Current status and future perspectives of leukocytapheresis for inflammatory bowel disease


Takayuki Matsumoto, Departments of Lower Gastroenterology and Internal Medicine, Hyogo College of Medicine, 1-1 Mukogawa, Nishinomiya, Hyogo 663-8501, Japan. Email:


Ulcerative colitis (UC) and Crohn's disease (CD) comprise the idiopathic inflammatory bowel diseases (IBD) of the gut. The etiology of IBD is poorly understood, but an autoimmune disturbance has been suggested to play an important role in this incurable disease. Extracorporeal leukocytapheresis (CAP) is an additional adjunct for IBD patients refractory to other conventional therapies, including steroids. The primary aim of CAP should be to suppress such unwanted immunological response by removing circulating inflammatory cells from the blood stream. The first decade has been passed since CAP was approved by Japanese social health insurance policy. It is therefore now an appropriate opportunity to upgrade and summarize our current understandings and/or future perspectives of this unique non-pharmacological and non-surgical strategy for IBD patients. According to several clinical and basic research reports, an early introduction of CAP should produce higher efficacy as compared with CAP applied sometime after a clinical relapse. Likewise, CAP therapy adjusted to patients' body-weight as well as two treatment sessions per week (intensive regimen) should benefit the efficacy rate. The etiology of IBD is not fully elucidated yet. As a result, the major therapeutic strategies in the Western world have been immunosuppressive therapy, including biologics. CAP is an unusual treatment modality for IBD because it seems to have both effectiveness and safety, which should generally be balanced in this type of illness. We now have to develop future strategies with and without combining biologics to improve the quality of life of IBD patients.


Ulcerative colitis (UC) together with Crohn's disease (CD) are the major phenotypes of idiopathic inflammatory bowel disease (IBD), which afflicts millions of individuals throughout the world with symptoms that impair quality of life (QOL) and ability to function.1 Currently, the etiology of IBD is not well understood, but mucosal tissue edema, increased gut epithelial cell permeability, and extensive infiltration of the intestinal mucosa by leukocytes of the myeloid lineage are major pathologic features of this immune disorder.2 Accordingly, an extra-strategy of removing these peripheral leukocytes by an extracorporeal circulation technique, cytapheresis (CAP), has been developed in Japan, where it is now recognized as a non-pharmacologic adjunct therapy to alleviate the inflammatory response in patients with active IBD.

The first decade has been passed since the selective depletion of peripheral granulocyte and monocyte/ macrophage by extracorporeal adsorption (GMA) with an Adacolumn and the filtration leukocytapheresis (LCAP) have been accepted by the Ministry of Health, Labour and Welfare of Japan as treatments for patients with active UC who are steroid refractory.3,4 In 2009, GMA was also accepted as an adjunct therapeutic strategy for active CD patients according to the superior results obtained from a nationwide multicenter trial.5 It is now therefore an appropriate opportunity to upgrade and summarize our current understandings and/or future perspectives of this unique non-pharmacological and non-surgical strategy of CAP for IBD patients.

Current available leukocytapheresis techniques for IBD patients in Japan

Filtration leukocytapheresis and GMA are the most used CAP techniques for intractable UC patients with acute flare. According to a national survey in Japan, the total number of UC patients has been expanding gradually, and it has now reached to over 100 000. Among them, almost 50% of patients have been facing active flare more severely than moderate; and, approximately 30% of them were diagnosed as “intractable”, meaning either treatment resistance or dependent characteristics for conventional steroid therapy (Fig. 1). Patients with intractable active UC flare are potential candidates for applying an adjunct strategy, including immunosuppressant, biologics, and CAP.

Figure 1.

Number of patients with ulcerative colitis (UC) in Japan (data have been modified from 2001 Annual Report of National registration of the Ministry of Health, Labour and Welfare of Japan).

We have developed both LCAP and GMA, and the current tasks for them should be to determine the appropriate therapeutic regimen in order to obtain the maximum clinical efficacy of these unique non-pharmacological and non-surgical interventions.

Filtration leukocytapheresis.  Filtration leukocytapheresis is performed using a specially designed leukocyte removal column, Cellsorba EX (Asahi Kasei Kuraray Medical, Tokyo, Japan), set on a simple one-way hemofiltration circuit.3,6,7 A roller pump drains the patient's peripheral blood from an antecubital vein under constant flow rate of 50 mL/min. An optimal amount of Nafamostat mesilate (NM; Futhan; Torii Pharmacology, Tokyo, Japan) or heparin is mixed with saline and added to the drained peripheral whole blood as anticoagulant before infusion into the column (Fig. 2). Polyester non-woven leukocyte removal filter was installed into the polycarbonate outer shell of Cellsorba. Approximately 35% of platelets are expected to be removed by LCAP from processed peripheral blood, together with almost 100% of granulocytes and monocytes and 64% of lymphocytes (Fig. 3a).7

Figure 2.

Basic circuit diagram for extracorporeal leukocytapheresis. The structure and patient preparation are almost the same for filtration leukocytapheresis (LCAP) and granulocyte/monocyte apheresis (GMA). An inlet pressure monitor is installed in the circuit for LCAP because of the characteristics of the Cellsorba column (see text).

Figure 3.

(a) Outer shape of Cellsorba E (the column, especially designed for filtration leukocytapheresis [LCAP]), and typical scanning electron micrographs (SEM) showing leukocytes adsorped onto the non-woven polyester fibers. Photos are kindly provided by Asahi Kasei-Kuraray Medical, Tokyo, Japan. (b) Outer shape of Adacolumn (the column, especially designed for granulocyte/monocyte apheresis [GMA]), and typical SEM showing monocyte, neutrophil adsorption to the Adacolumn carrier beads. Photo of the column is kindly provided by JIMRO, Gunnma. SEM courtesy of Dr A Saniabadi.

Adsorptive granulocyte/monocyte apheresis.  Granulocyte/monocyte apheresis is performed with the Adacolumn (JIMRO, Takasaki, Japan). The circuit diagram for GMA is almost the same as that of LCAP. Peripheral whole blood drained from the patient's body is passed at 30 mL/min, a flow speed created by an external roller pump with optimal amount of NM or heparin as an anti-coagulant. The Adacolumn is filled with cellulose acetate beads, which serve as the column adsorptive leukocytapheresis carriers. The carriers in the column selectively adsorb about 65% of granulocytes, 55% monocytes/macrophages and a smaller fraction of lymphocytes. These are the leukocytes that bear the so-called FcγR and complement receptors (Fig. 3b).4,8–10

Leukocytapheresis for active UC in Japan

Granulocyte/monocyte apheresis (2000) and LCAP (2001) have been accepted by the ministry of Labour and Welfare of Japan for active UC patients. During this decade, these CAP techniques have been recognized by gastroenterologists and integrated into the Japanese national therapeutic guidelines for UC patients. According to the accepted inclusion criteria, CAP is permitted for use if the patient has been diagnosed to have UC with severe or moderate activity in spite of giving 1.0–1.5/kg per day of injection prednisolone for 2 weeks, or relapsing if their dosage of peroral prednisolone has been reduced less than 10 mg/day. According to the previous trials, proceeding volume of each session has been decided as 3000 mL in LCAP and 1800 mL in GMA.

The therapeutic mechanism of CAP has never been fully elucidated, but frequency and quantity of the CAP session should play an important role because of their individual characteristics. Originally, the therapeutic schedule of CAP for moderate to severe active UC patients consisted of two series at maximum, and one series consists of five weekly sessions. However, because a significant superior efficacy of intensive, two times per week, GMA has been proven superior to the conventional weekly GMA in a nationwide multicenter clinical trial,11 the official health insurance policy has been approved to modify the frequency of CAP by physician's decision. On the other hand, the total number of CAP sessions has been fixed as 10 times maximum for a single episode of UC flare. This perceived requirement might, at least in part, indicate why a GMA study in the USA did not show significant clinical efficacy in active UC.12

Leukocytapheresis for active CD patients

Because the primary aim of CAP is to deplete effector cells of the intestinal inflammatory response, from the beginning of its development we have aimed to apply this therapy for not only active UC, but also active CD patients. In a preliminary clinical trial of LCAP for CD patients, significant clinical efficacy together with recovering peripheral immune response has been reported.13 On the other hand, GMA has also been reported to show significant clinical efficacy for active CD patients.14,15

With the above results in mind, a multicenter study has been conducted of GMA in Japan for active CD patient refractory to > 1200 kcal/day of elemental nutrition therapy. Significant improvements in CDAI, IOIBD, and IBDQ scores were observed at week 7 of weekly GMA therapy.5 According to this evidence, GMA has been recognized to have certain potential for inducing remission and improving quality of life in patients with active CD that is refractory to conventional therapy (although no biologic agents had been approved in Japan at the time of its acceptance).

In 2009, GMA received government approval in Japan. The approved therapeutic regime of GMA for CD patients has been determined as 1800 mL/session of weekly GMA for a maximum 10 consecutive weeks. Although intensive GMA, which has been approved for UC, has never been accepted as reimbursable, treatment for CD, approximately patients with 300 active CD have been enrolled for this novel strategy during these 2 years (2009–2011) according to the manufacturer's survey.

Current knowledge for CAP

Although both LCAP and GMA have been become popular and widely used in Japan as an effective therapeutic option for active IBD patients, our current level of knowledge about the mechanism of this unique therapy remains limited. Because of its basic leukocyte removal strategy, CAP has been recognized as a potential immune-modulation therapy by directly reducing peripheral immune active cells from the patient's blood stream.

Filtration leukocytapheresis

Clinical evidences of LCAP for UC.  As described in the earlier section, LCAP has been approved in Japan only for UC. As pivotal clinical evidence, a multicenter randomized controlled trial of LCAP for active UC patients has been reported.3 The results indicate that LCAP exhibits significant efficacy for steroid-resistant and relapsing UC patients compared with conventional high-dose steroid injection therapy (h-PSL) (LCA vs h-PSL = 74% vs 32%, P < 0.05) although no significant difference has been obtained between LCAP and h-PSL in the clinical efficiency for steroid naïve UC. Simultaneously, the safety characteristics of LCAP were favorable; there were no patients who experienced significant adverse effects from LCAP. Matsumoto et al.16 has conducted a multicenter open-labeled trial of weekly LCAP therapy for active UC patients. Based on their observations, they have proposed the following significant factors correlated with the rapid LCAP response: (i) steroid resistance (P < 0.05); (ii) severe disease indicated by a clinical activity index (CAI) score greater than 11 (P = 0.05); (iii) disease duration of less than one year (P < 0.05); and (iv) high C-reactive protein levels before treatment (P < 0.01).

Therapeutic mechanism of LCAP for UC.  Immune modulation induced during LCAP has been reported previously, especially from the point of view of cytokine production. It has been shown that LCAP enhances the ability of peripheral blood lymphocytes to produce interleukin (IL)-4, an anti-inflammatory cytokine.17 Hanai et al. has reported that LCAP has been shown to decrease IL-6 release (a pro-inflammatory cytokine) in the patients' peripheral blood concomitantly with increasing IL-10, which has been reported to markedly inhibit the protein and mRNA expression of another pro-inflammatory cytokine, IL-1 during the procedure.18

Recently, the immune pathology in patients with IBD has been thought to reflect an inadequate regulatory T-cell (Treg) function in these patients. Treg constitutes 5–10% of peripheral T cells in normal naive mice, and in humans, and the CD4+ T cell phenotype expressing CD25high and forkhead box protein 3 (FoxP3) has been recognized as its functional representative.19,20 Andoh et al. has investigated the alterations in circulating T cell subsets after LCAP therapy in UC patients. LCAP increased the ratio of CD25+ CD4+-cells (Treg)/CD25- CD4+-cells significantly after therapy;21 however, another report failed to detect a significant change in the CD25BrightCD4+/total CD4+ ratio.22 These authors showed that LCAP selectively removed CD14dullCD16+ monocytes, which preferentially produce tumor necrosis factor-α (TNF-α), which has also been demonstrated during the GMA procedure.23 According to this evidence, some of the immuno-suppressive effects of LCAP therapy may be associated with a modulation of circulating T cell subsets.

On the other hand, we should focus on the platelet removal performance of LCAP since platelet removal characteristics could be a major difference between LCAP and GMA. LCAP removes approximately 35% of peripheral blood platelets, which adhere onto the surface of polyester filter of Cellsorba.7 It is believed that circulating platelets are important cells not only in hemostasis, but also in a variety of inflammatory responses.24 An increase of the peripheral platelet count has been recognized as a common feature during the active phase of IBD.25 It has also been reported previously that higher platelet numbers correlate well with disease severity.26 According to this evidence, we have hypothesized that this significant platelet removal achieved during LCAP might have an active role in downregulating severe immunological reactions in patients with UC with an acute flare. We have also proved that activated platelets may have potential to predict clinical efficacy of LCAP in severe UC patients.27

Evidence for efficacy of GMA in IBD

Granulocyte/monocyte apheresis for UC.  A milestone in the clinical history of GMA for UC was a report presenting the results from a multicenter trial of GMA for active UC patients.4 The results indicated that GMA was significantly more effective for relapsing UC patients compared with conventional h-PSL therapy (GMA vs h-PSL = 55% vs 40%, P < 0.05) In the same study, GMA was associated with fewer adverse events compared with h-PSL (< 0.001). Since then, several studies28–31 have reported very significant clinical efficacy in patients with UC following a course of GMA.

The most recent evidence for this novel intervention is in a report by Sakuraba et al.11 that found both remission induction rate (71.2% vs 54.0%; P < 0.03), and the time to remission (14.9 days vs 28.1 days; P < 0.01) were significantly better in twice a week GMA sessions compared with the routine weekly GMA therapy. Thus, the clinical efficacy of GMA could be frequency dependent. In 2011, an Italian group reported the results of a multicenter trial,32 which enrolled 230 patients (UC 194/CD 36) from 24 Gastroenterology Units. Patients received five GMA sessions and a positive outcome at 3 months was achieved in 78% of UC patients (72% remission, 5.7% clinical response) and 61% of CD patients (55% remission, 6.5% clinical response). The cumulative proportion of positive outcomes at 12 months was 87% for UC patients (84% remission, 3.4% clinical response) and 77% for CD patients (74% remission, 3.2% clinical response). Sakata et al.33 has conducted a small prospective randomized trial to compare the clinical efficiency for active UC between LCAP and GMA; however, they could not detect any significant difference between them.

Therapeutic mechanism of GMA for UC.  The therapeutic mechanism of GMA for UC can be judged from the following background. In patients with active IBD, peripheral blood granulocyte and monocytes/macrophage levels are elevated, and cells show activation behavior and increased survival time.34–39 As these leukocytes are a major source of inflammatory cytokines,40,41 the level of neutrophil infiltration into the mucosal tissue in patients with active IBD has been directly related to the severity of intestinal inflammation and clinical relapse.42,43 Adacolumn has been developed to “tame the exuberant immune system” in patients in whom an overactive immune system, namely elevated peripheral blood neutrophils, is associated with disease progression.34 However, interestingly, the observed clinical efficacy cannot be fully explained by the effects of the procedure on peripheral blood leukocytes per se. We have proven that peripheral Treg (CD25HighCD4+ T-cells) expression, which has been suppressed in active UC compared with healthy controls, was significantly increased after a single GMA session.44 The increase in CD25High+CD4+ regulatory T-cells after GMA should contribute to improved immune function of the patient. Moreover, we have proven that the number of CD4+/FoxP3+ mucosal Treg in GMA responders decreased significantly after the fifth GMA session compared with the baseline level.45 It seems possible, therefore, that GMA might impact the circulating as well as the mucosal levels of Tregs. Likewise, several other investigators have reported favorable immunological observations associated with GMA.23,46,47

Reactivation of cytomegalovirus (CMV) infection has often exacerbated UC refractory to immunosuppressive therapies. Yoshino et al. reported the clinical effect of GMA therapy for UC patients with concomitant mucosal CMV infection, and they have proposed that GMA might be a safe and effective treatment for UC patients positive for CMV because the procedure does not induce CMV reactivation.48

Future perspectives of CAP in the management of IBD patients

Optimization of processing conditions.  Clinical effectiveness of LCAP and GMA should be regulated by blood volume for the procedure (Pv), procedure time, and procedure frequency (Qf). Pv can be calculated as: Pv = Blood flow speed (Qb) × Procedure time (Qt).

Basically, slower Qb should reinforce the leukocyte removal performance of the column. Cellsorba, the LCAP column, is unsuitable for proceeding under 20 mL/min of slow Qb conditions since the platelet removal characteristics of the column could cause formation of thromboses in the Cellsorba column.49 On the other hand, Adacolumn, the GMA column, aims to adsorb granulocytes, monocytes/macrophages and a smaller fraction of lymphocytes from the patient's peripheral whole blood on cellulose acetate beads that fill the device. These are the leukocytes that bear the so-called FcγR and complement receptors.4,8–10 Therefore, GMA should be suitable for processing under slow Qb conditions because GMA appears to remove fewer platelets than LCAP.

In spite of these facts, the Pv per extracorporeal hemofiltration session has been ignored in clinical settings. According to recent national surveys, the average body weight of US citizens is 78.9 kg;50 as compared with 58.2 kg for Japanese.51 The Pv per CAP session might be an equally relevant factor bearing in mind that the main function of CAP is to deplete elevated and activated leucocytes of the myeloid lineage, like CD14+CD16+ monocytes which are a major source of TNF-α.22,23 Up to now, however, CAP has been performed at a fixed PV of 3000 mL/session in LCAP and 1800 mL/session in GMA regardless of patient body weight (BW). Recently, we have published the first report for evaluating this point. We conducted open label prospective trials for evaluating the clinical response of BW-adjusted LCAP (BWA-LCAP)49 and GMA (BWA-GMA).52 The results showed that the average Pv in the BWA-LCAP, which was determined as 30 mL/kg × BW (1971 ± 330 mL) per session, provided significant improvements in both the clinical and endoscopic disease activity of UC. Further, these scores after 10 weekly sessions were not significantly different between the BWA-LCAP group and the conventional fixed 3000 mL/session group. However, a significantly higher incidence of adverse event was observed in the 3000 mL LCAP group as compared with the BWA-LCAP group (P < 0.01).49

Conversely, in order to determine the optimum Pv for GMA, 33 UC patients were successfully induced to remission with five weekly GMA sessions at a standard Pv of 1800 mL, and then divided into three groups according to their BW; high body weight (HBW) (≥ 65 kg, n = 11), 50 kg ≤ middle body weight (MBW) < 65 kg (n = 12), and low body weight (LBW) (≤ 50 kg, n = 10). The results indicated that, by the clinical activity index for UC, a significantly higher remission rate was achieved in the LBW (80%) versus MBW (33%) or HBW (27%) at 6 weeks after beginning weekly GMA (P < 0.03). Therefore, we have reported that the lower-limit of optimum Pv/kg should be higher than 39 mL/kg per session for BWA-GMA.52

Recently, Yoshimura et al. reported that GMA could achieve a significant higher clinical efficacy by up to twofold higher processed volume (≥ 60 mL/kg) without any safety concerns.53 We have to optimize the optimum Pv for GMA by adjusting Qt since Qb should be 30 mL/min because of its adsorption mechanism. However, then, we have to consider patient patience for longer Qt because the clinical performance of GMA should be inverse proportion to Qt.

Figure 4 is our hypothesized optimum therapeutic regime of CAP for active UC patients according to the obtained evidence. The optimum Pv of a single CAP session for UC patients should be 30 mL/kg in LCAP and 40 mL/kg in GMA. On the other hand, since established evidence indicates that both functional suppression of circulating leukocytes and the quantitative removal of activated leukocytes contribute to the efficacy of this non-pharmacological therapy,11 it has been hypothesized that there might be an inverse proportion between Qf and immunological effect of CAP. Therefore, twice or more a week procedure of GMA and LCAP (intensive GMA or LCAP) has been recommended, especially for patients experiencing a severe flare. Although there is not sufficient evidence obtained from CD patients, the optimum procedure condition for them should presently be the same as for UC. Nationwide multicenter trials have be planned and started so as to test this hypothesis.

Figure 4.

Our hypothesized therapeutic regimen of extracorporeal leukocytapheresis (CAP) for active ulcerative colitis (UC) patients is shown. According to the result from a Japanese nationwide multicenter trial,11 therapeutic frequency should be fixed at twice (or more) a week. In addition, the procedure volume (Pv) has to be adjusted in each case according to their body weight (BW). We propose the optimum Pv values for filtration leukocytapheresis (LCAP) and granulocyte/monocyte apheresis (GMA) are 30 mL/kg per session49 and 40 mL/kg per session,52 respectively (this figure indicates a UC case with 60 kg BW).

Extracorporeal leukocytapheresis for a long-term maintenance therapy.  In the clinical setting, there is a need to establish an effective therapeutic strategy for long-term maintenance of remission without compromising safety. Further, it is reasonable if one can work with a strategy that is very effective as remission induction therapy and then use the same intervention as maintenance therapy as well. This approach has been used with Infliximab.54 CAP has the potential to achieve these intentions. There is evidence to support the clinical efficacy for monthly CAP as an adjunct maintenance therapy in UC patients with steroid-refractory background.3 With this background in mind, we have designed a prospective, single centre, randomized, sham-controlled, double-blind one-year trial with three arms to see if monthly GMA can suppress UC relapse in a population of patients who had achieved remission with a series of weekly GMA sessions.55 At week 48, the avoiding relapse rates (%AR) in True, Sham, and Control were 40%, 9.1%, and 18%, respectively. Interestingly, in patients who could taper their PSL dose to < 20 mg/day during remission induction, the %AR in True was better versus Shan (P < 0.03) or Control (P < 0.05). Future multicenter trials in large cohorts are needed to further strengthen our concept.


The first decade has passed since CAP became accepted by the Japanese social health insurance policy. During this period, several lines of evidence for understanding the therapeutic mechanism of this unique strategy have been obtained from several sites around the world from both clinical and basic science/disease mechanism standpoints. The etiology of IBD is far from fully elucidated; therefore, immunosuppressive therapy, including biologics, has shared the main part of therapeutic strategy in the Western world to control this intestinal disorder. By comparison, CAP stands out has having both effectiveness and safety, the balance of which could be favorable compared with pharmacological/biomodulator approaches. Since a strong advantage of CAP should be its safety characteristics, future strategies with56 and without combining biologics to improve the quality of life of IBD patients should now be studied.


The authors are grateful to Asahi Kasei-Kuraray Medical and JIMRO for providing fine photos. Also, we should like to thank Dr Abbi R Saniabadi of JIMRO for providing beautiful artwork for this contribution. The authors have no conflict of interest in connection with the publication of this manuscript.