Background
Hyperemesis gravidarum is associated with adverse neuropsychiatric and mental health outcomes. We aimed to identify relative risks of incident neuropsychiatric and mental health outcomes during the first year following a hyperemesis gravidarum diagnosis.
Methods
For this retrospective cohort study of anonymised electronic medical records, we used data from the TriNetX network. Women with pregnancy coding and a diagnosis of hyperemesis gravidarum (excessive vomiting in pregnancy [ICD10 code O21]) were recruited to the hyperemesis gravidarum cohort. A propensity-scored matched control cohort was recruited including all women with pregnancy coding without hyperemesis gravidarum. The primary outcome was the cumulative 1-year incidence of 24 neuropsychiatric and mental health outcomes (Wernicke's encephalopathy; refeeding syndrome; any mental, behavioural, or neurodevelopmental disorder; depression; postpartum depression; affective psychoses; puerperal psychosis; non-affective psychoses; bipolar affective disorder (non-psychotic); post-traumatic stress disorder; obsessive compulsive disorder; phobic anxiety disorders; adjustments disorders; other anxiety disorders; eating disorders; substance use disorders; dissociative and conversion disorders; somatoform disorders; suicidal ideation; suicide attempt; intentional self-harm; antidepressants; antipsychotics; elective termination of pregnancy) were compared with propensity-score matched control cohorts.
Findings
Between Jan 1, 2010, and April 11, 2025, 476 857 pregnant women in the TriNetX database were diagnosed with hyperemesis gravidarum across 135 health-care providers in 18 high-income and middle-income countries. Among the women with hyperemesis gravidarum, mean age was 27·0 years (SD 6·1), 243 112 (51·0%) were White, and 113 058 (23·7%) were Black. Significant relative risks were observed for 18 of 24 outcomes. The five greatest relative risks were seen in Wernicke's encephalopathy (2·39 [95% CI 1·25–4·56]), refeeding syndrome (2·57 [2·32–2·84]), postpartum depression (2·70 [2·61–2·79]), eating disorders (2·06 [1·85–2·29]), and antipsychotic prescription (2·23 [2·19–2·26]). Risk of termination of pregnancy was reduced (0·77 [0·73–0·80]). No association was observed for non-affective psychoses, bipolar affective disorder (non-psychotic), intentional self-harm, or suicide attempt.
Interpretation
Our study suggests that although hyperemesis gravidarum is associated with higher risk for poorer mental health outcomes compared with pregnancy without the condition, the relationship between hyperemesis gravidarum severity and psychiatric burden is complex. We propose that the finding of significantly reduced risk of depression in hyperemesis gravidarum with metabolic disturbance compared with the mild form of the disease might reflect differences in levels of input received from health-care services. This finding has implications for service development and provision for a historically underserved population.
