Upadacitinib Treatment Outcomes in Crohn’s Disease: A Retrospective Analysis of Clinical Response and Perianal Fistula Resolution

Abstract

Introduction

Upadacitinib (UPA), a Janus kinase inhibitor approved for the treatment of Crohn’s disease (CD), has demonstrated high efficacy in clinical trials; however, real-world data on its outcomes remain limited. This study evaluated UPA outcomes in patients with CD in a real-world setting.

Methods

We retrospectively analyzed patients who initiated UPA treatment between June 2023 and June 2024. The primary endpoints were clinical response and remission at 12 and 54 weeks. The secondary endpoints included changes in the C-reactive protein (CRP) levels, serum albumin levels, and perianal fistula response.

Results

Thirteen patients (mean age, 32.0 years; 92.3% male; mean disease duration, 114.1 months) were included. At baseline, three (23.1%) patients used oral corticosteroids, and 11 (84.6%) had undergone prior biologic therapy, with nine (81.8%) exposed to anti-tumor necrosis factor-alpha antibodies. The mean follow-up period was 53.1 weeks, and the 54-week continuation rate was 83.3%. The clinical response rates were 46.2% and 53.8%, and the remission rates were 69.2% and 84.6% at 12 and 54 weeks, respectively. The mean Crohn’s Disease Activity Index scores decreased from 226.8 at baseline to 73.7 at 54 weeks, and the average CRP and albumin levels improved correspondingly. Among the 7 patients (53.8%) with draining perianal fistulas, six (85.7%) achieved cessation of pus discharge by week 8 (mean, 4.4 weeks). No significant adverse events were observed.

Conclusion

UPA appears to be an effective treatment for CD and may facilitate the early resolution of perianal fistulas, supporting its role in managing perianal disease.

Introduction

Crohn’s disease (CD) is a chronic, debilitating inflammatory bowel disease characterized by transmural inflammation that can affect any part of the gastrointestinal tract, from the mouth to the anus [1]. The disease course is often unpredictable, characterized by periods of relapse and remission, resulting in significant impairment of patient quality of life (QOL) [2, 3]. Common complications, including strictures, fistulas, and perianal disease, often necessitate surgical intervention. Among these, anal fistulas are particularly common and challenging to manage, significantly impacting patient QOL [4].

While conventional therapies, such as aminosalicylates, corticosteroids, and immunomodulators, have been used for CD management, the advent of advanced therapies (ADTs), particularly anti-tumor necrosis factor-alpha (TNFα) antibodies, has revolutionized the treatment landscape for moderate-to-severe cases [5]. However, a substantial proportion of patients either exhibit primary non-response or secondary loss of response to these therapies, highlighting the persistent need for novel and effective treatment options [6, 7].

Janus kinase (JAK) inhibitors represent a new class of small-molecule drugs that target the intracellular signaling pathways of various proinflammatory cytokines implicated in CD pathogenesis [8]. Upadacitinib (UPA, Rinvoq^®), an orally administered drug, was approved in May 2023 as the first JAK inhibitor for CD in Japan [9]. Pivotal phase 3 clinical trials, such as U-EXCEL (NCT03345849), U-ENDURE (NCT03345836), and U-ACCOMPLISH (NCT03345823), have demonstrated the efficacy of UPA in inducing and maintaining clinical remission in patients who failed or were intolerant to conventional or ADTs [9].

While clinical trials have demonstrated high efficacy, real-world data remain limited, particularly in areas such as optimal patient selection and treatment timing [10, 11]. Real-world studies offer invaluable insights into treatment persistence, safety, and effectiveness outside controlled trial environments, thereby informing clinical decision-making and optimizing treatment strategies. Furthermore, data on the efficacy of UPA for anal fistulas remain scarce, despite the significant impact of these lesions on patient QOL [10, 12]. Therefore, in this study, we aimed to evaluate the use of UPA for CD in a real-world setting and its therapeutic effects on fistulizing perianal lesions. Specifically, we assessed the clinical response and remission rates, changes in biochemical markers, and response of perianal fistulas to UPA therapy.

Methods

Study Design

This single-center, retrospective, observational study was conducted at Nagasaki University Hospital and reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.

Patients

We retrospectively reviewed the medical records of patients with CD who initiated UPA therapy between June 2023 and June 2024. Patients were eligible if they had a confirmed diagnosis of CD and received UPA therapy. The only exclusion criterion was the lack of comprehensive clinical data. The observation period for each patient was defined as the interval from UPA initiation to the last available follow-up visit, resulting in variable follow-up durations depending on the treatment start time.

Treatment

All patients received UPA treatment. Dosage and treatment duration were determined by the treating physicians based on clinical guidelines and patient characteristics, as detailed in their medical records. According to published data [9] and the approved prescribing information, the standard induction dose is 45 mg, administered orally once daily for 12 weeks, followed by a maintenance dose of 30 or 15 mg, administered orally once daily. The use of concomitant medications, including oral corticosteroids, was recorded.

Data Collection

Clinical and laboratory data were retrospectively extracted from the patients’ electronic medical records. The data included patient demographics (age, sex, and body mass index [BMI]), disease characteristics (disease duration and type according to the Montreal classification), history of perianal fistulas (presence, type, and drainage status), prior ADT exposure (naïve, one agent, two or more agents, secondary non-response to anti-TNFα antibody), and concomitant medication (oral steroid) use. Clinical disease activity was assessed using the Crohn’s Disease Activity Index (CDAI). The mucosal inflammation in CD is generally evaluated using the Simple Endoscopic Score for Crohn’s Disease (SES-CD) [13]. The laboratory parameters included the C-reactive protein (CRP) and serum albumin levels. Adverse events that occurred during the observation period were also recorded.

Definitions of Outcomes

The primary endpoints were clinical response and clinical remission rates at 12 and 54 weeks after UPA initiation. Clinical response was defined as a decrease of at least 100 points in the CDAI score from baseline [14]. Clinical remission was defined as a CDAI score of <150 points [15].

The secondary endpoints included changes in disease activity measured by the CDAI score; the UPA continuation rate at weeks 12 and 54 to evaluate treatment persistence; changes in CRP and serum albumin levels at weeks 0, 12, and 54; and perianal fistula response, defined as the cessation of pus discharge, confirmed by the patient’s subjective complaints or objective macroscopic findings. The average time to cessation of pus discharge was also evaluated in this study. Active fistulae were defined as lesions with ongoing purulent discharge confirmed by physical examination, with or without corresponding patient-reported symptoms.

Perianal fistulae were classified as complex if they involved high trans-sphincteric or suprasphincteric tracts, multiple external openings, associated abscesses, or significant involvement of the anal sphincter complex, in line with the European Crohn’s and Colitis Organisation’s (ECCO) guidelines and related consensus reports [16, 17].

Statistical Analysis

Continuous variables are presented as medians and interquartile ranges (IQRs), while categorical variables are presented as frequencies and percentages. Patient characteristics were compared using the chi-square test. Changes in continuous variables, such as CDAI score, CRP level, and serum albumin level over time were analyzed using the Wilcoxon signed-rank test. All statistical analyses were performed using JMP Clinical 17 software (SAS Institute Inc., Cary, NC, USA).

Results

Patient Characteristics

This study included 13 patients with CD who initiated UPA therapy between June 2023 and June 2024. As shown in Table 1, the mean age was 32.0 years (IQR 21–42), the mean disease duration was 51.0 months (IQR 27–219), and the mean observation period was 62.6 weeks (IQR 34.3–71.6). The variation in observation duration reflects differences in the timing of treatment initiation among patients during the inclusion period. According to the Montreal classification, the disease location was ileal in 1 patient (7.7%), ileocolonic in 11 patients (84.6%), and colonic in 1 patient (7.7%). At baseline, the median CDAI score was 222.0 (IQR 158.7–247.2), and the median SES-CD was 6 (IQR 0–12). The median CRP level was 0.67 (IQR 0.32–1.18) mg/dL, and median serum albumin level was 3.6 (IQR 3.4–3.8) g/dL. Four (30.8%) patients had undergone bowel resection, and five (38.5%) had undergone fistula surgery. Prior systemic steroid use was observed in 8 patients (61.5%), and 3 patients (23.1%) were treated with oral corticosteroids at the time of UPA initiation. Regarding prior ADT, 2 patients (15.4%) were ADT-naïve, while 11 (84.6%) were switched to another ADT. The biologics used immediately before switching to UPA were anti-TNFα antibodies in 5 patients (45.5%), vedolizumab in 3 patients (27.3%), ustekinumab in 2 patients (18.2%), and risankizumab in 1 patient (9.1%). Regarding prior ADT history, 9 of 11 patients (81.8%) had previously received anti-TNFα antibodies, highlighting that UPA demonstrated clinical efficacy even in this difficult-to-treat population with secondary non-response to anti-TNFα antibodies. Non-response to prior biologic therapy was determined at the discretion of the treating physician based on the persistence or worsening of clinical symptoms and the absence of improvement in objective disease activity markers, such as CDAI, CRP, or endoscopic findings.

Table 1.

Baseline characteristics of patients with CD at the initiation of UPA

Variable	13 cases
Age (median [IQR]), years	32.0 [21–42]
Sex, n (%)	​
Male	12 (92.3)
Female	1 (7.7)
BMI (median [IQR]), kg/m2	22.1 [19.7–22.8]
Disease duration (median [IQR]), months	51.0 [27–219]
Observation period (median [IQR]), weeks	62.6 [34.3–71.6]
Classification, n (%)	​
Ileal type	1 (7.7)
Ileocolonic type	11 (84.6)
Colonic type	1 (7.7)
CDAI score (median [IQR])	222.0 [158.7–247.2]
SES-CD score (median [IQR])	6.0 [0–12]
CRP (median [IQR]), mg/dL	0.67 [0.32–1.18]
Serum albumin (median [IQR]), g/dL	3.6 [3.4–3.8]
History of surgery, n (%)	​
Bowel resection	4 (30.8)
Fistula surgery	5 (38.5)
Steroid use at initiation, n (%)	3 (23.1)
Previous use of advanced therapy, n (%)	2 (15.4)
ADT-naïve	11 (84.6)
ADT experienced	6 (54.5)
Two or less three or more	5 (45.5)
ADT details before UPA switch, n (%)	​
Anti-TNFα antibodies	5 (45.5)
Vedolizumab	3 (27.3)
Ustekinumab	2 (18.2)
Risankizumab	1 (9.0)

Table 1 summarizes the demographic and clinical characteristics of the 13 included patients. Baseline disease activity was assessed using the CDAI, and inflammatory status was evaluated using CRP and serum albumin levels. Prior ADT exposure and surgical history were also recorded.

IQR, interquartile range; ADT, advanced therapy; BMI, body mass index; CDAI, Crohn’s Disease Activity Index; CRP, C-reactive protein; UPA, upadacitinib.

Treatment Outcomes

The clinical response and remission rates, which were the primary endpoints, are shown in Figure 1. Clinical response was observed in 46.2% (6/13) and 69.2% (9/13) of patients at weeks 12 and 54, respectively, while clinical remission was achieved in 69.2% (9/13) and 84.6% (11/13) of patients at the corresponding time points.

Fig. 1.

Clinical response and remission rates during UPA therapy. Clinical response was defined as a ≥100-point decrease in the Crohn’s Disease Activity Index (CDAI) score from baseline, and clinical remission was defined as a CDAI score <150. Response and remission rates were evaluated at weeks 12 and 54. a Clinical response rates were 46.2% and 69.2% at weeks 12 and 54, respectively. b Clinical remission rates were 69.2% and 84.6% at weeks 12 and 54, respectively.

As secondary endpoints, disease activity, as measured by the CDAI score, markedly improved, with median CDAI scores of 226.8 at baseline, 118.6 at week 12, and 73.7 at week 54 (Fig. 2). A mean decrease of 100 points was observed after 12 weeks of therapy. The UPA continuation rate was favorable at 100% at week 12 and 83.3% at week 54 (Fig. 3). In this analysis, an “event” was defined as treatment discontinuation due to inadequate response or adverse events, and censored cases were patients who completed the observation period without discontinuing their treatment. Serum markers improved in parallel (Fig. 4); the average CRP level decreased from 1.16 mg/dL at baseline to 0.58 and 0.20 mg/dL at weeks 12 and 54, respectively, whereas the average serum albumin level increased from 3.5 g/dL to 3.9 and 4.1 g/dL at the corresponding time points (“X” shows average marker).

Fig. 2.

Changes in the CDAI score during UPA therapy. The mean Crohn’s Disease Activity Index (CDAI) score decreased rapidly after upadacitinib (UPA) initiation: 226.8, 127.7, and 86.3 at weeks 0, 12, and 54, respectively. Error bars indicate the standard error (SE). Patients with “active fistulae” were defined as those with visible purulent discharge confirmed by both patient complaint and clinical examination.

Fig. 3.

Kaplan-Meier analysis of UPA treatment continuation. The Kaplan-Meier curve shows the continuation rate of upadacitinib (UPA) therapy over a mean follow-up period of 53.1 weeks. An event was defined as treatment discontinuation due to inadequate response or adverse events, while censored cases represented patients who completed the observation period without treatment discontinuation. Continuation rates were 100% and 83.3% at weeks 12 and 54, respectively. Tick marks indicate censored cases. The numbers at risk, events, and censored cases are shown in the plot for clarity of follow-up status.

Fig. 4.

Changes in serum biomarkers during UPA therapy. Serum biomarkers were assessed at baseline (week 0), week 12, and week 54. a Mean C-reactive protein (CRP) levels decreased from 1.16 mg/dL at baseline to 0.58 mg/dL at week 12 and 0.20 mg/dL at week 54. b Mean serum albumin levels increased from 3.5 g/dL at baseline to 3.9 g/dL at week 12 and 4.1 g/dL at week 54. Data are presented as box-and-whisker plots showing the IQR, median, minimum, and maximum values. Mean values are indicated by “×”.

Perianal Fistula Response

Seven patients (53.8%) presented with active perianal fistulas at baseline (Table 2). The mean age was 33.8 (IQR 17–66) years, and 85.7% were male. Compared with the entire cohort, these patients had a slightly lower mean BMI of 19.7 (IQR 17.3–25.4) kg/m². The mean disease duration was 71.9 (IQR 11–253) months, the mean observation period was shorter, at 48.6 (IQR 10–79) weeks, and the mean CDAI score was higher, at 247.7 (IQR 144.7–344.8) points. Most patients (85.7%) had complex perianal fistulas, defined by the presence of high trans-sphincteric or multiple tracts. No patient had rectovaginal fistulas, and no cases of active proctitis were documented at baseline. Surgical interventions, including incision and drainage or seton placement, were performed in patients with abscess formation or signs of local infection. Setons were generally retained during UPA therapy, and removal was not routinely performed during the observation period unless both cessation of drainage and radiologic improvement were confirmed in the patient. All patients had undergone at least one prior fistula-related surgery (incision and drainage, seton placement, or sphincter-preserving surgery), and prior exposure to biologics, including anti-TNF agents, vedolizumab, or ustekinumab, is detailed in Table 2. The definition of non-response to prior biologics was not standardized but was based on the treating physician’s assessment of insufficient symptomatic or objective improvements. Fistula drainage cessation (defined as macroscopic cessation of pus discharge) was achieved in 85.7% of patients within 8 weeks of UPA initiation, and the mean time to fistula drainage cessation was 4.4 weeks. Figure 5 shows the Kaplan-Meier curve for the cumulative incidence of persistent fistula drainage. In this analysis, an “event” referred to persistent fistula drainage at a given time point, whereas censoring indicated the end of the observation period. The curve shows that most cases achieved cessation of fistula drainage within the first 10 weeks and no recurrence was observed during the 54-week follow-up period.

Table 2.

Baseline characteristics of patients with perianal fistulas at the initiation of UPA

Variable	7 cases
Age (median [IQR]), years	36.3 [17–66]
Sex, n (%)	​
Male	6 (85.7)
Female	1 (14.3)
BMI (median [IQR]), kg/m2	19.7 [17.3–25.4]
Disease duration (median [IQR]), months	71.9 [11–253]
Observation period (median [IQR]), weeks	48.6 [10–79]
CDAI score, average, (median [IQR])	247.7 [144.7–344.8]
CRP (median [IQR]), mg/dL	1.18 [0.09–3.94]
Serum albumin (median [IQR]), g/dL	3.5 [2.3–3.9]
Types of fistulas, n (%)	​
Simple fistula	1 (14.3)
Complex fistula	6 (85.7)
History of fistula surgery, n (%)	​
Incision and drainage	2 (28.6)
Seton surgery	3 (42.8)
Sphincter-preserving surgery	2 (28.6)
Antibiotics use at initiation, n (%)	4 (57.1)
Steroid use at initiation, n (%)	3 (42.8)
Previous use of advanced therapy, n (%)	​
ADT-naïve	2 (28.6)
ADT experienced	5 (71.4)
ADT details before UPA switch, n (%)	​
Anti-TNFα antibodies	3 (60.0)
Vedolizumab	1 (20.0)
Ustekinumab	1 (20.0)

Table 2 summarizes the demographic and clinical characteristics of the 7 patients who presented with active perianal fistulas at baseline. Disease activity was assessed using the CDAI, and inflammatory status was evaluated using CRP and serum albumin levels. The presence and complexity of fistulas were determined by clinical examination. Treatment history, including prior fistula surgery and the use of antibiotics and corticosteroids, was also recorded.

IQR, interquartile range; ADT, advanced therapy; BMI, body mass index; CDAI, Crohn’s Disease Activity Index; CRP, C-reactive protein; UPA, upadacitinib.

Fig. 5.

Kaplan-Meier analysis of persistent fistula drainage in patients with perianal CD treated with UPA. An event was defined as cessation of macroscopic fistula drainage, and censored cases represented patients who completed the observation period without achieving drainage cessation. Most patients achieved cessation of fistula drainage within 8 weeks (mean, 4.4 weeks), and no recurrence was observed during the 54-week follow-up period. Patients with “active fistulae” were defined as those with visible purulent discharge confirmed by both patient complaint and clinical examination. Tick marks indicate censored cases. The numbers at risk, events, and censored cases are shown in the plot for clarity of follow-up status.

Adverse Events

No serious adverse events leading to treatment discontinuation were observed during the observation period. Mild adverse events were reported in 2 patients, which included abdominal pain in 1 patient and diarrhea in another. Both events were transient and considered related to CD activity rather than drug-induced effects. In addition, 2 patients discontinued UPA due to insufficient therapeutic response. No infectious events or liver dysfunction were observed during UPA treatment.

Discussion

This single-center retrospective study of the real-world effectiveness of UPA in a cohort of 13 patients with CD revealed favorable clinical response and remission rates, significant improvements in biochemical markers, and a remarkable early therapeutic effect on perianal fistulas. These results provide valuable real-world data on UPA, particularly in a population with a high prevalence of prior ADT exposure, including secondary non-response to anti-TNFα antibody therapy. The observed clinical remission rates of 61.5% at week 12 and 84.6% at week 54, along with clinical response rates of 38.5% and 53.8%, respectively, align well with the high efficacy demonstrated in pivotal phase 3 clinical trials of UPA in CD [9]. Notably, in our study, the clinical remission rate was higher than the clinical response rate. This finding may be explained by the relatively low baseline CDAI scores in several patients, making it easier to achieve clinical remission (CDAI <150 points) than to meet the criterion for clinical response (a decrease of at least 100 points in the CDAI score from the baseline). Importantly, our cohort included a high proportion of patients with ADT experience (84.6%), most of whom (81.8%) had previously received anti-TNFα antibodies. Achieving such high response and remission rates in this predominantly difficult-to-treat population underscores the robustness of UPA efficacy, even outside the stringent criteria of randomized controlled trials [10, 17].

Among the 7 patients with draining perianal fistulas, 85.7% achieved cessation of pus discharge by week 8, with an average time to cessation of 4.4 weeks. Perianal fistulas are a common and debilitating complication of CD, significantly impacting patient QOL and often proving refractory to conventional therapies, including ADT [18, 19]. In this context, infliximab remains the only agent with high-quality evidence from randomized controlled trials demonstrating its efficacy in complex perianal fistulas [20] and is strongly recommended as a first-line treatment in the ECCO guidelines [21]. Adalimumab has shown modest efficacy, with studies such as ADAFI reporting improved outcomes when combined with ciprofloxacin [22, 23]. For patients with inadequate response to anti-TNFα antibodies, ustekinumab and vedolizumab have been evaluated in cohort studies; however, current evidence is limited, and the ECCO guidelines provide only weak recommendations for their use [24]. Against this backdrop, the early and pronounced clinical response observed with UPA in our cohort, despite the high rate of prior ADT exposure, is noteworthy. This suggests that UPA, through selective JAK1 inhibition, may provide a novel therapeutic advantage in perianal CD by targeting the key inflammatory pathways implicated in fistula pathogenesis [25–28]. These observations are consistent with the recent post hoc analysis by Colombel et al. [29], which evaluated the efficacy of upadacitinib in patients with perianal fistulizing CD using data from the U-EXCEED (NCT03345849) and U-ENDURE (NCT03345836) phase 3 trials. In that analysis, 143 patients with active perianal fistulas were included, and the rate of fistula drainage closure was numerically higher in both the UPA 15 mg and 30 mg groups than in the placebo group, with statistical significance achieved for the 30 mg dose at week 52. The safety profile of this subgroup was comparable to that of the overall study population. In line with these findings, our cohort demonstrated a high rate of early clinical response, with cessation of pus discharge achieved in 85.7% of patients by week 8 (mean, 4.4 weeks). While Colombel et al. [29] focused on the long-term efficacy under controlled trial conditions, our data provide complementary real-world evidence of the rapid onset and feasibility of UPA treatment for perianal CD. Taken together, these results suggest that selective JAK1 inhibition may represent a promising therapeutic strategy for treating this difficult-to-treat phenotype.

These findings have important clinical implications. In the current Japanese clinical practice guidelines [30], anti-TNFα antibodies remain the only biologics recommended for patients with perianal disease. Nevertheless, in our cohort, two biologic-naïve patients received UPA as first-line therapy owing to their preference to avoid injectable agents, and informed consent was obtained after explaining the limited evidence for UPA in perianal disease. Taken together, these data suggest that UPA may serve as a valuable treatment option for patients with CD, particularly those who have failed ADTs or prefer oral therapy. Its rapid onset of action, favorable safety profile with no significant adverse events, and potential efficacy in perianal fistulas may help reduce the need for repeated surgical interventions and improve QOL [31, 32]. Despite these encouraging results, some limitations of this study must be acknowledged. First, the retrospective, single-center design may limit generalizability and introduce a selection bias. Second, the small sample size restricts statistical power, although the limited availability of UPA-treated perianal CD cases underscores the relevance of our findings as preliminary data points. Third, the relatively short mean observation period (53.1 weeks) highlights the need for a long-term follow-up. Fourth, fistula assessment was based on clinical symptoms and drainage findings without objective radiological confirmation, such as MRI. Finally, the absence of a control group prevented the definitive attribution of outcomes to UPA. Future multicenter, prospective studies with larger cohorts, longer follow-ups, and incorporation of objective measures (e.g., endoscopic and radiological healing) are needed to establish the long-term role of UPA in both luminal and perianal CD [33, 34].

In conclusion, this real-world study demonstrated the favorable clinical outcomes of UPA in patients with CD, including those with prior ADT failure. Notably, UPA has shown rapid and significant therapeutic effects on draining perianal fistulas, which are particularly challenging complications. Although current guidelines do not define its role in perianal disease, the promising findings of our study support UPA as a potential therapeutic option for selected patients with CD. Larger prospective studies are warranted to validate these results and define optimal patient selection and treatment strategies for this patient population.

Acknowledgment

The authors would like to thank Editage (www.editage.com) for English language editing.

Statement of Ethics

This study was approved by the Clinical Research Ethics Committee of Nagasaki University Hospital (Approval No.: 25031310). Owing to the retrospective nature of the study, the requirement for written informed consent was waived, and an opt-out approach was employed, as approved by the Ethics Committee. This study was conducted in accordance with the Declaration of Helsinki.

Conflict of Interest Statement

The authors have no conflicts of interest to declare.

Funding Sources

This study was not supported by any sponsor or funder.

Author Contributions

Conceptualization, writing – original draft, and formal analysis: J.S.; data curation, investigation, and methodology: J.S., M.K., and K.M.; project administration: K.M.; supervision: Y.A. and H.M.; validation: H.I., T.A., M.T., K.H., N.Y., Y.A., and M.H.; and writing – review and editing: M.K. and K.M. All authors will be informed concerning each step of manuscript processing, including submission, revision, and revision reminders via emails from our system or the assigned Assistant Editor.

Funding Statement

This study was not supported by any sponsor or funder.

Data Availability Statement

The data that support the findings of this study are not publicly available due to patient privacy considerations and restrictions in accordance with institutional ethical guidelines but are available from the corresponding author upon reasonable request.