A total of 1150 patients conformed to sepsis-3 and the first measurement of CVP after admission to the ICU being greater than 12 mmHg, and the data on these patients were extracted from the database (Fig. 1). Patients included in the analysis comprised 847 survivors and 303 non-survivors. Figure 2 shows the amounts of fluids used in the treatment (Fig. 2). Non-survivors had higher age, CVP, SOFA score, Apache IV score, respiratory rate, heart rate, creatinine, tracheal intubation rate, and vasoactive drug usage rate than the survivors (P < 0.05). Non-survivors had lower blood pressure and BMI than the survivors (P < 0.05). There were no significant differences in gender, history of chronic heart failure, Hb, cardiotonic drug use, and WBC count between the two groups (Table 1).
Table 1
Differences in the baseline characteristics between survivors and non-survivors.
Baseline characteristics | Survivor (n = 847) | Non-survivor (n = 303) | P |
Age (years) | 65 (54, 74) | 71 (58, 79) | < 0.001 |
Male % (n) | 52.25% (451) | 53.14% (161) | 0.973 |
BMI (kg/m2) | 29.45 (24.80, 36.58) | 27.51 (23.02, 33.85) | < 0.001 |
CVP (mmHg) | 15 (13, 19) | 16 (14, 20) | 0.002 |
SOFA | 7(5, 10) | 11 (8, 13) | < 0.001 |
APACHE IV | 71.12 (57, 91) | 98 (78, 121) | < 0.001 |
MAP (mmHg) | 54 (46, 118) | 50 (42, 127) | < 0.001 |
RR (beats per minute) | 29 (24, 34) | 30 (26, 35) | 0.002 |
HR (bpm) | 112.57 (98, 127) | 121 (105, 135) | < 0.001 |
WBC (109/L) | 14.4 (10.16, 20.9) | 16.06 (10.8, 23) | 0.056 |
Hb (g/dL) | 9.2 (10.5, 12) | 9.2 (10.6, 12.2) | 0.261 |
Cre (mg/dl) | 1.6 (1.02, 2.66) | 2.2 (1.40, 3.24) | < 0.001 |
CHF (n/%) | 18.06% (153) | 22.44% (68) | 0.097 |
Venti (N = 1128) (%) | 53.67% (446) | 79.80% (177) | < 0.001 |
Shock (N = 1128) (%) | 32.80% (254) | 48.18% (146) | < 0.001 |
Cardiac agents (N = 1128) (%) | 4.01% (34) | 5.94% (18) | 0.166 |
Abbreviations: BMI, Body Mass Index; CVP, Central Venous Pressure; SOFA, Sequential Organ Failure Assessment; MAP, Mean Arterial Pressure; RR, Respiratory Rate; HR, Heart Rate; WBC, White blood cell count; Hb, Hemoglobin; Cre, creatinine; CHF, Chronic Heart Failure; Venti, ventilation. |
Fluid balance and mortality in sepsis patients and different subgroup
The fluid balances at 6, 24, and 48 h in survivors were significantly less than those in the non-survivors (Fig. 3, Table 2). In the sepsis subgroup without shock, we found that the fluid balances of the survivors at 24 and 48 h were significantly less than those of non-survivors (P < 0.05), but there were no significant differences at 6 h (Fig. 4, Table 2). In patients with septic shock, there were no significant differences in the fluid volume between survivors and non-survivors at 6, 24, and 48 h (P > 0.05) (Fig. 4, Table 2). In patients without chronic heart failure, we found that survivors had significantly lower fluid balance than non-survivors (P < 0.05) (Fig. 4, Table 2). In patients with chronic heart failure, we found that there were no significant differences at 6, 24, and 48 h (Fig. 4, Table 2).
Table 2
Fluid balance in survivors and non-survivors.
Time points | Survivors | Non-survivors | P |
6 h | 2.56 ± 10.78 | 4.42 ± 11.25 | 0.008 |
24 h | 5.08 ± 33.72 | 15.25 ± 36.54 | < 0.001 |
48 h | 2.91 ± 52.16 | 20.47 ± 53.53 | < 0.001 |
subgroup analysis |
Non-shock |
6 h | 1.18 ± 10.24 | 1.89 ± 8.91 | 0.425 |
24 h | -0.87 ± 30.64 | 5.49 ± 30.12 | 0.021 |
48 h | -6.63 ± 47.55 | 6.51 ± 45.45 | 0.002 |
septic shock |
6 h | 5.80 ± 11.39 | 7.16 ± 12.84 | 0.272 |
24 h | 18.97 ± 36.43 | 25.76 ± 39.87 | 0.084 |
48 h | 25.12 ± 55.28 | 33.99 ± 59.69 | 0.079 |
Non-CHF |
6 h | 2.65 ± 10.93 | 5.10 ± 12.13 | 0.004 |
24 h | 5.67 ± 34.43 | 17.36 ± 39.13 | < 0.001 |
48 h | 3.72 ± 53.16 | 23.36 ± 56.43 | < 0.001 |
CHF |
6 h | 2.16 ± 10.22 | 2.11 ± 7.18 | 0.976 |
24 h | 2.41 ± 30.25 | 7.97 ± 24.51 | 0.184 |
48 h | -0.76 ± 47.34 | 10.50 ± 40.80 | 0.09 |
CHF, Chronic Heart Failure. |
Fluid Balance And Mortality In Sepsis Patients
It was found that fluid balance and mortality in patients with sepsis showed an inverted U-shaped relationship, and the highest mortality was at 0 ml/kg fluid balance. With the increase of negative balance, mortality decreased gradually. At 6 h, mortality was significantly lower at < -5 ml/kg fluid balance compared to the other time intervals. At 24 h, mortality was significantly lower when the fluid balance was at -40 ~ 0 ml/kg compared to other fluid balance intervals. At 48 h, significantly lower mortality was observed in patients with < -40 ml/kg fluid balance compared to the other fluid balance intervals (Fig. 5, Fig. 6).
Fluid Balance And Mortality In Sepsis Patients Without Shock
Similar results were obtained in the subgroup analysis of sepsis patients without shock. The fluid balance and mortality in septic patients at 6, 24, and 48 h showed an inverted U-shaped relationship, and the highest mortality was observed at 0 ml/kg fluid balance. Mortality was significantly lower at < -5 ml/kg at 6 h compared to the other intervals. At 24 h, mortality was significantly lower at -40~-20 ml/kg compared to the other intervals. At 48 h, mortality was significantly lower in patients with − 60~-40 ml/kg compared to the other intervals (Fig. 7, Fig. 8A).
Fluid Balance And Mortality In Septic Shock Patients
Subgroup analysis of patients with septic shock showed that there was an L-shaped relationship between fluid balance and mortality at 6 h. With an increase in the fluid volume, mortality risk gradually decreased. However, at 6 h, the OR of mortality in patients with fluid balances of < -5 ml/kg, ~-5 ~ 0 ml kg, 0 ~ 5 ml/kg, 5 ~ 10 ml/kg, 10 ~ 20 ml/kg, and ≥ 20 ml/kg were not statistically different. At 24 and 48 h, fluid balance and mortality showed an inverted U-shaped relationship. At 24 h, significantly lower mortality was observed at 30 ~ 50 ml/kg compared to other intervals. At 48 h, significantly lower mortality was observed at ≥ 80 ml/kg compared to other intervals. (Fig. 8B, Fig. 9).
Fluid balance and mortality in sepsis patients without heart failure
Fluid balance and mortality showed an inverted U-shaped relationship in the subgroup of sepsis patients without heart failure at 6, 24, and 48 h but there were no significant differences in the fluid balances at -10~-5 ml/kg, ~-5 ~ 0 ml/kg, 0 ~ 5 ml/kg, 5 ~ 10 ml/kg, 10 ~ 20 ml/kg, 20 ~ 30 ml/kg, and ≥ 30 ml/kg at 6 h. At 24 h, significantly lower mortality was observed at -40~-20 ml/kg compared to the other intervals. At 48 h, significantly lower mortality was observed at -60~-40 ml/kg compared to other intervals (Fig. 8C, Figure S1).
Fluid balance and mortality in sepsis patients with heart failure
Subgroup analysis of sepsis patients with heart failure showed that there was significantly lower mortality at < -5 ml/kg at 6 h compared to the other intervals. At 24 h, there were no significant differences in the fluid balance between the < -20 ml/kg, -20~-10 ml/kg, -10 ~ 0 ml/kg, 0 ~ 10 ml/kg, 10 ~ 30 ml/kg, and ≥ 30 ml/kg groups. Mortality was significantly lower with the fluid balance of < -40 ml/kg at 48 h compared to the other intervals (Fig. 8D, Figure S2).