重症不同营养状况的血浆代谢组变化：队列研究
营养不良与重症死亡率增加有相关性。代谢动态失衡常见于重症，其特征为大量代谢途径严重破坏。波士顿布莱根女子医院、斯坦福大学医学中心、纽约长老会医院假设重症监护病房（ICU）收治的重症患者，有营养不良与无营养不良相比，循环血浆代谢分布不同，入组85例全身炎性反应综合征或脓毒血症的ICU患者进行队列研究，发现营养不良与血浆中谷胱甘肽和嘌呤的代谢产物增加显著相关，营养不良患者28天死亡率是无营养不良组的3倍。此外，将与营养不良相关的代谢产物，综合患者的年龄、是否患脓毒血症或恶性肿瘤，可提高对重症患者28天死亡率预测的准确性。
JPEN J Parenter Enteral Nutr. 2016;40(4):115-117.
Plasma Metabolome Alteration in Critical Illness According to Nutrition Status: A Cohort Study.
Mogensen K, Lasky-Su J, Rogers A, Baron R, Fredenburg L, Rawn J, Robinson M, Massaro A, Choi A, Christopher K.
Brigham and Women's Hospital, Boston, MA, USA; Stanford University Medical Center, Palo Alto, CA, USA; New York-Presbyterian Hospital, New York, NY, USA.
Purpose: Malnutrition is associated with increased mortality in critical illness. Loss of metabolic homeostasis is a common in critical illness and characterized by a severe disruption of a multitude of metabolic pathways. We hypothesize that circulating plasma metabolic profiles at intensive care unit (ICU) admission would differ in critically ill patients with malnutrition relative to those without malnutrition.
Methods: We performed a cohort study in a 20-bed medical ICU of a Boston teaching hospital. We generated metabolomics profiles in plasma collected at ICU admission in 85 adult patients with systemic inflammatory response syndrome or sepsis from 2008-2010. To identify metabolites and metabolite pathways, we performed an integrated discovery approach that included gas and liquid chromatography, mass spectroscopy, logistic regression, partial least squares-discriminant analysis, and metabolite set enrichment analysis where metabolite sets are ranked according to Holm P value. We removed metabolites with the lowest IQR of variability and those related to medications, leaving 281 metabolites. All raw metabolite concentrations were log2 transformed and normalized via cubic root transformation and Pareto scaling. Nutrition status was determined by a registered dietitian with data related to anthropometric measurements, biochemical indicators, clinical signs of malnutrition, malnutrition risk factors, and metabolic stress. Malnutrition was considered to be present if the patient was diagnosed by a registered dietitian at ICU admission with nonspecific protein-calorie malnutrition, specific (mild, moderate, or severe) protein-calorie malnutrition, or marasmus.
Results: Most patients were male (56%) and white (80%). The mean ± SD age at ICU admission was 55.0 ± 14.5 years; 38% of cohort patients were diagnosed with malnutrition; and 66% of cohort patients had sepsis. The mean APACHE II score was 25.7 ± 9.9. The 28-day mortality of the cohort was 33.3%. There were significant differences in age and malignancy status between patients with malnutrition and those without. In our cohort, pyroglutamine (a glutathione pathway metabolite) and hypoxanthine (a purine metabolite) were identified as predictive of malnutrition in logistic regression analysis. Pyroglutamine and hypoxanthine were also identified in partial least squares-discriminant analysis as having high importance for malnutrition classification. Furthermore, quantitative metabolite set enrichment analyses show that increases in the glutathione pathway metabolites (cysteinylglycine, pyroglutamine, L-cysteine) and increases in purine metabolites (adenosine monophosphate, hypoxanthine, xanthine, and uric acid) were significantly associated with malnutrition. Furthermore, malnutrition was associated with 28-day mortality. When adjusted for age, sepsis, malignancy status, and APACHE II, patients diagnosed with malnutrition had a 3-fold odds of 28-day mortality (OR, 3.26; 95% CI, 1.12-9.52) relative to patients without malnutrition. Addition of metabolites associated with malnutrition to a 28-day mortality model including age, sepsis, malignancy status, and APACHE II improved model discrimination for mortality (area under the curve, 0.78-0.83).
Conclusions: Malnutrition is associated with differential metabolic profiles early in critical illness. Common to all of our metabolome analyses, glutathione and purine metabolism—which play principal roles in cellular redox regulation and accelerated tissue ATP degradation, respectively—were significantly altered with malnutrition. Furthermore, the addition of metabolites related to malnutrition to mortality prediction models can improve the discrimination for 28-day mortality in the critically ill.
Financial support: This work was supported by the A.S.P.E.N. Rhoads Research Foundation.
