This week: Kathryn Gray told Peter Attia that pregnancy is the hardest physiological test most bodies ever sit, and the result sheet is read decades later. We paired it with the meta-analysis behind the headline number, then with the trial that shows what to do about a bad reading. Plus Rhonda Patrick on how many blueberries count, Elissa Epel on stress as a challenge, and a Cochrane review of nearly 30,000 people trying to quit smoking.

GM.

Seven point four. That is how many times likelier a woman is, on average, to develop type 2 diabetes after gestational diabetes than after a pregnancy with normal blood sugar. Which raises the question this issue answers: what is that nine-month test telling you, and what are you meant to do with the result?

⏱ ~5 min read

Issue 27 argued that cardiovascular risk starts accumulating decades before anyone measures it. This week finds one of the places it gets measured early, and it is a maternity ward. A maternal-fetal medicine specialist explains what pregnancy reveals, and the research behind it shows what to do with the answer.

7.4×
How much likelier type 2 diabetes is, on average, in women who had gestational diabetes than in women whose pregnancies had normal blood sugar.

Bellamy and colleagues pooled 20 cohort studies covering about 675,000 women and roughly 10,900 cases of type 2 diabetes. The result, published in The Lancet in 2009, was a pooled relative risk of 7.43 (95% CI 4.79 to 11.51). The largest single study, with more than 659,000 women, reported 12.6.

What it means: gestational diabetes does not end at delivery. It is a readout of how a body handles a glucose load, taken while insulin resistance is being pushed up on purpose. The wide interval says the studies disagree about the size, and these are averages rather than a forecast for any one woman. The direction is not in dispute, and the useful part is that it arrives years early.

Bellamy et al., The Lancet, 2009 · LSHTM record

Pregnancy is a stress test with a long memory.
Kathryn Gray on what the body's most demanding nine months reveal about the decades after.
Kathryn Gray, M.D., Ph.D. · The Peter Attia Drive #410 · 5 October 2026

Gray is a maternal-fetal medicine specialist at the University of Washington School of Medicine, and she says plainly that her views are her own, not the university's. Women's health has been the thinnest shelf in this newsletter's archive, so this is a gap we are glad to close.

🎙️ Listen → The Peter Attia Drive #410, Kathryn Gray, M.D., Ph.D., October 2026 · Bellamy et al., BMJ, 2007

The bad reading comes with a plan

A 7.4-fold risk invites the question of whether it is destiny. A trial says it is not.

The Diabetes Prevention Program randomised adults at high risk of diabetes to placebo, metformin or an intensive lifestyle programme, whose standard goals were 7% weight loss and 150 minutes of activity a week. Ratner and colleagues analysed the 350 women in it with a history of gestational diabetes: 122 on placebo, 111 on metformin, 117 in the lifestyle arm. Over about three years, lifestyle cut the incidence of diabetes by 53.4% compared with placebo, and metformin by 50.4%.

Two honest limits. Three years is short, so this shows delay rather than a lifetime guarantee. And the authors note that their comparisons between women with and without gestational diabetes were underpowered, so we should not claim the effect was bigger for them, only that it clearly worked.

That is the other half of Gray's stress test. The reading is a flag with a known response attached: ask how often to check fasting glucose or HbA1c, and treat the walking and weight advice as the treatment it is.

Ratner et al., J Clin Endocrinol Metab, 2008

P.S. Hit reply and tell us which of this week's three actually changed something: the glucose check, the cup of blueberries, or the stress reframe. We are voting for the blueberries.

What did you think of today's BrainBite?