OBJECTIVE: To determine characteristics and pregnancy outcomes in women with primary adrenal insufficiency (PAI). DESIGN: Retrospective multicentre cohort study. SETTING: Twenty-three maternity units in the UK and Ireland. SAMPLE: Seventy-nine women with PAI who had 101 pregnancies. METHOD: Retrospective chart analysis. MAIN OUTCOME MEASURES: Adrenal crisis, pregnancy outcomes. RESULTS: We obtained data on 101 pregnancies in 79 women with PAI. Most (51, 64.1%) had autoimmune disease, 8 (10.3%) had prior adrenal infarction/surgery/haemorrhage, 2 (2.6%) had congenital adrenal hyperplasia, and 18 (21.3%) were unclassified. 19 (24%) women experienced a crisis during pregnancy (18.8% of pregnancies). One woman died postpartum. Although all women had recorded endocrinology input during pregnancy, steroid emergency cards were only reportedly carried in 40 (39.6%) pregnancies and 9/19 (47.4%) of those with an adrenal crisis in pregnancy. Compared with the pre-pregnancy dose, only 41% of women received an increased hydrocortisone dose in pregnancy. The caesarean section rate was higher than the UK average: 62/97 (63.9%). The preterm birth rate was 21.2% (21/99) and 12.8% (12/94) of neonates had a birthweight < 10th centile. CONCLUSION: Whilst the obstetric outcome of pregnancy with PAI is generally favourable, there are high rates of caesarean birth and prematurity. A high number of women experienced adrenal crisis and further exploration is warranted. Recommendations regarding third trimester increases in hydrocortisone need consideration and potentially strengthening, in light of further evidence. Pregnant women with adrenal insufficiency should carry an NHS steroid warning card; this should be reinforced both by endocrine and obstetric teams because of the increased risk of life-threatening adrenal crisis.