In Patients with Severe COVID-19, the Profound Decrease in the Peripheral Blood T-Cell Subsets Is Correlated with an Increase of QuantiFERON-TB Gold Plus Indeterminate Rates and Reflecting a Reduced Interferon-Gamma Production

Increased rates of indeterminate QuantiFERON-TB Gold Plus Assay (QFT-Plus) were demonstrated in patients hospitalized with Coronavirus Disease (COVID)-19. We aimed to define the prevalence and characteristics of hospitalized COVID-19 patients with indeterminate QFT-Plus. A retrospective study was performed including hospitalized COVID-19 patients, stratified in survivors and non-survivors, non-severe and severe according to the maximal oxygen supply required. Statistical analysis was performed using JASP ver0.14.1 and GraphPad Prism ver8.2.1. A total of 420 patients were included, median age: 65 years, males: 66.4%. The QFT-Plus was indeterminate in 22.1% of patients. Increased rate of indeterminate QFT-Plus was found in non-survivors (p = 0.013) and in severe COVID-19 patients (p < 0.001). Considering the Mitogen-Nil condition of the QFT-Plus, an impaired production of interferon-gamma (IFN-γ) was found in non-survivors (p < 0.001) and in severe COVID-19 patients (p < 0.001). A positive correlation between IFN-γ levels in the Mitogen-Nil condition and the absolute counts of CD3+ (p < 0.001), CD4+ (p < 0.001), and CD8+ (p < 0.001) T-lymphocytes was found. At the multivariable analysis, CD3+ T-cell absolute counts and CD4/CD8 ratio were confirmed as independent predictors of indeterminate results at the QFT-Plus. Our study confirmed the increased rate of indeterminate QFT-Plus in COVID-19 patients, mainly depending on the peripheral blood T-lymphocyte depletion found in the most severe cases.


Introduction
In December 2019, the first cases of pneumonia caused by a new coronavirus, named Severe Acute Respiratory Syndrome Coronavirus-2 (SARS-CoV-2), was reported in Wuhan, province of Hubei, China, and quickly spread to the rest of the world, causing the Coronavirus Disease (COVID)-19 pandemic. [1][2][3]. The virus can cause a wide variety of clinical manifestations, from asymptomatic disease to severe pneumonia with respiratory failure, the retrospective nature of the study the requirement for informed consent was waived by the Ethics Committee, as permitted by the local legislation.
In our hospital, QTF-Plus assay was routinely performed in all COVID-19 hospitalized patients to identify the presence of latent tuberculosis infection (LTBI). On hospital admission, peripheral blood lymphocyte absolute counts were also routinely assessed in COVID-19 patients. Both tests were used to better characterize COVID-19 patients, considering the possibility of starting a treatment with anti-IL-6 receptor and other immunomodulatory therapies.
Patients who fulfilled the following criteria were enrolled in the study: 1. hospitalized at the infectious diseases ward due to SARS-CoV-2 infection, as documented by a positive reverse transcription-polymerase chain reaction (RT-PCR) for SARS-CoV-2 RNA detection on a nasopharyngeal swab (NPhS), either symptomatic or asymptomatic; 2. QFT-Plus assay performed during hospitalization within 30 days from the first positive NPhS. Asymptomatic patients were hospitalized in the infectious disease ward because of concomitant clinical conditions that required inpatient care associated with SARS-CoV-2 transmission prevention, although they did not show any sign or symptom related to SARS-CoV-2 infection. In the manuscript, both symptomatic and asymptomatic patients are referred to as COVID- 19.
During the study period, demographics, clinical, and laboratory data of hospitalized COVID-19 patients were collected. The following information were recorded in an electronic ad hoc created database: age, sex, outcome, need for Intensive Care Unit (ICU), comorbidities, date of symptom onset, and first positive RT-PCR for SARS-CoV-2 on a NPhS, oxygen support, treatments, QFT-Plus assay results, peripheral blood lymphocyte subset assessment, neutrophil and lymphocyte counts, inflammation-related parameters (C-reactive protein (CRP), IL-6, tumor necrosis factor (TNF)-α, fibrinogen, ferritin, D-dimer). Neutrophils and lymphocyte counts, as well as inflammation-related parameters registered in the database and included in the statistical analysis were performed on admission to the Infectious Disease ward.

Definitions
The QFT-Plus assay consists of four vacutainer tubes: TB1 detects CD4+ T-cell response; TB2 is optimized for detection of CD4+ and CD8+ T-cell responses; the Mitogen tube represents the positive control, while the Nil tube represents the negative control.
In a typical QFT-Plus assay, the level of INF-γ in the Nil condition should be <8.0 IU/mL, and is subtracted to the other conditions (TB1-Nil, TB2-Nil, Mitogen-Nil). The test gives 3 possible results: a. positive when INF-γ level in Nil condition is ≤8.0 IU/mL and in TB1-Nil and/or TB2-Nil is ≥0.35 IU/mL and ≥25% of Nil; b. negative when INF-γ level in Nil condition is ≤8.0 and in Mitogen-Nil is ≥0.5 and in both TB1-Nil and TB2-Nil INF-γ levels are either ≤0.35 IU/mL or ≥0.35 IU/mL but ≤25% of Nil; and c. indeterminate when Nil is >8.0 IU/mL or Nil is ≤8.0 IU/mL and Mitogen-Nil < 0.5 IU/mL and in both TB1-Nil and TB2-Nil INF-γ levels are either ≤0.35 IU/mL or ≥0.35 IU/mL but ≤25% of Nil [20]. Positive and negative results to the QFT-Plus were considered as determinate results, in contrast to indeterminate results.
According to the outcome, patients were divided into non-survivors, if death occurred during hospitalization, or survivors, if they were discharged.
As for oxygen support, five groups were identified, based on maximal oxygen support required during the hospitalization; ambient air (AA), Venturi Oxygen Mask (VMK), Non-Rebreather Mask (NRM), Non-Invasive Ventilation (NIV), or Orotracheal Intubation (OTI) for mechanical ventilation. Oxygen groups were linked to disease severity, and patients were classified as non-severe (AA and VMK) and severe (NRM, NIV, and OTI). COVID-19 patients on oxygen-therapy received steroid treatment according to national and international guidelines available at the time of hospitalization. Furthermore, a number of patients were treated with anti-IL-6 receptor monoclonal antibodies, following the Italian Medicines Agency (AIFA) guidelines and the evidence from international literature [28][29][30].

Statistical Analysis
All data were analyzed using JASP (Version 0.14.1) and GraphPad Prism (Version ver8.2.1). Categorical variables are presented as absolute frequency and percentages (%), while quantitative variables are presented as medians and interquartile ranges (IQR). Differences between groups were assessed with the two-tailed Chi-square test for categorical data, and the non-parametric Mann-Whitney test (2 groups) or Kruskal-Wallis test (>2 groups) for quantitative data. Correlations were assessed using the non-parametric Spearman's test. Multivariable regression analysis was used to assess the factors associated to QFT-Plus indeterminate results. With this aim, all the factors significantly associated with an indeterminate QFT-Plus assay at the univariable analysis, were included in the multivariable regression model.
For all the tests, the level of statistical significance was <0.05.

Study Population
During the study period, 424 patients hospitalized for COVID-19 in the Infectious Disease Clinic of the University Hospital Policlinico Tor Vergata of Rome, were included in the study. A total of 150 were hospitalized from March to May 2020 (First Wave of COVID-19) and 274 were hospitalized from September to December 2020 (Second Wave of COVID-19). All patients had at least one positive RT-PCR for SARS-CoV-2 RNA detection on a NPhS. Four patients were excluded from the analysis because the QFT-Plus assay was performed more than 30 days after the first positive SARS-CoV-2 NPhS; therefore, 420 patients were evaluated. For 18 patients, peripheral blood T-, B-, NK-lymphocyte absolute counts were not available.
The median number of days between the onset of symptoms and the first positive NPhS for SARS-CoV-2 RNA detection was five days (IQR 1-8), while the median number of days between the onset of symptoms and the blood sampling for the QFT-Plus assay was 10 days (IQR 7-14). Median days from the first positive NPhS for SARS-CoV-2 RNA detection and the blood sampling for the QFT-Plus assay was four days (IQR 2-7). For 20 patients, the exact onset of symptoms was unknown, or they were asymptomatic for SARS-CoV-2 related symptoms. In these cases, hospitalization in the infectious disease ward was due to the presence of concomitant diseases (pneumological, hematological, solid tumors, myocardial infarction, cerebrovascular accidents) requiring inpatient care, or the need of hemodialysis, in a setting where SARS-CoV-2 containment protocols have been adopted. Statistical analyses after excluding asymptomatic patients are shown in supplementary Tables S1-S5. Peripheral blood T-, B-and NK-lymphocyte assessment was performed on the same day of the QFT-Plus assay (median elapsed time [IQR] between the two tests: 0.0 (0.0-3.0) days) ( Table 1).

Laboratory Parameters and Inflammation Markers
The analysis of inflammation markers assessed at baseline (on admission to the Infectious Disease Ward) showed increased levels of CRP (60.1 mg/L, IQR: 25.1-113.5 mg/L), Ddimers (942 ng/mL, IQR: 519-1601 ng/mL), fibrinogen (550 mg/dL, IQR: 418-684 mg/dL), and ferritin (635 ng/mL, IQR: 282-1252 ng/mL) in the overall cohort of COVID-19 patients, compared to the reference values of the central laboratory (Table 2). Table 2. Laboratory parameters at Infectious diseases ward admission of the study cohort, overall and after stratification for outcome (survivors vs. non-survivors) and COVID-19 severity (non-severe vs. severe).  After stratifying patients according to the outcome and disease severity, lymphocyte counts were reduced in non-survivors and severe patients compared to survivors and non-severe patients (p < 0.001 and p < 0.001, respectively). Conversely neutrophils counts were increased in non-survivors and severe patients compared to survivors and non-severe patients (p < 0.001 and p < 0.001, respectively). Therefore, the neutrophil-to-lymphocyte (N/L) ratio was significantly increased in non-survivors and severe patients compared to survivors and non-severe patients (p < 0.001 and p < 0.001, respectively). CRP, IL-6, D-dimers, fibrinogen, and ferritin were significantly different after comparing survivors with non-survivors and non-severe with severe COVID-19 patients (Table 2). Although TNF-α did not significantly differ after comparing survivors vs. non-survivors, its levels were lower in severe than non-severe COVID-19 patients (p < 0.001) ( Table 2).

QFT-Plus Assay Results
The QFT-Plus assay results were defined as determinate (either positive or negative) in 327 patients, with 292 (70%) negative and 35 (8%) positive results. A total of 93 patients (22%) had an indeterminate QFT-Plus assay. After stratifying patients according to the outcome and disease severity, higher rates of indeterminate results were found Median values of IFN-γ production for TB1 and TB1-Nil, TB2 and TB2-Nil, Mitogen and Mitogen-Nil tubes of the QFT-Plus assay are reported in Table 3. Overall, reduced levels of IFN-γ production were found in TB1, TB1-Nil, TB2 and TB2-Nil conditions in nonsurvivors and severe patients compared to survivors and non-severe COVID-19 patients (Table 3). We focused on the Mitogen-Nil condition, and significantly lower levels of IFN-γ production were found in non-survivors compared to survivors (1.3 IU/mL vs. 5.5 IU/mL, p < 0.001) and in severe compared to non-severe patients (1.7 IU/mL vs. 8.4 IU/mL, p < 0.001) ( Table 3).
After stratifying patients according to the maximal oxygen supply received during hospitalization, the rates of indeterminate QFT-Plus assay were significantly increased in patients treated with non-invasive or invasive ventilation compared to patients treated with Venturi Mask or not needing oxygen supplementation (QFT-Plus determinate: 38.5%, 19.6%, 7.3%, 23.2%, and 11.3%; QFT-Plus indeterminate: 4.3%, 20.4%, 11.8%, 37.6%, and 25.8%, for AA, VMK, NRM, NIV, and OTI, respectively; Chi 2 test p < 0.001) (Figure 1). Furthermore, a decreasing trend of IFN-γ production in the Mitogen-Nil condition was found proceeding from patients who did not need oxygen therapy (AA, 9.63 IU/mL), through patients with VMK (2.68 IU/mL), NRM (1.90 IU/mL) and NIV (1.92 IU/mL), to patients who underwent OTI (0.97 IU/mL) for mechanical ventilation (p < 0.001). Additionally, similar results were obtained after removing from the analysis patients with an indeterminate result at the QFT-Plus assay (p < 0.001). Thus, increasing disease severity was associated with decreasing IFN-γ levels produced in the Mitogen-Nil condition of the QFT-Plus assay (Figure 1). through patients with VMK (2.68 IU/mL), NRM (1.90 IU/mL) and NIV (1.92 IU/mL), to patients who underwent OTI (0.97 IU/mL) for mechanical ventilation (p < 0.001). Additionally, similar results were obtained after removing from the analysis patients with an indeterminate result at the QFT-Plus assay (p < 0.001). Thus, increasing disease severity was associated with decreasing IFN-γ levels produced in the Mitogen-Nil condition of the QFT-Plus assay (Figure 1).

QFT-Plus Mitogen-Nil Results and Correlation with Neutrophil and Lymphocyte Counts, Lymphocyte Subset Absolute Counts and Inflammation Markers
COVID-19 patients were stratified according to the QFT-Plus assay results in determinate and indeterminate. Table 4 reports characteristics of the two subpopulations. Patients with an indeterminate QFT-Plus were prevalently males (p = 0.041), older (p = 0.003), and showed higher ICU admission rates (p = 0.002) than patients with a determinate result. Importantly, time of blood sampling for QFT-Plus, T-, B-, NK-lymphocyte subset assessment, and NPhS for SARS-CoV-2 RNA detection did not significantly differ between the two groups (Table 4).

QFT-Plus Mitogen-Nil Results and Correlation with Neutrophil and Lymphocyte Counts, Lymphocyte Subset Absolute Counts and Inflammation Markers
COVID-19 patients were stratified according to the QFT-Plus assay results in determinate and indeterminate. Table 4 reports characteristics of the two subpopulations. Patients with an indeterminate QFT-Plus were prevalently males (p = 0.041), older (p = 0.003), and showed higher ICU admission rates (p = 0.002) than patients with a determinate result. Importantly, time of blood sampling for QFT-Plus, T-, B-, NK-lymphocyte subset assessment, and NPhS for SARS-CoV-2 RNA detection did not significantly differ between the two groups (Table 4).

Discussion
The main findings of this work have been the evidence of increased rate of indeterminate QFT-Plus assay in COVID-19 patients, with higher percentages observed in patients with a worse outcome and more severe disease. Indeterminate results of the QFT-Plus assay were due to an impaired IFN-γ production upon PHA stimulation. Furthermore, IFN-γ levels assessed in the mitogen tube of the QFT-Plus assay were directly correlated with the absolute count of CD3+ T-lymphocytes and inversely correlated with the CD4/CD8 ratio.
In this study, the prevalence of indeterminate QFT-Plus assay in COVID-19 hospitalized patients (93/420, 22%) was higher than it would be expected if compared to the general population, according to previous studies involving both immune competent and immunocompromised patients [31][32][33]. Conversely, compared to other publications on COVID-19 subjects, the rate of indeterminate QFT-Plus results reported in our work is slightly inferior (22% compared to 35-65%) [16,17,21,22]. A possible explanation could be the different selection criteria for the execution of the QFT-Plus assay. In other studies, only a limited number of patients were tested, with possible selection biases [16,17,22], while in our study all the patients hospitalized because of COVID-19 were tested with the QFT-Plus assay on hospital admission, for the diagnosis of latent tuberculosis infection. This systematic approach, reduced the selection biases, allowing us to analyze a large cohort of patients (420 subject with a QFT-Plus assay performed on hospital admission), with a good representation of all the degree of severity observed in COVID-19 hospitalized patients. The median time between the onset of symptoms and the blood sampling for the QFT-Plus assay was 10 days (IQR [7][8][9][10][11][12][13][14], in line with previous studies [16].
Interestingly, the results of the present study showed an association between indeterminate QFT-Plus assay, assessed early during hospitalization, and COVID-19 severity, defined by the maximal oxygen support required during the entire hospitalization. Moreover, we confirmed the association of a QFT-Plus indeterminate result with in hospitalmortality [16,22]. Furthermore, an association has emerged between indeterminate QFT-Plus results, age > 65 years and male sex, which are well known risk factors for severe forms of COVID-19 [3]. A higher prevalence of indeterminate QFT-Plus assay in our series was observed in patients who needed intensive care, due to a greater COVID-19 severity thus evidencing a limitation of the QFT-Plus assay for the identification of latent tuberculosis in acute COVID-19 critically ill patients.
While other studies showed an association between indeterminate QFT-Plus assay and immunosuppressive treatment with monoclonal antibodies against IL-6-receptor [16] and corticosteroids [16,[33][34][35], in our series we could not identify such a statistical association. This could be due to different aspects: (1) QFT-Plus assay was generally performed before steroid and anti-IL-6 receptor administration; (2) a limited number of patients included in the analysis received IL-6 receptor inhibitors (<10%, 41/420). After removing from the analysis all the patients who received steroids and IL-6-receptor inhibitors before the blood sampling for QFT-Plus assay, an increased rate of indeterminate results was confirmed in patients with more severe disease and poorer outcome, as shown by other authors [22]. It should be considered that the QFT-Plus assay was generally performed on admission to the Infectious Disease Ward, before patients underwent steroid or IL-6 receptor inhibitors, although in many cases a previous low dose home steroid treatment cannot be excluded with certainty.
It is known that the QFT-Plus assay results are influenced by a dysregulated immune system, such as in patients with inflammatory bowel disease on treatment with corticosteroid and/or TNF-α inhibitors [34,35] and in HIV-positive patients [36,37].
Moreover, several studies have shown a link between indeterminate QFT-Plus assay results and lymphopenia [38,39]. Lymphopenia often characterizes the most severe forms of SARS-CoV-2 infection [4,26,27] supporting the hypothesis of a direct correlation with increased rate of indeterminate QFT-Plus assay [16,17,21]. Considering that peripheral blood lymphocyte subsets were studied in almost all the patients included in the analysis, we could demonstrate that patients with indeterminate QFT-Plus assay had reduced circulating numbers of total CD3+, CD3+CD4+, CD3+CD8+ T-lymphocytes, and CD3negCD16+CD56+ NK-cells. No differences were observed for B-lymphocytes. Interestingly, patients with indeterminate results had also an increased CD4/CD8 ratio. Furthermore, total T-lymphocyte, CD4 and CD8 subset absolute counts were directly correlated to the IFN-γ production after PHA stimulation.
In some studies, the increased rate of QFT-Plus indeterminate results was attributed to the "cytokine storm" associated with the severe forms of COVID-19 [22]. Our results do not support this hypothesis, for several reasons. Although inflammation markers (CRP, D-dimers, fibrinogen, ferritin) were predictive of and found increased in patients with indeterminate QFT-Plus at the univariable analysis, only CD3+ T-cells count and CD4/CD8 ratio remained as independent predictors of an indeterminate result of the QFT-Plus assay, at the multivariable analysis. Moreover, IL-6 levels did not significantly differ in COVID-19 patients with a determinate or indeterminate QFT-Plus assay. TNF-α levels were even reduced in patients with an indeterminate result at the QFT-Plus assay compared to patients with a determinate result. Finally, in all the cases of indeterminate results, in the Nil tube (representing the basal unstimulated IFN-γ production) the levels of IFN-γ were not increased, while in the Mitogen-Nil condition the IFN-γ levels were below 0.5 IU/mL, demonstrating an impairment in the production of this cytokine, rather than a dysregulated hyperproduction. Taken together all these elements lead to the conclusion Life 2022, 12, 244 13 of 16 that in our cohort the increased rate of indeterminate QFT-Plus results depends on the reduction in T-lymphocyte absolute counts found in the most severe cases of SARS-CoV-2 infection, with an imbalance of CD4/CD8 ratio due to a prevalence of CD4 and reduction in CD8 T-cells. Nevertheless, we cannot completely rule out a possible role played by lymphocyte dysfunction.
Besides the reduction in circulating T-lymphocyte absolute counts, some authors demonstrated the presence of phenotypic and functional abnormalities in COVID-19 patients [11]. De Biasi et al. showed several alterations involving naïve, central memory, effector memory, and terminally differentiated T-cells, as well as regulatory T-cells and PD1+CD57+ exhausted T-cells [40]. An impairment in T-lymphocyte function, with reduced capacity to produce IFN-γ and TNF-α upon stimulation with anti-CD3/anti-CD28 monoclonal antibody has been evidenced in COVID-19 patients. Interestingly, IL-17 seemed to restore in vitro IFN-γ production by T-lymphocyte of COVID-19 patients [4]. Another consideration is the fact that in HIV patients with CD4 cell count below 100 cell/µL, the rate of indeterminate results of the QuantiFERON-TB Gold test is inferior to the value we and other groups have observed in COVID-19 patients [36], thus indicating the concomitant contribution of T-lymphocyte dysregulation together with absolute count reduction.
Our results demonstrated a reduced production of IFN-γ by T-lymphocytes upon PHA stimulation in COVID-19 patients with a more severe disease. Furthermore, a negative correlation between inflammatory markers (such as CRP, ferritin, D-dimers) and IFN-γ production by peripheral blood T-lymphocyte was also identified. These results seem to contrast with the evidence obtained by Karki et al., who demonstrated that the combination of IFN-γ and TNF-α was able to recapitulate the cellular and tissue damages observed in severe COVID-19 patients with acute respiratory distress syndrome (ARDS), both in vitro cultures and in vivo animal models. Moreover, IFN-γ and TNF-α blocking was able to increase the survival of SARS-CoV-2 infected mice [41]. T-cell dysregulation during COVID-19 disease is characterized by complex and heterogeneous series of events, which encompass both T-cell hyperactivation and exhaustion [42]. Moreover, T-cell count reduction and impaired IFN-γ production in peripheral blood can reflect activated T-cell margination, recruitment and sequestration in lung tissue and lymphoid organs with local hyperinflammation and secretion of inflammatory cytokines [42]. Considering this complexity, more studies comparing T-cell number, phenotype and function in both peripheral blood and lung tissue are needed, to clarify the role of these immune cells in the pathogenesis of COVID-19.
The strengths of this study are represented by the large number of COVID-19 patients included and the systematic approach of performing the QFT-Plus assay on hospital admission, avoiding selection biases. Some limitations are represented by the retrospective design and the cross-sectional assessment of the QFT-Plus assay. It would be useful to repeat the QFT-Plus after the acute phase of the disease, in patients who survived COVID-19 and normalized peripheral blood lymphocyte counts. Being a monocentric study, our results need to be confirmed and verified in other settings with different cohort of COVID-19 patients.

Conclusions
In conclusion, our study confirmed an increased prevalence of indeterminate QFT-Plus assay in a large cohort of patients hospitalized because of SARS-CoV-2 infection, and demonstrated a direct link between the impaired IFN-γ production in the Mitogen-Nil condition and the reduction in peripheral blood T-lymphocytes in COVID-19 patients.

Informed Consent Statement:
Considering the retrospective nature of the study the requirement for informed consent was waived by the Ethics Committee, as permitted by the local legislation.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy and ethical concerns.