OBJECTIVE: To describe the prevalence of iron depletion (ID), iron-deficiency anaemia (IDA) and risk of haemoconcentration during pregnancy and at delivery and to assess the influence of initial Fe stores and Fe supplementation on that prevalence. DESIGN: Longitudinal study. SETTING: Hospital Universitari Sant Joan de Reus (Catalonia, Spain). SUBJECTS: Two hundred and eighty-five pregnant women. Serum ferritin and Hb were measured in the first, second and third trimesters and at delivery. Women were classified according to initial Fe stores as ID or no ID (serum ferritin ≥12 μg/l) and according to Fe supplement use as supplemented or non-supplemented. RESULTS: Initial ID was 16·2 %. At delivery, 45·7 % had ID, 13·5 % IDA and 13·3 % had risk of haemoconcentration. Initial ID and non-supplemented groups had significantly higher prevalences of ID and IDA and lower risk of haemoconcentration at delivery than the other groups. In the multiple logistic models, no initial ID and Fe supplementation exerted a protective effect against ID at delivery (adjusted OR = 0·28; 95 % CI 0·13, 0·58 and adjusted OR = 0·39; 95 % CI 0·22, 0·69, respectively). Moderate Fe supplementation did not seem to clearly prevent IDA (adjusted OR = 0·91; 95 % CI 0·42, 1·96) or to enhance the haemoconcentration (adjusted OR = 1·42; 95 % CI 0·58, 3·50). CONCLUSIONS: The prevalence of ID and IDA was high in late pregnancy in healthy pregnant women, particularly in those with initial ID and/or those not taking supplements. Starting pregnancy with no ID and/or taking moderate Fe supplementation decreased the likelihood of ID at delivery. The risk of haemoconcentration was high at delivery, but did not seem to be promoted by Fe supplementation. Further research is necessary to determine the most appropriate nutritional advice for pregnant women.