Key Takeaways
Extended Takeaways
- This analysis drew from a very large EHR cohort of 894,986 women, but medication-exposed groups were much smaller (3,465 prescribed stimulants and 575 prescribed nonstimulants), so clinicians should read these findings as observational signal-detection evidence rather than causal safety estimates.
- Gestational diabetes was a notable exception among stimulant-exposed pregnancies: stimulant use showed no significant effect versus no medication, whereas nonstimulant use was associated with increased risk relative to no medication and also higher risk than stimulant use.
- Placental outcomes may help differentiate treatment discussions, because stimulant exposure was associated with higher odds of placenta previa than both no medication and nonstimulant exposure, while stimulant use was associated with lower odds of placental abruption than both comparators.
- Large for gestational age did not follow the expected pattern of stimulant-related fetal growth restriction: stimulant use was associated with increased risk versus no medication and higher odds than nonstimulant use, while nonstimulant medication was associated with lower odds of large for gestational age relative to no medication.
- The nonstimulant group appeared clinically more complex at baseline, with particularly high rates of mood disorders and problematic substance use and only 14.8% having a documented ADHD diagnosis, which may signal prescribing for heterogeneous indications and complicate direct interpretation of nonstimulant risk.
- Exposure classification required 2 or more EHR medication records during pregnancy, a design choice intended to capture repeated documentation of treatment, but it does not establish adherence, continuous exposure, or specific formulation risk; this matters because 3,039 (87.7%) of the stimulant group had more than 1 stimulant medication entry recorded during pregnancy.
