Concomitant pharmacologic medications influence the clinical outcomes of granulocyte and monocyte adsorptive apheresis in patients with ulcerative colitis: A multicenter retrospective cohort study

Granulocyte and monocyte adsorptive apheresis (GMA) with Adacolumn has been used as a remission induction therapy for patients with active ulcerative colitis (UC). Herein, we investigated the influence of concomitant medications in the remission induction of GMA in patients with active UC.


| INTRODUCTION
Ulcerative colitis (UC) is an idiopathic, chronic inflammatory disorder of the colonic mucosa and is characterized by alternating periods of relapse and remission. 1 Therefore, patients with UC require life-long medications, including immunosuppressive therapy.
Granulocyte and monocyte adsorptive apheresis (GMA) with an Adacolumn (JIMRO Co., Takasaki, Japan) has been applied as a nonpharmacological treatment strategy and is widely used in the remission induction therapy for patients with active UC in Asia and European Union. The mechanism underlying GMA involves the Adacolumn, which is filled with cellulose acetate (CA) beads, interacting with fragment crystallizable-gamma receptor (FcγR) expressed at the surface of activated leukocytes and selectively adsorbing granulocytes and monocytes from the systemic circulation. 2 Numerous reports have described the clinical efficacy and safety of GMA in patients with active UC. [3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20][21][22] Previous reports have demonstrated that patients who were most likely to respond to GMA tended to be first attack cases and were steroid-naïve with short disease duration and low disease activity. [9][10][11][12][13][14] However, GMA has been reported to enable the evaluation of clinical efficacy for refractory UC such as steroid-dependent UC and immunosuppressor (IM)-resistant UC. 6,15,16 Moreover, recent studies have described the combined efficacy of GMA with biologics 17,20 and tofacitinib 21 in patients with refractory UC. In addition, the combination of GMA with antitumor necrosis factor (TNF) alpha agents has decreased clinical activity and biomarkers even in cases with lost responses to antiTNF-alpha agents. 22 Yokoyama et al. demonstrated that patients with inflammatory bowel disease (IBD) experiencing a loss of response to infliximab appeared to regain clinical response to infliximab after undergoing GMA, as shown by a decrease in the antibody-to-infliximab level. 23 However, no clinical study has compared each concomitant medication in GMA therapy.
Thus, GMA with Adacolumn is considered to potentially exert sufficient effectiveness not only in mild active UC cases, but also in refractory cases, such as those dependent on steroids, those resistant to IMs and/or biologics, and even those that lost response to biologics. However, an appropriate GMA treatment strategy related to concomitant medication in patients with active UC has not been established. In this retrospective multicenter cohort study, we investigated whether concomitant pharmacologic medications influence the clinical outcomes of GMA with Adacolumn in patients with active UC.

| Study design and ethics
This multicenter retrospective cohort study included all patients with active UC who underwent GMA with Adacolumn in five independent institutions in Hokkaido, Japan, from January 2011 to July 2021. These institutions included Nayoro City General Hospital, Engaru Kosei General Hospital, Asahikawa Kosei General Hospital, Asahikawa City Hospital, and Asahikawa Medical University.
The research protocol was reviewed and approved by the ethics committees of Asahikawa Medical University (approval no. 21004) and of each institution in accordance with the Helsinki Declaration (1964, and later versions).

| GMA treatment strategy with Adacolumn
GMA with Adacolumn (JIMRO Co., Takasaki, Japan) is approved as a remission induction therapy by the Japan Ministry of Health for all patients with active UC. Evidencebased clinical practice guidelines for IBD 2020 have been published by the Japanese Society of Gastroenterology. 27 The treatment strategy for IBD in our five institutes is in accordance with this guideline. In these guideline, 5-ASA and corticosteroid are recommended as a basic medication and GMA is mostly recommended for patients with steroid dependent and/or refractory disease. In our institutes, GMA is used for patient with active UC with both 5-ASA insufficiency and/or intolerance and steroid dependent and/or refractory disease. Moreover, these guidelines recommend not only GMA, but also IMs, calcineurin inhibitors, and biologics for patients with steroid dependent and/or refractory disease. In our institutes, monotherapy or combination therapy is selected based on physicians' judgements for patients with steroid dependent and/or refractory disease. In this study, patients received 5-ASA, corticosteroids, IMs, biologic or calcineurin inhibitors at the time of admission, and the dose and frequency of these medications were adjusted based on physicians' judgements in each institute. Medications initiated during GMA were not included as concomitant medications. When a patient's condition worsened or remained unchanged, GMA was canceled and another therapy including corticosteroids, IMs, biologics, or surgery was initiated. Maintenance therapy, including 5-ASA, IMs, and biologic, was administered after GMA in accordance with these guidelines. In our institutes, patients with maintenance of remission receive monotherapy or combination therapy based on physicians' judgement. In this study, patients receiving 5-ASA, IMs, or biologic after GMA were considered as having received maintenance therapy. The contraindication of GMA in our institutes were neutrophil count <1500/mm 3 ; hemoglobin <10 g/dL; a history of allergic reaction to anticoagulant; or a serious cardiac, pulmonary, hepatic, or renal disorder.
The standard GMA treatment plan in our institutes involved 10 sessions with an Adacolumn twice a week for 5 consecutive weeks. In each patient, the blood via venipuncture of an antecubital vein entered the Adacolumn and then returned to the patient via the column outflow line. The GMA regimen consisted of a standard protocol, filtering 1800 mL per session at a rate of 30 mL/min. All GMA data including total number of sessions, frequency, amount of filtering, and adverse events, were collected from medical records.

| Endpoints and definitions
Clinical outcomes were assessed within 7 days after GMA (in other words, within 6 weeks from the beginning of GMA) and at 52 weeks thereafter. The primary endpoint was the influence of concomitant medication on the CR after GMA. The secondary endpoints were the influence of maintenance therapy on the sustained CR after GMA up to 52 weeks. To clarify these endpoints, the clinical factors associated with the CR and endoscopic remission (ER) after GMA and up to 52 weeks were statistically investigated. Patients were included even if the clinical information needed for this study was missing at a certain point.
CR was defined as PMS of ≤2 and ≤1 for all subscores. Relapse was defined PMS of ≥3. ER, which refers to mucosal healing, was defined as an MES 0 or 1. 5-ASA intolerance was defined as discontinuation of 5-ASA because of allergy or adverse events. A primary nonresponse to biologics occurs when a patient did not respond well to biologics during the remission induction therapy. Loss of response to biologics was defined as relapse despite continuous administration of biologics in the maintenance therapy.

| Statistical analyses
Numerical data are presented as medians with minimummaximum or interquartile ranges (IQR). The comparison of demographic characteristics between the CR and nonCR groups was performed using the Mann-Whitney U-test or chi-squared test. Factors associated with CR and ER after GMA were evaluated using a univariate logistic regression analysis. Relevant demographic features with P < .10 in the univariable analysis were entered into the multivariate model to identify independent predictors of CR and ER. Sustained CR was estimated by the Kaplan-Meier estimator graphs and the log-rank test. Post GMA demographic parameters that affect the duration of sustained CR were assessed by Cox proportional hazards model. Relevant demographic features with P < .10 in the univariable analysis were entered into the multivariate model to identify independent predictors of sustained CR; P < .05 was considered significant. Each statistical analysis was performed using the SPSS Statistics for Windows (SPSS Inc., Chicago, Illinois).

| Baseline characteristics of all patients
A total of 133 patients with active UC who had undergone GMA with Adacolumn were included in this retrospective cohort study. Details of these 133 patients were obtained. Of these 133 patients, 46 were from Asahikawa Medical University, 67 from Asahikawa City Hospital, 13 from Asahikawa Kosei General Hospital, 5 from Engaru Kosei General Hospital, and 2 from Nayoro City General Hospital. The annual change in GMA from 2011 to 2021 is shown in Figure 1 4 This large cohort enrolled 697 patients from 53 medical institutions between 1999 and 2006. The frequency of GMA therapy for active UC in this large cohort was similar to that in our retrospective cohort study. Therefore, we consider that the number of patients in our study was based on real world clinical data in Japan and had a large enough scale for analysis.
3.2 | Overall clinical efficacy and safety of GMA GMA was discontinued in 2 patients because of severe adverse events, and the remaining 131 patients were analyzed in this study. Clinical assessment was performed on all 131 patients. Of these 131 patients, 74 achieved CR and 57 did not achieve CR (nonCR) after GMA. The overall CR rate was 55.6%. Furthermore, endoscopic assessment was performed in 107 patients (81.6%) within 7 days after GMA. Of these 107 patients, 41 achieved ER and 66 did not achieve ER (nonER) after GMA. The overall ER rate was 38.3%. In patients with CR, 35 patients achieved ER and 23 patients did not achieve ER, and in patients with nonCR, 6 patients achieved ER and 43 patients did not achieve ER. (Figure 2) Furthermore, the CR rates in patients who underwent GMA concomitant with 5-ASA, corticosteroids, IMs, biologics, and calcineurin inhibitor were 61%, 62.5%, 38%, 46.5%, and 87.5%, respectively.
In this study, 11 patients (8.2%) experienced adverse events during GMA. All adverse events are shown in Table 2. In nine patients (6.7%), adverse events were mild, and they could continue all GMA sessions with symptomatic therapy. However, in two patients (1.5%), GMA was discontinued after the first GMA session because of severe headache and/or high fever. These two patients were recovered soon after the GMA was canceled without any trouble. No adverse events were reported after GMA.

| The overall number of GMA sessions
Patients were divided into the CR and nonCR groups following the assessment of the clinical activity based on the PMS after GMA. All GMA data including the total number of sessions, frequency, and amount of filtering, are summarized in Figure 3. The median total number of GMA sessions (min-max) was 10 (3-10). Regarding the GMA regimen, 125 patients (95.4%) received GMA twice per week and 129 patients (98.4%) received filtering of 1800 mL per session at a rate of 30 mL/min. No Loss of response for antiTNF-alpha agent N (%) 29 (21.8) significant difference was found in the frequency and total amount of filtering between the CR and nonCR groups. The total number of GMA sessions depended on the disease activity of UC, patient tolerance, and response to GMA therapy. Therefore, we did not evaluate the impact of the total number of GMA sessions on the clinical efficacy of GMA. In the CR group, 46 patients (62.1%) completed all 10 sessions of GMA and the remaining patients were withdrawn from GMA because they achieved CR before all 10 sessions were completed. In the nonCR group, 25 patients (43.8%) completed all 10 sessions of GMA, the GMA was canceled for the remaining patients because their conditions could not improve.

| Identification of clinical factors associated with CR after GMA
The clinical factors associated with CR after GMA are shown in Table 3. In the univariable analysis, an MES of 2 and concomitant medications with 5-ASA were significantly associated with higher CR rates (P = .026 and P = .007, respectively); furthermore, concomitant medications with IMs, and loss of response to antiTNF-alpha agent were significantly associated with lower CR rates (P = .004 and P = .042, respectively). In the multivariable analysis, an MES of 2 and concomitant medication with 5-ASA remained as significantly independent positive factors (P = . after GMA are shown in Table 4. In the univariable analysis, concomitant medication with 5-ASA was significantly associated with higher CR rates (P = .085). Furthermore, concomitant medication with IMs and antiTNF-alpha agents and loss of response to antiTNF-alpha agents were significantly associated with lower CR rates (P = .010, P = .082, and P = .058, respectively). In the multivariable analysis, concomitant medication with IMs remained as a significant independent negative factor of CR after GMA (P = .042, OR 0.354, 95% CI 0.130-0.962).

| Comparison of demographic variables between patients with IMs in the CR and nonCR groups
We compared the body weight and laboratory data in patients with concomitant medication with IMs, as they may have a potential to influence the effectiveness of GMA. The result is shown in Table 5. No significant difference was found in the body weight, WBC count, lymphocyte count, MCV, TP, ALB, and IgG between the groups; however, the CRP value was significantly higher in the nonCR group (P = .021).

| Identification of clinical factors associated with ER after GMA
The clinical factors associated with ER after GMA are shown in Table S1. In the univariable analysis, patient with first attack had significantly higher ER rates (P = .017). Other clinical factors and concomitant medications were not associate with ER after GMA.

| PostGMA characteristics of patients with CR
A total of 74 patients with active UC who underwent GMA achieved CR. The postGMA characteristics of the 74 eligible patients are shown in Table 6

| Rate of patients sustained CR after GMA
Seventy-four patients who achieved CR after GMA were followed up for 52 weeks. CR was sustained in 24 of the 74 patients. The proportion of patients in remission was 41.6% and the median remission time was 30 weeks ( Figure 4A).

| Clinical factors associated with sustained CR after GMA
The patients were divided into the CR and relapse groups following the assessment of the clinical activity based on the PMS after GMA up to 52 weeks. Clinical factors associated with sustained CR after GMA are shown in Table 7.
In the univariable analysis, WBC count and maintenance therapy with IMs were significantly associated with a higher sustained CR rate (P = .034 and P = .023, respectively). In the multivariable analysis, maintenance therapy with IMs remained as a significant independent positive factor of sustained CR up to 52 weeks (P = .038, OR 2.214, 95% CI 1.046-4.638, respectively).

| Rate of sustained CR among patients with IM after GMA
CR was sustained in 14 of the 27 patients with IMs, and the proportion of sustained CR in patients with IMs was significantly higher than that in patients without IMs (64.1% vs 28.8%, P = .017). ( Figure 4B) We also evaluated the effectiveness of IMs on maintenance therapy with 5-ASA or biologics. CR was sustained in 11 of the 19 patients with 5-ASA and IMs and the proportion of sustained CR in patients with 5-ASA and IMs was not different when compared with that in patients with 5-ASA only (60.7% vs 37.6%, P = .248). ( Figure 4C) CR was sustained in 6 of the 8 patients with biologics and IMs, and the proportion of sustained CR in patients with biologics and IMs was significantly higher than that in patients with biologics only (72.9% vs 0%, P = .002). ( Figure 4D) Furthermore, CR was sustained in 4 of the 5 patients with primary nonresponse to biologics and IMs, and the proportion of sustained CR in patients with primary nonresponse to biologics and IMs was significantly higher than that in patients with primary nonresponse to biologics only (80.0% vs 0%, P = .043). ( Figure 4E) CR was sustained in 2 of the 3 patients with loss of response to biologics and IMs, and the proportion of sustained CR in patients with loss of response to biologics and IMs was significantly higher than that in patients with loss of response to biologics only (66.7% vs 0%, P = .041) ( Figure 4F).

| DISCUSSION
In this multicenter retrospective cohort study, we evaluated whether concomitant pharmacologic medications influence the rates of CR and sustained CR in patients with active UC who underwent GMA with Adacolumn.
To the best of our knowledge, this is the first study to demonstrate that concomitant medication with IMs was a negative factor for the remission induction therapy by GMA. Moreover, maintenance therapy with IMs after GMA was a positive factor for sustained CR, especially in patients treated with biologics. In this study, patients with moderately active and refractory UC were largely included based on the baseline  with primary nonresponse to biologics and IMs and in 0 of 5 patients (0%) with primary nonresponse to biologics only (P = .043). The rate of sustained CR in patients with primary nonresponse to biologics and IMs was significantly higher than that in patients with biologics only (P = .043). (F) CR was sustained in 2 of 3 patients (66.7%) with loss of response to biologics and IMs and in 0 of 7 patients (0%) with loss of response to biologics only. The rate of sustained CR in patients with loss of response to biologics and IMs was significantly higher than that in patients with biologics only (P = .041) Adacolumn has sufficient effectiveness and a favorite safety not only in mild active UC cases, but also in moderate active and refractory cases. Furthermore, the CR rates in patients who underwent GMA concomitant with 5-ASA, corticosteroids, IMs, biologics, and calcineurin inhibitors were 61%, 62.5%, 38%, 46.5%, and 87.5%, respectively.
The clinical effectiveness of GMA and medications has been investigated by several study groups. Kakimoto et al reported that the CR rate in patients who did not receive corticosteroids was 25.4% in a prospective study. 14 The CR rate in refractory cases, such as steroid-dependent and IMresistant cases, was reported to be 35.7%-71% 6,10,11 and 36%, 15 respectively. Regarding biologics, Song et al reported that a combination therapy involving GMA and anti-TNF-alpha agents showed a higher CR rate than GMA monotherapy (63.3% vs 43.3%). 18 On the other hand, in patients with insufficient response/intolerance to biologics, the CR rate was reported to be 33%. 16 Moreover, in patients with loss of response to anti-TNF alpha agents, the response rate was reported to be 32%. 22 In our previous small prospective study, no significant difference in the ratio of concomitant medications, such as 5-aminosalicylic acid, corticosteroids, immunosuppressants, and biologics, was observed between the CR and nonCR groups. 28 Furthermore, past exposure to corticosteroids or biologics has been demonstrated as a risk factor for CR in several studies. [11][12][13] On the contrary, Matsuda et al reported that the effect of GMA with concomitant corticosteroids and that of GMA without corticosteroids were not different, and GMA was effective irrespective of corticosteroid administration. 29 The results of numerous previous suggest that the clinical effectiveness of GMA for patients with UC, especially refractory cases, shows wide variability, and the impact of past exposure and concomitant corticosteroids on the clinical effectiveness of GMA is still controversial. This restrictive clinical effectiveness of GMA might be influenced by various clinical factors, including a variety of concomitant medications during GMA. However, no clinical study has compared each concomitant medication during GMA therapy.
In the multivariable analysis of factors associated with CR after GMA, concomitant medication with 5-ASA or IMs and MES at the start of GMA were demonstrated as predictors of CR after GMA. In a large cohort study, Yamamoto et al. reported that patients with moderately to severely active UC who had severe endoscopic activity did not respond well to GMA. 11 Thus, the endoscopic activity, with deep mucosal lesions and extensive loss of the mucosal tissue, is already known as a negative factor of CR after GMA. 8,11 To evaluate the influence of concomitant medications, patients with severe endoscopic activity (MES of 3) that decreases the induction rate of GMA were excluded, and the subgroup of only patients with moderate endoscopic activity (MES of 2) was additionally investigated. In the multivariable analysis, concomitant medication with IMs was only demonstrated as a significant independent negative factor of CR after GMA. On the other hands, in univariable analysis of factor associated with ER after GMA, patient with first attack were demonstrated as predictors of ER after GMA. Concomitant medications were not associate with ER after GMA. In this study, the predictor for ER was different for patients with CR. In previous reports, the predictors for endoscopic efficacy were similar to those for clinical efficacy. 11,12 Our previous report has showed that when the FC value is ≤1150 mg/kg after GMA, patients had a high probability of achieving ER within 24 weeks. 28 In this study, the median FC value was 469 mg/kg in patients with CR; however, only 35 patients (60.3%) with CR who underwent endoscopic assessment achieved ER. Endoscopic assessment in this study, which was performed within 7 days from the last GMA, might have been conducted earlier than the appropriate timing of the mucosal healing stage.
In GMA, Adacolumn is filled with CA beads. The mechanisms of disease modification by CA beads, IgG, and C3-derived active complement fragments mediated leukocyte adhesion to CA beads. These plasma proteins on the CA beads interact with FcγR and/or leukocyte complement receptor-like CR3 expressed at the surface of activated granulocytes and monocytes. Thus, CA beads selectively adsorb granulocytes and monocytes from the systemic circulation. [30][31][32] Activated granulocytes and monocytes absorbed on the CA beads degranulate and induce apoptosis in the Adacolumn. Consequently, GMA would decrease the number of neutrophils in the intestinal mucosa 33 and macrophages through the decline of proinflammatory monocytes in the peripheral blood. 34,35 Furthermore, apoptotic neutrophils that returned to the systemic circulation from Adacolumn are processed in secondary lymphoid organs, such as the spleen 36 and phagocytosed by CD19(+) B-cells. CD19 (+) B-cells, which phagocytosed the apoptotic neutrophils are differentiated into regulatory B-cell and inhibit the activated lymphocytes by increasing the production of the immunosuppressive cytokine interleukin-10. 37,38 Regulatory T-cells are induced by dendritic cells that capture these apoptotic neutrophils. 39,40 These regulatory T-cells migrate into the inflammation site and then modulate the immune response by suppressing lymphocyte activation. 41 Thus, GMA is considered to exert a long-term anti-inflammatory effect in patients with active UC. On the contrary, IMs, including azathioprine and a prodrug of 6-mercaptopurine (6-MP), are used to maintain remission and reduce antidrug antibody formation in the monoclonal antibody therapy of patients with IBD. Orally administered IMs are absorbed from the gut and metabolized nonenzymatically to 6-MP in the body. 6-MP is eventually metabolized to 6-thioguanosine triphosphate (6-TGTP) and 6-thiodeoxyguanosine diphosphate (6-TdGDP). 6-TGTP is incorporated into the RNA, and 6-TdGDP is incorporated into the DNA, inhibiting RNA transcription and DNA replication, respectively, and leading to apoptosis. 42 6-TGTP also causes apoptosis of lymphocytes by inhibiting GTPase Rac1. 43 Thus, IMs strongly exert immunosuppressive effects on patients with active UC. As mentioned above, a similar anti-inflammatory effect between GMA and IMs might cause a lower induction rate in patients who underwent GMA concomitantly with IMs.
The measurement method of the concentration of IMs is not established; therefore, we always adjust the dosage of IMs while referring to the WBC count, lymphocyte count, and MCV value in clinical practice. In this study, concomitant medication with IMs was demonstrated as a negative factor for remission induction by GMA. Regarding the comparison between patients with IMs in the CR and nonCR groups, no significant difference was found in body weight, WBC count, lymphocyte count, MCV, TP, ALB, and IgG. However, the CRP value was significantly higher in the nonCR group. In this study, the medications initiated during GMA were not included as concomitant medications; therefore, IMs were continued before the start of GMA. Patients with IMs in the nonCR group might have decreased WBC count; nevertheless, the CRP value increased, because of the sufficient immunosuppressive effect of IMs. In this study, 42 patients have undergone GMA concomitant with IMs, and the CR rate in patients with IMs was 38% (16/42). In a previous clinical study, the CR rates in patients with resistance to IMs were 36% and 34%, respectively. 15,16 Thus, the clinical efficacy of GMA in patients with IMs in our study is not lower than that of patients in a previous clinical report. However, in our study, GMA that was conducted concomitantly with IMs was significantly associated with a lower CR rate than GMA with other concomitant medications. Thus, patients who experienced relapse under treatment with IMs did not respond well to GMA compared with those using other concomitant medications because of a similar anti-inflammatory effect that restrained the inflammatory immune response through the suppression of activated lymphocytes between GMA and IMs.
Moreover, in this study, we investigated the clinical factors associated with sustained CR up to 52 weeks. CR was sustained in 24 of 74 patients, and the median remission time was 30 weeks. In the multivariable analysis of factors associated with sustained CR up to 52 weeks in this study, maintenance therapy with IMs was demonstrated as a significant independent positive factor of sustained CR up to 52 weeks. The rate of sustained CR in patients with IMs was significantly higher than that in patients without IMs. However, IMs are usually used to maintain remission and reduce antidrug antibody formation in monoclonal antibody therapy in patients with IBD. Therefore, the additional effectiveness of IMs on maintenance therapy with 5-ASA or biologics was investigated. As a result, the additional effectiveness of IMs was confirmed especially in patients with maintenance therapy using biologics, in both cases of primary nonresponse and loss of response to biologics. Ishiguro et al. reported that GMA was effective for patients who are corticosteroid-naïve, and the efficacy was sustained longer in those not receiving IMs during GMA. 12 This prospective study included patients with UC who had received GMA as remission induction therapy after showing an inadequate response to 5-ASA preparations as first-line medications. Moreover, the rate of sustained CR in patients with 5-ASA and IMs was also not different from that in patients with 5-ASA only. These results indicate that IMs have no additional effects on 5-ASA preparation as maintenance remission therapy after the remission induction of GMA. On the contrary, Panaccione et al. demonstrated that patients with antiTNF alpha agent-naïve UC who received infliximab and azathioprine were more likely to achieve corticosteroid-free remission than those receiving either monotherapy because IMs are considered to reduce the antidrug antibody formation in monoclonal antibody therapy in patients with IBD. 44 Even in the GMA therapy, apoptotic neutrophils that returned to the systemic circulation from Adacolumn activate the differentiation of regulatory B-cells and inhibit the activated lymphocytes. 37,38 Yokoyama et al. demonstrated that an antibody-to-infliximab level in the serum has decreased after GMA therapy in patients with IBD experiencing a loss of response to infliximab. 23 In our study, all patients with biologics as a maintenance therapy without IMs were in relapse within 26 weeks; therefore, we assumed that the clinical effectiveness through the reduction of antidrug antibody by GMA would be attenuated after GMA. Thus, we consider that the maintenance therapy with biologics and IMs after GMA well sustained a CR because of a similar immunosuppressive effect that reduces the production of antidrug antibodies from lymphocytes between GMA and IMs.
The strength of this study is attributed to the analysis of large data from multiple facilities, and it is the first to compare concomitant medications in GMA. However, it was associated with some limitations. Specifically, it was a retrospective cohort study, and patients whose clinical information needed for this study was missing for a certain point and each number of concomitant medications is low were also included. Moreover, concomitant medication was collected only the medications included only the medications administered at the time of admission; the medications initiated during GMA were not included. In addition, the dose, and the frequency of each concomitant medication during GMA were not investigated. Further investigations with a prospective observation study are needed to establish the therapeutic strategy.

| CONCLUSION
To the best of our knowledge, this is the first study to demonstrate that concomitant medication with IMs was a negative factor for the remission induction therapy by GMA, and the maintenance therapy with IM after GMA was a positive factor for sustained CR, especially in patients treated with biologics. We suggest that GMA was more effective on patients with active UC who were in relapse despite treatment without IM; furthermore, the addition of IMs should be considered in patients under maintenance therapy with biologics after GMA. This therapeutic strategy may contribute to making better decisions in the clinical treatment of patients with active UC.

ETHICS STATEMENT
This multicenter retrospective cohort study was conducted by Nayoro City General Hospital, Engaru Kosei General Hospital, Asahikawa Kosei General Hospital, Asahikawa City Hospital, and Asahikawa Medical University. The research protocol was reviewed and approved by the ethics committees of Asahikawa Medical University (approval no. 21004) and of each institution in accordance with the Helsinki Declaration (1964, and later versions).

PATIENT CONSENT STATEMENT
This study is a retrospective observational cohort study and does not necessarily require written informed consent because it is a study, which uses only information such as medical records without using samples. By disclosing information related to clinical research on our website, we will provide the subject with the opportunity to withdraw from using their own information.