Low BALF CD4 T cells count is associated with extubation failure and mortality in critically ill covid-19 pneumonia

Abstract Background Critically ill COVID-19 pneumonia is one of the main causes of extubation failure and mortality. Understanding clinical characteristics, laboratory profiles and bronchoalveolar lavage fluid (BALF) immunopathology may help improve outcomes in critically ill COVID-19 pneumonia. We aimed to describe clinical characteristics, laboratory profiles and BALF immunopathology based on lung severity in critically ill COVID-19 pneumonia patients. Materials and methods Forty critically ill severe pneumonia patients requiring invasive mechanical ventilation in Cipto Mangunkusumo General (National Tertiary Referral Hospital), Indonesia within November 2020–January 2021 were enrolled in this study. Early BALF collection was performed after patients’ intubation. Clinical characteristics, laboratory profiles and BALF biomarkers (sTREM-1, alveolar macrophage amount and function, IL-6, IL-17, CD4 T-cells, Tregs, SP-A and Caspase-3) were observed and analysed. Outcomes were measured based on extubation failure (within 19 days) and 28-days mortality. Univariate and bivariate analyses were performed. Results Early bronchoscopy was performed in an average of 4 h (SD = 0.82) after patients’ intubation. Twenty-three and twenty-two patients had extubation failure (within 19 days) and 28-days mortality, respectively. In the baseline clinical characteristics of critically ill COVID-19 patients, we found no significant differences in the extubation and mortality status groups. In the laboratory profiles of critically ill COVID-19 patients, we found no significant differences in the extubation status groups. In critically ill COVID-19 pneumonia patients, there was a significant high D-dimer levels in survived group (p = .027), a significant low BALF CD4 T-cells count in the right lung (p = .001) and a significant low BALF CD4 T-cells count (p = .010 and p = .018) in severely affected lung with extubation failure and mortality. Conclusions BALF CD4 T-cells count evaluation of severely affected lung is associated with early extubation failure and mortality in critically ill COVID-19 pneumonia patients. KEY MESSAGE Few studies have been conducted during the peak COVID-19 period analysing combined bronchoalveolar lavage fluid (BALF) immunopathology biomarkers within four hours of intubation to assess extubation failure and mortality. In this study, we reported eight BALF immunopathology biomarkers (sTREM-1, alveolar macrophage, IL-6, IL-17, CD4 T-cells, Tregs, SP-A and Caspase-3). We found significantly low BALF CD4 T-cells count in the right lung, and low BALF CD4 T-cells count in severely affected lung of critically ill COVID-19 pneumonia patients in extubation failure and mortality.

The main aim of this study is to describe BALF immunopathology based on severity in right and left lung of critically ill COVID-19 pneumonia patients on mechanical ventilation hospitalized in Resuscitation Emergency Unit/Intensive Care Unit (REU/ICU). We analysed BALF biomarkers which may be associated with extubation failure (within 19 days) and 28days mortality.  (amount and function), IL-6, IL-17, CD4 Tcells, Tregs, SP-A and Caspase-3) were recorded and analysed. This study has been approved by Ethics Committee Universitas Indonesia (approval number: KET-171/UN2.F1/ETIK/PPM.00.02/2020) and Ethics Committee of Health Research, Cipto Mangunkusumo Hospital. Informed consent was approved by ethics committee (approval number: 0995/rev00/KEP/2020). Informed consent for bronchoscopy was signed by patient's family member (patients were intubated).

Study design
The present study was designed as an observational cross-sectional study, in which eligible patients were classified into extubation (success and failure) group and mortality (survived and non-survived) group. Critically ill COVID-19 pneumonia patients will be further analysed to evaluate BALF immunopathology.

Definition of extubation failure and mortality
The 19-days extubation failure was selected based on the study reported by Gamberini et al. (median 20days extubation failure) [6], and 28-days mortality observation was chosen according to the reported pneumonia studies on mortality prediction [5,9]. Extubation failure was defined as patients being reintubated and/or reusing ventilator after successful extubation in 48 h, and/or death within 19 days [6,41]. Twenty-eight days mortality was defined as patients being recorded died in REU/ICU. Patients' mortality outside REU/ICU (i.e. inward, after a hospital discharged) was not recorded as 28-days mortality.

Assessment of lung severity
Radiographic chest X-ray findings were reported by radiologists prior to BALF collection. Severely and less severely affected lung were discussed by two respirologist and critical illness consultants, plus one internist, based on the affected lung chest imaging severity scoring proposed by Feng et al. [42].

Bronchoalveolar lavage fluid collection
Early BALF collection was performed after patients' intubation. Intravenous midazolam and propofol were given by the anaesthesiologist in order to give optimum sedation during bronchoscopy. The order of BAL suctioning was initially performed from the less severely affected lung and proceeded to the severely affected lung, from the subsegment of the right middle lobe and lingula of the left lung. Severely and less severely affected lung BALF were analysed separately. Bronchoalveolar lavage was performed (standard guidance [43]) by serial 20 mL fractions 0.9% of normal saline solution to a total volume of 100 mL (room temperature). Minimum of 60-70% of lavage volume was retrieved by gentle syringe suction collected to the mucus extractor in a wedge position and processed for further examination within 2 h. Patients were observed for one-hour post-procedure.

Bronchoalveolar lavage fluid preparation
The BALF specimen containers were inserted in sterilized medical plastic bag prior to transfer to the Integrated Laboratory of medical faculty, Universitas Indonesia. Specimen containers were gathered in a ventilated room. All specimen handling was coordinated by experienced laboratory staff with a sufficient protective equipment including protective, disposable aprons, N95 masks, goggles and double layer hand gloves. Specimen volume was evaluated for appearance, colour, clearness and contamination with intrabronchial blood. Bronchoalveolar lavage fluid specimens were collected to 50 mL tube and then centrifuged at 1000Âg for 10 min. Bronchoalveolar lavage fluid supernatant was separated to analyse sTREM-1, IL-6, IL-17 and SP-A and frozen at À80 C. Bronchoalveolar lavage fluid pellet was suspended in 2 mL PBS (phosphate buffer saline) to analyse alveolar macrophage, Tregs and Caspase-3. For Caspase-3, sample was frozen at À80 C prior to transfer to Integrated Laboratory of Medical Faculty, Universitas Indonesia.

Bronchoalveolar lavage fluid analysis
Flow cytometry was used to analyse alveolar macrophage, CD4 T-cells and Tregs. Enzyme-linked immunosorbent assay (ELISA) was used in duplicate to analyse sTREM-1 (MyBioSource ELISA kits, San Diego, CA), IL-6 and IL-17 (R&D systems quantikine ELISA kits, Minneapolis, MN), SP-A (LSBio's ELISA kits, Seattle, WA) and Caspase-3 (Cusabio ELISA kits, Houston, TX). Interleukin-6, IL-17 and SP-A were observed from BALF supernatant. Caspase-3 was observed from cell pellet homogenate and were extracted by using freeze-thawing method (this process was done for two times). Since BALF protein level is in the same concentration as in blood, total protein levels were used as an index of BALF dilution [44]. Bradford technique (Bio-RAD, Hercules, CA) was used to measure normalized protein in BALF supernatant or cell pellet.

Flow cytometry BALF analysis
The obtaining cell pellet (alveolar macrophage and Tregs) was incubated at room temperature in the dark, with monoclonal antibodies (mAbs) in 5 mL polystyrene tubes for 15 min, followed by insertion of FACS lysing solution for 15 min. The working panel of mAbs at five colour assays used for alveolar macrophage and Tregs BALF evaluation was the following: anti-human CD206 PE, anti-human HLA-DR FITC, antihuman CD11b APC, anti-human CD45 PerCP-Cy5.5 and anti-human CD169 BV421 (BD Biosciences, Franklin Lakes, NJ). The resulting cell pellet of CD4 T-cells was incubated at room temperature in the dark, with mAbs in BD TruCount TM Tubes for 15 min, followed by insertion of FACS lysing solution for 15 min at room temperature with following mAbs working panel: CD45þ/CD3þ/CD4þ/CD8þ (BD Biosciences, Franklin Lakes, NJ). Specimens were acquired with a FACSCanto TM II flow cytometer (BD Biosciences, Franklin Lakes, NJ) after been washed. BD FASCDiva TM software version 6.1.3 (BD Biosciences, Franklin Lakes, NJ) was used to perform the cytometric analysis.
Amount and function of alveolar macrophage were analysed according to procedure proposed by Nayak et al. [45] Alveolar macrophage (amount) was analysed according to percentage macrophage cells/CD45 þ , HLA-DR þ and CD11b þ . Alveolar macrophage (function) was analysed according to the mean fluorescence intensity (MFI) of CD169 þ .

Statistical analysis
Numerical and categorical variables were reported through mean, and standard deviation or median, and interquartile range (IQR), and proportions. Data normality was assessed based on Shapiro-Wilk's test. A t-test was used to compare normal distribution variables (parametric data), and a Mann-Whitney Wilcoxon test was used to compare non-normal distribution continuous variables (non-parametric data). A Chisquare test was used for comparing categorical variables. Univariate analysis was performed for subject characteristics (demographic, clinical variables and laboratory values) and bivariate analysis was performed for biomarker variables (sTREM-1, alveolar macrophage, IL-6, IL-17, CD4 T-cells, Tregs, SP-A and Caspase-3). All p values <.05 were considered statistically significant. SPSS Statistic version 26 software (IBM Corporation, Armonk, NY) was used to analyse all recorded data.

Results
Based on our national guidelines, all critically ill COVID-19 pneumonia patients in this study received intravenous dexamethasone, N-acetyl cysteine, low dose vitamin D supplementation, empirical antibiotics; and four patients received tocilizumab. Bronchoalveolar lavage fluid collection was performed in an average of 4 h (SD ¼ 0.82) after patients' intubation. Twenty-three and twenty-two patients had extubation failure and mortality, respectively.
Among 40 severe pneumonia patients, 28 critically ill COVID-19 pneumonia patients were further analysed to evaluate BALF immunopathology. Twelve patients were excluded (non-COVID-19) from this study. The baseline clinical characteristics of critically ill COVID-19 patients based on extubation and mortality status are shown in Tables 1 and 2, respectively. In the extubation status, we found no significant differences between group of extubation success and extubation failure of critically ill COVID-19 patients, as reported in Table 1.
In the mortality status, we found no significant differences between group of survived and non-survived of critically ill COVID-19 patients, as reported in Table 2.
Radiographic findings were reported by Radiologists in Cipto Mangunkusumo General (National Tertiary Referral Hospital). Chest X-ray findings in critically ill COVID-19 patients predominantly showed bilateral consolidation (n ¼ 27) with severely affected right lung (n ¼ 19), as shown in Appendix 3. The laboratory profiles of critically ill COVID-19 patients based on extubation and mortality status are shown in Tables 3 and 4, respectively. In the extubation status, we found no significant differences between group of extubation success and extubation failure of critically ill COVID-19 patients, as reported in Table 3.
In the mortality status, we found a significant difference of D-dimer between group of survived and nonsurvived of critically ill COVID-19 patients (12,646 vs. 1800 lg/L; p ¼ .027), as reported in Table 4.
In the analysis of severely and less severely affected right and left lung, there were significantly lower BALF CD4 T-cells count in the right lung of critically ill COVID-19 patients (14 vs. 28 cells/lL; p ¼ .001), as reported in Table 5.
Bronchoalveolar lavage fluid immunopathology in severely affected lung of critically ill COVID-19 pneumonia with extubation success and failure Bronchoalveolar lavage fluid immunopathology profiles are shown in Tables 6 and 7. In the group of extubation failure, we found significantly lower BALF CD4 T-cells count (56 vs. 11 cells/lL; p ¼ .010), as reported in Table 6.

Bronchoalveolar lavage fluid immunopathology in severely affected lung of survivors and nonsurvivors critically ill COVID-19 pneumonia
In the group of non-survivors, we found significantly lower BALF CD4 T-cells count (40 vs. 11 cells/lL; p¼.018), as reported in Table 7.
Bronchoalveolar lavage fluid immunopathology biomarkers in less severely affected lung of critically ill COVID-19 are reported in Appendixes 1 and 2.

Discussion
The role of biomarkers has been suggested to guide clinicians for ventilator management in determining the potential risk of extubation failure and mortality [12]. This study implicates that the role of BALF CD4 T-cells as a biomarker is reliable for early management prediction in critically ill COVID-19 pneumonia patients (i.e. ventilator management, early tracheostomy and the need for aggressive treatment) which could shorten the length of hospital stays and to determine the readiness for extubation in patients who are at high risk for extubation failure.
Based on sample size calculation, 1010 patients were required in this study. Due to pandemic beta variant COVID-19 in Indonesia (November 2020 to January 2021), after the enrolment of 40 patients, an internal discussion was held, and the decision was to discontinue patients' enrolment. To our knowledge, only a few studies were conducted during the peak COVID-19 period analysing BALF immunopathology to assess extubation failure and mortality [46]. Additionally, this is the first study to perform early bronchoscopy after patients' intubation (average of  Based on extubation and mortality status, we also observed clinical characteristics and laboratory profiles of critically ill COVID-19 pneumonia patients. D-dimer levels were significantly higher in the group of survived rather than non-survived. The role of elevated D-dimer in critically ill COVID-19 has been associated with disease severity and mortality [51]. Other studies have found elevated    D-dimer to be associated with extubation failure and death [52,53]. Our study found a different result for elevated D-dimer probably due to small sample size, no serial D-dimer evaluation, and selection bias. Several studies reported bilateral lung consolidation as frequent abnormalities findings in COVID-19 patients [54,55]. In this study, the majority of patients' chest radiographs were characterized by bilateral diffuse consolidation, with the right lung being the most severely affected. Differences in anatomical structure led to a higher likelihood of foreign particles and pathogens entering the lower respiratory tract of the right lung [56][57][58]. Hence, the severity of inflammation will predominate in the right lung (marked by significantly low BALF CD4 T-cells count), as shown in this study. In the severe inflammation process, CD4 T-cells will differentiate into several subsets, including Tregs [59]. This study also displayed a high level of Tregs in extubation success and survived groups, albeit statistically not significant. This result indicates the vital function of Tregs as an immune suppressor in helping the adaptive immune process [60], and its role in the lung epithelial regeneration and surfactant production [61], as well as the potential effect on the resolution of the lung [62]. Future studies are required to investigate the detailed mechanisms of Tregs and the association with extubation success or extubation failure, which may better explain this present study.
Limited data are available for assessing extubation failure in critically ill COVID-19 pneumonia patients [7]. Information focussing on extubation failure in severe pneumonia patients has been reported by Yu et al. [10] Critically ill COVID-19 pneumonia patients need more prolonged intubation time [63], and Ionescu et al. [7] reported extubation failure in critically ill COVID-19 pneumonia was strongly associated with hospital mortality. Length of stays, requirement of tracheostomy and mortality were associated with extubation failure [64].
A BALF study by Saris et al. found that critically COVID-19 pneumonia patients showed a reduced levels of BALF CD4 T-cells count in non-survivor compared to the survivor group [46]. This study reports a significant low levels of BALF CD4 T-cells count in severely affected lung of critically ill COVID-19 pneumonia patients in extubation failure and non-survived group.
It is recommended for clinicians to evaluate BALF in severe pneumonia patients [65]. However, limited BALF immunopathology studies were found in severe pneumonia patients. Meta-analysis by Zeng et al. [66] reported serum immunopathology in critically ill COVID-19 pneumonia is associated with disease severity and high mortality. Sputum and BALF inflammatory biomarkers, IL-6, IL-8, IL-10, IL-17, IL-1b, IL-1ra, TNF-a and IFN-c, have been reported in several studies [1,13]. Bronchoalveolar lavage fluid immunopathology biomarkers have been described to predict clinical outcomes in COVID-19 pneumonia patients [67]. Eight BALF immunopathology biomarkers were assessed in this study, sTREM-1, alveolar macrophage, IL-6, IL-17, CD4 T-cells, Tregs, SP-A and Caspase-3. Based on the immunopathophysiology, the selection of eight biomarkers in this study (sTREM-1, alveolar macrophage amount and function, IL-6, IL-17, CD4 T-cells, Tregs, SP-A and Caspase-3) could provide a better comprehensive in explaining the role of local innate and adaptive immune responses and its association with extubation failure and mortality. The presence of infectious disease can be marked by the presence of sTREM-1 [68]. Alveolar macrophage as a phagocyte plays a role in homeostasis, body protection, response to foreign particles, and tissue regeneration. Additionally, alveolar macrophage function regulates local immunology, especially lung immunity. The interaction of each cell will mediate the immune response. Alveolar macrophages will secrete cytokines and eicosanoid for neutrophil polymorphonuclear recruitment [69]. Kooguchi et al. [22] found that downregulation of alveolar macrophages will increase the severity of lung injury and mortality of pneumonia, and a study conducted by Puren et al. [23] found that a decrease/increase in cytokine (IL-6 and IL-17) levels are associated with severity of infection and inflammation. A study by Jedynak et al. [70] found that an increase in lung infection will stimulate sTREM-1 to activate monocyte and neutrophil and produce more inflammatory cytokines. Insufficient protection of alveolar macrophages to the infection will initiate adaptive immunity and activate lymphocytes (T cells), which then induce apoptosis (upregulated Caspase-3 and downregulated CD4 T-cells) [71]. Bielosludtseva et al. reported that reduction of T cell levels is associated with high mortality in pneumonia [30]. Additionally, a study by Shi et al. [38], reported that an increase in Caspase-3 is associated with increased alveolar epithelial apoptosis. Persistent severe lung injury will dysregulate lung epithelium regeneration (Tregs) and reduce surfactant production [61]. A study by Mukherjee et al. [36] reported that downregulated surfactant (SP-A) production will cause lung atelectasis and gas exchange disturbance. Previous studies explained that the reduction of immune cells and mediators will cause individual be more prone to diseases (especially infectious diseases) and/or will aggravate the existed disease condition [72]. This study reports significantly low levels of BALF CD4 T-cells count in extubation failure (p ¼ .010) and non-survived (p ¼ .018) critically ill COVID-19 pneumonia patients. Jiang et al. [8] reported a significantly low serum CD4 T-cells count in non-survived COVID-19 pneumonia patients. Vedder et al. [73] described normal BALF CD4 T-cells in ventilated COVID-19 pneumonia patients (BALF was collected within two days after mechanical ventilation), whereas in this study BALF was collected in an average of 4 h (SD ¼ 0.82) after patients' intubation. No significant differences were found in other BALF immunopathology biomarkers (sTREM-1, alveolar macrophage, IL-6, IL-17, Tregs, SP-A and Caspase-3) for both groups (extubation failure and mortality) in this study.
Some study limitations included: (1) the small number of study subjects may result in other biomarkers that were not statically significant. The lack of statistical significance in some BALF biomarkers may be explained by our current sample size being too small to detect the difference. Therefore, future studies with larger sample size are suggested to investigate the differences in other BALF biomarkers. (2) Chest CT scan imaging was not performed; hence, we could not assess other abnormalities that might be presented in the study patients. (3) No serial evolution time for BALF biomarkers to assess the dynamic changes.

Conclusions
This study reports the association of low BALF CD4 Tcells count with early extubation failure and mortality in the severely affected lung of critically ill COVID-19 pneumonia patients. BALF CD4 T-cells count evaluation of severely affected lung can be an option to clinicians for assessment of early extubation failure and mortality in critically ill COVID-19 pneumonia patients. Illness, Internal Medicine Department); Utami (Research Coordinator unit) for their contributions in conducting this study. Registration

Patient consent
Informed consent for bronchoscopy was signed by patient's family member (patients were intubated).

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

Data availability statement
Our study data are available on request from the corresponding author (G.S.). The data are not publicly available due to their containing information that could compromise the privacy of study subjects.

Funding
The author(s) reported there is no funding associated with the work featured in this article.