Setting Maternity clinics in ten Finnish hospitals participating in the PREDO Project.
Sample A total of 152 women with risk factors for pre-eclampsia and abnormal uterine artery Doppler velocimetry.
Methods Participants were randomised to start either aspirin 100 mg/day or placebo at 12 + 0 to 13 + 6 weeks + days of gestation. Because of the limited power of this trial, we also conducted a meta-analysis of randomised controlled trials that included data on 346 women with abnormal uterine artery Doppler flow velocimetry, and aspirin 50–150 mg/day started at or before 16weeks of gestation.
Main outcome measure Pre-eclampsia, gestational hypertension and birthweight standard deviation (SD) score. Outcome measures for the meta-analysis were pre-eclampsia, severe pre-eclampsia, preterm (diagnosed <37 + 0 weeks of gestation) and term pre-eclampsia.
Results From the 152 randomised women, 121 were included in the final analysis. Low-dose aspirin did not reduce the rate of pre-eclampsia (relative risk [RR] 0.7, 95% CI 0.3–1.7); gestational hypertension (RR 1.6, 95% CI 0.6–4.2); early-onset pre-eclampsia (diagnosed <34 + 0 weeks of gestation) (RR 0.2, 95% CI 0.03–2.1); or severe pre-eclampsia (RR 0.4, 95% CI 0.1–1.3); and the results were not statistically significant in an intention-to-treat analysis. However, our meta-analysis, including the current data, suggested that low-dose aspirin initiated before 16 weeks of gestation reduces the risk of pre-eclampsia (RR 0.6, 95% CI 0.4–0.8) and severe pre-eclampsia (RR 0.3, 95% CI 0.1–0.7).
Conclusions Our trial showed no statistically significant effect of aspirin in preventing pre-eclampsia in high-risk women. However, our meta-analysis suggested that aspirin may reduce the incidence of pre-eclampsia.
Pre-eclampsia remains one of the most important challenges in obstetrics. The disorder affects 3–5% of pregnancies and is defined according to new-onset hypertension and proteinuria, which appear after 20 weeks of gestation.1 It is a multisystem disease with adverse short-term and long-term outcomes to both the mother and the fetus. In total, pre-eclampsia and related complications account for 63 000 maternal deaths worldwide every year, 12% of all maternal deaths.2 The onset and clinical course are unpredictable, and there is a strong need for tools to predict and prevent the disorder.
The aetiology of pre-eclampsia remains unknown, although placental dysfunction, which is due to early placental developmental abnormality, is central in the disease process. The early placental disease is followed months later by the clinical manifestations of pre-eclampsia, which reflect widespread endothelial dysfunction, resulting in vasoconstriction, end-organ ischaemia and increased vascular permeability.3 Many of the proposed prediction and prevention strategies are based on processes involved in placental development in early pregnancy, although none of these has been established in clinical practice.
Antiplatelet agents, such as aspirin (acetylsalicylic acid), are among the most promising candidates for prevention of pre-eclampsia. They have a positive effect on the balance between prostacyclin, a vasodilator, and thromboxane, a vasoconstrictor and stimulant of platelet aggregation. This process plays a key role in the development of the disease and is believed to result from shallow placental invasion and ischaemia that occur shortly after implantation. A recent meta-analysis, based on 27 trials on 31 678 women, concluded that aspirin is effective in preventing pre-eclampsia, although the effect was too modest to warrant routine use in all women.4 However, if started early in pregnancy, in high-risk women, the treatment may be effective,5,6 although studies are few and results are inconsistent.7–9
Our aim was to study the effect of aspirin started at 12 + 0 to 13 + 6 weeks + days of gestation on prevention of pre-eclampsia and intrauterine growth restriction in high-risk women identified by abnormal uterine artery flow. We performed the trial in conjunction with the multidisciplinary PREDO Project, which we also describe to set the study in context. In addition, we combined our data with data from similar previous trials in meta-analysis.
The Predo Project
The multidisciplinary PREDO Project ‘Prediction and Prevention of Pre-eclampsia’ had three arms: obstetric (including the present aspirin trial), genetic and psychological. The project was carried out between September 2005 and December 2009. We recruited 947 pregnant women with risk factors for pre-eclampsia and 117 pregnant women without known risk factors as a comparison group at 12 + 0 to 13 + 6 weeks + days of gestation (Figure 1). The recruitment took place when these women attended the first ultrasound screening in one of ten hospital maternity clinics participating in the PREDO Project; Women’s Hospital, Kätilöopisto Maternity Hospital and Jorvi Hospital at Helsinki University Central Hospital, Kanta-Häme Central Hospital, Päijät-Häme Central Hospital, Tampere University Hospital, Kuopio University Hospital, Northern Karelia Central Hospital and Iisalmi Hospital. A written informed consent was obtained from all participants. We also enrolled the spouse of each study participant (= biological father of the child) for the genetic arm of the PREDO Project.
Inclusion criteria and definitions
The inclusion and exclusion criteria of the aspirin trial are presented in Table 1. Women with one or more risk factors for pre-eclampsia were invited in arrival order to participate unless any of the exclusion criteria was present.
Table 1. Inclusion criteria in women randomised to aspirin or placebo in the ‘Prediction and Prevention of Pre-Eclampsia’ (PREDO) Project
Aspirin (n = 61)
Placebo (n = 60)
An additional inclusion criterion was systemic lupus erythematosus but this was not present in any of the participants.
The exclusion criteria were allergy to aspirin; tobacco smoking (during this pregnancy); multiple pregnancy; and a history of asthma, peptic ulcer, placental ablation, infammatory bowel diseases (Crohn’s disease, colitis ulcerosa), rheumatoid arthritis, haemophilia or thrombophilia (previous venous or pulmonary thrombosis or coagulation abnormality).
In all, 36 women fulfilled more than one inclusion criteria.
Age under 20 years
Age over 40 years
Obesity (body mass index over 30 kg/m2)
Chronic hypertension (≥140/90 mmHg or medication for hypertension before 20 weeks of gestation)
A history of one of the following conditions:
Pre-eclampsia (blood pressure ≥140 mmHg systolic or ≥90 mmHg diastolic and proteinuria ≥0.3 g/day or dipstick equivalent in two consecutive measurements)
Small for gestational age (birthweight <−2SD)
Fetus mortus (fetal death after 22 weeks of gestation or >500 g weight in a previous pregnancy)
Ultrasound measurements and aspirin
We measured uterine artery blood flow by colour Doppler ultrasound transvaginally from all participants at 12 + 0 to 13 + 6 weeks of gestation. The uterine artery was identified at the level of the internal cervical os, as it approached the uterus laterally. We defined the second-degree uterine artery notch as a notch in the beginning of diastole at least as deep as the end diastolic notch (see Supplementary material, Figure S1).10,11 Women who had bilateral second-degree notch were allocated to the medication group. They were randomised (see paragraph ‘Randomisation and Blinding’ below) to start either aspirin 100 mg/day or placebo, which were continued until 35 + 0 weeks of gestation or delivery, whichever occurred first. Those women who did not fulfil the criteria for the medication group were taken into the follow-up group as described in the flow chart (Figure 1).
Primary outcomes were pre-eclampsia (blood pressure ≥140 and/or 90 mmHg in two consecutive measurements and proteinuria ≥0.3 g/24 hours),1 gestational hypertension (new onset hypertension after 20 weeks of gestation),1 and birthweight SD score as a continuous variable calculated according to Finnish standards.12
Secondary outcomes were early-onset pre-eclampsia (pre-eclampsia diagnosed before 34 + 0 weeks of gestation), severe pre-eclampsia (blood pressure ≥160 systolic and/or ≥110 diastolic and/or proteinuria ≥5 g/24 hours),1 preterm pre-eclampsia (pre-eclampsia diagnosed before 37 + 0 weeks of gestation), small for gestational age (SGA) (birthweight <−2SD),12 and length of gestation (continuous variable).
Each individual outcome diagnosis was set by a jury, which consisted of two physicians and a study nurse. They met face-to-face and reviewed the hospital and maternity clinic records of each participant.
Randomisation and blinding
This was an investigator-initiated, randomised, placebo-controlled, double-blinded trial. The Tampere University Hospital Pharmacy performed the randomisation. As a paid service, the aspirin and placebo tablets were prepared by a pharmaceutical company (Orion®, Espoo, Finland) to appear identical. Tampere University Hospital Pharmacy repacked and randomised the tablets. The randomisation was made in blocks of tens by the pharmacists not otherwise involved in the study. The randomisation code of each participant was sealed in an envelope and was opened after the outcome diagnoses of all participants had been set by the jury, as described above.
A computerised search was conducted from January 1965 through January 2012 of the MEDLINE database and the Cochrane library using the search terms: aspirin, antiplatelet, asa, acetylsalicylic, eclamp*, hypertens*, intrauterine growth restriction, SGA, toxaemia, PIH, pregnancy-induced hypertension, Doppler, ultrasound, notch, uterine artery. We did not exclude any manuscript based on language.
We included in the meta-analysis the prospective, randomised, controlled trials, which met the following criteria: 1) included women with abnormal uterine artery Doppler flow velocimetry, and 2) started aspirin at or before the 16 weeks of gestation, with dose between 50 and 150 mg/day. The control group had to be allocated either to placebo or no treatment. Through the literature search we identified 1414 eligible studies. In addition to our study, only two studies fulfilled the inclusion criteria.7,9 The results were available for a total of 346 women.
The outcome measures for the meta-analysis were pre-eclampsia, severe pre-eclampsia, preterm and term pre-eclampsia. Definition for pre-eclampsia was consistent between studies (blood pressure ≥140/90 mmHg, and proteinuria ≥0.3 g/24 hours). Pre-eclampsia was defined preterm when diagnosed before 37 completed weeks of gestation. Severe pre-eclampsia was defined if in addition to above criteria one or more of the following criteria were present: blood pressure ≥160 mmHg systolic and/or ≥110 mmHg diastolic, or severe proteinuria (definition between studies ranging from ≥2 g to ≥5 g/24 hours). Oliguria <500 ml/24 hours, elevated liver enzymes, and platelet count <100 000/mm3, or fetal growth restriction.
Continuous variables were tested for normality. Highest proteinuria concentrations per day were log transformed to attain normality. Continuous variables between study groups were compared using the independent sample t-test and categorical variables by chi-square test. Two-tailed P-values <0.05 were considered statistically significant. Relative risks were calculated to compare the risk of each outcome between aspirin and placebo groups. Based on a previous study,7 we expected an incidence of 25% for pre-eclampsia among the study participants. We calculated that with a power of 0.80 and an α of 0.05 we would be able to confirm or exclude a reduction in incidence to 10% in groups of 80 participants each. For groups of 60 and 61 participants, which was the number included in analysis, the corresponding power is 0.62.
Studies included in the meta-analysis were combined and analysed using comprehensive meta-analysis V 2.0 software (Biostat Inc., Englewood, NJ, USA). Individual relative risks were calculated for each study, and pooled for global analysis with 95% confidence intervals (CI). Global RR were calculated according to Der Simmonian and Laird random effect models in case of heterogeneity. Heterogeneity between studies was analysed with Higgins’I2 and considered to be high if >50%. A random effects models was used for all outcomes, because heterogeneity for both term and preterm pre-eclampsia was 75%. For pre-eclampsia heterogeneity was 14%, and for severe pre-eclampsia, it was 0%. Because of the small number of studies, a funnel plot analysis to assess publication bias was not conducted.
Out of the 947 women recruited, 152 (16.0%) with bilateral second-degree diastolic notch in the uterine artery flow were allocated into the aspirin trial. One hundred and twenty-one women completed the trial.
Characteristics of the women participating in the aspirin trial
The baseline characteristics of the 61 women allocated randomly into the aspirin group and the 60 women in the placebo group are presented in Table 2. Table 3 shows their pregnancy characteristics.
Table 2. Baseline characteristics
Aspirin group (n = 61)
Placebo group (n = 60)
BMI, body mass index.
Continuous data presented as mean (standard deviation, SD).
Age, years (SD)
BMI before pregnancy, kg/m2 (SD)
Height, cm (SD)
Primiparous, n (%)
Elementary or less
High school or vocational school
Table 3. Pregnancy characteristics
Aspirin group (n = 61)
Placebo group (n = 60)
Continuous data presented as mean (SD).
**No umbilical artery pH was below 7.00.
Antihypertensive medication, n(%)
Before 20 weeks of gestation
After 20 weeks of gestation
Weight gain during pregnancy, kg (SD)
Gestational diabetes, n(%)
Oral glucose tolerance test not performed, n (%)
Highest systolic blood pressure, mmHg (SD)
Highest diastolic blood pressure, mmHg (SD)
Highest proteinuria, g/day*
Mode of delivery, n(%)
Elective caesarean section
Caesarean section during labour
Apgar score at 5 min
Umbilical artery pH below 7.15,**n (%)
Newborn birthweight, g (SD)
Placental weight, g (SD)
Subjects who discontinued the trial
Of the 152 women initially recruited into the aspirin trial, 31 women were left out of the study (see Supplementary material, Figure S2). Four of these women had a miscarriage, three in the aspirin group and one in the placebo group. Two of these miscarriages took place at 14 weeks of gestation and one at 19 weeks of gestation in the aspirin group and one at 18 weeks of gestation in the placebo group. Eleven women were lost to follow up or discontinued for various nonmedical reasons; seven of these were in the aspirin group and four in the placebo group. Five women decided to discontinue the aspirin trial because of a medical condition. Three of these women were in the placebo group and two in the aspirin group. Eleven participants were additionally excluded from analysis because of noncompliance with the study protocol. The pregnancy outcome of all of these 31 women is known. Three women had one of our primary or secondary outcomes. One woman, who cancelled her involvement in the trial 1 day after the entry, and did not start the medication, subsequently developed early pre-eclampsia. Another woman with Factor V Leiden mutation started low-molecular-weight heparin, and had to discontinue the trial; she gave birth to a SGA newborn. Both of these women were randomised to aspirin group. One woman from the placebo group, who discontinued the trial because of thrombocytopenia, developed gestational hypertension. We conducted an intention-to-treat analysis, in which we included all randomised women, except the ones that had a miscarriage. The results of our intention-to-treat analysis do not differ from the results of the analysis made without these excluded women; risk ratios (RR) in the aspirin group were as follows for: pre-eclampsia 0.8 (95% CI 0.4–1.8), gestational hypertension 1.4 (95% CI 0.6–3.5), early pre-eclampsia 0.5 (95% CI 0.1–2.6), preterm pre-eclampsia 0.8 (95% CI 0.2–2.8), severe pre-eclampsia 0.5 (95% CI 0.2–1.6), SGA newborn 0.5 (95% CI 0.1–1.9), and severe diagnosis 0.6 (95% CI 0.2–1.6). None of these associations were statistically significant.
As shown in Table 4, 19 (15.7%) women were diagnosed with pre-eclampsia, eight in the aspirin group and 11 in the placebo group (RR 0.7, 95% CI 0.3–1.7). Sixteen women were diagnosed with gestational hypertension, ten in the aspirin and six in the placebo group (RR 1.6, 95% CI 0.6–4.2). Birthweight SD score in the aspirin group was – 0.1 (SD = 1.1) and in the placebo group – 0.3 (SD = 1.3) (P = 0.3). These were not statistically significant.
Table 4. Study outcomes in aspirin and placebo groups, and relative risk (RR) with 95% confidence intervals (CI) for each outcome.
*****Early pre-eclampsia and/or severe pre-eclampsia and/or small for gestational age.
Small for gestational age****
There was one woman with early-onset pre-eclampsia in the aspirin group and four in the placebo group (RR 0.2, 95% CI 0.03–2.1). Severe pre-eclampsia was diagnosed in three women in the aspirin group and in eight women in the placebo group (RR 0.4, 95% CI 0.1–1.3). In the aspirin group there were two newborns diagnosed as SGA compared with placebo group with six (RR 0.3, 95% CI 0.1–1.6). These diagnoses were in part seen in the same participants. Four women in the aspirin group and ten in the placebo group had one or more of these severe diagnoses (early-onset pre-eclampsia and/or severe pre-eclampsia and/or SGA) (RR 0.4, 95% CI 0.1–1.2). Three women in the aspirin group and five women in the placebo group (RR 0.6, 95% CI 0.2–2.4) developed preterm pre-eclampsia (diagnosed before 37 + 0 weeks of gestation). Mean gestational age in the aspirin group was 39.1 weeks (SD = 0.8) and in the placebo group 38.9 weeks (SD = 3.0) (P = 0.6). None of these differences were statistically significant (Table 4). One woman in the placebo group had HELLP syndrome (haemolysis, elevated liver enzymes, and low platelets) with early pre-eclampsia. None had eclamptic seizures.
There was no difference in the incidence of pre-eclampsia between the aspirin and placebo groups when women with body mass index over 30 kg/m2 were analysed separately.
Among the 795 women who were included in the study but whose uterine artery Doppler ultrasound did not fulfil the criteria of the aspirin trial (follow-up groups), 66 (8.3%) developed pre-eclampsia, 24 (3.0%) of these women were diagnosed with severe pre-eclampsia, and 16 (2.0%) with early-onset pre-eclampsia. Eighty-nine women were diagnosed with gestational hypertension (11.1%). Twenty-four (3.0%) newborns were born SGA and ten (1.3%) women both gave birth to an SGA newborn and were diagnosed with pre-eclampsia.
One participant reported sudden deafness in one ear at 24 weeks of gestation. The medication was discontinued and the randomisation code was opened: this participant had received placebo. No other adverse effects were reported.
Our meta-analysis included two additional studies.7,9 In the meta-analysis aspirin started at or before 16 weeks of gestation in women whose uterine artery Doppler ultrasound indicated an increased risk, significantly reduced the risk of pre-eclampsia (RR 0.6, 95% CI 0.37–0.83) (Figure 2), and severe pre-eclampsia (RR 0.3, 95% CI 0.11–0.69) (Figure 3). Aspirin did not reduce the risk of preterm pre-eclampsia (RR 0.2, 95% CI 0.02–1.26) (Figure 4) or term pre-eclampsia (RR 1.0, 95% CI 0.25–4.26) (Figure 5).
We did not find statistically significant benefit for the effect of low-dose aspirin in preventing pre-eclampsia or related traits in women identified by clinical risk factors and bilateral uterine artery second-degree notch in early pregnancy. However, our meta-analysis showed that aspirin may be effective in preventing pre-eclampsia.
Comparisons with previous trials
In recent meta-analyses aspirin and other antiplatelet agents have shown a moderate but consistent reduction in the risk of pre-eclampsia. In the Paris collaboration meta-analysis of 32 217 mothers,4 which included randomised studies regardless of their inclusion criteria, the relative risk of pre-eclampsia was 0.9 in women receiving antiplatelet agents compared with control women. Whereas this reduction was not sufficient to warrant treatment for all pregnant women, the authors recommended low-dose aspirin started in early pregnancy to women with high-risk of pre-eclampsia. However, specific criteria for a high-risk group could not be identified based on the reviewed literature. Cochrane review of antiplatelet agents for prevention of pre-eclampsia,13 updated in 2007, demonstrated a 17% reduction in the incidence when combining studies of different design. Bujold et al.5 conducted a meta-analysis (11 348 women) of 27 studies in which the time of start of aspirin could be identified. They found a significant reduction of the incidence of pre-eclampsia. When aspirin was started at 16 weeks of gestation or earlier (n = 764), 36 women developed pre-eclampsia, and in the control group 80 women developed pre-eclampsia (relative risk 0.47, 95% CI 0.34–0.65), with little if any heterogeneity between the studies. This finding is consistent with our meta-analysis. If aspirin was started after 16 weeks of gestation (n = 10 584) there was no reduction of the incidence of pre-eclampsia (relative risk 0.81, 95% CI 0.63–1.03).
The results of our meta-analysis, together with the results of the meta-analyses by Bujold et al.5,6 are in agreement with previous suggestions that aspirin in prevention of pre-eclampsia should be started in early gestation, before the second active phase of trophoblast invasion, which takes place from 14 weeks of gestation onwards.14 During that phase the trophoblast invasion is completed. Although in our study aspirin was started between 12 + 0 and 13 + 6 weeks of gestation, an even earlier start of treatment might carry more benefits. This was suggested by a recent study,15 in which women received aspirin or placebo from the time of in vitro fertilisation until 12 weeks of gestation. The incidence of hypertensive complications was lower in the aspirin group (3.6% versus 26.9%, P = 0.02). However, this was not confirmed in another study16 in which aspirin was also started before pregnancy and the incidence of hypertensive pregnancy complications did not differ significantly between the low-dose aspirin (n = 52) and placebo (n = 52) groups (15.4% versus 18.2%, P = 0.7).
The most important limitation to this aspirin trial was the relatively small sample size. The main reason for this was the use of bilateral second-degree uterine artery notch as a criterion in selecting women into the medication group from the group with clinical risk factors. We hypothesised that this criterion would distinguish those women with highest risk for pre-eclampsia. However, the number of women fulfilling the criterion for medication, 16%, was unexpectedly small. Further, the number of women who developed pre-eclampsia was again small, as compared with previous studies with a similar design. In hindsight a more lax Doppler criterion, perhaps with focus on nulliparous women, could have been better in assessing the effect of aspirin in high-risk women. In the trial by Vainio et al.7 women with clinical risk factors similar to those in our study were allocated into the medication group if they had a bilateral first-degree notch in the uterine artery flow velocimetry. In that study those randomised to aspirin started at 12–14 weeks of gestation showed a significant reduction of pregnancy-induced hypertension (11.6% versus 37.2%, RR 0.31, 95% CI 0.13–0.78) and pre-eclampsia (4.7% versus 23.3%, RR 0.2, 95% CI 0.05–0.86) with aspirin started at 12–14 weeks of gestation. In Vainio’s study 70% of women with clinical risk factors had bilateral first-degree notch in the uterine artery flow. In general, caution should be exercised in evaluating trials with test–treatment combinations.17 Our trial and others with similar design do not reveal how Doppler ultrasound measurement performs in prediction of pre-eclampsia. Nor does this kind of study design find out whether women with normal uterine artery finding would benefit from aspirin. The rationale of conducting meta-analysis was to overcome the small sample size of the present study. The main limitation for the meta-analysis was the small number of eligible studies. Moreover, the heterogeneity for term and preterm pre-eclampsia suggests variability between studies. However, great homogeneity for pre-eclampsia and severe pre-eclampsia suggests valid findings.
Placental dysfunction is a result of the shallow invasion of trophoblasts into the placental bed spiral arteries,18 which leads to reduced placental perfusion and ischaemia. This activates platelets and causes an imbalance of the prostacyclin–thromboxane ratio in favour of vasoconstrictive and aggregatory thromboxane. Prostacyclin is produced by endothelial cells and is vasodilatory and anti-aggregatory. The hypothesis of aspirin in preventing pre-eclampsia is based on its effect on prostaglandin production. Low-dose aspirin inhibits thromboxane production of platelets but the production of prostacyclin by endothelial cells stays intact. The dosage of 0.5–2.0 mg/kg of aspirin significantly inhibits the production of thromboxane but leaves prostacyclin production unaffected.19 It is however of note that this process is most likely to be active in early-onset, severe pre-eclampsia. The meta-analyses, performed by us and others, support the hypothesis that aspirin started early is effective in preventing pre-eclampsia. However, further studies are needed, especially to assess the effectiveness of aspirin on early-onset pre-eclampsia. Aspirin may delay the onset of pre-eclampsia.
Our study included by design a larger proportion of women with a history of select pregnancy disorders (implying an excess of multiparous women), obesity and chronic hypertension, and who would be expected to be at an increased risk of the late-onset form of the disease.20 That said, it should also be noted that because there is excess inflammatory action in obesity, obese women might benefit from the anti-inflammatory action of aspirin,21 which however was not seen in our study.
Many women who develop severe pre-eclampsia are nulliparous with no other known clinical risk factor. To include them in a prevention trial, an early predictive biochemical, or genetic marker, or a combination of predictive tests would be necessary. Although we did not use a biochemical risk marker in the present study, this area has developed substantially in previous years. The risk calculating programs, which take into account risk factors, one or more biochemical measurements and ultrasound measurements, hold promise.21,22
In conclusion, although both early-onset and severe pre-eclampsia, as well SGA newborns, were more common in the placebo group than in the group receiving 100 mg daily dose of aspirin from approximately 12–35 weeks of gestation, the differences between the two groups were not statistically significant. However, supported by the results of our trial and meta-analysis, the role of aspirin in prevention of pre-eclampsia warrants further investigation. We focused on one specific treatment regimen in women with clinical risk factors for pre-eclampsia. Together with the recent rapid development of tools for predicting pre-eclampsia, our results encourage the use of biochemical risk markers, possibly in combination with assessment of uterine artery flow, for early identification of women at risk for future trials to prevent pre-eclampsia.
Disclosure of interest
There are no conflicts of interests.
Contribution to authorship
The authors made the following substantial contributions to this work: conception and design: PV, EK, KR, AKP, EH, MV, PT, HL; acquisition of data: PV, MV, PT, AAT, AMH, VKH, TK, LKN, EK, MK, RM, PS, RS, TS, SSK, SS, VMU, JU, TV, TW, HL, or analysis and interpretation of data: PV, EK, HL; drafting the article: PV, EK, HL; and revising it critically for important intellectual content: all authors. Final approval of the version to be published was given by all authors.
Details of the ethics approval
The PREDO Project was approved by the Ethics Committee of Obstetrics and Gynaecology, Hospital District of Helsinki and Uusimaa (Dnro HUS 3/E8/05). It is registered as an International Standard Randomised Controlled Trial number ISRCTN14030412.
This work was supported by Academy of Finland (KR, A-KP, EK, HL), Clinical Graduate School in Paediatrics and Obstetrics/Gynaecology, University of Helsinki (PV), Finnish Medical Society Duodecim (PV), Emil Aaltonen Foundation (EK), Finnish Concordia Fund (PV), Finnish Foundation For Paediatric Research (EK), Finnish Medical Foundation (EK, HL), Signe and Ane Gyllenberg Foundation (KR, EK), Sigrid Juselius Foundation (EK), Government Special Subsidy for Health Sciences at Helsinki and Uusimaa Hospital District (PT, HL, PREDO Project), Jane and Aatos Erkko Foundation (HL), Orion Foundation (PV), Päivikki and Sakari Sohlberg Foundation (PV, HL), Yrjö Jahnsson Foundation (KR).
We thank all our study nurses for their enthusiastic efforts in recruitment, data acquisition and data recording. We are grateful to the women who participated in the PREDO Project.
Commentary on ‘Aspirin in the prevention of pre-eclampsia in high-risk women’
Why is BJOG publishing a randomised trial of low-dose aspirin for the prevention of pre-eclampsia involving only 152 women when there is already, not just a Cochrane review (Duley et al. Cochrane Database Syst Rev 2007;2:CD004659) involving 37 560 women dealing with the same question, but also an individual patient meta-analysis (Askie et al. Lancet 2007;369:1791–8) involving 32 217 women? These both showed with considerable precision, that aspirin reduces pre-eclampsia, preterm delivery and severe adverse outcomes.
Did participants really understand the results of the previous trials? How was ethical approval justified in light of previous results? The trial recruitment started in 2005 and the trial registration document (www.controlled-trials.com/ISRCTN14030412/) showed a planned sample size of 1000 participants, but only 152 were eventually randomised. No less than 31 (20%) of participants were excluded from analysis for various reasons after randomisation, leaving only 121 in the analysis group. The small sample size of the present trial is not the only problem. In the trial registration document no < 43 primary outcomes are listed! The paper reports three primary outcomes, only one of which was also listed in the trial registration document. No explanation is given for either discrepancy in the paper.
Informed readers may raise their eyebrows, but what can we learn. The trial is now completed and nothing can be changed. We learn first that women had participated and the data they had generated should not be wasted, so there is an ethical ground for publication. Second, the trial results could be pooled with two other small trials into a new meta-analysis of the effect of aspirin in a subgroup of high-risk women who also have abnormal uterine artery Doppler waveforms at 14 weeks. We learn that it differs little from its effect in other high-risk groups. Third, we learn that overall results of the kind reported in previous comprehensive reviews (Cochrane Database Syst Rev 2007;2:CD004659; Lancet 2007;369:1791–8) cannot be trumped by trials or meta-analyses within subgroups. We hope that through this, the scientific community can learn that there comes a time to draw a line, and for aspirin use in pregnancy that time has come.
Conflict of interest
Division of Obstetrics, Gynaecology & Child Health, Nottingham City Hospital, Nottingham, UK
The PREDO Study group members are A Aitokallio-Tallberg, A-M Henry, VK Hiilesmaa, T Karipohja, R Meri, S Sainio, T Saisto, S Suomalainen-König, V-M Ulander, T Vaitilo (Department of Obstetrics and Gynaecology, University of Helsinki and Helsinki University Central Hospital, Helsinki, Finland). L Keski-Nisula (Kuopio University Hospital, Kuopio Finland). E Koistinen, T Walle, R Solja (Northern Karelia Central Hospital, Joensuu, Finland). M Kurkinen (Päijät-Häme Central Hospital, Lahti, Finland). P Staven (Iisalmi Hospital, Iisalmi, Finland). J Uotila (Tampere University Hospital, Tampere, Finland).