Active UC is debilitating, affects function and quality of life. The chronic nature of the disease means patients require medications throughout life. Conventional drugs often cause adverse side effects, adding to the disease complexity. It would be reasonable to assume that one major factor for physicians and patients opting for a non-pharmacological treatment intervention like GMA is to avoid adverse side effects of drugs. Ideally, physicians and in particular, patients hope treatment to be safe and effective. However, the reality of clinical practice would show that any given therapeutic intervention is effective only in a fraction of patients treated similarly for the same condition. The size of the responder fraction reflects the efficacy rate. In Japan and Europe, GMA with an Adacolumn in patients with UC has been applied as a non-pharmacological treatment strategy, but hitherto, the efficacy rate has been different, ranging from an 85% [9, 12–14, 25] to a statistically insignificant level . Treatment failure reflects waste of resources and an increase in morbidity time for the non-responder patients. However, there is evidence that the clinical response to GMA is dictated by patients’ demographic variables at baseline [26, 27] which should be evaluated. With this background in mind, in this study, we looked for predictive factors, which could guide us to select future patients who are most likely to respond to GMA.
The outcome of the present investigation might be summarized as follows. Forty-three outpatients who had received consecutive weekly GMA at one session per week, up to 10 sessions for an active UC were retrospectively reviewed. All 43 patients had active disease in spite of receiving conventional medications, the majority on PSL. At the assessment time point, the patients could be divided into remission group and non-remission group. During GMA therapy, the PSL dose was tapered and many responders became steroid free. Typically, the non-responders had extensive deep ulcers with near total loss of the mucosal tissue at the affected sites or had a long duration of UC and exposure to multiple conventional medications. Further, almost all patients who maintained their clinical remission for 54 weeks had achieved mucosal healing during the course of GMA therapy. Therefore, mucosal healing was a predictive factor for better prognosis [28–30]. Our search for predictive factors of clinical response to GMA showed that at baseline, the interval between relapse and the first GMA session was significantly shorter in the remission group as compared with the non-remission group. Additionally, the cumulative PSL dose before the first GMA session was a significant independent factor, and a high cumulative PSL dose appeared to negatively impact the efficacy of GMA. The application of logistic regression analyses indicated that GMA at an early stage following a clinical relapse was likely to induce remission of an active UC. Another significant predictive factor was total WBC at baseline. However, the percentage of granulocyte in WBC had increased in all patients (data not shown).
The publication of the first clinical experience with GMA in patients with UC by Shimoyama, et al.  generated significant interest in this non-pharmacological treatment intervention for patients with inflammatory diseases associated with activated myeloid lineage leucocytes [5–19, 23–27, 31–38]. However, as already mentioned, there is no shortage of contrasting efficacy outcome reports in the literature. In 2004, Suzuki, et al.  reported an 85% efficacy rate or 17 of 20 patients they treated with GMA as the first line medication. Then a controlled trial by Sands, et al.  reported no statistically significant difference between the sham arm and the GMA arm. The patients included in the latter study had a very different demography, had not responded to conventional medications. In contrast, all of the patients in Suzuki’s study were steroid naïve, most with a short UC duration and only needed first line medication [9, 12]. Suzuki, et al.9 reported that the only 3 GMA non-responders in that study had defective mucosal tissue and deep ulcers. In a subsequent study, Suzuki, et al.  reported that all patients with a short duration of UC (about 3 months), and first episode cases readily responded to GMA, while all of the 8 GMA non-responders in that study had a very long duration of UC together with a long history of exposure to multiple conventional medications. The observations by Suzuki, et al. [9, 27] were independently confirmed by Tanaka and colleagues in much larger cohorts of UC patients [16, 17, 25].
In the present study, the frequency of GMA was one session per week at 30 mL/min for 60 min. This is equivalent to 1800 mL of processed blood volume per GMA session. This treatment protocol might be inadequate as Sakuraba, et al.  found that 2 GMA sessions per week produced significantly higher efficacy rate in a shorter time as compared with one session per week. Similarly, Yoshimura, et al.  reported that more than 3000 mL of processed blood volume per GMA session was significantly more effective than routinely applied 1800 mL per session. Additionally, our experience suggests that unlike drug based medication, GMA demands operational skills, and maintenance of a steady blood flow.