A randomised trial of low dose folic acid to prevent neural tube defects. The Irish Vitamin Study Group.

Randomized, block-randomized, stratified trial in women with a previous pregnancy affected by a neural tube defect who were planning or entering a subsequent pregnancy at 12 Irish hospitals. A total of 354 eligible women were randomized to folic acid alone, multivitamins without folic acid, or folic acid plus multivitamins; 281 had completed a pregnancy and 257 women had a first trial pregnancy outcome with neural tube defect status ascertainable.

Women were assigned to one of three oral tablet regimens taken three times daily for at least two months before conception and until the third missed period: folic acid alone, multivitamins without folic acid, or folic acid plus multivitamins. The folic acid groups received 0.36 mg/day total folic acid (0.12 mg/tablet); the multivitamin groups received vitamin A 4000 IU, calciferol 400 IU, thiamine hydrochloride 1.5 mg, riboflavine 1.5 mg, pyridoxine hydrochloride 1 mg, nicotinamide 15 mg, ascorbic acid 40 mg, calcium phosphate 480 mg, and ferrous sulphate 252 mg daily.

The folic acid-containing regimens showed the most favorable pattern, with no neural tube defect recurrences reported in the folic acid groups. Among informative pregnancies, there was no recurrence in the 85 infants/fetuses in the folic acid-only group, while there was 1 recurrence in the 89 infants/fetuses in the multivitamin-only group; the report also states no recurrence among the 172 infants/fetuses in the two folic acid groups combined. Compared with non-randomized controls, the folic acid groups had a statistically significant lower recurrence rate, but the authors cautioned about the validity of that comparison. Overall, the trial suggested that folic acid may be protective, possibly at a lower dose than in the MRC trial, but the evidence was not definitive.

Small numbers of outcome events limited precision and made between-group comparisons unstable. A non-randomized control group was used for some comparisons, which weakens causal inference. Not all randomized women had an ascertainable pregnancy outcome, and the study relied on recurrence outcomes from a specific high-risk population, limiting generalizability.