Objective: To evaluate oral pyridoxine in conjunction with standard therapy in women hospitalized for hyperemesis gravidarum (HG). Methods: Patients with HG were randomized at hospitalization to 20 mg oral pyridoxine thrice daily or to placebo. Intravenous rehydration, metoclopramide and oral thiamine were also administered. Metoclopramide and thiamine were continued for 2 weeks after discharge. Rehospitalization for HG in the 2-week study period was ascertained, vomiting was recorded by diary, and nausea was evaluated with a 10-point visual analogue scale (VAS) at enrolment, after hospital discharge and on week 1 and 2 reviews. Results: Ninety-two women were included in the analysis. The rehospitalization rate was 37.5 versus 21.1% (relative risk 1.8, 95% confidence interval 0.9–3.7, p = 0.14) for pyridoxine and placebo, respectively. Daily vomiting episodes after discharge were not different (mean ± SD): 1.9 ± 2.4 versus 1.4 ± 1.1, p = 0.28 during week 1 and 1.4 ± 1.3 versus 1.4 ± 1.6, p = 0.98 during week 2. Repeated-measures analysis of variance showed no difference in the nausea visual analogue scale (p = 0.59). Conclusion: The use of oral pyridoxine in conjunction with metoclopramide during the inpatient stay and during the 2 weeks after hospital discharge for hyperemesis gravidarum did not improve the rehospitalization rate, the vomiting frequency or the nausea score. Larger studies need to be done.