Next Article in Journal
The Low Energy Availability in Females Questionnaire (LEAF-Q) as a Useful Tool to Identify Female Triathletes at Risk for Menstrual Disorders Related to Low Energy Availability
Previous Article in Journal
Associations between Conventional and Emerging Indicators of Dietary Carbohydrate Quality and New-Onset Type 2 Diabetes Mellitus in Chinese Adults
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Nutrients
Volume 15
Issue 3
10.3390/nu15030649
nutrients-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Combining Phenylalanine and Leucine Levels Predicts 30-Day Mortality in Critically Ill Patients Better than Traditional Risk Factors with Multicenter Validation

by
Yi-Liang Tsou
1,
Chao-Hung Wang
1,2,*,
Wei-Siang Chen
1,3,
Huang-Ping Wu
4,
Min-Hui Liu
1,5,
Hsuan-Ching Lin
1,3,
Jung-Jung Chang
6,
Meng-Shu Tsai
6,
Tien-Yu Chen
7,
Cheng-I Cheng
2,7,
Jih-Kai Yeh
8 and
I-Chang Hsieh
8
1
Heart Failure Research Center, Division of Cardiology, Department of Internal Medicine, Chang Gung Memorial Hospital, Keelung 20401, Taiwan
2
School of Medicine, Chang Gung University, Taoyuan 33302, Taiwan
3
Intensive Care Unit, Division of Cardiology, Department of Internal Medicine, Chang Gung Memorial Hospital, Keelung 20401, Taiwan
4
Division of Pulmonary, Critical Care and Sleep Medicine, Chang Gung Memorial Hospital, Keelung 20401, Taiwan
5
Department of Nursing, Keelung Chang Gung Memorial Hospital, Keelung 20401, Taiwan
6
Division of Cardiovascular Disease, Department of Internal Medicine, Chang Gung Memorial Hospital, Chiayi 61363, Taiwan
7
Division of Cardiology, Department of Internal Medicine, Kaohsiung Chang Gung Memorial Hospital, Kaohsiung 83301, Taiwan
8
Division of Cardiology, Department of Internal Medicine, Chang Gung Memorial Hospital, Linkou 33305, Taiwan
*
Author to whom correspondence should be addressed.
Nutrients 2023, 15(3), 649; https://doi.org/10.3390/nu15030649
Submission received: 5 January 2023 / Revised: 20 January 2023 / Accepted: 24 January 2023 / Published: 27 January 2023
(This article belongs to the Section Clinical Nutrition)
Browse Figures
Review Reports Versions Notes

Abstract

:
In critically ill patients, risk scores are used; however, they do not provide information for nutritional intervention. This study combined the levels of phenylalanine and leucine amino acids (PLA) to improve 30-day mortality prediction in intensive care unit (ICU) patients and to see whether PLA could help interpret the nutritional phases of critical illness. We recruited 676 patients with APACHE II scores ≥ 15 or intubated due to respiratory failure in ICUs, including 537 and 139 patients in the initiation and validation (multicenter) cohorts, respectively. In the initiation cohort, phenylalanine ≥ 88.5 μM (indicating metabolic disturbance) and leucine < 68.9 μM (indicating malnutrition) were associated with higher mortality rate. Based on different levels of phenylalanine and leucine, we developed PLA scores. In different models of multivariable analyses, PLA scores predicted 30-day mortality independent of traditional risk scores (p < 0.001). PLA scores were then classified into low, intermediate, high, and very-high risk categories with observed mortality rates of 9.0%, 23.8%, 45.6%, and 81.8%, respectively. These findings were validated in the multicenter cohort. PLA scores predicted 30-day mortality better than APACHE II and NUTRIC scores and provide a basis for future studies to determine whether PLA-guided nutritional intervention improves the outcomes of patients in ICUs.

1. Introduction

Recent decades have seen substantial improvements in critical care. This improvement is associated with advanced therapeutic modalities and development of precise risk stratification models, such as the Acute Physiology and Chronic Health Evaluation II (APACHE II), Sequential Organ Failure Assessment (SOFA), and Nutrition Risk in the Critically Ill (NUTRIC) [1,2,3]. Nevertheless, the mortality rate among critically ill patients admitted to intensive care units (ICUs) remains high, approaching 27% in one month [4,5]. Precision nutrition therapy is essential to improve short- and long-term outcomes in different phases of the patient’s journey during critical illness [6,7]. The phases of critical illness, however, are not well-defined. A risk stratification tool for critically ill patients that can also provide precise information for nutritional intervention would be useful.
Previously, we and others found that phenylalanine can provide prognostic value in patients with infection [8], coronavirus disease 2019 (COVID-19) [9,10], cardiovascular diseases [11,12,13], or critical illness [13,14]. In adults, the elevation of phenylalanine is related to tissue breakdown [11,15], liver and kidney dysfunction [11,14,16], systemic inflammation and tremendous oxidative stress [15], as well as insufficient tetrahydrobiopterin (BH4) [14], the co-factor of the enzyme that metabolizes phenylalanine [11,14,16]. Stress hyperphenylalaninemia has been associated with a three-fold increase of mortality in ICU patients [13,14]. On the other hand, a U-shaped relationship between leucine levels and mortality in ICU patients was also noted [17,18]. High leucine levels may represent large endogenous release of energy and amino acids in the acute phase of critical illness and anabolic resistance, while low leucine levels may imply malnutrition. Although the interpretation of the changes in these amino acids needs to be further investigated, measuring them may provide information that adds to traditional mortality risk assessment and could be useful for anti-oxidant therapy and precise nutritional supplementation at the right time while avoiding over-feeding.
Thus, the aims of this study were as follows: (1) to verify the prognostic value of phenylalanine and leucine, and to investigate whether they can be combined to develop a phenylalanine leucine amino acid (PLA) score independent of traditional risk scores for predicting 30-day mortality in ICU patients; (2) to investigate whether information provided by these two amino acids and PLA scores was superior for predicting mortality to traditional nutritional biomarkers, such as albumin, pre-albumin, and transferrin; and (3) to validate use of this score in an independent cohort. The findings of this study provide a foundation for future studies to assess whether PLA-based nutritional interventions improve the outcomes of patients with critical illness.

2. Materials and Methods

2.1. Patient Enrollment

For the initiation cohort from April 2017 to November 2021, patients in critical condition were consecutively enrolled at the medical and cardiac ICUs based on these inclusion criteria: (1) they had an APACHE II score of ≥15 or were intubated due to respiratory failure; (2) they were over 20 years old; and (3) they needed to stay in the ICU > 48 h. The exclusion criteria were (1) patients who died before baseline phenylalanine measurement or (2) patients with other comorbidities that were not related to the main cause for admission and might compromise their survival in 3 months, such as cancer at terminal stage. The study was designed and carried out in accordance with the principles of the Declaration of Helsinki and had approval from the Ethics Review Board of Chang Gung Memorial Hospital. All patients provided informed consent.
For validation, an independent cohort was recruited from multiple centers, including Lin-ko, Chia-Yi, Kaohsiung, and Keelung Chang Gung Memorial Hospitals from November 2021 to August 2022 (Figure 1). The patient number ratio of the initiation cohort to validation cohort was 4:1.

2.2. Scoring Systems

On the day of admission to the ICU, APACHE II [1], SOFA [2], and NUTRIC [3] scores were calculated to assess disease severity.

2.3. Blood Sampling and Examination

In the early morning of the day after obtaining informed consent, blood samples were collected in ethylenediaminetetraacetic acid (EDTA)-containing tubes. Plasma levels of phenylalanine and leucine were measured by ultra-performance liquid chromatography (UPLC). Measurement of pre-albumin, transferrin, albumin, estimated glomerular filtration rate (eGFR), C-reactive protein (CRP), and hemoglobin was conducted in the central laboratory of each hospital.

2.4. Phenylalanine and Leucine Measurement

Protein precipitation in the plasma samples (100 μL) was conducted with sulfosalicylic acid (10%). After centrifugation, derivatization was performed by aminoquinolyl-carbamyl (AQC) in acetonitrile. Phenylalanine and leucine were analyzed using the ACQUITY UPLC System (Waters Corp., Milford, MA, USA), consisting of a Sample Manager, a Binary Solvent Manager, and a Tunable UV detector. To control the system and collect data, EmpowerTM 2 software was used. A 2.1 mm × 100 mm ACQUITY BEH C18 column was used for separation at a flow rate of 0.70 mL/min. The detection limit was 3.3 μM for phenylalanine and 0.9 μM for leucine. The linear range was 25–500 μM. The average intra-assay coefficient of variation was 2.6% for phenylalanine and 4.5% for leucine. The total coefficient of variation was 2.7% for phenylalanine and 4.1% for leucine.

2.5. Follow-Up Program

The primary endpoint was 30-day all-cause mortality. Follow-up data were obtained from hospital records, personal communication with the patients’ physicians, and telephone interviews with patients/family. Patients were prospectively followed until death or for a maximum of 30 days.

2.6. Development of the Phenylalanine-Leucine Amino Acid (PLA) Score

From the initiation cohort, we used the concentration of plasma phenylalanine and leucine to generate PLA scores for a 30-day mortality prediction model (Figure 1). We used Youden’s index to define the cutoff values for phenylalanine and leucine by their ability to distinguish 30-day mortality and survival. To develop the PLA score, we used Cox multivariable analysis to select the range of phenylalanine and leucine concentrations that had statistical significance for predicting 30-day mortality. Different categories of each amino acid were defined based on normal ranges and risk ranges published in our previous studies, Youden’s index, and appropriate gradients. The regression coefficient for each category was converted into points by dividing it by the smallest regression coefficient in the model [19,20]. The PLA score was the sum of the points for each category.

2.7. Statistical Analysis

Results are expressed as the mean ± standard deviation (SD), the median [interquartile range (IQR)], and the number (percentage) when appropriate. Data were analyzed using t-test or the Mann–Whitney U test. The linear trend of the data distribution across study groups was tested using linear contrast in general linear model for continuous variables and Cochran–Armitage chi-square analysis for categorical variables. Receiver operating characteristic (ROC) curves and Youden’s index were used to identify the cutoff value of variables. A univariate Cox proportional hazards model was used to determine the predictive value of variables on mortality. Kaplan–Meier analyses with a log-rank test were used to compare time-dependent outcomes. Variables with a p value of <0.05 in the univariate analysis were recruited in the multivariable analysis. Based on Cox multivariable analysis, we identified independent predictors of mortality by adjusting for significant variables in the univariate analysis. We calculated hazard ratios (HRs) and 95% confidence intervals (CIs). All statistical analyses were two-sided using SPSS software (version 22.0, SPSS, Chicago, IL, USA) and R software (version 3.5.1, The R foundation for Statistical Computing, Vienna, Austria). A p value of <0.05 was considered significant.

3. Results

3.1. Baseline Characteristics and Laboratory Data

The baseline characteristics for the initiation cohort (n = 537) are shown in Table 1. The mean age was 71.5 years. Of the 537 patients, 323 (60.1%) were male, 415 (77.3%) were on ventilators, and 209 (38.9%) were receiving inotropic agents. The average APACHE II, SOFA, and NUTRIC scores were 18.8, 7.1, and 5.3, respectively.
These patients were admitted to the ICU for the following conditions: 292 patients (54.4%) for cardiovascular disease [including 165 patients (56.5%) due to heart failure]; 177 patients (33.0%) for infection; 29 patients (5.4%) for pulmonary diseases; 33 patients (6.1%) for gastrointestinal bleeding; and 6 patients (1.1%) for other conditions. Phenylalanine and leucine concentrations ranged from 38.1 μM to 645 μM and from 28.3 μM to 683 μM, respectively.
During the 30-day follow-up period, 150 patients (27.9%) died. Patients in the death group showed higher likelihood of being male, having atrial fibrillation, using a ventilator, and receiving inotropic agents. Higher APACHE II, SOFA, and NUTRIC scores were also noted in the death group. Compared with the survival group, CRP levels were higher in the patients who died. The levels of cholesterol, albumin, pre-albumin, and transferrin were lower in the non-survivors (Table 1).
For the validation cohort, 139 patients were recruited from 4 centers. The validation cohort had no significant differences in the demographic and laboratory variables when compared to the initiation cohort (Table S1). In the 30 days after enrollment, 41 patients (29.5%) died.

3.2. The Relationship between Mortality and Phenylalanine and Leucine Levels

In the univariate analysis for the initiation cohort, higher phenylalanine levels were linearly correlated with higher 30-day mortality (each increase of 10 μM of phenylalanine, HR = 1.09, 95% CI = 1.07–1.10, p < 0.001). Based on Youden’s index, the cutoff for phenylalanine was set at 88.5 μM. Phenylalanine ≥ 88.5 μM was associated with a higher 30-day mortality rate, compared to phenylalanine < 88.5 μM (HR = 3.49, 95% CI = 2.47–4.92, p < 0.001) (Table S2). Next, we analyzed the prognostic value of leucine. Since a U-shaped relationship was noted in our previous study, the cutoff points for the U-curve were set at 68.9 μM and 165.6 μM, based on Youden’s index. In the univariate analysis, patients with leucine > 165.6 μM and < 68.9 μM had a higher 30-day mortality rate (HR = 1.82, 95% CI = 1.24–2.67, p = 0.002 and HR = 1.85, 95% CI = 1.14–3.01, p = 0.012, respectively), compared to patients with leucine levels between 68.9 μM and 165.6 μM. In multivariable analysis, only phenylalanine ≥ 88.5 μM and leucine < 68.9 μM were independently associated with a higher mortality rate (HR = 3.85, 95% CI = 2.67–5.55, p < 0.001 and HR = 2.76, 95% CI = 1.68–4.54, p < 0.001, respectively). In addition, we noted that, in patients with phenylalanine < 88.5 μM, a higher leucine level was linearly correlated with a lower 30-day mortality (each increase of 10 μM of leucine, HR = 0.92, 95% CI = 0.84–0.99, p = 0.03). Based on these findings, we built up the PLA score model.

3.3. The Building of PLA Score

Based on the above findings, we stratified the concentrations of phenylalanine into eight levels and leucine into four levels (Table 2). Next, we conducted Cox multivariable analysis to calculate the points for each level of amino acids. The regression coefficient (B) was determined for every level of amino acids. The smallest B (0.485) was used as base constant so that the points were derived from the formula B/0.485. Table 2 shows the points for each level of the two amino acids. The PLA score was the sum of the points of phenylalanine and leucine. In the initiation cohort, the mean PLA score was 2.75 ± 1.91 (ranging from 0 to 10.3 points).
For clinical application, patients were classified into four risk categories according to their PLA scores and the observed mortality. In the initiation cohort, 0–1 point indicated low risk; 1.1–4 points, intermediate risk; 4.1–5 points, high risk; and >5 points, very-high risk. The observed 30-day mortality rate in each risk group was 9.0%, 23.8%, 45.6%, and 81.8%, respectively. Table 3 presents the characteristics of the four risk categories. The APACHE II, SOFA, and NUTRIC scores increased as PLA scores increased. As PLA score risk categories increased, the levels of CRP became elevated, and cholesterol, albumin, pre-albumin, and transferrin declined. Figure 2A shows the Kaplan–Meier curves of the four risk groups according to PLA scores (log-rank = 167.8, p < 0.001). Using the low risk group as the reference for 30-day mortality, the very-high risk group had a relative risk of 19.5 (95% CI = 9.91–38.5, p < 0.001) (Table S3). Based on the grid of phenylalanine and leucine concentrations, Figure 2B shows the relative risk of 30-day mortality associated with different scores compared to the reference. In the validation cohort, the Kaplan–Meier curves show that PLA score-based risk classification still significantly predicted the risk of mortality (log-rank = 19.2, p < 0.001) (Figure 2C).

3.4. The Prognostic Value of PLA Scores, Other Scores, and Biomarkers

In the univariate analysis, 30-day mortality was associated with higher PLA, APACHE II, SOFA, and NUTRIC scores, higher CRP levels, but lower levels of cholesterol, albumin, pre-albumin, and transferrin (Table 4). Since the component of NUTRIC score contains APACHE II and SOFA scores, significant collinearity exists. No collinearity was present for all other variables (all variance inflation factors were lower than 1.6). We performed two models of multivariable analysis. In model 1, when all these factors were recruited in the model (except NUTRIC score), there were only two independent predictors, including PLA (HR = 1.46, 95% CI = 1.33–1.60) and APACHE II scores (HR = 1.07, 95% CI = 1.03–1.11). In model 2 (without APACHE II and SOFA scores), there were also two independent predictors, including PLA (HR = 1.48, 95% CI = 1.35–1.62) and NUTRIC scores (HR = 1.18, 95% CI = 1.07–1.30). Based on ROC curves (Figure 2D), the area under the curves were 0.763 (95% CI = 0.725–0.799), 0.682 (95% CI = 0.640–0.721), and 0.668 (95% CI = 0.626–0.707) for PLA, APACHE II, and NUTRIC scores, respectively (PLA vs. APACHE II, p = 0.019; PLA vs. NUTRIC, p = 0.003).

4. Discussion

In this study, we focused on patients who were in critical condition with an APACHE II score ≥ 15 or intubated due to respiratory failure. Our study verified the predictive value of phenylalanine and leucine levels for 30-day mortality in patients in the ICU. Furthermore, we integrated these two amino acids to develop the PLA score. PLA scores predicted 30-day mortality independently of traditional risk factors, risk scores such as APACHE II, SOFA, and NUTRIC scores, and nutritional biomarkers. In addition, the prognostic value of the PLA score was superior to APACHE II, SOFA, and NUTRIC scores, as well as traditional nutritional biomarkers, including albumin, pre-albumin, and transferrin. The value of the PLA score was validated in an independent cohort recruited from multiple centers.

4.1. The Meaning of Elevated Phenylalanine Levels

Recent studies demonstrated that stress hyperphenylalaninemia predicts mortality in critically ill patients [9,10,13,14]. Atila et al. reported that in patients with severe COVID-19, the level of phenylalanine increased significantly compared with the control group; however, the levels of all other amino acids decreased [9,10,13,14]. Phenylalanine elevation in critical stress has been associated with depletion of BH4, a co-factor of phenylalanine hydroxylase, and with genetic variants on the genes for BH4 synthesis and recycling [14].
Sepsis or systemic inflammatory responses to infectious or non-infectious stimuli are increasingly common causes of mortality among patients in ICUs. Previously, Wischmeyer et al. [6] and Berger et al. [21] divided these patients into three phases, including acute, chronic, and recovery phases. In the acute phase, an overwhelming inflammatory mediator response to infective or non-infective stimuli results in excessive production of oxidative stress. Reactive oxygen species consume tremendous amounts of BH4 and activate the production process of BH4 either by de novo synthesis or recycling [22,23]. Genetic variants on these pathways cause inadequate repletion of BH4, leaving phenylalanine unmetabolized. Stress hyperphenylalaninemia may be used as an indicator of substantially increased oxidative stress and may indicate the timing of anti-oxidant therapy. However, the benefit of anti-oxidant therapy during the acute phase has not been well proven [24]. Whether phenylalanine level-guided anti-oxidant or BH4 supplementation is effective in lowering the mortality risk requires further investigation [21].
Although the elevation of plasma phenylalanine concentration might be a byproduct of metabolism in response to critical stress, recent reports have suggested that elevated phenylalanine is toxic to inflammatory organs [25] and that ectopic phenylalanine catabolism in the heart is associated with cardiac dysfunction [26]. On the other hand, hyperphenylalaninemia represents the decompensation of alternative phenylalanine catabolism pathways, which produce tremendous amounts of phenylalanine-derived phenolic acids with oxidative stress such as phenylpyruvate, phenyllactate, and phenylacetate [27,28,29]. Stress hyperphenylalaninemia also indicates anabolic resistance, as demonstrated by overloaded amino acid and lipid downstream metabolites in the circulation [13,18]. The perturbed phenylalanine metabolism can interfere with the production of catecholamine and cause unstable hemodynamics. Moreover, prior studies reported that inadequate BH4 bioavailability is associated with impaired immunity [30], uncoupling of nitric oxidase, and microvascular hypoperfusion [22], all related to poor outcomes.

4.2. The Meaning of Leucine Levels

As mentioned by Wischmeyer et al. [6], early in the acute phase of critical illness, endogenous supplies and release of energy were the main sources of energy supply. However, in the recovery phase, exogenous energy delivery becomes the primary energy source [31]. It seems that the amount of protein and non-protein calories supplied should differ in different phases. However, different guidelines have provided different suggestions [32]. In addition, reduced calorie support during the acute phase seems not appropriate for malnourished patents since they do not have the metabolic reserve to generate enough endogenous energy [3,32]. These contradictions show that we should have a scientific way to judge what phase the patients are at, which is currently performed by counting days after admission to the ICU, and better ability to assess their nutritional status as well.
A previous study reported that low total amounts of essential amino acids were associated with poor outcomes in patients with advanced disease [33]. It has been shown that leucine levels correlate well with the total amount of essential amino acids [18]. Recently, we noted a U-shaped relationship between leucine concentrations and ICU mortality [17,18]. High leucine levels represent anabolic resistance and massive endogenous release of energy in the acute phase, characterized by increased unmetabolized acylcarnitines in the circulation due to mitochondrial dysfunction with impaired beta-oxidation [15,18]. However, low leucine levels represent malnutrition and are correlated with high mortality risk two weeks after ICU admission. This study confirmed these findings with cutoffs noted at 165.6 μM and 68.9 μM, respectively. However, when phenylalanine concentrations were considered in the multivariable analysis model, the prognostic value of high leucine levels was replaced by high phenylalanine levels, suggesting that elevated phenylalanine represents metabolic disturbance along with co-factor deficiency more specifically than leucine level does. In summary, integrating leucine and phenylalanine levels can identify patients at a high risk of mortality and also help investigate how nutritional needs differ at different phases of critical illness following ICU admission.

4.3. Comparisons between PLA Score and Traditional Biomarkers

Our data demonstrated that each 1-unit increase in PLA score was associated with a 46% increase in 30-day mortality risk, independent of APACHE II, SOFA, and NUTRIC scores, traditional risk factors, and nutritional biomarkers. These findings support that PLA score is not a direct indicator of the nutritional status of the critically ill patients in ICUs. Although APACHE II is a traditional risk score independently associated with 30-day mortality in our multivariable model, it does not translate into therapeutic recommendations. Regarding nutritional indexes, severity of hypo-albuminemia does not further discriminate high-risk patients within the overall hypo-albuminemic ICU population in the multivariable analysis. Although pre-albumin and transferrin have been found to be prognostic for mortality, their turnover rate, half-life, and potential interference by inflammation and anemia substantially limit their universal use [34,35,36]. In the current study, all these traditional biomarkers, except for the NUTRIC score, were not significantly associated with 30-day mortality. In addition, our previous study noted that hypo-albuminemia and hypo-prealbuminemia were also observed in patients with high phenylalanine and leucine levels, and can cause incorrect interpretation of nutrition status, leading to overfeeding [13].

4.4. Clinical Implications of the PLA Score

Based on the scoring, we separated patients into four categories: low, intermediate, high, and very-high risk of 30-day mortality. Mortality risk increased 19.5-fold in patients with very-high risk compared to those with low risk. Based on the grid of phenylalanine and leucine concentrations, patients with very-high risk had a 92-fold increase in mortality compared to the reference patients. These findings were not only present in the initial cohort. In the multicenter cohort recruiting patients from different hospitals located from northern to southern Taiwan, the predictive value of the PLA score was repeatedly demonstrated. In addition to higher mortality risk, a higher PLA score was associated with higher APACHE II, SOFA, and NUTRIC scores, higher CRP levels, but lower cholesterol, albumin, pre-albumin and transferrin levels. In a clinical setting, when poor prognosis was identified by available risk stratification tools and malnutrition was identified by traditional nutritional parameters, PLA scores added to the accuracy of outcome prediction. In addition, since one third of patients in our study were admitted to the ICU due to heart failure and there have been a number of recent publications demonstrating the value of amino acid profiles in heart failure, PLA score is potentially applicable in this population.
The levels of phenylalanine and leucine provide clues for management in patients with critical illness at acute, chronic, and recovery phases as defined by Wischmeyer et al. [6] and Berger et al. [21]. When a leucine level is >165.6 μM, actions to avoid overfeeding should be taken. Along with a phenylalanine level of >72.2 μM or even higher, phenylalanine-free diet, BH4, micronutrients, or anti-oxidants may be considered. When a leucine level is <96.6 μM, protein supplementation should be optimized. However, in patients with high PLA score characterized by low leucine but high phenylalanine levels, phenylalanine-free high-protein diet and anti-oxidant strategies may be an option. Short-term follow-up of PLA score could appropriately fine-tune the nutritional plan. Our findings provide future perspectives to improve the outcomes of patients in ICUs by nutritional interventions based on phenylalanine and leucine.

4.5. Limitations

There are a few limitations in this study. First, the measurement of phenylalanine and leucine depended on a central laboratory to perform UPLC. Inter-laboratory variation should be tested for clinical application if the PLA score is to be used nationwide or worldwide. Second, UPLC is time-consuming and not appropriate for bulk samples. Other convenient methods for measuring amino acids, such as enzymatic assays, should be tested. Third, measurement of multiple amino acids, rather than leucine only, might provide additional nutritional information; however, that would substantially increase the cost and may not be clinically applicable. Finally, whether phenylalanine and leucine-based nutritional interventions can improve the outcomes of patients in ICUs needs further investigation.

5. Conclusions

Based on a simple amino acid profile, PLA scores predicted 30-day mortality independently of traditional risk factors, risk scores such as APACHE II, SOFA, and NUTRIC scores, and nutritional biomarkers. Future clinical intervention trials are warranted to test whether phenylalanine and leucine-based nutritional interventions can improve the outcomes for critically ill patients in the ICU.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/nu15030649/s1, Table S1. Demographic and laboratory data for the initiation and validation cohorts. Table S2. Analysis of the prognostic value of combining phenylalanine and leucine levels to assess 30-day mortality in the initiation cohort. Table S3. Thirty-day mortality prediction value of different risk levels defined by PLA scores.

Author Contributions

Conceptualization, C.-H.W., Y.-L.T., W.-S.C. and M.-H.L.; formal analysis, C.-H.W., Y.-L.T., W.-S.C. and M.-H.L.; investigation, Y.-L.T., H.-P.W., C.-I.C. and I.-C.H.; data curation, H.-P.W. and H.-C.L.; writing–review and editing, C.-H.W., Y.-L.T. and M.-H.L.; funding acquisition, C.-H.W.; validation, J.-J.C., M.-S.T., T.-Y.C., C.-I.C., J.-K.Y. and I.-C.H. All authors have read and agreed to the published version of the manuscript.

Funding

This study was supported in part by the Ministry of Science and Technology of Taiwan (MOST109-2314-B-182-056-MY3) and Chang Gung Memorial Hospital (CORPG2J0023 and G2J0013, and CMRPG2L0141 and G2M0011).

Institutional Review Board Statement

The study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Review Board of Chang Gung Memorial Hospital (IRB approval number: 201507968B0, date of approval: 2015; 201701750B0, date of approval: 2017; 201801514B0, date of approval: 2018; 202000831B0, date of approval: 2020; 202100466B0, date of approval: 2021).

Informed Consent Statement

Informed consent was obtained from all subjects involved in the study.

Data Availability Statement

The datasets used and analyzed during the current study are available from the corresponding author upon reasonable request.

Acknowledgments

The authors thank the Healthy Aging Research Center, Chang Gung University, from the Featured Areas Research Center Program within the Framework of the Higher Education Sprout Project by the Ministry of Education in Taiwan for technical support.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Knaus, W.A.; Draper, E.A.; Wagner, D.P.; Zimmerman, J.E. APACHE II: A severity of disease classification system. Crit. Care Med. 1985, 13, 818–829. [Google Scholar] [CrossRef] [PubMed]
  2. Vincent, J.L.; Moreno, R.; Takala, J.; Willatts, S.; De Mendonça, A.; Bruining, H.; Reinhart, C.K.; Suter, P.M.; Thijs, L.G. The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine. Intensive Care Med. 1996, 22, 707–710. [Google Scholar] [CrossRef] [PubMed]
  3. Heyland, D.K.; Dhaliwal, R.; Jiang, X.; Day, A.G. Identifying critically ill patients who benefit the most from nutrition therapy: The development and initial validation of a novel risk assessment tool. Crit. Care 2011, 15, R268. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  4. Fuchs, L.; Chronaki, C.E.; Park, S.; Novack, V.; Baumfeld, Y.; Scott, D.; McLennan, S.; Talmor, D.; Celi, L. ICU admission characteristics and mortality rates among elderly and very elderly patients. Intensive Care Med. 2012, 38, 1654–1661. [Google Scholar] [CrossRef]
  5. Capuzzo, M.; Volta, C.; Tassinati, T.; Moreno, R.; Valentin, A.; Guidet, B.; Iapichino, G.; Martin, C.; Perneger, T.; Combescure, C.; et al. Working Group on Health Economics of the European Society of Intensive Care Medicine. Hospital mortality of adults admitted to Intensive Care Units in hospitals with and without Intermediate Care Units: A multicentre European cohort study. Crit. Care 2014, 18, 551. [Google Scholar] [CrossRef] [Green Version]
  6. Wischmeyer, P.E. Tailoring nutrition therapy to illness and recovery. Crit. Care 2017, 21, 316. [Google Scholar] [CrossRef]
  7. van Zanten, A.R.H.; De Waele, E.; Wischmeyer, P.E. Nutrition therapy and critical illness: Practical guidance for the ICU, post-ICU, and long-term convalescence phases. Crit. Care. 2019, 23, 368. [Google Scholar] [CrossRef] [Green Version]
  8. Huang, S.S.; Lin, J.Y.; Chen, W.S.; Liu, M.H.; Cheng, C.W.; Cheng, M.L.; Wang, C.H. Phenylalanine- and leucine-defined metabolic types identify high mortality risk in patients with severe infection. Int. J. Infect. Dis. 2019, 85, 143–149. [Google Scholar] [CrossRef] [Green Version]
  9. Luporini, R.L.; Pott-Junior, H.; Di Medeiros Leal, M.C.B.; Castro, A.; Ferreira, A.G.; Cominetti, M.R.; de Freitas Anibal, F. Phenylalanine and COVID-19: Tracking disease severity markers. Int. Immunopharmacol. 2021, 101, 108313. [Google Scholar] [CrossRef] [PubMed]
  10. Atila, A.; Alay, H.; Yaman, M.E.; Akman, T.C.; Cadirci, E.; Bayrak, B.; Celik, S.; Atila, N.E.; Yaganoglu, A.M.; Kadioglu, Y.; et al. The serum amino acid profile in COVID-19. Amino Acids 2021, 53, 1569–1588. [Google Scholar] [CrossRef] [PubMed]
  11. Delles, C.; Rankin, N.J.; Boachie, C.; McConnachie, A.; Ford, I.; Kangas, A.; Soininen, P.; Trompet, S.; Mooijaart, S.P.; Jukema, J.W.; et al. Nuclear magnetic resonance-based metabolomics identifies phenylalanine as a novel predictor of incident heart failure hospitalisation: Results from PROSPER and FINRISK 1997. Eur. J. Heart Fail. 2018, 20, 663–673. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  12. Würtz, P.; Havulinna, A.S.; Soininen, P.; Tynkkynen, T.; Prieto-Merino, D.; Tillin, T.; Ghorbani, A.; Artati, A.; Wang, Q.; Tiainen, M.; et al. Metabolite profiling and cardiovascular event risk: A prospective study of 3 population-based cohorts. Circulation 2015, 131, 774–785. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  13. Chen, W.S.; Wang, C.H.; Cheng, C.W.; Liu, M.H.; Chu, C.M.; Wu, H.P.; Huang, P.C.; Lin, Y.T.; Ko, T.; Chen, W.H.; et al. Elevated plasma phenylalanine predicts mortality in critical patients with heart failure. ESC Heart Fail. 2020, 7, 2884–2893. [Google Scholar] [CrossRef]
  14. Wang, C.H.; Chen, W.S.; Liu, M.H.; Lee, C.Y.; Wang, M.Y.; Liang, C.Y.; Chu, C.M.; Wu, H.P.; Chen, W.H. Stress hyperphenylalaninemia is associated with mortality in cardiac ICU: Clinical factors, genetic variants, and pteridines. Crit. Care Med. 2022, 50, 1577–1587. [Google Scholar] [CrossRef]
  15. Cheng, C.W.; Liu, M.H.; Tang, H.Y.; Cheng, M.L.; Wang, C.H. Factors associated with elevated plasma phenylalanine in patients with heart failure. Amino Acids 2021, 53, 149–157. [Google Scholar] [CrossRef]
  16. Hillert, A.; Anikster, Y.; Belanger-Quintana, A.; Burlina, A.; Burton, B.K.; Carducci, C.; Chiesa, A.E.; Christodoulou, J.; Đorđević, M.; Desviat, L.R.; et al. The genetic landscape and epidemiology of phenylketonuria. Am. J. Hum. Genet. 2020, 107, 234–250. [Google Scholar] [CrossRef] [PubMed]
  17. Wang, M.Y.; Wang, C.H.; Chen, W.S.; Chu, C.M.; Wu, H.P.; Liu, M.H.; Lin, Y.T.; Kao, K.C.; Liang, C.Y.; Chen, W.H.; et al. U-shape relationship between plasma leucine level and mortality in the intensive care unit. Dis. Markers 2022, 2022, 7389258. [Google Scholar] [CrossRef]
  18. Wang, C.H.; Cheng, M.L.; Liu, M.H. Simplified plasma essential amino acid-based profiling provides metabolic information and prognostic value additive to traditional risk factors in heart failure. Amino Acids 2018, 50, 1739–1748. [Google Scholar] [CrossRef]
  19. Han, K.S.; Kim, S.J.; Lee, E.J.; Shin, J.H.; Lee, J.S.; Lee, S.W. Development and validation of new poisoning mortality score system for patients with acute poisoning at the emergency department. Crit. Care 2021, 25, 29. [Google Scholar] [CrossRef]
  20. Gatti, G.; Perrotti, A.; Obadia, J.F.; Duval, X.; Iung, B.; Alla, F.; Chirouze, C.; Selton-Suty, C.; Hoen, B.; Sinagra, G.; et al. Simple scoring system to predict in-hospital mortality after surgery for infective endocarditis. J. Am. Heart Assoc. 2017, 6, e004806. [Google Scholar] [CrossRef]
  21. Berger, M.M.; Pantet, O.; Schneider, A.; Ben-Hamouda, N. Micronutrient deficiencies in medical and surgical inpatients. J. Clin. Med. 2019, 8, 931. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  22. Bendall, J.K.; Douglas, G.; McNeill, E.; Channon, K.M.; Crabtree, M.J. Tetrahydrobiopterin in cardiovascular health and disease. Antioxid Redox Signal 2014, 20, 3040–3077. [Google Scholar] [CrossRef] [Green Version]
  23. Nishijima, Y.; Sridhar, A.; Bonilla, I.; Velayutham, M.; Khan, M.; Terentyeva, R.; Li, C.; Kuppusamy, P.; Elton, T.S.; Terentyev, D.; et al. Tetrahydrobiopterin depletion and NOS2 uncoupling contribute to heart failure-induced alterations in atrial electrophysiology. Cardiovasc. Res. 2011, 91, 71–79. [Google Scholar] [CrossRef] [Green Version]
  24. Gudivada, K.K.; Kumar, A.; Shariff, M.; Sampath, S.; Varma, M.M.; Sivakoti, S.; Krishna, B. Antioxidant micronutrient supplementation in critically ill adults: A systematic review with meta-analysis and trial sequential analysis. Clin. Nutr. 2021, 40, 740–750. [Google Scholar] [CrossRef] [PubMed]
  25. Xu, J.; Pan, T.; Qi, X.; Tan, R.; Wang, X.; Liu, Z.; Tao, Z.; Qu, H.; Zhang, Y.; Chen, H.; et al. Increased mortality of acute respiratory distress syndrome was associated with high levels of plasma phenylalanine. Respir. Res. 2020, 21, 99. [Google Scholar] [CrossRef] [PubMed]
  26. Czibik, G.; Mezdari, Z.; Murat Altintas, D.; Bréhat, J.; Pini, M.; d’Humières, T.; Delmont, T.; Radu, C.; Breau, M.; Liang, H.; et al. Dysregulated phenylalanine catabolism plays a key role in the trajectory of cardiac aging. Circulation 2021, 144, 559–574. [Google Scholar] [CrossRef] [PubMed]
  27. Hadjiconstantinou, M.; Neff, N.H. Enhancing aromatic L-amino acid decarboxylase activity: Implications for L-DOPA treatment in Parkinson’s disease. CNS Neurosci. Ther. 2008, 14, 340–351. [Google Scholar] [CrossRef]
  28. Williams, R.A.; Mamotte, C.D.; Burnett, J.R. Phenylketonuria: An inborn error of phenylalanine metabolism. Clin. Biochem. Rev. 2008, 29, 31–41. [Google Scholar]
  29. Fedotcheva, N.I.; Kazakov, R.E.; Kondrashova, M.N.; Beloborodova, N.V. Toxic effects of microbial phenolic acids on the functions of mitochondria. Toxicol. Lett. 2008, 180, 182–188. [Google Scholar] [CrossRef]
  30. Cronin, S.J.F.; Seehus, C.; Weidinger, A.; Talbot, S.; Reissig, S.; Seifert, M.; Pierson, Y.; McNeill, E.; Longhi, M.S.; Turnes, B.L.; et al. The metabolite BH4 controls T cell proliferation in autoimmunity and cancer. Nature 2018, 563, 564–568. [Google Scholar] [CrossRef]
  31. Oshima, T.; Deutz, N.E.; Doig, G.; Wischmeyer, P.E.; Pichard, C. Protein-energy nutrition in the ICU is the power couple: A hypothesis forming analysis. Clin. Nutr. 2016, 35, 968–974. [Google Scholar] [CrossRef]
  32. McClave, S.A.; Taylor, B.E.; Martindale, R.G.; Warren, M.M.; Johnson, D.R.; Braunschweig, C.; McCarthy, M.S.; Davanos, E.; Rice, T.W.; Cresci, G.A.; et al. Guidelines for the provision and assessment of nutrition support therapy in the adult critically ill patient: Society of Critical Care Medicine (SCCM) and American Society for Parenteral and Enteral Nutrition (A.S.P.E.N.). JPEN J. Parenter. Enteral. Nutr. 2016, 40, 159–211. [Google Scholar] [CrossRef] [PubMed]
  33. Cheng, M.L.; Wang, C.H.; Shiao, M.S.; Liu, M.H.; Huang, Y.Y.; Huang, C.Y.; Mao, C.T.; Lin, J.F.; Ho, H.Y.; Yang, N.I. Metabolic disturbances identified in plasma are associated with outcomes in patients with heart failure: Diagnostic and prognostic value of metabolomics. J. Am. Coll. Cardiol. 2015, 65, 1509–1520. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  34. Ferrie, S.; Tsang, E. Monitoring nutrition in critical illness: What can we use? Nutr. Clin. Pract. 2018, 33, 133–146. [Google Scholar] [CrossRef] [PubMed]
  35. Geisler, J.P.; Linnemeier, G.C.; Thomas, A.J.; Manahan, K.J. Nutritional assessment using prealbumin as an objective criterion to determine whom should not undergo primary radical cytoreductive surgery for ovarian cancer. Gynecol. Oncol. 2007, 106, 128–131. [Google Scholar] [CrossRef] [PubMed]
  36. Guerra, L.T.; Rosa, A.R.; Romani, R.F.; Gurski, R.R.; Schirmer, C.C.; Kruel, C.D. Serum transferrin and serum prealbumin as markers of response to nutritional support in patient. Nutr. Hosp. 2009, 24, 241–242. [Google Scholar] [PubMed]
Nutrients 15 00649 g001 550
Figure 1. Flow diagram of the study.
Figure 1. Flow diagram of the study.
Nutrients 15 00649 g001
Nutrients 15 00649 g002 550
Figure 2. Prognostic value of phenylalanine leucine amino acid (PLA) scores. (A) The Kaplan–Meier curves for the four risk levels in the initiation cohort (for all-cause 30-day mortality). (B) Phenylalanine and leucine grid reflects the risk of 30-day mortality by intensity of coloring (green: low risk, yellow: intermediate risk, orange: high risk, red: very-high risk). Each cell represents the relative risk compared to the reference cell. “Blank” indicates the condition is not present in our study. (C) The Kaplan–Meier curves for the four risk levels in the validation cohort (for all-cause 30-day mortality). (D) The ROC curves for PLA, APACHE II, and NUTRIC scores. Abbreviations: Phe, phenylalanine; Leu, leucine; PLA, phenylalanine leucine amino acid score; APACHE II score, Acute Physiology and Chronic Health Evaluation II score; NUTRIC score, Nutrition Risk in the Critically Ill score; ROC curve, receiver operating characteristic curve.
Figure 2. Prognostic value of phenylalanine leucine amino acid (PLA) scores. (A) The Kaplan–Meier curves for the four risk levels in the initiation cohort (for all-cause 30-day mortality). (B) Phenylalanine and leucine grid reflects the risk of 30-day mortality by intensity of coloring (green: low risk, yellow: intermediate risk, orange: high risk, red: very-high risk). Each cell represents the relative risk compared to the reference cell. “Blank” indicates the condition is not present in our study. (C) The Kaplan–Meier curves for the four risk levels in the validation cohort (for all-cause 30-day mortality). (D) The ROC curves for PLA, APACHE II, and NUTRIC scores. Abbreviations: Phe, phenylalanine; Leu, leucine; PLA, phenylalanine leucine amino acid score; APACHE II score, Acute Physiology and Chronic Health Evaluation II score; NUTRIC score, Nutrition Risk in the Critically Ill score; ROC curve, receiver operating characteristic curve.
Nutrients 15 00649 g002
Table
Table 1. Demographic and laboratory data for the initiation cohort.
Table 1. Demographic and laboratory data for the initiation cohort.
AllSurvivalDeath
n = 537n = 387n = 150p Value
Age (years)71.5 ± 13.671.4 ± 14.072.0 ± 12.80.659
Male (%)323 (60.1)222 (57.4)101 (67.3)0.034
APACHE II score18.8 ± 5.817.7 ± 5.221.8 ± 6.3<0.001
SOFA score7.1 ± 3.36.4 ± 3.09.0 ± 3.5<0.001
NUTRIC score5.3 ± 1.95.0 ± 1.96.2 ± 1.8<0.001
Co-morbidity
    Diabetes mellitus (%)256 (47.7)185 (47.8)71 (47.3)0.922
    Hypertension (%)350 (65.2)249 (64.3)101 (67.3)0.514
    Coronary artery disease (%)231 (43.0)172 (44.4)59 (39.3)0.283
    Atrial fibrillation (%)80 (14.9)50 (12.9)30 (20.0)0.039
    Chronic kidney disease (%) *156 (29.1)114 (29.5)42 (28.0)0.739
Ventilator use (%)415 (77.3)282 (72.9)133 (88.7)<0.001
Inotropic agent use (%)209 (38.9)124 (32.0)85 (56.7)<0.001
Days in ICU (days)11.2 ± 7.411.2 ± 7.411.2 ± 7.20.957
Laboratory data
    Hemoglobin (g/dL)10.8 ± 2.810.9 ± 2.810.6 ± 2.80.247
    eGFR (mL/min/1.73 m2)45.4 ± 43.647.9 ± 45.739.0 ± 37.10.020
    C-reactive protein (mg/L)46.8 (11.1–104)36.7 (8.1–93.0)61.7 (24.7–146)<0.001
    Cholesterol (mg/dL)132.0 ± 52.9138.9 ± 54.8114.4 ± 43.2<0.001
    Triglyceride (mg/dL)109 (81.5–152)108 (79–150)115 (84–158)0.270
    Albumin (g/dL)3.19 ± 0.623.28 ± 0.592.95 ± 0.65<0.001
    Pre-Albumin (mg/dL)15.1 ± 8.216.3 ± 8.712.0 ± 5.7<0.001
    Transferrin (mg/dL)154.2 ± 49.9162.3 ± 49.1133.1 ± 45.9<0.001
Data are expressed as the mean ± SD for variables with normal distribution, as the median [interquartile range (IQR)] for variables with skewed distribution, and as a number (percentage) for categorical variables. * Chronic kidney disease is defined as eGFR < 60 mL/min/1.73 m2. Abbreviations: APACHE II score, Acute Physiology and Chronic Health Evaluation II score; SOFA score, Sequential Organ Failure Assessment score; NUTRIC score, Nutrition Risk in the Critically Ill score; ICU, intensive care unit; eGFR, estimated glomerular filtration rate.
Table
Table 2. Multivariable Cox logistic regression for the PLA score calculation to predict 30-day mortality.
Table 2. Multivariable Cox logistic regression for the PLA score calculation to predict 30-day mortality.
B *Points = B/0.485 **HR (95% CI)p Value
Phenylalanine (μM)
    <72.2Reference01
    72.2–<88.50.8431.72.32 (1.26–4.27)0.007
    88.5–<1121.5563.24.74 (2.65–8.48)<0.001
    112–<1301.7933.76.01 (3.02–11.94)<0.001
    130–<1502.0964.38.13 (3.73–17.74)<0.001
    150–<2002.403511.06 (5.05–24.21)<0.001
    200–<2803.418730.50 (14.07–66.15)<0.001
    ≥2804.5019.390.13 (38.62–210.37)<0.001
Leucine (μM)
    ≥96.6Reference01
    68.9–<96.60.48511.62 (1.08–2.44)0.020
    50.0–<68.91.3592.83.89 (2.14–7.09)<0.001
    <501.6333.45.12 (2.12–12.36)<0.001
* B, regression coefficient. ** Base constant B was the smallest regression coefficient in the model, which was 0.485. Abbreviations: PLA score, phenylalanine leucine amino acid score; HR, hazard ratio; CI, confidence interval.
Table
Table 3. Characteristics of patients with different PLA scores in the initiation cohort.
Table 3. Characteristics of patients with different PLA scores in the initiation cohort.
PLA Scorep for Trend
0–11.1–44.1–5>5
Variablen = 122n = 303n = 68n = 44
Age (years)69.8 ± 14.972.0 ± 13.173.0 ± 13.471.0 ± 14.10.529
Male (%)64 (52.5)192 (63.4)38 (55.9)29 (65.9)0.217
APACHE II score17.5 ± 5.618.3 ± 5.521.2 ± 5.522.6 ± 6.7<0.001
SOFA score5.5 ± 2.97.1 ± 3.08.5 ± 3.710.0 ± 3.5<0.001
NUTRIC score4.7 ± 2.05.3 ± 1.85.9 ± 2.16.5 ± 1.7<0.001
Co-morbidity
    Diabetes mellitus (%)61 (50.0)143 (47.2)33 (48.5)19 (43.2)0.515
    Hypertension (%)71 (58.2)210 (69.3)42 (61.8)27 (61.4)0.755
    Coronary artery disease (%)52 (42.6)137 (45.2)25 (36.8)17 (38.6)0.420
    Atrial fibrillation15 (12.3)39 (12.9)14 (20.6)12 (27.3)0.008
    Chronic kidney disease (%) *34 (27.9)91 (30.0)16 (23.5)15 (34.1)0.807
Ventilator use (%)80 (65.6)240 (79.2)57 (83.8)38 (86.4)0.001
Inotropic agent use (%)31 (25.4)112 (37.0)36 (52.9)30 (68.2)<0.001
Days in ICU (day)10.2 ± 6.711.8 ± 7.612.1 ± 7.18.9 ± 7.10.379
Laboratory data
    Hemoglobin (g/dL)10.8 ± 2.410.9 ± 2.910.8 ± 2.89.7 ± 3.10.025
    eGFR (ml/min/1.73 m2)59.6 ± 54.843.4 ± 40.638.8 ± 34.529.8 ± 29.6<0.001
    C-reactive protein (mg/L)34 (7–67)47 (12–100)73 (18–153)51 (16–148)0.010
    Cholesterol (mg/dL)144.5 ± 66.8134.6 ± 47.0121.0 ± 43.694.7 ± 40.8<0.001
    Triglyceride (mg/dL)112 (77–153)112 (84–153)115 (82–156)87 (68–121)0.104
    Albumin (g/dL)3.26 ± 0.563.25 ± 0.612.90 ± 0.632.97 ± 0.75<0.001
    Pre-Albumin (mg/dL)17.1 ± 6.915.4 ± 6.911.7 ± 6.210.2 ± 5.8<0.001
    Transferrin (mg/dL)160.7 ± 44.4159.5 ± 50.0138.1 ± 49.0125.9 ± 48.4<0.001
Data are expressed as the mean ± SD for variables with normal distribution, as the median [interquartile range (IQR)] for variables with skewed distribution, and as a number (percentage) for categorical variables. * Chronic kidney disease is defined as eGFR < 60 mL/min/1.73 m2. Abbreviations: PLA score, phenylalanine leucine amino acid score; APACHE II score, Acute Physiology and Chronic Health Evaluation II score; SOFA score, Sequential Organ Failure Assessment score; NUTRIC score, Nutrition Risk in the Critically Ill score; ICU, intensive care unit; eGFR, estimated glomerular filtration rate.
Table
Table 4. Cox univariate and multivariable analysis for predicting 30-day mortality in the initiation cohort.
Table 4. Cox univariate and multivariable analysis for predicting 30-day mortality in the initiation cohort.
UnivariateMultivariable (Model 1) *Multivariable (Model 2) †
HR (95% CI)p ValueHR (95% CI)p ValueHR (95% CI)p Value
PLA score1.62 (1.51–1.75)<0.0011.46 (1.33–1.60)<0.0011.48 (1.35–1.62)<0.001
APACHE II score1.12 (1.09–1.15)<0.0011.07 (1.03–1.11)<0.001
SOFA score1.20 (1.15–1.26)<0.0011.04 (0.98–1.11)0.061
NUTRIC score1.34 (1.23–1.46)<0.001 1.18 (1.07–1.30)0.001
Age (years)1.00 (0.99–1.01)0.728
Sex (male)1.43 (1.02–2.01)0.0401.37 (0.96–1.96)0.0861.33 (0.93–1.90)0.118
Atrial fibrillation1.50 (1.01–2.24)0.0461.06 (0.69–1.62)0.7930.95 (0.61–1.47)0.816
C-reactive protein (log)1.78 (1.36–2.32)<0.0011.29 (0.96–1.72)0.0871.29 (0.96–1.74)0.086
eGFR (mL/min/1.73 m2)0.99 (0.99–1.00)0.0341.00 (0.99–1.01)0.7180.99 (0.99–1.00)0.804
Cholesterol (mg/dL)0.99 (0.98–0.99)<0.0010.99 (0.99–1.01)0.5510.99 (0.99–1.00)0.261
Albumin (g/dL)0.48 (0.37–0.62)<0.0010.88 (0.64–1.20)0.4110.85 (0.62–1.15)0.284
Pre-Albumin (mg/dL)0.92 (0.89–0.94)<0.0010.99 (0.96–1.02)0.6260.99 (0.98–1.02)0.937
Transferrin (mg/dL)0.99 (0.98–0.99)<0.0010.99 (0.99–1.01)0.2770.99 (0.99–1.00)0.117
* Model 1 includes all significant variables in the univariate analysis, except NUTRIC score; † Model 2 includes all significant variables in the univariate analysis, except APACHE II and SOFA scores. Abbreviations: HR, hazard ratio; CI, confidence interval; PLA score, phenylalanine leucine amino acid score; APACHE II score, Acute Physiology and Chronic Health Evaluation II score; SOFA score, Sequential Organ Failure Assessment score; NUTRIC score, Nutrition Risk in the Critically Ill score; eGFR, estimated glomerular filtration rate.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Tsou, Y.-L.; Wang, C.-H.; Chen, W.-S.; Wu, H.-P.; Liu, M.-H.; Lin, H.-C.; Chang, J.-J.; Tsai, M.-S.; Chen, T.-Y.; Cheng, C.-I.; et al. Combining Phenylalanine and Leucine Levels Predicts 30-Day Mortality in Critically Ill Patients Better than Traditional Risk Factors with Multicenter Validation. Nutrients 2023, 15, 649. https://doi.org/10.3390/nu15030649

AMA Style

Tsou Y-L, Wang C-H, Chen W-S, Wu H-P, Liu M-H, Lin H-C, Chang J-J, Tsai M-S, Chen T-Y, Cheng C-I, et al. Combining Phenylalanine and Leucine Levels Predicts 30-Day Mortality in Critically Ill Patients Better than Traditional Risk Factors with Multicenter Validation. Nutrients. 2023; 15(3):649. https://doi.org/10.3390/nu15030649

Chicago/Turabian Style

Tsou, Yi-Liang, Chao-Hung Wang, Wei-Siang Chen, Huang-Ping Wu, Min-Hui Liu, Hsuan-Ching Lin, Jung-Jung Chang, Meng-Shu Tsai, Tien-Yu Chen, Cheng-I Cheng, and et al. 2023. "Combining Phenylalanine and Leucine Levels Predicts 30-Day Mortality in Critically Ill Patients Better than Traditional Risk Factors with Multicenter Validation" Nutrients 15, no. 3: 649. https://doi.org/10.3390/nu15030649

APA Style

Tsou, Y. -L., Wang, C. -H., Chen, W. -S., Wu, H. -P., Liu, M. -H., Lin, H. -C., Chang, J. -J., Tsai, M. -S., Chen, T. -Y., Cheng, C. -I., Yeh, J. -K., & Hsieh, I. -C. (2023). Combining Phenylalanine and Leucine Levels Predicts 30-Day Mortality in Critically Ill Patients Better than Traditional Risk Factors with Multicenter Validation. Nutrients, 15(3), 649. https://doi.org/10.3390/nu15030649

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop