Predictive Value of Serum β2 ‐ Microglobulin for Outcomes in Patients with Acute Respiratory Distress Syndrome Caused by Bacterial Infection

Acute respiratory distress syndrome (ARDS) is a heterogeneous disease with extremely high mortality. We hypothesized that the serum β2-microglobulin (β2MG) level would be elevated and be an independent risk factor for 28-day mortality in patients with ARDS. We who from 1, 2015 to 28, 2021. Patients were followed for up to 28 days from diagnosis and were divided into a survival group and non-survival group according to their clinical outcomes. The serum β2MG levels and other clinical data were collected. The relationship between serum β2MG levels and 28-day mortality was explored by performing a Cox regression analysis adjusted for age, updated Charlson comorbidity index, disorders of consciousness, septic shock, albumin level, cardiac troponin I level, procalcitonin level, lactic acid level, prothrombin time, and partial pressure of arterial oxygen/fraction of inspired oxygen ratio. Data are the mean ± SD, median (IQR), or n (%). P values comparing the Survivor and non-Survivor groups are from a 2-sample t-test, Mann-Whitney U test, or χ² test. Differences with values of P < 0.05 were considered statistically signicant. non-Survivor subcategories. mmHg; 100 mmHg PaO mechanical Failure interquartile A Cox proportional hazards analysis was performed. Data are the HR (95% CI). Adjusted for age, updated CCI, disorders of consciousness, septic shock, albumin, cTnI, procalcitonin, prothrombin time, lactic acid, and PaO 2 /FiO 2 ratio. Differences with values of P < 0.05 were considered statistically signicant. syndrome; β2MG, β2-microglobulin; CCI, Charlson comorbidity index; cTnI cardiac troponin I; PCT, procalcitonin; PT, prothrombin time; PaO 2 , partial pressure of arterial oxygen; FiO 2 , fraction of inspired oxygen; BMI, body mass index; Scr, serum creatinine; Ccr, creatinine clearance rate; BUN, blood urea nitrogen; AKI, acute kidney injury; TBIL, total bilirubin; ALT, alanine aminotransferase; NT-proBNP, N-terminal pro-brain natriuretic peptide; FPG, WBC, white blood cell; CRP, C-reactive protein; APACHE, Acute Physiology and Chronic Health Evaluation; SOFA, Sequential Organ Failure Assessment; HR, hazard ratio; CI, A Cox proportional hazards analysis was performed. Data are the HR (95% CI). Adjusted for age, updated CCI, disorders of consciousness, septic shock, albumin, cTnI, procalcitonin, prothrombin time, lactic acid, and PaO 2 /FiO 2 ratio. Differences with values of P < 0.05 were considered statistically signicant. ARDS, acute respiratory disease syndrome; β2MG, β2-microglobulin; CCI, Charlson comorbidity index; cTnI cardiac troponin I; PCT, procalcitonin; PT, prothrombin time; PaO 2 , partial pressure of arterial oxygen; FiO 2 , fraction of inspired oxygen; BMI, body mass index; Scr, serum creatinine; Ccr, creatinine clearance rate; BUN, blood urea nitrogen; AKI, acute kidney injury; TBIL, total bilirubin; ALT, alanine aminotransferase; NT-proBNP, N-terminal pro-brain natriuretic peptide; FPG, fasting plasma glucose; WBC, white blood cell; CRP, C-reactive protein; Acute and Chronic Health Evaluation; SOFA, Sequential Organ Failure Assessment; HR, hazard ratio; CI, condence , , APACHE,


Statistical Analyses
Categorical variables are described as numbers and percentages (%), and continuous variables are described as the mean and standard deviation (SD) or the median and interquartile range (IQR). The Shapiro-Wilk test was used to verify normality. Differences between the survival and non-survival groups were assessed by the two-sample t-test for normally distributed continuous variables, the Mann-Whitney U test for non-normally distributed continuous variables, or the χ 2 test for categorical variables. A Spearman rank correlation analysis was used to analyse the correlation between β2MG levels and other basic variables. Both univariate and multivariate Cox regression analyses were applied to evaluate the relationship between risk factors and 28-day mortality. Results are presented as hazard ratios (HRs) with 95% con dence intervals (CIs). Serum β2MG levels were adjusted for age, updated CCI, disturbance of consciousness, septic shock, serum albumin level, cTnI level, PCT level, PT, lactic acid level, and PaO 2 /FiO 2 ratio in the multivariate Cox regression analysis. Survival rates grouped by β2MG tertile are presented as cumulative survival curves adjusted for the above-mentioned variables. Receiver operating characteristic (ROC) analyses were performed to calculate the sensitivity and speci city of risk factors for predicting 28-day mortality. The areas under receiver operating curves (ROC-AUCs) for different risk factors were compared using the method of DeLong et al. (1988) by MedCalc. All other statistical analyses were performed using SPSS version 21.0 (Statistical Package for the Social Sciences, Chicago, IL USA). All tests were two-tailed; differences with a value of P < 0.05 were considered statistically signi cant.

Patient Enrolment
A total of 257 patients with ARDS were included in this study. A flow chart of patient enrolment and outcomes is shown in Fig. 1. Table 1 shows the demographic and clinical characteristics of the enrolled patients. In this cohort, 161 patients survived and 96 patients died within 28 days after the diagnosis of ARDS, yielding a 28-day mortality rate of 37.4%. The median level of β2MG for all patients with ARDS, regardless of their 28-day survival, was 4.6 (IQR: 2.9 -8.5) mg/L. Compared with the non-survivors, the survivors were younger (P = 0.002) and had lower updated CCIs (P < 0.001), lower serum β2MG levels (P < 0.001), lower Scr levels (P < 0.001), lower BUN levels (P < 0.001), lower cTnI levels (P = 0.001), lower NT-proBNP levels (P < 0.001), lower PCT levels (P = 0.008), lower lactic acid levels (P < 0.001), shorter PTs (P = 0.001), higher albumin levels (P < 0.001), and higher PaO 2 /FiO 2 ratios (P < 0.001). Survivors also had lower APACHE II scores (P < 0.001) and lower SOFA scores (P < 0.001) than did non-survivors. More non-survivors than survivors had AKI (P < 0.001) or acute myocardial injury (P = 0.003). There was no difference in the duration of mechanical ventilation (MV) between survivors and non-survivors (P = 0.959).

Univariate and Multivariate Survival Analyses
A univariate Cox regression analysis revealed that the level of serum β2MG is a predictor of 28-day mortality in patients with ARDS (HR: 1.096; 95% CI: 1.064 -1.128; P < 0.001) ( Table 2). Other predictors of 28-day mortality in these patients included age, updated CCI, septic shock, Scr level, BUN level, Ccr, albumin level, PCT level, lactic acid level, PT, PaO 2 /FiO 2 ratio, APACHE II score, and SOFA score (P < 0.05 for each).
To reduce data duplication, we did not include the APACHE II score or the SOFA score in our multivariate Cox proportional hazards analysis. Additionally, because we found that the serum β2MG level was positively correlated with the Scr level (Spearman correlation coe cient: 0.815), the BUN level (Spearman correlation coe cient: 0.723), AKI (Spearman correlation coe cient: 0.683), and the NT-proBNP level (Spearman correlation coe cient: 0.564) and was negatively correlated with the Ccr (Spearman correlation coe cient: -0.811) in our cohort (P < 0.001 for each) ( Table 3), we did not include these variables in the multivariate survival analysis either. The metabolism of NT-proBNP is in uenced by renal function, and therefore we further analysed the correlation between the serum β2MG level and NT-proBNP level in patients with a Ccr of > 60 mL/min and found a positive correlation between them (Spearman correlation coe cient: 0.338; P < 0.001). Of the 257 patients in this study, 173 (67.3%) had a CRP level equal to 120 mg/mL (the upper limit value for CRP testing in our laboratory at that time), and therefore we further analysed the 84 patients with a CRP level of <120 mg/mL and found a signi cant positive correlation between the serum β2MG level and CRP level (Spearman correlation coe cient: 0.562; P < 0.001).

Discussion
Our study observed that the levels of β2MG in patients with ARDS were elevated and were signi cantly higher in non-survivors than in survivors. A multivariate Cox proportional hazards analysis revealed that the β2MG level is an independent predictor for 28-day mortality in patients with ARDS, after adjusting for age, updated CCI, disorders of consciousness, septic shock, albumin level, cTnI level, PCT level, PT, lactic acid level, and PaO 2 /FiO 2 ratio. To our knowledge, this is the rst report suggesting that the serum β2MG level might have a predictive value for the outcomes of patients with ARDS.
As a low molecular-weight protein, β2MG is an ideal endogenous biomarker for estimating the glomerular ltration rate and AKI [8,12] and is also associated with a number of clinical states. Several previous studies have shown that levels of serum β2MG are higher in patients with in ammatory bowel disease or systemic lupus erythematosus than in healthy controls, suggesting that it might also be a useful biomarker for the assessment of these autoimmune diseases [25,26]. In addition, elevated levels of serum β2MG have been also observed in patients with haemato-oncological pathology and solid tumours despite their preserved renal function [15,27]. Some research has suggested that β2MG is probably a general biomarker that re ects acute or chronic changes during in ammation and infection [28,29]. Levels of serum β2MG are independently associated with major cardiovascular events in the general population as well as in patients with asymptomatic carotid atherosclerosis, patients with isolated systolic hypertension, and patients with acute heart failure who do not have severe renal insu ciency [10,[30][31][32]. Mao et al. reported that β2MG levels are associated with poor outcomes in patients with exacerbated chronic obstructive pulmonary disease [18].
ARDS is a clinical syndrome with extremely high mortality, characterized by severe hypoxemia and an overwhelming in ammatory response, accompanied by multiple organ dysfunctions. Kohanpour et al. These ndings suggest that elevated serum β2MG levels are associated with hypoxemia. Some studies have found that an increase in serum β2MG levels is also present during infectious diseases as well as in in ammatory responses [28,29]. In our study, the serum β2MG levels were found to be elevated in patients with bacterial infection-induced ARDS. A rank correlation analysis revealed that the serum β2MG levels were negatively correlated with the PaO 2 /FiO 2 ratio and positively correlated with the PCT and CRP levels. These correlations suggest that elevated serum β2MG levels during ARDS may be associated with hypoxemia and infection as well as with in ammation. In addition, ARDS is often accompanied by multiple organ dysfunctions, such as AKI and myocardial injury, which are signi cantly associated with a poor prognosis in patients with ARDS [35,36]. Previous studies have shown that serum β2MG levels are correlated with renal injury [8,12] as well as with cardiac function [32, 37, 38]. Both AKI and myocardial injury were present in approximately half of the patients in our cohort, and our correlation analysis revealed that the serum β2MG levels are positively correlated with AKI, NT-proBNP levels, and cTnI levels and are negatively correlated with the Ccr and left ventricular ejection fraction. An analysis conducted after strati cation of the patients according to their Ccrs [39] showed that serum β2MG levels are also positively correlated with NT-proBNP levels in patients with a Ccr of > 60 mL/min. These conditions suggest that, in the case of ARDS caused by bacterial infection, severe hypoxemia, infection, and in ammatory responses, as well as the impairment of organ function, result in increased β2MG production and decreased renal ltration, ultimately leading to elevated serum β2MG levels, which are positively correlated with disease severity and sensitively predict an elevated risk of death.
The ROC curves generated from our data show that the predictive value of serum β2MG levels for patient outcome is superior to that of AKI and Ccr, which may additionally illustrate that the correlation between serum β2MG levels and mortality is not solely a consequence of renal impairment. Further comparison of these ROC curves showed that serum β2MG levels are not inferior to currently applied critical illness scores, such as the APACHE II score and SOFA score, for predicting 28-day mortality in patients with ARDS caused by bacterial infection. Therefore, the serum β2MG level may be an ideal screening tool that can be reliably and cost effectively measured.
We applied the updated CCI [19] to assess patient comorbidities, including coronary heart disease, congestive heart failure, cerebrovascular disease, diabetes mellitus, dementia, connective tissue disease, liver disease, and kidney disease, and found it was signi cantly associated with 28-day mortality in patients with bacterial infection-induced ARDS. This result is consistent with previous research [40]. We also found that the lactic acid level was an independent risk factor for 28-day mortality in patients with bacterial infection-induced ARDS. Lactic acid is directly produced by anaerobic glucose metabolism and is the product of pyruvate transformation through glycolysis. An acceleration of lactate synthesis may be observed under conditions of increased glucose uptake from circulation, of increased glycogenolysis and glycolysis owing to enhanced epinephrine secretion, of inhibition of pyruvate dehydrogenase or of glycogen synthesis during sepsis and, nally, during tissue hypoxia. Therefore, the lactic acid level is considered to be a sensitive biomarker, which can re ect the oxygen supply in cells and the perfusion of surrounding tissues in the early stage of disease, and can be used to assess disease severity and to predict the occurrence of and death risk from septic shock and multiple organ dysfunction syndrome [21,41]. Demirel reported that the lactate level is a good predictor of in-hospital mortality in pneumonia cases [42]. During ARDS, owing to severe hypoxemia and varying degrees of tissue perfusion insu ciency, glucose anaerobic metabolism increases, resulting in increased lactic acid production, which indicates a poor prognosis [5,6].
Severe hypoxemia is a characteristic manifestation of ARDS. The PaO 2 /FiO 2 ratio is an integral part of the assessment of patients with ARDS and is an important criterion for severity grading in the Berlin standard [1]. Although some studies have found that the PaO 2 /FiO 2 ratio is not a good prognostic factor for ARDS [4], as an important indicator of the severity of lung injury, however, most studies have shown that the decreased PaO 2 /FiO 2 ratio is associated with increased mortality or failure of non-invasive MV in patients with ARDS [21,43,44]. Similarly, our study showed that the PaO 2 /FiO 2 ratio was a protective factor for the prognosis of patients with ARDS.
There are some limitations to our study. First of all, this research was conducted in a single centre, which could have biased its results. Second, owing to the small sample size in this study, to avoid over tting, only a limited number of clinical variables were entered into the Cox regression analysis, and it is possible that potentially relevant variables were not evaluated. Third, owing to the retrospective nature of this study, we could not simultaneously assay the levels of β2MG in the urine, and thus it was not possible to determine how much of the increase in serum β2MG levels can be attributed to renal injury and how much of the increased production is a consequence of the disease state. Fourth, although the ROC curve in our study showed predictive value for the outcome of patients with ARDS, we could not verify the applicability of this clinical indicator because of the small sample size. Future prospective studies will be necessary to identify and verify the prognostic value of serum β2MG levels in patients with ARDS caused by bacterial infection.

Conclusions
This prospective study showed that the level of serum β2MG, measured within 24 hours after the diagnosis of ARDS, was elevated and may be a promising early biomarker of adverse outcomes in patients with ARDS caused by bacterial infection. Further prospective research will be necessary to verify this nding, which may help clinicians undertake timely and effective programmes to improve the outcomes of these patients.

Abbreviations
ARDS, acute respiratory distress syndrome; β2MG, β2-microglobulin; AKI, acute kidney injury; Scr, serum creatinine; BUN, blood urea nitrogen; TBIL, total bilirubin; ALT, alanine aminotransferase; FPG, fasting plasma glucose; NT-proBNP, N-terminal pro-brain natriuretic peptide; cTnI, cardiac troponin I; WBC, white blood cell; CRP, C-reactive protein; PCT, procalcitonin; PT, prothrombin time; BMI, body mass index; Ccr, creatinine clearance rate; PaO 2 , partial pressure of arterial oxygen; FiO 2 , fraction of inspired oxygen; PEEP, positive end-expiratory pressure; CPAP, continuous positive airway pressure; CCI, Charlson comorbidity index; APACHE, Acute Physiology and Chronic Health Evaluation; SOFA, Sequential Organ Failure Assessment; MV, mechanical ventilation; SD, standard deviation; IQR, interquartile range; HR, hazard ratio; CI, con dence interval; ROC, receiver operating characteristic. Declarations NC, JW and LMZ contributed to the conception and design of the study. LMZ and JW took part in managing the research. XKF and CGJ contributed to the acquisition of data. All authors were involved in data analysis and interpretation and development of the manuscript. All authors read and approved the nal manuscript. LMZ and JW contributed equally to this article and shared corresponding authorship.

Funding
This research was supported by the multicenter clinical research veri cation of nasal high ow humidi cation oxygen therapy equipment (Grant NO. 2019YFC0121704 to Dr Jing Wang).

Availability of data and materials
All data analysed during the study are presented in the main manuscript. The anonymous dataset is available from the corresponding author upon reasonable request.

Ethics and approval and consent to participate
This retrospective study involving human participants was approved by the ethics committee of the Beijing Chao-Yang Hospital, Capital Medical University (2020-ke-429) and was in accordance with 1964 Helsinki Declaration and its later amendments or comparable ethical standards.

Consent for publication
Not applicable.

Figure 3
Prediction of 28-day mortality in patients with ARDS. β2MG showed a diagnostic accuracy for mortality screening that is superior to AKI (P = 0.001) and Ccr (P = 0.032) and not inferior to the APACHE II Score (P = 0.153) and SOFA Score (P = 0.114). * Comparing between β2MG and other variables. Abbreviations: ARDS, acute respiratory disease syndrome; ROC, receiver operating characteristic; AUC, area under curve; β2MG, β2-microglobulin; AKI, acute kidney injury; Ccr, creatinine clearance rate; APACHE, Acute Physiology and Chronic Health Evaluation; SOFA, Sequential Organ Failure Assessment.