3. Results
3.1 Study selection
Four hundred sixty-one titles and abstracts were screened, 159 were retrieved as full text articles. A total of 5 studies (published in 7 articles) met eligibility criteria and were included in the review [3,4,32-36]. The PRISMA flowchart, depicting the process of study selection, is shown in Figure 1.
3.2 Risk of bias and quality of evidence
Each of the domains for the five included studies were deemed low risk of bias. The detailed assessments or risk of bias and the ratings for quality of evidence are displayed in supplementary figures S3 and S4 and supplementary table SI.
3.3 Study characteristics
Five double-blinded RCTs evaluating 2,899 patients with IBD initiating treatment with VDZ were included in this review [3,4,32,33,34]. In one study, VDZ was administered as essentially first line treatment [32] whereas in the remaining studies, VDZ was given as atleast second line treatment [3,4,33,34]. Follow-up time ranged from 6 to 76 weeks with a median follow-up time of 52 weeks. All of the studies were multicenter RCTs; one study took place in Japan alone, one in the USA and Canada, one across 5 continents, and in more than 30 countries in the remaining two studies. Four out of the 5 included studies were funded by pharmaceutical companies [3,4,33,34]. Table I summarizes the characteristics of the included studies.
3.4 Study participants
The number of randomized patients in the included studies ranged from 181 to 1115 [3,4,32,33,34]. In 4 of the listed studies, mean age ranged from 36 to 43 years [3,4,32,34]. In the fifth study, a median age of 36 was provided [33]. In 3 of the studies, the majority of participants were male [3,32,34], while in the remaining, the majority were female [4,33]. Table II provides a more detailed summary of the baseline characteristics of the participants in the included studies.
3.5 Data from RCTs
In 2005, Feagan et al [32] published an RCT to evaluate the efficacy of VDZ and to assess optimal dosing in patients with UC. One hundred eighty-one patients were included, in which patients were evenly split into three arms- placebo and two different doses of VDZ. Included patients had evidence of active disease as measured by both clinical and endoscopic parameters and were required to be naïve to anti-tumor necrosis factor (TNF) therapies, and had only previously received glucocorticoids or mesalamine. The study met its primary endpoint, demonstrating that after 6 weeks, patients receiving VDZ (33% in the 0.5mg/kg dose group and 32% in the 2mg/kg dosage) experienced clinical remission significantly more than those who were receiving placebo with an overall p value of 0.03.
AEs did not differ significantly among the groups (p=0.50). Arthralgias were experienced by 4 (7%) in the low dose group, 7 (12%) in the high dose group, and 5 (8%) in the placebo group without a detectable statistically significant difference among all those treated with VDZ compared with placebo, as calculated by the authors using Fisher’s test (p= 0.75). Further analysis yielded an OR of 1.19 with a 95% CI of 0.40-3.60 [32].
In 2013, the GEMINI 1 study group [3] conducted 2 integrated RCTs to assess the effect of VDZ on induction and maintenance of active disease in patients with UC. Prior TNF exposure was permitted. In the trial of induction therapy, patients were either randomized to receive VDZ or placebo at weeks 0 and 2, with evaluation of disease activity at the sixth week. In the subsequent maintenance trial, participants in either cohort who were deemed to have responded to VDZ at week 6 were randomized to receive VDZ every 4 or 8 weeks or to receive placebo through week 52. The study met its primary endpoint defined by a reduction in the Mayo Clinic score of atleast 3 points and a decrease of atleast 30% from baseline, with either a decrease of at least 1 point on the rectal bleeding subscale or an absolute rectal bleeding score of 0 or 1. The study also met its designated primary outcome for the maintenance phase, clinical remission at week 52.
The authors evaluated AEs by comparing all patients who had received VDZ as maintenance therapy with the group that were given placebo during the maintenance phase, whether they had received placebo or VDZ during induction. They did not find significant differences in the category of any AE (p= 0.23) as well as among serious AEs (p= 0.06) in the induction trial. Similarly, there was no difference in any AEs in those receiving VDZ every 8 weeks and every 4 weeks compared with placebo (p=0.65 and 0.49, respectively) nor any difference in serious AEs (p=0.06 and 0.09, respectively) during the maintenance phase. While there was an increased rate of development of arthralgias in those receiving VDZ during the trial, this was not statistically significant (OR 1.01; 95% CI 0.61-1.65: p= 0.98) [3].
The GEMINI 2 study group [4] conducted a trial with an identical design to that of the GEMINI 1 group assessing VDZ response in patients with active CD instead. The study met part of its primary endpoint in the trial of induction therapy evaluating for clinical remission (CDAI ≤150) at week 6 (p= 0.02) though not for clinical response (decrease in CDAI-100 response ≥100) (p= 0.23). In the maintenance phase, the primary endpoint of clinical remission at week 52 was achieved in patients receiving VDZ every 8 weeks (p <0.001) and every 4 weeks (p= 0.004).
There were no significant differences in any reported AEs (p= 0.56) as well as serious AEs (p= 0.29) in the induction trial. Similarly, there was no difference in any AE in those receiving VDZ every 8 weeks and every 4 weeks compared with placebo (p=0.32 and 0.86, respectively) nor any difference in serious AEs (p=0.46 and 0.77, respectively) in the maintenance phase. Similarly, the authors evaluated arthralgias in all those who received VDZ in the maintenance phase compared with those who had received placebo as maintenance, regardless of what they had received during the induction, and did not find a statistically significant increased rate of arthralgias (OR 1.02; 95% CI 0.69-1.50: p= 0.92) [4].
The GEMINI 3 study group [33] conducted a phase 3 RCT to evaluate the efficacy of VDZ in patients with moderately to severely active CD with an objective toward assessing those who had previously failed anti-TNF therapy. In this trial, patients were randomized to receive VDZ or placebo (1:1) at weeks 0,2, and 6 and were assessed at week 6. This trial did not meet its primary outcome assessing clinical remission, as defined by CDAI≤ 150 at week 6, among those participants with prior anti-TNF failure (p= 0.433).
There were no significant differences in the AEs (p= 0.418) and in the drug-related AEs (p= 0.965) in patients who received VDZ compared with placebo. Similarly, no significantly increased risk of arthralgias was identified at week 6 in those who had received VDZ (OR 1.11; 95% CI 0.44-2.78: p= 0.831). This study also evaluated musculoskeletal pain apart from arthralgias and found that the while there was an increased risk of myalgias, this difference was not significant (OR 9.09; 95% CI 0.49-169.88: p= 0.140) [33].
Most recently, Motoya et al [34] performed a phrase 3 RCT in Japan looking at patients with active UC, who were permitted to have previously used anti-TNF agents. This study was split into a double-blinded cohort, who were randomized to receive VDZ or placebo, and an open label cohort who were assessed following completion of induction, at week 10. All patients who showed a clinical response to VDZ during induction were randomized 1:1 to receive VDZ or placebo every 8 weeks through week 54 with clinical evaluation at week 60. In addition, anyone could receive re-induction open label VDZ followed by treatment every 8 weeks up to 94 weeks if they had not had a clinical response to VDZ or placebo during the induction, experienced disease worsening, had received rescue treatment during the maintenance phase, or completed week 60 of maintenance, with an evaluation at 16 weeks following the last dosage received. This study did not reach its primary endpoint for the induction phase, a clinical response at week 10, as defined by a reduction in the full Mayo score of > 3 and atleast 30% from baseline as well as > 1 on the rectal bleeding subscore or an absolute rectal bleeding subscore< 1 (p= 0.272). It did however meet its primary endpoint for the maintenance phase, designated as clinical remission (full Mayo score < 2 and no subscore > 1) at week 60 (p= 0.021).
There was no significant difference in the development of AEs in the VDZ group compared with placebo in the induction phase (p= 0.720) nor in the maintenance phase (p= 0.2668). Similarly, there was no difference in the development of serious AEs in the induction (p= 0.409) and maintenance phases (p= 0.670). In addition, there was no significant difference in the development of arthralgias during the induction phase (OR 1.57; 95% CI 0.17-14.29: p= 0.688) nor during the maintenance phase (OR 10.20; 95% CI 0.53-195.78: p= 0.123) [34].
3.6: Post-hoc analyses
A post-hoc analysis of the GEMINI 2 trial [35], conducted by some of the members of the original study group, divided patients into those that were anti-TNF naïve and those who had previously failed anti-TNF therapy. By stratifying the participants, the authors found that, the risk of developing arthralgias when treated with VDZ compared with placebo during the induction was not greater among those who had formerly received anti-TNF agents (TNF-naïve: OR 1.26; 95% CI 0.26-6.00: p= 0.774 and TNF-failure: OR 1.24; 95% CI 0.64-2.41: p= 0.523). During the maintenance phase, on the other hand, there was an increased risk of developing arthralgias in those who had previously used anti-TNF treatments, though that difference was not statistically significant (TNF-naïve: OR 0.83; 95% CI 0.37-1.84: p= 0.648 and TNF-failure: OR 1.24; 95% CI 0.59-2.60: p= 0.566).
A second post-hoc analysis [36], also co-authored by several members of the GEMINI study group and published 2 years later, evaluated data from the 3 GEMINI trials. Prevalence of baseline arthritis/arthralgias, which was the most common EIM at baseline in both patients with CD and UC, was similar in all of the study arms assessed in GEMINI 2 and 3. As part of further analysis, the co-authors re-divided the patients into three groups based on whether they received: VDZ throughout, VDZ in the induction phase followed by placebo for maintenance, and placebo alone. They evaluated the probability of sustained resolution of arthritis/arthralgias with prevalences of 11.9%, 15.3%, and 11.2% according to the respective groupings. They found no significant difference among the groups (VDZ vs placebo: hazard ratio (HR), 0.99; 95% CI, 0.52–1.90; VDZ followed by placebo vs placebo alone: HR, 1.57; 95% CI, 0.70–3.28).
Further scrutiny revealed that in the GEMINI 2 trial, the probability of sustained resolution of arthritis/arthralgia was 51.2% in those who received VDZ during both phases, 41.4% in those who had received VDZ during induction followed by placebo in the maintenance phase, and 35.5% in those had had been given placebo throughout. Differences among groups were not statistically significant (VDZ alone vs. placebo alone: HR, 1.56; 95% CI, 0.93–2.59; VDZ followed by placebo vs placebo alone: HR, 1.40; 95% CI, 0.75–2.64). Their analysis also revealed that clinical response and clinical remission at weeks 6 and 52 significantly correlated with sustained resolution of baseline arthritis/arthralgia (p  < 0.05). Evaluation of GEMINI 3 revealed sustained resolution of baseline arthritis/arthralgia at week 10 in 22% of the VDZ group and in 16% of the placebo group with no significant difference detected between the 2 arms (HR, 1.40; 95% CI, 0.73–2.67) [36].
A Cox analysis of GEMINI 2 found that when separating the participants into those who received VDZ alone, placebo alone, and VDZ followed by placebo, there was a significant increase in the incidence of arthritis/arthralgias in those who received VDZ during both trial phases and in the group who received VDZ during induction followed by placebo compared with those who remained on placebo throughout (VDZ alone vs. placebo alone: HR, 0.55; 95% CI 0.36–0.84); and (VDZ followed by placebo vs. placebo alone: HR, 0.45; 95% CI, 0.26–0.81). Moreover, patients with prior anti-TNF use were significantly more likely to develop new joint complaints compared with those who were anti-TNF naïve (HR, 2.20; 95% CI, 1.56–3.11) [36].
On the other hand, a Cox analysis of GEMINI 2 evaluating both new or worsening arthritis/arthralgias revealed that such findings were less likely in those who received VDZ and VDZ followed by placebo compared to those who received placebo alone (VDZ alone vs. placebo alone: HR, 0.63; 95% CI, 0.44–0.89; and VDZ followed by placebo vs. placebo alone: HR, 0.54; 95% CI, 0.34–0.87). Additionally, patients with prior anti-TNF failure were generally more likely to experience new and or worsening arthritis/arthralgia than naïve patients (HR, 1.81; 95% CI, 1.37–2.38) [36].