The present study shows clinical information of 125 hospitalized COVID-19 patients. The mean age of patients was 51.93 ± 19.5 years old. The most of patients were male (57%; n = 71) and married (57%; n = 71). The mean BMI was 26.34 ± 5.4. The highest amount of co-morbidities was attributed to COPD (31%; n = 39), dyslipidemia and hypertension (27%; n = 34 for each) and diabetes (25%; n = 31). Table 1.
The mean baseline cystatin C level was 1.42 ± 0.93 mg/L, baseline creatinine was 1.38 ± 0.86 mg/L, and baseline NLR was 6.17 ± 4.50. The level of other biochemical and respiratory indicators is illustrated in Table 2.
The results of Spearman’s correlation coefficient test show that baseline cystatin C level has a direct and highly significant linear relationship with baseline creatinine level of patients (P < 0.001; r: 0.926). ). The relationship between baseline cystatin C level and baseline level of other inflammatory factors was not significant, Table 3.
Lymphopenia was reported in 41.6% (n = 52). The mean baseline cystatin C level in patients with lymphopenia was 1.20 ± 0.45 mg/L and had no significant difference from other patients (P = 0.182). The baseline level of other inflammatory factors, including creatinine, ferritin, NLR, LDH, and CRP, according to lymphopenia, is shown in Table 4. As indicated, the average level of ferritin, NLR, and CRP in patients with lymphopenia was significantly higher than in other patients.To identify possible confounding variables, the demographic and clinical characteristics of patients with and without lymphopenia are compared in Table 4. The mean age of patients with lymphopenia was 63.50 ± 16.92 years which was about 20 years higher than the average age (43.69 ± 16.94) in people who did not have lymphopenia (p < 0.001).
Moreover, patients with lymphopenia were overweight (mean of BMI = 27.95 ± 5.51 kg/m2), while other patients were almost not overweight (25.19 ± 5.11 kg/m2) (p = 0.007). Most patients in both groups were men (P = 0.574). Comparing the distribution of comorbidities in two groups, a significant difference was observed in cardiovascular disease and asthma. Several patients with cardiovascular disease were reported more in the group with lymphopenia, while this number was lower in the case of asthma. On the other hand, considering that no significant relationship was observed between the demographic and clinical factors of the patients with ferritin, NLR, and CRP, it was not necessary to repeat the analysis to control the confounding variables.
The average score of the severity of lung involvement was 31.42 ± 10.80. Also, there is a direct and highly significant linear relationship between baseline cystatin C level and lung involvement severity score (r = 0.890, P < 0.001). The correlation between the severity of lung involvement and the baseline values of other inflammatory factors, including creatinine, ferritin, NLR, LDH, and CRP, are shown in Table 5. It was significant only in the case of creatinine (r = 0.851, p < 0.001).
The simultaneous effect of baseline cystatin C and creatinine values on the severity of lung involvement were evaluated using a generalized linear model. Results showed that baseline cystatin C has a higher diagnostic power in predicting the severity of lung involvement (B = 3.88 ± 1.74, p = 0.026). So, for each one unit increase in the baseline values of cystatin C and creatinine, the average score of the severity of lung involvement will increase by 14 and 11 units, respectively (p < 0.001 for both).
Out of 125 patients, 53% (n = 66) were intubated and 47% (n = 59) were not intubated. The mean baseline cystatin C level in intubated patients was 1.48 ± 1.09 mg/L and in non-intubated patients was 1.35 ± 0.69 mg/L. And in this regard, there was no significant difference between the two groups of patients (P = 0.899).
The analysis of the basic values of other inflammatory factors, including creatinine, ferritin, NLR, LDH, and CRP, according to the intubation status of the patients, is shown in Table 6. As can be seen, the average level of ferritin, LDH, and CRP in intubated patients was significantly higher than in other patients.To identify possible confounding variables, the demographic and clinical characteristics of patients with and without intubation are compared in Table 6. The mean age of intubated patients was about 8 years, and the mean BMI was about 3 units higher than non-intubated patients (P-value equal to 0.017 and < 0.001, respectively). Comparing the comorbidities distribution in these two groups of patients was significant in terms of dyslipidemia and cardiovascular disease. So, the number of patients with dyslipidemia and cardiovascular disease in intubated patients was 38% and 33%, respectively, and in non-intubated patients, 15% and 12%, respectively (P-value in both comparisons equal to 0.005).
On the other hand, the relationship between demographic and clinical factors of patients with ferritin, NLR, and CRP was significant only for CRP and dyslipidemia (P = 0.020). So the mean CRP in patients with dyslipidemia was 102.85 ± 50.50 mg/dl; in other patients, it was 85.14 ± 59.14.
Based on this, using logistic regression, the relationship between CRP and intubation status was still significant by adjusting the patients’ dyslipidemia status as a confounding variable. By keeping other conditions constant, for each unit increase in CRP, the chance of intubation increased by 1.02 times (P < 0.001). Moreover, keeping other conditions constant, the chance of intubation in patients with dyslipidemia was significantly 2.82 times that of patients without dyslipidemia (P = 0.033).
The average ventilation time of intubated patients was 2.72 days, and the relationship between baseline cystatin C level and the average ventilation time of intubated patients was insignificant (r=-0.162, P = 0.195). However, the average ventilation time of intubated patients showed a direct and incomplete linear relationship with the baseline values of NLR and CRP (P-value equal to 0.012 and 0.024, respectively), Table 7.
The acute kidney injury (AKI) report was positive in 11% of patients (n = 14). The mean baseline cystatin C level in patients with AKI was 2.41 ± 1.43 mg/L and significantly higher than patients without AKI (P > 0.001).
Baseline values of other inflammatory factors, including creatinine, ferritin, NLR, LDH, and CRP, according to AKI status, are shown in Table 8. Among the above factors, the average baseline creatinine level in patients with AKI was significantly higher by 1.27 units than in other patients (P < 0.001).
The average hospitalization duration of the patients was 10.70 ± 5.74 days, and there was no significant relationship between the baseline cystatin C level of the patients and hospitalization duration (r=-0.049, P = 0.586). A direct and incomplete linear relationship was observed between ferritin (r = 0.219, P = 0.014) and CRP (r = 0.241, P = 0.007) with hospitalization duration.
34.4% (n = 43) of patients expired in the hospital, and the mean baseline cystatin C level of this group of patients was 1.58 ± 0.90 mg/L which was significantly higher than other patients (1.35 ± 0.94 mg/L, P = 0.002). Moreover the mean baseline values of creatinine (1.32 ± 0.89 mg/L vs. 1.49 ± 0.78 mg/L, p = 0.012), ferritin (4.98 ± 3.45 mg/L vs. 8.69 ± 4.19 mg/L, p < 0.001), LDH (5.91 ± 2.49 mg/L vs. 8.29 ± 3.73 mg/L, p < 0.001) and CRP (76.54 ± 57.18 mg/L vs. 115.53 ± 48.55 mg/L, p < 0.001) in patients who died in the hospital was significantly higher than other patients.
Backward logistic regression analysis showed that the simultaneous effect of baseline values of cystatin C, creatinine, ferritin, LDH, and CRP on the occurrence of death in the hospital only remained significant for ferritin and LDH; So that for each unit of increase in baseline values of ferritin and LDH, the chance of dying in the hospital increased one-fold (P-value < 0.001 and 0.010 respectively), Table 9.
To identify possible confounding variables, the demographic and clinical characteristics of patients who died in the hospital are compared with other patients. Considering that there was no significant difference between the two groups of patients in terms of demographic and clinical characteristics, it was not necessary to repeat the analysis in an adapted manner to check the relationship between the baseline values of cystatin C and other inflammatory factors.
Among the 82 patients discharged from the hospital, 9.75% (n = 8) expired within one month after discharge. The average baseline cystatin C level of this group of patients was 1.58 ± 1.68 mg/L, which was not significantly different from the level of patients who survived during this period (1.41 ± 0.86 mg/L, P = 0.593). Comparing the mean values of other inflammatory factors in patients who died within one month after discharge with those who survived showed that it was significant only in ferritin (P = 0.004). The mean baseline ferritin level in patients who died within one month after discharge was about 4 units higher than in patients who survived during this period (8.38 ± 2.45 mg/L vs. 4.61 ± 3.35 mg/L, respectively).
Among the 74 patients who survived one month after discharge from the hospital, 4 patients (5%) died within 3 months after discharge. The average baseline cystatin C level of this group of patients was 1.00 ± 0.021 mg/L was not significantly different from the patients who survived during this period (1.44 ± 0.94 mg/L, P = 0.510). Other inflammatory factors were not significantly different between the two groups.
In the investigation of the relationship between baseline values of cystatin C and other inflammatory factors with death within one month and 3 months after discharge, there was not enough sample in the subgroup of expired patients to compare with alive patients. Therefore the observed relationship is subject to random error.
The relationship between cystatin C level and baseline respiratory factors of patients showed that there is only a direct and incomplete linear relationship between baseline breathing rate and cystatin C level (r = 0.202, P = 0.024).
Generalized linear models (GLM) showed that for each unit increase in the baseline breathing rate, the breathing rate of the patients increased by 0.23 units in the follow-up measurement.
Results showed no correlation between baseline cystatin C level and follow-up respiratory factors, Table 10.
