A Porcine Sepsis Model With Numerical Scoring for Early Prediction of Severity

Abstract

Introduction: Sepsis can lead to organ dysfunctions with disturbed oxygen dynamics and life-threatening consequences. Since the results of organ-protective treatments cannot always be transferred from laboratory models into human therapies, increasing the translational potential of preclinical settings is an important goal. Our aim was to develop a standardized research protocol, where the progression of sepsis-related events can be characterized reproducibly in model experiments within clinically-relevant time frames.

Methods: Peritonitis was induced in anesthetized minipigs injected intraperitoneally with autofeces inoculum (n = 27) or with saline (sham operation; n = 9). The microbial colony-forming units (CFUs) in the inoculum were retrospectively determined. After awakening, clinically relevant supportive therapies were conducted. Nineteen inoculated animals developed sepsis without a fulminant reaction. Sixteen hours later, these animals were re-anesthetized for invasive monitoring. Blood samples were taken to detect plasma TNF-α, IL-10, big endothelin (bET), high mobility group box protein1 (HMGB1) levels and blood gases, and sublingual microcirculatory measurements were conducted. Hemodynamic, respiratory, coagulation, liver and kidney dysfunctions were detected to characterize the septic status with a pig-specific Sequential Organ Failure Assessment (pSOFA) score and its simplified version (respiratory, cardiovascular and renal failure) between 16 and 24 h of the experiments.

Results: Despite the standardized sepsis induction, the animals could be clustered into two distinct levels of severity: a sepsis (n = 10; median pSOFA score = 2) and a septic shock (n = 9; median pSOFA score = 8) subgroup at 18 h of the experiments, when the decreased systemic vascular resistance, increased DO₂ and VO₂, and markedly increased ExO₂ demonstrated a compensated hyperdynamic state. Septic animals showed severity-dependent scores for organ failure with reduced microcirculation despite the adequate oxygen dynamics. Sepsis severity characterized later with pSOFA scores was in correlation with the germ count in the induction inoculum (r = 0.664) and CFUs in hemocultures (r = 0.876). Early changes in plasma levels of TNF-α, bET and HMGB1 were all related to the late-onset organ dysfunctions characterized by pSOFA scores.

Conclusions: This microbiologically-monitored, large animal model of intraabdominal sepsis is suitable for clinically-relevant investigations. The methodology combines the advantages of conscious and anesthetized studies, and mimics human sepsis and septic shock closely with the possibility of numerical quantification of host responses.

Keywords: SOFA score; fecal peritonitis; inflammatory markers; organ dysfunction; pig model; sepsis; septic shock.

Figure 1

Experimental stages (A), experimental protocol, groups, interventions and assessments (B). (A) Experimental stages: (1) baseline measurements (under temporary anesthesia); (2) sepsis induction with intraperitoneal (ip) fecal inoculum; (3) sepsis progression (in awake animals); and (4) re-anesthesia, instrumentation, monitoring and scoring (B). Groups: The animals were randomly assigned into sham-operated and septic groups. Based on disease severity and vasopressor requirement (at 18 h after sepsis induction), animals in the septic group were allocated into sepsis and septic shock subgroups.

Figure 2

Average values for basic pSOFA scores in individual animals during the 8-h invasive monitoring period (during t = 16–24 h) (A) in the sham-operated group (open circles) and the sepsis (gray square) and septic shock subgroups (black triangle). Changes in 3D-pSOFA (B) and 5D-pSOFA (C) values as a function of time in the sham-operated group and the sepsis and septic shock subgroups. The plots demonstrate the median values and the 25th (lower whisker) and 75th (upper whisker) percentiles. ^XP < 0.05 vs. sham-operated group; ^#P < 0.05 vs. sepsis group; *P < 0.05 vs. t = 16 h.

Figure 3

Changes in mean arterial pressure (A), urine output (B), PaO₂/FiO₂ ratio (C), platelet count (D), plasma bilirubin (E), De Ritis (AST/ALT) ratio (F), plasma creatinine (G), and plasma albumin levels (H) in the sham-operated group (open circles) and the sepsis (gray square) and septic shock subgroups (black triangle). The plots demonstrate the median values and the 25th (lower whisker) and 75th (upper whisker) percentiles. ^XP < 0.05 vs. sham-operated group; ^#P < 0.05 vs. sepsis group; (A–D) *P < 0.05 vs. t = 16 h; (E–H) *P < 0.05 vs. t = 0 h.

Figure 4

Changes in white blood cell count (A) and whole blood lactate levels (B) in the sham-operated group (open circles) and the sepsis (gray square) and septic shock subgroups (black triangle). The plots demonstrate the median values and the 25th (lower whisker) and 75th (upper whisker) percentiles. ^XP < 0.05 vs. sham-operated group; ^#P < 0.05 vs. sepsis group; *P < 0.05 vs. t = 0 h.

Figure 5

Changes in heart rate (A), cardiac index (B), stroke volume index (SVI) (C), and systemic vascular resistance index (SVRI) (D) in the sham-operated group (open circles) and the sepsis (gray square) and septic shock subgroups (black triangle). The plots demonstrate the median values and the 25th (lower whisker) and 75th (upper whisker) percentiles. ^XP < 0.05 vs. sham-operated group; *P < 0.05 vs. t = 16 h.

Figure 6

Changes in oxygen delivery (DO₂) (A), oxygen consumption (VO₂) (B), oxygen extraction (Ex O₂) (C), and the proportion of perfused vessels (PPV) (D) in the sham-operated group (open circles) and the sepsis (gray square) and septic shock subgroups (black triangle). The plots demonstrate the median values and the 25th (lower whisker) and 75th (upper whisker) percentiles. ^XP < 0.05 vs. sham-operated group; *P < 0.05 vs. t = 16 h.

Figure 7

Correlation between the concentration of injected microorganism in feces and the pSOFA score values (A). Correlation between the microbial concentration in the hemoculture and the pSOFA score values (B).

Figure 8

Changes in plasma tumor necrosis factor alpha (TNF-α) (A), interleukin-10 (IL-10) (B), big endothelin (bET) (C), and high-mobility group box 1 protein (HMGB1) (D) levels in the sham-operated group (open circles) and in the sepsis (gray square) and septic shock subgroups (black triangle). The plots demonstrate the median values and the 25th (lower whisker) and 75th (upper whisker) percentiles. ^XP < 0.05 vs. sham-operated group; ^#P < 0.05 vs. sepsis group; *P < 0.05 t = 0 h.

Figure 9

Correlations between plasma concentrations of TNF-α at 6t = h (A), interleukin-10 (IL-10) (B), big endothelin (bET) (C), high-mobility group box 1 protein (HMGB1) at t = 16 h (D), and 3D-pSOFA and 5D-pSOFA scores at t = 24 h.