The plasma turnover of transfused antithrombin concentrate inpatients with acquired antithrombin deficiencyP. L. Harper,* G. R. Park† and R. W. Carrell*Departments of Haematology* and Anaesthesia,†

Addenbrookes Hospital, Cambridge CB2 2QQ, UK.

Received 13 March 1995; accepted for publication 11 July 1995

SUMMARY. Antithrombin concentrate, prepared fromhuman plasma, has been used as replacement therapyin 35 patients with acquired antithrombin deficiency.The inhibitory activity of the concentrate, measured bychromogenic assay, correlates well with the manufac-turer’s quoted activity. The meanin vivo recovery of theproduct was 0:0124 iu mLÿ1 per iu of antithrombin (AT)concentrate administered by kilogram body weight. Therecovery was similar in all diagnostic groups studiedand did not vary during the course of treatment.Consumption of the antithrombin concentrate wasmonitored by measuring the production of thrombin–antithrombin complexes and the loss of plasma antith-

rombin activity. The mean concentration of thrombin–antithrombin complexes was elevated (23 ng mLÿ1) atthe time of admission to the intensive care unit and fellprogressively over the next 4 days. The mean time for thedecay of half the antithrombin activity was 23 h duringthe first 24 h of therapy and rose to 42.1 h after day 1.The recovery and half-life measurements are necessaryto plan an appropriate dosage regimen for the adminis-tration of this antithrombin concentrate in acquireddeficiency states.

Key words: antithrombin deficiency, antithrombin con-centrate.

Acquired antithrombin deficiency is frequently seen inpatients with disseminated intravascular coagulation(DIC) due to sepsis (Mammenet al., 1985), trauma(Gitel et al., 1979) or following major surgery (Jorgensenet al., 1980; Seyteret al., 1981). The role of antithrombinsupplements in the management of these conditionsremains controversial (Schipperet al., 1978; Buller &Ten Cate, 1983; Hellgrenet al., 1984; von Krieset al.,1985; Vinazzer, 1989) as any clinical benefit appears tobe marginal and is based primarily on a small numberof pilot studies and anecdotal reports (Hellgrenet al.,1984; Harperet al., 1991). Nevertheless, antithrombinconcentrates have been recommended in the manage-ment of sepsis and DIC and are widely used in Europe(Hellgrenet al., 1984; Seitset al., 1988; Vinazzer, 1989).

In view of the small number of trials published to date,recommendations on dosage regimen remain unclear.Guidelines on treatment are further complicated by thefact that there are a number of different antithrombinpreparations available, each prepared by a differenttechnique (Hoffman, 1989). Antithrombin recoveryand half-life data have been reported in patients with

congenital antithrombin deficiency, but less data areavailable from patients with an acquired deficiency(Schwartz et al., 1989; Menacheet al., 1990). Wereport here studies of the plasma recovery and turnoverrate of one antithrombin preparation (Bio ProductsLaboratory). These results will prove useful in planninga dosage schedule for this product for use in acquireddeficiency states.

PATIENTS AND METHODS

Patients

Over a 5-month period the plasma antithrombin activitywas measured in all patients admitted to AddenbrookesHospital, Intensive Care Unit. These assays were per-formed as part of a prospective randomized trial toexamine the effects of antithrombin supplements oncritically ill patients. The study was approved by thelocal ethics committee. Patients with an admissionantithrombin activity of less than 70% were randomizedto receive either antithrombin or to act as controls.Patients randomized to treatment were given a loadingdose of antithrombin concentrate based on the followingformula; dose (iu)� [120ÿ admission antithrombin

Transfusion Medicine, 1996,6, 45–50

45# 1996 Blackwell Science Ltd

Correspondence: Dr P. L. Harper, Department of Haematology, WestSuffolk Hospital, Bury St Edmunds IP33 2QZ, UK.

activity (%)]�weight (kg). The loading dose was givenintravenously by slow infusion over 5 min; this wasadministered as soon as possible after admission to theintensive care unit. The first maintenance dose (half theloading dose) was administered 12 h later. All furtherantithrombin doses were calculated from the half-life andrecovery.

Methods

Measurement of antithrombin recovery.Antithrombinactivity was measured both before and 10 min after eachdose. Incrementalin vivo recovery was defined as theincrease in plasma antithrombin activity (iu mLÿ1) perunit antithrombin concentrate administered per kilogramof the patient’s weight. For example: in a 50-kg man arise in antithrombin activity of 0:2 iu mLÿ1 following anantithrombin dose of 1000 iu gives a recovery of0:2=�1000=50� � 0:01 iu mLÿ1 iu administeredÿ1 kgÿ1.In clinical practice antithrombin activity is often referredto as a percentage of normal plasma activity where1 iu mLÿ1

� 100% activity. Therefore, in the exampleabove 0:01 iu mLÿ1 iuÿ1 administeredÿ1 kgÿ1

� 1% risein activity iu administeredÿ1 kgÿ1.

Decay of plasma antithrombin activity.Antithrombinconsumption was assessed by measuring the time takenfor the plasma activity to fall by 50% following eachdose of concentrate. This is calculated in the same way asplasma half-life. This is not a true half-life measurementas the baseline antithrombin activity in patients withacquired deficiency does not remain constant. We referto this approximate half-life asT a

1=2. The initial plasmahalf-life (T a

1=2) was calculated from measurements takenduring the first 24 h of treatment.

The antigenic concentration of antithrombin wasmeasured before treatment was commenced and dailyin a post-treatment sample thereafter in all patients.

Thrombin–antithrombin complexes were measureddaily in all patients treated with antithrombinconcentrate.

Antithrombin concentrate.Antithrombin concentratewas prepared from pooled normal plasma as previouslydescribed (Smithet al., 1985). The product was pas-teurized at 608C for 10 h under the protection of 0.7M

sodium citrate. The product was provided on a namedpatient basis only for the study.

Antithrombin measurements in plasma.Antithrombinconcentration was estimated by enzyme-linked immuno-sorbent assay (ELISA) (Edgaret al., 1989) andby immunoelectrophoresis. A normal plasma pool wasarbitrarily given a value of 100%. Assays of antithrombinactivity were carried out by amidolytic methods usinga synthetic chromogenic substrate (Abilgaardet al.,1977). A normal range (mean� 3SD) of 78–122%

was established. Thrombin–antithrombin complexeswere assayed by ELISA (Enzygnost-TAT, Behring).

Antithrombin activity in the concentrate.Anti-thrombin activity in the concentrate is defined in inter-national units (iu). One international unit antithrombinactivity is approximately equivalent to the antithrombinactivity in 1 mL of normal plasma, i.e. 1 iu mLÿ1 shouldrepresent an activity of 100%. Antithrombin concentratewas diluted to a concentration of 1 iu mLÿ1 in NaCl(0.9%) based on the activity stated for the freeze-driedpreparation. Antithrombin activity in the concentrate wasmeasured using a chromogenic assay.

Heparin sepharose chromatography of antithrombinconcentrate.Heparin affinity chromatography of anti-thrombin concentrate was carried out using a1:5-cm� 10-cm heparin sepharose column attached toa Waters 600E advanced protein purification system.Samples were loaded in 20 mM NaCl, 20 mM Tris HCl,pH 7.5, at a flow rate of 1 mL minÿ1. The column waswashed with 300 mM NaCl, 20 mM Tris HCl, pH 7.5, toremove low-affinity material. Elution of antithrombinwas achieved by a 2-h linear gradient from 300 mM NaCl,20 mM Tris HCl, pH 7.5, to 1.0M NaCl, 20 mM Tris HCl,pH 7.5, at a flow rate of 1 mL minÿ1. Antithrombinactivity and antigenic concentration were measured inall peaks eluted.

Purification of normal antithrombin.Normal anti-thrombin was prepared from pooled normal plasmausing heparin sepharose chromatography. The conditionsof the column were the same as those described above.

Statistical analysis.Comparison of means was madeusing Student’st-test.

RESULTS

Patients

A total of 35 patients received antithrombin replacement.The diagnoses of the patients treated were as follows; 11post-operative, 11 post liver transplantation, four trauma,five with septicaemia and four others.

The plasma turnover of antithrombin

Loading dose.All patients received the loading doseof antithrombin concentrate within 6 h of admission tothe intensive care unit. Before the loading dosethe antithrombin activity ranged from 6 to 69% (median49%). Ten minutes after the loading dose the antithrom-bin activity ranged from 75 to 183% (median 108%).The median loading dose was 3950 iu (range 2760–5520 iu), equivalent to: median 56:5 iu kgÿ1 (range39:4ÿ83 iu kgÿ1). The antithrombin concentrate was

46 P. L. Harperet al.

# 1996 Blackwell Science Ltd,Transfusion Medicine, 6, 45–50

administered at between 50 and 100 iu mLÿ1 by slowinfusion over 5 min.

First maintenance dose.The first maintenance dose(half the loading dose) was given approximately 12 hafter the loading dose. The median antithrombin activityprior to the first maintenance dose was 91% (range 43–138%). The median antithrombin activity immediatelyafter the maintenance dose was 115% (range 82–193%).

Frequency of treatment.Four patients received theloading dose alone. The remaining patients receivedbetween two and 38 doses (median six). Maintenancetherapy was continued whilst the patient remained onICU. Maintenance treatment was calculated from thehalf-life (T a

1=2) and recovery for each patient. Examplesof the dosage regimen in three patients are shown inFig. 1.

Recovery of antithrombin concentrate.The mean risein antithrombin activity was 0:0124 iu mLÿ1 (1.24%activity) for each iu of AT concentrate administeredper kg body weight. This recovery was based on 158measurements. Recovery was not significantly differentin various diagnostic groups and initial recovery was notsignificantly less than later measurements (Table 1).

Decay in plasma antithrombin activity.The mean timefor antithrombin activity to fall by 50% is significantlyless during the first 24 h of treatment than later measure-ments in all diagnostic groups (Table 2). There was nosignificant difference between diagnostic groups. Theinitial half-life (T a

1=2) ranged from 12 to 60 h (mean23.2 h).

Antigenic concentration of antithrombin.The meanantithrombin antigenic concentration prior to treatmentwas 71% (SEM� 9%) rising to a mean of 128%(SEM� 20%) on day 1 and 140% (SEM� 13%) onday 2.

Thrombin–antithrombin complexes.The mean throm-bin–antithrombin complex antigen measured in 20patients shows a progressive fall over the first 4 dayson the intensive care unit (Fig. 2).

Activity of antithrombin concentrate.The anti-thrombin concentrate was diluted to a concentration of1 iu mLÿ1 based on the stated antithrombin activity onthe vial. This diluted material had an antithrombinactivity of 1:009� 0:00146 iu mLÿ1 (mean � SEM)measured in 20 samples from two batches of anti-thrombin concentrate. There was no significantdifference between batches (t-test).

Acquired antithrombin deficiency and plasma turnover47

# 1996 Blackwell Science Ltd,Transfusion Medicine, 6, 45–50

Fig. 1. Plasma antithrombin activity and dosage regimenin three patients with acquired antithrombin deficiency.(a) Patient with trauma; (b) post-operative patient and (c) patientwith septicaemia.

Heparin sepharose chromatography.Heparin sepha-rose separation of the heat-treated concentrate resolvedas two peaks, one with low heparin affinity which elutedat 0.3M NaCl and a second peak with higher heparinaffinity which eluted at 0.7M NaCl.

The first peak contained 37% of the total proteineluted.This material had low specific thrombin inhibitoryactivity of 1.26% compared to normal antithrombin withactivity of 100%. The second peak contained 63% oftotal protein. This had a specific thrombin inhibitoryactivity of 98.4%. Both peaks had similar antithrombinantigenic concentrations. Therefore, the low-affinitypeak was assumed to be inactive antithrombin eitherdenatured or in an altered conformation. In a further fivebatches of antithrombin concentrate studied, 16–40% ofthe material had a low heparin affinity and a specificactivity of less than 2%.

DISCUSSION

During disseminated intravascular coagulation (DIC)there is rapid activation of coagulation with consumptionof antithrombin. Theoretically, antithrombin replace-ment should help to halt the coagulopathy by neutralizingany free thrombin. In practice, however, it has proveddifficult to confirm that antithrombin has any role in

clinical management (Mammenet al., 1985); this haspartly been due to difficulties designing large appropriateclinical studies. Nonetheless, there is growing evidencefrom both animal studies and small clinical trials thatantithrombin does have a role in the treatment of DIC andthe shock syndromes. Clearly a large clinical trial isnecessary. The design of such a trial will be importantand the information gained from smaller studies, such asthe one reported here, should prove useful in the planningprocess.

It is clear from animal studies that antithrombin needsto be administered as near the onset of DIC as possible(Mammenet al., 1985; Emersonet al., 1987), but theappropriate dosage remains unclear. Animal studies haveshown that a high concentration, in excess of five timesnormal, is necessary to confer a survival benefit (Tayloret al., 1988); however, such a large concentration hasnot been used in clinical studies. In animal studiesthis high concentration did not appear to have a detri-mental effect. Therefore, in planning future clinical trialsit may be appropriate to administer larger doses ofantithrombin in order to achieve supranormal concentra-tions. Clearly safety would be of paramount importance;from our results reported here concentrations almosttwice normal have been achieved without adverseeffects.

48 P. L. Harperet al.

# 1996 Blackwell Science Ltd,Transfusion Medicine, 6, 45–50

Recovery after Recovery afterDiagnosis loading dose subsequent doses Overall recovery

All patients 1:20� 0:36 (30) 1:26� 0:7 (128) 1:24� 0:66 (158)

Liver transplants 1:14� 0:3 (10) 1:09� 0:4 (41) 1:12� 0:4 (51)Trauma 1:44� 0:5 (3) 1:78� 1:0 (12) 1:71� 0:9 (15)Surgery 1:29� 0:4 (10) 1:30� 0:8 (45) 1:30� 0:7 (55)Sepsis 0:92� 0:3 (4) 1:18� 0:7 (16) 1:13� 0:6 (20)Others 1:22� 0:3 (3) 1:28� 0:7 (14) 1:26� 0:7 (17)

Rise in percentage activity per iu administered per kg body weight. Mean� standard deviation(number of measurements).

Table 1. Recovery of antithrombinconcentrate in various diagnostic groups

Initial Measurement OverallDiagnosis measurement after 24 h measurement

All patients 23:2� 2:5 (23) 42:1� 3:6 (130) 39:2� 3:2 (153)

Liver transplants 23:2� 3:2 (7) 34:4� 6:0 (43) 32:8� 5:2 (50)Trauma 28:6� 3:6 (3) 57:6� 11:0 (16) 52:9� 9:8 (19)Surgery 17:4� 3:0 (8) 36:9� 6:2 (41) 33:7� 5:3 (49)Sepsis 27:3� 1:0 (3) 41:4� 5:8 (16) 39:2� 4:9 (19)Others 32:1� 20 (2) 63:9� 14:8 (14) 59:9� 13:5 (16)

Mean� standard error of the mean in hours (number of measurements).

Table 2. Half-life of antithrombinconcentrate in various diagnostic groups

A second factor to consider is the dosage regime. Thisdepends largely on the half-life and recovery of thetransfused material. Recovery in the patients studiedwas similar to earlier reports (Menacheet al., 1990)and was unaffected by the underlying disease process(Table 1). The mean half-life of antithrombin concentratein patients with a congenital deficiency of antithrombinhas been reported to be of the order of 60 h (Collenet al.,1977). In the critically ill patient, however, antithrombinconsumption is likely to be more rapid. This was demon-strated in our patients by the short plasma survival ofantithrombin (mean half-life of 23.2 h) (Table 2) and thehigh concentration of thrombin–antithrombin complexes[mean 23 ng mLÿ1 (normal <4 ng mLÿ1)] during thefirst day of treatment (Fig. 2). Subsequently, the rateof consumption fell, with the plasma survival of anti-thrombin significantly longer and the concentration ofthe thrombin–antithrombin complexes lower by thesecond day of treatment. Failure to appreciate thisrapid consumption meant that, in our study, many treatedpatients had prolonged periods with antithrombin activ-ity below the normal range, especially during the first24 h of treatment (Fig. 1). Clearly, in order to maintainadequate plasma concentrations antithrombin needs tobe administered frequently during the acute phase of theillness when consumption is most rapid.

Finally, in planning a trial it should be recognizedthat there are a number of commercially availableconcentrates. These are prepared by different techniquesbut at some stage in preparation all undergo heat treat-ment and freeze-drying. These procedures would beexpected to damage some of the active material. Theproduct used in our study contained a variable amount ofinactive antithrombin. The proportion of this inactivematerial ranged from 16 to 40% of the total proteinadministered. This probably represents denatured or

structurally altered antithrombin. As yet, it is not clearif this inactive material is clinically significant. Thishigh concentration of altered material in part explainswhy the antigenic concentrations of antithrombinmeasured on day 1 and day 2 are significantly higherthan the activity measurements. Clearly, monitoringreplacement therapy with antigenic assays would bemisleading and unreliable.

The role of antithrombin replacement in acquireddeficiency states can only be clarified by further clinicaltrials. The lesson learnt from this study is that theconsumption of antithrombin is considerably faster inan intensive care patient than in a healthy individual.Therefore, in order to plan an appropriate study usingantithrombin replacement it should be recognized thathigh concentrations of antithrombin are necessary tomaintain activity within the normal range, particularlyduring the first few days of treatment. In addition, inview of the different production methods used forthe manufacture of antithrombin concentrates and thevariable amount of inactive material in these prepara-tions, it is essential when planning a study using aspecific antithrombin concentrate to have data onthe plasma recovery and turnover of that particularproduct.

ACKNOWLEDGEMENTS

We are grateful to Dr J. K. Smith and BPL forthe provision of Antithrombin Concentrate. The clinicalstudies were supported by the British Heart Foundation.

REFERENCES

Abilgaard, U., Lie, M. & Odegard, O.R. (1977) Antithrombin(heparin cofactor) assay with new chromogenic substrate(S2238 and chromozym TH).Thrombosis Research, 11,549–552.

Buller, H.R. & Ten Cate, J.W. (1983) Antithrombin III infusionin patients undergoing peritoneovenous shunt operations;failure in the prevention of disseminated intravascularcoagulation.Thrombosis and Haemostasis, 49, 128–131.

Collen, D., Schetz, J., De Cock, F., Holmer, E. & Verstraete, M.(1977) Metabolism of antithrombin III (heparin of cofactor)in man: effects of venous thrombosis and of heparinadministration.European Journal of Clinical Investigation,7, 27–35.

Edgar, P., Jennings, I. & Harper, P.L. (1989) Enzyme linkedimmunosorbent assay for measuring antithrombin III.Jour-nal of Clinical Pathology, 7, 985–987.

Emerson, T.E., Fournel, T.E., Leach, W.J. & Redens, T.B.(1987) Protection against intravascular coagulation and deathby antithrombin III in theE coli endotoxaemic rat.Circula-tory Shock, 21, 1–13.

Acquired antithrombin deficiency and plasma turnover49

# 1996 Blackwell Science Ltd,Transfusion Medicine, 6, 45–50

Fig. 2.Mean thrombin–antithrombin complex concentration in20 patients during the first 4 days of antithrombin replacementtherapy.

Gitel, S.N., Salvati, E.A., Wessler, S., Robinson, H.A. &Worth, M.H. (1979) The effect of total hip replacementand general surgery on antithrombin III in relation tovenous thrombosis.Journal of Bone & Joint Surgery, 61-A, 653–656.

Harper, P.L., Williamson, L., Park, G., Smith, J.K. & Carrell,R.W. (1991) A pilot study of antithrombin replacementin intensive care management: the effects on mortality,coagulation and renal function.Transfusion Medicine, 1,121–128.

Hellgren, M., Javelin, L. & Hagnevik, K. (1984) AntithrombinIII concentrate as an adjuvant in disseminated intravascularcoagulation patients: a pilot study in 9 severely ill patients.Thrombosis Research, 35, 459–466.

Hoffman, D.L. (1989) Purification and large scale preparationof antithrombin III.American Journal of Medicine, 87,235–265.

Jorgensen, K.A., Seffensen, E., Sorensen, P.J., Ingeberg, S.,Huttel, M. & Ahlbom, G. (1980) Alterations in plasmaantithrombin III following total hip replacement and electivecholecystectomy.Scandinavian Journal of Haematology, 24,101–104.

von Kries, R., Stannigel, H. & Gobel, U. (1985) Anticoagulanttherapy by continuous heparin– antithrombin III infusion innewborns with disseminated intravascular coagulation(DIC). European Journal of Pediatrics, 144,191–194.

Mammen, E.F., Miyakawa, T., Phillips, T.F., Assarim, G.S.,Brown, J.M. & Murano, G. (1985) Human antithrombinconcentrates and experimental disseminated intravascularcoagulation.Seminars in Thrombosis and Hemostasis, 11,373–383.

Menache, D., O’Malley, J.P., Schorr, J.B., Wagner, B. &Williams, C. (1990) Evaluation of safety, recovery, half-life, and clinical efficacy of antithrombin III (Human) inpatients with hereditary antithrombin III deficiency.Blood,75, 33–39.

Schipper, H.G., Jenkins, C.S.P., Kahle, L.H. & ten Cate, J.W.(1978) Antithrombin III transfusions in disseminated intra-vascular coagulation.Lancet, i, 854–856.

Schwartz, R.S., Bauer, K.A., Rosenberg, R.D., Ellin,J.K., Davies, D.A. & Bogdanoff, D.A. (1989) Clinicalexperience with Antithrombin III concentrate in treatmentof congenital and acquired deficiency of antithrombin.Amer-ican Journal of Medicine, 87, Suppl. 3B, 53S–60S.

Seits, R., Wolf, M., Egbring, R. & Havemann, K. (1988)Substitution of antithrombin III and plasma in septicshock. Effects on proteolysis and survival.Haemostasis,18, 7–8.

Seyter, A.E., Seaber, A.V., Dombrose, F.A. & Urbaniak, J.R.(1981) Coagulation changes in elective surgery and trauma.American Journal of Surgery, 193,210–213.

Smith, J.K., Winkelman, L., Evans, D.R., Haddon, M.E.& Sims, G.A. (1985) A pasteurised antithrombin III con-centrate for clinical use.Vox Sanguinis, 48, 325–332.

Taylor, F.B., Emerson, T.E., Chang, J.R. & Blick, K.E. (1988)Antithrombin III prevents lethal effects of E coli infusion inbaboons.Circulatory Shock, 26, 227–235.

Vinazzer, H. (1989) Therapeutic use of antithrombin IIIin shock and disseminated intravascular coagulation.Semi-nars in Thrombosis and Haemostasis, 3, 128–131.

50 P. L. Harperet al.

# 1996 Blackwell Science Ltd,Transfusion Medicine, 6, 45–50