A total of 76 severe COVID-19 patients (age 63.5 ± 14.7 years) were included in the study. Among these 45 (59%) were males and 31 (41༅) were female (p-value = 0.89). Among the dead, all the subjects were older than 50 years (72.6 ± 12.1 years). This also included 6 patients that were 50–65 years old (31.5%). A total of 13 patients were over 65 years (68.5%) and approximately half of the patients had at least one chronic disease on admission. The most common diseases were hypertension (n = 29,38%), cardiovascular disease (n = 14,18%), diabetes (n = 11,15%), lung disease (n = 3,4%), tumor (n = 3,4%), tuberculosis (n = 3,4%), hepatic disease (n = 4,5%), cerebrovascular disease (n = 1,1%), connective tissue diseases (n = 1,1%).
The most common symptoms at the onset of illness were fever (53 [70%] of 76 patients), cough (36 [47%]), and myalgia or fatigue (4 [5%]). Relatively less common symptoms were headache (1 [1%]), diarrhea (3[4%]). Fourteen (14) patients (18%) had developed dyspnea.
Peripheral white blood cell counts were normal in severe and critically severe patients (p-value > 0.05). A decrease in the number of lymphocytes was noted in non-survivors but not in survivors (p-value > 0.05). The value of c-reactive protein for all the patients was above the normal range (p-value > 0.05). The average levels of platelets, blood potassium, blood calcium, blood phosphorus and blood glucose in 76 patients were normal (p-value > 0.05). All the patients on admission showed hyponatremia (p-value > 0.05) and hypoalbuminemia (average 36.74 g/L). These values were lower in non-survivors as compared to the survivors (p-value = 0.006) (Table 1).
Table 1
Demographic, Clinical and Laboratory Findings of Patients on Admission
Characteristics | All patients (n = 76) | Survivors(n = 57) | Nonsurvivors(n = 19) | p value |
Age(years, mean ± SD ) | 63.5 ± 14.7 | 60.5 ± 14.4 | 72.6 ± 12.1 | 0.001 |
Sex | | | | 0.89 |
Male | 45(59%) | 34(60%) | 11(58%) | |
Female | 31(41%) | 23(40%) | 8(42%) | |
Charlton’s weighted index of capabilities (Median༈Quartile low, Quartile up)༉a | 3(2,3) | 2(0,3) | 4(3,6) | 0.000 |
Acute physiology and chronic health evaluation II score ( mean ± SD)b | 9.4 ± 4.8 | 8.2 ± 4.4 | 13.3 ± 3.9 | 0.000 |
Comorbidity | | | | |
Hypertension | 29(38%) | 21(36%) | 8(42%) | 0.68 |
Lung disease | 3(4%) | 2(4%) | 1(5%) | 0.73 |
Hepatic disease | 4(5%) | 3(5%) | 1(5%) | 1 |
Diabetes | 11(15%) | 10(18%) | 1(5%) | 0.19 |
Cerebrovascular disease | 1(1%) | 0 | 1(5%) | 0.08 |
Connective tissue diseases | 1(1%) | 0 | 1(5%) | 0.08 |
Tumor | 3(4%) | 1(2%) | 2(10%) | 0.09 |
Tuberculosis | 3(4%) | 2(4%) | 1(5%) | 0.73 |
Cardiovascular disease | 14(18%) | 9(16%) | 5(26%) | 0.34 |
Signs and symptoms | | | | |
Fever | 53(70%) | 39(63%) | 14(74%) | 0.67 |
Cough | 36(47%) | 28(49%) | 8(42%) | 0.59 |
Myalgia or fatigue | 4(5%) | 4(7%) | 0 | 0.24 |
Headache | 1(1%) | 1(2%) | 0 | 0.56 |
Diarrhea | 3(4%) | 1(2%) | 2(10%) | 0.09 |
Dyspnea | 14(18%) | 9(16%) | 5(26%) | 0.31 |
> 24 breaths per min | 36(47%) | 23(40%) | 13(68%) | 0.03 |
Mean arterial pressure, mm Hg | 97.6 ± 10.2 | 97.3 ± 10.2 | 96.6 ± 10.4 | 0.79 |
Laboratory parameters | | | | |
Lymphocyte count(10^9/L) | 0.85 ± 0.65 | 0.89 ± 0.71 | 0.71 ± 0.44 | 0.29 |
Albumin,g/L | 36.74 ± 4.93 | 37.45 ± 4.43 | 34.60 ± 5.80 | 0.02 |
Creatinine,µmol/L | 65.47 ± 20.85 | 63.16 ± 20.08 | 72.41 ± 22.08 | 0.09 |
White blood cell count, × 10⁹ per L | 5.4 ± 2.68 | 5.08 ± 2.03 | 6.35 ± 3.98 | 0.07 |
Neutrophil count, × 10^9 per L | 4.11 ± 2.54 | 4.39 ± 2.76 | 3.25 ± 1.50 | 0.09 |
Hemoglobin, g/L | 129.63 ± 30.04 | 126.68 ± 23.19 | 138.47 ± 44.45 | 0.13 |
Platelet count, × 10⁹ per L | 150.80 ± 57.79 | 157.72 ± 58.82 | 130.05 ± 50.48 | 0.07 |
C-reactive protein, mg/L | 57.72 ± 36.48 | 55.55 ± 37.86 | 64.22 ± 32.03 | 0.37 |
Potassium, mmol/L | 4.01 ± 0.55 | 3.99 ± 0.53 | 4.05 ± 0.61 | 0.65 |
Sodium, mmol/L | 130.95 ± 30.93 | 130.96 ± 30.67 | 131.46 ± 32.53 | 0.93 |
Phosphate, mmol/L | 0.97 ± 0.23 | 0.95 ± 0.21 | 1.04 ± 0.26 | 0.15 |
Blood sugar,mmol/L | 9.58 ± 15.20 | 10.19 ± 17.49 | 7.73 ± 2.61 | 0.55 |
SD = standard deviation. |
Data are median (IQR), n (%), or n/N (%). |
aCharlton’s weighted index of Capabilities was assessed within 24 hr of admission to hospital |
bAcute Physiology and Chronic Health Evaluation II score was assessed within 24 hr of admission to hospital |
The survivor group had a median WIC of 2 (0,3) while the median for the non-survivor group was 4(3,6). 14 patients (18.4%) had WIC of 0 and survived, 2 patients (2.6%) had WIC of 1, 20 patients (26.3%) had WIC of 2, 22 patients (28.9%) had WIC of 3, 7 patients (9.2%) had WIC of 4, and 4 patients (5.3%) had WIC of 5. The group with WIC greater than or equal to 6 had 7 patients and all of them died. An increasing WIC was found to be associated with higher mortality. (Fig. 1)
The group whose APACHE Ⅱ score ranged from 0 to 5 had 21 patients and only one died. There were 28 patients with APACHE Ⅱ scores ranging from 6 to 10. A total of 20 patients were part of the group with APACHE II values ranging from 11 to 15 and 6 patients in the group with values ranging from 16 to 20. Only one patient had score higher than 20 and he/she did not survive. (Fig. 2)
We also found a correlation (Medium) between the APACHE Ⅱ score and WIC score (r = 0.565; p-value < 0 .000). Multivariate Logistic regression analysis was performed using age, WIC score, APACHE Ⅱ score, albumin and respiratory rate > 24 breaths per min. These were indicated as covariates by univariate logistic regression (variable inclusion criteria p-value < 0.05). The multivariate logistic regression model identified APACHE Ⅱ score and the WIC as independent risk factors of COVID-19 death. (Table 2)
Table 2
Risk Factors Associated with In-hospital Death
Demographics and clinical characteristics | Univariable OR(95% CI) | p value | Multivariable OR (95% CI) | p value |
Age(years) | 1.07(1.02–1.12) | 0.003 | 0.98(0.91–1.06) | 0.681 |
Female sex(vs male) | 0.93(0.324–2.667) | 0.890 | - | - |
Comorbidity present (vs not present) | | | | |
Hypertension | 0.80(0.28–2.31) | 0.683 | - | - |
Lung disease | 0.66(0.06–7.65 | 0.735 | - | - |
Hepatic disease | 1(0.09–10.23) | 1.000 | - | - |
Diabetes | 3.83(0.46–32.11) | 0.216 | - | - |
Tumor | 0.15(0.01–1.78) | 0.133 | - | - |
Tuberculosis | 0.66(0.06–7.65) | 0.735 | - | - |
Cardiovascular disease | 0.53(0.15–1.82) | 0.310 | - | - |
> 24 breaths per min | 0.312(0.10–0.94) | 0.039 | 2.31(0.51–4.19) | 0.277 |
Charlton’s weighted index of capabilities | 2.53(1.54–4.16) | 0.000 | 2.18(1.14–4.19) | 0.019 |
Acute physiology and chronic health evaluation II score | 1.30(1.13–1.49) | 0.000 | 1.28(1.08–1.51) | 0.004 |
laboratory parameters | | | | |
Lymphocyte count(10^9/L) | 0.32(0.07–1.49) | 0.146 | - | |
Albumin | 0.88(0.78–0.99) | 0.034 | 1.04(0.89–1.22) | 0.558 |
Creatinine,µmol/L | 1.02(0.99–1.05) | 0.097 | - | - |
White blood cell count,x10^9 per L | 1.13(0.88–1.44) | 0.336 | - | - |
Neutrophil count, × 10^9 per L | 0.76(0.55–1.04) | 0.090 | - | - |
Haemoglobin, g/L | 1.01(0.99–1.03) | 0.190 | - | - |
Platelet count, × 10⁹ per L | 0.99(0.98-1.00) | 0.077 | - | - |
C-reactive protein, mg/L | 1.01(.099-1.02) | 0.369 | - | - |
Potassium, mmol/L | 1.25(0.49–3.19) | 0.645 | - | - |
Sodium, mmol/L | 1.00(0.98–1.02) | 0.933 | - | - |
Phosphate, mmol/L | 5.26(0.55–50.63) | 0.151 | - | - |
Blood sugar,mmol/L | 0.97(0.874–1.09) | 0.630 | - | - |
Univariable and multivariate logistic regression analysis of risk factors associated with hospitalized patients. |
OR = Odds Ratio.CI = Confidence Interval. |
When the WIC score and the APACHE Ⅱ score were substituted into Logit (P) =-5.724 + 0.814xWIC + 0.199xAPACHE Ⅱ and the predicted probability of the equation was saved as the combined value of the two scores. The area under the ROC curve (AUC) of the WIC score, APACHE Ⅱ score, and their combined predicted probabilities were 0.814, 0.854, and 0.891, respectively. This further indicates that they are highly valuable to predict the prognosis of the COVID-19. The AUC of joint detection was the largest indicating that the accuracy rate of predicting survival outcomes was greater than any single indicator (Fig. 3).
The cutoff value for Charlton’s WIC as revealed by the ROC curve was 2.5 while the APACHE Ⅱ was 10.5. The sensitivity and specificity of Charlton’s WIC were 84.2% and 57.9% while the APACHE Ⅱ was 84.2% and 64.9%, respectively. The ROC curve of the joint detection is 0.891. The sensitivity and specificity were 94.7% and 68.4%, respectively (Table 3). The diagnostic value of the joint detection was better than WIC (p-value = 0.002) and APACHE II (p-value = 0.042).
Table 3
Cut-off Value,95% CI, P value, Cut-off value, Sensitivity and Specificity of Predictors of Mortality
Predictors of Mortality | AUCa | 95% CI | P value | cut-off value | sensitivity(%) | specificity(%) |
WICb | 0.814 | 0.705–0.923 | 0.000 | 2.5 | 84.2 | 57.9 |
APACHE Ⅱc | 0.854 | 0.753–0.963 | 0.000 | 10.5 | 84.2 | 64.9 |
Joint detectiond | 0.891 | 0.818–0.966 | 0.000 | 0.947 | 94.7 | 68.4 |
aAUC= Area under the receiver operating characteristic curves. |
bCharlton’s weighted index of Capabilities were assessed within 24 hr of admission to hospital |
cAcute Physiology and Chronic Health Evaluation (APACHE) II score was assessed within 24 hr of admission to hospital. WIC = Charlton’s weighted index of comorbidities |
d Predicted probability of the equation by substituting WIC score and the APACHE Ⅱ score into Logit (P) =-5.724 + 0.814xWIC + 0.199xAPACHE Ⅱ |