Aspirin and prevention of vascular complications: there are still indications

Three articles in this first issue of the year 2000 1–3 raise, directly or indirectly, the question of continuing interest in prescribing aspirin to some patients during pregnancy. This idea, like so many others, followed the classic stages of development: first, enthusiasm, second doubt and third a return to reason. We are currently in the second of these phases, and it is clearly time to review the many and contradictory studies that have been published on this subject. We note immediately that the first articles 4–6 recommending this treatment for preventing pre-eclampsia, intrauterine growth restriction (IUGR), or both, always stressed that their results concerned only patients presenting with a specific indication (tested in each of the articles) and that these results could not be extrapolated to broader indications, such as primiparity. Subsequent trials which were generally well conducted, despite some flaws, tested aspirin for these broader indications 7–9 and reported negative results that have subsequently thrown discredit both on the use of aspirin in general and on its use for some specific indications. Although some of the latter are admittedly rare, the results for them have not been invalidated. The argument, now used by detractors of aspirin, about the size of the trials with negative conclusions, compared with those of the trials reaching positive results, is in fact consistent with these earlier reports. Those rejecting the use of aspirin 10 report ‘anecdotally, aspirin continues to be used for women at risk of early onset pre-eclampsia. Although it seems unlikely that this confers any benefit, at least we can reassure these women that it is probably doing no harm’. Such authors rely on moderation after the legitimate and interesting reaction of readers 11 who do not seem surprised over the claim that aspirin lacks efficacy in preventing pre-eclampsia in high-risk populations 12 after having established1 that this same product significantly reduces the pre-eclampsia rate in populations at lower risk 13–16 . Similarly, these analyses ignore some results from the CLASP study 7 which suggested that the severity of disorders played a role in the efficacy of aspirin and that prematurity was significantly decreased among patients receiving it. In the next few months the meta-analyses prepared by ‘obstetric epidemiologists’ threaten to bury still more of the positive results found in well-selected small populations from explanatory studies under truckloads of so-called ‘pragmatic trials’. In fact, this passionate, and sometimes surreal, debate is mixing a variety of responses to questions that are no less varied. To study objectively what has been learnt about aspirin and pregnancy we must analyse each trial according to at least three of the variables or characteristics likely to have an important influence on the result, namely: 1. The indication and its severity. For example, one single previous IUGR or two previous consecutive pregnancies with IUGR, late pre-eclampsia or severe early pre-eclampsia. 2. The time at which treatment began, in particular the latest possible time from preconception to 32 weeks of gestation. 3. The dose of aspirin, varying from 50 mg to 150 mg per day. We shall return to this issue later. Without reviewing in detail the above factors in each of the trials published, we can note the following: 1. Further study of aspirin, especially combined with heparin, in antiphospholipid disorders is not warranted17,18. 2. The use of aspirin, often before conception, is recommended, associated with corticosteroids and heparin in some cases17,18. 3. Aspirin does not appear useful in cases of chronic hypertension or diabetes. However, the results from twin pregnancies lack consensus and therefore perhaps deserve further study, especially in the effects of early or preconception treatment. This is especially true for cases using assisted-reproduction technology as some trials show that aspirin is useful in cases involving in vitro fertilization19 . We can therefore infer that aspirin may play a favorable role in implantation which is, as we know, related to the genesis of pre-eclampsia and some cases of IUGR. 4. The efficacy of aspirin is directly linked to the type of indication. This has been recently demonstrated by our team 20. The preventative effect was significantly better when the indication was a previous early severe pre-eclampsia or severe IUGR, defined as < 3rd centile. In the EPREDA trial 14 we even observed a stratification of preventative efficacy according to whether the indication involved one or more than one consecutive past incident. This indication, which we continue to think is the most important, is only rarely considered in large trials. It is logically related to the question of how early treatment starts. Only indications based on individual history allow so early a treatment. The importance of gestational age at the commencement of treatment appears clear when we remember that implantation is essentially completed by 14–18 weeks 21. Treatment begun after this time, and sometimes as late as 32 weeks of gestation—long after the onset of the most severe forms of vascular complications—is thus doomed to partial action, if any at all, after 22 weeks of gestation. Sullivan et al. 22 observed that patients in whom treatment was initiated in the first trimester of pregnancy, at 13 weeks, had significantly better outcomes than those in whom treatment was initiated in the second trimester. The importance of gestational age of less that 17 weeks was observed in our retrospective work 20 as well as in the meta-analysis by LEITICH 23. The late onset of treatment reported in several publications is probably the, or at least a, reason for the negative or only partially effective results observed when the indication for treatment was an abnormal uterine Doppler examination. It is, moreover, probable that the type of Doppler anomaly—of high resistance, notching or both—is also important. Obviously indications such as abnormal umbilical Doppler are also observed too late to be useful. In a prospective trial McCowan et al.24 did not observe an increase in birth weight in pregnancies where the fetus had an abnormal umbilical Doppler and was thought to be small for gestational age. However, the average gestational age at recruitment was 31 weeks 6 days in the aspirin-treated group and 32 weeks 1 day in the placebo group. The article published by Harrington et al.1 in this issue clearly shows that aspirin improved outcome by reducing the overall incidence of complications associated with utero–placental insufficiency in cases of abnormal uterine artery Doppler screenings. Moreover, this study should be repeated with aspirin prescribed for earlier Doppler anomalies and/or in cases of bilateral uterine notching as this group has reported that Doppler findings as early as 14–16 weeks are predictive of placental insufficiency 25. Coleman et al.2 also observe in this issue that Doppler findings were predictive, but dismiss the use of aspirin with an off-handedness incompatible with a real study of all the data. ‘One-third of women in this study received low-dose aspirin from early in pregnancy. This study was not designed to determine whether aspirin treatment could improve pregnancy outcome in high-risk women. However, no difference in the rates of pre-eclampsia of SGA babies was seen between women treated and those not treated with aspirin’2. This conclusion does not take into account other publications studying uterine Doppler and aspirin26,27. It is probable that this product, although safe, both generally 28 and for the ductus arteriosus 3 will remain ineffective if it continues to be prescribed only after abnormal Doppler results obtained at 22 weeks and 24 weeks of gestation. Two French trials 29 currently in press also reported negative results when first prescribed at so late a gestational age. The article by Clarke et al. is especially interesting in its use of a slow-release form of aspirin 30 and opens up the possibility of also combining aspirin with other products. This study also used a 100-mg dose of aspirin leading us to our third comment—the importance of dose. This is a significant, even determinant, factor in several multiparameter analyses. The importance of dose, evoked in a communication by Wang et al.31 was also found by Dumont et al.20. Wang and colleagues showed that a dose of aspirin of < 100 mg might be inadequate to inhibit placental thromboxane. The success of the treatment depends, as Huath and co-workers showed 32, on effective thromboxane inhibition. Where inhibition was ≥ 50% the following results were observed: pre-eclampsia 1.9 vs. 5.7 (P = 0.0016), IUGR 2.9 vs. 7% (P = 0.023). We showed 20 that, in patients with a positive effect from treatment, the increase in bleeding time was significantly higher (more than 2 min increase before and after treatment) than among the patients with no effect. Furthermore, putting aside weight and probable personal sensitivity, it is likely that an aspirin dose of 50 or 60 mg is totally ineffective against platelet aggregation in some patients. One issue therefore seems certain to us. We will not resolve this problem by recruiting thousands more women with vaguely-defined characteristics at ‘between 12 and 32 weeks’ in a new mega-trial with 60 mg of aspirin and doubtful endpoints, all labelled as pre-eclampsia. It is time instead for very focused and explanatory studies, even if they are disdainfully ‘small’ at Oxford or in Memphis. Such trials might allow us to discover not who is right but rather who aspirin may or may not help. Moreover, the combination of aspirin with other products, such as Calciparine, Ketanserin, calcium, corticoids or nitric oxide donors, deserves more study. Finally another modality also merits careful study—the prescription of aspirin before conception. We are currently applying this regime intuitively to patients who have not had successful pregnancy outcomes, despite early aspirin treatment. In conclusion, aspirin remains appropriate for certain indications. Its use in new modalities, either prior to conception or in combination with other products, should be explored. Contributors:Ph. MervielO. KadochA. DumontN. BerkaneM. BeaufilsG. BreartM. Uzan