Medical nutrition therapy and clinical outcomes in critically ill adults: a European multinational, prospective observational cohort study (EuroPN).

BACKGROUND: Medical nutrition therapy may be associated with clinical outcomes in critically ill patients with prolonged intensive care unit (ICU) stay. We wanted to assess nutrition practices in European intensive care units (ICU) and their importance for clinical outcomes. METHODS: Prospective multinational cohort study in patients staying in ICU ≥ 5 days with outcome recorded until day 90. Macronutrient intake from enteral and parenteral nutrition and non-nutritional sources during the first 15 days after ICU admission was compared with targets recommended by ESPEN guidelines. We modeled associations between three categories of daily calorie and protein intake (low: < 10 kcal/kg, < 0.8 g/kg; moderate: 10-20 kcal/kg, 0.8-1.2 g/kg, high: > 20 kcal/kg; > 1.2 g/kg) and the time-varying hazard rates of 90-day mortality or successful weaning from invasive mechanical ventilation (IMV). RESULTS: A total of 1172 patients with median [Q1;Q3] APACHE II score of 18.5 [13.0;26.0] were included, and 24% died within 90 days. Median length of ICU stay was 10.0 [7.0;16.0] days, and 74% of patients could be weaned from invasive mechanical ventilation. Patients reached on average 83% [59;107] and 65% [41;91] of ESPEN calorie and protein recommended targets, respectively. Whereas specific reasons for ICU admission (especially respiratory diseases requiring IMV) were associated with higher intakes (estimate 2.43 [95% CI: 1.60;3.25] for calorie intake, 0.14 [0.09;0.20] for protein intake), a lack of nutrition on the preceding day was associated with lower calorie and protein intakes (- 2.74 [- 3.28; - 2.21] and - 0.12 [- 0.15; - 0.09], respectively). Compared to a lower intake, a daily moderate intake was associated with higher probability of successful weaning (for calories: maximum HR 4.59 [95% CI: 1.5;14.09] on day 12; for protein: maximum HR 2.60 [1.09;6.23] on day 12), and with a lower hazard of death (for calories only: minimum HR 0.15, [0.05;0.39] on day 19). There was no evidence that a high calorie or protein intake was associated with further outcome improvements. CONCLUSIONS: Calorie intake was mainly provided according to the targets recommended by the active ESPEN guideline, but protein intake was lower. In patients staying in ICU ≥ 5 days, early moderate daily calorie and protein intakes were associated with improved clinical outcomes. Trial registration NCT04143503 , registered on October 25, 2019.

Chances and challenges of machine learning-based disease classification in genetic association studies illustrated on age-related macular degeneration.

Imaging technology and machine learning algorithms for disease classification set the stage for high-throughput phenotyping and promising new avenues for genome-wide association studies (GWAS). Despite emerging algorithms, there has been no successful application in GWAS so far. We establish machine learning-based phenotyping in genetic association analysis as misclassification problem. To evaluate chances and challenges, we performed a GWAS based on automatically classified age-related macular degeneration (AMD) in UK Biobank (images from 135,500 eyes; 68,400 persons). We quantified misclassification of automatically derived AMD in internal validation data (4,001 eyes; 2,013 persons) and developed a maximum likelihood approach (MLA) to account for it when estimating genetic association. We demonstrate that our MLA guards against bias and artifacts in simulation studies. By combining a GWAS on automatically derived AMD and our MLA in UK Biobank data, we were able to dissect true association (ARMS2/HTRA1, CFH) from artifacts (near HERC2) and identified eye color as associated with the misclassification. On this example, we provide a proof-of-concept that a GWAS using machine learning-derived disease classification yields relevant results and that misclassification needs to be considered in analysis. These findings generalize to other phenotypes and emphasize the utility of genetic data for understanding misclassification structure of machine learning algorithms.