Participant selection
In this study, 625 TB patients who developed sepsis were identified on the basis of the third international consensus definition. This study included 537 patients who were screened according to the inclusion and exclusion criteria. Figure 1 illustrates the flowchart for selecting study participants.
Baseline characteristics
The baseline characteristics of the patients are presented in Table 1. The enrolled patients were divided into three groups by the tertials of BAR as follows: Q1 group, 1.0–4.52 mg/g; Q2 group, 4.53–7.85 mg/g; and Q3 group, 7.86–60.52 mg/g. The median age of all participants was 63.0 (48.0, 72.0) years; 76.7% were men. There were some differences between the BAR groups with respect to various covariates (i.e., sex, age, alcohol abuse, CKD, diabetes, hypertension, AST, BUN, ALB, creatinine, APACHE II score, and SOFA score) (P < 0.05). Additionally, the overall 28-day all-cause mortality rate was 45.8%; the higher BAR group had a worse prognosis (P < 0.05).
Table 1
Characteristics of study participants by blood urea nitrogen to albumin ratio (BAR) tertiles.
Variables | Total | BAR (mg/g) | P value |
Q1 (1.0-4.52) | Q2 (4.53–7.85) | Q3 (7.86–60.52) |
No. | 537 | 179 | 179 | 179 | |
Patient characteristics | | | | | |
sex | | | | | 0.002 |
Women | 125 (23.3) | 55 (30.7) | 43 (24) | 27 (15.1) | |
Men | 412 (76.7) | 124 (69.3) | 136 (76) | 152 (84.9) | |
Age (years) | 63.0 (48.0, 72.0) | 56.0 (41.0, 67.0) | 65.0 (51.5, 73.5) | 67.0 (54.0, 75.5) | < 0.001 |
Education level (years) | | | | | 0.183 |
< 9 | 321 (59.8) | 98 (54.7) | 115 (64.2) | 108 (60.3) | |
≥ 9 | 216 (40.2) | 81 (45.3) | 64 (35.8) | 71 (39.7) | |
BMI (Kg/m2) | 19.0 (18.3, 21.0) | 19.3 (18.6, 21.7) | 19.1 (18.3, 20.5) | 19.0 (18.0, 20.3) | 0.081 |
Smoking | 299 (55.7) | 86 (48) | 98 (54.7) | 115 (64.2) | 0.008 |
Alcohol abuse | 199 (37.1) | 54 (30.2) | 66 (36.9) | 79 (44.1) | 0.024 |
Comorbidities | | | | | |
CKD | 34 (6.3) | 0 (0) | 3 (1.7) | 31 (17.3) | < 0.001 |
Chronic liver disease | 30 ( .6) | 6 (3.4) | 11 (6.1) | 13 (7.3) | 0.252 |
Diabetes | 103 (19.2) | 25 (14) | 34 (19) | 44 (24.6) | 0.039 |
Hypertension | 114 (21.2) | 29 (16.2) | 35 (19.6) | 50 (27.9) | 0.02 |
HIV | 17 (3.2) | 5 (2.8) | 6 (3.4) | 6 (3.4) | 0.941 |
TB epidemiology | | | | | |
Retreated TB | | | | | 0.153 |
No | 428 (79.7) | 137 (76.5) | 140 (78.2) | 151 (84.4) | |
Yes | 109 (20.3) | 42 (23.5) | 39 (21.8) | 28 (15.6) | |
DR-TB | 87 (16.2) | 37 (20.7) | 29 (16.2) | 21 (11.7) | 0.072 |
laboratory test results | | | | | |
ALT (U/L) | 23.0 (13.0, 42.0) | 21.0 (12.0, 36.0) | 24.0 (14.0, 43.0) | 26.0 (12.0, 45.0) | 0.331 |
AST (U/L) | 35.0 (23.0, 65.0) | 29.0 (20.0, 48.5) | 40.0 (24.0, 68.0) | 43.0 (25.5, 75.5) | < 0.001 |
TBIL (umol/L) | 10.3 (6.9, 16.4) | 9.7 (7.0, 14.4) | 11.1 (6.8, 17.3) | 10.3 (7.2, 17.4) | 0.251 |
BUN (mg/dL) | 16.3 (11.6, 25.4) | 9.7 (7.9, 12.0) | 16.3 (13.9, 19.0) | 31.0 (24.2, 47.5) | < 0.001 |
ALB (g/dL) | 2.8 (2.4, 3.1) | 3.0 (2.7, 3.3) | 2.8 (2.4, 3.2) | 2.6 (2.2, 2.9) | < 0.001 |
Creatinine (umol/L) | 57.0 (45.0, 81.7) | 46.1 (38.1, 56.8) | 57.2 (47.5, 74.1) | 86.6 (57.0, 163.7) | < 0.001 |
PF ratio (mmHg) | 162.7 (129.2, 193.9) | 166.3 (136.0, 190.9) | 165.0 (131.7, 189.4) | 150.8 (117.6, 202.0) | 0.498 |
Disease severity score | | | | | |
APACH II score | 17.0 (13.0, 21.0) | 16.0 (13.0, 20.0) | 16.0 (13.0, 20.0) | 19.0 (16.0, 23.5) | < 0.001 |
SOFA score | 4.0 (3.0, 6.0) | 3.0 (3.0, 4.0) | 4.0 (3.0, 6.0) | 6.0 (4.0, 8.0) | < 0.001 |
outcome | 246 (45.8) | 45 (25.1) | 84 (46.9) | 117 (65.4) | < 0.001 |
Data are presented as median (IQR) or N (%).
BAR, blood urea nitrogen-to-albumin ratio; IQR, interquartile range; BMI, body mass index; CKD, chronic kidney disease; DR-TB, drug-resistant TB; ALT, alanine aminotransferase; AST, aspartate aminotransferase; TBIL, total bilirubin; BUN, Blood Urea Nitrogen, ALB, albumin; PF ratio, PaO2/FiO2 ratio; APACHE II, Acute Physiology and Chronic Health Evaluation; SOFA, Sequential Organ Failure Assessment.
Outcomes
The overall 28-day all-cause mortality was 45.8%. Table 1 shows the 28-day all-cause mortality rates in different BAR groups. The 28-day all-cause mortality rates in groups 1–3 were 25.1%, 46.9%, and 65.4%, respectively. Kaplan-Meier curves showed that the higher the BAR, the higher the risk of 28-day all-cause mortality (log-rank test: P = 0.0001, Fig. 2).
BAR and 28-day all-cause mortality
The results of the multivariate COX regression models are shown in Table 2, which were used to assess the relationship between the BAR and 28-day all-cause mortality in TB patients complicated by sepsis, are shown in Table 2. In the model 1 adjusted for sex and age, the BAR was positively associated with 28-day all-cause mortality (hazard ratio [HR], 1.04; 95% confidence interval [CI], 1.02–1.05; P < 0.001). Despite adjusting for various covariates, the risk of 28-day all-cause mortality increased by 4–5% for each 1 mg/g increase in BAR. At the same time, in models 1 to 5, which were also adjusted for covariates (Table 2), the categorized BAR in the multivariate COX regression model seemed to confirm a linear relationship between BAR and 28-day all-cause mortality. The 1.0-4.52 mg/g BAR group had the lowest 28-day all-cause mortality rate.
Table 2
Relationship between BAR and the 28-day all-cause mortality in TB patients complicated by sepsis.
Exposure | Model 1 | Model 2 | model 3 | Model 4 | Model 5 |
HR (95% CI) | P value | HR (95% CI) | P value | HR (95% CI) | P value | HR (95% CI) | P value | HR (95% CI) | P value |
BAR (mg/g) | 1.04(1.02 ~ 1.05) | < 0.001 | 1.04(1.02 ~ 1.05) | < 0.001 | 1.05(1.03 ~ 1.07) | < 0.001 | 1.05(1.03 ~ 1.07) | < 0.001 | 1.04(1.03 ~ 1.07) | < 0.001 |
BAR tertiles (mg/g) | | | | | | | | | | |
Q1(1.0-4.52) | 1(Ref) | | 1(Ref) | | 1(Ref) | | 1(Ref) | | 1(Ref) | |
Q2(4.53–7.85) | 2.14 (1.49 ~ 3.09) | < 0.001 | 2.23 (1.55 ~ 3.23) | < 0.001 | 2.19 (1.52 ~ 3.16) | < 0.001 | 2.09 (1.45 ~ 3.03) | < 0.001 | 2.03 (1.41 ~ 2.95) | < 0.001 |
Q3(7.86–60.52) | 3.45 (2.43 ~ 4.9) | < 0.001 | 3.41 (2.37 ~ 4.89) | < 0.001 | 3.25 (2.24 ~ 4.72) | < 0.001 | 3.13 (2.15 ~ 4.56) | < 0.001 | 3.02 (2.06 ~ 4.41) | < 0.001 |
P for trend | | < 0.001 | | < 0.001 | | < 0.001 | | < 0.001 | | < 0.001 |
Model 1 adjusted for sex and age; Model 2 adjusted for covariates from model 1 plus CKD, chronic liver disease, diabetes, hypertension, HIV, retreated TB, and drug resistant TB; Model 3 adjusted for covariates from model 2 plus aspartate aminotransferase, creatinine, and PaO2/FiO2 ratio (PCOX <0.1); Model 4 adjusted for variables from model 2 plus education level, BMI, alcohol abuse, creatinine, and PaO2/FiO2 ratio (effect value > 10%); Model 5 adjusted for sex, age, education level, BMI, smoking, alcohol abuse, CKD, chronic liver disease, diabetes, hypertension, HIV, retreated TB, DR-TB, ALT, AST, TBIL, creatinine, and PF ratio.
HR, hazard ratio.
Subgroup analyses
The results of the subgroup analyses are shown in Fig. 3. BAR was associated with 28-day all-cause mortality among men (HR, 1.05; 95% CI, 1.03–1.07), those aged 45 to 64 years (HR,1.08; 95% CI, 1.04–1.12), those without diabetes (HR 1.06; 95% CI, 1.03–1.08), these without hypertension (HR, 1.06, 95% CI, 1.03–1.1), retreated TB (HR,1.05; 95% CI, 1.03–1.07), and those without drug-resistant TB (HR, 1.04; 95% CI, 1.02–1.06). There was no association among women, participants aged 18 to 44 years and 65 to 80 years, diabetes, hypertension, retreated TB, and drug-resistant TB.
Non-linear relationship between BAR and 28-day all-cause mortality
After adjusting for sex, age, education level, BMI, smoking, alcohol abuse, CKD, chronic liver disease, diabetes, hypertension, HIV, retreated TB, drug-resistant TB, ALT, AST, TBIL, creatinine, and PF ratio, we observed a non-linear relationship between BAR and 28-day all-cause through a restricted cubic spline (P for non-linearity was 0.001, Fig. 4). Using a two-piecewise Cox regression model adjusted for sex, age, education level, BMI, smoking, alcohol abuse, CKD, chronic liver disease, diabetes, hypertension, HIV, retreated TB, drug-resistant TB, ALT, AST, TBIL, creatinine, and PF ratio, we found that the BAR threshold was 10.168 mg/g (Table 3). Below the threshold, the 28-day all-cause mortality rose rapidly (HR, 1.166; 95% CI, 1.082–1.257; P < 0.001; Table 3); above the threshold, the 28-day all-cause mortality did not rise rapidly, wherein the estimated dose-response curve appeared to be within a consistent horizontal line (HR = 1.017; 95% CI, 0.99–1.046; P = 0.2179; Table 3). This suggests that when the BAR was below 10.168 mg/g, the risk of 28-day all-cause mortality increased by 16.6% per 1 mg/g increase in BAR. In contrast, when the BAR value was greater than 10.168 mg/g, the risk of 28-day all-cause mortality increased by only 1.7% for every 1 mg/g increase in BAR.
Table 3
Threshold effect analysis of the relationship of BAR and the 28-day all-cause mortality in TB patients complicated by sepsis.
BAR (mg/g) | No. | Adjusted Model |
HR(95%CI) | P-Value |
< 10.168 | 467 | 1.166 (1.082,1.257) | < 0.001 |
≥ 10.168 | 70 | 1.017 (0.99,1.046) | 0.2179 |
Likelihood Ratio test | | - | 0.001 |
Note: Adjusted for sex, age, education level, BMI, smoking, alcohol abuse, CKD, chronic liver disease, diabetes, hypertension, HIV, retreated TB, drug-resistant TB, ALT, AST, TBIL, creatinine, and the PF ratio. All the data are displayed. |
Sensitivity analyses
Although we excluded patients with comorbidities, such as CKD, chronic liver disease, or both, the multivariate COX regression model, after adjusting for each covariate, still demonstrated a 4% to 9% increase in the 28-day all-cause mortality for each 1-mg/g increase in BAR (Table 4). Additionally, our analysis using restricted cubic splines revealed a non-linear relationship between BAR and 28-day all-cause mortality in TB patients complicated by sepsis in the ICU (All P values for non-linearity < 0.05).
Table 4. Sensitivity Analyses.
Analysis
|
Unweighted participants/
total participants, No.
|
Adusted HR (95%CI)a
|
P value
|
Excluding CKD
|
223/503
|
1.09(1.05-1.12)
|
P<0.001
|
Excluding chronic liver disease
|
237/507
|
1.04(1.02-1.06)
|
P<0.001
|
Excluding CKD and chronic liver disease
|
215/474
|
1.08(1.04-1.11)
|
P<0.001
|
Abbreviations: HR, Hazard ratio; CI, confidence interval; CKD, chronic kidney disease.
aAdjusted for sex, age, education level, BMI, smoking, alcohol abuse, diabetes, hypertension, HIV, retreated TB, drug resistant TB, ALT, AST, TBIL, creatinine, and PF ratio.