SHORT REPORT

Risk of venous thromboembolismassociated with a G to A transition at position 20210in the 30-untranslated region of the prothrombin gene

KAREN BROWN, ROGER LUDDINGTON, DAVID WIL LIAMSO N, PE TER BAKER AND TREVOR BAGLIN

Department of Haematology, Addenbrooke’s NHS Trust, Cambridge

Received 26 March 1997; accepted for publication 20 June 1997

Summary. The odds ratio for the FII 20210G/A mutation in504 patients with venous thromboembolism compared tocontrols was 2.0 (95% CI 1.0–4.0) and, for factor V Leiden,5.8 (95% CI 3.3–10.3). 3/504 patients were heterozygousfor both mutations. None of the patients had combinednatural anticoagulant deficiency and the FII 20210G/A

mutation. We conclude that the FII 20210G/A mutation ispresent in 2.6% of the population and the relative risk ofvenous thromboembolism in carriers is 2.0.

Keywords: thrombosis, embolism, thrombophilia, prothrom-bin gene.

The prothrombin gene has recently been examined as acandidate gene for venous thrombosis risk. A novel sequencevariation in the 30-untranslated region of the gene, a G to Atransition at nucleotide position 20210, was found in 18% ofselected patients with a personal and family history of deepvein thrombosis and 6.2% of unselected consecutive patientswith a first episode (Poort et al, 1996). This mutation waspresent in 2.3% of healthy control subjects. The odds ratiofor the mutation in unselected patients with deep veinthrombosis compared to control subjects was 2.8 (95% CI1.4–5.6). After exclusion of patients with other genetic riskfactors or lupus anticoagulant activity, the odds ratio for themutation was 2.7 (95% CI 1.3–5.6). We have calculated oddsratios for the mutation in a large number of patients withvenous thromboembolism investigated in the U.K. We havealso analysed the relative risk of double heterozygosity for thismutation and the factor V Leiden (factor V R506Q) mutation.

METHODS

Patients and controls. DNA was extracted from 504consecutive citrated blood samples from patients with ahistory of venous thromboembolism. These samples weresent from hospitals in the South-East of England to thehaemostasis unit in Addenbrooke’s Hospital, Cambridge, forinvestigation of thrombophilia. 343 were from patients withdeep vein thrombosis and 161 with pulmonary embolus.

Deep vein thrombosis was diagnosed by compressionultrasonography or contrast venography and pulmonaryembolus by ventilation-perfusion lung scanning. 177 weremale and 326 female. The age range of patients withdeep vein thrombosis was 9–89 (median 47) years and forpatients with pulmonary embolus 16–82 (median 43) years.DNA was also extracted from citrated blood samples takenfrom 508 anonymous blood donors from the same geo-graphical area. Samples were centrifuged at 2500 g for10 min on arrival in the laboratory. The buffy coat layer wasthen removed and stored at ¹808C until DNA extraction.

Factor II genotyping was performed by gene amplificationand HindIII digestion as described (Poort et al, 1996). Theoligonucleotide primer contains a nucleotide substitutionclose to the 30 end such that in the presence of the geneticabnormality a new restriction enzyme cleavage site is created.A 345 bp fragment from the 30-UT region of the prothrombingene was amplified by PCR using the primers 50-TCTA-GAAACAGTTGCCTGGC-30 (nucleotides 19889–19908) and5-ATAGCACTGGGAGCATTGAAGC-30 (nucleotides 20233–20212). The normal allele lacks a HindIII restriction siteand generates only a 345 bp fragment. A new HindIII site(¹A/AGCTT), introduced into the amplified fragment fromthe mutant allele yielded two fragments of 322 bp and 23 bpafter enzyme digestion.

Factor V genotyping was performed as previously described(Beauchamp et al, 1994). PCR amplification was performedwith a primer pair spanning a 147 bp fragment encodingthe APC cleavage site (primer FV3, position nt 1623 to

British Journal of Haematology, 1997, 98, 907–909

907q 1997 Blackwell Science Ltd

Correspondence: Dr T. Baglin, Department of Haematology,Addenbrooke’s NHS Trust, Cambridge CB2 2QQ.

nt 1642, sequence 50-CATGAGAGACATCG CCTCTG-30 andprimer FV6, position IVS-10 nt 45 to nt 68, sequence 50-GACCTAACATGTTC TAGCCAGAAG-30). Incubation of thePCR product with Mnl I (Biolabs) produced fragments of25, 37 and 85 bp when A was present at position 1691,whereas the substitution nt1691G/A mutation resulted intwo fragments of 25 and 122 bp.

Natural anticoagulants and lupus anticoagulant activity.Levels of natural anticoagulants and lupus anticoagulantactivity were measured as previously described with normalranges as previously determined (Lee et al, 1994).

Statistics. Proportions, odds ratios and 95% confidenceintervals were calculated by the logit method using Con-fidence Interval Analysis software (Gardner & Altman,1989) (British Medical Journal (CIA), BMA House, TavistockSquare, London WC1H 9JR).

RESULTS (Table I)

Factor II 20210G/A13/508 healthy donors were heterozygous for the FII20210G/A mutation (2.6%, 95% CI 1.4–4.5) and none washomozygous. 25/504 patients were heterozygous (5.0%, 95%CI 3.2–7.2) and none was homozygous. Compared to controlsthe odds ratio for FII 20210G/A heterozygosity in all patientswith venous thromboembolism was 2.0 (95% CI 1.0–4.0).

Factor V Leiden15/508 healthy donors were heterozygous for the factorV Leiden mutation (3.0%, 95% CI 1.7–4.8) and none washomozygous. Assuming the alleles are in Hardy-Weinbergequilibrium, the expected frequency of homozygosity is0.00022 (1 in 4545). 76/504 patients were heterozygous(15.1%, 95% CI 12.0–18.2) and four were homozygous.Compared to controls the odds ratio for factor V Leidenheterozygosity in patients with venous thromboembolismwas 5.8 (95% CI 3.3–10.3).

Combined defectsNone of the control subjects were heterozygous for bothFII 20210G/A and factor V Leiden. 3/504 patients wereheterozygous for both mutations.

Antithrombin levels were normal in all patients with theFII 20210G/A mutation and protein C and S levels werenormal in 16/25 patients with FII 20210G/A not takingwarfarin. One patient had a strong lupus anticoagulant inassociation with a high IgG anticardiolipin titre.

DISCUSSION

The prevalence of the factor V Leiden mutation in our studyis similar to that seen in the original report of this mutation(Rosendaal et al, 1995). The prevalence of the FII 20210G/Amutation in the healthy control subjects in the LeidenThrombophilia Study (LETS) was 2.3% (Poort et al, 1996)and in our study it was 2.6%. The prevalence of thismutation is therefore the same in the U.K. as in theNetherlands. The prevalence of the mutation in our patientgroup was similar to that in the Leiden Study (5% v 6.2%).Our similar results indicate that the FII 20210G/A mutationis common but is a relatively low risk factor for venousthromboembolism. Three patients were double heterozygotesfor the FII 20210G/A mutation and the factor V Leidenmutation. No patient was a double heterozygote for FII20210G/A and deficiency of a natural anticoagulant,although it was not possible to test 9/25 patients for proteinC and S deficiency as some samples were from patientsreceiving oral anticoagulant therapy. None of the 25 patientswith FII 20210G/A were antithrombin deficient. Althoughwe did not find natural anticoagulant deficiency in anypatient with FII 20210G/A, a larger number of patients withnatural anticoagulant deficiency should now be screened forthis mutation. This would determine categorically whetherthe mutation increases the risk of thrombosis in patientswith other forms of thrombophilia.

Case–control studies evaluating recurrence rates forthromboembolism in affected individuals and primarythrombosis rates after exclusion of the proband from affectedfamilies are now required to determine the predictive valueof this mutation. Until this information is available thelow risk of thrombosis associated with this mutation doesnot necessarily justify its detection as part of the routineinvestigation of patients with an episode of venous thrombo-embolism.

908 Short Report

q 1997 Blackwell Science Ltd, British Journal of Haematology 98: 907–909

Table I. Frequency of FII 20210 and FV Leiden.

Frequency in venous Odds ratio in venousGenotype Frequency in control subjects thromboembolism thromboembolism

FII 20210G/A heterozygous 13/508 [2.6% (1.4–4.5)] 25/504 [5.0% (3.2–7.2)] 2.0 (1.0–4.0)

FV Leiden heterozygous 15/508 [3.0% (1.7–4.8)] 76/504 [15.1% (12.0–18.2)] 5.8 (3.3–10.3)FV Leiden homozygous 0/508 (0.00022*) [0.022%] 4/504 [0.8% (0.2–2.0)] —

FV Leiden heterozygous and FII 20210G/A 0/508 (0.00076*) [0.076%] 3/504 [0.6% (0.1–1.7)] —FV Leiden homozygous and FII 20210G/A 0/508 0/504 —

Frequencies and odds ratios for FII 20210G/A and FV Leiden in control subjects and the patients with venous thromboembolism. 95%confidence intervals are in parentheses.

* Frequencies calculated assuming Hardy-Weinberg equilibrium and independent allele inheritance.

REFERENCES

Beauchamp, N., Daly, M., Hampton, K., Cooper, P., Preston, F. &Peake, I. (1994) High prevalence of a mutation in the factor Vgene within the U.K. population: relationship to activatedprotein C resistance and familial thrombosis. British Journal ofHaematology, 88, 219–222.

Gardner, M. & Altman, D. (1989) Statistics with Confidence: ConfidenceIntervals and Statistical Guidelines. British Medical Journal, BMAHouse, London.

Lee, L., Jennings, I., Luddington, R. & Baglin, T. (1994) Markers of

thrombin and plasmin generation in patients with inheritedthrombophilia. Journal of Clinical Pathology, 47, 631–634.

Poort, S., Rosendaal, F., Reitsma, P. & Bertina, R. (1996) Acommon genetic variation in the 30-untranslated region of theprothrombin gene is associated with elevated prothrombinlevels and an increase in venous thrombosis. Blood, 88, 3698–3703.

Rosendaal, F., Koster, T., Vandenbrouck, J. & Reitsma, P. (1995)High risk of thrombosis in patients homozygous for factor V Leiden(activated protein C resistance). Blood, 85, 1504–1508.

909Short Report

q 1997 Blackwell Science Ltd, British Journal of Haematology 98: 907–909