Demographic and clinical characteristics of COVID–19 patients
A total of 154 patients were enrolled in this study, and were divided into two groups according to the published guidelines, including 122 mild/moderate cases (79.2 %, referred to as mild cases thereafter) and 32 severe/critical cases (20.7 %, referred to as severe cases). Among 122 patients with mild symptoms, 8 patients had no chest CT imaging abnormality. Among the 32 patients with severe symptoms, 4 patients required ICU care and ventilator-aided breathing.
The enrolled COVID–19 patients consisted of 86 males (55.8%) and 68 females (44.1%) (Table 1). The mean age of the patients was 42.4 years. The medium age of the severe symptom group (52.2 ± 14.0 years) was older than that of the mild group (39.9 ± 15.0 years). A total of 52 (33.8%) patients in both groups had basline diseases, including hypertension (23 [14.9%]), diabetes (11 [7.1%]), chronic liver diseases (12 [7.8%]), coronary heart diseases (3 [1.9%]), malignant tumors (3 [1.3%]), cerebrovascular diseases (2 [1.3%], COPD (2 [1.3%]) and AIDS (1 [0.65%] (Table 1). Two severe group patients had mixed fungal and 6 bacterial infections, respectively. At admission, 31 (96.9%) of severe group patients and 86.4% of mild group patients had fever, with a significant difference in body temperature (Table 1). The levels of the inflammatory markers ESR (P = 0.008), C-reactive protein (CRP, P<0.001), serum amyloid A protein (SAA, P = 0.039), interleukin–6 (IL–6, P<0.001), and other markers including albumin (P<0.001), aspartate transaminase (p = 0.003), alanine transaminase (P = 0.043), lactate dehydrogenase (p<0.001), fibrinogen (p = 0.004) and D-dimer (P<0.001) in the peripheral blood of the severe group patients were significantly higher at admission compared to the mild group patients (Table 2). No significant differences in the numbers of red blood cells and platelets, nor in the serum levels of procalcitonin, total bilirubin, creatinine, creatine kinase and PT were observed between the two groups (Table 2).
The association of the severity of COVID–19 with comorbidities in our patient cohort was also analyzed (Figure 1). Overall, patients with underlying diseases were more prone to developing severe COVID–19. However, patients with comorbidities such as chronic hepatitis B virus infection, liver cirrhosis, chronic respiratory diseases, HIV infection, and cancers (comorbidity group 1, CG–1) which are known to impair immunological functions were significantly more frequently affected compared to those with comorbidities unrelated to immune functions (comorbidity group 2, CG–2) (Figure 1A), in the age-matched groups (Figure 1B).
Viral levels and the correlation with disease severities
The Ct values of RT-PCR reactions performed were used to assess viral RNA levels in patient samples. An absolute quantification of viral RNA loads could not be obtained due to a lack of proved standard material. The amplification reaction for f the N region regularly resulted in lower Ct values compared with that for the ORF1ab region, indicating a better sensitivity with N region amplification. Therefore, the Ct values of N-specific RT-PCR tests were analyzed to assess the dynamic changes in viral RNA levels (Figure 2). For the detection of SARS-CoV–2 RNA, sputum or throat-swab specimens were used. Overall, sputum specimens gave higher viral levels indicated by lower Ct values (Ct values up to 17) than throat-swab specimens (Ct value up to 21), which was consistent with an early report [12]. Thus, the viral load data generated by using sputum or throat-swab specimens were analyzed separately. We compared the measured SARS-CoV–2 RNA levels in sputum specimens from COVID–19 patients at admission among groups divided according to age, sex, underlying diseases and disease severity (Figure 2A). The Ct values varied from >38 up to 17 and no difference between patient groups was found. However, viral RNA levels in severe cases were found to be significantly higher in sputum specimens but not in throat-swab specimens between days 3 to 5 after disease onset compared to those in mild cases (Figure 2B). Nevertheless, a number of mild cases had relatively high viral loads with a Ct value of up to 20. The viral RNA levels in the patients decreased and became undetectable (Ct > 38) over time (Figure 2C). The mild cases became negative for viral RNAs with a median of 21 days after the disease onset, which was significantly earlier than severe cases with a mean of 35 days based on the test results using sputum specimens (Figure 2D). No significant difference was found between the mild and severe cases at 28 and 25 days after disease onset if throat-swap specimens were used (Figure 2D).
Kinetic analysis of lymphocyte subsets in the peripheral blood of COVID–19 patients
Lymphopenia was observed in 47.4% (73/154) of patients in our cohort, 40.2% (49/122) and 75% (24/32) in mild and severe cases at the onset of disease respectively (Table 2). As shown in Table 2, the absolute counts of lymphocytes in the peripheral blood of patients subsequently developing severe COVID–19 was significantly lower at hospital admission and disease onset respectively (Table 2, Figure 3). Sustained lower lymphocyte counts in severe cases were observed, compared to those in mild cases (Figure 3C). The absolute counts of total white blood cells (WBC) and neutrophils (NE) were initially comparable in both patient groups but significantly increased only in severe cases at day 9 up to day 17 after disease onset (Figure 3A and B). A significant difference in monocyte counts was observed between the two groups only at days 3 to 5 but not seen afterwards (Table 2, Figure 3D). The kinetic changes in different lymphocyte subsets in the peripheral blood of COVID–19 patients were monitored from disease onset up to 18 days. Flow cytometric analysis for CD3+ T cells, CD4+ and CD8+ T cell subsets, B cells and NK cells was performed (Figure 4). Firstly, we analyzed initial lymphocyte counts and CD3+, CD4+ and CD8+ T cell counts at admission in patients by age and presence of comorbidities (Figure 4A and B). Clearly, patients of older age and with comorbidities had significantly lower counts for lymphocytes and T cell subsets. Consistent with the findings about lymphocytes, sustained significantly lower CD3+, CD8+ and CD4+ T cell counts were found in patients with severe disease compared to those of the mild cases during clinical observation (Figure 4C). The lowest CD3+, CD4+ and CD8+ T cell counts in patients with severe COVID–19 were generally found 3–5 days after disease onset. However, they recovered and reached comparable levels to those in the mild cases after day 18 (Figure 4D, Figure Suppl. 1). As severe COVID–19 cases had high SARS-CoV–2 RNA levels and low T, CD4+, and CD8+ cell counts, these parameters were tested for possible correlations. However, our analyses did not show any significant correlation between SARS-CoV–2 RNA level and T, CD4+or CD8+ T cell counts (Figure Suppl 2).
No significant differences in B cell and NK cell counts were observed between the two groups during the whole disease course (Figure 4D).
Kinetic analysis of serum inflammatory markers in COVID–19 patients
A previous study demonstrated a marked elevation of inflammatory cytokine levels such as IL–6 in the sera of COVID–19 patients [2]. Therefore, we further characterized the kinetic changes of inflammatory markers, including CRP, IL–6, SAA, and PCT in the sera of the patients. The serum levels of these markers in patients with mild disease showed minor fluctuations, while significant elevations of these markers were found severe cases (Figure 5). CRP reached its peak in serum at 6 to 8 days after disease onset (Figure 5A). IL–6 levels showed sustained increases in the severe group compared to the mild group over 18 days, even when some other markers returned to normal levels (Figure 5B). The IL–6 levels in some patients with severe COVID–19 remained high as late as day 12 to 14 after disease onset, even though these patients had clinically recovered (Figure 5E).
Significant increases in serum SAA and PCT levels were observed in the severe group after disease onset but became only significantly different to those in the mild group at certain time points (Figure 5C and D). The 3 examined markers CRP, SAA and PCT reached similar levels between the severe and mild patient groups at day 18 after disease onset (Figure 5). A statistical analysis showed that SARS-CoV–2 RNA levels in patients did not correlate with the levels of serum inflammatory markers. No correlation was found between the initial levels of lymphocyte subsets in peripheral blood and serum inflammatory markers in the late period between days 12 to 14 after disease onset (not shown).
Factors associated with severe COVID–19
Next, we examined the possibility of using the above-mentioned parameters as prognostic factors for identifying severe cases in COVID–19 patients. PCA was firstly performed with R package “factoextra” to identify correlating variables for distinguishing severe patients from mild patients (Figure Suppl 3A). The lymphocyte and CD3+ T cell counts were found to be the two most important contributing variables among immunological markers. Previously, other factors such as neutrophil-to-CD8+ T cell ratio (N8R), neutrophil-to-lymphocyte ratio (NLR), neutrophil counts (NEC) and white blood cells counts (WBCC) had been proposed as potential prognostic factors for severe COVID–19 [13]. To assess the diagnostic value of these parameters, the receiver operating characteristic (ROC) curve and the area under ROC curve (AUC) were calculated by R package “pROC” (Figure Suppl 3B), and the AUC values for lymphocyte and CD3+ T cell counts were found to be 0.739 and 0.793 respectively. At the same time, the cut-off values were calculated from the ROC curves, with a value of 0.85 for lymphocyte counts (L) (Specificity: 60.6%, Sensitivity: 77.9%) and 665 for T cell counts (T) (665, 83.3%, 67.9%) (Figure Suppl 3B). Further, the ROC curves were calculated for other parameters: CD4+ T cell counts (CD4), CD8+ cell counts (CD8), and combined CD4+ and CD8+ T cell counts (CD4xCD8) with the AUC values 0.767 (Cut-off: 269, Specificity: 60.0%, Sensitivity: 87.2%), 0.772 (236, 73.3%, 73.4%), and 0.800 (115520, 90.0%, 62.4%), respectively (Figure Suppl 3B). Overall, the immune related parameters were found to have associations with the chance to develop severe COVID–19 but their diagnostic values need to be evaluated in larger cohorts.