Patients’ characteristics
As shown in Table 1, the mean age of the patients was 63.0 ± 14.8 years and 60.4% were men (32 men and 21 women). The mean Acute Physiology and Chronic Evaluation (APACHE) II score was 27 ± 7.6, mean Sequential Organ Failure Assessment (SOFA) score was 9 ± 3.9, and mean acute DIC score was 2 ± 2.5 at 1 day after admission to the ICU. Twenty-nine (54.7%) patients presented with septic shock, 34 (64.2%) with ARDS, and 25 (47.2%) with DIC. ARDS in 20 patients was due to coronavirus disease 2019 (COVID-19). Of these patients, 12 (22.6%) died within 28 days.
Table 1
Patients’ characteristics. Data are reported as the mean ± standard deviation.
Factor | All n = 53 | Survivor n = 41 | Non-Survivor n = 12 | P value |
Age, years | 63.6 ± 14.8 | 61.9 ± 15.4 | 69.4 ± 10.7 | 0.125 |
Male/Female, (%) | 32/21 (60.4/39,6) | 24/17 (58.5/41.5) | 8/4 (66.7/33.3) | 0.612 |
Site of infection | 53 | | | |
- Lung | 30 | 24 | 6 | |
- Abdomen | 15 | 11 | 4 | |
- Urinary tract | 3 | 3 | 0 | |
- Soft Tissues | 5 | 3 | 2 | |
SOFA | 9.7 ± 3.9 | 9.2 ± 3.7 | 11.4 ± 3.9 | 0.08 |
APACHE II | 27.7 ± 7.6 | 26 ± 7.4 | 33.5 ± 4.9 | 0.002* |
Acute DIC score | 2.6 ± 2.5 | 2.6 ± 2.7 | 2.7 ± 1.7 | 0.9 |
Platelet (x103) | 167.5 ± 90.8 | 171.3 ± 87.1 | 154.3 ± 105.3 | 0.573 |
ICU stays, days | 14.6 ± 13.3 | 15.3 ± 14.6 | 12.5 ± 7.7 | 0.529 |
Diagnosis |
Septic shock, (%) ARDS, (%) DIC, (%) | 29 (54.7) 34 (64.2) 25 (47.2) | 19 (46.3) 25 (61) 16 (39) | 10 (83.3) 9 (75) 9 (75) | 0.04* 0.5 0.04* |
Association between serum and plasma HRG concentrations
We analyzed plasma and serum HRG concentrations in blood samples from patients and volunteers. Plasma and serum HRG concentrations showed a positive correlation, with a coefficient (r) of 0.9486. (Y = 0.911X + 4.57, P < 0.0001) (Fig. 1).
Changes in serum HRG concentrations in survivors and non-survivors
In survivors, mean HRG concentrations were 32.5, 28.6, 30.3, and 32.6 µg/mL on days 1, 3, 5, and 7, respectively (Fig. 2). In non-survivors, mean HRG concentrations were 36.3, 23.3, 19.7, and 20.3 µg/mL on days 1, 3, 5, and 7, respectively. In non-survivors, HRG concentrations were significantly lower on days 3 (P < 0.05), 5 (P < 0.01), and 7 (P < 0.05) than on day 1. HRG concentrations were significantly lower in non-survivors than in survivors on days 5 and 7.
Changes in HRG concentrations in patients with septic shock, DIC, and ARDS
HRG concentrations on days 1, 3, 5, and 7 were 35.3, 25.0, 24.1 and 24.3 µg/mL in patients with shock and 31.0, 30.4, 34.0, and 34.4 µg/mL in patients without shock. There was a significant decrease in HRG concentrations in patients with septic shock on days 3, 5, and 7 compared to day 1 (all P < 0.01). In addition, the HRG concentrations were significantly lower in patients with septic shock than in those without septic shock on days 5 and 7 (Fig. 3A).
Mean HRG concentrations were 29.0, 22.7, 22.3, and 22.1 µg/mL in patients with DIC and 35.5, 31.0, 32.5, and 34.9 µg/mL in patients without DIC on days 1, 3, 5, and 7, respectively. In patients with DIC, there was no significant difference in HRG concentrations on days 3, 5, or 7 compared with day 1. The patients with DIC showed significantly lower HRG concentrations than patients without DIC on days 3, 5, and 7 (Fig. 3B).
Mean HRG concentrations were 31.3, 27.6, 29.4, and 28.9 µg/mL in patients with ARDS and 36.9, 27.2, 26.0, and 29.9 µg/mL in patients without ARDS on days 1, 3, 5, and 7, respectively. Although the mean HRG concentration was significantly lower on day 5 than on day 1 in patients without ARDS, no significant difference was found in patients with ARDS. There was no significant difference in HRG concentrations at any time point between patients with ARDS and those without ARDS (Fig. 3C).
HRG concentrations between survivors and non-survivors in patients with septic shock, DIC, and ARDS
Mean HRG concentrations were 35.3, 28.0, 27.4, and 29.6 µg/mL in survivors with septic shock and 35.4, 19.0, 16.6, and 16.4 µg/mL in non-survivors with septic shock on days 1, 3, 5, and 7, respectively. HRG concentrations were lower on days 3, 5 and 7 than on day 1 in non-survivors with septic shock. HRG concentrations were also significantly lower in non-survivors than in survivors with septic shock on days 5 and 7 (both P < 0.05, Fig. 4A).
Mean HRG concentrations were 29.5, 26.0, 26.4, and 27.0 µg/mL in survivors with DIC and 28.1, 14.5, 16.6, and 16.4 µg/mL in non-survivors with DIC on days 1, 3, 5, and 7, respectively. HRG concentrations in non-survivors with DIC were lower than those in survivors with DIC, but significant differences were only observed on days 3 and 5 (both P < 0.05, Fig. 4B).
Mean HRG concentrations were 29.3, 28.9, 32.5, and 34.3 µg/mL in survivors with ARDS and 36.3, 24.2, 20.6, and 20.3 µg/mL in non-survivors with ARDS on days 1, 3, 5, and 7, respectively. HRG concentrations were significantly lower on days 5 and 7 than on day 1 in non-survivors with ARDS, but there was no difference between the time points in survivors with ARDS (Fig. 4C). HRG concentrations were also significantly lower in non-survivors with ARDS than in survivors with ARDS on days 5 and 7.
Receiver operator characteristic analysis to predict the prognosis of critically ill patients with sepsis
We analyzed the predictive ability of HRG concentrations for the prognosis using receiver operator characteristic (ROC) curves (Fig. 5a). The areas under the curve of HRG concentrations were 0.566, 0.651, 0.764, and 0.798 on days 1, 3, 5, and 7, respectively. The sensitivity and specificity of HRG concentrations were 0.77 and 0.75, respectively, on day 5, with a cut-off level of 25.5 µg/mL The sensitivity and specificity of HRG concentrations were 0.75 and 0.81, respectively, on day 7, with a cut-off level of 25.7 µg/mL. It was very close in two cut-off levels. Therefore, Kaplan–Meier analysis was performed by using an HRG concentration of 25.5 µg/mL on day 5. Patients with an HRG concentration < 25.5 µg/mL showed a significantly lower survival curve (log-rank test, P < 0.05) (Fig. 5b).