A big number about plastic and preterm birth. And a quieter number that matters more.

Early in 2026, a research team put a startling number into print. Roughly 1.97 million preterm births worldwide in 2018 were statistically tied to a common plastic softener called DEHP ^[1]. That is close to 8% of all preterm births on the planet, along with about 74,000 newborn deaths.

The headline is easy to feel. But the number sitting right next to it may be the one worth carrying home.

The replacement chemical looks almost identical on the scoreboard

DEHP is a plasticizer — the ingredient that makes rigid PVC flexible enough to bend, stretch, and package. Because regulators kept flagging it, industry gradually swapped it out for a cousin called DiNP, marketed as the safer alternative.

Then the same team ran the same math on DiNP. The modeled result: about 1.88 million attributable preterm births and 64,000 newborn deaths ^[1].

That is nearly the same number, from the molecule that was supposed to fix the problem.

This is what people in environmental health call “regrettable substitution.” Swap a suspect chemical for a structurally similar one, deploy it everywhere the first one used to be, and end up in roughly the same place. The DiNP line in that results table is the whole story in miniature.

What “attributable” actually means

It is tempting to read “1.97 million preterm births attributable to DEHP” as “DEHP caused 1.97 million preterm births.” Those are not the same sentence.

An attributable-fraction model works like this. Researchers take a measured association between a chemical and an outcome — here, higher phthalate levels in a pregnant person’s urine linked to higher odds of preterm delivery. They combine that association with an estimate of how exposed the world’s populations are. Then they ask: in a counterfactual world with no exposure, how many of these events would not have happened?

It is a policy-facing calculation. It says: if the association is truly causal, and if the exposure estimates are close to right, this is roughly the scale of the harm.

Both of those “ifs” are load-bearing. The uncertainty ranges around the central estimates span roughly a fourfold range for DEHP and closer to a tenfold range for DiNP ^[1]. The headline number is the middle of a wide band, not a census count.

Where the underlying data comes from

Most of the association data feeding the global model was collected in American pregnancy cohorts ^[2]. The largest pooled analysis followed more than 6,000 US pregnancies and linked DEHP breakdown products in maternal urine to roughly 12 to 16% higher odds of preterm birth ^[3]. A follow-up pointed to the first and second trimesters as the most sensitive windows ^[4]. An independent study saw a similar overall pattern but flagged the third trimester as the strongest signal, and also implicated DINCH, yet another replacement plasticizer ^[5]. That kind of disagreement is normal in this field — a real tension, not a scandal.

To go global, modelers layered the US-derived numbers onto biomonitoring data from every region they could find. The math is defensible. But a pregnancy in Karachi or Cairo does not necessarily look like a pregnancy in Boston.

Why the DiNP result is the one to sit with

A systematic review has concluded that DiNP shows hormone-disrupting activity in a similar range to DEHP at today's exposure levels ^[6]. So the modeled preterm burden has a plausible biological reason behind it.

What the near-parity between the DEHP and DiNP numbers really tells us is a story about volume. DiNP is now present nearly everywhere DEHP used to be, and in some product categories, more places. Multiply a slightly lower per-unit hazard by a slightly higher population-level exposure and you can land in roughly the same neighborhood.

Regulators are catching up unevenly. In January 2026, the US EPA finalized risk evaluations for DEHP and four related phthalates, preliminarily determining that they present unreasonable risks to human health ^[9]. DiNP was not in that batch. If the pattern holds, the next “safer” replacements are already moving through the supply chain.

The particle question sitting alongside

Phthalates are the chemicals that leach out of plastic. Micro- and nanoplastics are the plastic itself, broken down. They are related exposures — and the pregnancy literature is starting to look at both together.

Placentas from preterm deliveries tend to carry more micro- and nanoplastic particles than term placentas do ^[7]. Another study found particles in most amniotic fluid samples and reported that higher counts tracked with shorter pregnancies ^[10].

These studies are small and correlational, and the measurement methods are still maturing. They do not prove that particles cause preterm birth. But they raise a question the single-chemical model cannot fully answer: what if the particles are, in part, delivery vehicles for the very phthalates the burden estimate is measuring?

Notes on the science

The 1.97 million and 1.88 million figures are central estimates from a modeling exercise, not counts of documented cases. Uncertainty intervals span roughly four-fold for DEHP and ten-fold for DiNP ^[1]. The exposure-response data comes largely from US cohorts and may not translate cleanly to other regions. And tissue-level particle measurements in placenta and amniotic fluid remain methodologically constrained — small samples, varied detection platforms, inconsistent contamination control. Read those counts as directional, not definitive.

Where Winnow fits in

If you are pregnant, planning to be, or love someone who is, a headline like “1.97 million preterm births” lands hard. It should. The response, though, is largely upstream of any one household — better regulation of the whole class of plasticizers, safer packaging defaults, and honest labeling of what replaces what. Winnow’s work sits alongside that broader response: educational writing about what the science actually shows, and a probiotic formulation shown in laboratory testing to bind micro- and nanoplastics within the gut lumen. If you want the reproductive-health arc laid out more fully, our piece on fertility, pregnancy, and plastic exposure is the right next read.