Introduction

Chronic obstructive pulmonary disease (COPD), characterized by persistent airflow obstruction and respiratory symptoms, affects approximately 400 million people and is expected to become the third leading cause of mortality and disability worldwide by 2030.1,2 It is related to genetic, developmental, and social factors, and exposure to inhaled particulate matter (mainly cigarette smoke and air pollutants).3 Acute exacerbations of chronic obstructive pulmonary disease (AECOPD) are episodes of symptoms of COPD that have significant adverse consequences for patients.4 It is estimated that 22–40% of COPD patients suffer at least one moderate or severe exacerbation each year, while 9–16% of COPD patients suffer more than one. COPD patients with acute exacerbation experience accelerated lung function decline and increased risk of myocardial infarction, cerebrovascular events and death.5 Therefore, it is essential to identify accurate predictors of prognosis in AECOPD patients.

During the hospitalization of AECOPD patients, fluid volume management is an extremely important treatment strategy. Fluid overload can lead to expiratory flow limitation (EFL), a dynamic state of maximal expiratory flow value, which results in a higher mortality risk in the ICU.6,7 Clinically, plasma volume (PV) can evaluate the fluid balance of patients, and has been suggested to be related to the prognosis of critically ill patients in ICU by several epidemiological studies.8–10 Traditionally, PV is measured by tracer dilution technology, but these methods are expensive, time consuming and complicated.11 Hence, a noninvasive and easily available tool is necessary to assess the PV status. Estimated PV status (ePVS) is such a parameter based on hemoglobin and hematocrit, which is proposed to estimate the deviation of a patient’s PV from the ideal PV.12 Currently, there are two formulas to calculate ePVS, Duarte-ePVS and Kaplan–Hakim-ePVS (KH-ePVS), which have been proven to have different predictive abilities in predicting the outcome of ICU patients. He et al13 suggested higher KH-ePVS is related to the poor prognosis of acute respiratory distress syndrome (ARDS), while observing no significant association between Duarte-ePVS and the outcome of ARDS patients. However, the associations of ePVS, calculated by Duarte and Kaplan–Hakim, and the risk of in-hospital mortality in AECOPD patients admitted to the ICU were not understood.

Herein, based on the electronic Intensive Care Unit Collaborative Research Database (eICU-CRD), we explored the association of ePVS and the in-hospital mortality in AECOPD patients admitted to the ICU. This study aims to find accurate markers for identifying poor prognosis of AECOPD and assist in the clinical management and risk stratification of AECOPD patients admitted to the ICU.

Methods Data Source and Study Outcome

Data of this retrospective cohort study were extracted from eICU-CRD (version 2.0) (https://eicu-crd.mit.edu/), which was established by Philips Healthcare in collaboration with the Massachusetts Institute of Technology (MIT) Computational Physiology Laboratory. This database contains demographic, vital sign measurements, diagnostic information, and treatment information for more than 200,000 patients in 355 ICUs at 208 hospitals in the US during the period 2014–2015.14 The present study used anonymous data available in the eICU database and was preapproved by the Institutional review board at the MIT.15 The requirement of ethical approval for this was waived by the Institutional Review Board of Jian Yang Hospital of Traditional Chinese Medicine, because the data was accessed from eICU-CRD (a publicly available database). The need for written informed consent was waived by the Institutional Review Board of Jian Yang Hospital of Traditional Chinese Medicine due to the retrospective nature of the study.

The inclusion criteria were: (1) subjects aged ≥18 years old; (2) subjects diagnosed as AECOPD at the ICU admission; (3) subjects with examination of hematocrit, hemoglobin and weight upon ICU admission. The exclusion criteria were: (1) patients stayed in the ICU less than 24 h; (2) patients missing survival data.

Study Outcome

The study outcome of the present study was in-hospital mortality among AECOPD patients admitted to the ICU. Patient mortality information was accessed from the US Social Security Death Index.

The ePVS Calculation

The Duarte and Hakim formulas were utilized to calculate the ePVS in the present study. The Duarte formula included hematocrit and hemoglobin as follows: ePVS = [100-hematocrit (%)/hemoglobin (g/dL)].16 The Hakim formula17 was derived by comparing actual volume (aPV) to ideal plasma volume (iPV). The aPV = (1-hematocrit) ×[a+(b× body weight)] (males: a = 1530; b = 41.0. females: a = 865; b = 47.9). The iPV = c× body weight (males: c = 39; females: c = 40). KH-ePVS = [(aPV-iPV)/iPV] × 100%. In the present study, Duarte-ePVS and KH-ePVS was divided into three levels. The low level of Duarte-ePVS was <96.81; medium Duarte-ePVS was 96.81–96.94; high Duarte-ePVS was ≥96.94. The low level of KH-ePVS was <−12.26; medium KH-ePVS was −12.26- −1.21; high level of KH-ePVS was ≥−1.21.

Potential Covariates

The potential covariates include demographic information [age (<65 and ≥65 years old), gender (male and female) and ethnicity (African American, Caucasian and other)], vital signs [heart rate, blood pressure, respiratory rate, temperature and oxyhemoglobin saturation (SPO2)], vital signs [white blood cell (WBC), lymphocyte count, neutrophil count, platelet count, hematocrit, hemoglobin, red blood cell distribution width (RDW-CV), bilirubin, eGFR, international normalized ratio (INR), prothrombin time (PT), partial thromboplastin time (PTT), albumin, blood urea nitrogen (BUN), glucose, calcium, anion gap and arterial blood gas test], interventions [vasopressors, mechanical ventilation, diuretic, erythropoietin and angiotensin converting enzyme inhibitor/angiotensin receptor antagonist (ACEI/ARB), complications (congestive heart failure, cardiogenic shock, sepsis, pneumonia, polycythemia, hematological system disease)] and scoring systems [Glasgow Coma Scale (GCS), age Above or Below 65 Years (CURB-65) and Acute Physiology, Age, Chronic Health Evaluation IV score (APACHE IV)]. If vital signs were collected multiple times or patients received more than one laboratory test during the hospitalization, the initial test data with the first 24 h after ICU admission were extracted for subsequent analysis.

Statistics Analysis

For the baseline characteristics, the normality of quantitative data was tested by skewness and kurtosis measurements, and the homogeneity of variance was tested by Levene’s test. Normally distributed measurement data were described by [Mean (±SD)]. The t-test was used for comparison between groups with equal variance, and comparisons between groups with uneven variances were performed by t-test. Non-normal data were presented by the median and interquartile, and the Wilcoxon rank sum test was used for comparison between alive and dead groups. Count data was described as the number and percentage [n (%)], and Chi-square test or Fisher’s exact test was used for comparison between alive and dead group. Multivariate imputation by chained equations (MICE) was used for missing data imputation. Sensitivity analysis was performed before and after missing data imputation (Supplementary Table 1).

The univariate and multivariate Cox regression models were utilized to explore the association of ePVS and KH-ePVS with in-hospital mortality among AECOPD patients admitted to the ICU, with hazard ratios (HRs) and 95% confidence intervals (CIs). Model 1 was a crude model without adjusting any covariates. Model 2 was adjusted gender, heart rate, ICU stay duration, sepsis, Apache score, CURB-65 and vasopressor. All statistical analyses were performed by SAS 9.4 (SAS Institute Inc., Cary, NC, USA). Two-sided P-value <0.05 was considered statistically significant.

Results Characteristics of AECOPD Patients

Figure 1 shows the flow chart for the screening of AECOPD patients. A total of 3,886 AECOPD patients aged ≥18 years old were screened from the EICU. Among them, 236 patients were missing hematocrit examination, 33 patients were missing hemoglobin examination, and 113 were missing weight examination upon ICU admission. 703 patients with the duration of ICU stay less than 24 h, and 28 patients missing survival data were excluded. Finally, 2,773 eligible AECOPD patients were included, with the mean age of 67.53 (±10.62). Of whom, 219 (7.90%) patients passed away and 2,554 patients survived within 6.24 (4.01–9.26) days. The proportion of patients with higher KH-ePVS levels was significantly higher in the dead group compared with survival group (49.77% vs 31.95%). No significant differences in ePVS levels were found between the alive and dead groups (P>0.05). The clinical and biochemical characteristics of the AECOPD patients are shown in Table 1.

Table 1 Characteristics of AECOPD Patients

Figure 1 The flow chart of population screening.

Multivariable Cox Proportional Hazard Analysis of in-Hospital Mortality

As shown in Supplementary Table 2, it can be found that the risk of in-hospital mortality among AECOPD patients was related to gender, heart rate, ICU stay duration, sepsis, Apache score, CURB-65 and vasopressor (all P<0.05).

KH-ePVS, ePVS and in-Hospital Mortality Among AECOPD Patients

Two Cox proportional hazard models were utilized to explore the association of KH-ePVS and ePVS with in-hospital mortality among AECOPD patients admitted to the ICU, as depicted in Table 2 and Figure 2. After adjusted gender, heart rate, ICU stay duration, sepsis, Apache score, CURB-65 and vasopressor, patients with high KH-ePVS level had a 53% increased risk of in-hospital mortality compared with those with low KH-ePVS level (HR = 1.53, 95% CI: 1.05–2.24); while, no statistically significant associations were observed between Duarte-ePVS and in-hospital mortality (P>0.05).

Table 2 KH-ePVS, Duarte-ePVS and in-Hospital Mortality of AECOPD Patients

Figure 2 The Kaplan Meier curve of KH-ePVS and in-hospital mortality among AECOPD patients.

Association of KH-ePVS with in-Hospital Mortality Based on Different Age, Gender and Complications

Subgroup analysis were performed to further explore the association of KH-ePVS with in-hospital mortality among AECOPD patients based on age, gender, the history of sepsis and congestive heart failure. The results are shown in Figure 3. After all covariates were adjusted, we found the association of KH-ePVS and in-hospital mortality was still robust, especially among AECOPD patients without a history of sepsis (HR = 1.78, 95% CI: 1.11–2.85) and congestive heart failure (HR = 2.19, 95% CI: 1.34–3.56).

Figure 3 Subgroup analysis of KH-ePVS and in-hospital mortality among AECOPD patients.

Discussion

Based on the eICU-CRD database, we explored the association between ePVS assessed by two different methods, Duarte-ePVS and KH-ePVS, and the risk of in-hospital all-cause mortality in AECOPD patients. Our findings suggested that high KH-ePVS was related to high in-hospital mortality risk in AECOPD patients, while no significant statistical association of Duarte-ePVS and in-hospital mortality was found. There was potential benefit in monitoring KH-ePVS levels to assist in the clinical management and prediction of prognosis of AECOPD patients in the ICU.

In patients admitted to the ICU, fluid management is essential in order to restore cardiac output, systemic blood pressure, and renal perfusion. Clinically, patients with COPD were often accompanied by EFL, resulting in a higher risk of death. Direct quantification of PV, representing the status of fluid balance, has been proven to have clinical utility to reveal volume overload in several diseases, however, this methodology has not been readily available to clinicians due to its being costly, time consuming and complex.11 Some hemodialysis equipment used hematocrit to monitor volume during hemodialysis, but this methodology is less precise in assessing PV status.18 Previous studies reported that the ePVS, calculated from blood count and/or body weight has been associated with the poor outcome of pulmonary disease.10,12 Duarte-ePVS was first derived by Duarte et al16 and the prognostic value of Duarte-ePVS in severe respiratory disease was suggested.10,12 A retrospective, multicenter observational study reported that a lower ePVS on admission related to a greater chance of survival in COVID-19-related ARDS patients,12 however, there was no significant association found between KH-ePVS and outcome of ARDS patients, which was inconsistent with our results. Previous studies reported the similar accuracy of Duarte-ePVS and KH-ePVS for predicting in-hospital death in ICU patients.19,20 However, our study found high KH-ePVS, but not Duarte-ePVS, was associated with poor outcome in AECOPD patients admitted to the ICU. KH-ePVS may be a more powerful predictor of AECOPD patients admitted to the ICU.

Over-activation of the renin-angiotensin-aldosterone system may be associated with high plasma load and increased mortality risk in patients with AECOPD.21 The angiotensin-converting enzyme (ACE) is abundant in the lung. When AECOPD occurred, ACE was activated by acute alveolar hypoxia, increasing the production of angiotensin II (Ang II), an important mediator of lung injury and apoptosis.22 It can mediate inflammatory response by interfering with cytokine production, inflammatory cell migration, epithelial apoptosis, oxidative stress, activation of tissue mast cells and lung fibrosis.23,24 Ang II also can induce bronchoconstriction, leading to the deterioration of lung function.25 Basic experiments in the next step are need to further explore the association of renin-angiotensin-aldosterone system and high plasma load in patients with AECOPD.

To the best of our knowledge, this was the first retrospective cohort study to explore the association of Duarte-ePVS, KH-ePVS and in-hospital mortality in AECOPD patients admitted to the ICU. KH-ePVS is calculated by the KH formula established by Kaplan-Hakim et al based on hematocrit, bodyweight and gender. Our results observed that high KH-ePVS remained significantly connected to in-hospital mortality after multivariate adjustment. Our study has certain clinical significance. Clinicians can relatively objectively judge the PV status of patients with AECOPD by calculating KH-ePVS and formulate standardized personalized treatment plans. In clinical practice, compared with methods such as measuring PV using radioisotope analysis and evaluating the patient’s congestion status by echocardiography, KH-ePVS has the advantages of low-cost, strong practicality, and is not affected by the level of equipment and operators. It is expected to become a useful indicator for the short-term prognosis of AECOPD patients in the ICU. However, several limitations should be recognized in the present study. First, only US data were included for exploring the association of ePVS and the prognosis of AECOPD patients, which could potentially limit the generalizability of the association extrapolated to other races. Second, we did not directly measure PV by radioisotope assays, although these techniques were relatively impractical and expensive in routine clinical practice. The consistency with estimated and measured PV remains controversial.11 Third, due to the nature of the single-center retrospective study, limited sample size and data bias were inevitable despite strong statistical corrections. However, the inclusion criteria were performed strictly so that the cases included reflected the actual conditions as accurately as possible. Additionally, the present study could provide only the relationship between KH-ePVS and in-hospital mortality in AECOPD patients admitted to the ICU rather than causality. Future research should aim to expand sample sources to include a wider range of regions and patient types to improve the generalizability of the findings.

Conclusion

Our study suggested that high KH-ePVS is associated with high in-hospital mortality among ICU patients with AECOPD. Moreover, KH-ePVS appears to have a more accurate predictive value than Duarte-ePVS in predicting the prognosis of AECOPD patients admitted to the ICU. However, the role of ePVS in guiding fluid management should be investigated in a future large-scale prospective cohort study.

Ethics Approval and Informed Consent

The requirement of ethical approval for this was waived by the Institutional Review Board of Jian Yang Hospital of Traditional Chinese Medicine, because the data were accessed from eICU-CRD (a publicly available database). The need for written informed consent was waived by the Institutional Review Board of Jian Yang Hospital of Traditional Chinese Medicine due to the retrospective nature of the study.

Funding

This study received no funding.

Disclosure

The authors report no conflicts of interest in this work.

References

1. Christenson SA, Smith BM, Bafadhel M, Putcha N. Chronic obstructive pulmonary disease. Lancet. 2022;399(10342):2227–2242. doi:10.1016/S0140-6736(22)00470-6

2. Ritchie AI, Wedzicha JA. Definition, causes, pathogenesis, and consequences of chronic obstructive pulmonary disease exacerbations. Clin Chest Med. 2020;41(3):421–438. doi:10.1016/j.ccm.2020.06.007

3. Labaki WW, Rosenberg SR. Chronic obstructive pulmonary disease. Ann Intern Med. 2020;173(3):Itc17–itc32. doi:10.7326/AITC202008040

4. MacLeod M, Papi A, Contoli M, et al. Chronic obstructive pulmonary disease exacerbation fundamentals: diagnosis, treatment, prevention and disease impact. Respirology. 2021;26(6):532–551. doi:10.1111/resp.14041

5. Bertoletti L, Couturaud F, Sanchez O, Jimenez D. Pulmonary embolism and chronic obstructive pulmonary disease. Semin Thromb Hemost. 2023;49(8):809–815. doi:10.1055/s-0042-1756190

6. Junhasavasdikul D, Telias I, Grieco DL, et al. Expiratory flow limitation during mechanical ventilation. Chest. 2018;154(4):948–962. doi:10.1016/j.chest.2018.01.046

7. Volta CA, Dalla Corte F, Ragazzi R, et al. Expiratory flow limitation in intensive care: prevalence and risk factors. Crit Care. 2019;23(1):395. doi:10.1186/s13054-019-2682-4

8. Kobayashi M, Girerd N, Duarte K, et al. Estimated plasma volume status in heart failure: clinical implications and future directions. Clin Res Cardiol. 2021;110(8):1159–1172. doi:10.1007/s00392-020-01794-8

9. Marawan A, Qayyum R. Estimated plasma volume and mortality: analysis from NHANES 1999-2014. Clin Res Cardiol. 2020;109(9):1148–1154. doi:10.1007/s00392-020-01606-z

10. Niedermeyer SE, Stephens RS, Kim BS, Metkus TS. Calculated plasma volume status is associated with mortality in acute respiratory distress syndrome. Crit Care Explor. 2021;3(9):e0534. doi:10.1097/CCE.0000000000000534

11. Fudim M, Miller WL. Calculated estimates of plasma volume in patients with chronic heart failure-comparison with measured volumes. J Card Fail. 2018;24(9):553–560. doi:10.1016/j.cardfail.2018.07.462

12. Balasubramanian P, Isha S, Hanson AJ, et al. Association of plasma volume status with outcomes in hospitalized Covid-19 ARDS patients: a retrospective multicenter observational study. J Crit Care. 2023;78:154378. doi:10.1016/j.jcrc.2023.154378

13. He C, Zhang S, He H, et al. Predictive value of plasma volume status for contrast-induced nephropathy in patients with heart failure undergoing PCI. ESC Heart Fail. 2021;8(6):4873–4881. doi:10.1002/ehf2.13681

14. Pollard TJ, Johnson AEW, Raffa JD, Celi LA, Mark RG, Badawi O. The eICU collaborative research database, a freely available multi-center database for critical care research. Sci Data. 2018;5:180178. doi:10.1038/sdata.2018.178

15. Li S, Huang Q, Nan W, He B. Association between serum phosphate and in-hospital mortality of patients with AECOPD: a retrospective analysis on eICU database. Heliyon. 2023;9(9):e19748. doi:10.1016/j.heliyon.2023.e19748

16. Duarte K, Monnez JM, Albuisson E, Pitt B, Zannad F, Rossignol P. Prognostic value of estimated plasma volume in heart failure. JACC Heart Fail. 2015;3(11):886–893. doi:10.1016/j.jchf.2015.06.014

17. Hakim RM, Schulman G, Churchill WH Jr, Lazarus JM. Successful management of thrombocytopenia, microangiopathic anemia, and acute renal failure by plasmapheresis. Am J Kidney Dis. 1985;5(3):170–176. doi:10.1016/S0272-6386(85)80046-9

18. Movilli E, Cancarini GC, Cassamali S, et al. Inter-dialytic variations in blood volume and total body water in uraemic patients treated by dialysis. Nephrol Dial Transplant. 2004;19(1):185–189. doi:10.1093/ndt/gfg494

19. Lin Y, Xue Y, Liu J, et al. Prognostic value of estimated plasma volume in patients with chronic systolic heart failure. J Investig Med. 2021;69(2):338–344. doi:10.1136/jim-2020-001538

20. Chen J, Shen J, Cai D, et al. Estimated plasma volume status (ePVS) is a predictor for acute myocardial infarction in-hospital mortality: analysis based on MIMIC-III database. BMC Cardiovasc Disord. 2021;21(1):530. doi:10.1186/s12872-021-02338-2

21. Ruan Z, Li D, Hu Y, Qiu Z, Chen X. The association of renin-angiotensin system blockades and mortality in patients with acute exacerbation of chronic obstructive pulmonary disease and acute respiratory failure: a retrospective cohort study. Int J Chron Obstruct Pulmon Dis. 2022;17:2001–2011. doi:10.2147/COPD.S370817

22. Shrikrishna D, Astin R, Kemp PR, Hopkinson NS. Renin-angiotensin system blockade: a novel therapeutic approach in chronic obstructive pulmonary disease. Clin Sci. 2012;123(8):487–498. doi:10.1042/CS20120081

23. Kaparianos A, Argyropoulou E. Local renin-angiotensin II systems, angiotensin-converting enzyme and its homologue ACE2: their potential role in the pathogenesis of chronic obstructive pulmonary diseases, pulmonary hypertension and acute respiratory distress syndrome. Curr Med Chem. 2011;18(23):3506–3515. doi:10.2174/092986711796642562

24. Benigni A, Cassis P, Remuzzi G. Angiotensin II revisited: new roles in inflammation, immunology and aging. EMBO Mol Med. 2010;2(7):247–257. doi:10.1002/emmm.201000080

25. Raiden S, Nahmod K, Nahmod V, et al. Nonpeptide antagonists of AT1 receptor for angiotensin II delay the onset of acute respiratory distress syndrome. J Pharmacol Exp Ther. 2002;303(1):45–51. doi:10.1124/jpet.102.037382