A one year prospective study was conducted on 82 infants between 1 and 12 months of age in the Eastern area of Málaga-Axarquía, of whom 40 of them had been diagnosed with non-IgE-mediated CMA (with suggestive symptoms and positive response to milk withdrawal), 12 not diagnosed with CMA, and 30 of them were the control group. FCP was measured at three different times: time of diagnosis, and one and three months later. ANOVA for repeated measures, nominal logistic regression and ROC curves were prepared using the SPSS.20 package and Medcalc.
Differences between diagnostic and control groups were assessed: there was a statistically significant relationship (P < .0001) between high FCP levels and infants suffering CMA, as well as the levels at time of diagnosis, 1 and 3 months (P < .001). A ROC curve was constructed between FCP levels and diagnosis of CMA, with 138 μg/g, with the best cut-off being with an area under the curve of 0.89. However, it is only 0.68 to predict a clinical response.
FCP levels lower than 138 μg/g could be useful to rule out non-IgE-mediated CMA diagnosis. Calprotectin is not a good test to predict clinical response to milk withdrawal.