Crohn's disease (CD) and ulcerative colitis (UC) are the two main forms of IBD. CD is characterized by chronic, patchy granulomatous inflammation with skip lesions, affecting any part of the gastrointestinal tract, especially the terminal ileum and colon. The inflammation is transmural which can lead to fibrosis, stricture, and fistula. In contrast, UC is characterized by continuous mucosal inflammation extending from the rectum proximally toward the colon. Differentiating these two conditions is important as each has diverse prognoses and differential responses to treatment (7). The clinical presentations of IBD include recurrent abdominal pain, bloody diarrhea, and mucus in the stool. Patients with CD can present with intestinal obstruction, recurrent fistulas, and other perianal findings. Systemic symptoms include fatigue, weight loss, fever, and symptoms of anemia. The standardized mortality ratio for CD ranges from 1.2 to 1.9 times the general population (8). The prognosis of IBD has improved in recent decades due to therapeutic advances.

Amongst patients with IBD, extraintestinal manifestations are common, including musculoskeletal (axial and peripheral arthropathy and arthritis), ocular (uveitis, scleritis and episcleritis), and skin. Inflammatory arthropathies are reported up to 40% of patients with IBD (9). While asymptomatic sacroiliitis may be seen in up to three-quarters of IBD patients, the reported prevalence of seronegative SpA ranges from 18–45%, and ankylosing spondylitis (AS) 3–9.9% (10, 11). Peripheral arthritis is reported in 7–16% of IBD patients. Peripheral arthritis is mainly asymmetrical and oligoarticular, usually acute and occurs during IBD exacerbations, and self-limiting. However, it may also persist for months or years. Its onset usually coincides with or after IBD but may precede IBD. Enthesitis and dactylitis were reported in 2–4% of patients (12).

Amongst patients with SpA, IBD is common (13). Patients with PsO, PsA and AS have a 1–4 fold increased risk of IBD compared to the general population (14–18). Among patients with SpA, 30–42 % have endoscopic (macroscopic) gastrointestinal inflammation (19–22) while 46–58 % have histologic (microscopic) inflammation (20, 21, 23). The presence of macroscopic or microscopic inflammatory lesions poorly correlates with symptoms (19). In patients with axial SpA, the severity of microscopic inflammation was significantly associated with severity of bone marrow edema on magnetic resonance imaging, indicating a link between mucosal inflammation and progressive disease (24).These subclinical gastrointestinal inflammatory lesions may predispose SpA patients to develop IBD, with a lifetime IBD risk of between 4–8% (25–28). Among patients with PsO and PsA, IBD is more common in patients with more severe PsA (29). IBD is also more common in patients with axial-PsA than in those with peripheral-only PsA (30).

Uveitis is the inflammation of the uveal tract of the eye which comprises of the iris, ciliary body, and choroid. Adjacent structures including retina, optic nerve, vitreous, and sclera may also be affected. Clinically, uveitis is categorized anatomically – anterior, intermediate, posterior, or panuveitis (31). There is an increased association of ocular manifestations amongst patients with PsD (32, 33). Other presentations like vitritis, retinal vasculitis, and cystoid macular edema involving the posterior chamber are sight-threatening (34, 35). The prevalence of uveitis increases with disease duration, lifelong prevalence is over 40%. Among patients with SpA, acute anterior uveitis (AAU) is most common (26) and its prevalence varied with the type of SpA: 33% in AS; 37% in IBD-associated arthritis; 26% in reactive arthritis; 25% in PsA; and 13% in undifferentiated SpA (36, 37). In both Asian and Western populations, uveitis is common in patients with severe PsO and those with PsA (38, 39). Uveitis in SpA usually presents with a ‘unilateral alternating' pattern, sudden in onset, confined to the anterior chamber, and completely resolves between episodes (40). In contrast, uveitis in PsA is insidious in onset, bilateral with a chronic relapsing course. PsA patients with both uveitis and axial arthropathy tend to be male and HLA-B^*27 positive (41, 42). HLA-DR^*13 positivity is also associated with uveitis in patients with PsA (43). Uveitis may also precede the development of PsA in patients with PsO (44).

A combination of genetic and environmental factors contributes to pathogenesis of PsA (Figure 1). Genetic component in PsA is strong (45). HLA class I alleles such as HLA-B^*27:05:02 haplotype is widely reported to be positively associated with enthesitis, dactylitis, and sacroiliitis while the HLA-B^*08:01:01–HLA-C^*07:01:01 haplotype is positively associated with joint fusion, deformity and asymmetrical sacroiliitis. In contrast, the B^*44:03:01–C^*16:01:01 haplotype may be protective against enthesitis (46). Additional HLA haplotypes associated with susceptibility to PsA were HLA-B^*38, and HLA-B^*39 (47–51). Non-HLA PsA susceptibility loci related to inflammatory pathways have been implicated. IL-23 receptor (IL-23R) polymorphisms are associated with risk of PsA (52). Tumor necrosis factor receptor-associated factor 3-interacting protein 2 (TRAF3IP2), encoding nuclear factor-κB (NFκB) activator protein 1 (Act1) which is an adaptor protein for interleukin-17 (IL-17) receptor (53–55), IL-23A, IL-12B, and TYK2 (Tyrosine Kinase 2) are other examples, highlighting the importance of IL-23/IL-17 axis in the pathogenesis of PsA (56).

In a genetically predisposed individual, environmental factors including mechanical stress may trigger enthesitis – a hallmark clinical presentation of SpA including PsA (57, 58). Mechanical stress and trauma release damage-associated molecular patterns (DAMPs), triggering the production of prostaglandin E2 (PGE2) (59) by resident mesenchymal cells, which recruit innate immune cells to perpetuate inflammation. PGE2 also induces T cell secretion of IL-17, a key driver in PsA pathogenesis (58, 60). Innate immune cells such as dendritic cells (DCs), monocytes/macrophages, neutrophils, and innate lymphoid cells (ILCs) corroborate with adaptive immune cells to perpetuate inflammation in PsA (61). Additionally, plasmacytoid dendritic cells (pDCs) infiltrate the synovium to act as antigen presenting cells (APCs), triggering downstream expression of TNF-α, IFN-γ, and IL-2 from CD68+ macrophage-like-synoviocytes that mediate synovial inflammation and bone erosions (62, 63). TNFα synergizes with DCs to activate and polarize Th17 cells (64). In addition to Th17 cells, type 3 innate lymphoid cells (ILCs) (65, 66), mucosal-associated variant T (MAIT) cells (67, 68), and γδ T cells (69) are recruited to the synovium and produce IL-17A upon stimulation (70). In short, the IL-23/IL-17 axis is the central driver of inflammation in PsA (71–75).

Genetic predisposition increases the risk of developing IBD amongst patients with PsA and SpA. HLA-B27 is the major HLA associated with IBD risk. 25–78% of patients with AS and IBD are HLA-B27 positive (9, 76, 77). HLA-DRB1^*01:03 is also common between AS and IBD (76, 78–80). Non-HLA polymorphisms such as nucleotide-binding oligomerization domain-containing protein 2 (NOD2) polymorphisms increase the risk of CD about 4–40 times and is associated with sacroiliitis amongst patients with IBD. NOD2, an intracellular receptor expressed by immune and intestinal cells, is involved in the activation of NFκB and inducing pro-inflammatory genes in innate immune cells (81–85). IL-23R polymorphisms modify susceptibility to IBD, where a loss-of-function mutation may have protective effect against IBD (86). Polymorphisms in genes related to the IL-23/IL-17 axis such as IL-12B, signal transducer and activator of transcription 3 (STAT3), and caspase recruitment domain family member 9 (CARD9) are associated with CD (87). Once again, this highlights the IL-23/IL-17 axis as a major pathogenetic pathway for IBD manifestations in patients with PsA.

The microbiome plays an important role in gastrointestinal health, and dysbiosis of the microbiota is observed in patients with IBD. Microbiota in IBD patients is less diverse compared to healthy controls. Gastrointestinal bacteria may invade the sterile inner colonic adherent mucus layer, disrupt epithelial architecture, and allow leakage of bacterial antigen into the systemic circulation to induce and perpetuate inflammation (88–90). A “gut-joint axis” has been proposed where immune cells traffic between the two domains (91, 92). Fecal microbiota transplantation (FMT) has shown promising results in the treatment of UC in a Cochrane Database systematic review (93). Positive clinical outcomes are associated with higher dosage and delivery of FMT via lower gastrointestinal tract (94), and may be dependent on stool donor (95). However, a recent RCT on FMT in 31 patients with active PsA randomized to FMT vs. sham treatment was not efficacious for arthritis (96). Further study is required.

In patients with IBD, the number of IL-17-secreting MAIT cells (97), was increased in the gastrointestinal tract as compared to the peripheral blood, echoing PsA studies showing depleted MAIT cells in blood, and increased MAIT cells in inflamed synovia and psoriatic skin (67, 68). γδ T cells are found in colonic mucosa and represent around 40% of intraepithelial lymphocytes (98). In contrast to PsA, the presence of γδ T cells appears to be protective and anti-inflammatory in patients with IBD. Different subtypes of γδ T cells may behave differently in different cytokine environments, explaining the diverse observations of γδ T cells in PsA and IBD (99, 100). As with PsA, Th17 cells are major players in IBD (101). The chemokine receptor CCR6 is the main surface marker of the Th17 lineage. CCL20, a ligand for CCR6, is elevated in IBD gut epithelium and likely contributes to recruitment of CCR6+ type 3 ILC, Th17, and dendritic cells (102, 103). Due to high levels of IL-17 and IL-23 in IBD gut epithelia, the IL-23/IL-17 pathway was thought to be a therapeutic target (104–106). However IL-23 inhibition showed efficacy in patients with IBD but IL-17 inhibition lead to disease exacerbation (107). A possible explanation for this paradox is that IL-17 plays a role in maintaining intestinal barrier and microbial defense (108–110).

The HLA-B^*27 is associated with increased risk of AAU (111), and is a common risk locus for PsA (and other SpA) and uveitis (112). HLA haplotypes such as HLA-A^*02 (113), HLA-DRB1^*08-03 (114), HLA-B^*58 (115) were also associated with development of uveitis. Other non-HLA susceptibility loci are major histocompatibility complex class I chain-related gene A (MICA) (116, 117), IL-10 (118), TNF (119), killer immunoglobulin receptor (KIR) (120), and polymorphisms in IL-23R, which all participate in immune response (121). Nonetheless, positive risk polymorphisms do not necessarily translate to uveitis. Other environmental and undiscovered factors are likely required to initiate uveitis in patients with SpA.

The eye is an immunologically privileged organ with a local inhibitory microenvironment, entailing immune ignorance and tolerance to prevent inflammation. The blood-retina barrier and absence of efferent lymphatics reduces exposure of the eye to the circulating immune system (122). In uveitis, infiltration of immune cells into the eye and disrupts the immunologically quiescent environment. However, the trigger of this infiltration is undetermined (123). Some evidence implicates the perturbation of the gut microbiome to SpA-associated uveitis. Animal studies demonstrated trafficking of leucocytes from intestine to eye, supporting the concept of a gut-eye axis (124). Further evidence from retina-specific TCR transgenic mice reared under germ-free conditions showed that the severity of uveitis was reduced in the absence of gut microbiota. This reduction of severity was associated with a reduction in Th17 cells in the lamina propria of the intestine. Reconstitution of gut microbiota increased retina-specific T cell signaling (125). McGonagle et. al (2015) proposed that anterior uveal structures are analogous to entheses due to their mechanical and structural roles in lens suspension. The repeated contractions and relaxations of these structures expose them to mechanical stress just like musculoskeletal entheses, thus providing the initial stimuli for inflammation (126). Like entheseal mesenchymal cells in enthesitis, cells in ciliary body express IL-23R, suggesting receptiveness to IL-23 signaling (127). In patients with uveitis, serum IL-17A levels were elevated during active disease (128). Association between Th17 and the development of uveitis has been observed in animal and clinical studies highlighting the importance of the IL-23/IL-17 axis in driving inflammation in PsA and uveitis (129–132). However, clinical trials have yet to demonstrate the efficacy of IL-17 inhibition in uveitis (133).

The treatment targets for patients with IBD are clinical remission, mucosal healing, and restoring quality of life (137, 138). The importance of mucosal healing defined as restitution of the intestinal lining and regression or disappearance of endoscopic lesions has been emphasized. Achievement of this target is associated with reduced risk of relapse, reduced hospitalization rates, steroid-free remission, and resection-free status (139–141). With medical advancements, the need for bowel resection is substantially reduced (142).

Corticosteroids can be used to induce clinical remission. It is given either topically as ileal-release budesonide for active mild-to-moderate CD or systemically for moderate-to-severe CD (132). However, systemic corticosteroid should not be used for maintenance (143, 144). Early initiation of corticosteroid-sparing immunomodulators such as azathioprine (AZA), mercaptopurine or methotrexate (MTX) for maintenance should be considered, although the level of evidence supporting efficacy of these drugs is relatively low (144, 145).

Monoclonal antibody targeting TNFα (TNF inhibitors, TNFi) has become the standard of care for patients with moderate-to-severe, active CD. Infliximab (IFX), adalimumab (ADA), and certolizumab (CZP) have demonstrated efficacy in inducing remission and maintenance in RCTs, and well supported by meta-analyses (146, 147). We summarized the major RCTs supporting the efficacies of TNFi in IBD Table 2. In a Cochrane Database Systematic review, CD patients who responded to induction by TNFi were more likely to maintain remission at 52 weeks with TNFi compared to placebo (147). Continued treatment with TNFi reduces surgery and hospitalization for CD (168, 169). Combination therapy of IFX with AZA was more efficacious than either agent alone in achieving response, inducing clinical and histological remission (156), suggesting synergistic effect. TNFi appears to be more effective when given at earlier stage of disease, with higher rates of response and remission, than given at later stage of disease (170, 171). Early escalation to TNFi treatment should be considered for patients with extensive disease and poor prognostic factors (144, 145).

Vedolizumab (VZD) is a monoclonal antibody targeting α4β7 integrin, which reduces lymphocytes trafficking to the gastrointestinal tract by blocking lymphocyte surface α4β7 binding to the mucosal addressin cell adhesion molecule-1 (MAdCAM-1). The efficacies of VZD in induction and maintenance in CD have been demonstrated in the GEMINI-2 (172) and GEMINI-3 trials (173) (Table 1). In a meta-analysis involving 1716 patients with CD, VZD was more effective than placebo for inducing clinical remission (RR 1.71 [95% confidence interval, CI: 1.25, 2.34], p = 0.0008), and maintaining clinical remission (RR 1.75 (95% CI: 1.25, 2.44), p <0.001).

Ustekinumab (UST) is an antibody targeting the IL-12/23 p40 subunit. The efficacy of UST in inducing remission in CD has been shown in UNITI-1 and UNITI-2 trials in patients with inadequate response to TNFi, and without prior TNFi failure, respectively. Responders from both studies were randomized to the IM-UNITI maintenance study and demonstrated significantly higher clinical remission rates [high dose: 53%, P = 0.005; low dose: 49%, P = 0.04)] compared to placebo (36%) at week 44 (162). There is no head-to-head study comparing efficacies between TNFi, VZD and UST. The choice of biological treatment is a shared decision between clinician and patient, and according to the individual risk–benefit preferences.

Risankizumab (RZB), an IL-23/p19 inhibitor met the primary remission induction endpoints in CD in two phase III RCTs, ADVANCE (NCT03105128) and MOTIVATE (NCT03104413) (174). Patients in remission from ADVANCE and MOTIVATE were recruited to the Phase III open-label maintenance study, FORTIFY, showing RZB 360 mg every 8 weeks achieved the co-primary endpoints of clinical remission and endoscopic response at 52 weeks (175).

Blocking IL-17, however, has not been effective in CD. In a phase II trial evaluating safety and efficacy of brodalumab (BRO), a monoclonal antibody targeting IL-17 receptor, the primary induction endpoint was not met. The trial was terminated early due to a disproportionate number of cases of worsening CD (160). In a phase II RCT, two doses of 10 mg/kg secukinumab (SEC) given intravenously on days 0 and 22, failed to meet the primary endpoint and had more adverse events compared to placebo (176). However, the use IL-17i is not associated with increased incidence of IBD. Data from the SEC development program pooling 7,355 patients with a cumulative exposure of 16,227 person-years of patients exposed to SEC for PsO, PsA or SpA, no increase in exposure adjusted incidence rates of IBD was observed (15). Similarly, events of IBD remained low in the ixekizumab development program that pooled data from 15 RCTs in PsO and PsA (177).

Phase II RCT results for the Janus kinase inhibitor (JAKi), upadacitinib (UPA), in CD are promising. Endoscopic but not clinical remission increased with dose during the induction period (167). However, in a phase II trial for JAKi, tofacitinib (TOF), no statistically significant differences in clinical responses between TOF and placebo were observed at week 4 (164) (Table 2).

Oral 5-ASA (5-aminosalicylic acid) is the standard therapy for induction in mild-to-moderately active UC. For those failing 5-ASA or with moderate-to-severe UC, a short 6- to 8-week course of oral corticosteroid is indicated. 5-ASA and thiopurines can be used as maintenance therapy. Like the treatment strategy for CD, early escalation to biologic therapies should be considered for those who failed induction therapy with corticosteroid, or failed maintenance with immunomodulators, and those with poor prognostic factors. TNFi [IFX, ADA, golimumab (GOL)], α4β7 integrin inhibitor (VZD) and IL12/23i (UST) are approved treatments for induction and maintenance of UC (Table 3). A combination of TNFi with an immunomodulator is more effective. In the UC-SUCCESS trial, patients treated with IFX and AZA were more likely to achieve corticosteroid-free remission at 16 weeks than those receiving either monotherapy (181). In a head-to-head study (VARSITY) comparing TNFi and VZD in patients with moderate-to-severe UC, 769 patients with moderate-to-severe active UC were randomized to receive VDZ or ADA (185). Only 26% of patients in either group were on concomitant immunomodulators. At week 52, a higher percentage of patients achieved clinical remission (31.3 vs. 22.5%; P = 0.006), and endoscopic improvement (39.7 vs. 27.7%; P <0.001) in VDZ compared to ADA group. Whilst corticosteroid-free clinical remission occurred at a higher rate in the group receiving ADA compared to VDZ (21.8 vs. 12.6%; difference, −9.3 percentage points; 95% CI, −18.9 to 0.4). Despite a slight superiority of VDZ over ADA, more data are pending for consistency and class effect. The choice of biologics is, again, a shared decision between between clinician and patient.

Due to the ineffectiveness of IL-17i for CD, there have not been trials for their use in UC. As for IL-23i, there is an ongoing phase II/III trial of RZB for UC (NCT03398148).

In contrast to CD, the JAKi, TOF, was approved for use in moderate-to-severe UC based on three pivotal phase III OCTAVE studies, showing a significantly greater percentage of clinical remission at week 8 for induction, and remission at week 52 for maintenance in TOF compared to placebo group (186). UPA met the clinical remission, endoscopic improvement and histological improvement endpoints in a phase III induction trial for moderate-to-severe UC (187).

Prompt control of inflammation using topical corticosteroid is the first-line treatment for anterior uveitis in SpA. Typically, prednisolone acetate 1% eye drops are used as for severe AAU while milder corticosteroids such as dexamethasone 0.1% may be used for maintenance. A mydriatic drug is often prescribed together to reduce the development of posterior synechia and reduce pain from ciliary body spasm. Periocular corticosteroid injections or intravitreal implants can be used for more chronic cases. Adverse effects of corticosteroid in the eyes include cataract and ocular hypertension in up to 30% of patients. Oral corticosteroids may be used for acute management of severe and sight-threatening posterior uveitis such as vasculitis and cystoid macular edema, however, immunotherapy should be considered early in these cases to reduce recurrences (188). Traditional immunomodulators such as sulfasalazine (SSZ) and MTX may be tried although few data have supported their efficacy. Monoclonal antibody-TNFi including IFX, ADA, GOL and CZP are considered as effective treatment options for both acute flares and reducing recurrences of AAU (189). We summarize the major RCTs of therapeutic options for uveitis in Table 4. In a post-hoc analysis pooling data from four RCTs with TNFi in AS, the frequency of AAU flares was substantially lower among IFX or etanercept (ETN) treated than placebo treated patients. Lower frequency of AAU flares was seen in the open-labeled extension phase compared to the placebo phase of the trial (TNFi: 6.8 flares per 100 patient-years compared to PBO: 15.6 per 100 patient-years, p = 0.01) (198). ADA is the only TNFi licensed for treatment of non-infectious uveitis in adult following favorable results in 2 phase III RCTs. In the VISUAL I study, patients with active non-infectious intermediate, posterior uveitis or panuveitis were randomized to receive ADA or placebo after a prednisolone burst (60 mg) with tapering course. Patients treated with ADA were less likely than those treated with placebo to have treatment failure (hazard ratio, HR, 0.50; 95% CI, 0.36 to 0.70; P <0.001). The VISUAL II study recruited 226 patients with inactive, non-infectious intermediate, posterior, or panuveitis controlled by 10–35 mg/day of prednisone were randomized to ADA vs. placebo. All patients underwent a mandatory prednisolone tapering to 0 mg by week 19. The time to treatment failure was significantly longer in ADA compared to placebo arm (median >18 months vs. 8.3 months; HR 0.57, 95% CI 0.39–0.84; p = 0.004) (191). ADA is also licensed for juvenile idiopathic arthritis-related uveitis. In an open-label study in 93 AS patients with history of uveitis, GOL reduced uveitis flares compared to patients' historical control 12-month prior to initiation of GOL (192). There is an ongoing phase III 96-week open-label study for CZP in 115 patients with axial SpA and recurrent uveitis. In the 48-week interim analysis of 85 patients, uveitis flares were substantially reduced during the CZP treated period compared to the historical rates (64.0 and 31.5% respectively) (193). The use of ETN in the management of uveitis has diminished in favor of other TNFi because of its weaker ability in preventing flares.

Despite implicated in the pathogenesis of uveitis, inhibiting IL-17A was not effective for uveitis. In three RCTs, SEC failed to meet the primary efficacy endpoints (194). In another RCT comparing three doses of SEC, statistical higher response rates and remission on day 57 for the high dose regimen (30 mg/kg intravenously Q2W for 4 doses) was seen compared to the other two lower dose regimens, suggesting a higher dose intravenous regimen may be required to deliver SEC in therapeutic concentrations (195). Results are awaiting for two trials using UST in active sight-threatening uveitis (STAR) (196) and Behçet uveitis (STELABEC) (197), which may provide insight for its potential use in PsA related uveitis.

Minimal data exist for use of JAKi in uveitis. One phase 2 RCT evaluating filgotinib (FIL) in patients with active non-infectious uveitis (NCT03207815) is ongoing.