Perioperative Low-Dose Prednisolone Treatment Has Beneficial Effects on Postoperative Recovery and Anastomotic Healing in a Murine Colitis Model

Abstract

Background and Aims

High-dose glucocorticoid treatment has been identified as a risk factor for anastomotic leakage in patients with inflammatory bowel disease [IBD] undergoing bowel resection surgery. By contrast, active disease during surgery is also associated with elevated morbidity. Perioperative low-dose treatment might be beneficial regarding postoperative outcomes by controlling disease activity. The present study is the first to investigate the dose-dependent effect of perioperative prednisolone therapy in a murine IBD model combining dextran sodium sulphate [DSS] colitis with intestinal anastomosis surgery.

Methods

In 84 10-week-old wild-type mice, a colorectal anastomosis was performed using a microsurgical technique. Half the animals received induction of chemical colitis with 2% DSS via drinking water prior to surgery. In both groups, one-third of the animals received daily oral administration of high-dose [0.533 mg/kg] and one-third low-dose [0.133 mg/kg] prednisolone. Evaluation was performed on postoperative days 3 and 7.

Results

While high-dose prednisolone treatment led to an increased anastomotic leakage rate in mice under colitis, low-dose prednisolone treatment limited preoperative disease activity and did not influence the leakage rate. Histological examination showed a beneficial effect of low-dose prednisolone treatment on microscopic abscess formation at the anastomotic site in DSS mice as well as an increased anastomotic healing score.

Conclusions

We demonstrate a beneficial effect of perioperative short-term low-dose prednisolone treatment on intestinal anastomotic healing in the context of colitis. Perioperative use of short-term low-dose prednisolone treatment might be beneficial in IBD patients who need to undergo surgery during active disease.

1. Introduction

Despite significant innovations in medical therapy for inflammatory bowel disease [IBD], a large proportion of patients still require bowel resection due to disease-related complications, such as fibrosis-associated intestinal stenosis, interenteric fistula, intestinal perforation, colitis-associated colorectal cancer or resistance to medical therapy.^1,2 There is evidence that patients with Crohn’s disease are at higher risk for developing anastomotic leakage.³ Data from cohort studies further suggest that preoperative steroid therapy in IBD patients increases the rate of postoperative intra-abdominal septic complications [IASCs] including anastomotic leakage and intra-abdominal abscess formation. ^4-7 IASCs not only lead to general postoperative morbidity and high rates of relaparotomy and other interventions but also increase the risk of postoperative recurrence after ileocecal resections for Crohn’s disease.⁸

Disease complications in IBD patients requiring surgical intervention often arise in phases of acute disease flare-ups that require glucocorticoid treatment. As high daily doses of prednisolone have been established as a risk factor for postoperative IASCs and anastomotic leakage following bowel resection surgeries in IBD patients,^5,9 it has been recommended to reduce the preoperative dosage of prednisolone as low as possible while still controlling colitis activity.^10,11 A dose reduction below 10–20 mg daily has been recommended by national and international guidelines and is widely accepted as relatively safe by colorectal IBD surgeons¹².

However, no randomized controlled trial has examined the effect of perioperative prednisolone treatment in IBD patients on postoperative IASCs and anastomotic leakage or anastomotic healing in general.¹³ As there is no routinely used treatment option so far to reduce the risk of anastomotic leakage in patients undergoing bowel resection surgery except for diverting ileostomy,¹⁴ it is of great importance to optimize perioperative medical therapy, especially in IBD patients. Regarding the management of preoperative glucocorticoid treatment in IBD patients undergoing bowel resection surgery, it remains uncertain whether steroid therapy should only be reduced or completely discontinued prior to surgery.^13,15

We have recently developed a murine IBD surgery model combining experimental dextran sodium sulphate [DSS] colitis with a colorectal anastomosis to investigate the effects of perioperative medical therapy on postoperative anastomotic complications.^16,17 Using this murine model, we aimed to investigate whether perioperative short-term low-dose glucocorticoid treatment might be beneficial regarding anastomotic healing in the context of colitis. The present study is thus the first to investigate the dose-dependent effect of perioperative prednisolone therapy in a murine IBD model combining experimental colitis with intestinal anastomosis surgery.

2. Materials and Methods

2.1 Ethical considerations

All animal experiments were approved by the local animal welfare committee of the administration of Upper Bavaria [No. 55.2-2532.Vet_02-17-203]. Reporting of in vivo experiments was performed according to the ARRIVE guidelines.¹⁸

2.2 Murine experiments

To evaluate the effect of perioperative prednisolone treatment on intestinal anastomotic healing, we used a murine colitis model in combination with abdominal surgery and colorectal anastomosis formation. To model an acute disease flare-up, we used the acute DSS colitis model by administering 2% DSS via drinking water for 7 days prior to surgery. Perioperative prednisolone therapy was adapted to commonly used dosages in human patients. To evaluate the dose-dependent effect of prednisolone on postoperative recovery and anastomotic healing after intestinal anastomosis surgery, we chose a dose equivalent of 40 mg per day in a 75-kg patient [0.53 mg/kg body weight] and a dose equivalent of 10 mg per day in a 75-kg patient [0.13 mg/kg body weight] [Figure 1A]. The anastomosis was then performed in the descending colon by sparing the arterial blood supply and ensuring the integrity of the anastomosis by using 12–13 single stitches, which led to sufficient closure of the anastomotic line in all 84 animals [Figure 1B].

All murine experiments were performed using female wild-type BALB/c mice, aged 9–11 weeks and with a starting body weight of 16–21 g [mean 19.70 g, SD 1.15 g]. Mice were bought from Charles River. Mice were kept at the local specific pathogen-free animal facility with standardized housing conditions including a 12-h light/dark cycle at 22°C, 45–60% humidity and free access to water and standard chow. Environmental enrichment was achieved by providing a nesting house, nesting material and scattering food in the bedding. Before starting the experiments, mice were randomly assigned to their cages and were acclimatized to the local conditions for at least 1 week. Mice were monitored daily for weight [score 0–4], appearance of the fur [0–1], behaviour [0–4], posture [0–4], pain during abdominal palpation [0–2], appearance of the surgical wound [0–4], signs of dehydration [0–1] and stool consistency [0–1]. A priori exclusion criteria were defined as a single score of 4 or a combined score of >8 points. All animals were kept in the same housing area and were handled by the same animal keepers and researchers to avoid potential confounders. All animals underwent intestinal anastomosis surgery as described in the following. Evaluation time-points were postoperative day [POD] 3 and POD7. For evaluation of the anastomosis, endoscopy was performed under general isoflurane inhalation anaesthesia before killing the animals under general anaesthesia by cervical dislocation with preceding cardiac puncture for blood sampling.

2.3 Study design

A total of 84 mice were used in this study, 14 mice in each experimental group [seven mice per group for a total of two evaluation time-points]. Group sizes were calculated using the least number of animals to achieve statistically significant results based on our previous studies using the histological anastomotic healing score as the primary outcome as described in the following sections. The following experimental groups were defined: CTL [control], CTL + LD [control plus low-dose prednisolone treatment], CTL + HD [control plus high-dose prednisolone treatment], DSS [dextran sodium sulphate colitis], DSS + LD [DSS colitis plus low-dose prednisolone treatment] and DSS + HD [DSS colitis plus high-dose prednisolone treatment] [Figure 1A].

2.4 Induction of acute DSS colitis

DSS colitis was induced in the DSS, DSS + LD and DSS + HD experimental groups. These mice were exposed to 2% [w/v] DSS [colitis grade, molecular weight 36 000–50 000 Da; MP Biomedicals] via drinking water for 7 days prior to surgery as previously described.¹⁶ The pathomechanism of DSS colitis is based on epithelial damage in the distal colon, which compromises the epithelial barrier function and consequently leads to colitis.¹⁹ This colitis model was chosen as it is precisely controllable, which is especially important in combination with the surgical procedure of intestinal anastomosis formation. Disease activity was evaluated using the disease activity index [DAI] that includes evaluation of weight loss, blood in stool and stool consistency [see Supplementary Table 1].

2.5 Prednisolone treatment

The following experimental groups were exposed to perioperative prednisolone treatment: CTL + LD, CTL + HD, DSS + LD and DSS + HD. The prednisolone derivate Okrido [1 mL ≙ 6 mg prednisolone as prednisolone dihydrogenphosphate disodium; Pharmapol Arzneimittelvertrieb] was applied via oral gavage daily for 7 days preoperatively until POD3 or POD7 respectively. Mice without prednisolone treatment received a daily gavage of placebo [water]. Prednisolone doses were set to 0.13 mg/kg body weight daily for low-dose treatment [equivalent to 10 mg prednisolone daily in a 75-kg human patient] and to 0.53 mg/kg body weight daily for high-dose treatment [equivalent to 40 mg prednisolone daily in a 75-kg human patient]. The last dose of prednisolone was applied prior to final evaluation at POD3 or POD7. Prednisolone serum levels were measured from serum samples at POD3 or POD7 using liquid chromatography with tandem mass spectrometry [LC-MS/MS]. The chromatographic separation was carried out on a reversed-phase chromatography [RP] column with a solvent gradient. Prednisolone was measured in Multiple Reaction Monitoring Mode [MRM] with specific transitions.

2.6 Surgical procedure

Surgery was performed under general inhalation anaesthesia using isoflurane [CP Pharma]. Perioperative analgesia was managed using meloxicam [0.1 mg/kg body weight per dose; Boehringer Ingelheim] and buprenorphine [0.01 mg/kg per dose; Indivior]. During surgery, body temperature was preserved at 37°C using a heated surgery platform. The intestinal anastomosis surgeries were performed in microsurgical technique by using an operating microscope [Carl Zeiss Meditec] as previously described.¹⁶ In summary, after laparotomy, the colon was exposed and transected at the level below the left kidney above the main vasculature to prevent ischaemia. An end-to-end anastomosis was then performed using 12–13 single stitches [Ethilon, BV-2, USP 9-0, #2809; Ethicon] [Figure 1B]. After anastomosis formation the peritoneum and the abdominal wall were closed in continuous suture technique [Prolene, P3, USP 6-0, #8695H; Ethicon]. After surgery, mice were monitored closely as described above. Surgeries were performed by two surgeons in alternation with each surgeon operating on the same number of mice per experimental group.

2.7 Endoscopy

Colonoscopy was performed under general isoflurane inhalation anaesthesia using a 1.9-mm telescope [#64301 AA 0° telescope with #61029 C protection sheath and H3-Z 22220055 camera head; all Karl Storz] prior to final evaluation at POD3 or POD7 respectively. The anastomosis was evaluated endoscopically using an endoscopic healing score ranging from 0 to 4 points with 0 points indicating a completely smooth anastomotic line and 4 points indicating a primary leaking anastomosis [see Supplementary Table 2 and Figure 2E].

2.8 Bursting pressure measurement

Bursting pressure was measured after dissecting the descending colon including the anastomosis from the mouse using the MPR 1 Datalogger with an Omnibar E5F precision pressure catheter [both Raumedic]. The pressure probe was inserted into the rectum while a plastic canula was inserted into the oral end of the colon. Both were fixed with a polyfilament suture. After calibration, the lumen was manually filled with isotonic saline solution until bursting of the anastomosis or the adjacent colon. The maximum pressure was defined as the bursting pressure.

2.9 Macroscopic anastomosis evaluation

The anastomosis was evaluated macroscopically with regard to criteria for anastomotic leakage as well as the number and severity of adhesions around the anastomosis using an adhesion score [Figure 2D]. Criteria for anastomotic leakage were [i] visible dehiscence of the anastomosis and [ii] abscesses around the anastomosis. Adhesion severity was scored according to the number of adhesions from different organs as well as removability of adhesions [see Supplementary Table 3].

2.10 Tissue harvesting and histology

The descending colon was resected from the abdomen including the anastomotic region and longitudinally opened along the incision of the mesocolon. The gonadal fat pads were resected and weighed, and the spleen and adrenal glands were resected and photographed along a ruler for subsequent measurement. Tissue for histological analysis was fixed in 4% paraformaldehyde for 24 h, then dehydrated and embedded in paraffin (formalin-fixed, paraffin-embedded [FFPE]). Sections were cut from FFPE blocks with 2.5 µm thickness prior to standard H&E and Masson trichrome staining [Morphisto] according to the manufacturer's instructions. Slides were scanned at 400× magnification [Aperio GT 450; Leica].

Histological scoring was performed by a pathologist and a surgeon blinded to the treatment groups using pseudonymization of individual samples. Microscopic abscesses were defined as an abundance of inflammatory cells, especially neutrophil granulocytes, over mesenchymal cells and fibroblasts using H&E and Masson trichrome-stained slides [Figure 3B]. Scoring of the anastomosis was performed adapted to the histological anastomotic healing score by Philipps et al.²⁰ as well as a functional healing score¹⁶ [see Supplementary Table 4]. Adipocyte sizes were analysed using H&E-stained sections from gonadal fat pads. Seven representative single images per scanned slide were extracted using Aperio ImageScope v12.3.3.5048 [Leica Biosystems] and analysed using the Adiposoft plugin²¹ in FIJI/ImageJ [v1.53s].²²

2.11 Data analysis and statistics

Of 84 mice undergoing surgery in this study, four were excluded from analysis according to the a priori exclusion criteria. Of these four mice, one in the CTL, one in the CTL + HD and one in the DSS + HD group were excluded due to weight loss >20%. After examination of the abdomen, they showed apparent signs of anastomotic leakage. One mouse in the DSS + HD group died within 48 h after surgery, probably due to anaesthesia complications and did not show signs of anastomotic leakage during obduction. However, the mice with anastomotic leakage that had to be removed from the analysis due to a priori exclusion criteria were included in the overall anastomotic leakage rate [see Figure 2A] but not in any other analysis.

Statistical analysis and data visualization were performed using GraphPad Prism [Version 9.3.0, GraphPad Software]. Data are presented as mean ± SD or SEM as indicated in the figures. Shapiro–Wilk’s test was used to test for a normal distribution of data points. Statistical differences between the CTL and DSS groups were analysed using an unpaired t-test for parametric, normally distributed parameters or Mann–Whitney U-test for non-parametric parameters. Statistical differences within the CTL groups [CTL, CTL + LD and CTL + HD] and the DSS groups [DSS, DSS + LD, DSS + HD] were analysed using the one-way ANOVA with Tukey’s correction for multiple comparisons for normally distributed parameters or Kruskal–Wallis test with Dunn’s correction for multiple comparisons for non-parametric parameters. The test applied to the individual analyses is indicated in the figures accordingly. A p-value of <0.05 was considered as statistically significant.

3. Results

3.1 Dose-dependent effects of perioperative prednisolone treatment on colitis activity and postoperative recovery

LC-MS/MS was used to measure serum prednisolone concentrations. To reduce stress on the animals, blood samples were only taken before the final evaluation at POD3 or POD7. No prednisolone could be detected in sera from mice not receiving prednisolone therapy. The maximum serum concentration for mice receiving high-dose prednisolone treatment was 95.89 ng/ml 10 min after oral gavage and for mice receiving low-dose prednisolone treatment 22.21 ng/ml 15 min after oral gavage [Figure 1C]. As blood samples were taken at different time points after final gavage of prednisolone for each mouse, we used non-linear regression analysis to model serum concentration/time profiles. While high-dose treatment led to higher serum concentrations, the half-life was similar for both dosages [t_½ = 29.53 min for low-dose and t_½ = 27.73 min for high-dose treatment].

To evaluate the colitis disease activity, we used a clinical disease activity index [DAI] combining scores for weight loss, stool consistency and intestinal bleeding [Supplementary Table 1]. On the day of surgery, mice in the DSS group had a significantly higher DAI than mice in the CTL group [median DAI 0.67 vs 0.00; p = 0.005; Figure 1E]. Prednisolone treatment had no significant effect on the DAI in the control groups [CTL + LD and CTL + HD], but while low-dose prednisolone treatment resulted in a slightly lower DAI in mice with DSS colitis (median DAI 0.67 [DSS] vs 0.33 [DSS + LD], p = 0.17) high-dose prednisolone treatment resulted in a significantly higher DAI (median DAI 1.0 [DSS + HD]; p = 0.009; Figure 1E). Regarding the sub-scores of the DAI, the most prominent difference between the CTL and DSS groups was seen in the incidence of rectal bleeding as measured using the haemoccult test [Supplementary Fig. 1A]. The DAI was generally higher at POD3 as the postoperative weight loss was included in the index. The main differences, however, remained with a higher DAI for mice with DSS colitis compared to controls. Furthermore, mice with DSS colitis and high-dose prednisolone presented with a higher score than those without any prednisolone treatment [Figure 1E].

During daily body weight measurements we saw a decrease in body weight in mice during the induction of DSS colitis while control mice on average did not lose body weight [Figure 1D, F]. Mice with DSS colitis weighed significantly less on day 5 and 6 of colitis induction, but the difference was no longer statistically significant on the day of surgery, which might be due to the intended overall small effect of DSS colitis on body weight in our model [Supplementary Fig. 1B]. After surgery, body weight declined dramatically in all experimental groups and weight loss peaked at POD3 with a mean weight loss of 12.5% for the CTL group and 14.9% for the DSS group. Interestingly, during macroscopic examination we noted that the gonadal fat pads differed in size for the individual groups. We thus further examined the gonadal fat pads as indirect indicators of the catabolic response after surgery. We found that in the DSS group, the gonadal fat pads were significantly lighter than in the control group at POD3 (mean weight 0.06 g [DSS] vs 0.13 g [CTL]; p = 0.001; Figure 1G). Perioperative prednisolone treatment did not affect the gonadal fat pad weight in mice without colitis, but in mice with DSS colitis, low-dose prednisolone treatment prevented the decline in gonadal fat pad mass (mean weight 0.11 g [DSS + LD]; Figure 1G). At POD7, there was no longer any difference in gonadal fat pad weights between the treatment groups. Histomorphometric analysis of adipocyte size [exemplary histological images in Figure 1I] within the gonadal fat pads showed a similar pattern with a larger mean adipocyte size in the CTL group compared to the DSS group (mean adipocyte size 647.2 µm² [CTL] vs 430.3 µm² [DSS], p = 0.011) while low-dose prednisolone treatment prevented adipocyte size loss in mice with DSS colitis (mean adipocyte size 657.8 µm² [DSS + LD], p = 0.018; Figure 1H).

In summary, adequate serum levels of prednisolone were achieved in our model. Mice with DSS colitis showed clinical signs of colitis and had delayed postoperative recovery and elevated fat catabolism compared to controls. High-dose prednisolone treatment worsened the effect of colitis in combination with surgery on disease activity and fat catabolism while low-dose prednisolone treatment showed beneficial effects.

3.2 Dose-dependent effect of prednisolone treatment on anastomotic leakage during DSS colitis

We used macroscopic evaluation and endoscopy to define anastomotic leakage in our model. If one of the following criteria was met, anastomotic leakage was assumed: [i] apparent leakage upon macroscopic examination of the anastomosis [Figure 2C IV], [ii] abscess formation around the anastomosis [Figure 2C III] and [iii] obvious leakage upon endoscopic examination [endoscopic score = 4; Figure 2D III and IV]. One mouse in the CTL group [7.14%], no mouse in the CTL + LD [0%], one mouse in the CTL + HD group [7.14%], two mice in the DSS and the DSS + LD groups [14.29%] and four mice in the DSS + HD group [28.57%] developed anastomotic leakage [Figure 2C]. Thus, low-dose prednisolone treatment did not affect leakage rates in mice with DSS colitis while high-dose prednisolone treatment doubled the anastomotic leakage rate. We quantified adhesion formation around the anastomosis as previously described,¹⁶ mainly considering the number of different organs forming adhesions around the anastomosis as well as the removability of the adhesions [Supplementary Table 3]. At POD3, there was no significant difference between the treatment groups regarding the adhesion score. At POD7, there were generally more adhesions that were mostly not removable, and thus the adhesion score was generally higher at POD7 [Figure 2B]. Interestingly, the adhesion score was significantly lower in the DSS + HD group compared to the DSS + LD group. We used endoscopic evaluation of the anastomosis to score fibrin coverage of the anastomosis [Figure 2D, I and II] as well as determine anastomotic leakage endoscopically [Figure 2D, III and IV]. At POD3, the median endoscopic score was 2 for the CTL and DSS group, 0 for the CTL + LD, CTL + HD and DSS + LD groups, and 2.5 for the DSS + HD groups, but the differences were not statistically significant. At POD3, two mice [one from the DSS and one from the DSS + LD group] were diagnosed with anastomotic leakage during endoscopy [endoscopy score IV], but also demonstrated to have abscess formation upon obduction. At POD7, the median endoscopic score was 2–2.5 with no significant differences between the experimental groups. No anastomotic leakage was detected via endoscopy at POD7.

We further evaluated the anastomosis functionally using bursting pressure measurements. There was no statistically significant difference in bursting pressure at POD3 or POD7 between the experimental groups, but a trend towards a lower bursting pressure in the DSS group compared to CTL could be detected (mean bursting pressure 89.4 mmHg [CTL] vs 50.4 mmHg [DSS], p = 0.11; Figure 2F). Bursting pressure was generally higher at POD7 and the bursting location was not the anastomosis itself but rather the colon proximal or distal of the anastomosis, and thus conclusions from bursting pressure measurements on anastomotic strength were limited.

In summary, perioperative short-term low-dose prednisolone treatment seems safe in combination with DSS colitis as it does not lead to a higher anastomotic leakage rate in mice with DSS colitis, nor does it negatively influence adhesion formation around the anastomosis, the endoscopic score or the bursting pressure. High-dose prednisolone treatment in mice with DSS colitis, however, leads to higher anastomotic leakage rates and weakens adhesion formation around the anastomosis.

3.3 Histological evaluation of anastomosis reveals beneficial effect of low-dose prednisolone on anastomotic healing

We used H&E as well as Masson trichrome staining of longitudinal sections of the anastomosis to evaluate healing of the anastomosis microscopically. In a first step, we analysed whether microscopic abscess formation was present or not. Microscopic abscess formation was defined as an abundance of inflammatory cells, especially neutrophil granulocytes, over mesenchymal cells and fibroblasts with or without consecutive necrosis within the anastomotic site [Figure 1B]. Combining histological sections from POD3 and POD7, we found microscopic abscess formation in 15.38% of CTL, 28.57% of CTL + LD, 53.85% of CTL + HD, 50.00% of DSS, 28.57% of DSS + LD and 61.54% of DSS + HD mice [Figure 3A]. Thus, in mice without colitis, prednisolone treatment led to a dose-dependent increase in microscopic abscess formation while in mice with DSS colitis, low-dose prednisolone prevented microscopic abscess formation.

During the course of anastomotic healing, a dramatic change in the microscopic image between POD3 and POD7 could be seen. At POD3, a thickening of the submucosa and increase in immune cell infiltration was seen at the anastomotic site with no connecting tissue between the intestinal ends, leading to the hypothesis that at POD3, the anastomosis is still held together by the surgical sutures. At POD7, however, the anastomotic region was covered with a seal of extracellular matrix and cell infiltrates of immune and spindle-shaped mesenchymal cells and fibroblasts to various degrees. To quantify the microscopic healing of the anastomosis, we applied a semi-quantitative healing score combining a functional healing score with the well-established Philipps score²⁰ as previously described.¹⁶ When applying the functional healing score, no statistically significant difference between the experimental groups was detected, but some trends were visible [Figure 3D]. At POD3, the Philipps score was significantly higher in the CTL group compared to the DSS group (mean score 3.29 [CTL] vs 1.33 [DSS], p = 0.0004; Figure 3E). Furthermore, low-dose prednisolone treatment prevented impaired healing as quantified by the Philipps score and led to a significantly higher mean Philipps score of 2.86 at POD3 for the DSS + LD group compared to DSS alone [p = 0.006; Figure 3E]. As expected, the Philipps score was higher for all experimental groups at POD7, but differences between the experimental groups became apparent. DSS colitis led to a significantly lower score compared to controls (mean score 8.25 [CTL] vs 4.71 [DSS], p < 0.001). Low-dose prednisolone treatment again prevented an impaired healing score due to colitis (mean score 7.43 [DSS + LD] vs 4.71 [DSS]; p = 0.001), whereas high-dose prednisolone treatment did not influence the healing score compared to DSS alone [mean score 5.42; Figure 3E]. In mice without colitis, high-dose prednisolone treatment resulted in a significantly lower score [mean score 4.58] compared to no prednisolone treatment [mean score 8.25, p < 0.001] and to low-dose prednisolone treatment [mean score 7.00, p = 0.015]. Combining the functional score with the Philipps score showed a similar pattern [Figure 3F].

Taken together, perioperative short-term low-dose prednisolone treatment prevented microscopic abscess formation at the anastomotic site in mice with DSS colitis and led to a better microscopic healing as shown using two semi-quantitative histological anastomotic healing scores. High-dose prednisolone treatment led to higher rates of microscopic abscesses in mice without DSS colitis and higher leakage rates in mice with DSS colitis, showing a dose-dependent effect of perioperative prednisolone treatment on intestinal anastomotic healing. Collectively, the data presented here indicates that perioperative short-term low-dose prednisolone treatment might be beneficial in the context of anastomotic healing and consecutive colitis and that high-dose prednisolone treatment leads to higher anastomotic complications such as leakage, which confirms clinical data suggesting a dose-dependent effect of perioperative prednisolone treatment in the context of IBD on anastomotic leakage.²³

4. Discussion

The present study is the first to combine a murine IBD model with intestinal anastomosis formation and a perioperative steroid therapy to evaluate the dose-dependent effect of prednisolone on intestinal anastomotic healing in the presence of colitis. In our study, high-dose prednisolone treatment led to a two-fold increase of anastomotic leakage in mice with DSS colitis compared to low-dose and no perioperative prednisolone treatment as assessed by endoscopic and macroscopic evaluation of the anastomosis. Furthermore, mice with DSS colitis and low-dose prednisolone treatment showed significantly higher histological healing scores and lower rates of microscopic abscesses compared to mice with DSS colitis without and with high-dose prednisolone treatment. Thus, our data even suggest a beneficial effect of low-dose prednisolone treatment on anastomotic healing in the context of colitis. Additionally, we found a dose-dependent effect of perioperative prednisolone treatment on disease activity in mice with DSS colitis, with a consecutive increased loss in body weight after surgery in mice with high-dose prednisolone treatment while low-dose prednisolone treatment did not increase preoperative disease activity or postoperative weight loss. Furthermore, gonadal fat weight and adipocyte size were retained by low-dose prednisolone treatment in DSS mice, indicating a beneficial effect on postoperative metabolism and suggesting a prevention of overwhelming postoperative catabolism.

In mice without DSS colitis, prednisolone treatment led to a higher rate of microscopic abscesses in a dose-dependent manner, increasing the microscopic abscess rate from 15.38 to 28.57% [low-dose] and 53.85% [high-dose]. Also, the histological healing score was significantly decreased in mice with high-dose prednisolone treatment compared to those without and with low-dose prednisolone treatment. These results on the effect of prednisolone treatment on anastomotic healing concur with the results from Baca et al.²⁴ who found a negative effect of methylprednisolone treatment on anastomotic healing in their rat model. However, they did not include a colitis model in their study and their dose for low-dose methylprednisolone treatment was 0.28 mg/kg/day, which is an equivalent to 20 mg/day in a 75-kg patient. In our model we used 0.13 mg/kg/day [low-dose] and 0.53 mg/kg/day [high-dose], which is equivalent to a daily dose of 10 and 40 mg prednisolone in a patient of 75 kg body weight respectively. We chose those dosages according to the established cut-off of 20 mg prednisolone per day.¹⁰ Our results thus support the hypothesis that perioperative short-term low-dose prednisolone treatment [i.e. an equivalent of 10 mg/day] does not increase the risk of anastomotic leakage in colitis patients but might even be beneficial regarding the healing process by dampening disease activity.

In our study, mice with DSS colitis and low-dose prednisolone treatment presented with a lower DAI on the day of surgery compared to mice without prednisolone treatment, although the difference was not statistically significant (median DAI 0.67 [DSS] vs 0.33 [DSS + LD], p = 0.17). Mice with high-dose prednisolone treatment, however, presented with a significantly higher DAI after DSS colitis induction. It has been shown that glucocorticoids, despite exerting anti-inflammatory effects on DSS colitis, dampen epithelial barrier function and adequate immune responses towards mucosal injury and thereby lead to increased weight loss and rectal bleeding in the course of DSS colitis induction. ²⁵ These findings by Ocón et al. thus correspond to our results and might explain the increased preoperative DAI score in the mice receiving high-dose prednisolone in our experiment. However, the lowest budesonide dose in their study was 1 µm/day, equivalent to 50 mg/kg in a 20-g mouse, which corresponds even after adjusting for dose equivalence to prednisolone in a daily dose that is higher than the high-dose treatment in our study.

High-dose as well as long-term steroid therapy has been identified as a risk factor for postoperative IASCs and intra-abdominal abscess formation in IBD patients undergoing bowel resection surgery, but this evidence is based on retrospective studies with considerable heterogeneity.^4,5,9,26 No randomized-controlled trial exists so far to evaluate the effect of prednisolone or another steroid in the context of IBD on the postoperative outcome with regard to anastomotic healing and anastomosis-related IASCs. A dose-dependent effect of glucocorticoid treatment has been hypothesized in the context of post-surgical complications in patients undergoing surgery for IBD-related complications or drug-resistant disease courses. Hence, most guidelines recommend a reduction of prednisolone treatment to below 20 mg/day preoperatively without compromising disease control in individual patients.^10,11,27,28 However, the question remains whether the continuation of low-dose prednisolone treatment might even be beneficial in IBD patients undergoing bowel resection surgery or if it should be completely discontinued.

As our results indicate, short-term perioperative exposure to low-dose prednisolone treatment is beneficial with regard to intestinal anastomotic healing during colitis, potentially by controlling disease activity, and we propose that evaluation of disease activity preoperatively might enhance the clinical outcomes of IBD patients requiring bowel resection surgery. Accordingly, Lee et al. found that high preoperative disease activity measured by the Crohn’s Disease Activity Index [CDAI] score can predict postoperative complications ²⁹. Furthermore, some studies suggest that inflamed resection margins are associated with higher rates of postoperative anastomotic leakage in Crohn’s disease. ^30,31 These findings might additionally favour a perioperative short-term low-dose prednisolone therapy in patients with Crohn’s disease.

In conclusion, our data suggest a beneficial effect of perioperative short-term low-dose prednisolone treatment on anastomotic healing after intestinal anastomosis surgery in the context of colitis. Thus, we suggest evaluating colitis activity prior to surgery and adapt preoperative prednisolone dosage accordingly.

Funding

This project was financially supported by institutional funds of the Department of Surgery, TUM School of Medicine, Technical University of Munich, Munich, Germany. MCW received funding from the Clinician Scientist Program [KKF; H-17] funded by the Technical University of Munich, Munich, Germany.

Conflict of Interest

None of the authors have a conflict of interest related to this article.

Acknowledgements

We would like to thank all animal keepers and veterinarians from the Zentrum für Präklinische Forschung, TUM School of Medicine, for their support during this project.

Author Contributions

MCW, JB, SR, HF and PAN were responsible for experimental design and conceptualization of the study. MCW, JB, AB and ZC performed in vivo experiments and were responsible for data acquisition and analysis. RO performed analysis of prednisolone serum levels. AK assisted in histological analysis. MCW, JB and PAN were responsible for the original draft preparation. ZC, AB, RO, SR, AK, DW and HF critically revised the manuscript. All authors have read and agreed to the published version of the manuscript.

Data Availability

The data underlying this article are available in the article and in its online supplementary material.

References