The Cord Tissue Knew the Dose. It Didn't Know the Withdrawal.

A pregnant patient on methadone asks a reasonable question: if my dose is lower, will my baby withdraw less? The instinct behind it is old and widespread, and clinicians have at times acted on it, trimming a stable dose in the third trimester to protect a newborn who has not arrived yet.

A cohort study published Oct. 29 in JAMA Network Open tested the premise and could not find it. Among 92 pregnancies and 94 newborns enrolled across sites in Maryland, Connecticut and Missouri between July 2016 and December 2025, neither the mother's medication dose nor the opioid concentration measured in umbilical cord tissue at delivery predicted how severe the newborn's withdrawal would be.

Severe neonatal opioid withdrawal syndrome — needing pharmacologic treatment, elevated Finnegan scores, or failure of the Eat-Sleep-Console assessment — occurred in 41 of the 94 neonates, 43.6 percent. Methadone carried more than twice the rate of buprenorphine: 33 of 53 methadone-exposed newborns, 55.6 percent, against 11 of 40 buprenorphine-exposed newborns, 27.5 percent. The gap held after adjustment for gestational age, maternal smoking and study site, with an adjusted odds ratio of 3.42.

What the cord tissue measured, and what it didn't

Cord tissue is an appealing biomarker because it accumulates exposure across the third trimester rather than capturing a single day, and the team quantified methadone, buprenorphine and their metabolites by mass spectrometry at a certified laboratory.

Maternal methadone dose tracked cord methadone closely, with a Spearman correlation of 0.64. Buprenorphine dose did not reach significance against cord buprenorphine, at 0.30 with a p value of 0.08.

The cord concentrations themselves were statistically indistinguishable between newborns who developed severe withdrawal and those who did not: p = 0.42 for methadone, p = 0.84 for buprenorphine. Median maternal doses were likewise similar across outcome groups — 127.5 against 120 milligrams for methadone, 16 against 16 for buprenorphine. Peak Finnegan scores showed no meaningful correlation with either dose or cord level, and a sensitivity analysis restricted to the 22 infants who actually received pharmacologic treatment returned the same answer.

Why a negative result matters here

The clinical temptation the study addresses is specific. Methadone exposure does produce more withdrawal than buprenorphine, and it is intuitive to read that as an exposure-response relationship a lower dose could flatten. The data say the between-drug difference is not a dose difference. The authors point to pharmacologic properties, placental transport and fetal adaptation instead, and write plainly that their findings do not support dose reduction as a strategy — a recommendation landing against a practice that circulated for years without a trial behind it. Dose reduction also carries a cost the study does not measure: a lower dose in a patient whose tolerance has grown is a relapse risk, and relapse in pregnancy carries overdose and obstetric danger that no withdrawal score captures.

Ninety-two pregnancies is small, and medication choice was not randomized — methadone versus buprenorphine is decided on maternal stability, so confounding by indication is built into the comparison. The cohort was 88 percent White, and withdrawal assessment varied by site between Finnegan scoring and Eat-Sleep-Console, though results held across definitions.

For clinicians managing medication-assisted treatment in pregnancy, the message is to keep titrating to maternal stability rather than to a number presumed safe for the nursery. For scale, newborn withdrawal diagnoses nationally rose from 4.0 to 7.3 per 1,000 birth hospitalizations between 2010 and 2017, reaching 53.5 per 1,000 in West Virginia — a rise that ran alongside wider opioid use disorder treatment, not narrower dosing.