Serum inflammatory biomarkers can predict clinical outcomes in patients with sepsis-associated gastrointestinal dysfunction

Abstract

BACKGROUND

Effective predictive indicators for clinical outcomes in septic patients with gastrointestinal dysfunction (GID) remain lacking.

AIM

To evaluate the relationship between serum inflammatory biomarkers (SIBs) and clinical outcomes in patients with sepsis-induced GID.

METHODS

A total of 117 sepsis-induced GID patients were selected (January 2021 to January 2025). They were grouped into poor [n = 76; Acute Physiology and Chronic Health Evaluation II (APACHE II) score > 10] and good (n = 41; APACHE II ≤ 10) prognosis groups based on outcomes. Clinical variables (sex, age, body mass index, smoking history, diabetes, hypertension, and infection site) were collected. SIBs, encompassing procalcitonin (PCT), C-reactive protein (CRP), interleukin (IL)-6/10, and tumor necrosis factor (TNF)-α, were analyzed. Correlation analyses were conducted to assess the associations of SIBs with Sequential Organ Failure Assessment (SOFA) and APACHE II scores. Receiver operating characteristic curves visualized SIBs’ predictive performance, and multivariate analysis identified independent outcome predictors.

RESULTS

The groups were similar in age, sex, body mass index, comorbidities, and major infection site. The poor prognosis group exhibited elevated APACHE II, SOFA, PCT, CRP, IL-6, IL-10, and TNF-α than the good prognosis cohort. SIBs correlated positively with both APACHE II and SOFA scores. When used to predict outcomes individually, SIBs yielded an area under the curve range of 0.660-0.780 in septic patients with GID, whereas combined biomarker analysis increased the area under the curve to 0.906. APACHE II, SOFA, PCT, CRP, IL-6, and TNF-α acted as independent predictors of prognosis.

CONCLUSION

SIBs correlate intimately with clinical outcomes in sepsis-induced GID patients.

Key Words: Serum inflammatory biomarkers; Procalcitonin; C-reactive protein; Interleukin-6/10; Sepsis; Gastrointestinal dysfunction; Clinical outcome; Predictive value

Core Tip: This study analyzed 117 patients with sepsis-associated gastrointestinal dysfunction and identified procalcitonin, C-reactive protein, tumor necrosis factor-alpha, and interleukin-6/10 as key biomarkers of disease severity and prognosis. Elevated C-reactive protein and tumor necrosis factor-alpha levels and higher Sequential Organ Failure Assessment scores increased the likelihood of poor outcomes, while combining multiple biomarkers significantly improved predictive accuracy.

INTRODUCTION

Life-threatening sepsis is caused by the imbalance of host response to infection, and the abnormal inflammatory response and immune activation disorder constitute its core pathophysiological mechanism[1,2]. It is among the principal determinants of morbidity and mortality among intensive care unit (ICU) inpatients and across both low- and high-income countries[3]. Epidemiological data indicate that sepsis affects approximately 50 million people globally annually, with a mortality rate of approximately 30% among those with complications[4,5]. Gastrointestinal dysfunction (GID) has been identified as a driving factor of disease progression in sepsis. Sepsis can induce epithelial apoptosis, increase intestinal permeability, and exacerbate enterogenic infections, leading to multiple organ failure[6]. The incidence of GID in ICU patients approaches 85% and exhibits a close connection with poor outcomes[7]. However, no effective predictive index currently exists for clinical outcome evaluation in sepsis-induced GID patients[8]. Identifying effective indicators could therefore significantly improve clinical management and outcomes in this population.

Procalcitonin (PCT), C-reactive protein (CRP), interleukin (IL)-6/10, and tumor necrosis factor (TNF)-α are commonly studied serum inflammatory biomarkers (SIBs). PCT and CRP are associated with apoptotic pathways and bacterial infection, and their elevated expression has prognostic significance for short-term mortality (30 days) in bloodstream infections[9]. They are also linked to bacterial gastroenteritis, suggesting potential utility in predicting GID among septic patients[10]. Previous studies have shown that IL-6, PCT, and CRP serve as early signs of hyperinflammatory activity, with precise diagnostic relevance for pediatric sepsis, while PCT and IL-10 may help predict early organ dysfunction[11]. TNF-α not only stimulates IL-6 production but also correlates strongly with sepsis severity, both being central to systemic acute inflammatory responses[12]. Despite these insights, the relationship between SIBs and clinical prognoses in sepsis-induced GID is yet to be clarified. This paper therefore intended to clarify these associations and contribute to improved prognostic assessment in such patients.

MATERIALS AND METHODS

General data

The participants were 117 sepsis-induced GID patients admitted to The First Hospital of Hunan University of Chinese Medicine (February 2021 to February 2025). They were grouped into poor [Acute Physiology and Chronic Health Evaluation II (APACHE II) score > 10; n = 76] and good (APACHE II ≤ 10; n = 41) prognosis groups based on outcomes.

Patient enrollment criteria

Inclusion criteria: Confirmed sepsis by referring to the National Institute for Health and Care Excellence guidelines[13]; GID secondary to sepsis; complete laboratory and clinical data; enrollment within the acute disease phase (≤ 2 weeks post-onset); no medication use before admission that could influence results; and aged ≥ 18 years. Exclusion criteria: Renal failure or hospitalization < 7 days; GID caused by gastrointestinal surgery, chronic gastrointestinal disorders, or other etiologies; pregnancy or lactation; acute abdominal conditions (e.g., peritonitis, biliary tract infection); pre-enrollment treatment for sepsis; malignancy or connective tissue disease; hematologic disorders, immunodeficiency, or immunosuppressive therapy; severe psychiatric illness; ventilator dependence.

Analysis indexes

Fasting venous blood draws (4 mL) were collected upon patient admission. Serum was separated through centrifugation and stored at low temperature before enzyme-linked immunosorbent assay-based PCT, CRP, IL-6/10, and TNF-α level quantification. Absorbance (450 nm wavelength) was measured, with concentrations determined using standard curves. All experimental steps strictly followed manufacturer protocols. The APACHE II scale (0-71 points) includes acute physiology, age, and chronic health subscales, with higher scores indicating greater in-hospital mortality risk[14]. The Sequential Organ Failure Assessment (SOFA) score[15] evaluates seven parameters, including cardiovascular performance, platelet levels, and urine output, each graded from 0 to 4 points; higher totals indicate more severe organ dysfunction.

Statistical analysis

Data analyses utilized SPSS 20.0. The n (%) was used to statistically describe categorical data, whose group differences were assessed using χ² tests. All continuous variables were subjected to normality (Kolmogorov-Smirnov test) and homogeneity of variance (Bartlett’s test) testing. Those with homogeneity of variance and an approximate normal distribution were presented in the form of the mean ± standard deviation (SD), with between-group comparisons made with independent t-tests; if non-normally distributed, the continuous measures were shown as the median (interquartile range) [M(Q1, Q3)], and the Mann-Whitney U test was used to identify inter-group differences. We used Pearson’s or Spearman’s correlation coefficients to associate SIBs with APACHE II/SOFA scores. Cox regression identified factors influencing clinical outcomes in sepsis. The SIBs’ performance for predicting sepsis-induced GID was visualized by receiver operating characteristic (ROC) curves, with area under the curve (AUC), confidence interval (CI), sensitivity, and specificity reported. P < 0.05 was deemed statistically significant.

RESULTS

Patient clinical data

Table 1 presents an inter-group comparison of clinical characteristics (poor vs good prognosis). The groups were similar in sex, age, body mass index, smoking history, diabetes, hypertension, and infection site (P > 0.05). However, poor prognosis group was higher than the good prognosis cohort in APACHE II and SOFA scores (P < 0.001).

Table 1 Clinical data of two groups, mean ± SD/n (%).

Indicators	n	Poor prognosis group (n = 76)	Good prognosis group (n = 41)	χ2/t/Z	P value
Sex				0.656	0.418
Male	77	52 (68.42)	25 (60.98)
Female	40	24 (31.58)	16 (39.02)
Age (years)	117	62.58 ± 7.58	60.59 ± 7.63	1.352	0.179
Body mass index (kg/m2)	117	23.00 (20.00, 25.00)	22.00 (20.00, 25.00)	-0.322	0.748
Smoking history				0.158	0.691
No	57	36 (47.37)	21 (51.22)
Yes	60	40 (52.63)	20 (48.78)
Diabetes				0.779	0.377
No	62	38 (50.00)	24 (58.54)
Yes	55	38 (50.00)	17 (41.46)
Hypertension				0.617	0.432
No	57	35 (46.05)	22 (53.66)
Yes	60	41 (53.95)	19 (46.34)
Infection location				2.145	0.342
Respiratory system	96	60 (78.95)	36 (87.80)
Hematologic system	8	7 (9.21)	1 (2.44)
Others	13	9 (11.84)	4 (9.76)
APACHE II (points)	117	20.50 (15.00, 24.00)	15.00 (11.00, 19.00)	-3.676	< 0.001
SOFA (points)	117	8.46 ± 3.30	6.17 ± 3.46	3.521	< 0.001

APACHE II: Acute Physiology and Chronic Health Evaluation II; SOFA: Sequential Organ Failure Assessment.

SIBs

As shown in Figure 1, PCT, CRP, IL-6/10, and TNF-α were elevated in patients with poor prognoses compared with those with good prognoses (P < 0.01).

Figure 1 Serum inflammatory biomarkers. A: Procalcitonin comparison (good vs poor prognosis groups); B: Inter-group comparison of C-reactive protein; C: Interleukin-6 measurements; D: Interleukin-10 disparities; E: Tumor necrosis factor-alpha differences. ^bP < 0.01 vs good prognosis group. PCT: Procalcitonin; CRP: C-reactive protein; IL: Interleukin; TNF-α: Tumor necrosis factor-alpha.

Correlation between SIBs and APACHE II/SOFA scores

Table 2 presents the correlation analyses between SIBs and both APACHE II and SOFA scores. PCT correlated positively with APACHE II (r = 0.379, P < 0.001) and SOFA (r = 0.350, P < 0.001). CRP was linked to APACHE II (r = 0.320, P < 0.001) and SOFA (r = 0.376, P < 0.001). IL-6 and IL-10 were also correlated with APACHE II (r = 0.419 and r = 0.464, respectively; both P < 0.001) and SOFA (r = 0.392, P < 0.001; r = 0.220, P = 0.017). TNF-α exhibited a positive link to APACHE II and SOFA (r = 0.467 and r = 0.383, respectively; both P < 0.001).

Table 2 Correlation between serum inflammatory biomarkers and Acute Physiology and Chronic Health Evaluation II and Sequential Organ Failure Assessment Sequential Organ Failure Assessment.

Indicators	APACHE II	SOFA
PCT	r = 0.379, P < 0.001	r = 0.350, P < 0.001
CRP	r = 320, P < 0.001	r = 0.376, P < 0.001
IL-6	r = 0.419, P < 0.001	r = 0.392, P < 0.001
IL-10	r = 0.464, P < 0.001	r = 0.220, P = 0.017
TNF-α	r = 0.467, P < 0.001	r = 0.383, P < 0.001

APACHE II: Acute Physiology and Chronic Health Evaluation II; SOFA: Sequential Organ Failure Assessment; PCT: Procalcitonin; CRP: C-reactive protein; IL: Interleukin; TNF-α: Tumor necrosis factor-alpha.

Predictive value of SIBs for outcomes (ROC analysis)

We plotted ROC curves to assess these SIBs’ prognosis prediction accuracy, individually or combined. The calculated AUCs were as follows. PCT: 0.661 (95%CI: 0.564-0.759), CRP: 0.672 (95%CI: 0.574-0.771), IL-6: 0.737 (95%CI: 0.648-0.827), IL-10: 0.773 (95%CI: 0.685-0.860), TNF-α: 0.780 (95%CI: 0.697-0.864); combined detection: 0.906 (95%CI: 0.853-0.958). Further analysis revealed that at a PCT threshold of 3.96 ng/mL, specificity was the lowest (40.79%) but sensitivity was the highest (92.68%). Other single indicators had specificity between 55.0% and 68.0% and sensitivities between 70.0% and 86.0%. The combined index performed best, achieving 81.58% specificity, 90.24% sensitivity, and 84.64% overall accuracy (Figure 2, Table 3).

Figure 2 Receiver operating characteristic analysis of serum inflammatory biomarkers’ predictive value for outcomes in sepsis-induced gastrointestinal dysfunction. PCT: Procalcitonin; CRP: C-reactive protein; IL: Interleukin; TNF-α: Tumor necrosis factor-alpha.

Table 3 Prognostic value of serum inflammatory biomarkers in septic patients exhibiting gastrointestinal dysfunction.

Indicators	AUC	95%CI	Cut-off value	Specificity, %	Sensitivity, %	Accuracy, %
PCT (ng/mL)	0.661	0.564-0.759	3.96	40.79	92.68	58.97
CRP (mg/L)	0.672	0.574-0.771	138.0	63.16	70.73	65.81
IL-6 (ng/L)	0.737	0.648-0.827	34.5	57.89	80.49	65.81
IL-10 (pg/mL)	0.773	0.685-0.860	439.5	63.16	85.37	70.94
TNF-α (μg/L)	0.780	0.697-0.864	176.5	67.11	82.93	72.65
Panel	0.906	0.853-0.958	0.68	81.58	90.24	84.62

AUC: Area under the curve; CI: Confidence interval; PCT: Procalcitonin; CRP: C-reactive protein; IL: Interleukin; TNF-α: Tumor necrosis factor-alpha.

Multivariate analysis of clinical outcome predictors in septic patients with GID

Table 4 presents the Cox multivariate analysis of clinical outcome predictors in sepsis-induced GID cases. Clinical outcomes were independently influenced by CRP [odds ratio (OR) = 1.006, 95%CI: 1.001-1.010], TNF-α (OR = 1.006, 95%CI: 1.002-1.010), and SOFA scores (OR = 1.077, 95%CI: 1.011-1.148).

Table 4 Multivariate analysis of clinical outcome predictors in septic patients with gastrointestinal dysfunction.

Indicators	β	SE	Wald	P value	OR	95%CI
PCT (ng/mL)	0.102	0.078	1.699	0.192	1.107	0.950-1.291
CRP (mg/L)	0.006	0.005	5.742	0.017	1.006	1.001-1.010
IL-6 (ng/L)	0.016	0.009	3.357	0.067	1.016	0.999-1.034
IL-10 (pg/mL)	0.003	0.002	2.366	0.124	1.003	0.999-1.007
TNF-α (μg/L)	0.006	0.002	9.603	0.002	1.006	1.002-1.010
APACHE II (points)	0.027	0.022	1.516	0.218	1.028	0.984-1.074
SOFA (points)	0.074	0.032	5.214	0.022	1.077	1.011-1.148

PCT: Procalcitonin; CRP: C-reactive protein; IL: Interleukin; TNF-α: Tumor necrosis factor-alpha; APACHE II: Acute Physiology and Chronic Health Evaluation II; SOFA: Sequential Organ Failure Assessment; OR: Odds ratio; CI: Confidence interval.

DISCUSSION

Sepsis represents a major global health concern involving complex interactions among immune and endothelial cells, cytokines, complements, and the coagulation system[16]. GID, a frequent complication of sepsis, impairs nutrient absorption, immune function, and prognosis[17]. This study assessed how specific SIBs, PCT, CRP, IL-6/10, and TNF-α, influence outcomes in sepsis-associated GID. Clinical outcomes were unaffected by sex, age, body mass index, smoking status, diabetes, hypertension, or infection site but were strongly associated with APACHE II and SOFA scores, consistent with Li et al’s research results[18]. These indices have been well-validated for predicting mortality and disease severity[19,20]. Additionally, SIBs were markedly elevated in sepsis + GID patients with poor prognoses, consistent with Guo et al[21], who demonstrated elevated IL-6, PCT, and CRP levels in rat models of sepsis-induced lung injury. Similarly, Zhang et al[22] observed increased serum and hippocampal TNF-α and IL-6/10 concentrations in septic rats. Notably, although IL-10 is conventionally anti-inflammatory, higher levels were found in patients with adverse gastrointestinal outcomes, possibly reflecting the AA homozygous genotype that heightens susceptibility to gram-negative infection[23].

SIBs showed significant positive correlations with APACHE II and SOFA scores, indicating potential predictive value. Individually, their AUCs ranged from 0.660 to 0.780, whereas their combined assessment achieved superior performance (AUC = 0.906, specificity 81.58%, sensitivity 90.24%, and accuracy 84.62%). Zhou et al[24] likewise found that the combined detection of PCT, IL-6, high-sensitivity CRP, and presepsin optimized mortality prediction in ICU-hospitalized septic patients. Yan et al[25] reported a comparable IL-6 AUC of 0.704 for ICU outcomes, close to the 0.737 observed in this study for poor prognosis in sepsis-related GID. In hematology/oncology cohorts, Lin et al[26] found IL-10 (AUC = 0.757) predictive of bloodstream infection and PCT (AUC = 0.620) predictive of mortality, mirroring the outcomes observed in our investigation. Finally, the Cox multivariate analysis confirmed CRP (OR = 1.006, 95%CI: 1.001-1.010), TNF-α (OR = 1.006, 95%CI: 1.002-1.010), and SOFA (OR = 1.077, 95%CI: 1.011-1.148) as independent prognostic indicators in sepsis-induced GID.

This study has several limitations: First, no cost analysis was conducted for using various SIBs to predict poor prognosis in sepsis-induced GID; future supplementation of such data could enhance clinical applicability. Second, dynamic longitudinal data on SIBs were unavailable, preventing the evaluation of peak concentrations or decline slopes; incorporating these analyses would better reflect inflammation intensity and recovery rate. Third, the absence of external validation limits generalizability, highlighting the need for further investigations to refine CRP-, TNF-α-, and SOFA-based prognostic risk stratification and strengthen clinical reliability.

CONCLUSION

Patients with sepsis-induced GID and poor outcomes exhibit significantly PCT, CRP, IL-6/10, and TNF-α up-regulation, which increase with disease severity. Combined evaluation of these biomarkers enables robust predictive value for clinical outcomes, while elevated CRP, TNF-α, and SOFA scores significantly increase the risk of adverse prognosis.