Objective The optimal iron hypothesis posits a trade‐off in iron nutrition—iron deficiency restricts iron available to infectious agents, protecting against severe infection, but also compromises immune defense—such that mild‐to‐moderate iron deficiency may be more adaptive than either iron‐replete or severe deficiency in environments with high infectious disease load. This hypothesis has not been tested among adults. Materials and Methods A secondary analysis of data and specimens from 220 lactating mothers in northern Kenya was conducted. Elevated serum C‐reactive protein (CRP > 2 or >5 mg/l) was utilized to identify prevalent subclinical infection/inflammation. Iron deficiency was identified with transferrin receptor in archived dried blood spots (TfR > 5.0 mg/l). The absence of iron deficiency or anemia (Hemoglobin < 12 g/l) defined the iron replete state. Iron‐deficient erythropoiesis (IDE, mild‐to‐moderate iron deficiency) was defined as iron deficiency without anemia; iron deficiency anemia (IDA, severe iron deficiency) as iron deficiency with anemia; and noniron‐deficiency anemia (NIDA) as anemia without iron deficiency. Results The prevalence of elevated inflammation (subclinical infection) was lowest in IDE. In logistic regression, IDE was inversely associated with inflammation (for CRP > 2 mg/l: adjusted odds ratio, aOR = 0.30; p = 0.02; for CRP > 5 mg/l: aOR = 0.27; p = 0.10), compared to the iron replete state. The protective effect of IDE differed in the presence of vitamin A deficiency or underweight. Conclusions We interpret these patterns as tentative support for the optimal iron hypothesis in breastfeeding women in the infectious disease ecology of northern Kenya. Iron deficiency may interact in important ways with other forms of malnutrition that are known to affect immune protection.