Appraise screening for and aspirin prophylaxis of pre-eclampsia

In one line

Pre-eclampsia prevention is a screen-then-treat problem: identify the woman at risk early, start aspirin 150 mg at night before 16 weeks, and continue to 36 weeks — the single intervention with a real, dose- and timing-dependent effect on preterm pre-eclampsia. The pathophysiology and acute management are in pre-eclampsia basics; this chapter covers getting the prophylaxis right, why the regimen is the regimen, and the arguments on dose, timing, screening method and the WHO calcium recommendation.

Why this is a prevention problem at all

This chapter assumes the two-stage placental model from pre-eclampsia & HELLP basics and the early-vs-late phenotype split developed in Early-onset severe pre-eclampsia. The idea that unifies everything below is this: aspirin acts on Stage 1 (placentation), not Stage 2 (the maternal syndrome) — which is why it works only when given before spiral-artery remodelling is complete (the <16-week window), why it prevents the placental phenotype (preterm pre-eclampsia and its travelling companion, FGR) and not the maternal-constitutional term phenotype, and why it is useless once the disease is established. Every dose, timing and screening argument that follows is downstream of that one mechanistic commitment.

Aetiology / mechanism — why aspirin, why 150 mg, why at night

The prostacyclin–thromboxane story is assumed here. What matters is the pharmacology that justifies the exact regimen, because that is where the guideline disagreements live.

The COX-1 / thromboxane mechanism, at the level that explains the dose. Low-dose aspirin (<300 mg) irreversibly acetylates serine-529 of cyclo-oxygenase-1 (COX-1) in the anucleate platelet. Because the platelet cannot resynthesise the enzyme, its thromboxane-A₂ output is suppressed for the platelet's whole ~10-day lifespan; the vascular endothelium, being nucleated, regenerates COX and keeps making the vasodilator/antiaggregant prostacyclin (PGI₂). The pre-eclamptic placenta sits on the wrong side of a thromboxane-dominant imbalance, so selectively crippling platelet thromboxane while sparing endothelial prostacyclin restores the vasodilator-to-vasoconstrictor balance and, more importantly, improves the deep placentation that is failing in Stage 1. This is why the effect is antiplatelet, not analgesic — and why the relevant dose is the antiplatelet dose, not the anti-inflammatory one. Aspirin must be on board while trophoblast is still invading, which is the entire basis of the <16-week rule.

Why 150 mg and not 75–81 mg — the pharmacodynamic argument. The dose question is not arbitrary tablet-counting; it is about incomplete platelet COX-1 inhibition at the low end. A meaningful minority of pregnant women — disproportionately those with raised BMI and accelerated platelet turnover — show non-optimal platelet inhibition on 75–81 mg, with residual thromboxane production ("biochemical aspirin non-response"). Higher mg and a circadian-aligned dose both push more women across the threshold of full platelet inhibition. The 150-vs-81 mg difference is therefore not "the Americans use a smaller pill" but a pharmacodynamic dose-response one: 150 mg captures responders that 81 mg misses, which is why the trials favour it.

Why at night — chronotherapy, not folklore. Platelet reactivity and the thromboxane:prostacyclin ratio follow a circadian rhythm, and the maternal blood-pressure and angiogenic milieu vary across the 24 hours. Aspirin taken in the evening more completely suppresses the overnight/early-morning thromboxane surge and the nocturnal blood-pressure profile than the same dose taken on waking. The bedtime instruction in the SA NDoH and FIGO regimens is therefore a deliberate chronotherapeutic choice with a mechanistic rationale — worth stating as such rather than as a quirk.

Why it cannot rescue established disease. Once Stage 1 is over and the anti-angiogenic Stage-2 cascade (sFlt-1 scavenging PlGF/VEGF → endothelial dysfunction) is running, an antiplatelet agent acting upstream on placentation has nothing left to modify. This is the mechanistic reason aspirin is prophylaxis, never treatment, and the reason it does little for term (maternal-phenotype) disease where placentation was relatively normal to begin with.

Assessment — who screens positive, and how

Two screening philosophies compete, and they are best contrasted by where each one fails and how the method matches the resource setting, not by the checklists themselves.

1. Risk-factor (checklist) screening — the SA and NICE default. Treat any woman with one high-risk factor OR two or more moderate-risk factors (NICE NG133).

• High risk: prior hypertensive disease in pregnancy, chronic hypertension, chronic kidney disease, type 1/2 diabetes, autoimmune disease (SLE, antiphospholipid syndrome).
• Moderate risk: nulliparity, age ≥40, BMI ≥35 kg/m² at booking, family history of pre-eclampsia, pregnancy interval >10 years, multifetal pregnancy.

The SA national hypertension-in-pregnancy guideline frames the indication slightly differently — it triggers aspirin on any single early-onset risk factor (prior pre-eclampsia, chronic hypertension, multiple gestation, pre-gestational diabetes, BMI >33, APS/SLE, assisted reproduction), reflecting a lower threshold to treat in a high-mortality setting.

The checklist is a deliberately blunt instrument, with three failure modes. (i) It is dichotomous — a woman with one weak moderate factor and a woman with severe chronic kidney disease are treated identically, with no gradation of risk. (ii) It cannot integrate continuous biophysical data (a high MAP, an abnormal uterine-artery waveform) that materially change risk. (iii) It mostly flags the maternal-constitutional factors (age, BMI, parity) that predict term disease, while under-detecting the placental phenotype that drives the dangerous preterm disease aspirin actually prevents — so it is weakest exactly where the stakes are highest. In effect the checklist screens for the wrong half of the disease.

2. First-trimester combined (FMF) screening. At 11–13⁺⁶ weeks, combine maternal factors with mean arterial pressure (MAP), uterine artery pulsatility index (UtA-PI) and serum placental growth factor (PlGF) through a Bayesian (competing-risks) algorithm, treating those with a risk of ≥1 in 100 for preterm pre-eclampsia. This is the FIGO 2019 "best-practice" model.

The competing-risks model's logic runs as follows: it treats gestational age at delivery with pre-eclampsia as a continuously distributed survival outcome and shifts that distribution earlier as the biophysical markers worsen, so it estimates risk specifically for preterm (and early-onset) disease — which is why it outperforms the checklist precisely on the phenotype that matters. Each marker indexes a different facet of placentation: a high MAP captures the maternal cardiovascular substrate, a raised UtA-PI is the Doppler signature of unconverted high-resistance spiral arteries (failed Stage 1), and a low PlGF is the earliest circulating evidence of the anti-angiogenic shift. The decisive evidence is SPREE: at a matched 10% screen-positive rate, the NICE checklist detected only 40.8% of preterm pre-eclampsia versus 82.4% for the full combined test.

FIGO's pragmatic compromise for limited resources is contingent screening. Stage one runs maternal factors + MAP (± UtA-PI/PAPP-A) on everyone to stratify into high-, intermediate- and low-risk bands; PlGF (the expensive reflex assay) is then measured only in the intermediate band to recompute a final risk, sparing the cost of universal PlGF while retaining most of the detection of the full test. This three-tier reflex design is the model that actually fits an SA budget — universal cheap inputs, selective expensive ones. In most SA public-sector units neither PlGF assay nor reliable first-trimester UtA-PI is available, so risk-factor screening is the operative reality, accepting that it under-performs and that the contingent fallback is what would be built if PlGF were funded.

Management — immediate, ongoing, long-term

Immediate — start aspirin correctly