Recurrent miscarriage and long-term thrombosis risk: a case–control study

S. Quenby1,3, R.G. Farquharson2, F. Dawood2, A.M. Hughes2 and J. Topping2

1 School of Developmental and Reproductive Medicine, University of Liverpool, Liverpool, L8 7SS and 2 Liverpool Women's Hospital Trust, Liverpool, L8 7SS, UK

3 To whom correspondence should be addressed. Email: squenby{at}liv.ac.uk


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
BACKGROUND: Recurrent miscarriage has been associated with antiphospholipid syndrome (APS) and other prothombotic conditions. We tested the hypothesis that women diagnosed as having APS as an aetiological factor for their miscarriages were at higher risk of thrombosis than those with idiopathic recurrent miscarriage. METHODS: A retrospective case–control study was performed using validated questionnaires. A total of 141 women with recurrent miscarriage and APS alone were matched with 141 women with idiopathic recurrent miscarriage for age, number and type of pregnancy loss and number of years of follow-up. A subgroup of eight women included those who initially presented with recurrent miscarriage, thrombosis and APS. RESULTS: The mean length of follow-up was 7.3 years and response rate 74%. The incidence of thrombosis was similar in the recurrent miscarriage and APS women (6/1000 women-years) and in the idiopathic recurrent miscarriage women (2/1000 women-years) (P=0.18). All eight women who presented with recurrent miscarriage, APS and thrombosis reported further thrombotic events. CONCLUSIONS: Both idiopathic and APS-associated recurrent miscarriage were associated with a similar long-term risk of thrombosis.

Key words: antiphospholipid syndrome/recurrent miscarriage/thrombosis


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
The antiphospholipid syndrome (APS) consists of a clinical history of recurrent miscarriage, thrombosis or thrombocytopenia, accompanied by persistently positive tests for antiphospholipid antibodies (Hughes and Khamashta, 1994Go). Antiphospholipid antibodies (aPL) have been characterized as a heterogeneous group of antibodies. It is possible that subtypes of aPL associated with thrombosis are different from the aPL subtypes associated with recurrent miscarriage. Antiphospholipid antibodies detected in standard assays include some that are directed against anionic phospholipid, some that are directed against phospholipid binding proteins including various parts of the {beta}2-glycoprotien 1 protein (Roubey, 1994Go) and some against prothrombin (Pierangeli et al., 2001Go), annexin V (Pierangeli et al., 2001Go) and factor XII (Jones et al., 2000Go). Furthermore, there was a marked variation in antiphospholipid antibodies when measured longitudinally in pregnancy (Topping et al., 1999Go). Some women who had positive tests when not pregnant had negative test in late pregnancy (Topping et al., 1999Go). We do not know how many women remain positive for aPL in the years after their reproductive life.

The risk of thrombosis following a positive test for APS in young women is unknown (Cuadrado and Khamashta, 2000Go; Navarrete et al., 2000Go) but has recently been considered to be ‘moderate’ (Lockshin and Erkan, 2003Go). This means that clinicians have no firm evidence on which to base advice regarding need for long-term thromboprophylaxis. The aim of this project was to assess the long-term risk of subsequent thrombosis in young women attending a tertiary recurrent miscarriage clinic. We tested the hypothesis that women diagnosed as having APS as an aetiological factor for their miscarriages were at higher risk of thrombosis than those with idiopathic RM.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Following research ethics committee approval, a questionnaire was designed to elicit a medical history. In order to ensure that the details of any thrombotic episodes were as comprehensive as possible, the questionnaire included both open and closed questions. Furthermore, patients were asked to indicate whether they would consent to a telephone call from the miscarriage clinic nurse specialist, to clarify any inconsistencies. Questionnaire validation occurred in two stages. Initially, the questionnaire was given to 10 non-clinical members of the Liverpool Women's Hospital staff. A medical history was taken from these staff; the verbal- and questionnaire-elicited histories correlated exactly. However, when the members of staff were asked to comment regarding the acceptability of the questionnaire, three said that it might be frightening for patients to receive. Therefore, the questions were modified using less threatening language (Table I). The questionnaire was then given to five women who had APS and follow-up appointments. In each case it successfully elicited an accurate medical history and was felt to be non-threatening by patients.


View this table:
[in this window]
[in a new window]
 
Table I. Questionnaire sent to women

 
In 2003, women were then selected for the study who had attended the Liverpool Women's Hospital recurrent miscarriage clinic more than 5 years previously (1991–1998), from the clinic database. A questionnaire, information leaflet, consent form and a stamped addressed envelope were posted to 201 women whose APS was diagnosed when attending the miscarriage clinic. Inclusion criteria for the APS group were in accordance with the Sapporo agreement (Wilson and Gharavi, 1999Go), i.e. two positive tests for IgG or IgM anticardiolipin antibodies (>3 SD above the normal range) or lupus anticoagulant (Dilute Russell's Viper Venom time >1.1 with a >20% correction for the platelet neutralization step) and one or more unexplained fetal deaths beyond 10 weeks gestation or three or more unexplained consecutive miscarriages before 10 weeks gestation. Exclusion criteria included any other form of autoimmune disorder, i.e. secondary APS. Women who had already suffered a thrombosis at their first clinic visit were treated as a separate group.

When questionnaires were returned, controls were selected from the idiopathic recurrent miscarriage section of the database. Idiopathic recurrent miscarriage was diagnosed for those for whom the following investigations failed to detect an aetiology for their pregnancy losses: IgG or IgM anticardiolipin, lupus anticoagulant, parental karotype, auto-antibody screen, random blood glucose, thyroid function tests, pelvic ultrasound scan, hysteroscopy, cervical swabs for myoplasma and ureaplasma, high vaginal swab for bacterial vaginosis, and serology for toxoplasmosis, parovirus, rubella and cytomegalovirus (Drakeley et al., 1998Go). The cases were matched to controls for age, number of years of follow-up and number and type of pregnancy losses (Table II). If the selected controls failed to answer the questionnaire, a further matched control was selected and the questionnaire was sent to them. When 141 case-matched women with idiopathic recurrent miscarriage had answered the questionnaire, recruitment was discontinued.


View this table:
[in this window]
[in a new window]
 
Table II. Demographics at first clinic visit

 
Women whose questionnaires were difficult to interpret were telephoned and asked to clarify their medical history. Any reported thrombotic event was verified by checking both hospital medical and local general practitioner records. The results of the questionnaire were entered into a database, then analysed. Thromboses occurring during or within 6 weeks of a pregnancy were excluded from the analysis. Sixteen per cent of the questionnaires were returned as the patients had moved house.

The risk of thrombosis between the APS and control group was compared using a statistical test for two independent proportions using StatsDirect (Sale, Cheshire) software for personal computer. P<0.05 was accepted as statistically significant.


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
In all, 149 out of 201 women with APS sent back the questionnaire, thus there was a 74% response rate. These were divided into 141 with APS (a history at the first clinic visit of recurrent miscarriages alone) and eight with APS, RM and previous thrombosis. The APS patients were successfully matched to the controls for the pre-specified criteria with no significant difference detected for age, number and type of loss or live birth history at diagnosis (Table II).

Of the 141 women with APS who had no past history of thrombosis prior to referral, six reported a subsequent thrombosis (Table III). No women in the control group had a thrombosis prior to referral, but two women in this group subsequently reported a thrombotic event (Table III). The two women reporting a thrombotic episode were subsequently contacted and agreed to a thrombophilia screen. One was found to be heterozygous for the Leiden factor V mutation and one homozygous for the prothrombin gene mutation 20210A. There was no statistically significant difference in the rate of thrombosis in the APS group [six thromboses per 1000 women-years of follow-up (95% CI 2–13/1000)] and control group [two thromboses per 1000 women-years (95% CI 0.2–7/1000)]. There was one non-thrombotic death in each group, reported by the women's partner (Table III). All eight women in the APS subgroup who had had a thrombotic episode prior to referral (Table II) had subsequent thrombotic episodes (Table III).


View this table:
[in this window]
[in a new window]
 
Table III. Results of questionnaire: self-reported conditions

 
There were more smokers in the idiopathic RM group and very few women in any group took the oral contraceptive pill (Table IV). Twenty-two women in the APS group and eight in the control group took low-dose aspirin for long-term prophylaxis (Table IV). Within the case and control groups, none of the women suffering a primary thrombosis had taken any thromboprophylaxis, the oral contraceptive pill or smoked. Of the eight women who had a secondary thrombosis in the APS group, seven were taking low-dose aspirin and two warfarin.


View this table:
[in this window]
[in a new window]
 
Table IV. Results of questionnaire: current drugs

 
Three women from each group (2%) reported being distressed because of their receipt of the questionnaire. None of these women who experienced distress at receiving the questionnaire had ever had a live birth.


    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
The incidence of thrombosis found in our study for women with recurrent miscarriage and APS, 6/1000 women years (CI 2–13/1000), was not significantly different from that for women with ‘idiopathic’ recurrent miscarriage, 2/1000 women years (CI 0.2–7/1000).

Our study has collected important and challenging data suggesting that women presenting to a recurrent miscarriage clinic do have a long-term thrombosis risk, particularly if APS has been diagnosed. The thrombosis risk in the idiopathic recurrent miscarriage group was found to reflect underlying thrombophilia in women with ‘idiopathic’ recurrent pregnancy. These thrombophilias were not detected initially as the patient's first visit to the miscarriage clinic pre-dated the discovery of the association between pregnancy loss and thrombophilia (Preston et al., 1996Go). There are many risk factors for thrombosis which include smoking, oral contraceptive pill use, obesity and hypertension. However, these potential confounding factors were not present in the women reporting thrombosis subsequent to a history of recurrent miscarriage alone.

The incidence of venous thromboembolism in our study was 2/1000 women-years for all women with recurrent miscarriage. It is difficult to compare this rate of thrombosis with that in the published literature because of differences in the methodologies used to detect thrombosis in our study and those previously published. However, a retrospective case–control study based on general practitioner records in the UK found that baseline rate of venous thromboembolism for women of reproductive years was 0.27/1000 women-years (Seaman et al., 2003Go), significantly lower than that detected in our study.

Our study had a good response rate of 74%; however, it is possible that the non-responders had more cerebral thrombosis than the responders so that they were unable to answer the questionnaire. If this were the case then we may have underestimated the thrombosis risk. However, we did know that the most common reason for failing to respond to the questionnaire was change of address and that our postal questionnaires even detected patient death.

Cerebral ischaemia associated with aPL was previously reported to be the most common arterial manifestation of APS (Krnic-Barrie et al., 1997Go; Shah et al., 1998Go; Navarrete et al., 2000Go). Most of the thrombotic episodes reported by the patients in the current study were indeed cerebral (Table III). However, the importance of aPL as a risk factor for stroke is controversial. Some studies support an association between aPL and cerebral thrombosis (Levine et al., 1990Go; Kushner, 1990Go; Camerlingo et al., 1995Go; D'Olhaberriague et al., 1998Go; Brey et al., 2001Go) and others do not (Muir et al., 1994Go; Antiphospholipid Antibodies and Stroke Study Group, 1997Go; Tanne et al., 1999Go). The data presented here suggest that there is an association between APS and cerebral thrombosis.

The efficacy of long-term thromboprophylaxis and its associated risk of bleeding is a complex problem. A recent randomized controlled trial found that high-intensity warfarin was not superior to moderate-intensity warfarin for the prevention of thrombosis in patients with aPL and previous thrombosis (Crowther et al., 2003Go). Furthermore, the authors reported an annual risk of major bleeding of 3.6% with high-intensity warfarin versus 2.2% with moderate-intensity warfarin. A retrospective study has suggested that low-dose aspirin may be effective in the prevention of thrombosis for asymptomatic APS patients (Erkan et al., 2002Go). However, low-dose aspirin has been associated with significant side-effects: a risk of 9% for gastrointestinal bleeding has been reported (Newby et al., 2003Go). Therefore, randomized controlled trials are needed to ascertain the efficacy and side-effects of thromboprophylaxis in patients with a history of recurrent miscarriage with and without APS (Lockshin and Erkan, 2003Go). Avoidance of other risk factors for thrombosis—oral contraceptive pill, smoking, obesity and hypertension—seems to be sensible advice for women with recurrent miscarriage (Cuadrado and Lopez-Pedrera, 2003Go).

Presentation with previous thrombosis, recurrent miscarriage and APS was a significant risk factor for recurrent thrombosis. The subgroup results are similar to that of Silver et al. (1994)Go who also found a significant incidence of cerebral thrombosis in a follow-up study of a similar group of women with APS. Patients with such a high risk of recurrent thrombosis should be referred routinely to a specialist haematology clinic.

In conclusion, our data suggest that a history of recurrent miscarriage is a risk factor for subsequent thrombosis in the long term. In light of these findings, a randomized controlled trial is needed to assess efficacy and bleeding risk for long-term thromboprophylaxis in women with recurrent miscarriage.


    Acknowledgements
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
We would like to thank Lupus UK for funding this study and the patients for returning the questionnaires.


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Antiphospholipid Antibodies and Stroke Study Group (APASS) (1997) Anticardiolipin antibodies and the risk of recurrent thrombo–occlusive events and death. Neurology 48, 91–94.[Abstract]

Brey RL, Abbott RD, Curb JD, Sharp DS, Ross GW, Stallworth CL and Kittner SJ (2001) beta(2)-Glycoprotein 1-dependent anticardiolipin antibodies and risk of ischemic stroke and myocardial infarction: the Honolulu heart program. Stroke 32, 1701–1706.[Abstract/Free Full Text]

Camerlingo M, Casto L, Censori B, Drago G, Frigeni A, Ferraro B, Servalli MC, Radice E and Mamoli A (1995) Anticardiolipin antibodies in acute non-hemorrhagic stroke seen within six hours after onset. Acta Neurol Scand 92, 69–71.[ISI][Medline]

Crowther MA, Ginsberg JS, Julian J et al. (2003) A comparison of two intensities of warfarin for the prevention of recurrent thrombosis in patients with the antiphospholipid antibody syndrome. New Engl J Med 349, 1133–1138.[Abstract/Free Full Text]

Cuadrado MJ and Khamashta MA (2000) The anti-phospholipid antibody syndrome (Hughes syndrome): therapeutic aspects. Baillières Best Pract Res Clin Rheumatol 14, 151–163.[CrossRef][ISI][Medline]

Cuadrado MJ and Lopez-Pedrera C (2003) Antiphospholipid syndrome. Clin Exp Med 3, 129–139.[CrossRef][ISI][Medline]

D'Olhaberriague L, Levine SR, Salowich-Palm L, Tanne D, Sawaya KL, Aurora TK, Perry M, Day M, Spencer T and Schultz L (1998) Specificity, isotype, and titer distribution of anticardiolipin antibodies in CNS diseases. Neurology 51, 1376–1380.[Abstract]

Drakeley A, Quenby S and Farquharson R (1998) Mid-trimester pregnancy loss-appraisal of a screening protocol. Hum Reprod 13, 1975–1980.[Abstract]

Erkan D, Merrill JT, Yazici Y, Sammaritano L, Buyon JP and Lockshin MD (2002) High thrombosis rate after fetal loss in antiphospholipid syndrome: effective prophylaxis with aspirin. Arthr Rheum 46, 1126–1127.[CrossRef][ISI][Medline]

Hughes GR and Khamashta MA (1994) The antiphospholipid syndrome. J R Coll Phys Lond 28, 301–304.[ISI][Medline]

Jones DW, Gallimore MJ, MacKie IJ, Harris SL and Winter M (2000) Reduced factor XII levels in patients with the antiphospholipid syndrome are associated with antibodies to factor XII. Br J Haematol 110, 721–726.[CrossRef][ISI][Medline]

Kushner MJ (1990) Prospective study of anticardiolipin antibodies in stroke. Stroke 21, 295–298.[Abstract]

Krnic-Barrie S, O'Connor CR, Looney SW, Pierangeli SS and Harris EN (1997) A retrospective review of 61 patients with antiphospholipid syndrome. Analysis of factors influencing recurrent thrombosis. Arch Intern Med 157, 2101–2108.[Abstract]

Levine SR, Deegan MJ, Futrell N and Welch KM (1990) Cerebrovascular and neurologic disease associated with antiphospholipid/antibodies: 48 cases. Neurology 40, 1181–1189.[Abstract]

Lockshin MD and Erkan D (2003) Treatment of the antiphospholipid syndrome. New Engl J Med 349, 1177–1179.[Free Full Text]

Muir KW, Squire IB, Alwan W and Lees KR (1994) Anticardiolipin antibodies in an unselected stroke population. Lancet 344, 452–456.[CrossRef][ISI][Medline]

Navarrete MG, Brey RL and Levine SR (2000) In Khamashta MA (ed.) Hughes Syndrome. Springer, Berlin, pp. 43–58.

Newby LK, Bhapkar MV, White HD, Moliterno DJ, LaPointe NM, Kandzari DE Verheugt FW, Kramer JM, Armstrong PW and Califf RM, (2003) for the SYMPHONY and 2nd SYMPHONY Investigators Aspirin use post-acute coronary syndromes. intolerance, bleeding and discontinuation. J Thromb Thrombolys 16 119–128.[CrossRef][ISI][Medline]

Pierangeli SS, Gharavi AE and Harris EN (2001) Testing for antiphospholipid antibodies: problems and solutions. Clin Obstet Gynecol 44, 48–57.[CrossRef][ISI][Medline]

Preston FE, Rosendaal FR, Walker ID et al. (1996) Increased fetal loss in women with heritable thrombophilia. Lancet 348, 913–916.[CrossRef][ISI][Medline]

Roubey RA (1994) Autoantibodies to phospholipid-binding plasma proteins: a new view of lupus anticoagulants and other "antiphospholipid" autoantibodies. Blood 89, 2854–2867.

Seaman HE, de Vries CS and Farmer RDT (2003) The risk of venous thromboembolism in women prescribed cytproterone acetate in combination with ethinyl estradiol; a nested cohort analysis and case–control study. Hum Reprod 18, 522–526.[Abstract/Free Full Text]

Shah NM, Khamasthta MA, Atsumi T and Hughes GRV (1998) Outcome of patients with anticardiolipin antibodies: a 10 year follow-up of 52 patients. Lupus 7, 3–6.[CrossRef][ISI][Medline]

Silver RM, Draper ML, Scott JR, Lyon JL, Reading J and Branch DW (1994) Clinical consequences of antiphospholipid antibodies: an historic cohort study. Obstet Gynecol 83, 372–377.[Abstract]

Tanne D, D'Olhaberriague L, Schultz LR, Salowich-Palm L, Sawaya KL and Levine SR (1999) Anticardiolipin antibodies and their associations with cerebrovascular risk factors. Neurology 52, 1368–1373.[Abstract/Free Full Text]

Topping J, Quenby S, Farquharson R, Malia B and Greaves M (1999) Marked variation in results of antiphospholipid during pregnancy: relationship to pregnancy outcome. Hum Reprod 14, 224–228.[Abstract/Free Full Text]

Wilson AW and Gharavi AE (1999) International consensus statement on preliminary classification criteria for definite antiphospholipid syndrome. Arthr Rheum 7, 1309–1311.[CrossRef]

Submitted on November 4, 2004; resubmitted on January 31, 2005; accepted on February 11, 2005.