Quantified proteinuria predicts how long you have, and how much maternal trouble is coming

Design

secondary analysis of two randomised trials; parametric time-to-event and Poisson regression

Population

291 women expectantly managed for pre-eclampsia diagnosed at 26 to 32 weeks, Cape Town

Primary outcome

pregnancy latency and composite adverse maternal or perinatal outcomes by 24-hour proteinuria

Effect

median latency 15.3 vs 10.0 vs 4.9 days across <3 g, 3–5 g and ≥5 g; maternal composite 8.8% vs 16.1% vs 27.9%

A secondary analysis of two randomised trials at Tygerberg Hospital, Cape Town, looked at 291 women managed expectantly for pre-eclampsia diagnosed between 26 and 32 weeks. The question was whether the 24-hour protein figure — often collected, rarely used for anything beyond diagnosis — predicts anything useful.

It predicts latency. Median prolongation was 15.3 days (IQR 4.3 to 25.6) below 3 g per 24 hours, 10.0 days (5.0 to 19.6) between 3 and 5 g, and 4.9 days (3.8 to 11.7) at 5 g or above. Composite adverse maternal outcomes rose in step: 8.8 per cent, 16.1 per cent and 27.9 per cent across the same bands. There was no association with composite perinatal outcomes. The sFlt-1/PlGF ratio predicted latency somewhat better (AUC 0.78 vs 0.67) but maternal events about the same (0.67 vs 0.64), and both were poor for perinatal outcomes (0.61 and 0.56).

That last comparison is the practical point. Where angiogenic marker testing is unavailable or unaffordable — which is most of India outside large private centres — a quantified 24-hour protein gives comparable information about maternal risk. It changes what you tell a woman at 28 weeks about how long the steroid window and the transfer plan have to work with, and it is available tonight.

• Quantify proteinuria at diagnosis rather than stopping at 'significant'
• Use the 3 g and 5 g bands to frame the latency conversation with the woman and her family
• At 5 g or more, plan transfer and steroid timing around days, not weeks
• Do not use the protein figure to predict neonatal outcome — it did not track
• Where sFlt-1/PlGF is unavailable, record that quantified protein is carrying the same maternal prognostic load

Why it matters

It gives units without angiogenic markers a maternal prognostic tool they already have in the lab book.

Don't overread it

This was a secondary analysis at one South African centre, and the association with perinatal outcomes was absent.

The statistics, in plain English

AUCs of 0.67 and 0.78 mean modest and moderate discrimination — better than guessing, nowhere near a test you would act on alone. Interquartile ranges this wide (4.3 to 25.6 days) mean the bands describe a population, not an individual prediction: a woman below 3 g can still deliver in four days.

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.