We read with great interest the comments from Motosko and colleagues (Reference Motosko, Brown and Kwatra2012) regarding our paper, in which we reported that insulin-like growth factor-1 (IGF-1) was not associated with delirium in our prospective study of critically ill patients (Morandi et al., Reference Morandi2011). Based on these results and those of the other four published studies examining IGF-1 and delirium, we take issue with the statement that a large body of evidence supports IGF-1's role in delirium pathogenesis. As pointed out by Adamis and Meagher (Reference Adamis and Meagher2011) in their recent systematic review, these five studies were small and produced conflicting results, with only three of the five finding an association between IGF-1 and delirium. Important differences in the patients studied may explain, in part, the conflicting results. Investigations of older medical inpatients who were not critically ill found IGF-1 to be associated with delirium, but neither a study of hip surgery patients (Lemstra et al., Reference Lemstra, Kalisvaart, Vreeswijk, Gool and Eikelenboom2008) nor our study of intensive care unit (ICU) patients found an association. The relative importance of IGF-1 to delirium might be diminished during critical illness given that a multitude of risk factors are involved in the pathogenesis of delirium in this vulnerable population. On average, medical ICU patients are subject to more than ten delirium risk factors; changes in IGF-1 in this setting might therefore contribute less to delirium than in other settings.

Appropriate validation and use of the assay employed to analyze levels of IGF-1, as highlighted by Motosko and colleagues, is essential when studying this hormone. Similar to the four studies that preceded ours, our study used a commercially available IGF-1 assay (ALPCO Diagnostics; Salem, NH, USA). The characteristics and validation of this radioimmunoassay – which is calibrated against the WHO International Reference Standard preparation of IGF-1 (NIBSC Code 02/254) and was designed to avoid errors caused by IGF binding protein interference – are described on the ALPCO Diagnostics website (http://www.alpco.com/pdfs/22/22-IGF-R21.pdf). The Hormone Assay and Analytical Services Core at Vanderbilt University, where our samples were assayed, has extensive experience with the ALPCO kit; before using the assay, the Core lab established inter- and intra-assay coefficients of variability, confirmed that varying concentrations of exogenous IGF-1 could be accurately recovered, and ran a set of controls. Though we agree that the additional procedures recommended by Frystyk et al. (Frystyk et al., Reference Frystyk, Freda and Clemmons2010) are reasonable suggestions, a consensus guideline for IGF-1 assays has yet to be established and none of the published studies of IGF-1 and delirium reported adherence to these suggestions.

Our finding that IGF-1 concentrations were not associated with delirium during critical illness does not rule out the possibility that IGF-1 plays a role in the development of cognitive deficits affecting ICU patients. Long-term cognitive impairment, for example, is common among ICU survivors (Hopkins and Jackson, Reference Hopkins and Jackson2006) and may be associated with derangements in IGF-1. Motosko and colleagues' comments are important and should lead to more thorough reporting in future publications regarding the role of IGF-1 in neurologic outcomes of critically ill patients.