To the Editor,

Among the limitations of their study not cited by Loeffen and colleagues [1] is the failure to acknowledge the results of a study in Central America of the impact of nutritional status at diagnosis on clinical outcomes in children with an unselected variety of cancers. This prospective study enrolled almost 3000 subjects of whom 1787 had their nutritional status categorized on the basis of arm anthropometry; a sample approximately sevenfold greater than that in the study by Loeffen et al. (N = 269). Malnourished children had an event-free survival that was inferior to that of other children [2].

This omission, of which the Dutch authors are aware (their reference 21), is compounded by a mis-representation of the data; the authors of the Central American study did not “promote the use of BMI as an indicator of nutritional status” in children with cancer, preferring arm anthropometry for reasons articulated previously [3]. Loeffen and colleagues may also take note of a study conducted subsequently in Guatemala that reported on the restoration of good survival prospects for children with acute lymphoblastic leukemia who were malnourished at diagnosis but normalized their nutritional status during therapy [4].

The plea from Loeffen and colleagues, to undertake prospective studies on large numbers of subjects, has been pre-empted and acted upon already.