www.thelancet.com Vol 369 April 28, 2007 1429 However, this fi nding comes from only one trial in which breastfeeding was stopped before the procedure. Even if this fi nding was confi rmed by other studies, it might not necessarily be clinically signifi cant. In this regard, we would like to summarise our fi ndings on nonpharma cological analgesia in new born babies. We studied 197 healthy term infants (birthweight 2500–4500 g) at the time of blood sampling by heel prick for metabolic screening. They were randomly assigned to one of four groups: (1) breastfeeding 2 min before and during the procedure; (2) 2·0 mL 20% glucose 2 min before and during the procedure; (3) sensory stimulation by holding, speaking, and fondling; or (4) wrapping. Pain was evaluated with the premature infant pain profi le (PIPP score). As in Carbajal’s study, we found no diff erence between breastfeeding and glucose (table). In both studies, mean PIPP scores were very similar and quite low. It might well be that sucrose plus breastfeeding will further lower the pain score. However, in the average nursery, this can represent an additional burden in an already busy routine. One might wonder whether lowering an already low score is worth the eff ort.