The 194 patients included in this research were divided into two groups, younger age group(<60 years old) and older age group(≥60 years old). Among them, older age group (125 patients) were more than younger age group (69 patients). The M±SD of Hospitalization days, C-reactive protein, lymphocyte count were calculated. CRP and L were grouped according to the normal reference range (CRP: <3 mg/L, 3-100mg/L, >100mg/L; L: 1.1-3.2 × 109 / L,<1.1× 109 / L,>3.2× 109 / L). For quantitative data, the average of Hospitalization days in the two age groups was 20.71 days and 20.27 days, respectively. The average CRP level of older age group is higher than younger age group, with mean values of 55.06 and 39.59. The average L level of younger age group is closer to the normal reference range than older age group. In the classified data, 90.21% of patients were above the normal level of CRP and 34.54% of patients had L levels deviated from the normal range (See Table 1).
Table 1 Demographic characteristics of the patient
|
Age stratification
|
Total
(n=194)
|
<60
(n=69)
|
≥60
(n=125)
|
HOD
|
20.71±6.63
|
20.27±6.27
|
20.43±6.39
|
CRP
|
39.59±47.15
|
55.06±49.72
|
49.56±49.26
|
<3
|
14(7.22%)
|
5(2.58%)
|
19(9.79%)
|
3-100
|
46(23.71%)
|
93(47.94%)
|
139(71.65%)
|
>100
|
9(4.64%)
|
27(13.92%)
|
36(18.56%)
|
L
|
1.87±3.60
|
0.91±0.59
|
1.25±2.24
|
<1.1
|
32(16.49%)
|
30(15.46%)
|
62(31.96%)
|
1.1-3.2
|
34(17.53%)
|
93(47.94%)
|
127(65.46%)
|
>3.2
|
3(1.55%)
|
2(1.03%)
|
5(2.58%)
|
Note: HOD: Hospitalization days; Normal reference values: CRP (<3mg / L); L (1.1× 109 / L -3.2 × 109 / L)
In Figure 1, patients were divided into moderate, severe & critical groups. As can be seen from Figure 1, the mean values of L and CRP in older age group were lower than younger age group in the Moderate group. In the Severe & Critical group, the mean CRP of older age group was higher than younger age group, and the mean L of older age group was lower than younger age group. Mann-whitney test was used to analyze the differences of CRP and L levels in patients of different ages in the two groups. In Figure 1A, Patients of different ages with Moderate clinical classification showed no significant difference in CRP (p > 0.05), but patients of different ages with Severe & critical clinical classification showed significant statistical difference in CRP (p < 0.01). In Figure 1B, there were significant statistical differences in L levels between younger age group and older age group in the Moderate and Severe & Critical groups (P <0.05).
In Figure 2, the Kaplan-Meier method of survival analysis was used to describe the cumulative survival rate at the end of events during hospitalization. With the increase of observation time in hospital for all patients, the survival rate curve gradually showed a step-like decline from 1-0. During in-hospital clinical observation, older age group had earlier death events compared with younger age group. The log-rank test showed that P <0.05(P=0.013). The survival rate of patients in the two groups was significantly different and the difference was statistically significant, so the risk of death in older age group was greater than younger age group.
In Table 2, univariate logistic regression analysis showed that age was an independent influence factor of clinical classification and clinical outcome (P <0.05). younger age group had a severe clinical classification and a 0.471 times higher risk of critical type than older age group (OR=0.471,95% CI: 0.259-0.856),and the risk of death was 0.246 times that of older age group (OR=0.246,95% CI: 0.082-0.740).
Therefore, 125 samples of elderly patients (over 60 years old) were extracted in this study and the relationship between characteristics of elderly patients (demographic characteristics, clinical indicator characteristics) and clinical classification and clinical outcome was studied.
Table 2 Univariate regression analysis of independent factors influencing Clinical classification and clinical outcome
|
Univariate Logistic regression analysis of Clinical classification
|
B
|
SE
|
Wald c2
|
p
|
Exp (B)
|
95%CI
|
Age< 60
|
-0.754
|
0.305
|
6.092
|
0.014
|
0.471
|
(0.259,0.856)
|
Age≥ 60
|
|
|
|
|
1
|
|
|
Univariate Logistic regression analysis of Clinical outcome
|
B
|
SE
|
Wald c2
|
p
|
Exp (B)
|
95%CI
|
Age< 60
|
-1.402
|
0.562
|
6.231
|
0.013
|
0.246
|
(0.082,0.740)
|
Age≥ 60
|
|
|
|
|
1
|
|
|
|
|
|
|
|
|
|
|
|
|
In Table 3, 125 patients over 60 years old were tested with chi-square test (Fisher's Exact Test was used for theoretical frequency T<1). The statistical test results showed that the different levels of CRP and L were not identical with the distribution of clinical classification and clinical outcome, and there were significant statistical differences between the levels of CRP and L and the clinical classification and clinical outcome of patients admitted to hospital (p<0.05).
Table 3 Correlation between CRP and L and clinical classification and clinical outcome of the elderly over 60 years old
Characteristic
|
Clinical typing at admission
|
|
Outcome the ending
|
|
Moderate
|
Severe & critical type
|
Total
|
Improvement
|
Discharge
|
Aggravation
|
Death
|
Total
|
|
CRP
|
32.10
±30.29
|
70.88
±54.30
|
55.06
±49.72
|
57.81
±45.86
|
37.47
±34.24
|
82.65
±69.71
|
94.24
±60.38
|
55.06
±49.72
|
|
<3
|
2
|
3
|
5
|
0
|
5
|
0
|
0
|
5
|
|
3-100
|
46
|
47
|
93
|
20
|
59
|
3
|
11
|
93
|
|
>100
|
2
|
24
|
27
|
5
|
5
|
3
|
14
|
27
|
|
Total
|
51
|
74
|
125
|
25
|
69
|
6
|
25
|
125
|
|
c2
|
c2=12.744a
|
c2=31.441b
|
|
p
|
P=0.001**
|
P=0.001**
|
|
L
|
1.10
±0.61
|
0.79
±0.54
|
0.91
±0.59
|
1.10
±0.95
|
0.98
± 0.46
|
0.65
±0.34
|
0.61
±0.32
|
0.91
±0.59
|
|
<1.1
|
32
|
61
|
93
|
4
|
23
|
1
|
2
|
30
|
|
1.1-3.2
|
18
|
12
|
30
|
19
|
46
|
5
|
23
|
93
|
|
>3.2
|
1
|
1
|
2
|
2
|
0
|
0
|
0
|
2
|
|
Total
|
51
|
74
|
125
|
25
|
69
|
6
|
25
|
125
|
|
c2
|
c2=6.369b
|
c2=15.641b
|
|
p
|
P=0.029*
|
P=0.020*
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Note: Normal reference values: CRP (<3mg / L); L (1.1× 109 / L -3.2 × 109 / L)
a Chi-square test (double-tail test);b Fisher's exact test(double-tail test)
*At level 0.05 (double-tailed), **At level 0.01 (double-tailed), the correlation was significant.
In the binary Logistic regression analysis of whether the elderly patients had severe or critical clinical type, we used univariate Logistic regression analysis to preliminarily determine the factors affecting clinical type and the degree of risk, and the results showed that CRP, L level had a significant impact on the clinical type of the patients. For each unit increase of CRP index, the risk of disease severity was increased 1.022 times, and the 95% interval of OR value was greater than 1, which constitute the conditions of risk factors. For every unit increase in L level, the risk of disease severity increased by 0.359 times. In order to correct the effects of confounding factors, we will be more variable into the multi-factor Logistic regression analysis, the results showed that CRP index for every rise in unit, 1.019 times the risk of disease severity ascension, OR 95% of the value range are greater than 1, constitute the conditions of the dangerous factors, L level for every rise in unit, the risk of disease severity increase 0.478 times. (See Table 4)
Table 4 Logistic regression analysis of influencing factors of clinical classification in patients over 60 years old
Factor
|
Univariate Logistic regression analysis
|
|
B
|
SE
|
Wald c2
|
p
|
Exp(B)
|
95%CI
|
|
Sex(male)
|
0.650
|
0.383
|
2.881
|
0.090
|
1.915
|
(0.904,4.053)
|
|
HOD
|
-0.042
|
0.030
|
2.006
|
0.157
|
0.959
|
(0.904,1.016)
|
|
CRP
|
0.021
|
0.005
|
15.844
|
0.000
|
1.022
|
(1.011,1.032)
|
|
L
|
-1.024
|
0.396
|
6.962
|
0.008
|
0.359
|
(0.168,0.769)
|
|
Factor
|
Multivariate Logistic regression analysis
|
0.017多因素Logistic回归分析
|
B
|
SE
|
Wald c2
|
p
|
Exp(B)
|
95%CI
|
|
Sex(male)
|
0.268
|
0.431
|
0.386
|
0.534
|
1.307
|
(0.562,3.039)
|
|
HOD
|
-0.027
|
0.034
|
0.627
|
0.428
|
0.973
|
(0.910,1.041)
|
|
CRP
|
0.019
|
0.006
|
11.376
|
0.001
|
1.019
|
(1.008,1.031)
|
|
L
|
-0.738
|
0.356
|
4.290
|
0.038
|
0.478
|
(0.238,0.961)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Note: HOD: Hospitalization days;
In the multivariate Logistic regression analysis of hospitalization days, gender, CRP, L and clinical outcomes of elderly patients, the elderly patients discharged from hospital were compared with elderly patients who died,when each unit increase in CRP, the risk of death in discharged patients was 0.981 times that of those who died. Compared with the elderly patients who died, the increased length of stay in hospital constituted the risk factor for death, and the OR value was 1.330, 1.396 and 1.377. (See Table 5)
Table 5 multivariate Logistic regression analysis of the influencing factors of mortality outcome in patients
Factors
|
B
|
SE
|
Wald c2
|
p
|
Exp(B)
|
95%CI
|
Discharged
|
|
|
|
|
|
|
|
HOD
|
0.285
|
0.069
|
16.854
|
0.000
|
1.330
|
1.161
|
1.524
|
CRP
|
-0.019
|
0.007
|
7.700
|
0.006
|
0.981
|
0.968
|
0.994
|
L
|
2.003
|
0.943
|
4.511
|
0.034
|
7.412
|
1.167
|
47.062
|
[sex = male]
|
0.451
|
0.686
|
0.433
|
0.510
|
1.570
|
0.410
|
6.019
|
[sex = female]
|
0
|
|
|
|
|
|
|
Improvement
|
|
|
|
|
|
|
|
HOD
|
0.334
|
0.076
|
19.159
|
0.000
|
1.396
|
1.202
|
1.621
|
CRP
|
-0.004
|
0.007
|
0.322
|
0.570
|
0.996
|
0.982
|
1.010
|
L
|
2.477
|
0.986
|
6.314
|
0.012
|
11.907
|
1.724
|
82.215
|
[sex = male]
|
-0.037
|
0.765
|
0.002
|
0.961
|
0.963
|
0.215
|
4.312
|
[sex = female]
|
0
|
.
|
.
|
.
|
.
|
.
|
.
|
Exacerbation
|
|
|
|
|
|
|
|
HOD
|
0.320
|
0.105
|
9.249
|
0.002
|
1.377
|
1.121
|
1.693
|
CRP
|
0.000
|
0.010
|
0.000
|
0.991
|
1.000
|
0.981
|
1.020
|
L
|
1.091
|
1.509
|
0.523
|
0.470
|
2.977
|
0.155
|
57.253
|
[sex = male]
|
0.941
|
1.270
|
0.549
|
0.459
|
2.563
|
0.213
|
30.901
|
[sex = female]
|
0
|
.
|
.
|
.
|
.
|
.
|
.
|
Note: HOD: Hospitalization days
In the analysis of the prognostic diagnostic value of CRP and L on death outcome in elderly patients, we used non-death outcome and death outcome as the basis for positive classification. With AUC=0.5 as the null hypothesis, the significance of CRP and L in ROC curve analysis was less than 0.05. It can be seen in the ROC curve that both CRP and L have good diagnostic value for death outcomes. The area under the curve (AUC) of CRP is greater than 0.7, that is 0.751, and the area under the curve (AUC) of L is 0.720. The optimal cutoff values for CRP and L as indicators to determine the outcome of death were calculated using the maximum Youden index, which were 91.5 and 0.615. (See Figure 3)
CT is very important in the clinical diagnosis and typing of COVID-19. This research investigated the CT imaging changes of a patient over 60 years old at different periods, and the results showed that the patient's lung condition gradually improved during the clinical process, and CRP decreased with the improvement of CT imaging. In the above analysis, it was concluded that an increase in CRP index would increase the severity of the disease, which was consistent with CT results. Therefore, in the process of monitoring patients' condition, combining the results of CT imaging to judge the changes of the disease can make the judgment of the condition more accurate. (See figure 4)