Natalizumab is a recombinant humanised monoclonal IgG4 antibody against the integrin subunit α4 that blocks both α4β7 and α4β1. The α4β7/MAdCAM-1 interaction is gut-specific, whereas the α4β1–VCAM-1 interaction interferes with lymphocyte migration to the central nervous system [9]. Natalizumab was the first monoclonal antibody to be approved for the treatment of relapsing–remitting multiple sclerosis demonstrating considerable and sustained efficacy [14]. Mechanistic support for its use as induction therapy in active IBD came from the finding that VCAM-1 and MAdCAM-1 are increased in gut inflammation and that natalizumab interferes with this interaction [15, 16].

Natalizumab was first assessed in 30 patients with mild to moderate active CD in a randomised double-blind placebo-controlled trial. A single infusion of natalizumab 3 mg/kg showed superior efficacy in inducing remission at week 2 and was well tolerated [17]. The effect was short-lived with majority of patients requiring rescue therapy at a median of 22 days post-infusion.

In a subsequent double-blind, placebo-controlled trial, natalizumab was administered to 248 patients with moderate to severe CD [18]. Patients were randomly assigned to receive one of four treatments: two infusions of placebo, one infusion of 3 mg/kg natalizumab followed by placebo, two infusions of natalizumab 3 mg/kg or two infusions of natalizumab 6 mg/kg. The group receiving two infusions of 3 mg/kg achieved the highest remission at 44% and a high response rate at 71% at week 6 with reduction in CRP and improvement in quality of life [18]. The efficacy of natalizumab to induce remission was evaluated further in the ENACT-1 and ENACT-2 (Efficacy of natalizumab as Active Crohn’s Therapy) and ENCORE (Efficacy of natalizumab in Crohn’s Disease Response and Remission) trials [19, 20]. In the first trial (ENACT-1), 905 patients were randomly assigned to receive natalizumab or placebo at weeks 0, 4 and 8. Rates of response (56% and 49%, respectively; p = 0.05) and remission (37% and 30%, respectively; p = 0.12) for drug and placebo were similar at 10 weeks. In the second trial (ENACT-2), 339 patients who had responded to natalizumab were randomly assigned to receive 300 mg natalizumab or placebo every 4 weeks through week 56. Although there was no difference between natalizumab and placebo in the first trial (ENACT-1), continuing natalizumab in patients who had a clinical response resulted in higher rates of sustained response (61% vs. 28%, p < 0.001) and remission (44% vs. 26% p < 0.003) than placebo at week 36 [19].

The ENCORE study evaluated the efficacy of natalizumab therapy in patients with moderate to severely active CD and elevated CRP concentrations in a randomised, placebo-controlled trial [20]. Of 509 patients enrolled, 48% demonstrated a sustained response at week 8 through week 12 as compared to 32% of patients treated with placebo (p < 0.001) while sustained remission was noted in 26% given natalizumab and 16% who received placebo [20].

A recent meta-analysis showed that natalizumab was superior to placebo for the induction of remission in CD (RR, 0.86; 95% CI, 0.80–0.93), being equally efficacious for anti-TNF-naive (RR, 0.87; 95% CI, 0.75–1.00) and anti-TNF-exposed (RR, 0.86; 95% CI, 0.76–0.99) patients. Anti-α4 integrins were effective in inducing clinical response and improving quality of life, with no significant differences between natalizumab and vedolizumab. Rates of serious adverse events, infusion reactions, infections and treatment discontinuation were similar [21].

Data on the efficacy of natalizumab in UC are limited. In a pilot study evaluating the efficacy of a single infusion of natalizumab (3 mg/kg), a significant decrease in the median Powell-Tuck score was noted at 2 and 4 weeks (7.5 and 6, respectively) compared to median baseline scores of 10. Reduction in median CRP (6 mg/L) was achieved at 2 weeks from pretreatment levels (16 mg/L), but rescue medication was needed for 2 (20%), 3 (30%) and 8 (80%) patients by weeks 2, 4 and 8, respectively [22].

Despite data confirming efficacy of natalizumab and also its safety profile, further use in clinical practice was limited by the death of a patient, treated with natalizumab in the ENACT study, from progressive multifocal encephalopathy (PML) [23]. Meanwhile, two other cases of PML occurred in patients with multiple sclerosis receiving concomitant natalizumab and interferon β1a [24, 25]. PML is caused by the reactivation of the latent polyoma JC virus and is related to decreased immune surveillance and impaired diapedesis of lymphocytes across the blood-brain barrier [26]. The earliest clinical manifestations are cognitive impairment and behavioural changes, which progress to visual and language disturbances and also seizures, cortical spinal syndrome and motor weakness [27]. Optic neuritis and spinal cord involvement are rare. Unlike classical PML however, gadolinium-enhancing lesions are observed at presentation in approximately 43% of patients with natalizumab-associated PML; these are diffuse and subcortical and rarely involve the periventricular region [28]. The diagnosis is confirmed by quantitative detection of JCV DNA in the cerebrospinal fluid (CSF) using an ultrasensitive assay [27]. Serum JCV PCR is not a useful test for either screening or diagnosis of PML. Management options for natalizumab-induced PML are limited [29]. Natalizumab must be discontinued at the first clinical suspicion of PML. Plasmapheresis to remove natalizumab followed by accelerated desaturation of the targeted α4 integrin receptor and restoration of leucocyte migration are recommended for up to five sessions. Antiviral therapy with cytosine arabinoside or cidofovir and serotonin receptor antagonists may be considered [29]. Rapid reversal of immunosuppression in cases of natalizumab-associated PML may result in an “immune reconstitution syndrome ” which targets JCV in the central nervous system but may result in a paradoxical worsening of PML symptoms for which high-dose corticosteroid therapy may be required [29]. The outcome of PML is dismal, with a reported mortality of 60% in patients with at least 6 months of follow-up [26].

All cases of natalizumab-induced PML occurred in patients who were JCV antibody positive. The seroprevalence of JCV-specific IgG in healthy blood donors is estimated at 50% by 30 years and 68% by 70 years [30]. The main risk factors for PML in natalizumab-treated patients in addition to JCV virus seropositivity, however, are the duration of natalizumab treatment (more than 2 years) and prior use of immunosuppressive therapy [31]. Further clinical experience with natalizumab has been limited by the PML risk, with approval in the USA, under strict vigilance of the TOUCH programme. It is also available in Russia and Switzerland but not in the European Union [1, 32–35].

AJM300 is an orally active and highly specific α4 integrin inhibitor with demonstrated efficacy in a murine model of colitis [36]. Takazoe and colleagues randomised 71 CD patients to receive placebo, oral AJM300 at 40 mg, 120 mg and 240 mg, three times daily for 8 weeks [38]. CDAI reduction at week 4, in AJM300 groups, was higher than in the placebo group, but differences were not statistically significant. The drug was well tolerated at doses of 120 mg and 240 mg three times daily [38]. Yoshimura and colleagues conducted a randomised double-blind placebo-controlled phase IIA trial in patients with moderately active UC [37]. A clinical response (primary endpoint) was achieved in 62.5% in the AJM300 group vs. 25.5% given placebo. Clinical remission rates (Mayo Clinic score ≤ 2 and no subscore >1) were 23.5% and 3.9% in the AJM300 group and placebo groups, respectively (OR = 7.81; 95% CI: 1.64–37.24; P = 0.0099), and rates of mucosal healing (endoscopic subscores of 0 or 1) were 58.8% and 29.4% (OR = 4.65; 95% CI: 1.81–11.90; P = 0.0014). No serious adverse events including progressive multifocal leucoencephalopathy were observed [37].

Vedolizumab is a humanised monoclonal IgG1 antibody which selectively binds to the α4β7 integrin and has been approved for the treatment of patients with moderate to severe UC and CD, by both FDA and the European Medicines Agency [38, 39]. Feagan and colleagues reported the first multicentre, double-blind, placebo-controlled trials of MLN 0002 in two separate studies of similar design [40, 41]. Patients received intravenous infusion of MLN 0002 at 0.5 mg/kg, 2.0 mg/kg or placebo on days 1 and 29. In the UC study with 181 patients, clinical remission at 6 weeks was achieved in 33%, 32% and 14% for the group receiving MLN 0002 at 0.5 mg/kg, 2.0 mg/kg and respectively (P = 0.03) [40]. Corresponding clinical response rates were 66%, 53 and 33%, respectively (P = 0.007). Endoscopic remission was achieved in 28% of patients receiving 0.5 mg/kg MLN 0002 and 12% of patients receiving 2.0 mg/kg, compared with 8% of those receiving placebo (p = 0.007) [40].

In the CD study in 185 patients, the primary efficacy endpoint of clinical response (>70-point decrement in the CDAI score) at 6 weeks was achieved in 53%, 49% and 41% in the 2.0 mg/kg, 0.5 mg/kg and placebo groups, respectively [41]. Clinical remission (secondary endpoint, CDAI < 150) was achieved in 37%, 30% and 21%, respectively. Clinically significant anti-vedolizumab antibody levels (titres >1:125) at day 57 were noted in 12% and 34% of patients in the 2.0 mg/kg and 0.5 mg/kg groups, respectively [41].

The GEMINI phase III studies were each of similar design in the induction phase. Randomised patients received intravenous VDZ 300 mg or placebo at weeks 0 and 2 [42, 43]. A separate open-label group received the same induction regimen. Clinical response was assessed at week 6, and responders were then randomly assigned to continue receiving VDZ (300 mg) every 8 weeks, every 4 weeks or placebo, for up to 52 weeks. All groups included patients with active inflammation despite conventional therapy (corticosteroids, immunosuppressive agents, anti-TNF therapy) and were stratified accordingly.

The GEMINI I study enrolled patients with active UC [42]. The primary endpoint for induction was clinical response at week 6 (a reduction in the Mayo score of ≥3 points and a decre ase of at least 30% from baseline, with a decrease of ≥1 point on the rectal bleeding subscore, absolute score 0–1). The primary endpoint for maintenance therapy was clinical remission at week 52. Of 374 patients randomised to VDZ or placebo, clinical response at week 6 was achieved in 47.1% of the VDZ group versus 25.5% of the placebo group (95% confidence interval 11.6–31.7, p < 0.001). At week 52, 41.8% of patients assigned to VDZ 8 weekly, 44.8% assigned to VDZ 4 weekly and 15.9% of patients assigned to placebo were in clinical remission (8 weekly and 4 weekly compared with placebo, respectively). A Cochrane systematic review on the efficacy of VDZ included 606 patients from four studies [44]. Vedolizumab was significantly superior to placebo for induction of remission (RR = 0.86; 95% CI, 0.80–0.91), clinical response (RR = 0.82; 95% CI, 0.75–0.91), endoscopic remission (RR = 0.82; 95% CI, 0.75–0.91) and achieving remission at 52 weeks in week 6 responders (RR = 2.73; 95% CI, 1.78–4.18) [44].

The GEMINI II trial enrolled patients with moderate to severely active CD with objective evidence of inflammation (CRP > 2.87 mg/L, colonoscopic ulceration or faecal calprotectin >250 μg/g stool plus evidence of ulcers on imaging) [43]. The co-primary endpoints for induction were clinical remission (CDAI ≤150 points) and a CDAI-100 response (≥100-point decrease in CDAI) at week 6. The primary endpoint for maintenance therapy was clinical remission at week 52. Of 368 patients randomised to induction, clinical remission was achieved in 14.5% on VDZ versus 6.8% on placebo (p = 0.02). A CDAI-100 response was achieved in 31.3% on VDZ versus 25.7% on placebo (p = 0.23). At week 52, 39% receiving VDZ 8 weekly, 36.4% receiving VDZ 4 weekly and 21.6% receiving placebo were in clinical remission [43]. A Cochrane systematic review found vedolizumab to be superior to placebo for induction of remission (RR, 0.87; 95% CI, 0.79–0.95). Vedolizumab was efficacious for anti-TNF-naive (RR, 0.86; 95% CI, 0.79–0.94) and anti-TNF-exposed (RR, 0.89; 95% CI, 0.78–1.01) patients [21].

The GEMINI III trial enrolled patients with moderately to severely active CD, the majority of who (76%) had failed anti-TNF therapy [45]. The primary endpoint was clinical remission at week 6 in the anti-TNF failure subgroup. Secondary endpoints were clinical remission at week 10 and a CDAI-100 response at week 6 and week 10. Of 315 patients with CD and anti-TNF intolerance or failure, 15.2% on VDZ versus 12.1% on placebo achieved clinical remission at week 6 (p = 0.433). At week 10, more patients on VDZ achieved remission compared with placebo (26.6% versus 12.1%; 95% CI, 1.3–3.6; p < 0.001) [45]. Taken together, these three trials indicate that VDZ is moderately effective both for UC and CD in a group of patients refractory to conventional therapy including anti-TNF agents. It is noteworthy that the onset of action is relatively slow, often requiring 10 weeks or more of therapy.

Data on clinical efficacy and safety from prospectively followed cohorts are now available. In a recently reported GETAID study, patients with active IBD (CD = 173 and UC = 121), with an inadequate or loss of response to conventional therapy or at least 1 anti-TNF agent, received standard induction and maintenance doses of vedolizumab [46]. Concomitant use of corticosteroids, thiopurines or methotrexate was permitted. At week 14, 31% of patients with CD were in steroid-free clinical remission, and 51% had a response. Among patients with UC, 36% were in steroid-free clinical remission, and 50% had a response. Severe adverse events occurred in 24 patients (8.2%), including 15 (5.1%) that led to vedolizumab discontinuation (pulmonary tuberculosis in one patient and rectal adenocarcinoma in another). No deaths were reported [46]. Integrated long-term safety data (May 2009–June 2013) from the vedolizumab studies [42, 43, 45, 47, 48] have recently been published and show promising results [49]. Of 2830 patients with 4811 PYs of vedolizumab exposure (median exposure range, 1–1977 days), there was no increased risk associated with vedolizumab exposure. Clostridial infections, sepsis and tuberculosis were reported infrequently (≤0.6% of patients). Independent risk factors for serious infection in UC were prior failure of a TNF-α antagonist and narcotic analgesic use, and in CD these were younger age and corticosteroid or narcotic analgesic use. Eighteen vedolizumab-exposed patients (<1%) were diagnosed with a malignancy including non-melanoma skin cancer, malignant melanoma, colon cancer, breast cancer and renal, liver and lung cancer, with nearly all patients (except one with renal cancer) having had prior exposure to thiopurines and or anti-TNF agents [49]. Vedolizumab demonstrated a favourable safety profile over an extended period [49]. A recent systematic review did not detect any significant increase in either opportunistic infections or malignancy with either non-gut-specific or gut-specific anti-integrin antibodies compared to placebo [50]. Reassuringly, no cases of PML have been reported.

Much of the intrinsic appeal for vedolizumab lies in its gut selectivity without systemic immunosuppression. This was elegantly demonstrated in a randomised trial showing reduced seroconversion following oral cholera vaccination against cholera toxin but no attenuation of serological response to parenteral hepatitis B vaccination following a single 750 mg dose of vedolizumab [51]. In the GEMINI trials, enteric infections (Clostridium difficile in six patients, Campylobacter in three and Salmonella in one) occurred after vedolizumab but not placebo [42, 43, 45]. Although the real potential for gut-specific immune inhibition to predispose to enteric infection will be borne out in the fullness of time, clinicians must remain vigilant with patients living in or travelling to the tropics and possibly in patients with risk factors for Clostridium difficile infection. Until recently there were no data that existed on the transmission of infection by live vaccines in patients receiving vedolizumab. The FDA label indicates that patients receiving the medication should receive live vaccines only if the benefits outweigh the risks. Wichmann et al. recently reported a case of a patient with Crohn’s ileocolitis successfully vaccinated against measles virus while on vedolizumab [52]. This anecdotal success with a live vaccine on gut vedolizumab therapy despite making mechanistic sense needs to be studied further.

Vedolizumab has emerged as a viable, efficacious and indeed attractive option in the expanding biological armamentarium for IBD therapeutics. It is crucial for clinicians to understand how this drug will integrate into clinical practice with inevitable comparisons drawn with anti-TNF agents. Bayesian network meta-analyses aim to address this through indirect comparisons with a common comparator but are limited by the heterogeneity of patient populations studied and study design [53–55]. In one network meta-analysis of eight RCTs, the odds ratio for inducing remission in UC was comparable for anti-TNF agents and vedolizumab [55]. One network meta-analysis comparing vedolizumab to other biological therapies in CD found no significant differences [53], whereas another ranked infliximab as the most efficacious agent for induction (86%) and adalimumab for maintenance of remission (48%) [54].

Although induction efficacy of vedolizumab in Crohn’s disease at 6 weeks appears to be less compelling, clinicians must pause to consider certain caveats in extrapolating from these results. Indeed, although clinical remission in CD was superior to placebo, no difference in CDAI-100 response or CRP was noted following induction [43]. It seems likely that the timing of assessment was the limiting factor as evidenced by the GEMINI III trial, wherein vedolizumab was superior to placebo for induction of remission at 10 weeks but not at 6 weeks, in patients who had previously failed anti-TNF therapy [45]. For maintenance of remission at 52 weeks, vedolizumab demonstrated superiority over placebo, with a magnitude of effect generally similar to that seen in UC [42, 43, 45]. Thus, although induction data with CD from trials are less compelling, the clearly clinically meaningful effect after 30 weeks suggests that vedolizumab is an appropriate option in well-selected patients in whom the concomitant use of bridging strategies (such as co-induction with steroids) is possible and where surgery may not be more appropriate. Indeed, it might also be a first-line biologic option in patients where the focus is safety, for example, in young or elderly patients with IBD [56–58]. The role of vedolizumab in treatment of perianal disease is unclear. At 52 weeks in GEMINI II, 41.2% of the vedolizumab 8-weekly group achieved fistula closure compared with 22.7% of the vedolizumab 4-weekly group and 11.1% of the placebo group (p = 0.03, p = 0.32 versus placebo, respectively) [43]. This borderline significance needs further investigation. Indeed, a higher incidence of perianal abscesses was reported in preliminary data from the GEMINI-LTS extension study [49].

The safety and efficacy profile for UC may be regarded as more favourable, positioning vedolizumab as a potential first-line biologic for induction and maintenance of remission in outpatients with moderate to severe UC, who have failed or had an inadequate response to corticosteroids or immunosuppressant therapy. It cannot be recommended at the present time for the treatment of acute severe UC due to its relative slow onset of action and in the absence of data for this indication. There are no data for the perioperative safety and efficacy of vedolizumab in CD, and although the mechanism of action of vedolizumab , preventing early stages of inflammation, is appealing, clinical trials are needed to provide credible evidence. There are no data in patients with extra-intestinal manifestations of IBD. Although it seems implausible that a gut-selective agent should benefit those manifestations that do not parallel IBD activity (such as pyoderma gangrenosum or ankylosing spondylitis), manifestations associated with gut inflammation (e.g. erythema nodosum and episcleritis) may benefit, and this merits further study.

More promising is the prospect of treating primary sclerosing cholangitis (PSC) affecting 3–10% of IBD patients [59, 60]. Hepatic inflammation in PSC is driven by TNF-α and methylamines in the portal circulation and results in aberrant hepatic expression of MAdCAM-1 and the chemokine CCL25 [60]. This leads to enhanced recruitment of α4β7 and the CCL25 receptor CCR9. Randomised trials of VDZ in patients with IBD–PSC are under way (clinicaltrials.​gov NCT00783692 and NCT01316939).

Although vedolizumab has not been associated with an increased risk of malignancy, long-term experience is limited, and indeed patients with prior malignancy were excluded from trials. Diminished gastrointestinal immune surveillance may pose a theoretical concern for colorectal cancer complicating UC or small intestinal adenocarcinoma in CD, given their increased risk relative to the general population. Nonetheless, carcinogenesis is a likely consequence of inflammation, and it is not implausible that control of inflammation by vedolizumab may reduce this risk [61, 62]. More research is needed in this area.

Vedolizumab is a pregnancy risk category B drug with limited data on safety in pregnancy. In a series of 24 women exposed to vedolizumab during pregnancy, there were 12 live births, five elective abortions and four spontaneous abortions [63]. With a half-life of 25 days, any strategy of withholding dosing in the third trimester could result in significant vedolizumab concentration in the foetus and prolonged drug clearance in the neonate potentially extending to 6–12 months, the consequences of which are presently unknown. This could have implications on vaccination against enteric infections such as rotavirus (an oral vaccine) but possibly not parenteral agents commonly administered in the first year of life. No evidence-based recommendations can be made at the present time, and any intentional use in pregnancy would need to be discussed on a case-by-case basis. Data on vedolizumab use in the paediatric age group are limited to retrospective observational data in largely TNF-exposed patients, suggesting a remission rate of 100% at 14 weeks in three patients with UC, an improvement to the comparative 44% reported in nine Crohn’s patients [64]. A phase III study of vedolizumab of patients 15 years and older is currently ongoing (ClinicalTrials.​gov:NCT02039505), and a phase III PK/PD paediatric trial is about to start.