In This article

• A large electronic health record study examined 32,152 pregnancies among women who had used stimulant medication before pregnancy, including 8,394 who continued stimulants, 405 who used atomoxetine, and 23,353 who used neither medication during pregnancy. 
• Compared with women who discontinued ADHD medication, stimulant use during pregnancy was associated with modestly higher odds of gestational hypertension, pre-eclampsia/eclampsia, preterm delivery, and miscarriage.
• Stimulant exposure was also associated with lower odds of several outcomes, including fetal growth restriction, cesarean delivery, and induction of labor.
• The atomoxetine group had substantially higher rates of depression, anxiety, bipolar disorder, substance use, smoking, and alcohol use, making comparisons between atomoxetine and stimulants particularly difficult to interpret.
• Because this was an observational study, the findings do not establish that ADHD medications caused the adverse outcomes. Treatment decisions during pregnancy should be individualized, weighing potential medication risks against the consequences of inadequately treated ADHD.

Over the last two decades, there has been a significant increase in stimulant prescriptions among adults in the United States, particularly among women of reproductive age. As a result, clinicians are increasingly seeing women with questions about the use of ADHD medications during pregnancy. While data regarding the risk of congenital malformations with stimulant exposure are largely reassuring, information on maternal and fetal outcomes is more limited and has been somewhat inconsistent.

Studies examining pregnancy outcomes differ considerably in their methodology. With observational studies investigating medication use during pregnancy, outcomes in exposed pregnancies are compared to outcomes in unexposed pregnancies; however, women who elect to continue medication during pregnancy may differ from those who do not use or discontinue treatment in ways that can independently affect pregnancy outcomes. These differences can make it difficult to determine whether an observed association is attributable to the medication itself or to characteristics of the women receiving treatment.

Several recently published studies have examined the relationship between prenatal exposure to ADHD medications and pregnancy outcomes. In this three-part series, we review these studies in detail, with the goal of helping clinicians provide an accurate assessment of the potential risks and benefits of ADHD medication use during pregnancy.

• In Part 1 of this series, we reviewed a large claims-based study by Hasan and colleagues examining pregnancy outcomes according to the timing of stimulant exposure.
• In Part 2, we review a recent study by Tuan and colleagues examining pregnancy outcomes among women who continued stimulant treatment or switched to atomoxetine during pregnancy.
• In Part 3, we will review meta-analyses examining maternal and fetal outcomes in pregnancies exposed to ADHD medications.

Study Design

Tuan and colleagues conducted a retrospective observational study using electronic health record data from the TriNetX Research Network. The database includes more than 80 million patients from 103 participating healthcare organizations, predominantly in the United States. The researchers identified adult women who delivered between 2010 and 2022 and who had received at least two stimulant prescriptions before pregnancy.

An important feature of this study is that all women included in the analysis had received a prescription for stimulant treatment before pregnancy. This allowed the researchers to compare women who continued stimulant treatment during pregnancy with women who discontinued treatment or switched to atomoxetine, rather than comparing stimulant users with a population that had never used stimulants.

The final analysis included 32,152 pregnancies:

• 8,394 women (26.1%) continued prescribed stimulants during pregnancy
• 405 women (1.3%) switched to atomoxetine during pregnancy
• 23,353 women (72.6%) received neither stimulants nor atomoxetine during pregnancy

The stimulant group included women who received at least two stimulant refills during pregnancy and no atomoxetine. The atomoxetine group included women who received at least two atomoxetine prescriptions during pregnancy and no stimulants. Women who received only one prescription for either medication were excluded.

The researchers also calculated the proportion of days covered (PDC) during pregnancy to account for differences in the duration and consistency of medication exposure. The average PDC was 38.7% in the stimulant group and 22.5% in the atomoxetine group.

The following outcomes were assessed using electronic health record data:

• prenatal maternal complications, including gestational diabetes, gestational hypertension, and pre-eclampsia/eclampsia
• placental complications, including placental abruption and placenta previa
• fetal growth restriction and large-for-gestational-age infants
• labor and delivery complications, including cesarean section, instrumental delivery, shoulder dystocia, and induction of labor
• birth-related outcomes, including preterm delivery and spontaneous abortion/miscarriage

Propensity-score matching was used to reduce differences in baseline characteristics between the comparison groups, followed by logistic regression to estimate the likelihood of individual adverse pregnancy outcomes.

Outcomes Associated with Stimulant and Atomoxetine Exposure

While the authors report, “The stimulant group was 1.2–3.1 times more likely to develop prenatal maternal, placental, and birth-related complications than the atomoxetine and control groups,” this is an oversimplification of the findings. To better understand the findings and the clinical relevance of the study, it is essential to examine specific outcomes and comparison groups.

Stimulants vs. Unexposed Controls (n=8,394 vs. 23,353)

In this comparison, women who continued stimulant treatment during pregnancy were compared with women who had a history of stimulant treatment before pregnancy but received neither stimulants nor atomoxetine during pregnancy.

Rather than examining outcomes of all pregnancies, the study focused on women who had received treatment with a stimulant. Because the control group also had a history of stimulant treatment, this type of comparison may help to reduce some of the differences between women with ADHD or other indications for stimulant treatment and women who have never used stimulants. However, women who discontinue medication during pregnancy may still differ in important ways from those who continue treatment.

• Increased risk with stimulants:
  • Gestational hypertension (aOR 1.15, 95% CI 1.08–1.22)
  • Preeclampsia/eclampsia (aOR 1.36, 95% CI 1.27–1.45)
  • Placenta previa (aOR 1.31, 95% CI 1.19–1.45)
  • Preterm delivery (aOR 1.34, 95% CI 1.24–1.46)
  • Spontaneous abortion/miscarriage (aOR 1.48, 95% CI 1.41–1.55)
• Decreased risk with stimulants:
  • Gestational diabetes (aOR 0.90, 95% CI 0.83–0.97)
  • Fetal growth restriction (aOR 0.74, 95% CI 0.64–0.85)
  • Large for gestational age (aOR 0.87, 95% CI 0.78–0.97)
  • Cesarean section (aOR 0.87, 95% CI 0.83–0.92)
  • Induction of labor (aOR 0.58, 95% CI 0.48–0.71)

The increased risks associated with stimulant exposure were generally modest. For example, the odds of pre-eclampsia/eclampsia were approximately 36% higher and the odds of preterm delivery approximately 34% higher among women who continued stimulants compared with those who discontinued treatment.

Stimulants vs. Atomoxetine (n=8,394 vs. 405)

In this comparison, women taking stimulants during pregnancy were compared with women who switched from stimulant treatment before pregnancy to atomoxetine during pregnancy.

• Increased risk with stimulants:
  • Gestational diabetes (aOR 1.28, 95% CI 1.01–1.64)
  • Gestational hypertension (aOR 1.25, 95% CI 1.08–1.45)
  • Placental abruption (aOR 3.43, 95% CI 2.12–5.56)
  • Placenta previa (aOR 2.45, 95% CI 1.81–3.31)
  • Large for gestational age (aOR 1.76, 95% CI 1.29–2.40)
  • Spontaneous abortion/miscarriage (aOR 3.06, 95% CI 2.65–3.52)
• Decreased risk with stimulants:
  • Fetal growth restriction (aOR 0.46, 95% CI 0.39–0.54)
  • Cesarean section (aOR 0.71, 95% CI 0.62–0.81)
  • Instrumental delivery (aOR 0.63, 95% CI 0.47–0.85)
  • Induction of labor (aOR 0.24, 95% CI 0.18–0.30)

The differences between stimulants and atomoxetine appear larger than those observed when stimulants were compared with the control group. However, these findings should not be interpreted as evidence that stimulants are inherently more dangerous than atomoxetine. The atomoxetine group was very small and differed substantially from both the stimulant and control groups in important ways.

Atomoxetine vs. Unexposed Controls (n=405 vs. 23,353)

Compared with women who received neither medication during pregnancy, atomoxetine exposure was associated with increased risk of several outcomes:

• Increased risk with atomoxetine:
  • Fetal growth restriction (aOR 2.29, 95% CI 2.00–2.62)
  • Large for gestational age (aOR 1.94, 95% CI 1.71–2.21)
  • Instrumental delivery (aOR 15.0, 95% CI 11.7–19.3)
  • Induction of labor (aOR 9.49, 95% CI 7.90–11.4)
• Decreased risk with atomoxetine:
  • Shoulder dystocia (aOR 0.63, 95% CI 0.47–0.85)

The researchers found no significant differences between atomoxetine and the control group for other outcomes

How Should We Interpret These Findings?

The most consistent findings in this study were associations between stimulant exposure and increased risk of gestational hypertension, pre-eclampsia/eclampsia, preterm delivery, and miscarriage. The magnitude of these associations was generally modest when stimulants were compared with women who discontinued stimulant medications, with adjusted odds ratios ranging from approximately 1.15 to 1.48.

These findings are consistent with previous research suggesting that stimulant exposure during pregnancy may be associated with increased risk of hypertensive disorders and preterm birth. The biological mechanism is also plausible. Prescription stimulants can increase blood pressure and cause blood vessel constriction, which could potentially affect uteroplacental blood flow.

At the same time, this study cannot establish that stimulant exposure caused these outcomes. Women who continue stimulant treatment during pregnancy may differ from women who discontinue treatment in ways that are difficult to fully capture in electronic health record data.

The comparison with atomoxetine is particularly difficult to interpret. At first glance, the substantially larger odds ratios for several outcomes might suggest that stimulants carry considerably greater risk than atomoxetine. However, the atomoxetine group was small and had markedly different baseline characteristics. Women in the atomoxetine group were more likely to have depression, anxiety, bipolar disorder, and substance use disorders, as well as several medical comorbidities than those in the control and stimulant groups.

Thus, the larger differences observed when stimulants were compared with atomoxetine may reflect differences between the women receiving these medications rather than differences in the medications themselves.

Interestingly, stimulant exposure was associated with lower odds of fetal growth restriction than either atomoxetine exposure or the control group. The authors also found lower rates of cesarean section and induction of labor among women receiving stimulants. It should not be assumed that stimulants decrease risk; however, the findings may reflect differences in maternal characteristics, clinical care, ADHD symptom control, or other factors.

Other Associated Factors Affecting Risk of Adverse Outcomes

One of the major strengths of this study is its large sample size and its focus on women with documented stimulant treatment before pregnancy. However, there are important differences between the comparison groups that raise concerns about residual confounding.

Women who received stimulants during pregnancy were somewhat older than women in the control group. Women in both medication groups also had higher rates of several medical and psychiatric conditions. Most notably, psychiatric comorbidity was common.  The differences were particularly striking in the atomoxetine group. Approximately 64% had a diagnosis of depression, 80% had a diagnosis of anxiety, 41% had bipolar disorder, and 72% had a substance use disorder.

Smoking and alcohol use were also substantially more common among women receiving atomoxetine. Smoking was reported in 62.5% of the atomoxetine group compared with 33.9% of the stimulant group and 31.5% of controls. Alcohol use was reported in 26.4% of the atomoxetine group compared with 8.2% of the stimulant group and 8.8% of controls.

These differences are important because depression, anxiety, substance use, smoking, and other health behaviors may independently influence pregnancy outcomes. In addition, women who choose to continue stimulant treatment may have more severe or impairing ADHD symptoms than women who discontinue treatment.

The researchers used propensity-score methods to account for measured differences between the groups, but these methods cannot control for factors that were not measured or were incompletely captured in the electronic health record. In particular, the study could not fully account for ADHD severity, treatment indications, clinical decision-making, medication dose or formulation, or socioeconomic factors.

Another important limitation is that the study did not provide a detailed measure of medication dose. Although the researchers used the proportion of days covered or PDC to capture differences in treatment duration and consistency, prescription records cannot tell us whether the medication was actually taken as prescribed. Overall exposure rates appear to be relatively low;  the stimulant group had mean PDC of 38.7% (SD = 35.8%) during pregnancy, while the atomoxetine group had a lower PDC (mean = 22.5%, SD = 21.5%).  This raises the question of whether we are actually measuring the effects of exposure to medication or the effects of untreated illness in the mother or other associated factors.

Bottom Line

The current study observed that stimulant exposure during pregnancy was associated with a modestly increased risk of several outcomes, particularly gestational hypertension, pre-eclampsia, preterm delivery, and miscarriage. These findings are consistent with some previous studies but do not establish that stimulant treatment causes these complications.

The findings regarding atomoxetine are more difficult to interpret. Although atomoxetine was associated with increased odds of several fetal growth and labor/delivery outcomes, the atomoxetine group was small and had substantially higher rates of psychiatric and substance use comorbidity than the other comparison groups.

Although studies of in utero exposure to prescription stimulants have suggested associations with certain adverse outcomes, including preterm birth and pre-eclampsia, findings have not been consistent. A recent meta-analysis found no clear evidence of an increased risk of adverse pregnancy outcomes with prescription stimulant use (Li et al, 2020; Jiang et al, 2019).

Overall, the results of this study and recent meta-analyses, do not suggest that stimulant or atomoxetine treatment should automatically be discontinued during pregnancy. Rather, they add to the information that clinicians and patients should consider when making individualized treatment decisions. For women with significant ADHD symptoms, discontinuing treatment may also have consequences, including impaired functioning, increased risk of accidents, difficulty maintaining healthy behaviors and prenatal care, and worsening mood or anxiety symptoms.

When considering ADHD medication during pregnancy, the goal is not to eliminate all potential risk, but to balance the potential risks of medication exposure against the potential risks of inadequately treated ADHD. More research is needed to better understand how medication dose, duration of exposure, ADHD severity, psychiatric comorbidity, and other maternal characteristics influence pregnancy outcomes.

—Ruta Nonacs, MD PhD