Longer expectant management in preterm hypertensive disease, and what shows up years later

Design

secondary analysis of the prospective nuMoM2b Heart Health Study with adjudicated pregnancy outcomes and multivariable linear regression

Population

142 nulliparous participants with hypertensive disorder of pregnancy diagnosed before 37 weeks, without pregestational hypertension or diabetes

Primary outcome

AHA Life's Essential 8 health factor score (BMI, blood pressure, non-HDL cholesterol, HbA1c) 2-7 years after delivery

Effect

overall score beta -0.26 (95% CI -0.94 to 0.41, P = 0.44); non-HDL score beta -1.36 (-2.50 to -0.21); hs-CRP beta 0.05 log mg/dl (0.001-0.10)

Expectant management in preterm-onset hypertensive disorders of pregnancy buys gestational age for the fetus at some cost to the mother, and the size of that cost beyond the pregnancy has not been well described. This secondary analysis of the nuMoM2b Heart Health Study followed 142 nulliparous participants diagnosed with a hypertensive disorder before 37 weeks, without pre-existing hypertension or diabetes, and measured cardiovascular health 2 to 7 years later. Thirty had short latency of 2 to 7 days; 112 had long latency, median 17.5 days.

On the primary outcome — the American Heart Association Life's Essential 8 health factor score combining body mass index, blood pressure, non-HDL cholesterol and HbA1c — longer latency showed no association (beta -0.26, 95% CI -0.94 to 0.41). Two secondary measures did move: non-HDL scores were worse (beta -1.36, 95% CI -2.50 to -0.21) and high-sensitivity CRP higher (beta 0.05 log mg/dl, 95% CI 0.001 to 0.10). NT-proBNP and the other score components did not differ, and sensitivity analyses showed early-onset and severe subtypes did not explain the pattern.

Read this carefully before it changes anything. With 142 participants and two positive results out of several tested, one of which has a lower confidence bound of 0.001, the multiplicity problem is real and the findings are hypothesis-generating. Latency is also not randomly assigned — women managed expectantly for longer were women whose condition allowed it, which cuts against the finding rather than for it. What it reasonably supports is not shortening latency, but ensuring that every woman with a preterm hypertensive disorder gets structured cardiovascular follow-up afterwards, which most still do not.

• Arrange structured cardiovascular follow-up after any preterm-onset hypertensive disorder of pregnancy
• Do not shorten expectant management on the strength of this — the primary outcome was null and latency was not randomised
• Include a lipid profile in postpartum cardiovascular review, not only blood pressure
• Treat the CRP and non-HDL findings as subclinical markers, not as events
• 142 participants with several outcomes tested — two significant results is what chance alone can produce

The statistics, in plain English

The primary outcome interval (-0.94 to 0.41) comfortably includes zero, so on the measure the study was designed around, nothing was found. The two secondary findings have intervals that only just exclude zero — one bound sits at 0.001 — and when several outcomes are tested, results this marginal are exactly what would appear by chance. A high-sensitivity CRP difference expressed on a log scale is also hard to translate into anything a patient would notice. With 30 participants in the short-latency group, this analysis could only ever have detected a large difference.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.