Caecal ligation and puncture in the minipig – a model of sepsis induction

Summary

Background:
Sepsis belongs among the most serious conditions and animal models of sepsis are the basic tools to investigate the pathophysiological response to this condition.

Material and methods:
A total of 16 adult minipigs with identical baseline parameters were randomized into two groups. In the sepsis group (n = 10), sepsis was induced using caecal ligation and puncture (CLP). The control group (n = 6) underwent laparotomy without CLP. Selected clinical and laboratory parameters as well as histological findings between the sepsis and control group were subsequently compared. Results: All animals undergoing CLP developed diffuse peritonitis and sepsis. Compared to the control group, experimental animals showed significant increase of body temperature and heart rate (while) requiring noradrenaline to maintain their perfusion pressure. No significant differences in the monitored biochemical parameters (including C-reactive protein levels) between the two groups were found. Histological findings in organs of experimental animals were consistent with changes of organs seen in sepsis, i.e., centrilobular liver necroses, acute tubular renal necrosis, serous fibrinopurulent exudate, myocardial malacias, and pulmonary edema.

Conclusion:
Experimental caecal ligation with a predefined size of the perforation in the intestinal wall is a suitable model for assessing the pathophysiological changes occurring in the body in sepsis.

Key words:
sepsis, caecal ligation, experimental sepsis

Acknowledgment:


Supported by Research Project IGA NS10556-3/2009 of the Czech Ministry of Health.

Sepsis, severe sepsis, and septic shock pose a serious diagnostic and therapeutic challenge at the intensive care unit. Sepsis ranks second as the most common non-coronary causes of death at the intensive care unit, with its incidence increasing each year over the past two decades (1). Intra-abdominal sepsis belongs among the most serious conditions encountered in the department of surgery. A crucial role in the persistently high morbidity and mortality rates of these conditions is played by their late diagnosis and management. In terms of the body´s response, septic shock is characterized by immune deregulation with the body-triggered mediator ”torment“ leading to microvascular dysfunction and, eventually, to organ failure and death (2). The body´s mediator response is still poorly understood, which is why sepsis continues to be an attractive topic of basic and clinical research. Animal models of intra-abdominal sepsis belong among the basic tools allowing to monitor, step-by-step and in a reproducible manner, the development of intra-abdominal inflammation. Three models of sepsis are currently used: an exogenous toxin model, exogenous bacterial infection model, and host barrier disruption model. Our paper presents our own data obtained in a model of sepsis using a large laboratory animal (minipig), with sepsis induced by caecal ligation and perforation.

The study was approved by the Committee for the Prevention of Cruelty to Animals at the Institute for Clinical and Experimental Medicine (IKEM). Experiments were performed in the accredited Center for Experimental Medicine of IKEM in compliance with applicable European law and that of the Czech Republic regarding the keeping and use of experimental animals.

Methods

Animals A total of 19 adult minipigs with a mean weight of 39.5 kg (median 37 kg) were used in the experiment. Two animals were used in the pilot phase of the experiment. In the actual experiment, the animals were randomized into two groups. In the sepsis group (n = 10), sepsis was induced by caecal ligation and puncture (CLP). The control group (n = 6) had laparotomy without CLP (sham surgery) performed. One animal was excluded at the beginning of the experiment because of signs of lung infection.

Selected clinical and laboratory parameters as well as histological findings in both groups were compared.

Pilot study

As successful use of a functional model requires experience with the chosen technique appropriate to the complexity of the procedure, the technique of the procedure including monitoring of the selected parameters was tested in two animals as a pilot experiment. The purpose of the preparatory phase was to test the efficacy of the surgical technique of CLP in the development of the septic state. Two variants of the CLP model were employed, with each resulting in the development of peritonitis and septic response. An alternative procedure involving caecal puncture using a trocar was chosen to ensure simple and exact model reproducibility.

Technique Under general anesthesia, CLP was used to surgically induce stercoral peritonitis. This was achieved by midline laparotomy and caecal ligation at one fourth of its length (5 cm from Bauhin´s valve). Blood vascular supply was discontinued along the length of caecal ligation. The intestinal content from the ascending colon was massaged into the caecum prior to its ligation. Using a trocar, point perforation of the caecum was performed on the antimesenterial site in the tenial region, with the incision subsequently extended to a total of 3 cm (Figure 1). The abdominal cavity was contaminated with the intestinal content (Figure 2). This was followed by gastrostomy and cystostomy. The control (sham) group had only lapatoromy, gastrostomy, and cystostomy performed.

Anesthesia The animals were fasted for 12 hours preoperatively, receiving only liquids. On the day of surgery, the following premedication was administered i.m.: ketamine 10 mg/kg (Narketan, Vetoquinol), azaperone 5 mg/kg (Stresnil, Janssen), and atropine 0.1 mg/kg (Atropin, Biotica). Anesthesia was induced by i.v. propofol at a dose of 2 mg/kg (Propofol 1%, Fresenius) and fentanyl at a dose of 100 μg (Fentanil, Torrex) into to the auricular vein. The animals were intubated and mechanically ventilated (Servoventilator, Siemens, Sweden, FiO2 0.4, PEEP 5 cmH20, respiratory volume 10 ml/kg). Ventilation was set with respect to target arterial PCO2 in the range of 4.0–5.0 kPa. Relaxation was obtained by vecuronium (Norcuron, Organon) in a starting dose of 0.1 mg/kg, and continued at a rate of 0.8 µg/kg/min throughout the procedure. Anesthesia was maintained with fentanyl (10–15 µg/kg/h) and propofol (6–10 mg/kg/h). An 18 G catheter (Braun) was inserted into the femoral artery for blood pressure monitoring and blood sampling, a 7 Fr three-way central venous catheter (Arrow) into the left jugular vein for infusion and drug administration, and a 7 Fr thermodilution catheter (Arrow) into the pulmonary artery via the right jugular vein for the measurement of hemodynamic parameters and body temperature. Fluid loss was compensated with an infusion solution at a rate of 10–15 ml/kg/h (Plasmalyte, Baxter).

Postoperative care

Upon closure of the laparotomy, all animals remained mechanically ventilated and analgosedated with continuous infusion of thiopental (5 mg/kg/h) and fentanyl (5–10 μg/kg/h) throughout the experiment. Infusion solutions (Plasmalyte 10 ml/kg/h and 6% hydroxyethylstarch 130 kD/0.4, Voluven, Fresenius in bolus form) were administered continuously. In animals, where the mean arterial pressure (MAP) dropped below 65 mmHg not responding to volume therapy, noradrenaline (Noradrenalin, Zentiva) was administered in continuous infusion. Body temperature was controlled by cooling and warming with target values of 37–39°C. Glycemia was maintained within the range of 4.0 to 7.0 mmol/l with infusion of 20% glucose.

In experimental animals, the procedure was stopped upon the onset of septic shock, at least 24 hours postoperatively. In the control group the experiment was stopped 24 hours postoperatively. The experiment was completed by surgical revision of the abdominal cavity and collection of tissue samples. The animals were subsequently sacrificed using a bolus of an anesthetic.

Monitoring The following selected parameters were monitored (Marquette, USA): heart electrical activity (electrocardiogram), mean arterial pressure (MAP), central venous pressure (CVP), body temperature, and oxygen saturation in peripheral blood. Cardiac output (CO) and PC wedged pressure (PCWP) were determined, and cardiac index (CI) and systemic vascular resistance index (SVRI) calculated. Measurements were performed at the beginning of the experiment (time T0), at the end of the surgery (time T1), 8 hours (T8), 16 hours (T16), and 24 hours (T24) postoperatively.

Collection and analysis of blood samples, plasma samples and histological material At the beginning of the experiment, and at 6-hour intervals postoperatively, arterial blood samples were obtained for biochemistry (urea, creatinine, AST, ALT, bilirubin, ALP, GMT),  leukocyte count and prothromobin time. Upon anesthesia induction, 1 hour, and every 3 hours after the surgery arterial blood samples were obtained to determine plasma C-reactive protein levels [CRP; TA 901 (Tridelta Development Ltd.) – porcine C-reactive protein assay]. Additional blood samples were obtained at the end of the experiment for microbiological assays.

Histological examination

Prior to the completion of the experiment laparotomy and surgical revision of the abdominal cavity was performed, and biopsy samples were collected from the liver, kidney, spleen, and colon. Additional biopsy samples were also obtained from the heart and lungs. Tissues were fixed in 10% buffered formol and embedded in paraffin for histology. The samples embedded in paraffin were cut into 3-4 µm sections and stained with haematoxylin-eosin, Sirius red with elastica, PAS and Gramm stains. The sections were observed with Olympus BX 41 microscope.

Statistical analysis

Data were processed using Statistica version 8 software (StatSoft). Numerical values are given as mean ± standard deviation. The non-parametric Mann-Whitney test was used to compare the data of both groups.

Results

At the beginning of the experiment, no significant differences between the experimental and control animals in the monitored clinical and laboratory parameters were documented, nor did differ histological findings consistent with the pattern of normal tissues.

In the postoperative period, the control group was hemodynamically stable and, compared to the sepsis group, showed significantly lower consumption of colloidal solutions (Graf 1). (Note: As no differences were seen between the values measured at times T0 and T1 in individual animals, Figures 1–4 show only T1 value). No control animal required noradrenaline administration to maintain perfusion pressure. By contrast, all animals with induced intraabdominal sepsis by means of CLP developed, whitin 15-36 hours postoperatively, septic shock with a febrile response, tachycardia, and hypotension requiring continuous noradrenaline infusion to maintain MAP above 65 mmHg. The dose of noradrenaline in the septic animals was on average 0.7±0.6 μg/kg/min.

Compared to control group animals, heart rate of septic animals rose significantly during the experiment (Graf 2).

The differences in CI and SVRI did not reach statistical significance at individual time points (Graf 3).

A significant difference in body temperature between both groups was demonstrated in the postoperative course. While mean body temperature of the control group animals did not change significantly, it rose to 40.8±0.96ºC in the sepsis group (Graf 4).

The serum levels of ALT, AST, bilirubin, GMT, ALP, urea, and creatinine or prothromobin time did not differ significantly between the two groups throughout the experiment. No significant difference between both groups was shown in the levels of lactate in arterial blood. While the baseline mean lactate levels were 2.4±1.9 mmol/l in the control group, they were 3.6±1.5 mmol/l in the sepsis group (p=0.44). At the end of the experiment, the mean levels of lactate were 1.0±0.2 mmol/l and 1.9±0.7 mmol/l in the control and sepsis group animals, respectively (p=0.49). Comparing the two groups, the difference in the change of lactate levels at the beginning and end of experiment did not reach statistical significance (p=0.62).

Baseline leukocyte counts were 9.2±4.9 × 109/l and 8.4±5.1 × 109/l in the control and sepsis groups, respectively; the difference was not significant (p=0.32). While leukocyte count in the control group remained unchanged over time (7.3±3.7 × 109/l), septic animals developed leukopenia (3.5±3.2 × 109/l); however, the difference between the two groups was not significant at the end of the experiment (p=0.69).

No significant differences were found between both groups in the difference of CRP levels (∆ CRP) at the beginning and end of the experiment (p= 0.67) (Graf 5).

A mixture of intestinal bacteria (Escherichia coli, Enterococcus spp., Pseudomonas aeruginosa, Klebsiella pneumoniae, Burgholderia cepacia, Acinetobacter spp., Enterobacter cloacae, staphylococci) was documented in the blood cultures of sepsis group animals..

At the end of the experiment, all minipigs with CLP – unlike those in the control group – were shown to develop peritonitis with the presence of seropurulent secretion in the abdominal cavity, and signs of ischemic necrosis in the caecum (Obr. 3).

Histological pictures of tissue samples collected from organs were physiological in both groups at the beginning of the experiment. Changes consistent with the effect of sepsis, centrilobular liver necrosis, acute tubular necrosis and increased renal perfusion, edema, and serous fibrinopurulent exudate, with myocardial myomalatias and pulmonary edema were documented in all samples obtained from septic animals, as compared to the control group at the end of the experiment (Obr. 4).

Discussion

Sepsis continues to be one of the most serious acute conditions in medicine. Late diagnosis markedly deteriorates the outcome. Recognition of the dynamics of pathophysiological alterations is the only possibility for making early diagnosis with subsequent targeted therapy. The research of these changes in sepsis in human medicine is limited for many methodological reasons. These limits can be eliminated by use of animal models of sepsis, where the septic state is homogeneous and better reproducible. On the other hand, it cannot be exactly translated into human medicine (3).

At present, there are three experimental animal models of sepsis, i.e., an exogenous toxin model, exogenous bacterial infection model, and a model of host barrier disruption. However, none of these models is optimal, with each having its advantages and drawbacks compared to the natural course of sepsis.

Endogenous toxin model

In 1940, Andre Boivin was the first to isolate endotoxin from Gram-negative bacteria. Later, Bordon and Hall incriminated endotoxin as the factor inducing human septic shock (4). At present, lipopolysaccharide (LPS) is the endotoxin used most commonly to induce experimental sepsis.

The model of endotoxin-induced sepsis is relatively simple and easy to monitor. On the other hand, it is not fully identical with the natural course of sepsis.

After LPS administration, the hemodynamic changes occurring in men are not truly identical with those seen in the animal model of sepsis (5). The hemodynamic changes in the endotoxic model are not fully identical with those seen in septic patients. Most importantly, the hyperdynamic and hypodynamic phases are not clearly expressed (6).

Exogenous infection model

Infection caused by exogenous bacteria has a number of disadvantages, the main one is the exact definition of the lethal dose of bacteria. Significant differences have been reported in the mediator response of the body depending on the bacterial load (7). In addition, the cytokine response is modified by the infected compartment (vascular bed-peritoneal cavity). As a result, the model is currently used particularly to study the host response to a clearly identified pathogen in the given body compartment.

Model of host barrier disruption

This model uses the host physiological mixed bacterial flora as the infectious insult in artificially induced disruption of the host barrier. The most common models are the caecal ligation with incision (CLP) and a model involving a stent inserted into the ascending colon (CASP).

Model of caecal ligation

This model was first described by Irshad Chaudry in 1980 in a publication by Wichtermann (8) as a model of caecal ligation without intestinal wall injury in the mouse. The model was subsequently complemented with intestinal wall puncture to allow bacteria to enter the peritoneal cavity through the hole in the intestinal wall. This model mimics the clinical course of perforated appendicitis or diverculitis, and was regarded as the „gold standard“ for the study of early sepsis. The advantages of this model include simple reproducibility, relatively low costs, and the possibility to be used on both small and large animals. Disadvantages of this technique (of inducing sepsis) include the ability of the body to identify the focus of infection resulting in the formation of abscess focus without the development of  generalized septic state.

CASP model (stent insertion into the ascending colon)

According to literary reports, and compared to the model described above, the infectious content continues to escape into the peritoneal cavity via the stent inserted into the ascending colon, and there is a minimal possibility for an abscess focus to form (9).

In our experiment, the model of caecal ligation with a predefined size of the artificial perforation in the caecum was used. We chose this model as a condition most identical with the pathological states in human medicine (appendicitis perforation, diverculitis perforation). The puncture-induced impairment of the integrity of the intestinal wall was replaced by caecal perforation with a trocar measuring 12 mm in diameter, thereby allowing for standardization of the perforation perforation size.

Among the monitored clinical parameters of sepsis, there was no difference in cardiac output and systemic vascular resistance between the two groups; every effort was made to maintain MAP and cardiac output, and vascular resistance within the target values by administration of colloidal solution and noradrenaline as per protocol.

Despite the detectable development of sepsis, no significant elevation of lactate levels in arterial blood was seen, perhaps because of early fluid resuscitation of the sepsis animals.

During the experiment, no significant changes in the biochemical parameters of septic animals were seen, nor was there a significant increase of CRP levels in these animal. An explanation for this could be that the duration of the experiment (24 hours) was not sufficient for these changes to occur. Procalcitonin has been repeatedly shown to be superior to CRP as a predictor of bacterial sepsis. However, specific diagnostic kits for determination of procalcitonin levels in the pig are not available yet (10).

Conclusion

Our experiment has shown that the technique of caecal ligation with predefined size of the opening in the intestinal wall is a reproducible model of septic state. All experimental animals with CPL developed diffuse peritonitis, without any signs of abscess focus formation, and gradual development of sepsis. We conclude, that this model is optimal for future research of the development of the septic state (in experimental animals).