OBJECTIVE: To capture all data meeting a rigid definition of non-anaemic iron deficiency (NAID) and determine whether it is associated with poor outcomes compared with normal iron status and whether iron supplementation improves outcomes in NAID. DESIGN: Systematic review. DATA SOURCES: EMBASE, Medline, Web of Science, clinicaltrials.gov, International Clinical Trials Registry Platform (ICTRP) and Central from database inception to April 2014. ELIGIBILITY CRITERIA: Ferritin <16 μg/L (<12 μg/L if age <5 yr) in the absence of anaemia in observational studies or randomised trials. Where populations were deemed to be sufficiently similar, meta-analysis was undertaken. RESULTS: There were 21 studies included. NAID in pregnancy associated with reduction in birthweight (P = 0.028). Iron supplementation in NAID was associated with improvement in objective scores (P = 0.005) and self-rating (P = 0.03) of fatigue. Meta-analysis was limited and, where possible, was not statistically significant including the comparison of NAID with cardiovascular function in adults (VO2max P = 0.21, RERmax P = 0.68), educational attainment in children (P = 0.14), infant mental (P = 0.29) and psychomotor (P = 0.07) development, and iron supplementation in NAID with educational attainment in language (P = 0.31). CONCLUSIONS: There is emerging evidence that NAID is a disease in its own right, deserving of further research in the development of strategies for detection and treatment.