Karin Wasmann

The research presented in this thesis provides insight in the challenges of improving IBD surgical treatment on a national and international level. As described in chapter 1, the multicentre, randomised controlled PISA trial was the first study that directly compared the current standard treatment options for high perianal Crohn’s fistulas, which include i) chronic seton drainage, ii) anti-TNF medication and iii) surgical closure combined with anti-TNF. The study was powered to provide superiority of seton treatment compared to the other two groups. However, after inclusion of 44 out of the projected 126 patients, the PISA trial was preliminary terminated due to a higher re-intervention rate in the seton group (10/15, versus 6/15 anti-TNF and 3/14 surgical closure patients, P = 0.02). The results imply that chronic seton treatment should no longer be advised as the sole treatment for perianal Crohn’s fistulas. Patients who declined randomisation, due to a specific treatment preference, were included in a parallel prospective PISA registry cohort (n = 50). Interestingly, in the PISA prospective registry, inferiority of chronic seton treatment was not observed for any outcome measure.

The discrepancy, between the RCT and registry results, raised questions: which results should be used for clinical practice? Randomization in the PISA study may have had influence on participation and outcomes (e.g., external and internal validity). Following, in chapter 2, we aimed to assess the influence of patients’ preference in RCTs by analyzing partially randomized patient preference trials (RPPT); a RCT and preference cohort combined. We systematically reviewed all RPPTs published between 2005 and 2018, 44 of 3734 identified articles were included (24,873 patients). The outcomes of the RCT and preference cohort were compared. The results showed that patients preference led to the majority of patients refusing randomisation (randomisation refusal > 50% in 26 trials), hence decreasing external validity of the RCT cohort. The reported primary outcomes – reflecting internal validity – were comparable between both cohorts of the RPPTs, mean difference 0.093 (95% CI: -0.178;0.364, P = 0.502). Therefore, RPPTs could increase external validity compared with RCTs, without compromising the internal validity.

In contrast to colorectal cancer surgery, during surgery for CD, only macroscopically affected bowel is resected to prevent short bowel syndrome, as the impact of microscopic inflammation at the resection margins on recurrence rates is unclear. Chapter 3 and 4 searched for a basis to guide these resection margins. In chapter 3, both resection margins of 106 consecutive patients undergoing ileocecal resection for Crohn’s disease between 2002–2009 were revised and scored for inflammatory characteristics. The results indicated that only active inflammation at the distal colonic resection margin was an independent significant predictor for disease recurrence (88% vs. 43% vs. 51% respectively for distal, proximal, and no involved margins, P < 0.01). Hence, a more extensive resection aiming at a non-inflamed ileal margin will not be beneficial. Moreover, it revealed new insights, suggesting that active inflammation at the distal colonic resection identifies a high risk patient group with L3 disease (ileocolic phenotype) instead of L1 disease only (limited to the ileum). This patient group may benefit from postoperative medical treatment. In chapter 4, the inflammatory status of mesenteric macrophages in the mesorectum and the ileocecal mesentery in Crohn’s disease compared with non-Crohn’s disease was characterized. Proinflammatory and regulatory cells were mapped after sampling three standardised mesentery locations of 51 CD and 11 control patients (17 proctectomies and 45 ileocecal resections). Immune cells from these tissue specimens were analysed by flow cytometry for expression of CD206 in order to determine the inflammatory status. In the mesorectum, proinflammatory macrophages reside next to the inflamed rectal tissue and display a gradient to a more regulatory phenotype further away from the inflamed rectum. The ileocecal mesentery did not contain high amounts of proinflammatory macrophages adjacent to the inflamed ileal tissue. In contrast, creeping fat contained more regulatory macrophages. Therefore, there is currently no basis to perform an extended mesenteric ileocecal resection in Crohn’s disease patients. While striving to meet the quality standards for oncological care, hospitals prioritize oncological procedures more frequently, resulting in longer waiting times for surgery regarding benign diseases like IBD. Chapter 5 highlights the potential consequences of a longer interval to surgery for IBD patients compared to colorectal cancer surgery in the Amsterdam UMC, location AMC, between 2013–2015. The mean waiting time was more than 10 weeks for IBD patients, twice as long compared to colorectal cancer patients (5 weeks). While awaiting surgery, 1 out of 8 IBD patients had to undergo surgery in an (semi-)acute setting, 19% had disease complications (e.g., > 5% weight loss, abscess formation) and 44% needed additional health care (e.g., (telephone)outpatient clinic appointment, hospital admission). It highlights that the current waiting time for IBD surgery is not medically justified and creates a burden for health care resources. It is time to also set a maximally acceptable waiting time to surgery for IBD patients.

In chapter 6 and 7, we set up an international collaboration to compare the short- and long-term outcomes of the new transanal ileal pouch-anal anastomosis (ta-IPAA) technique with the standard transabdominal minimal invasive approach in UC. Ta-IPAA surgery resulted in lower morbidity rates and comparable long-term functional outcomes.

Chapter 8 focussed on the long-term functional outcomes of the novel endo-sponge® assisted early surgical closure (ESC) approach for IPAA leakage in 280 UC patients. Out of the 40 patients with anastomotic leakage, 18 were treated with ESC (2010–2017) and 22 (2002–2009) with conventional management. ESC resulted in comparable pouch function (P = 0.647) and comparable pouch failure rates (0/18 vs. 5/133, P > 0.99, resp.) versus control patients without leakage. Conventional management resulted in worse pouch function (P = 0.016) and a higher pouch failure rate (5/22 vs. 5/107, P = 0.013, resp.) compared to control patients. Therefore, ESC is associated with preservation of pouch function and might prevent pouch failure.

In the last chapter of this thesis, chapter 9, the impact of rectal stump inflammation after subtotal colectomy on both short- and long-term pouch outcomes for 204 UC patients operated between 1999 and 2017 was studied. Rectal stump inflammation (found in 82%) was not associated with an increased risk of anastomotic leakage (10.2% vs. non-inflamed 5.4%, P = 0.54). However, it was associated with a higher incidence of pouchitis (54.3% vs. non-inflamed 25.5%, Plog = 0.02). It was therefore suggested that patients with rectal stump inflammation have a more aggressive phenotype of UC.

Research questions addressed in this thesis

1. With respect to re-interventions, is seton treatment superior to anti-TNF treatment and surgical closure combined with anti-TNF for patients with a high perianal Crohn’s fistula?
Chronic seton treatment was not associated with lower re-intervention rates.

2. Is a partially randomised patient preference trial a valid alternative to a randomised controlled trial regarding internal and external validity?
A partially randomised patient preference trial is a valid alternative with a higher participation rate, thereby increasing external validity, while primary outcomes remain comparable, hence preserving internal validity.

3. What is the predictive value of microscopic inflammation at ileocecal resections margins for postoperative Crohn’s recurrence?
Inflammation at the distal colonic ileocecal resection margins is associated with an increased disease recurrence rate. Inflammation at the proximal ileal margin is not associated with an increase in disease recurrence. Therefore, more extended ileocecal resection does not seem to be beneficial.

4. Is there an anatomical variation in mesenteric macrophage phenotypes that can guide surgical resection margins in Crohn’s disease?
A gradient of pro-inflammatory macrophages in the mesorectum is associated with inflamed adjacent rectal tissue. Hence, resecting that part of mesorectum seems beneficial. A gradient of pro-inflammatory macrophages in the mesentery of the ileocolonic adjacent to inflamed ileal tissue was not observed. Moreover, the creeping fat contained a gradient of regulatory macrophages. Consequently, resecting the mesentery during ileocecal resection seems not beneficial.

5. Is a longer waiting time for IBD surgery associated with ‘waiting list complications’?
A longer waiting time for IBD surgery is associated with an increase of semi-acute surgery and non-surgical complications such as more than 5% weight loss, fistula or abscess formations requiring radiological intervention, dehydration and additional health care consumption.

6. Is transanal versus transabdominal minimally invasive pouch surgery in UC beneficial regarding short-term morbidity?
Transanal minimally invasive pouch surgery is associated with a reduction in post-operative morbidity.

7. Does transanal versus transabdominal minimally invasive pouch surgery in UC result in superior long-term pouch function?
Long-term functional outcome and quality of life after transanal and transabdominal minimally invasive pouch surgery were comparable.

8. Does Endo-sponge assisted early surgical closure of pouch leakage improve long-term pouch function?
Endo-sponge assisted early surgical closure was associated with preservation of pouch function and might prevent pouch failure, probably due to early and effective treatment of anastomotic leakage.

9. What is the impact of rectal stump inflammation on anastomotic pouch leakage and pouchitis?
Rectal stump inflammation after subtotal colectomy occurs in 80% of UC patients. It is not significantly associated with an increased anastomotic leakage rate of the pouch, but was an independent predictor for the development of (therapy refractory) pouchitis.

Discussion and future perspectives

Regarding research of IBD treatment strategies, a data gap exists for the comparison of surgical versus medical strategies. The PISA study showed how challenging such a comparison can be, yet also revealed considerable lessons learned. To properly translate clinical situations in a trial, some established assumptions should be scrutinised. First of all, surgery should not be seen only as a last resort but also as an alternative to medical treatment. Besides, trials comparing medical therapies in Crohn’s disease should not use surgical recurrence as an endpoint. The PISA registry also revealed that relatively few patients chose surgery as a treatment. It touches upon a more extensive problem that patients may not be well informed about the surgical treatment options. A fundamental factor driving this observation is probably due to the majority of Crohn’s fistula patients having a long medical history with a gastroenterologist, who advises the patients. Since, the gastroenterologist is probably less familiar with the surgical treatment options and its respective outcomes, shared decision making is likely to be impaired. A vital starting point would be a shift in patient counselling towards earlier visiting a surgeon, paving the way to talk about alternatives, instead of inevitable last resort surgery. Hopefully combined out-patient clinics regarding gastroenterology and IBD surgery, aiming at solid cooperation, will soon become entrenched in modern healthcare on a global scale.

Perianal Crohn’s disease
The optimal treatment for patients with perianal Crohn’s fistulas remains unknown. While designing the PISA study, results suggested comparable closure rates between the three treatment options. According to the most recent systematic review, the initial remission of drainage rate after anti-TNF treatment is 44%. Initial closure of fistulas in CD following surgical closure seems higher (65%). A future trial comparing these treatments head-to-head would be of great importance for these patients. Ideally all types of patients with perianal CD fistulas should be represented in large numbers. These recent closure rates also suggest that a substantial number of patients fail their therapy. For these patient, hyperbaric oxygen seems an option. To conduct such studies, consensus on the definition of a closed fistula should be reached. The definition of a fibrotic tract without collections on MRI can be correlated to patient reported outcomes to develop a firm endpoint.

Designing surgical trials
The PISA study results challenge the current dogma of the RCT being the ‘gold standard’. However, the assumption that trial participants are passive recipients of interventions is not valid. Patient preferences can influence RCT participation and outcomes. Additionally, an RCT is costly, time consuming and does not correct for learning curves. Modern research should try to adapt in order to find a healthy balance between limiting bias effects and drawing conclusions applicable for routine practice. Especially now that ‘big data’ is becoming more established in medical research, more pragmatic designs can be considered such as patient preference designs or a cohort-embedded RCT (also known as TWICS or randomised registry trial).

Ileocecal resection
Yet to be researched, but why not start with an ileocecal resection for patient with uncomplicated terminal ileitis, avoiding medical treatment? At a minimum, the short- and long-term results of the LIR!C study induce a shift in the current step-up treatment approach for uncomplicated terminal ileitis; ileocecal resection has shown to be an alternative treatment for anti-TNF instead of a last resort treatment. It’s likely that more ileocecal resections will be performed. The results of this thesis suggest that microscopically inflamed distal colonic resection margin is associated with a higher disease recurrence rate, as it identifies undiagnosed L3 disease (ileocolonic instead of ileum only Crohn’s disease). As the recurrence rate for L3 disease is significantly higher, the colon should be accurately scoped before surgery, in order for patients to be thoroughly counselled. Additionally, pathology reports should specifically address the inflammatory state of the distal resection margin, as patients with an inflamed distal margin should be considered for prophylactic treatment. Furthermore, a microscopically inflamed proximal resection margin and mapping macrophages phenotypes in the ileocecal mesentery did not result into a prognostic value. These findings intuitively support performing stricturoplasties for selected patients, in which the affected bowel is left in situ. However, ongoing research suggests that the mesentery does play a role in driving (recurrences of) CD. The specific role is probably dependent on Crohn’s location, phenotype and patient characteristics. Therefore a patient-tailored surgical approach would be desirable. In this regard a fluorescent-guided surgical approach demonstrating the extent of inflammation could be an interesting step forward. Bearing in mind that we stand at the beginning of understanding the IBD anatomy and the related role of the mesentery, we should foster the current cooperation between laboratory researchers and surgeons. Samples of resection specimens being directly analysed in the laboratory is a fertile soil for future research.

Waiting list complications
The observation that patients with IBD have to wait longer for surgery compared to patient with colorectal cancer illustrates that IBD care has taken a back seat. Quality criteria like regular multidisciplinary team meetings, centralisation of care and healthcare regulatory bodies setting the norm for time to treatment in IBD should become equally established as their respective counterparts in oncology. Public awareness must be raised to fuel these developments.

Pouch surgery
Innovation in pouch surgery is rising, as illustrated with the introduction of the ta-IPAA distal resection. To suppress the negative side effect of the learning curve, this thesis emphasises the importance of centralisation. For the treatment of pouch leakage endo-sponge assisted surgical closure is advised. This requires good collaboration between the gastroenterologist and the surgeon advocating for IBD referral centers. Also for pouch surgery fluorescent-guided surgery seems promising to further reduce the anastomotic leak rate. The prognostic value of an inflamed rectal stump on pouch outcomes should be further analysed. The results of the MIRACLE study, aiming to identify the best pouch practices throughout Europe, are eagerly awaited.

In conclusion, surgery could be introduced earlier and more often in the multidisciplinary management of IBD.