Introduction

Coronavirus disease 2019 (COVID-19), caused by infection with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), poses the most urgent threat to global health.1–5 As of 13 April 2023, the WHO reported that there were over 762 million confirmed cases of COVID-19, including over 6.8 million deaths.6 For patients hospitalized with COVID-19, one-quarter to one-third of patients will further develop COVID-19-associated acute respiratory distress syndrome (CARDS).7 Additionally, rapid disease progression and comorbidities can result in poor prognosis.8–14 Therefore, early detection of severe COVID-19 cases and timely treatment are of vital importance.

Pre-existing weakened immune system can result in out of control infections. Defective cellular immunity, an important part of host immune response dysregulation, plays a crucial role in the pathophysiology of COVID-199,15–19 and MERS-CoV,20 ICU-acquired infection21 and increased mortality of ICU patients.22–24 SARS-CoV-2 can alter host immune responses, including innate and adaptive immune responses.25 Overall, decreased lymphocytes and subsets, especially T and B subsets, were found16,17 and were strongly associated with disease progression and adverse outcomes in COVID-19 patients with diabetes mellitus (DM) or severe disease.17 One previous study found that a model based on NK cell and CD4+ T cell counts combined with interleukin-8 had good predictive value for COVID-19 patient mortality.26 Other previous studies found that the CD8+ T-cell level reflected disease severity, and decreased CD4+ T-cell counts were independently associated with increased in-hospital mortality.27 Meanwhile, lower T lymphocyte subsets were significantly associated with a higher occurrence of composite endpoint events in COVID-19 patients.28 An increase in CD4+ T cells, CD8+ T cells, and the administration of interferon were independent predictors of clinical response within the first week after admission.29 Significantly lower lymphocyte subpopulation counts (CD3+ T cells, CD3+CD4+ T cells, CD3+CD8+T cells, CD16+CD56+ natural killer (NK) cells and CD19+ B cells) were found in ICU patients with acquired infection compared those without acquired infection.21 At ICU admission lymphopenia and T cell depletion were associated with increased mortality of ICU patients.22–24

Those above findings supported lymphocytes and subsets as predictive biomarkers and potential therapeutic targets in intensive care medicine. However, most studies on COVID-19 had small sample sizes and did not report dynamic changes in total lymphocytes and their subsets during disease progression. Additionally, associations between different lymphocyte subsets and disease severity and prognosis remain unclear, and optimal values of cutoff points of different lymphocyte subsets are also unknown. Therefore, we conducted a cohort study of 773 COVID-19 cases to investigate the dynamic changes in total lymphocytes and their subsets, as well as their abilities and optimal cutoff points in predicting disease severity and prognosis in primary infection and unvaccinated COVID-19 patients.

MethodsSubjects

This study is a cohort study that recruited a total of 773 patients, including 718 primary infection and unvaccinated COVID-19 patients (COVID-19 group) from the hospital isolation ward16,18,19 and 55 patients without COVID-19 (control group)17 from the medical examination clinic who presented to the Public Health Clinical Centre of Chengdu from January 16, 2020, to February 28, 2021 (Figure 1).16,18,19

Figure 1 Patient data. Nonsevere refers to the clinical type of COVID-19 that is asymptomatic, light, and common. Severe refers to the clinical type of COVID-19 that is associated with severe and critical illness.

Inclusion and Exclusion Criteria

Both male and female patients who were primarily infected with COVID-19 and unvaccinated were included in the COVID-19 group without any age limitation. Patients with autoimmune diseases or who have used or are using immunosuppressive agents, and patients with previous infection or vaccination were excluded from this study.

Criteria for Disease Diagnosis, Clinical Typing and Cure Definition

The criteria of disease diagnosis, clinical typing and cure definition in COVID-19 patients were based on the seventh Trial Version of the Novel Coronavirus Pneumonia Diagnosis and Treatment Guidance.8

The diagnostic criteria were cases with one of the following etiological pieces of evidence: positive results of SARS-CoV-2 detected by real-time fluorescence reverse transcription polymerase chain reaction (RT‒PCR) and positive results from viral gene sequencing.

The clinical types were classified as asymptomatic infection, mild, common, severe and critical illness. Specific classification criteria were as follows: (1) asymptomatic infection means that there are no clinical symptoms and no pneumonia manifestations on imaging; (2) mild illness means that the clinical symptoms are mild, and there are no pneumonia manifestations on imaging; (3) common illness refers to patients with clinical symptoms including fever, symptoms of respiratory tract, and pneumonia can be seen on imaging; (4) severe illness refers to patients with any of the following items: ① respiratory distress, respiratory rate (RR) ≥30 times/min; ② oxygen saturation ≤93% in the resting state; ③ arterial blood oxygen partial pressure (PaO2)/oxygen concentration (FiO2)≤300 mmHg (1 mmHg = 0.133 kPa) (in areas with high altitude (over 1000 meters above sea level), PaO2/FiO2 should be corrected according to the following formula: PaO2/FiO2 * [atmospheric pressure (mmHg)/760]); ④ pulmonary imaging shows lesions with significant progress over 50% within 24–48 hours; (5) criteria for critical illness contains one of the following conditions: ① occurrence of respiratory failure which needs mechanical ventilation; ② occurrence of shock; ③ concurrence of other organ failure which requires intensive care unit.

Patients who meet all of the following criteria can be regarded as cured and can be discharged from the hospital: (1) body temperature returns to normal for more than 3 days; (2) respiratory symptoms improve significantly; (3) lung imaging shows a significant improvement in acute exudative lesions; and (4) two consecutive sputum, nasopharyngeal swab or other respiratory specimen tests are negative for nucleic acid (sampling intervals are at least 24 hours).

Grouping Standards

Among all 773 cases, there were 718 patients in the COVID-19 group18,19 and the remaining 55 in the control group17 (Figure 1). According to disease severity, these patients were classified into the asymptomatic infection group (237 cases), mild illness group (73 cases), common illness group (371 cases), severe illness group (18 cases) and critical illness group (19 cases). Furthermore, 681 patients with asymptomatic infection or mild or common illness were assigned to the nonsevere group, while the remaining 37 severe and critical illness patients were assigned to the severe group (Figure 1).

Additionally, 710 surviving patients were included in the survival subgroup, while 8 dead patients were included in the death subgroup (Table 1 and Figure 1). For the survival subgroup, specifically, there were 237 cases with asymptomatic infection, 73 cases with mild illness, 371 cases with common illness, 18 cases with severe illness and 11 cases with critical illness (Figure 1). For the death subgroup, all 8 patients who died were initially diagnosed with critical illness.

Table 1 Baseline Information of Controls and COVID-19 Patients (n=773)

Definition of Disease Severity and Prognosis

The prognosis was classified as survival or death within four weeks after admission or within five weeks after onset. The disease severity included nonsevere (COVID-19 patients with asymptomatic infection, mild or common illness clinic type) and severe (COVID-19 patients with severe or critical illness clinic type).

Detection Methods and Analysis Methods of Lymphocytes and Subsets

Peripheral blood (2 mL) samples with EDTA anticoagulation were collected from COVID-19 patients and controls and were tested by Beckman Coulter flow cytometry (FCM) within 6 hours after collection. Two million cells were resuspended in PBS and blocking buffer (2% bovine serum albumin, 2% rabbit serum, 2% human serum, 2% mouse serum, 2% rat serum) to block unspecific binding of antibodies. Briefly, CD19+ B cell, CD3+ T cell, CD4+ T cell, CD8+ T cell and CD56+ NK cell counts (cells/μL) were measured by multiple-color FCM with human monoclonal anti-CD19-phycoerythrin (PE), anti-CD3-fluorescein isothiocyanate (FITC), anti-CD4-PE, anti-CD8-allophycocyanin (APC) and anti-CD56-PE antibodies (BD Multitest) according to the manufacturer’s instructions. After 30 min of light-protected incubation, the cells were washed, centrifuged, and subjected to flow-cytometric analysis using a FACS Canto-II (BD, Heidelberg, Germany). Cell populations were defined after exclusion of doublets by antibody positivity in the following manner: T cells as CD3+CD56- (subsequent analysis for CD4 and CD8), natural killer cells as CD3-CD19-CD56+ (subsequent analysis of CD16 and CD56), and B cells as CD19+. A total of 1200 CD3+ cells were recorded per tube, and FCM data were analyzed using the Beckman application software CyExpert for DxFLEX.16,18,19,23,30

The samples from patients with COVID-19 were collected at different time points, including the admission day, the third day, the first week, the second week, the third week and the fourth week after admission or the day of onset, the third day, the first week, the second week, the third week, the fourth week and the fifth week after onset.

Data Collection

All data of 773 cases, including clinical and demographic information, lymphocytes and subsets at all follow-up time points, were collected to establish databases. Researchers strictly controlled the accuracy, completeness and authenticity of all data.

Statistical Analysis

SPSS 26.0 (SPSS, Chicago, IL, USA) and GraphPad Prism 8 (GraphPad, CA, USA) were used for statistical analyses. Measurement data with normal distributions are presented as the mean and standard deviation (SD), while those with nonnormal distributions are presented as the median and standard error. Categorical data are expressed as percentages or proportions.

Data with normal distributions and homogeneity of variance between multiple groups were compared using one-way or two-way ANOVA, and further comparison between two groups was performed by using the least significant difference (LSD) t-test. Data with normal distributions and homogeneity of variance between two groups were compared using the independent samples t-test. Enumeration data are presented as percentages or proportions, and data between two or multiple groups were compared using a chi-square test. Two-factor correlation analysis was performed with Spearman correlation analysis. Receiver operating characteristic (ROC) analysis was used to assess the capacity of different lymphocytes and their subsets to distinguish cases of the nonsevere group from the severe group, as well as those who survived from dead COVID-19 patients. A P<0.05 was considered statistically significant.

Patient and Public Involvement

Patients and the public were not involved in the design or conduct of this study. All patients recruited in this study received verbal and written information about this research. The burden of the intervention was carefully assessed by the investigators. Participants were assessed for eligibility, and data collection was performed. Dissemination of the general results (without personally identifying data) will occur on demand.

Our study was approved by the Ethics Committee of the Public Health Clinical Centre of Chengdu (ethics approval number: PJ-K2020-26-01). Written informed consent was waived by the Ethics Commission of the designated hospital because this study is related to emerging infectious diseases.

ResultsBaseline Information

A total of 718 COVID-19 patients (medium age, 38.48 years; 73.68% male) and 55 control patients (medium age, 55.54 years; 45.45% male) were included. Other baseline information, including duration, lymphocyte subsets at admission, virus-negative conversion time, in-hospital time, proportion of severe cases and proportion of deaths, is shown in Table 1 and Figure 1. Among 718 COVID-19 patients, the severity rate was 5.15% (37/718), and the overall mortality rate was 1.11% (8/718) (Table 1).

The time of onset and admission in the asymptomatic infection group was the same day. The average time from onset to admission in the mild illness group and the common illness group was approximately 3–4 days, and the average time from onset to admission in the severe illness group and the critical illness group was approximately 6–7 days.

COVID-19 patients in the nonsevere subgroup and in the survival group were significantly younger than healthy controls in the control group (Table 2). While COVID-19 patients in the severe subgroup were similar old as healthy controls in the control group, but those in the nonsurvival group were older than healthy controls in the control group (Table 2).

Table 2 Comparison of Baseline Conditions Between the Three Groups (n=773)15

For COVID-19 patients, those in the survival group were significantly younger than those in the death group (P<0.0001) (Table 2). Meanwhile, those in the severe subgroup were also older than those in the nonsevere subgroup (Table 2). There was also a difference in sex between the severe COVID-19 group, the survival COVID-19 group, the death COVID-19 group and the control group (all P<0.0001) (Table 2).

The Characteristics of Lymphocyte Subsets at Admission in Patients with COVID-19

At admission, different groups demonstrated different lymphocyte subsets. In terms of disease severity, compared with patients in the control group, the nonsevere group had higher lymphocyte counts and percentages (Figure 2A and B), CD3+ T cell counts (Figure 2C) (all P<0.05), similar CD3+ T cell percentages and CD19+ B cell counts (Figure 2D and E) (all P>0.05), while it showed lower CD19+ B cell percentages (Figure 2F), CD56+ NK cell counts and percentages (Figure 2G and H), also had higher CD3+CD4+ T cell counts (Figure 3A), CD3+CD8+ T cell counts and percentages (Figure 3C and D) (all P<0.05), and lower CD3+CD4+ T cell percentages (Figure 3B) and the ratio of CD4+ T cells to CD8+ T cells (Figure 3E) (all P<0.05). For the severe group, despite the age similarity, the lymphocyte counts and percentages, CD3+ T cell counts and percentages, CD19+ B cell counts and percentages, CD56+ NK cell counts, CD3+CD4+ T cell counts and percentages, CD3+CD8+ T cell counts and the ratio of CD3+CD4+ T cells to CD3+CD8+ T cells were lower than those in the control group (Figure 2A–G and Figure 3A–C, E) (all P<0.05), while the CD3+CD8+ T cell percentages (Figure 3D) was higher than that in the control subgroup (P<0.05). Moreover, compared with the nonsevere group, lymphocyte counts and percentages, CD3+ T cell counts and percentages, CD19+ B cell counts, CD56+ NK cell counts, CD3+CD4+ T cell counts and percentages, and CD3+CD8+ T cell counts (Figure 2A–E, H and Figure 3A–C) in the severe group were significantly lower (all P<0.05), while the CD56+ NK cell percentages (Figure 2G) was higher than that in the nonsevere subgroup (P<0.05).

Figure 2 Comparison of baseline levels of lymphocyte and T, B and NK cell counts and percentages among the control group, the nonsevere COVID-19 group and the severe COVID-19 group (n=773; control, severe and nonsevere groups, n=55, 37 and 681, respectively). Abbreviations: COVID-19, coronavirus disease 2019. (A) Lymphocyte count. (B) Lymphocyte percentage. (C) T lymphocyte count. (D) T lymphocyte percentage. (E) B lymphocyte count. (F) B lymphocyte percentage. (G) NK lymphocyte count. (H) NK lymphocyte percentage. The measurement data are expressed as χ±SD. Unpaired one-way ANOVA was used for intergroup comparisons (A–C, G and H), P all<0.0001; (E), P<0.001; (D and F), P all<0.01). Unpaired t-tests were used for comparisons with the control group or with the nonsevere COVID-19 group, NSP>0.05, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

Figure 3 Comparison of baseline T lymphocyte subset counts and percentages among the control group, the severe COVID-19 group and the nonsevere COVID-19 group (n=773; control, severe and nonsevere groups, n=55, 37 and 681, respectively). (A) CD3+CD4+ count. (B) CD3+CD4+ percentage. (C) CD3+CD8+ count. (D) CD3+CD8+ percentage. (E) Ratio of CD4+/CD8+ cells. The measurement data are expressed as χ±SD. Unpaired one-way ANOVA was used for intergroup comparisons (A–C), P all<0.0001; (D), P<0.01; (E), P<0.05). Comparisons with the control group or with the nonsevere COVID-19 group using unpaired t-tests, NSP>0.05, **P<0.01,***P<0.001, ****P<0.0001.

Regarding disease prognosis, compared with the control group, the survival group also demonstrated higher lymphocyte counts and percentages (Figure 4A and B), CD3+ T cell counts (Figure 4C) (all P<0.05), similar CD3+ T cell percentages and CD19+ B cell counts (Figure 4D and E), while it harbored lower CD19+ B cell percentages (Figure 4F), CD56+ NK cell counts and percentages (Figure 4G and H), and also had higher CD3+CD4+ T cell counts (Figure 5A), and CD3+CD8+ T cell counts and percentages (Figure 5C and D) (all P<0.05), lower CD3+CD4+ T cell percentages (Figure 5B) and the ratio of CD4+ T cells to CD8+ T cells (Figure 5E) (all P<0.05). For the death group, the lymphocyte counts and percentages (Figure 4A and B), CD3+ T cell counts (Figure 4C), CD19+ B cell counts and percentages (Figure 4E and F), CD56+ NK cell counts (Figure 4G), CD3+CD4+ T cell counts (Figure 5A) and CD3+CD8+ T cell counts (Figure 5C) were lower than those in the control group (all P<0.05), while the CD3+CD8+ T cell percentages (Figure 5D) was higher than that in the control subgroup (P<0.05). Furthermore, in the death group, lymphocyte counts and percentages, CD3+ T cell counts, CD3+CD4+ T cell counts and CD3+CD8+ T cell counts (Figures 4A–C and 5A, C) were significantly lower than the corresponding values in the survival group (all P<0.05).

Figure 4 Comparison of baseline levels of lymphocyte and T, B and NK cell counts and percentages among the control group, the death COVID-19 group and the survival COVID-19 group (n=773; control, death and survival groups, n=55, 8 and 710, respectively). (A) Lymphocyte count. (B) Lymphocyte percentage. (C) T lymphocyte count. (D) T lymphocyte percentage. (E) B lymphocyte count. (F) B lymphocyte percentage. (G) NK lymphocyte count. (H) NK lymphocyte percentage. The measurement data are expressed as χ±SD. Intergroup comparisons were performed using unpaired one-way ANOVA (A–C, G and H), all P all<0.0001; (D), P>0.05, (E and F), P<0.001). Comparisons with the control group or with the surviving COVID-19 group were performed using unpaired t-tests, NSP>0.05, ***P<0.001, ****P<0.0001.

Figure 5 Comparison of baseline levels of T lymphocyte subset counts and percentages among the control group, the survival COVID-19 group and the death COVID-19 group (n=773; control, survival and death groups, n=55, 710 and 8, respectively). (A) CD3+CD4+ count. (B) CD3+CD4+ percentage. (C) CD3+CD8+ count. (D) CD3+CD8+ percentage. (E) Ratio of CD4+/CD8+ cells. The measurement data are expressed as χ±SD. Intergroup comparisons were performed using unpaired one-way ANOVA (A and C), P all<0.0001; (B) P<0.001; (D) P<0.01; (E), P<0.05). Comparisons with the control group or with the surviving COVID-19 group using unpaired t-tests, NSP>0.05, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

Dynamic Characteristics of Lymphocyte Subsets Within Four Weeks After Admission in Patients with COVID-19

From admission to the fourth week after admission, significant differences between the severe group and the nonsevere group were noticed in lymphocyte counts, lymphocyte percentages, CD19+ B cell counts, CD56+ NK cell percentages, CD3+ T cell counts, CD3+CD4+ T cell counts and CD3+CD8+ T cell counts. In the nonsevere group, lymphocyte counts (Figure 6A) and lymphocyte percentages (Figure 6B) decreased gradually from admission to the third week and rose slightly in the fourth week. In the severe COVID-19 group, lymphocyte counts (Figure 6A) and lymphocyte percentages (Figure 6B) were always lower than those in the nonsevere COVID-19 group but showed a small increase from admission to the fourth week. CD19+ B cell counts (Figure 6C) remained stable at a low level in the severe COVID-19 group but showed some fluctuations around a higher level in the nonsevere COVID-19 group. In the nonsevere group, CD19+ B cell percentages (Figure 6D) and CD56+ NK cell counts (Figure 6E) remained stable from admission to the fourth week, but in the severe COVID-19 group CD19+ B cell percentages (Figure 6D) remained stable only from admission to the first week, then decreased gradually from the first week to the fourth week, and CD56+ NK cell counts (Figure 6E) increased gradually from admission to the fourth week. In contrast, CD56+ NK cell percentages (Figure 6F) stayed at a stable low level in the nonsevere COVID-19 group, while they were significantly higher in the severe COVID-19 group. In the nonsevere group, CD3+ T cell counts and percentages (Figure 7A and B), CD3+CD4+ T cell counts (Figure 7C) and CD3+CD8+ T cell counts (Figure 7E) decreased gradually from admission to the third week and rose slightly in the fourth week In the severe COVID-19 group, lymphocyte counts (Figure 6A), lymphocyte percentages (Figure 6B), CD3+ T cell counts (Figure 7A), CD3+CD4+ T cell counts (Figure 7C) and CD3+CD8+ T cell counts (Figure 7E) were always lower than those in the nonsevere COVID group but showed a small increase from admission to the fourth week. CD19+ B cell counts (Figure 6C) remained stable at a low level in the severe COVID-19 group but showed some fluctuations around a higher level in the nonsevere COVID-19 group. In contrast, CD56+ NK cell percentages (Figure 6F) stayed at a stable low level in the nonsevere COVID-19 group, while they were significantly higher in the severe COVID-19 group. Whether in the nonsevere COVID-19 group or in the severe COVID-19 group, CD3+CD4+ T cell percentages (Figure 7D) and CD3+CD8+ T cell percentages (Figure 7F) remained stable from admission to the fourth week, and the ratio of CD3+CD4+ T cells to CD3+CD8+ T cells (Figure 7G) also remained stable from admission to the third week then increased in the nonsevere COVID-19 group but decreased in the severe COVID-19 group in the fourth week.

Figure 6 Comparison of the dynamic characteristics of lymphocyte and B and NK cell counts and percentages between the nonsevere COVID-19 group and the severe COVID-19 group (n=718; nonsevere and severe groups, n=681 and 37, respectively). (A) Lymphocyte count. (B) Lymphocyte percentage. (C) CD19+ count. (D) CD19+ percentage. (E) CD56+ count. (F) CD56+ percentage. The measurement data are expressed as χ±SD. Two-way ANOVA was used for comparisons between two subgroups from admission to the fourth week after admission (interaction, (A and B), all P<0.0001, (D) P<0.05, (E) P<0.001, (C and F) all P>0.05; row factor, (A and F) all P>0.05, (B and C) all P<0.05, (D) P<0.01, (E) P<0.001; column factor, (A–F) all P<0.0001). Unpaired t-tests were used for comparisons between the nonsevere group and the severe group at the same time point, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

Figure 7 Comparison of dynamic characteristics of T lymphocyte subset counts and percentages among the severe COVID-19 group and the nonsevere COVID-19 group (n=718; severe and nonsevere groups, n=37 and 681, respectively). (A) CD3+ count. (B) CD3+ percentage. (C) CD3+CD4+ count. (D) CD3+CD4+ percentage. E. CD3+CD8+ count. (F) CD3+CD8+ percentage. (G) Ratio of CD4+/CD8+ cells. The measurement data are expressed as χ±SD. Two-way ANOVA was used for comparisons between two subgroups from admission to the fourth week after admission (interaction, (A–C, E and G), all P<0.0001, (D and F), all P>0.05; row factor, (B) P<0.0001, (G) P<0.01, (A) and (C–F), all P>0.05; column factor, (A–E and G), all P<0.0001, (F), P>0.05). Unpaired t-tests were used for comparisons between the nonsevere group and the severe group at the same time point, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

For the death group and survival group, significant differences were shown in lymphocyte counts, lymphocyte percentages, CD3+ T cell counts, CD3+CD4+ T cell counts and CD3+CD8+ T cell counts. In the survival group, the lymphocyte counts (Figure 8A) and lymphocyte percentages (Figure 8B) declined in the first three weeks after admission and increased slightly at the fourth week, CD19+ B cell counts and percentages (Figure 8C and D), CD56+ NK cell counts and percentages (Figure 8E and F), CD3+CD4+ T cell percentages (Figure 9D) and CD3+CD8+ T cell percentages (Figure 9F) and the ratio of CD3+CD4+ T cells to CD3+CD8+ T cells (Figure 9G) remained stable from admission to the fourth week, CD3+ T cell counts (Figure 9A), CD3+CD4+ T cell counts (Figure 9C) and CD3+CD8+ T cell counts (Figure 9E) declined in the first three weeks after admission and increased slightly at the fourth week, and CD3+ T cell percentages (Figure 9B) remained stable from admission to the first week then increased from the first week to the fourth week. For the death group, those parameters remained similar trends and were always remarkably lower than those in the survival group (Figure 8A–F and Figure 9A–G).

Figure 8 Comparison of dynamic characteristics of lymphocytes and B and NK cells and percentages among the surviving COVID-19 group and the nonsurviving COVID-19 group (n=718; the surviving COVID-19 group and the nonsurviving COVID-19 group, n=710 and 8, respectively). (A). Lymphocyte count. (B). Lymphocyte percentage. (C) CD19+ count. (D) CD19+ percentage. (E). CD56+ count. (F) CD56+ percentage. The measurement data are expressed as χ±SD. Two-way ANOVA was used for comparisons between two subgroups from admission to the fourth week after admission (interaction, (A–C and E), all P>0.05, (D) P<0.01, (F) P<0.05; row factor, (A–C, E and F), all P>0.05, (D) P<0.05; column factor, (A–C and E), all P<0.0001, (D), P<0.05, (F), P<0.001). Unpaired t-tests were used for comparisons between the surviving COVID-19 group and the nonsurviving COVID-19 group at the same time point, *P<0.05, **P<0.01, ***P<0.001.

Figure 9 Comparison of dynamic characteristics of T lymphocyte levels and percentages among the nonsurviving COVID-19 group and the surviving COVID-19 group (n=718; death and survival group, n=8 and 710, respectively). (A). CD3+ count. (B). CD3+ percentage. C. CD3+CD4+ count. (D). CD3+CD4+ percentage. (E) CD3+CD8+ count. F. CD3+CD8+ percentage. (G). Ratio of CD4+/CD8+ cells. The measurement data are expressed as χ±SD. Two-way ANOVA was used for comparisons between two subgroups from admission to the fourth week after admission (interaction, (A–E and G, all P>0.05, F, P<0.01; row factor, B, P<0.0001, G, P<0.05, A–F, all P>0.05; column factor, A–C, all P<0.0001, D–G, all P>0.05). Unpaired t-tests were used for comparisons between the surviving COVID-19 group and the nonsurviving COVID-19 group at the same time point, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

Dynamic Characteristics of Lymphocyte Subsets within Five Weeks After Onset in COVID-19 Patients with Different Disease Severity

The lymphocyte subsets among different groups with different disease severity also showed similar dynamic changes from onset to five weeks after onset. Lymphocyte counts (Figure 10A), lymphocyte percentages (Figure 10B), CD3+ T cell counts (Figure 10C) and CD19+ B cell counts (Figure 10E) were all significantly higher in the asymptomatic group than those in the critically ill group and severe group, while those in the common group and mild group stayed in the middle. In addition, all parameters above showed a slight decrease in the asymptomatic group but demonstrated a small increase in the critically ill group and severe group. In all groups CD3+ T cell percentages (Figure 10D), CD19+ B cell percentages (Figure 10F), CD56+ NK cell counts and percentages (Figure 10G and H) remained relatively stable or small fluctuations. In contrast, CD56+ NK cell percentages (Figure 10H) remained stable at a low level in the asymptomatic group but showed a higher level in the critically ill group and severe group. Moreover, CD3+CD4+ T cell counts (Figure 11A) and CD3+CD8+ T cell counts (Figure 11C) were all significantly higher in the asymptomatic group than those in the critically ill group and severe group, while those in the common group and mild group stayed in the middle. In all groups CD3+CD4+ T cell percentages (Figure 11B), CD3+CD8+ T cell percentages (Figure 11D) and the ratio of CD3+CD4+ T cell to CD3+CD8+ T cell (Figure 11E) remained relatively stable or small fluctuations. In addition, all parameters above showed a slight decrease in the asymptomatic group but demonstrated a small increase in the critically ill group and severe group.

Figure 10 Comparison of the dynamic characteristics of lymphocytes and T, B and NK cell counts and percentages among the asymptomatic infection, light, common, severe, and critical illness COVID-19 groups (n=718; the asymptomatic infection, light, common, severe, and critical illness groups, n=237, 73, 371, 18 and 19, respectively). (A). Lymphocyte count. (B). Lymphocyte percentage. (C) T lymphocyte count. (D) T lymphocyte percentage. (E) B lymphocyte count. (F) B lymphocyte percentage. (G) NK lymphocyte count. (H) NK lymphocyte percentage. The measurement data are expressed as χ±SD. Two-way ANOVA was used for comparisons among subgroups from onset to the fifth week after onset (interaction, (A, C–E and G), P all>0.05, (B and F), all P<0.01, (H), (G), P all>0.05, (B and F), P all<0.01, (D and H), P all <0.05; column factor, (A–F and H), P all<0.0001, (G), P<0.05). One-way ANOVA was used for comparisons among groups at the same time point, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

Figure 11 Comparison of dynamic characteristics of T lymphocyte subset counts and percentages among the asymptomatic infection, light, common, severe, and critical illness COVID-19 groups (n=718; the asymptomatic infection, light, common, severe, and critical illness groups, n=237, 73, 371, 18 and 19, respectively). (A). CD3+CD4+ count. (B). CD3+CD4+ percentage. (C) CD3+CD8+ count. (D). CD3+CD8+ percentage. (E). Ratio of CD4+/CD8+ cells. The measurement data are expressed as χ±SD. Two-way ANOVA was used for comparisons among groups from onset to the fifth week after onset (interaction, (B), P<0.05, (A and C–E), P>0.05; row factor, (B) P<0.05, (A and C–E), P>0.05; column factor, (A–C), P all<0.0001, (D and E), P all>0.05). One-way ANOVA was used for comparisons among subgroups at the same time point, *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

The Relationship of Disease Progression and Prognosis with Lymphocyte Subsets in Patients with COVID-19

Spearman correlation analysis (Table 3) reported that when considering patients’ lymphocyte subsets at admission, disease severity was negatively correlated with lymphocyte counts (r=−0.166, P<0.0001), lymphocyte percentages (r=−0.179, P<0.0001), CD3+ T cell counts (r=−0.173, P<0.0001), CD3+CD4+ T cell counts (r=−0.186, P<0.0001), CD3+CD8+ T cell counts (r=−0.127, P=0.004), CD3+ T cell percentages (r=−0.117, P=0.008), and CD3+CD4+ T cell percentages (r=−0.149, P=0.001). Meanwhile, when considering patients’ lymphocyte subsets at admission, the prognosis of death was negatively correlated with lymphocyte counts (r=−0.125, P=0.005), lymphocyte percentages (r=−0.121, P=0.006), CD3+ T cell counts (r=−0.123, P=0.005), CD3+CD4+ T cell counts (r=−0.123, P=0.005), and CD3+CD8+ T cell counts (r=−0.091, P=0.04). In contrast, the prognosis of death was positively related to CD56+ NK cell counts (r=0.204, P=0.033).

Table 3 Spearman Correlation Analysis of Disease Severity, Virus Negative Conversion Time, Prognosis and Lymphocytes and Subsets (n=718)

Receiver operating characteristic curve (ROC) analysis also found that lymphocyte subsets at onset had a good ability to distinguish severe cases and nonsevere cases, as well as surviving cases and dead cases. The cutoff points of different lymphocyte subsets, including lymphocyte counts, lymphocyte percentages, CD3+ T cell counts and percentages, CD3+CD4+ T cell counts, CD3+CD8+ T cell counts and percentages, CD19+ B cell counts and percentages and CD56+ NK cell counts, are shown in Tables 4 and 5.

Table 4 The Performance of Various Parameters at COVID-19 Onset for Distinguishing Between Severe Cases and Nonsevere Cases (n =718)

Table 5 The Performance of Various Parameters at COVID-19 Onset for Distinguishing Between Surviving Cases and Dead Cases (n =718)

Discussion

COVID-19, caused by a virus called SARS-CoV-2, can cause a systemic inflammatory response and significant tissue damage, leading to rapid disease progression and even death. In this condition, white blood cell populations, including monocytes, lymphocytes and neutrophils, play a critical role.

Our study found that nonsevere and severe COVID-19 patients, survivors and nonsurvivors of COVID-19 infection, demonstrated different lymphocyte subsets, whether at admission, four weeks after admission, or five weeks after disease onset. At admission, compared with nonsevere COVID-19 patients, severe COVID-19 patients demonstrated lower lymphocyte counts and percentages, CD3+ T cell counts and percentages, CD19+ B cell counts and percentages, CD56+ NK cell counts, CD3+CD4+ T cell counts and percentages and CD3+CD8+ T cell counts. Meanwhile, at admission, nonsurvivors also had lower lymphocyte counts and percentages, CD3+ T cell counts, CD3+CD4+ T cell counts, and CD3+CD8+ T cell counts. For the follow-up period after admission or onset, the nonsevere group showed higher lymphocyte counts and percentages, CD19+ B cell counts, CD56+ NK cell percentages, CD3+ T cell counts, CD3+CD4+ T cell counts and CD3+CD8+ T cell counts than the severe group. These phenomena were also found between survivors and nonsurvivors. Therefore, it seemed that the lymphocyte subsets mentioned above were correlated with disease severity and prognosis. Our Spearman correlation analysis proved that when considering patients’ lymphocyte subsets at admission, disease severity was negatively correlated with lymphocyte counts and percentages, CD3+ T cell counts and percentages, CD3+CD4+ T cell counts and percentages, and CD3+CD8+ T cell percentages. Meanwhile, when considering patients’ lymphocyte subsets at admission, the prognosis of death was negatively correlated with lymphocyte counts and percentages, CD3+ T cell counts, CD3+CD4+ T cell counts, and CD3+CD8+ T cell counts. Furthermore, our ROC analysis also found that lymphocyte subsets at COVID-19 onset had a good capacity to distinguish severe and nonsevere cases, as well as survivors and nonsurvivors. These findings showed that in primary infection and unvaccinated COVID-19 patients total lymphocytes and T cell, B cell and NK cell subsets at COVID-19 onset play valuable roles in predicting disease severity and prognosis.

Our results were consistent with previous studies. Several observational studies have suggested the relationship between immune parameters and clinical severity and disease prognosis. Lymphopenia seemed to be the leading characteristic of COVID-19-infected patients, as a great number of clinical studies have reported decreased lymphocyte numbers in a large proportion of COVID-19 patients.9,31–42 In addition, some studies further suggested that absolute numbers of lymphocytes were correlated with disease severity: significantly lower levels of lymphocyte numbers were noticed in severe and critically ill COVID-19 cases than in mild cases.31,43,44 Increases in CD4+ and particularly CD8+ T lymphocyte counts were also noticed in severe COVID-19 patients after treatment.45 Some studies have also suggested the role of a higher neutrophil-to-lymphocyte ratio (NLR) in predicting the severity of COVID-19.46,47

For subpopulations of lymphocytes, remarkable differences in CD4+ and CD8+ T cells were noticed between mild and severe/critical COVID-19 patients.42 Furthermore, some studies also suggested cut-off points for different lymphocyte subsets to predict further disease deterioration and in-hospital death. Diao et al48 reported that for nonintensive care unit patients, those with total T cell counts lower than 800/μL, CD4+ T cell counts lower than 400/μL, or CD8+ T cell counts lower than 300/μL were prone to further deterioration. Moreover, Xu et al49 also found that lower levels of T lymphocyte counts and their subsets (total CD3+ T cells<200/μL, CD4+ T cells<100/μL, and CD8+ T cells<100/μL) were remarkably associated with an increased risk of in-hospital death owing to COVID-19. Our ROC analysis also demonstrated that lymphocyte counts and their subsets at disease onset had a good capacity to predict future disease severity and prognosis. The cutoff points were as follows: lymphocyte counts 448/μL, lymphocyte percentages 5.575%, CD3+ T cell counts 340/μL, CD3+ T cell percentages 65.865%, CD3+CD4+ T cell counts 161/μL, CD3+CD8+ T cell counts 131/μL, CD3+CD8+ T cell percentages 65.865%, CD19+ B cell counts 30/μL, CD19+ B cell percentages 7.98%, and CD56+ NK cell counts 62/μL.

In contrast, with decreased levels of most lymphocyte subsets, including CD3+ T cells, CD4+ T cells, CD8+ T cells and CD19+ B cells, CD56+ NK cells were the only subset that had an elevated level in more severe COVID-19 cases. However, our findings were different from previous results that reported a decreased level of NK cell counts in patients with more severe disease.50–52 This inconsistency might be attributed to our small sample size of severe cases. Kram et al52 also found that NK cells were functionally impaired and expressed higher levels of the apoptosis markers active caspase-3 and CD95.

Those above findings supported lymphocytes and subsets as predictive biomarkers and potential therapeutic targets in primary infection and unvaccinated COVID-19 patients. The mechanism behind lower lymphocyte counts and subsets in more severe COVID-19 cases might be the result of inflammatory responses and cytokine storms induced by virus infection. Ni et al45 found that the levels of proinflammatory factors increased significantly in severe COVID-19 cases, and most of these factors, including IL-2R, IL-6, TNF-α, and CRP, decreased significantly after treatment. These proinflammatory factors can lead to lymphocyte activation initially; however, continuous active responses can subsequently contribute to lymphocyte exhaustion. A previous study also reported increased numbers of activated CD8+ T cells and NK cell populations in COVID-19 patients, and with increasing severity, failure of clonal expansion in CD8+ T effector and central memory cells and depletion of COVID-19 clonotypes were observed.53 Furthermore, associations with severity for exhaustion markers and specific activated NK and CD69+ MAIT cell populations were also noticed.53 Meanwhile, CD4+ and CD8+ cell exhaustion, with increased concentration of proinflammatory cytokines and chemokines and decreased T regulatory cells, could lead to an excessive inflammatory response that is out of control, which causes a vicious circle.54

It has been widely recognized that immune function declines with age.55,56 Our study found that healthy controls and COVID-19 patients, whether nonsevere COVID-19 patients or sever COVID-19 patients, whether survivors of COVID-19 infection or nonsurvivors of COVID-19 infection, all demonstrated different lymphocyte subsets at admission. Compared with healthy controls, nonsevere COVID-19 patients and survivors of COVID-19 infection all demonstrated higher lymphocyte counts and percentages, CD3+ T cell counts, CD3+CD4+ T cell counts, CD3+CD8+ T cell counts and percentages, lower CD19+ B cell percentages, CD56+ NK cell counts and percentages, CD3+CD4+ T cell percentages and the ratio of CD3+CD4+ T cells to CD3+CD8+ T cells, and nonsevere COVID-19 patients also demonstrated lower CD19+ B cell counts. Compared with healthy controls, severe COVID-19 patients and nonsurvivors of COVID-19 infection all demonstrated lower lymphocyte counts and percentages, CD3+ T cell counts, CD19+ B cell counts and percentages, CD56+ NK cell counts, CD3+CD4+ T cell counts and CD3+CD8+ T cell counts, higher CD3+CD8+ T cell percentages. Severe COVID-19 patients also demonstrated lower CD3+ T cell percentages, CD3+CD4+ T cell percentages and the ratio of CD3+CD4+ T cells to CD3+CD8+ T cells. These changes may be age-related, but not exclusively age-related, as nonsevere COVID-19 patients and survivors of COVID-19 infection were significantly younger than healthy controls. These changes showed that these nonsevere COVID-19 patients and survivors of COVID-19 infection did not respond to the infection with a robust immune response. Because of these nonsevere COVID-19 patients and survivors of COVID-19 infection were primary infection and unvaccinated COVID-19 patients, so these changes was not due to already pre-infection. Although these nonsevere COVID-19 patients and survivors of COVID-19 infection had higher lymphocyte and T cell subset, but B cell and NK cell subset, CD4+ T cell percentages and the ratio of CD4+ T cells to CD8+ T cells were decreased. Moreover, although severe COVID-19 patients were similar old as healthy controls, and nonsurvivors of COVID-19 infection were significantly older than healthy controls, severe COVID-19 patients and nonsurvivors of COVID-19 infection all had overall decreased lymphocyte and T cell, B cell and NK cell subsets, these changes showed that severe COVID-19 patients and nonsurvivors of COVID-19 infection responded to the infection with a robust immune response. Previous studies showed that in old age the innate immune system is characterized by chronic low-grade inflammation and a simultaneous reduction of an adequate response to pathogens, the adaptive immune system is characterized by a reduced hematopoiesis of new naïve lymphocytes and a reduction of diversity describe.56 Immunosenescence, as age-related impairment of immune function, play an important role in age-related morbidity and mortality, and infectious disease.One aspect of immunosenescence is age-related natural killer (NK) cell dysfunction, demonstrated decreased target cell cytotoxicity and reduced cytokine secretion, increased NK cell numbers, which play central actors in the immunosurveillance of senescent cells. NK cell dysfunction is implicated in the increasing burden of infection, malignancy, inflammatory disorders, and senescent cells with age.55

This study findings showed that in primary infection and unvaccinated COVID-19 patients total lymphocytes and T cell, B cell and NK cell subsets at COVID-19 onset play valuable roles in predicting disease severity and prognosis. But this is not COVID-19 specific, not even infection specific, and a general and well-known observation for patients ending up in ICU. A study found that Mpox patients showed as elevated CD3+CD8+T counts and inverted ratio of CD3+CD4+T cell to CD3+CD8+T cell, significantly decreased CD3+CD4+T counts in mpox patients co-infected with HIV compared to the pre-infection level.57 Other studies found that ICU patients with acquired infection had significantly lower lymphocyte subpopulation counts (CD3+ T cells, CD3+CD4+ T cells, CD3+CD8+T cells, CD16+CD56+ natural killer (NK) cells and CD19+ B cells) compared those without acquired infection, CD3+T cells and CD3+CD4+ T cells were independent significant risk factors for ICU-acquired infections,21 and lymphopenia and T cell depletion at ICU admission were associated with increased mortality of ICU patients.22–24 Those above findings supported lymphocytes and subsets as predictive biomarkers and potential therapeutic targets in intensive care medicine, yet the underlying mechanisms remain unknown.

It is clear that SARS-CoV-2 infection results in a good and robust immune response. However, not in all patients. The reason for this unknown, and likely not