- Top of page
Objective To determine whether prophylactic, low dose controlled-release aspirin improves outcome for pregnant women and their babies in Barbados.
Design Randomised placebo-controlled trial.
Setting The Queen Elizabeth Hospital, Barbados.
Population All women attending antenatal clinics between 12 and 32 weeks of gestation were eligible, if without specific contraindications to aspirin and unlikely to deliver immediately.
Methods Randomisation was computer-generated in the antenatal clinic; 1822 women were allocated to receive 75 mg controlled-release aspirin and 1825 matching placebo.
Main outcome measures Proteinuric pre-eclampsia, other pregnancy-induced hypertension, pregnancy duration, birthweight, stillbirths and neonatal deaths, major neonatal events.
Results All but three women from each group were followed up successfully. Forty-four percent were primigravid, and 8% had previous obstetric complications. There were no significant differences between the allocated treatment groups in the incidence of proteinuric pre-eclampsia (40 [2.2%] of those allocated aspirin, compared with 46 [2.5%] allocated placebo), of preterm delivery (255 [14.0%] vs 270 [14.8%]), of birthweight < 1500 g (32 [1.7%] vs 33 [1.8%]) or of stillbirth and neonatal death (44 [2.4%] vs 38 [2.1%]). Aspirin was not associated with excess risk of maternal orfetal bleeding.
Conclusions The results of this study in Barbados do not support the routine use of low dose aspirin for prevention of pre-eclampsia or its complications, confirming results of previous large trials in other settings.
- Top of page
Hypertensive disorders of pregnancy are associated with substantial maternal and perinatal morbidity and mortality in the Caribbean1–4. These conditions continue to be common, complicating an estimated 10–15% of deliveries, even among women of Caribbean origin who migrate elsewhere5. In Barbados, where hypertensive disorders account for at least 25% of hospital antenatal bed occupancy, perinatal mortality had declined to around 30/1000 total births in the 1980s and by 1990 was 22/10006. Many of these perinatal deaths are related to maternal hypertension, intrauterine growth retardation, premature delivery and low birthweight. Recent evidence from Jamaica shows an inverse relation between birthweight and blood pressure later in life7 similar to that shown among adults in Britain8. Thus effective intervention against pre-eclampsia and smallness for gestational age might reduce not only perinatal morbidity and mortality but also the high prevalence of later adult hypertension which is so common throughout the Caribbean and in adults of Caribbean origin elsewhere9–11.
The evidence for platelet activation early in protein uric pre-eclampsia and smallness for gestational age has been extensively reviewed elsewhere12–14, including evidence as far back as 1958 from the Caribbean15. Despite promising results from early small trials of antiplatelet therapy (usually aspirin), later larger trials have failed to confirm any substantial benefit16–21 Most of those trials were conducted largely in developed countries, where the general risk of pre-eclampsia and low birthweight is much less than in the Caribbean. In Barbados, with a population of 263,872 in the 1990 census, > 90% of whom are of African descent, the Queen Elizabeth Hospital is responsible for some 4000 of the island's total 4500 annual births. This setting provided an opportunity to conduct an intervention trial in a setting where almost all pregnant women in a generally high risk population would be eligible.
The antiplatelet effect of aspirin is thought to derive from inhibition of platelet production of thromboxane A, a potent vasoconstrictor and platelet agonist. However, coincidental inhibition of endothelial production of prostacyclin, with opposite effects on platelet function and vascular tone to those of thromboxane A, may limit the efficacy of conventional aspirin. Moreover, systemic levels of aspirin in pregnant women could expose the fetus and newborn to an increased risk of bleeding. Aspirin is subject to extensive first pass metabolism in the liver; a controlled-release 75 mg formulation however has been shown to inhibit platelet thromboxane A, while preserving prostacyclin production and without aspirin reaching the maternal systemic or fetal circulation22. This preparation was chosen for the Barbados Low dose Aspirin Study in Pregnancy (BLASP) as it might improve the imbalance between thromboxane A, and prostacyclin reported in pre eclampsia with less risk of any adverse effects from fetal exposure.
- Top of page
Women were recruited into the trial at the Queen Elizabeth Hospital, Barbados, between July 1992 and July 1994. Prior to beginning recruitment, the protocol was approved by the National Medical Ethics Committee of Barbados. All women between 12 and 32 weeks of gestation without contraindications were eligible for entry into the trial. Contraindications were an increased risk of bleeding, known allergy to aspirin, high likelihood of immediate delivery or previous placental abruption.
All women attending the Queen Elizabeth Hospital for a booking antenatal visit were given an information leaflet explaining the study. Literacy is over 98%, and written informed consent to participate in the trial was obtained, baseline details were recorded on an entry form and entered on a computer in the clinic. Blood pressure (diastolic phase V) was measured by trained midwives using an appropriate cuff size (large for arms ≥ 33 cm in diameter). The clinic computer then randomly allocated the treatment which was dispensed by the study pharmacist. All staff in Barbados were blind to the treatment allocation, the randomisation code being kept in Oxford.
Women were assigned to either 75 mg controlled release aspirin or a matching placebo. Each treatment pack contained seven one-month calendar sheets. Women were instructed to take one tablet daily until delivery, unless advised otherwise by her doctor, and to avoid any other aspirin-containing preparations.
Most women continued their antenatal care at parish polyclinics or their general practice, only returning to the hospital clinic between 32 and 34 weeks. A simple single page follow up form was completed by trial midwives at discharge after delivery, or at six weeks if either the woman or her infant were still in hospital. As well as collecting information about the outcome measures, women were asked the last known date of taking the trial tablets.
The main pre-specified outcomes were: proteinuric pre eclampsia, estimated duration of pregnancy; birthweight; stillbirths (fetal loss ≥ 24 weeks of gestation) and infant deaths before hospital discharge; number of days in the special care nursery; number of days from birth to discharge; and bleeding problems of the newborn. Other maternal events, such as use of antihypertensive and anticonvulsant drugs, antepartum and postpartum haemorrhage and caesarean section, were also considered in subsidiary comparisons. Severity of hypertension was defined as 1. nonproteinuric pre-eclampsia: initial diastolic blood pressure < 90 mmHg with an increase of ≥ 25 mmHg to ≥ 90 mmHg, or an initial pressure ≥ 90 mmHg with ≥ 15 mmHg increase, irrespective of antihypertensive agent use; 2. proteinuric pre-eclampsia: blood pressure changes as above, with more than trace proteinuria; 3. nonpre-eclampsia hypertension: women with diastolic pressure 90 mmHg not included in either (1) or (2); and 4. unclassifiable: persistent proteinuria pre-dating pregnancy. Preterm delivery was defined, as in the Collaborative Low dose Aspirin Study (CLASP)17, as occurring before 37 weeks of estimated gestation. Birthweight centiles for Barbados were not available, and so smallness for gestational age could not be defined; instead, birthweight < 2500 g and < 1500 g was used.
Based on previous data in Barbados, the expected rate of pre-eclampsia was 10–12% and of perinatal mortality was 30 deaths per 1000 births at the time of planning the study. It was estimated that a trial of 3000 to 4000 women would have a 95% probability of detecting at P < 0.02 a decrease of at least one-quarter in that rate of proteinuric pre-eclampsia, as well as an increase of 100 g in mean birthweight or one day in mean gestational duration, while double that number would allow detection of a 30% to 40% decrease in perinatal mortality from previously observed rates with 70% power. The analyses were to be by intention-to-treat for all women randomised appropriately. Two confidential interim analyses were provided to an independent data monitoring committee, but no reason to stop or alter the trial prematurely emerged.
- Top of page
Some 20% of eligible women did not agree to participate and 3697 women were randomised during a two year period. Of those randomised, 42 women were allocated packs which had been labelled in error with two different pack numbers so that their contents were unknown, and these women have been excluded from all analyses. Eight women were later found not to be pregnant and were also excluded. Of the remaining 3647 women, 1822 were allocated aspirin and 1825 placebo, with follow up available for 1819 and 1822, respectively (99.8% complete). Characteristics of the women at entry to the trial were well balanced between the groups (Table 1). Overall, 24% were younger than 20 years of age, 53% were randomised before 20 weeks of gestation, and 44% of women were primigravid. Information on compliance was available for 3268 randomised women (90%): 1361 (42%) took their tablets for 1 95% of the time between randomisation and delivery, and a further 413 (13%) took them for between 80% and 94% of the time, but 226 women (7%) never started their tablets.
Table 1. Characteristics of the women at trial entry. Values are given as n (%).
| ||Allocated group|
| ||Aspirin (n= 1822)||Placebo (n= 1825)|
|Woman's age (years)|
| <20||445 (24)||439 (24)|
| 20–29||1005 (55)||1011 (55)|
| 30–39||361 (20)||361 (20)|
| ≥40||11 (1)||14(1)|
|Gestational age (completed weeks)|
| <20||961 (53)||952 (52)|
| 20–28||712 (39)||735 (40)|
| ≥28||149 (8)||138 (8)|
|Diastolic blood pressure (mmHg)|
| <90||1810 (99)||1812 (99)|
| 90–109||12(1)||13 (1)|
|Proteinuria (≥ 1+)||48(3)||49(3)|
|Obstetric and medical history|
| Primigravidae||801 (44)||802 (44)|
| Previous problems||147 (8)||135 (7)|
| No previous problems||871 (48)||885 (49)|
| Multiple pregnancy||15 (1)||19(1)|
| Chronic hypertension||8 (0)||8(0)|
| Diabetes||13 (1)||8(0)|
Hypertensive disorders of pregnancy
Proteinuric pre-eclampsia developed in 40 women (2.2%) the allocated aspirin compared with 46 (25%) of those allocated placebo (Table 2). Although this represents a 13% (SD 20) reduction in the odds of developing proteinuric pre-eclampsia, the rates of proteinuric pre-eclampsia were low in both groups and this difference between the groups is not statistically significant, having a 95% confidence interval (CI) ranging from a 43% decrease to a 33% increase (Table 3). Nor was there any evidence that the effect differed between those in whom treatment was started earlier or later during gestation, or between nulliparae or multiparae. There was also no support for the hypothesis generated by the CLASP trial17 that aspirin might reduce early onset pre eclampsia: among women delivered before 32 weeks, pre-eclampsia occurred in 3/74 allocated aspirin and 3/76 allocated placebo. Nonproteinuric pre-eclampsia was reported more commonly, but still there was no significant difference between the treatment groups (1 26 [6.9%] aspirin vs 133 [7.3%] placebo). No differences were found between the groups in the medians of the highest systolic or diastolic blood pressures recorded between randomisation and the onset of labour (120 mmHg and 72 mmHg for both groups). Nor did the use of antihypertensive therapy (4.0% vs 4.8%) or anticonvulsant therapy (2.0%vs 1.7%) after randomisation differ.
Table 2. Effects of aspirin on outcome for the mother. Values are given as n (%) or mean [SD].
| ||Allocated group|
| ||Aspirin (n=1819)||Placebo (n=1822)|
|Pregnancy induced hypertension|
| Proteinuric pre-eclampsia||40 (2.2)||46 (2.5)|
| Nonproteinuric pre-eclampsia||36 (2.0)||39 (2.1)|
| Nonpre-eclamptic hypertension||126 (6.9)||133 (7.3)|
| TOTAL||202 (11.1)||218 (12.0)|
| Pregnancy duration (weeks)|
| <28||45 (2.5)||48 (2.6)|
| 28–31||29 (1.6)||28 (1.5)|
| 32–36||181 (10.0)||194 (10.6)|
| ≥37||1562 (85.9)||1552 (85.2)|
| Not known||2 (1.0)||0 (0.0)|
| Mean [SD]||38.7 [3.4]||38.5 [3.7]|
|Labour and delivery|
| Induced||330 (18.1)||307 (16.8)|
| Pre-labour caesarean section||126 (6.9)||152 (8.3)|
| Caesarean section during labour||120 (6.6)||100 (5.5)|
| Antihypertensive therapy||72 (4.0)||88 (4.8)|
| Anticonvulsant therapy||36 (2.0)||31 (1.7)|
| Placental abruption||9 (0.5)||14 (0.8)|
| Other antepartum bleed||65 (3.6)||76 (4.2)|
| Postpartum bleed ≥ 500 mL||178 (9.8)||175 (9.6)|
| Amount not known Transfusion||173 (9.5)||188 (10.3)|
|Other serious maternal problems|
| Eclampsia||4||0|Table 3. Effects of aspirin on proteinuric pre-eclampsia developing after randomisation. Values are given as n/total (%) ▪= 99% CI; ⋄= 95% CI
Duration of pregnancy
The mean duration of pregnancy was nearly a day and a half longer among women allocated aspirin than among those allocated placebo (38.7 weeks [SD 3.41 with aspirin vs 38.5 weeks [SD 3–71 with placebo), but this difference was not statistically significant. Aspirin did not significantly reduce the risk of delivery before 37 weeks of gestation (14.0% in the group allocated aspirin vs 14.8% allocated placebo; 2P > 0.1; 95% CI ranging from a 22% reduction to a 13% increase) (Table 4). Nor was there any evidence that the effect on preterm delivery was different among women entered earlier or later in gestation or in those who were nulliparous or multiparous.
Table 4. Effect of aspirin on pretenn delivery. Values are given as n/total (%); key as for Table 3.
Other maternal outcomes
Antenatal hospital admission was common, but did not differ between the allocated treatment groups (609 [33%] for women allocated aspirin, compared with 610 [33%] for those allocated placebo). Relatively few women stayed in hospital for five days or more (101 women [5.5%] allocated aspirin, compared with 93 [5.1%] of those allocated placebo). There were no significant differences between the groups in induction of labour or caesarean section either before or during labour, nor were there any significant differences in placental abruption, other antepartum haemorrhage, postpartum haemorrhage or transfusion (Table 2). One woman died and four developed eclampsia, all in the aspirin group.
The mean birthweight was not significantly different between the groups (3098 g [SD 5381 with aspirin vs3086 g [SD 5411 with placebo) (Table 5). Nor was the distribution of birthweights significantly different between the groups, with birthweight < 1500 g in 32 infants (1.7%) born to women allocated aspirin, compared with 33 (1.8%) of those allocated placebo.
Table 5. Effects of aspirin on outcome for the baby. Values are given as n, n (%) and mean [SD].
| ||Allocated group|
| ||Aspirin (n= 1834)||Placebo (n= 1841)|
| 1500–2499||131 (7)||152 (8)|
| Not known||12 (1)||12(1)|
| Mean [SD]||3098 ||3086 |
| Loss < 24 weeks||22 (1.2)||33 (14)|
| Stillbirth224 weeks||29 (1.6)||27 (1.5)|
| Neonatal death||14 (0.8)||1I (0.6)|
| Infant death||1 (0.0)||0 (0.0)|
|For all live births||1783||1781|
| Admission to special care nursery||272 (15.3)||293 (163)|
| Not known||0(0)||2 (0.1)|
| Bleeding problems||9 (0.5)||10* (0.6)|
Stillbirths and infant deaths before discharge
Fetal loss occurred before 24 weeks of gestation in 22 women (1.2%) allocated aspirin, compared with 33 (1.8%) allocated placebo (Table 5)-an absolute reduction of 0.6%, with a 99% confidence interval of a 1.6% to a 0.4% increase. There were 44 deaths (2.4%)-29 stillbirths, 14 neonatal deaths and one infant death among the women allocated aspirin, compared with 38 deaths (2.1%)-27 stillbirths and 11 neonatal deaths- among those allocated placebo. This overall increase of 17% in the odds of the infant dying after 24 weeks of gestation with allocation to aspirin was not statistically significant, with a 95% confidence interval ranging from a 85% increase to a 26% reduction (Table 6). There was no evidence of any difference in subgroups of gestational age and parity. Nor was there any difference in the number of such deaths that were associated with pre-eclampsia or smallness for gestational age (8 aspirin versus 9 placebo).
Table 6. Effect of aspirin on death of the baby after 24 weeks of gestation (including stillbirths > 24 weeks, neonatal deaths and infant deaths). Values are given as n/total (%); key as for Table 3.
Other outcomes for the baby
There were no significant differences between the groups in admission to the special care nursery or in bleeding problems (Table 5). Most liveborn infants were discharged from the hospital within five days of delivery (1623 [91.1%] aspirin versus 1608 [90.4%] placebo), and there were no significant differences between the allocated groups in any measure of duration of stay.
- Top of page
This trial recruited about 60% of the women who gave birth at the Queen Elizabeth Hospital over the two-year period and demonstrates the feasibility of conducting large simple trials in this type of setting. Nevertheless, recruitment was lower than anticipated and factors contributing to this probably included higher than expected proportions of women either booking after 32 weeks of gestation or arriving for delivery without any antenatal care. In addition, despite the efforts of two dedicated midwives, pressure of time in busy clinics meant that it was not possible to enroll all eligible women. The perinatal and neonatal mortality rates were close to those expected prior to the trial, but the incidence of hypertensive disorders of pregnancy, and in particular protein uric pre-eclampsia, was lower than expected. This may reflect the use of more rigorous definitions of these disorders within the trial than in previous hospital statistics, most notably, as in other similar the exclusion of oedema because of its highly variable determination by different observers. Birthweight centiles were not previously available for Barbados, and so the data collected in BLASP will be useful for defining size for gestational age in future studies in this community. The mean birthweights found in BLASP where women were predominantly of African-Caribbean origin are similar to those recently reported from the United States for US-born African-Americans (at 3089 g) but lower than those of both US-born whites (3446 g) and of infants of African-born black women in the United States, who were of higher socioeconomic status (333 g)23
In BLASP compliance was lower than in previous large-scale studies. For example, only 55% of women reported taking their tablets for more than 80% of the time between randomisation and delivery, compared with 88% of those in CLASP17. This probably reflects the recruitment of a general population, rather than selected high risk women who are likely to be more motivated to comply with treatment24–26. Such factors may account for some of the observed lack of effectiveness of aspirin in BLASP; however, as it was designed as a pragmatic trial, the results represent the likely out- come in routine care and real life.
Previously data were available for over 16,000 pregnant women entered into randomised trials evaluating aspirin and other antiplatelet agents21, and this trial brings the total to almost 20,000. It had been hoped that the controlled-release formulation of aspirin used in BLASP might be more effective in pre-eclampsia than other antiplatelet regimens. However, the results of BLASP are entirely consistent with the overview of previous trials. Overall, there appear to be reductions in proteinuric pre-eclampsia of about one-fifth (i.e. an incidence of 5.0% among all women allocated aspirin compared with 6.3% among all controls) and of about one-tenth in the incidence of preterm deliveries (17.6%vs 19.2%) and of low birthweight (6.8%vs 75%). This status (3333 g)23
In BLASP compliance was lower than in previous large-scale studies. For example, only 55% of women reported taking their tablets for more than 80% of the time between randomisation and delivery, compared with 88% of those in CLASP. This probably reflects the recruitment of a general population, rather than selected high risk women who are likely to be more motivated to comply with treatment24–26. Such factors may account for some of the observed lack of effectiveness of aspirin in BLASP; however, as it was designed as a pragmatic trial, the results represent the likely outcome in routine care and real life.
Previously data were available for over 16,000 pregnant women entered into randomised trials evaluating aspirin and other antiplatelet agents21, and this trial brings the total to almost 20,000. It had been hoped that the controlled-release formulation of aspirin used in BLASP might be more effective in pre-eclampsia than other antiplatelet regimens. However, the results of BLASP are entirely consistent with the overview of previous trials. Overall, there appear to be reductions in proteinuric pre-eclampsia of about one-fifth (i.e. an incidence of 5.0% among all women allocated aspirin compared with 6.3% among all controls) and of about one-tenth in the incidence of preterm deliveries (17.6% vs 19.2%) and of low birthweight (6.8% vs 7.5%). This does not appear to be reflected in any improvement in stillbirths and neonatal deaths. Moreover, the hypothesis generated by CLASP that aspirin might be particularly effective at preventing early onset pre-eclampsia is not supported by BLASP, but the number of such events was small.
The purpose of the controlled-release low dose aspirin regimen was to avoid aspirin reaching the fetal circulation, so preserving fetal platelet function and minimising the risk of fetal or neonatal bleeding. The BLASP results indicate that this controlled-release 75 mg daily aspirin regimen was safe.