Errata: Corrigendum Volume 13, Issue 2, 327, Article first published online: 12 January 2015
Johanna I.P. de Vries, Department of Obstetrics and Gynaecology, VU University Medical Center, PO box 7057, 1007 MB Amsterdam, the Netherlands. Tel.: +31 20 4443258; fax: +31 20 444 1485. E-mail: email@example.com
Summary. Background: Early-onset hypertensive disorders (HD) of pregnancy and small-for-gestational age infants (SGA) are associated with placental vascular thrombosis, these often recur and are also associated with inheritable thrombophilia. Aspirin reduces the recurrence risk. Objectives: Adding low-molecular-weight heparin (LMWH) to aspirin at < 12 weeks gestation reduces the recurrence of HD in women with previous early-onset HD (pre-eclampsia, hemolysis, elevated liver enzymes and low platelets [HELLP] syndrome and eclampsia) and/or SGA, in the context of inheritable thrombophilia without antiphospholipid antibodies. Patients/methods: In a multicenter randomized control trial (RCT), 139 women included were < 12 weeks gestation. Inclusion criteria: previous delivery < 34 weeks gestation with HD and/or SGA; inheritable thrombophilia (protein C deficiency, protein S deficiency, activated protein C resistance, factor V Leiden heterozygosity and prothrombin gene G20210A mutation heterozygosity); and no antiphospholipid antibodies detected. Intervention: either daily LMWH (dalteparin, 5000 IU weight-adjusted dosage) with aspirin 80 mg or aspirin 80 mg alone. Main outcome measures: Primary outcomes: recurrent HD onset (i) < 34 weeks gestation and (ii) irrespective of gestational age. Secondary outcomes: recurrent SGA, preterm birth, maternal/neonatal hospitalization, spontaneous abortion and individual HD. Analysis by intention-to-treat. Results: Low-molecular-weight heparin with aspirin reduced recurrent HD onset < 34 weeks gestation (risk difference [RD] 8.7%: confidence interval [CI] of RD 1.9–15.5%; P = 0.012; number needed to treat [NNT] 12). Recurrent HD irrespective of gestational age was not different between the arms. No women withdrew as a result of adverse effects. Trial Registration: http://www.isrctn.org) (isrctn87325378). Conclusions: Adding LMWH to aspirin at < 12 weeks gestation reduces recurrent HD onset < 34 weeks gestation in women with inheritable thrombophilia and prior delivery for HD/SGA <34 weeks. However, close monitoring of the mother and fetus remains important throughout pregnancy.
Hypertensive disorders of pregnancy (HD) and small-for-gestational age (SGA) infants are often associated with utero-placental thrombosis. The recurrence risk of HD (pre-eclampsia, hemolysis, elevated liver enzymes and low platelets [HELLP] syndrome and eclampsia) and SGA depends partly on the severity of the index complication. We previously reported a 35% recurrence rate after HD before 34 weeks gestation in women with inheritable thrombophilia . The cause of pre-eclampsia is unknown, but a multifactorial origin is generally accepted. Immune maladaptation (fetal–maternal) and genetics are probably involved while there is consensus that endothelial cell dysfunction/activation represents the final common pathway in the maternal syndrome . Given the frequency of placental thrombosis in HD and SGA, their association with thrombophilic disorders has been debated and explained by possible confounders, but this association has been confirmed in a meta-analysis dedicated to such confounders, particularly in relation to early-onset disease .
Measures to prevent endothelial damage associated with HD and SGA include aspirin, which reduces the recurrence of pre-eclampsia by 12–17% and of SGA associated with intra-uterine growth restriction by 9%, with more effect in women at high risk, and with a greater effect (up to 53%) where treatment is commenced ≤ 16 weeks gestation [4,5].
Women with antiphospholipid syndrome, an acquired thrombophilic state associated with recurrent pregnancy loss, have a better pregnancy outcome when treated with a combination of aspirin and unfractionated or low-molecular-weight heparin (LMWH) than with aspirin alone, although randomized data in respect of HD are few . Recently, the additional value of LMWH in the case of antiphospholipid syndrome has been further disputed . Some physicians have treated women with inheritable thrombophilia with similar combination therapy. However, the premise that the addition of LMWH to aspirin reduces the recurrence of HD more than aspirin alone in women with previous HD and inheritable thrombophilia is still debatable [1,8–10].
We therefore designed a randomized controlled trial (RCT) to test whether adding LMWH to aspirin before 12 weeks gestation reduces the recurrence of HD in women with previous early-onset HD and/or SGA, in the context of an inheritable thrombophilia without antiphospholipid antibodies. (The FRUIT trial, FRactionated heparin in pregnant women with a history of Utero-placental Insufficiency and Thrombophilia).
Patients and methods
The inclusion/exclusion criteria are shown in Table 1. The history of previous pregnancies was confirmed by chart review. Thrombophilia status was established as annotated in Table 1. All the functional tests were performed while the subject was not using oral contraceptives, and tested twice: at least 10 weeks postpartum and at least 6 weeks apart. Central confirmation assessment was not performed either at or after trial entry. All centers were participators in quality assessment programs (in the Netherlands: International Thrombophilia External Quality Assessment; Sweden: Swedish Board for Accreditation and Conformity Assessment; Australia: Royal College of Pathologists of Australia thrombophilia program).
Table 1. Inclusion and exclusion criteria
Pregnancy <12 weeks gestation
Maternal age >18 years
History of uteroplacental insufficiency and delivery before 34 weeks gestation
HD of pregnancy
Pre-eclampsia: pregnancy-induced hypertension (diastolic blood pressure of ≥90 mm Hg and ≥20 mm Hg increment as compared with first trimester diastolic blood pressure), together with proteinuria (≥300 mg per 24 h urine; or spot urine protein/creatinine ≥30 mg mmol−1) after 20 weeks of gestational age
HELLP syndrome: the presence of (i) hemolysis, defined by increased LDH (≥600 IU L−1) and (ii) elevated liver enzymes, defined as increased SGOT/SGPT (≥70 IU L−1); and (iii) thrombocytopenia (<100 × 109 per L)
Eclampsia: generalized convulsions in pregnancy, not caused by epilepsy
Total protein S <60%, protein S antigen <64%, free protein S <60% 
Activated protein C resistance
Activated protein C resistance ratio <2.0 measured once in case of positive FV Leiden mutation, measured twice in case of negative FV Leiden mutation 
Factor (F) V Leiden mutation (heterozygous)
Adenine-to-guanine mutation at nucleotide 506 in the FV gene (polymerase chain reaction)
Prothrombin gene G20210A mutation (heterozygous)
Guanine-to-adenine mutation at nucleotide 20 210 in the prothrombin gene (PCR)
A diastolic blood pressure of ≥90 mm Hg outside of pregnancy, or in the first trimester
Eleven university, six non-university/teaching hospitals
Exclusion criteria: One or more of: antithrombin deficiency, homozygosity for FV Leiden and Prothrombin G20210A mutation, diabetes mellitus, known malignancy, known peptic ulceration, severe renal or hepatic insufficiency, history of venous thrombo-embolism, hemorrhagic diathesis, idiopathic thrombocytopenia, earlier participation in the FRUIT trial and low-molecular-weight heparin (LMWH) use in an earlier pregnancy. HELLP, hemolysis, elevated liver enzymes and low platelets; HD, hypertensive disorders; LDH, lactate dehydrogenase; SGOT/SGPT, serum glutamic oxaloacetic transaminase/serum glutamic pyruvic transaminase; SGA, small-for-gestational age.
Hyperhomocysteinemia was tested for using a methionine loading test (MLT) performed at least 10 weeks after the index pregnancy: cut-off values were fasting homocysteine >15 μm and/or 6-h post load homocysteine > 51 μm. Such women were not excluded from the trial but were treated with folic acid 5 mg daily and pyridoxine 100 mg daily.
If antiphospholipid antibodies (lupus anticoagulant and/or cardiolipin IgG, and/or cardiolipin IgM antibodies) were also been found to be present, women were randomized into a separate study (still ongoing).
Settings and location
Women were recruited from all university hospitals in the Netherlands, two university hospitals in Australia and one university hospital in Sweden, as well as from six non-university/teaching hospitals in the Netherlands.
Eligible women were fully informed and only those who gave written consent were included in the study and subsequently randomized to one or other of the interventions.
Baseline demographic, obstetric and laboratory data were obtained from each subject at trial entry. This included enquiry as to any 1st degree family member with arterial (pre-existing hypertension, angina pectoris, myocardial infarction and cerebrovascular accident) and/or venous (previous thrombosis and pulmonary embolus) disease, and/or having a mother and/or a sister with pre-eclampsia, all summated as a ‘family history of vascular disease’.
Women who declined to participate were treated with aspirin alone. Their permission was asked to record and analyze their baseline data and pregnancy outcome.
The study was approved by the medical ethical committees of all participating hospitals. The study was registered on 11 September 2005 in the Netherlands Trial Register (NTR337) and on 20 December 2005 internationally (isrctn87325378).
Subjects were randomized to receive either daily LMWH, using dalteparin (Fragmin, Pfizer Inc., formerly Pharmacia, New York, NY, USA) 5000 IU subcutaneously, together with daily aspirin 80 mg orally or daily aspirin 80 mg alone. Australian subjects received aspirin 100 mg, and Swedish subjects received aspirin 75 mg, according to their guidelines. Placebo injections were not included. The daily dose of LMWH was adjusted for body weight: subjects below 50 kg received dalteparin 2500 IU, those above 80 kg 7500 IU  and further adjusted as the pregnancy proceeded and postpartum.
Low-molecular-weight heparin was commenced between 6 and 12 weeks gestation, after sonographic confirmation of a viable intrauterine pregnancy, and was continued until the onset of labor. Aspirin was also started before 12 weeks gestation and continued until 36 weeks gestation. After delivery, LMWH was prescribed to all subjects for 6 weeks. All subjects received instructions on self-injection. During follow-up visits, subjects were questioned about the administration of LMWH and about their intake of aspirin. Those who complained of problems at the injection site were inspected for hemorrhage or allergic reactions. Any side effects of LMWH and aspirin, as well as any thromboembolic events, were recorded.
In those subjects who developed local irritation, treatment with dalteparin was changed to enoxaparin and, if persistent, to nadroparin. Both groups received standard care according to national guidelines for high-risk pregnancies. Blood pressure, weight, blood and urine evaluation, and endpoints were serially recorded for each subject. Data were manually transcribed into a database.
All uncertainties/missing data were discussed with the project leader with review of the original source data and, where necessary, requested from the participating center.
Primary A reduction in HD (pre-eclampsia and/or HELLP syndrome, and/or eclampsia) (i) before 34 weeks gestation and (ii) irrespective of gestational age.
Secondary A reduction in SGA infants, spontaneous abortions, preterm births, length of maternal and neonatal admissions, various forms of HD and differences in side effects between the intervention groups.
Randomization was performed by an independent center, using a computer to select random permuted blocks of four, stratifying by hospital and presence/absence of chronic hypertension. Block length and randomization codes were concealed from the investigators.
Assignment was random. Neither study personnel nor participants were blinded to treatment assignment, as placebo injections were not considered to be ethically acceptable during pregnancy.
Percentages and means in both arms of the study were compared with chi-square and/or exact tests, as well as t-tests as appropriate, supplemented with absolute Risk Differences (RD) and 95% confidence intervals for risk difference (CI for RD) for the main outcomes, together with Mean Differences (MD) as appropriate. Results were considered significant at the 5% level. The intention-to-treat principle was followed, but per-protocol analyzes were also performed. A comparison of the primary outcome was corrected for possible confounding with multiple logistic regression. Exploratory analyzes with subgroups were performed to trace possible effect modifications by comparing RD between subgroups. Confounders were defined a priori as: maternal age above 36 years and body mass index (BMI) above 25 kg m−2. The recurrence of HD over time is presented graphically per arm using the actual numbers of HD. Women declining recruitment into the trial were compared with those included.
Sample size In all, 128 subjects were needed to detect a 50% reduction from a 35% recurrence rate of HD  (one-tailed, power 80%). Two-tailed calculations (sample size 262) were initially used, although not considered strictly necessary as no increase in HD and/or SGA infants was expected, nor any serious side effects of LMWH , and with the anticipation of sufficient recruitment, given an estimation of 1% severe HD of all pregnancies per year (180 000) in the Netherlands, 11–24% of which would have inherited thrombophilia [1,17]. The sample size was therefore recomputed using one-sided testing.
At a planned interim analysis in March 2006 no difference in the rate of recurrent HD was found between the arms, and there were no significant side effects. The study was continued.
Between December 2000–2009, 177 women were deemed eligible, 139 were included and 37 declined. One had already participated and was therefore excluded; 27 deviations of protocol occurred in 27 women (Fig. 1).
Baseline characteristics are demonstrated in Table 2. There was a difference in one primary outcome (Table 3). Low-molecular-weight heparin-with-aspirin reduced the recurrence of HD before 34 weeks gestation (RD 8.7%: CI for RD 1.9–15.5%; P = 0.012; NNT 12). There were no early recurrences in the LMWH-with-aspirin arm and six in the aspirin-only arm (at 19, 20, 22, 30 and two at 33 weeks). Of the three women with recurrent HD at 19–22 weeks, all had pre-existing hypertension, all were using antihypertensive therapy and all had had index pregnancies with deliveries between 24 and 28 weeks gestation. One was heterozygous for the Factor (F) V Leiden mutation; one had protein S deficiency; one with FXII deficiency had had a renal transplant 5 years earlier, had stable and optimal kidney function at inclusion, and was using azathioprine 50 mg daily, cyclosporine 250 mg daily and prednisolone 10 mg daily. One had a new partner. None of these subjects had had live births. Of the three other women with recurrent HD before 34 weeks, all required delivery within a few days of diagnosis. All three had had index pregnancies with deliveries between 27 and 29 weeks, all three were heterozygous for the FV Leiden mutation, one had pre-existing hypertension and was using antihypertensive therapy and one had a new partner.
Table 2. Baseline characteristics
LMWH with aspirin
n = 70
n = 69
Values represented as means ± SD or absolute numbers (%) as appropriate. *some women had more than one diagnosis during index pregnancy, aspect of vascular disease in the family, thrombophilia disorder HELLP, hemolysis, elevated liver enzymes and low platelets; APCr, activated protein C resistance; BMI, body mass index; SGA, small-for-gestational age.
Values represented as means ± SD or absolute numbers (%) as appropriate. *RD, risk difference; MD, mean difference; HELLP, hemolysis, elevated liver enzymes and low platelets; LMWH; low-molecular-weight heparin.
Gestational age at delivery – days spontaneous abortion excluded
266.0 ± 24.2
263.1 ± 33.4
Birth weight (grams)
3061.7 ± 740.7
2906.0 ± 971.6
Gestational age of HD at diagnosis (days)
259.3 ± 15.0
233.0 ± 44.4
−26.3 (−37.5, −15.1 days)
Medication started during pregnancy
5.8 (+0.3, +11.3)
Increase in pregnancy duration – days spontaneous abortion excluded
57.6 ± 25.5
56.8 ± 3.33
No difference between the arms was found for the second primary outcome, recurrent HD irrespective of gestational age. The recurrence of HD over time for each group is illustrated in Fig. 2.
The secondary outcomes were also different. In the aspirin-alone group, more subjects had pre-eclampsia before 34 weeks: in line with this, gestational age at delivery was earlier and there was an increased use of corticosteroids for fetal lung maturation.
Skin reactions were only seen in the LMWH-with-aspirin group (Table 3), but only two subjects had to change to an alternative LMWH, after which the reaction disappeared.
Two subjects in the aspirin-alone arm experienced thrombosis: one developed a superficial thrombophlebitis antepartum and one developed a deep venous thrombosis postpartum (both were FV Leiden heterozygous). The latter had not been prescribed LMWH as per-protocol in the postpartum period. There were no complaints (e.g. bone pain) suggestive of osteoporosis up to 6 weeks postpartum.
After correcting with logistic regression for the possible confounders (maternal age above 36 years and BMI >25 kg m−2), the effect of LMWH-with-aspirin on recurrent HD before 34 weeks gestation remained. Subgroup analysis for the same parameters revealed no significant effect modification. The stratification variables had no influence on the primary outcome.
When the results were analyzed per-protocol, after exclusion of eight subjects with an entry criteria protocol violation (one previous fetal death without HD/SGA, four previous pulmonary embolism, one FXII deficiency, two positive for antiphospholipid antibodies), comparison of the remaining 66 subjects in the LMWH-with-aspirin arm and 65 in the aspirin-alone arm revealed a similar reduction in HD before 34 weeks gestation (0 compared with five: P = 0.022, CI + 1.2, + 14.2%) and no effect on HD irrespective of age (P = 0.630, CI−10.1, +16.7%).
When the three subjects who delivered around 20 weeks gestation were excluded from the intention-to-treat analysis, the reduction in HD before 34 weeks gestation remained significant when tested one-sided (P < 0.05) but only tended towards significance when two-sided analysis was performed (P < 0.10).
Baseline characteristics and outcomes of the women who declined to enter the trial were generally not different from those of the included women (data not shown).
This randomized trial has shown that adding LMWH to aspirin before 12 weeks gestation reduces recurrent HD in women with previous early-onset HD and/or SGA, in the context of an inheritable thrombophilia without antiphospholipid antibodies. This reduction affects recurrent early-onset HD, and in particular pre-eclampsia, before 34 weeks gestation, with a highly favorable NNT. It results in longer gestation and less need for corticosteroid therapy for fetal lung maturation. No such difference was seen in the overall recurrence of HD irrespective of gestational age.
Side effects of LMWH occurred in 11% of subjects, but only 2/70 subjects had to change from the prescribed form of LMWH. Systemic thrombotic complications were only seen in the aspirin-alone arm. There were no complaints suggestive of osteoporosis, but bone density was not measured. Reassuring data regarding this have been provided from the Thrombophilia in Pregnancy Prophylaxis Study (TIPPS) trial .
Sources for potential bias have been examined. Having stratified by the presence of chronic hypertension and by treating hospital, evaluation of these factors revealed no influence on the primary outcome. The lack of a difference in baseline characteristics from those who declined recruitment suggests that a good representative sample was studied.
Logistic regression did not reveal potential confounders; subgroup analysis looking at effect modification demonstrated no impact on recurrent HD from the same parameters.
While this study shows a difference in outcome for the combination of LMWH and aspirin commenced before 12 weeks gestation in this carefully selected cohort, the numbers are small. Early-onset recurrent HD has been demonstrated in observational studies [1,8,19]. In a large, prospective cohort with unknown thrombophilia, the risks after a first pregnancy with early-onset pre-eclampsia were: 29% recurrent pre-eclampsia irrespective of gestational age and 6.8% recurrent early-onset pre-eclampsia . In a Dutch cohort with inheritable thrombophilia, recurrent HD occurred before 34 weeks in 2/26 subjects treated with LMWH and aspirin (at 31 and 33 weeks), and in 7/32 with aspirin alone or on no medication (at 21, 24, 28, 30, 31 and twice at 33 weeks) . In the present study, the three women with recurrent HD at 19–22 weeks had no autoimmune disease. All three had chronic hypertension. One had chronic renal disease, as well as FXII deficiency, which was not generally tested at the beginning of the study, and was therefore excluded in the per-protocol analysis. Chronic hypertension is a known risk factor for HD , as is highlighted by these three cases and the overall high prevalence of chronic hypertension in the FRUIT subjects. The difference in recurrent early-onset HD persisted after per-protocol analysis as well as after exclusion of the three outliers, although it was reduced.
To date there have been no completed trials studying recurrent early onset pre-eclampsia in the same patient population, although the TIPPS trial that is currently in progress is including such women in the inclusion criteria . One RCT in a population with the same obstetrical history but in women without known thrombophilia reported a beneficial effect of dalteparin on a composite outcome including recurrent severe pre-eclampsia, intrauterine growth restriction and a major abruption : there was a trend to benefit in the reduction in early-onset pre-eclampsia in this previous study. It may even be that our study represents a similar effect, and that thrombophilia is irrelevant to the outcome. Meanwhile, our view is that these data underline the importance of counseling women with the various forms of thrombophilia in the light of their specific obstetric history.
It is hard to explain the differential effect of LMWH in women with inheritable thrombophilia in the reduction in recurrent early onset HD but not in the overall frequency of recurrent HD. The exact mechanism of how LMWH causes such a reduction in HD is not currently known. In vitro studies have shown a differential effect of heparin and LMWH on angiogenesis in placental villi , but it is uncertain how this reflects the in vivo situation. The present data may also support the hypothesis that heparin might impact on the dysregulation of soluble vascular endothelial growth factor seen in pre-eclampsia . In the ongoing TIPPS trial, dalteparin, given in the same dosage as in the FRUIT study at least until 20 weeks and during the second wave of trophoblast invasion, did significantly increase plasma anti-Xa activity, but had no significant effect on the coagulation activation markers: thrombin-antithrombin complexes, prothrombin fragments 1 and 2, and D-dimer concentrations . The mechanisms of positive and negative influences of heparin on trophoblast invasion are increasingly studied and discussed . The present data suggest that at least the first part of the second wave of trophoblast invasion can be supported by LMWH treatment, leading to reduced early-onset HD.
In the FRUIT study, the dose of LMWH was adjusted to maternal weight . Other investigators have used an increased dose of LMWH after 20 weeks [24,26], whereas others have not  or have started with lower prophylactic dosage without an increase after 20 weeks . The dosage of dalteparin 5000 U day−1 is in line with a pharmacokinetic study during pregnancy of Ensom et al. . Further studies are required to elucidate the correct strategy.
The recurrence rate of HD in the present study was 20%, substantially lower than the 35% found in the Dutch retrospective study in the same centers with an identical frequency of inheritable thrombophilia . It is not clear why there should be such a large difference, beyond the usual caveats for comparing retrospective studies and prospective intervention trials.
The strengths of the present study are its strict inclusion criteria, the limitation of entry to women with previous early-onset HD and/or SGA, and the early initiation of treatment, in the context of inheritable thrombophilia, screened similarly across the eligible population, not limited to those with known genetic etiology, but also including functional deficiencies. On the other hand, it was only possible to mobilize a small population in 10 years of inclusion. Another limitation of the present study is the lack of a non-intervention control group: having such a group was considered unethical but it may have an impact on the interpretation of the results. Finally, laboratory testing for thrombophilia was not centralized and was not repeated after randomization.
In conclusion, women with a history of early-onset HD and/or SGA, with an inheritable thrombophilia and without evidence of antiphospholipid antibodies, have for the first time been shown to benefit from a combination of LMWH and aspirin started before the second trimester of pregnancy. Although the incidence of all recurrent HD irrespective of gestational age did not differ between the treatment groups, in the LMWH-with-aspirin group HD never developed before 34 weeks, thus reducing the considerable financial and emotional load associated with such adverse outcomes.
FRUIT investigators: J. I. de Vries, M. G. van Pampus, W. M. Hague, P. D. Bezemer, J. H. Joosten performed the conception and the design of the study, analysis and interpretation of the data, drafting of the manuscript. J. H. Bruinse, University Medical Center, Utrecht, the Netherlands: designed and performed research, collected data. J van Eyck, Isala Hospital, Zwolle, the Netherlands: designed and performed research, collected data. J. M. Middeldorp, Leiden University Medical Center, Leiden, the Netherlands: designed and performed research, collected data. J. T. Brons, Medical Spectrum Twente, Enschede, the Netherlands designed and performed research, collected data. W. Visser, Academic Hospital Rotterdam, Dijkzigt, the Netherlands: designed and performed research, collected data, revised the intellectual content. G. A. Dekker, University of Adelaide, Lyell McEwin Hospital, Adelaide, Australia: designed and performed research, collected data, obtained funding, revised the intellectual content. H. Wolf, Academic Medical Center, Amsterdam, the Netherlands: designed and performed research, collected data, revised the intellectual content. M. E. Spaanderman, Radboud University Nijmegen Medical Center, the Netherlands: designed and performed research, collected data. C. J. de Groot, Medical Center Haaglanden, The Hague, the Netherlands: designed and performed research, collected data, revised the intellectual content. S. G. Oei, Máxima Medical center, Veldhoven, the Netherlands: designed research, collected data. D. A. Smit, Twenteborg Hospital, Almelo, the Netherlands: designed research, collected data. K. Bremme, Karolinska Institute, Stockholm, Sweden: designed and performed research, collected data, revised the intellectual content.
The authors would like to thank all the women who volunteered to take part in the study. A. A. M. Hart, M. Chamuleau, R. A. Stigter and K. W. Bloemenkamp for their work in the interim analysis. A. M. S. van Poelgeest, D. L. van der Mast for data management. The Netherlands Journal of Medicine for allowing us to publish data not presented in Kalk et al. . J. J. Kalk for her contribution on the set-up of the study.
Disclosure of Conflict of Interest
The study was supported by a single 2-year investigator grant period 2000–2001 by Pfizer, formerly Pharmacia grant number 524E-CVD-9101-0001, annual Dutch investigators meetings, a single grant to support a midwife to recruit Australian subjects, and support for a local meeting in Sweden in 2004. Pharmacia was not the sponsor of the study. The authors state that they have no conflict of interest.