Prognostic Value of Serum Cholinesterase Levels for In-Hospital Mortality among Patients with Acute Exacerbation of Chronic Obstructive Pulmonary Disease

Abstract

Cholinesterase (ChE) is associated with the pathogenesis of chronic obstructive pulmonary disease (COPD), including chronic airway inflammation and oxidation/antioxidant imbalance. However, the relationship between serum ChE levels and survival outcomes of patients hospitalized with acute exacerbations of COPD (AECOPD) is unknown. In this retrospective single-center study, we investigated the ability of the serum ChE level to predict in-hospital death in patients hospitalized with AECOPD. The clinicopathological data, including serum ChE levels as well as clinical and biochemical indicators were extracted for 477 patients from the hospital records and analyzed. Our results demonstrated that AECOPD patients with lower serum ChE levels were associated with increased mortality, frequent hospitalization due to acute exacerbations (AE) in the past year, and longer hospital stay. The optimal cutoff value for the serum ChE level was 4323 U/L. The area under the ROC curve (AUC) values for predicting in-hospital mortality based on the serum ChE level was 0.79 (95% confidence interval (CI), 0.72–0.85). Multivariate logistic regression analysis demonstrated that serum ChE level ≤ 4323 U/L (odds ratio (OR) 9.09, 95% CI 3.43–28.3, p < 0.001), age-adjusted Charlson comorbidity index (aCCI), and the number of hospitalizations due to AE in the past year were independent risk factors for predicting the in-hospital mortality of AECOPD patients. In conclusion, our study demonstrated that low serum ChE levels were associated with significantly higher in-hospital mortality rates of patients hospitalized with AECOPD. Therefore, serum ChE level is a promising prognostic predictor of hospitalized AECOPD patients.

Keywords:

• AECOPD
• serum cholinesterase
• mortality
• prediction model
• biomarker

Introduction

Chronic obstructive pulmonary disease (COPD) is a highly prevalent chronic inflammatory lung disease characterized by respiratory symptoms such as breathing difficulty, cough, wheezing, and mucus production because of obstructed airflow in the lung airways and/or alveoli caused by long-term smoking and exposure to fumes and toxic gases [1]. COPD is a major disease burden globally that is associated with high morbidity and mortality rates [2]. Acute exacerbations of COPD (AECOPD) are caused by disease progression and are associated with reduced quality of life, frequent hospitalization, and death [3–5]. The hospitalization rates due to AECOPD are high [6]. Furthermore, the exacerbations and their outcomes are highly heterogeneous among patients with COPD [7]. The currently available clinical biomarkers are not satisfactory in predicting the prognosis of AECOPD patients. Therefore, there is an urgent need for reliable biomarkers that can be used for accurate prognostic prediction and early risk stratification of hospitalized AECOPD patients.

Cholinesterase (ChE) is a glycoprotein synthesized by the liver that can be measured in a fast and economical manner using the serum samples [8]. Previous studies have shown that serum ChE levels are associated with trauma severity and the level of systemic inflammation in patients with severe trauma [9] as well as the muscle mass and strength indices of the elderly [10]. Therefore, Serum ChE assessment has been recommended as an effective tool for evaluating nutritional and inflammatory status [11]. Furthermore, ChE is also associated with the pathogenic mechanisms of COPD, such as chronic airway inflammation and oxidation/antioxidant imbalance [12, 13].

Serum ChE levels are useful in predicting the prognosis of various human diseases [11, 14–19]. However, the relationship between serum ChE levels and in-hospital mortality of AECOPD patients is unclear. Therefore, in the present study, we conducted a retrospective study to investigate the prognostic value of serum ChE in hospitalized AECOPD patients.

Materials and methods

Study design, setting, and subjects

In our hospital, serum ChE levels have been estimated for all the hospitalized patients since 2015 as part of the standard liver function test package, which therefore offered us the opportunity to investigate the prognostic significance of serum ChE in hospitalized AECOPD patients by conducting this retrospective study. The medical records of AECOPD patients hospitalized at the Department of Respiratory and Critical Care Medicine of Wuhu Hospital of Traditional Chinese Medicine were reviewed between June 2016 and December 2021. The inclusion criteria were as follows: (1) age ≥ 40 years; (2) confirmed diagnosis of COPD based on pulmonary function test results (conducted previously or after the stabilization of clinical symptoms during the current hospitalization) defined as the forced expiratory volume in 1 s (FEV1)/forced vital capacity (FVC) <70% after inhalation of a bronchodilator [20], the reference values and the lower limit of normal (LLN) of spirometry in our hospital were based on the Global Lung Function Initiative (GLI) 2012 [21]; (3) hospitalization due to primary diagnosis of AECOPD with respiratory symptoms such as dyspnea, cough, and sputum pus that were severe enough to require hospitalization [20]; and (4) availability of serum ChE test results from the day of admission or the following day. The exclusion criteria were as follows: (1) patients with malignant tumors; (2) patients with active pulmonary tuberculosis confirmed by pathogen detection or histopathology after admission; (3) patients with a history of exposure to organophosphorus and pyrethroid pesticides including agricultural spraying within 3 months before admission; (4) patients with severe renal insufficiency (estimated glomerular filtration rates [eGFR] < 15 ml/min/1.73 m²); (5) patients with severe liver insufficiency; and (6) absence of relevant clinical data. For patients with multiple admissions during the study period, only the data from the first admission was included in the analysis.

This study was conducted in accordance with the Declaration of Helsinki (as revised in 2013) and was approved by the Medical Ethics Committee of the Wuhu Hospital of Traditional Chinese Medicine (LW 2022-09; May 13, 2022). Informed consent from study participants was waived because this study was conducted by collecting patient data in a retrospective and anonymous manner from the electronic medical records without influencing patient treatment.

Variables and data measurement

All the patient data was extracted from the electronic medical record system of the Wuhu Hospital of Traditional Chinese Medicine. This included parameters such as age, sex, the number of hospitalizations due to AE in the past year, underlying diseases (hypertension, diabetes mellitus, coronary heart disease, and heart failure), aCCI (age-adjusted Charlson comorbidity index), noninvasive ventilation (NIV), ICU admission, length of hospital stay, prognosis (survival/death), blood test results including coefficient of variation of red blood cell distribution width (RDW-CV), platelet counts (PLT), platelet-to-lymphocyte ratio (PLR), blood biochemistry data including serum albumin (Alb), serum creatinine (CREA), serum C-reactive protein (CRP), and serum ChE (butyryl thiocholine assay, normal reference range: 5000–12,000 U/L), and arterial blood gas analysis results including oxygenation index (OI), pH, partial pressure of carbon dioxide (PCO₂), and blood lactate (Lac). All the blood tests were conducted on the day of admission or the next morning after fasting, usually within 24 h or even shorter. Participants with missing data were removed from the data set because the proportion of the missing data is small.

Outcomes and definitions

This study’s primary endpoint was all-cause mortality during hospitalization (in-hospital death). The secondary endpoints included presence of hypercapnia, admission to the intensive care unit (ICU), and the use of noninvasive and/or invasive mechanical ventilation. Length of hospital stay was defined as the number of days admitted in the hospital, yet, subjects discharged or dying on the same day of admission are recorded as a one-day stay. The NLR and PLR was calculated as the absolute neutrophil count divided by the absolute lymphocyte count, and absolute platelet count divided by absolute lymphocyte count, respectively. The oxygenation index was calculated as arterial partial oxygen pressure (PaO₂)/fraction of inspired oxygen (FiO₂). Hypercapnia was defined as PCO₂ > 45 mmHg.

Statistical analysis

The continuous variables were expressed as the median and interquartile range (IQR). The categorical variables were expressed as frequencies and percentages. The chi-square test was used to analyze the categorical variables. The Mann–Whitney U test was used to analyze the continuous variables. Receiver operating characteristic (ROC) curve analysis was used to evaluate the sensitivity and specificity of the serum ChE levels in predicting in-hospital mortality. The area under the ROC curve (AUC) value was used to assess the prediction accuracy and determine the cutoff value. The AECOPD patients were divided into two groups based on the serum ChE cutoff value. Subsequently, the chi-square test (categorical variables) and the Mann–Whitney U test (continuous variables) were used to compare the differences in the various clinicopathological indicators between the two groups. Finally, multivariate logistic regression analysis was used to determine the performances of various parameters in predicting the in-hospital mortality of AECOPD patients. Variables selection in the multivariate regression model was based on univariate regression and then referred to previous studies. The final prediction model would include only variables reported to be associated with the patient’s prognosis and significant in the univariate regression. Because certain continuous variables were converted to categorical variables, we therefore further run a new regression model incorporating these variables as continuous variables to test whether the results would change (Table S1). The goodness-of-fit of the model was assessed using the calibration curve. The calibration curve (internal calibration) was drawn with “rms” package in R, then the bias-corrected calibration curve was further drawn to estimate out of sample calibration via a bootstrapping procedure (external calibration). The cutoff values for the serum albumin levels were determined using the clinical reference intervals that were used in the laboratory. Age and pH cutoff values were based on previous reports [22, 23]. To assess the performance of the prediction model, the Brier score (0.055) was calculated, and the clinical use-fullness (net benefit curve) was drawn (Figure S1). The statistical analysis was performed using the R (version 4.1.3) statistical software. A two-sided p value <0.05 was considered as statistically significant. GraphPad Prism 8.0 software (GraphPad Prism Software, San Diego, CA, USA) was used for plotting the graphs.

Results

Basic characteristics of AECOPD patients

Initially, 562 AECOPD patients were enrolled in this study based on the inclusion criteria. However, 85 patients were excluded after applying the exclusion criteria. Finally, 477 patients were included in this study (Figure 1). The median age of the study subjects was 77 (70–83) years and the in-hospital mortality rate was 10.1%. The study included 358 male patients (75.1%). The median length of hospital stay was 11 (8–15) days, and the median number of hospitalizations due to AE in the past year was 1 (1–2). Among the 477 patients, 205 (43%) showed hypercapnia (PCO₂ > 45 mmHg), 118 (24.7%) received noninvasive ventilation (NIV), and 48 (10.1%) were admitted to the ICU. The co-morbidities included diabetes mellitus in 70 patients (14.7%), hypertension in 203 patients (42.6%), coronary heart disease in 179 patients (37.7%), and heart failure in 213 patients (45.4%). The median age-adjusted Charlson comorbidity index (aCCI) was 5 (4–6). The serum ChE level distribution in the study subjects is shown in Figure 2. The average values for various clinical and biochemical indicators were as follows: RDW-CV, 14.1 (13.2–15.1); Alb, 36.7 (33.38–40.4) g/L; CREA, 75.1 (59.3–98.2) µmol/L; CRP, 17 (4.2–67.2) mg/L; OI, 267 (201–341); PCO₂, 45.00 (34.40–60.48) mmHg; pH, 7.4 (7.35–7.44); and Lac, 1.2 (0.9–1.6) mmol/L.

Figure 1. The flow chart shows the selection protocol of the study participants.

Figure 2. Distribution of the serum cholinesterase levels in patients with AECOPD. Abbreviations: ChE, cholinesterase; IQR, interquartile range.

Comparison of clinical characteristics between the non-survival and survival groups of AECOPD patients

Table 1 shows the differences in the clinical parameters between the non-survival and survival groups of AECOPD patients. In comparison with the patients in the survival group, those in the non-survival group were older (81.00 [77.00–86.00] years vs. 76.00 [70.00–82.00] years) and more frequently hospitalized due to AE in the past year (2.00 [2.00–3.00] vs. 1.00 [0.00–2.00]). The non-survival group also showed higher aCCI values and longer time of hospitalization. Furthermore, the non-survival group showed lower oxygenation index (OI) values, pH values, and higher proportion of patients with hypercapnia compared to the survival group. Coronary heart disease and heart failure were more prevalent in the non-survival group than in the survival group. The prevalence of diabetes mellitus and hypertension was similar between the two groups. The serum ChE levels were significantly lower in the non-survival group compared to the survival group (3211.5 [IQR 2922.75–4203.25] vs. 5295 [IQR 4116–6737], p < 0.001). The serum albumin levels were significantly lower in the non-survival group than those in the survival group. However, RDW-CV, PLT, PLR, CREA, CRP, and PCT values were comparable between the survival and non-survival groups.

Table 1. Baseline characteristics of hospitalized survival and non-survivals groups of AECOPD patients.

Characteristics	Survival	Non-survival	P value
Number of patients	429	48
Age (years)	76.00 [70.00–82.00]	81.00 [77.00–86.00]	<0.001
Sex			0.222
Males	318 (74.1)	40 (83.3)
Females	111 (25.9)	8 (16.7)
Number of hospitalizations due to AE in the past year	1.00 [0.00–2.00]	2.00 [2.00–3.00]	<0.001
Comorbidities
DM	64 (14.9)	6 (12.5)	0.815
Hypertension	183 (42.7)	20 (41.7)	1.000
CHD	152 (35.6)	27 (56.2)	0.008
Heart failure	186 (43.4)	27 (67.5)	0.006
aCCI	5.00 [4.00–6.00]	6.00 [5.00–6.25]	<0.001
Laboratory tests
RDW-CV (%)	14.10 [13.10–15.00]	14.50 [13.73–15.30]	0.138
PLT (×10^9/L)	165.00 [128.00–220.00]	163.50 [126.50–215.50]	0.871
PLR	190.74 [117.36–317.70]	226.65 [135.89–385.69]	0.169
ChE (U/L)	5295.00 [4116.00–6737.00]	3211.50 [2922.75–4203.25]	<0.001
Alb (g/L)	37.25 [34.00–40.50]	32.50 [29.70–35.93]	<0.001
CREA (μmol/L)	74.70 [59.30–96.85]	84.05 [59.15–123.70]	0.203
CRP (mg/mL)	15.90 [3.90–65.30]	37.10 [13.78–102.97]	0.057
PCT (ng/mL)	0.08 [0.05–0.56]	0.11 [0.05–0.86]	0.838
OI	274.28 [205.00–346.80]	212.00 [156.75–267.92]	<0.001
PCO2 (mmHg)	44.03 [34.20–59.70]	52.80 [36.55–66.85]	0.057
pH	7.40 [7.35–7.44]	7.36 [7.28–7.42]	0.004
Lac (mmol/L)	1.20 [0.90–1.60]	1.00 [0.78–1.82]	0.256
Secondary outcomes
Hypercapnia	175 (40.8)	30 (62.5)	0.006
NIV	96 (22.4)	22 (45.8)	0.001
MV	18 (4.2)	10 (20.8)	<0.001
ICU Admission	34 (7.9)	14 (29.2)	<0.001
LOS (days)	10.00 [8.00–14.00]	21.50 [11.75–34.50]	<0.001

Abbreviations: DM, diabetes mellitus; CHD, coronary heart disease; aCCI, age-adjusted Charlson comorbidity index; RDW-CV, coefficient of variation of red blood cell distribution width; PLT, platelet; PLR, platelet-to-lymphocyte ratio; ChE, serum cholinesterase; Alb, albumin; CREA, serum creatinine; CRP, serum C-reactive protein; PCT, procalcitonin; OI, oxygenation index; PCO₂, partial pressure of carbon dioxide; Lac, blood lactate; Hypercapnia, PCO₂ > 45 mmHg; NIV, noninvasive ventilation; MV, mechanical ventilation; IQR, interquartile range; LOS, length of stay.

Serum ChE levels show good prognostic prediction performance in the hospitalized AECOPD patients

ROC curve analysis of the prognostic performance of serum ChE level in hospitalized AECOPD patients is shown in Figure 3. The AUC value for the serum ChE levels in the AECOPD patients was 0.79 (95% CI [0.72–0.85]). The sensitivity and specificity values for serum ChE levels were 0.72 and 0.81, respectively, at the cutoff value of 4323 U/L. The serum ChE levels were > 4323 U/L (high ChE group) in 316 patients and ≤ 4323 U/L (low ChE group) in 161 patients. There were significant differences in age, the number of hospitalizations due to AE in the past year, the history of diabetes mellitus, the history of hypertension, the history of heart failure, length of hospital stay, the proportion of patients who received NIV, the proportion of patients who received invasive ventilation, ICU admissions, hypercapnia, and the PLR, RDW-CV, Alb, CRP, PCT, the OI, and PCO₂ values between the high ChE and low ChE groups (p < 0.05; Table 2). However, the high ChE and low ChE groups did not show any significant differences in sex, history of coronary heart disease, and the PLT, pH, and serum lactate values (p > 0.05) (Table 2).

Figure 3. ROC curve analysis of the clinical value of serum cholinesterase levels in predicting the in-hospital mortality of patients with AECOPD. Abbreviations: ROC, receiver operating characteristic; AUC, area under the ROC curve; CI, confidence interval.

Table 2. Patient characteristics stratified by the serum cholinesterase levels.

Characteristics	ChE ≤ 4323 U/L	ChE > 4323 U/L	P value
Number of patients	161	316
Age (years)	78.00 [73.00–83.00]	76.00 [69.00–82.25]	0.006
Sex			0.205
Male	127 (78.9)	231 (73.1)
Female	34 (21.1)	85 (26.9)
Number of hospitalizations due to AE in the past year	1.00 [1.00–2.00]	1.00 [0.00–2.00]	0.025
Comorbidities
DM	15 (9.3)	55 (17.4)	0.026
Hypertension	53 (32.9)	150 (47.5)	0.003
CHD	69 (42.9)	110 (35.0)	0.117
Heart failure	96 (61.9)	117 (37.3)	<0.001
aCCI	5.00 [4.00–6.00]	5.00 [4.24–6.00]	<0.001
Laboratory tests
RDW-CV (%)	14.70 [13.60–16.00]	13.90 [13.10–14.70]	<0.001
PLT (*10^9/L)	158.00 [115.00–238.00]	169.00 [131.00–212.50]	0.229
PLR	224.30 [148.55–369.86]	177.35 [113.21–283.72]	0.002
Alb (g/L)	32.75 [29.65–36.20]	38.40 [35.70–41.60]	<0.001
CREA (μmol/L)	77.30 [58.40–121.13]	74.50 [59.55–91.30]	0.048
CRP (mg/mL)	36.75 [9.85–115.53]	12.10 [2.90–48.00]	<0.001
PCT (ng/mL)	0.24 [0.05–0.86]	0.05 [0.05–0.21]	<0.001
OI	240.60 [168.00–289.00]	285.05 [207.88–358.22]	<0.001
OI level			<0.001
≥300	34 (21.1)	148 (46.8)
200–299	74 (46.0)	103 (32.6)
100–199	42 (26.1)	61 (19.3)
≤99	11 (6.8)	4 (1.3)
PCO2 (mmHg])	50.39 [35.10–66.90]	42.85 [34.20–57.58]	0.019
pH	7.39 [7.32–7.44]	7.40 [7.36–7.44]	0.187
Lac (mmol/L)	1.23 [0.90–1.84]	1.13 [0.90–1.60]	0.194
Secondary outcomes
Hypercapnia	85 (52.8)	120 (38.0)	0.003
NIV	62 (38.5)	56 (17.7)	<0.001
MV	21 (13.0)	7 (2.2)	<0.001
ICU Admission	33 (20.5)	15 (4.7)	<0.001
LOS (days)	12.00 [9.00–19.00]	10.00 [7.00–13.25]	<0.001

Abbreviations: ChE, serum cholinesterase; DM, diabetes mellitus; CHD, coronary heart disease; aCCI, age-adjusted Charlson comorbidity index; RDW-CV, coefficient of variation of red blood cell distribution width; PLT, platelet; PLR, platelet-to-lymphocyte ratio; Alb, albumin; CREA, serum creatinine; CRP, serum c-reactive protein; PCT, procalcitonin; OI, oxygenation index; PCO₂, partial pressure of carbon dioxide; Lac, blood lactate; Hypercapnia, PCO₂ > 45 mmHg; NIV, noninvasive ventilation; MV, mechanical ventilation; IQR, interquartile range, LOS, length of stay.

Binary multivariate logistic regression analysis

Table 3 shows the results of the univariate and multivariate logistic regression analyses to determine the risk factors associated with the in-hospital mortality of AECOPD patients. Univariate analysis results showed that aCCI, Alb ≤ 35 g/L, ChE ≤ 4323 U/L, pH <7.35, hypercapnia, and OI <300 were potential risk factors for in-hospital mortality in the AECOPD patients. The main principles for predictors’ selection in the multivariate regression analysis were definitive evidence from previous studies combined with its accessibility, low cost, efficacy, and simplicity. Furthermore, multivariate logistic regression analysis was performed with parameters such as aCCI, the number of hospitalizations due to AE in the past year, serum ChE, OI, hypercapnia, and serum Alb. Multivariate analysis demonstrated that ChE ≤ 4323 U/L (vs. serum ChE levels > 4323 U/L; odds ratio [OR] 9.09, 95% CI (3.43–28.3), p < 0.001), aCCI (OR 1.42, 95% CI (1.03–2.00), p = 0.037), number of hospitalizations due to AE in the past year (OR 2.37, 95% CI (1.70–3.41), p < 0.001), and Alb ≤ 35 g/L (OR 2.60, 95% CI (1.13–6.36), p = 0.029) were significantly associated with the in-hospital mortality of AECOPD patients. The p value from the Hosmer–Lemeshow goodness of fit test was 0.704, indicating adequate fit. The calibration curve of the multivariate logistic regression model also showed adequate goodness of fit (Figure 4).

Figure 4. Calibration curve of the multivariate logistic regression model. As shown, the calibration curve demonstrates satisfactory goodness of fit for the multivariate logistic regression model.

Table 3. Univariate and multivariate logistic regression analyses of potential risk factors associated with the in-hospital mortality of AECOPD patients.

Characteristics	Univariate		Multivariate
	OR (95% CI)	P value	OR (95% CI)	P value
Age (years)
>65/≤65	1.42 (0.54–4.85)	0.520
aCCI	1.74 (1.33–2.29)	<0.001	1.42 (1.03-2.00)	0.037
Number of hospitalizations due to AE in the past year	1.99 (1.55–2.59)	<0.001	2.37 (1.70-3.41)	<0.001
Alb (g/L)
≤35/>35	4.77 (2.55–9.32)	<0.001	2.6 (1.13-6.36)	0.029
ChE (U/L)
≤4323/>4323	10.90 (5.35–24.62)	<0.001	9.09 (3.43–28.3)	<0.001
Lac (mmol/L)
≥2/<2	1.37 (0.62–2.79)	0.402
pH (%)
7.20–7.35/≥7.35	2.58 (1.32–4.96)	0.005
≤7.20/≥7.35	11.45 (3.46–37.10)	<0.001
Hypercapnia	2.42 (1.32–4.55)	0.004	1.18 (0.51-2.78)	0.7
OI level		<0.001
200–299/≥300	3.73 (1.63–9.63)	0.003	1.88 (0.68-5.79)	0.2
100–199/≥300	2.98 (1.14–8.36)	0.028	1.74 (0.52-6.17)	0.4
≤99/≥300	21.8 (6.21–80.7)	<0.001	3.47 (0.57-20.3)	0.2

Abbreviations: aCCI, age-adjusted Charlson comorbidity index; Alb, albumin; ChE, serum cholinesterase; Lac, blood lactate; Hypercapnia, PCO₂ > 45 mmHg; OI, oxygenation index, OR, odds ratio; CI, confidence interval.

Discussion

This is the first clinical study to the best of our knowledge investigating the correlation between the serum ChE levels and the in-hospital mortality of AECOPD patients. The mortality risk was significantly higher for AECOPD patients with serum ChE levels ≤ 4323 U/L compared to those with serum ChE levels > 4323 U/L (OR 9.09, 95% CI, 3.43–28.3). The AUC value of the regression model using the cutoff value of 4323 U/L for the serum ChE levels was 0.79 (95% CI, 0.72–0.85). This suggested that the serum ChE levels were good predictors of in-hospital mortality for patients with AECOPD. Multivariate analysis also showed that aCCI (OR 1.42, 95% CI, 1.03–2.00, p = 0.037), the number of hospitalizations due to AECOPD in the past year (OR 2.37, 95% CI [1.70–3.41], p < 0.001), and Alb ≤ 35 g/L (OR 2.60, 95% CI [1.13–6.36], p = 0.029) were also significant predictors of in-hospital mortality for patients with AECOPD.

Our data showed that serum albumin levels were significantly lower in the non-survival group of AECOPD patients compared to those in the survival group. This finding was comparable to the findings reported by Koc et al. [24]. Crisafulli et al. [25] reported that age was an independent predictor of 90-day mortality in hospitalized AECOPD patients when age was analyzed as a continuous variable and a tertile variable. We did not include age as an independent variable in the multivariate logistic regression model. However, we analyzed aCCI, a composite variable dependent on age. Oster et al. [26] analyzed hospitalized patients with non-liver and non-kidney diseases and reported that patients with hypoalbuminemia were older and were associated with higher Charlson co-morbidity index (CCI) values, longer average length of hospital stay, higher 1-year re-admission rates, and higher 1-year mortality rates than patients with normal albuminemia. Moreover, Pellicori et al. [27] demonstrated that hypoalbuminemia was closely related to mortality in the AECOPD patients. In the present study, the multivariate logistic regression analysis also showed that serum albumin level ≤ 35 g/L was closely related with the in-hospital mortality of patients with AECOPD.

In the present study, the hospital stay was longer for patients in the non-survival group compared to those in the survival group. Furthermore, higher proportion of patients in the non-survival group required noninvasive or invasive ventilation compared to those in the survival group. Moreover, higher proportion of patients in the non-survival group were diagnosed with hypercapnia and lower OI. This suggested higher disease severity in the non-survival group of patients with AECOPD. García-Sanz et al. [28] reported that advanced age, longer hospital stay, requirement for invasive or noninvasive mechanical ventilation, early readmission, history of atrial fibrillation, and dementia were independent factors for predicting the 1-year and long-term mortality of hospitalized COPD patients. This suggested that AE severity was closely associated with mortality. The aCCI values can be used to predict the long-term mortality of patients belonging to different clinical populations, including internal medicine, surgery, ICU, trauma, and cancer patients [29]. Ho et al. [30] performed a population-based cohort study of 4204 hospitalized AECOPD patients and showed that aCCI was closely associated with in-hospital mortality and 1-year mortality of patients (OR: 1.08 per point; 95% CI: 1.01–1.15). These results were consistent with our findings. Hurst et al. reported that history of AE was the best predictor of AE recurrence in COPD patients [31]. Furthermore, Celli et al. reported that the frequency of AE was closely related to all-cause mortality [32]. Our study also demonstrated that the in-hospital mortality of AECOPD patients was significantly correlated with the number of hospitalizations due to AE in the past year (OR 2.37, 95% CI [1.70–3.41], p < 0.001).

Hu et al. reported that D-dimer was a strong and independent risk factor in the AECOPD patients for in-hospital mortality and 1-year mortality [22]. In our study, a considerable proportion of patients did not undergo D-dimer testing. Hence, we could not analyze the predictive value of the D-dimer. Furthermore, in the present study, we observed that the serum CRP values were statistically similar between the survival and non-survival group of AECOPD patients. This was contradictory to the reports by Gomez-Rosero et al. and Leuzzi et al. [33, 34]. However, other studies have also reported absence of correlation between the serum CRP values and the mortality of AECOPD patients [22, 35]. Therefore, the prognostic value of CRP in patients with AECOPD requires further study.

In the present study, hospitalized AECOPD patients with low serum ChE levels were associated with higher in-hospital mortality rates. This was in line with the results of previous studies. Seo et al. [36] reported that low serum ChE levels predicted adverse outcomes in patients with preserved ejection fraction but acute decompensated heart failure. Takano et al. [15] reported that low pre-operative serum ChE levels were associated with post-operative recurrence and poor prognosis after curative resection in patients with colorectal cancer. This result might partly be explained by systemic inflammation and malnutrition [37, 38], which are highly prevalent in AECOPD patients, especially those in the later stages of the disease [39]. Previous studies have shown that lower ChE levels are closely related to systemic inflammation, malnutrition, and sarcopenia [11], which are associated with poor prognosis in subjects with COPD [40, 41].

Some biomarkers are frequently used in clinical practice and proved to have a good predictive value for in-hospital mortality, including Neutrophil-to-lymphocyte ratio (NLR) [42], D-dimer [22], and IL-8 [43]. In the previous study, the AUC for serum D-dimer to predict in-hospital death was 0.748 (95% CI 0.641–0.854), lower than that of ChE in our study. However, it is well-recognized that no single factor accurately predicts adverse outcomes in patients after hospitalization with AECOPD. The present study has demonstrated that the serum ChE level could be a useful complementary tool. In addition, ChE holds enormous promise in many fields of COPD research, as it is reported to be closely related to inflammation and malnutrition [11]. Thus, our current study is a pilot study; the promising results found in the present study could be used as rationality to conduct further prospective studies.

This study has several limitations. First, this is a single-center retrospective study with a small sample size; biases are inevitable, like competing risks, and no blinding in the study. Moreover, the baseline lung function data were unavailable because patients with poor cardiopulmonary function could not complete the test; therefore, we could not incorporate the lung function in the final prediction model. Although we have used the oxygenation index and the level of carbon dioxide in the arterial blood as surrogates in the final model, which could still induce bias and should be interpreted cautiously. Additionally, we could not eliminate the inhomogeneous therapeutic measures during the study period, including the clinicians’ attitudes toward the intervention, patients’ economic status, and their tolerance and willingness to receive noninvasive or invasive ventilation, which undoubtedly impacts the patient’s prognosis. Second, all patients were from one large tertiary hospital, mainly dealing with patients with underlying diseases and in critical condition. Therefore, this result needs to be interpreted with caution, which might be influenced by the practices exclusive to this center, thereby limiting the generalizability of the study findings. Third, variables including in the final model were based on univariate regression. Although all these variables have been demonstrated well that were associated with patient’s outcomes in published literature, this post-hoc analysis may still somewhat introduce bias. Fourth, this study did not include various clinical indicators of nutritional status such as the geriatric nutritional risk index (GNRI), controlling nutritional status (CONUT) score, and prognostic nutritional index (PNI). Therefore, the correlation between serum ChE levels and the nutritional status of AECOPD patients could not be further analyzed. Fifth, we did not study whether adding the ChE to widely validated tools could improve the predictive power, like the DECAF score [44]. These studies are highly welcomed as which might improve the predictive power with the new updated tools. Finally, we did not conduct external validation, which could lead to over-fitting of the model and an over-optimistic estimate of the prediction accuracy.

Conclusions

This study demonstrated that low serum ChE levels were associated with significantly higher in-hospital mortality of patients with AECOPD. Therefore, our findings suggest that serum ChE level is a promising clinical biomarker for predicting the in-hospital mortality of patients with AECOPD.

Author contributions

Zhixiang Chen and Lei Zha conceived the idea for this report and are co-first authors. Zhixiang Chen, Huimin Xia, Milan Zhang, and Lu Li collected data for the study. Zhixiang Chen, Lei Zha, and Qiancheng Xu performed the statistical analyses and prepared figures and tables. Zhixiang Chen, Guohong Feng, Qian An, Fei Shi, Jingjing Xu, Huimin Xia, Milan Zhang, and Lu Li wrote the first draft of the manuscript. Zhixiang Chen, Qiancheng Xu, and Lei Zha reviewed, rated, and revised the manuscript. All authors approved the final version and are accountable for all aspects of the work.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data that support the findings of this study are available on request from the corresponding author, Zhixiang Chen.

Additional information

Funding

This work was supported by the [High-level Talents Project of the Wuhu Health Commission] under Grant [Number 2020-132].