Crohn’s disease is a chronic inflammatory bowel disease (IBD) characterised by a relapsing–remitting course with trans-mural inflammation of potentially any section of the digestive tract, leading to various intestinal and extra-intestinal manifestations.1 2

The pathogenesis of Crohn’s disease remains to be fully elucidated, but it has been suggested to develop in a genetically predisposed subject due to a dysregulated immune response to unknown antigens (probably environmental or infective, including endogenous microflora), resulting in continuous immune-mediated inflammation.3 4

Without a well-defined aetiological factor to target in IBD, current treatment procedures variably affect the complex inflammatory events that lead to intestinal damage.5 6 Indeed, no drug is, at present, curative, and a significant percentage of patients require repeated surgery to combat complications of the disease; moreover, a relevant subgroup of patients is refractory or intolerant to many pharmacological approaches.7^–9 Almost all drugs employed in the advanced forms of the disease, such as immunosuppressors and immunomodulators, have potential risks, both in short- and long-term use.10 11 Finally, the need to depend upon chronic maintenance treatment with drugs requiring frequent monitoring leads to low compliance in a large percentage of patients.12 This general dissatisfaction with current medical treatment would justify the search for more effective therapeutic approaches.

Autologous haematopoietic stem cell transplantation (HSCT) has recently been proposed as an innovative form of treatment in the more severe forms of refractory Crohn’s disease, showing good safety and promising efficacy. Following reports concerning clinical responses to allogeneic and autologous HSCT in patients with Crohn’s disease and concomitant malignancies,13^–20 recent reports from the Northwestern University Medical Center of Chicago described the first wide experience with autologous HSCT, specifically used to treat refractory Crohn’s disease, reporting enthusiastic preliminary results.21

Based on the same rationale, a phase I–II treatment protocol, including autologous HSCT, was started more than 1 year ago at the Gastroenterology Unit of “L. Sacco” University Hospital, Milan, a third-level centre for the diagnosis and treatment of approximately 3100 IBD patients, in collaboration with the Bone Marrow Transplantation Centre of “Fondazione Ospedale Maggiore Policlinico, Mangiagalli e Regina Elena”, University of Milan. The safety and efficacy of a protocol of autologous HSCT were explored for use in refractory Crohn’s disease, with unselected peripheral blood stem cells (PBSCs) in an effort to reduce procedure-related toxicity, as suggested by preliminary reports in other autoimmune diseases.22 23

The present report refers to the experience with the first four patients who have completed at least 6 months’ (range 11–20) follow-up. The primary aim was to investigate, for the first time, in the preliminary phase of our study, the clinical efficacy of HSCT in inducing remission, at the third month after transplantation, considered a good outcome for these patients in whom conventional treatment was not feasible.

MATERIALS AND METHODS

Study design

This pilot, phase I–II, study was designed to investigate the safety and efficacy of HSCT using high-dose cyclophosphamide (CTX), rabbit anti-thymocyte globulin (ATG) and autologous HSCT without CD34⁺ selection in patients with active Crohn’s disease, refractory or intolerant to conventional treatment protocols, including infliximab.

Patient selection

From November 2005, patients aged 18–65 years with active moderate–severe Crohn’s disease (Crohn’s Disease Activity Index (CDAI) ⩾250),24 refractory or intolerant to various conventional treatment schedules including corticosteroids and at least two immunosuppressors (usually azathioprine/methotrexate and infliximab), were considered eligible for HSCT. For eligibility, the presence of active anatomical lesions, confirmed endoscopically and/or radiologically, as well as a C-reactive protein level >1 mg/dl were also considered necessary. Enrolment in the study was dependent upon signing an informed consent form, which was approved by the Institutional Review Board and by the Ethics Committees of the “L. Sacco” University Hospital and “Fondazione Ospedale Maggiore Policlinico Mangiagalli e Regina Elena” of Milan.

Exclusion criteria included: serum bilirubin level greater than twice the upper reference value, serum transaminases greater than twice the upper reference value, creatinine clearance <50 ml/min, diffusion capacity for carbon monoxide <45%, acute coronary events within the last year, myocardial ejection fraction <50%, absolute neutrophil count <1500/μl, absolute CD4⁺-lymphocyte count <200/μl, platelet count <50 000/μl, active infections (except for fistulising perianal disease without abscesses), toxic megacolon, intestinal perforation, psychiatric disorders including alcoholism or drug dependency or any other disturbance not allowing expression of adequate informed consent.

All patients had undergone endoscopic and/or radiological assessment of disease activity within the 6 weeks prior to enrolment.

Transplantation protocol

Following mobilisation with CTX 1.5 g/m2 and G-CSF 10 μg/kg/day, peripheral blood stem cells (⩾3×106 CD34⁺ cells/kg of patient body weight) were collected by leukapheresis procedures and cryopreserved in DMSO until transplantation.

The conditioning regimen included CTX 50 mg/kg/day administered on days −5 to −2 and rabbit ATG 2.5 mg/kg administered on days −4 to −2. Haematopoietic stem cells were thawed and infused intravenously on day 0.

Patients were hospitalised and fully equipped in single laminar air flow rooms, during the entire conditioning treatment and up to haematological recovery. Anti-emetic and wide-spectrum anti-microbial therapy, parenteral nutrition and irradiated haemocomponent transfusions were delivered as supporting regimen. Anti-microbial prophylaxis included oral levofloxacin and fluconazole. Moreover, long-term antiviral prophylaxis with acyclovir was administered. Any significant febrile episode, during the neutropenic phase, was treated with an empirical poly-antibiotic combination of piperacillin/tazobactam and amikacin.

Evaluation of response

Neutrophil engraftment was defined as a neutrophil count >500/μl, while platelet engraftment was defined as a platelet count >50 000/μl. Each patient was asked to return to the centre 1, 2, 3 and 6 months after HSC re-infusion and, thereafter, every 6 months. At each evaluation, the clinical assessment included case history and physical examination, CDAI measurement, standard biochemical tests including CRP, weight/body mass index, bowel ultrasound sonography and anal ultrasound or other radiological tests, if clinically necessary. Endoscopic evaluation was performed at 3 and 12 months, while quality of life (QoL) was assessed using the Inflammatory Bowel Disease Questionnaire (IBD-Q)25 at 3, 6 and 12 months.

Primary endpoints were toxicity and clinical remission at 3 months. Secondary endpoints were clinical and endoscopic response at 3 months and toxicity, clinical and endoscopic remission/response at 12 months.

Clinical remission was defined as a CDAI <150, without symptoms and disease-related signs continuously for at least 1 month.

Clinical response was defined as a reduction in the CDAI of at least 100 points with respect to the basal value of >150. Clinical relapse was defined as the recurrence of disease-related symptoms with a CDAI ⩾150, increased by >100 points compared to basal values.

Endoscopic remission was defined as complete absence of mucosal lesions at endoscopic examination. Endoscopic activity was quantified by the Simple Endoscopic Score for Crohn’s Disease (SES-CD), in which selected endoscopic parameters (ulcer size, ulcerated and affected surface, stenosis) are scored from 0 to 3, with lower score indicating better endoscopic response.26

The IBD-Q, used to evaluate QoL, is a disease-specific QoL index, exploring 32 items concerning bowel symptoms, systemic symptoms, functional impairment, social impairment and emotional disorders. Total IBD-Q score ranges from 32 to 224. A higher score indicates better QoL, with a value >170 indicating significant improvement.25

RESULTS

Basal demographic and clinical characteristics

Baseline characteristics of the patients are outlined in table 1. Four white Caucasian patients (three male, one female), median age 31 years (range 26–45) were enrolled. By protocol, all patients presented with a moderate–severe disease, median CDAI of 319 (range 272–345). All patients were corticosteroid-resistant and showed refractoriness or intolerance to several conventional drug-lines (median 6, range 5–7), including thiopurines and infliximab. It is worthwhile pointing out that all patients were refractory to infliximab and methotrexate; azathioprine was ineffective in two patients (cases 1 and 4) and not tolerated in the other two (acute pancreatitis in cases 2 and 3), one did not tolerate 6-mercaptopurine (hepatitis and leukopenia in case 3) and one thalidomide (high sedation in case 1). The third case was refractory also to certolizumab pegol, while the first case was also refractory to multiple antibiotic lines (metronidazole, ciprofloxacin, amoxicillin, rifaximin), administered for recurrent small bowel bacterial overgrowth (see below). All drugs, including steroids, were gradually tapered within the mobilisation phase.

Three patients had already undergone surgery: intestinal resections (multiple ileal resections in case 1 and single colonic resection in case 4), multiple strictureplasties (case 1) and one fistulectomy and setons positioning for perianal disease (case 2).

Disease location was ileocolonic in 3/4 patients (cases 1, 2 and 4), and colonic in case 3. Among those presenting ileocolonic diseases, one patient (case 1) showed no signs of colonic activity at the time of enrolment, while case 4 had no documented ileal activity. All cases had severe mucosal damage, with ulcers observed at endoscopy, and a SES-CD ranging from 12 to 14.

Extra-intestinal manifestations were reported by three patients, with arthralgia in cases 2, 3 and 4 at the time of enrolment. The third patient had a previous diagnosis of HLA B27-negative ankylosing spondylitis.

Two patients (cases 2 and 4) had active perianal disease at the time of enrolment, while the third patient developed the condition after the mobilisation phase (see below); thus, 3/4 patients had a pattern of penetrating–fistulising disease. Only case 1 had a stenosing disease, requiring, in the past, surgery on three occasions, namely two major ileal resections and 15 ileal strictureplasties. On those occasions, active micro-bleeding from ileal ulcers (histologically reported to be Crohn’s-related) were identified at the site of previous stricturoplasties; it is worthwhile pointing out that during another operation an ileal haemangioma as the source of the bleeding was unexpectedly found and removed. Moreover, this patient presented recurrent positivity at glucose and/or lactulose breath tests for small bowel bacterial overgrowth, however, without the full clinical syndrome and showing refractoriness to multiple antibiotics. At the time of enrolment, this male patient showed no symptoms of intestinal stenosis, while the high activity index was related to severe refractory anaemia. This was caused by active chronically bleeding upper ileal ulcers revealed at videocapsule endoscopy prior to transplantation, although the same device stopped at the site of a previous strictureplasty, thus requiring a further surgery to remove it, during which ileal ulcers, localised at the site of at least four previous strictureplasties, were confirmed. After surgical removal of the capsule, the patient continued to present chronic bleeding resulting in very low haemoglobin levels (Hb 4–7.6), requiring multiple transfusions. This patient was the most difficult to treat, having been submitted several times to surgery and also being refractory to potential anti-angiogenetic treatment such as infliximab and thalidomide. Based on this complex scenario and the dependency upon chronic transfusions (1–2 U/week) the patient was, nevertheless, enrolled in the study and his low haematocrit and poor general conditions, related to severe anaemia, justified a high clinical activity index.

All patients refused surgery to treat their active disease. The IBD-Q was poor with a median score of 93 (range 81–144).

Mobilisation phase

Technical and clinical data related to the mobilisation phase are outlined in table 2. The mobilisation phase was technically satisfactory and, overall, well tolerated according to the usual clinical practice. Minor conventional side-effects were common, such as nausea, headache and arthralgia, but well tolerated and successfully submitted to symptomatic treatment. The median number of leukapheresis sessions was 1.5 (range 1–3). The median harvested and re-infused CD34⁺ cell counts were as follows: CD34⁺, 11.085×106/kg body weight (range 7.49–14.49) harvested and 4.565×106/kg body weight (range 3.62–7.38) infused. Only one patient (case 3) developed, 2 days after stem cell collection and for the first time in his clinical history, a perianal fistula, complicated by a perianal abscess, requiring surgical drainage. After resolution of this complication, the patient was maintained on antibiotics and, 3 months later, the transplantation phase was started.

During the week before hospitalisation for transplant, clinical evaluation was carried out in order to analyse the potential efficacy of cyclophosphamide in the first phase of the study. As shown in table 2 and fig. 1, no patients reached clinical remission (CDAI<150); on the contrary, three patients presented a slight deterioration in their activity index (median CDAI 339, range 258–404). In the only patient (case 4) showing an improvement in the clinical activity index, the CDAI value was maintained, however, in the range of moderate–severe disease.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Clinical efficacy. CDAI values pre- and post-transplantation.

Transplantation phase

After a median time of 1 month (range, 2 weeks to 3 months) since HSCT collection, all patients were hospitalised for the final phase including conditioning regimen and stem cell re-infusion (table 3). The median hospitalisation time was 24.5 days (range 19–32). Patients reached a neutrophil count ⩾500/μl within a median time of 11.5 days (range 10–13) and a platelet count ⩾50 000/μl within a median time of 8.5 days (range 0–9). The median number of packed red blood cells and single donor platelet transfusions was 1 (range 0–12) and 0.5 (range 1–2), respectively. Engraftment was prompt and complete already at the first month, and none of the patients developed late cytopenia.

All patients reported a progressive improvement in their intestinal symptoms already at the time of discharge. The haemorrhagic patient (case 1) reached a stable haemoglobin level of 14 mg/dl, without need of further transfusions. All patients were maintained without immunosuppressive treatment; topical mesalamine was continued only in one patient (case 2), while antibiotics (ciprofloxacin or metronidazole), for perianal fistula, were continued in two patients (cases 2 and 3).

Safety

Overall, the transplantation phase was well tolerated and no deaths or serious life-threatening complications occurred. All patients had culture-negative fever, either neutropenic or disease-related. One patient (case 1) developed chest pain related to pleural and pericardial effusions of unknown origin, rapidly resolved with conservative treatment. Another patient (case 4) developed BK virus-related macrohaematuria, a few weeks after discharge, which resolved completely with hyperhydration within 5 months of transplantation.

Disease response at 3 months

All patients were in complete clinical remission within the third month, with a median CDAI of 91 (range 56–102). Remission was reached within the first month in three patients (cases 1, 3 and 4). In case 2, a slower clinical response was observed and continuation of topical mesalamine and antibiotics for perianal fistula was necessary. However, at the third month, the patient had reached clinical remission, and all treatment was withdrawn (figs 1 and 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2 Overall clinical efficacy. Median CDAI pre- and post-transplantation.

Endoscopic assessment, performed in three patients (cases 2, 3, 4), revealed a mucosal response in all three (fig. 3). Complete endoscopic remission was achieved, for the first time in their clinical history, in cases 3 and 4 (fig. 4), the SES-CD falling from a baseline value of 14 and 12 to 0 and 2, respectively, after 3 months, while the female patient (case 2), even if not achieving complete mucosa healing, showed, nonetheless, improvement in colonic lesions, with a decrease of SES-CD from 13 to 7, after 3 months.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3 Endoscopic efficacy. Mucosal healing according to SES-CD scoring.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4 Endoscopic remission in case 3. A  =  pre-transplantation. B  =  3 months post-transplantation. C  =  12 months post-transplantation.

No data related to the mucosa were available in case 1 on account of the difficulty in studying the ileal mucosa due to the upper ileal location. Bowel enema analysis was not considered adequate to study the mucosal bleeding problem and ultrasound bowel sonography showed no signs of disease activity. Videocapsule endoscopy was not repeated due to the previously reported blockage in a strictureplasty. However, this patient maintained good haematological parameters (Hb 11 mg/dl) and no further episodes of macro- or microscopic bleeding were reported at the third month endpoint.

As far as perianal disease is concerned, anal sonography analysis showed complete fistula closure in case 4 (confirmed by magnetic resonance) and stable perianal fistula tract, with no other complications, in cases 2 and 3, even if pauci-symptomatic.

Extra-intestinal complications (arthralgia) disappeared in all affected patients. Overall QoL improved (fig. 5), with a median IBD-Q of 199 (range 142–213).

• Download figure
• Open in new tab
• Download powerpoint

Figure 5 Quality of life. IBD-Q results pre- and post-transplantation.

Survival and disease response during follow-up

After a median follow-up of 16.5 months (range 11–20), all patients are alive and 3/4 maintain remission without further treatment. The CDAI profile during the follow-up period is shown in figs 1 and 2.

Endoscopy findings were available, at 12 months, for cases 2 and 3, in which no lesions were observed (figs 3 and 4) and SES-CD decreased to 3 and 0, respectively. As far as case 2 is concerned, this was the first time in his clinical refractory phase. Perianal disease was resolved in all patients, as confirmed by magnetic resonance. Note that the recto-vaginal fistula in the female patient (case 2) was resolved after 6 months, as confirmed at ultrasound and magnetic resonance.

The only clinical relapse was observed in case 1, at the fourth month, when a recurrence of melena with severe anaemia was detected. Infliximab plus azathioprine treatment was recommenced, without effect, leading to the need of further weekly transfusions. Double balloon endoscopy, performed when this technique became available at our centre, showed the usual peri-strictureplasty ileal ulcers, with slight active bleeding. Histological examination of the ileal samples showed non-specific chronic inflammation not otherwise classifiable, while a lactulose breath test confirmed positivity for small bowel bacterial overgrowth, although conventional antibiotic treatment (metronidazole, ciprofloxacin, amoxicillin, rifaximin) was again ineffective, thus this patient remained a clinical dilemma, requiring recommencement of chronic blood transfusions.

With the exception of this patient, a good long-term QoL was maintained in the other three cases, with a maximum IBD-Q of 213. Unexpectedly, a higher IBD-Q value than those recorded during the entire follow-up was found in the anaemic patient at the third month (fig. 5).

DISCUSSION

Improved safety of autologous HSCT, over the last few decades, has been followed by increasing acceptance of HSCT as an experimental approach in the management of severe autoimmune diseases resistant to conventional treatment.23 27^–30 In the case of autologous transplantation for autoimmune diseases, the graft is usually depleted of T cells to reduce the likelihood of re-infusing presumed disease-causing T cells.31 32

The rationale of HSCT for Crohn’s disease was originally suggested by reports of a good clinical response using allogeneic or autologous HSCT in patients with concomitant Crohn’s disease and malignancies.13–20 23 Subsequent reports referred to HSCT specifically performed to cure refractory Crohn’s disease.21 33 34 The widest experience came from the Northwestern University Medical Center in Chicago where 12 patients with chronic, active, refractory Crohn’s disease were treated. Peripheral blood stem cells were used after T-cell depletion (ex vivo CD34⁺ cell selection) to deplete activated T cells. The principal aim of this clinical study was to prove the general concept that HSCT, as salvage treatment in refractory Crohn’s disease, is safe and feasible using very powerful means of induction. After a median follow-up period of 18.5 months (range 7–37), no death occurred, while recurrence of Crohn’s disease was reported in two patients (2 and 14 months after HSCT, respectively) and only two patients achieved a treatment-free clinical remission >3 years.21 35

These preliminary results, and subsequent international debates, suggest that autologous HSCT may offer a favourable risk/benefit ratio in difficult patients. The expected procedure-related mortality, about 1–8% from registry data, would be similar to the particularly severe refractory form of the disease.36 Although the reported mortality from Crohn’s disease does not appear to be excessively high (standardised mortality ratio ranging between 1.3 and 1.5),37 38 no specific data are available regarding the subgroup of severe refractory patients for which it may be presumed to be higher. On the other hand, it should be noted that the mortality data related to autologous HSCT is derived from clinical experience with patients affected by malignancies or other life-threatening conditions, where death rates are likely related to the severity of the underlying illness/organ dysfunction. Conversely, every effort is justified to reduce mortality in patients with non-neoplastic diseases, where available data are somewhat contradictory. Namely, the reported rates range from 0% in small size single centre series to figures comparable with those of patients with advanced neoplastic diseases.23

Our experience shows encouraging results, demonstrating overall safety and both clinical and endoscopic remission as well as improved QoL in the vast majority of patients (figs 1–5). The protocol used in the present study was adopted on account of the better tolerance related to a shorter engraftment period with less risk of infections, whereas T depletion in the graft can be effectively achieved by ATG even without positive CD34⁺ selection. This good tolerance coincides with clinical efficacy. Indeed, all patients reached complete clinical remission within a short period, demonstrating that HSCT can induce clinical and endoscopic remission in previously refractory patients. Despite the limited sample size and short follow-up period, promising data have emerged as far as concerns not only maintenance of clinical and endoscopic remission in most patients, but also resolution of perianal fistulas in a slow but effective manner as a result of the HSCT protocol.

The only negative aspect, in our study population, is related to the haemorrhagic patient, even if many confounding factors may contribute to explain the lack of maintenance of the (however) short-term extraordinary benefit. This patient, indeed, showed full remission for 4 months, followed by the recurrence of melena with anaemia and transfusion dependency. Double balloon endoscopy showed the usual peri-strictureplasty ileal ulcers observed at surgery, on previous occasions, with active minor bleeding. Although histological analysis failed to provide a clear classification of the lesions, a high faecal calprotectin level was found, suggesting involvement of an inflammatory component in the relapse. On the other hand, such a tenacity to relapse, with ulcers at constant peri-strictureplasty sites, may indicate additional pathological factors. It is worthwhile mentioning the presence of a positive test for small bowel bacterial overgrowth, although it is difficult to analyse the true clinical impact, while multiple antibiotic treatment was, moreover, ineffective. On the other hand, luminal bleeding is one of the most common complications of strictureplasty, having been reported in about 2% of the procedures, even if usually of short duration.39 40 It is worthwhile pointing out that during a previous surgical intervention, performed for stenosis and anaemia, an ileal haemangioma was unexpectedly found to be the source of this patient’s recurrent bleeding, thus suggesting speculation about possible undiscovered alterations in angiogenesis or in the vascular structure.

On the contrary, if we consider the other patients with active classical forms of inflammatory-fistulising Crohn’s disease, the clinical and endoscopic results are completely satisfactory. Indeed, transplantation induced clinical remission in each patient and endoscopic remission (a controversial parameter of greater efficacy) has been reached, after variable intervals of time, in all cases. Another very interesting aspect is the resolution, in all patients, of perianal disease and, in one case, also of recto-vaginal fistulas, which, although slower, were confirmed.

Compared with previous experience from Chicago, we observed no improvement and, indeed, even a slight clinical deterioration after mobilisation with G-CSF and low doses of CTX. In fact, all patients showed an increase in CDAI compared with the baseline, with the exception of one patient (case 4) in whom, however, high levels of CDAI were maintained in the range of moderate–severe active disease. It should be pointed out that, in order to permit adequate haematological and clinical stabilisation and to better analyse the possible clinical effect of mobilisation on disease activity, a median time of 1 month should elapse between the end of mobilisation and the start of the conditioning regimen followed by stem cell re-infusion. Previously published reports have not clearly addressed this topic, suggesting a shorter latency period between the two protocol phases. The observed inefficacy of mobilisation drugs may have emerged only on account of the latency period of our protocol during which the disease continued to be substantially untreated with a well-defined therapeutic approach. Although the drugs used during the mobilisation phase (CTX, G-CSF and even the leukapheresis procedure itself) had been reported to be effective in some studies,41^–48 they were used with different treatment schedules, higher dosage and more prolonged duration of treatment, thus accounting for the different clinical effect in our series. In Europe, a randomised trial on autologous stem cell transplantation in Crohn’s disease (ASTIC) is programmed which should provide an answer to the question of whether autologous transplantation significantly adds to the effect of immunosuppression/G-CSF alone. Patients have been randomised to receive mobilisation chemotherapy with G-CSF only versus mobilisation chemotherapy with G-CSF followed by autologous transplantation.49

In conclusion, our experience confirms the promising and dramatic benefits from the use of autologous stem cell transplantation in refractory Crohn’s disease, leading both to clinical and endoscopic remission, with maintenance even after more than 1 year, without need of further major treatment. Our data also show, for the first time, that clinical benefits can be reached with a protocol not using CD34⁺ cell selection, leading in theory to a better tolerance and shorter time to engraftment. It must be stressed, however, that this treatment should, at the moment, be reserved for refractory patients, in the context of clinical trials, at referral centres.50 Further experience and randomised controlled studies are mandatory. The need is also stressed to perform basic research to better investigate the pathogenetic aspect of the reported clinical effects and to establish the predictors of an optimal response for better recruitment.