Both of the calcineurin inhibitors, cyclosporine and tacrolimus, have been shown to be effective for refractory UC [3-7,12,14]. Whereas cyclosporine has to be administered intravenously due to the variable nature of blood concentration, tacrolimus can be used by peroral administration, usually twice a day, with sufficient efficacy. It has been reported that the optimal trough concentration of 10–15 ng/mL should be attained ‘quickly and safely’ for successful treatment. However, the doses required for reaching the target trough level differ largely among patients, and therefore, management of the trough concentrations is often difficult. Delay in reaching the target trough level would also delay treatment response, and may decrease the efficacy of the drug, whereas too rapid increase in blood concentration could cause adverse events. In this study, to ease the administration of tacrolimus, we looked for markers to predict the dose required for achieving the target blood concentration among materials already available in actual clinical practice.

It is known that tacrolimus is metabolized primarily by CYP3A4 and CYP3A5, and is converted to a substrate of ABCB1. Hirai et al. previously reported that a SNP of CYP3A5 affected pharmacokinetics of tacrolimus and the presence of the CYP3A5 *1 allele was a predictor of requirement of high dose administration. Moreover, they reported that patients without the CYP3A5 *1 allele were more likely to enter remission with tacrolimus therapy. In line with this report, we also showed that the SNP of CYP3A5 affected pharmacokinetics of tacrolimus and predicted the requiring dose of the drug to some extent. However, examination of SNPs prior to prescription of the drug is not practical due to the cost, unavailability at hospitals, and ethical problems. Hence, we meticulously analyzed the disease course and drug trough concentrations of the patients who were treated with tacrolimus, and found that the trough concentration measured 2 or 3 days after starting the treatment is a more precise predictor of dose requirement of the drug for remission induction therapy. This predictor is very simple and can be used without any other examinations or parameters. For instance, if a patient with body weight of 60 kg started tacrolimus therapy with 6 mg/day, and the trough concentration measured 2 or 3 days after initiation was < 6 ng/mL, the dose should be increased to 12 mg/day. On the other hand, if the trough concentration measured 2 or 3 days after initiation was > 6 ng/mL, the increase of the dose should be minimized (Figure 4). Thus, our results would be very useful in adjusting administration of tacrolimus for remission induction therapy in UC patients.

Patients with the CYP3A5 *1 allele are rich in the CYP3A5 protein and can metabolize tacrolimus more efficiently. Therefore, the results regarding the requirement of a higher dose to reach the target trough concentration in patients with CYP3A5 *1 allele appeared to be consistent and corroborative with those of the previous report by Hirai et al [12]. In contrast to the previous report, however, the remission rate within 4 weeks did not differ significantly between the two groups (13/24 (54%) vs. 10/13 (77%), p = 0.29). The discrepancy could be attributable to the following reasons.

First, the patients in the report by Hirai et al. had more severe disease than our cohort; all of their patients were hospitalized, and the average CRP appeared to be more than 5 mg/dL. In contrast, 14/47 (30%) of our patients were outpatients, and the median CRP of our patients was 1.4 mg/dL. Perhaps due to the difference in severity, the total remission rate was higher in our study than in the previous study (30/47 (64%) vs.14/45 (31%)). The higher remission rate of our study was probably one of the contributing factors for lack of statistical significance between patients with and without the SNP. Second, the frequency of adjustment of tacrolimus doses may be different. We measured trough concentrations of tacrolimus once every 2–3 days, while Hirai et al. appeared to measure the trough concentration only twice, on days 2–5 and 7–10, during the early period of therapy. The frequent adjustment of the dose of tacrolimus may have increased the remission rate of our patients. Whether or not our easy method for predicting the dose requirement of tacrolimus is valid in more severe UC patients should be verified in future studies.

It is expected that our results would make the handling of tacrolimus easier and would expand the use of the drug in the clinical situation. Because the treatment tools for refractory UC are still limited, tacrolimus could be a user-friendly option that would impact greatly on the clinical practice for refractory UC patients in several manners. First, while another calcineurin inhibitor, cyclosporine, requires continuous intravenous infusion, tacrolimus can be taken orally, indicating that the drug can be prescribed for outpatients. Our results would make the trough management of tacrolimus therapy for outpatients easier and reduce the frequency of clinic visits, possibly resulting in an increase in the success rate of the treatment. Second, in current clinical situations, anti-tumor necrosis factor alpha antibodies such as infliximab and adalimumab are usually used for refractory UC outpatients. Although such agents have also been shown to be effective for refractory UC patients [17,18], those are injected agents, thus, patients have to visit a clinic or hospital periodically for infusions or perform self-injection. In contrast, tacrolimus has an advantage that it is a peroral drug and easily acceptable for every type of patients. Third, appropriate management of the blood concentration of tacrolimus would reduce adverse events due to the drug. The relatively frequent adverse events of tacrolimus are nausea, headache, tremor, and renal dysfunction. Those events usually occur when blood concentrations of the drug are too high. In particular, nausea and headache are likely to occur when the blood concentration abruptly increases. Reduction of adverse events by appropriate monitoring of the blood concentration as per our results would also lead to increased success of the tacrolimus treatment.

The starting dose of tacrolimus for UC used in the previous phase III study was 0.05 mg/kg/day [6,14]. In this study, the average initial dose of tacrolimus was 0.10 mg/kg/day, although there was variation among patients due to the retrospective cohort. As shown in Figure 1, the dose required for the target trough level was more than 0.10 mg/kg/day in 91% of patients, thus, the initial dose of 0.10 mg/kg/day would be reasonable. In this situation, however, patients of the low dose group more frequently developed adverse events (23%) mostly due to excessive trough concentrations, although most of these events were mild. Therefore, our findings regarding the prediction of requirement of low or high dose within 2 or 3 days after starting tacrolimus would be helpful in view of avoidance of adverse events.

There are limitations to the present study. Because this was a retrospective clinical study, the intervals of trough measurement differed among patients. The initial trough concentration was not measured at 2 or 3 days after start of tacrolimus in several patients, and those patients were not included in the corresponding analyses. Trough management based on our results should be validated in prospective studies in the future. The SNP analysis also could not be performed in a few patients who did not give approval for the analysis.