Dr. Kelly asks whether spontaneous caloric and protein intake could have been modified by the ingestion of protein supplements. To avoid significant influence of the supplements on regular food consumption, patients enrolled in the trial were advised to consume the supplements in the evening. Protein intake was evaluated with a food-frequency questionnaire; thus, evaluation relied on the patient's memory. This questionnaire was administered three times during 1 year, the first time within 1 week after surgery. Mean spontaneous protein intake in the protein supplement group was 45.0 ± 2.4 g/d at baseline and 44.0 ± 2.8 g/d at the end of supplement use; in the controls, these values were 51.0 ± 3.0 g/d and 44.0 ± 3.2 g/d. There were no statistically significant differences between the groups and no change over time.

In response to the comments of Dr. Craig, the effects of protein-calorie supplements compared with no intervention have been addressed in a previous study by our group [1]. We subsequently showed that the benefits previously observed could be mostly ascribed to protein [2]. We agree that the low baseline levels of albumin and prealbumin were related to surgery or trauma, as well as to a poor nutritional state. Under these conditions, in a double-blind trial, higher incremental increase of prealbumin in the supplement group (P = 0.075) could be interpreted as a direct indication of compliance with treatment or indirect evidence of accelerated recovery.

The greater increase in insulin-like growth factor-I (IGF-I) levels in response to protein supplements remained significant when the analysis combined the dropouts in each group with assigned values of the other group, or when a Bonferroni criterion was applied. Furthermore, the IGF-I results are in agreement with many preclinical studies demonstrating a relation between IGF-I levels and protein intake [3]. The changes in IGF-I levels could therefore be considered significant and clinically relevant.