Objective: To examine zinc-protoporphyrin (ZPP) and haemoglobin levels, and to determine predictors of iron deficiency anaemia (IDA) in Zambian infants. Subjects and methods: Ninety-one women and their normal birth weight (NBW) infants were followed bi-monthly during the first 6 months of life, and iron status, food intake, malaria parasitaemia and growth were monitored. At 4 months, the infants were divided into two groups, and the data were analysed according to whether or not they were exclusively breastfed. Results: Almost two-third of infants were born with low iron stores as defined by ZPP levels, and this proportion increased with age. Over 50% had developed IDA by 6 months. Exclusive breastfeeding at 4 months could be a protective factor for IDA (odds ratio (OR): 0.2; 95% confidence interval (CI): 0.0-1.1). Exclusively breastfed infants had higher haemoglobin values at 4 and 6 months (mean difference 0.6; 95% CI: 0.1-1.2 g/dl and mean difference 0.9; 95% CI: 0.2-1.7 g/dl, respectively), compared with infants with early complementary feeding. In univariate analysis, past or chronic placental malaria appeared to be a predictor of IDA at 4 and 6 months, but the significance was lost in multivariate analysis. Conclusions: Zambian NBW infants are born with low iron stores and have a high risk to develop IDA in the first 6 months of life. Continuation of exclusive breastfeeding after 4 months is associated with a reduction of anaemia. The effect of placental malaria infection on increased risk of infant IDA could not be proven.

Due to the insidious nature of infant anaemia, this disorder frequently remains undetected and untreated by health-care workers in resource-poor settings. We assessed the accuracy of a low-cost and simple diagnostic tool, the haemoglobin colour scale (HCS), in estimating haemoglobin (Hb) values in infants between zero and four months of age. In a rural hospital in Zambia, blood samples were analysed for Hb concentration by HCS and HemoCue method. Bland-Altman plots were used to express agreement between the two methods. The mean difference between HCS and HemoCue at birth (n = 94), two months (n = 87) and four months (n = 69) was 0.39, 0.20 and -0.11 g/dL, respectively. Limits of agreement were -2.39 to 1.51, -1.80 to 2.20 and -1.98 to 1.75 g/dL, respectively. Disagreement with HemoCue measurements of more than 2 g/dL was noted in only 4% of all blood samples. We conclude that the HCS provides Hb estimations in infants aged 0-4 months that are sufficiently accurate to improve timely recognition of anaemia in settings where there is no laboratory.