McNicol GP. Preoperative predictionof postoperative deep vein thrombosis. Br J Med 1976; 2:910-912.

29. Nicolaides AN, Fernandes JF, Pollock AV. Intermittent sequen-tial pneumatic compression of the legs in the prevention ofvenous stasis and postoperative deep venous thrombosis. Sur-gery 1980; 87(1):69-76.

30. Nicolaides AN, Miles C, Hoare M, et al. Intermittent sequentialpneumatic compression of the legs and thromboembolism-de-terrent stockings in the prevention of postoperative deepvenous thrombosis. Surgery 1983; 94:21-25.

31. Freiman JA, Chalmers TC, Smith H, Kuebler RR. The impor-tance ofbeta, the type II error and sample size in the design andinterpretation ofthe randomized control trial. New Engl J Med1978; 299(13):690-694.