Thyroid hormone levels improve the prediction of mortality among patients admitted to the intensive care unit

Sir: We read with great interest the recent article by Chinga-Alayo et al. [1] regarding the prognostic significance of hormonal measurements in critically ill patients. This study assessed adrenal and thyroid function but also focused on other hormones, including growth hormone, dehydroepiandrosterone, and prolactin. This is highly important since there is evidence that the latter hormones modulate humoral and cellular immune functions [2, 3]. In this context it is surprising that the authors failed to show differences in growth hormone, dehydroepiandrosterone, or prolactin levels between survivors and nonsurvivors among critically ill patients (Table 2). We would like to present our experience on the topic of hormonal measurements in critically ill patients. We performed blood sampling in 150 severely ill patients within 24 h after ICU admission to investigate whether hormone levels differ between survivors and nonsurvivors. The preliminary analysis of our data revealed no differences between survivors and nonsurvivors in terms of cortisol, corticotropin, free thyroxine, and tri-iodothyronine levels, but survivors had higher concentrations of thyroidstimulating hormone, prolactin, and dehydroepiandrosterone sulfate than nonsurvivors [4]. The reason for the discrepant results with respect to prolactin and dehydroepiandrosterone levels is unclear. Possible explanations include the timing of sampling, the assays used, and more importantly the differences in patient population. Our ICU is a medical and surgical unit admitting a large number of polytraumatized patients, including those with severe traumatic brain injury while the cohort of Chinga-Alayo et al. comprised only a small number of trauma patients (n = 4) and possibly of surgical cases. Nevertheless, the study by Chinga-Alayo and coworkers highlights the importance of measuring hormones in critically ill patients. Endocrine alterations are frequently observed in severely ill patients, they relate to mortality, and of particular importance, as the current study showed, they improve our predictive ability when used in conjunction with well established severity scores. Further research is needed to validate such findings in a larger cohort comprising a wider range of ICU patients. In such a case it might be justified to incorporate a hormonal score to the already available scoring systems. References