Article Text

Download PDFPDF

Adalimumab for the treatment of refractory active and inactive non-infectious uveitis
  1. Jonathan TL Lee1,
  2. William B Yates2,3,
  3. Sophie Rogers1,
  4. Denis Wakefield4,
  5. Peter McCluskey2,3,
  6. Lyndell L Lim1,5
  1. 1 Centre for Eye Research Australia, University of Melbourne, East Melbourne, Victoria, Australia
  2. 2 Department of Ophthalmology, Sydney Eye Hospital, Sydney, New South Wales, Australia
  3. 3 Save Sight Institute, The University of Sydney, Sydney, New South Wales, Australia
  4. 4 School of Medical Sciences, University of New South Wales, Kensington, New South Wales, Australia
  5. 5 Ocular Immunology Clinic, Royal Victorian Eye and Ear Hospital, East Melbourne, Victoria, Australia
  1. Correspondence to Associate Professor Lyndell L Lim, Centre for Eye Research Australia, University of Melbourne, Royal Victorian Eye and Ear Hospital, East Melbourne, VIC 3002, Australia; limllp{at}


Background/aims To compare the efficacy of adalimumab in eyes with active and inactive non-infectious uveitis in the real-world setting.

Methods Multicentre, retrospective, chart review of patients with refractory non-infectious uveitis treated with adalimumab. Main outcome measures included reduction of prednisolone dose, ability to taper immunosuppressives and a composite endpoint of treatment failure encompassing active inflammatory chorioretinal or retinal vascular lesions, intraocular inflammation grade and visual acuity.

Results Thirty-seven eyes of 22 patients were studied. Mean follow-up was 20.1 months (median: 13). Most had either posterior or panuveitis (n=12, 55%). Mean duration of uveitis at baseline was 83.2 months (median: 61), where the majority (n=15, 68%) had already been treated with two or more conventional immunosuppressive agents in addition to prednisolone. Oral prednisolone was reduced to ≤10 mg/day in 9 of 12 patients (75%) by 6 weeks. At 6 months of therapy, nine (90%) of the active eyes achieved a 2-step improvement in anterior chamber inflammation, with six (60%) demonstrating a similar improvement in vitreous haze grade. Almost all (n=17, 94%) of the initially inactive eyes maintained clinical quiescence at this time point. The incidence rate of treatment failure during follow-up was 88 per 100 eye-years for the active eyes and 24 per 100 eye-years for the initially inactive eyes. There were no serious adverse effects.

Conclusion Adalimumab appears to reduce the corticosteroid burden in active and inactive non-infectious uveitis in the real-world setting. Inflammatory activity at the time of adalimumab commencement may determine long-term treatment success.

  • Inflammation
  • Adalimumab
  • Uveitis
  • Ophthalmology
  • Tumor Necrosis Factor-alpha
  • Retrospective Studies
  • Chronic Disease
  • Adult
  • Child
  • Humans
  • Treatment Outcome
  • Treatment Failure
  • antibodies
  • monoclonal
  • humanised

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Uveitis encompasses a diverse group of intraocular inflammatory diseases that accounts for approximately 10% of blindness in developed countries and is a common cause of visual disability in the working-age population.1 Non-infectious uveitis is characterised by a T-cell-mediated immune response, whereby there is an imbalance between regulatory and inflammatory mechanisms.2 Corticosteroids remain first-line therapy but their long-term use is limited by their systemic side-effects, and patients with chronic ocular inflammation benefit from the addition of a corticosteroid-sparing immunomodulatory agent or biological agent.3 4

Tumour necrosis factor alpha (TNF-α) is a proinflammatory cytokine which plays a critical role in many autoimmune diseases, including uveitis.5 6 Adalimumab (Humira, AbbVie, Chicago, Illinois, USA) is a completely human monoclonal IgG1 TNF-α antibody that was first approved by the US Food and Drug Administration (FDA) for the treatment of rheumatoid arthritis in 2002. It has since gained approval for use in patients with psoriatic arthritis, ankylosing spondylitis, ulcerative colitis, Crohn’s disease, chronic plaque psoriasis and juvenile idiopathic arthritis.7 Following two randomised placebo-controlled trials VISUAL-18 and VISUAL-2,9 adalimumab has now been approved for the treatment of adults with non-infectious intermediate, posterior and panuveitis by the European Medicines Agency10 and the FDA,11 although international experts suggest that adalimumab may also be useful in children and patients with anterior uveitis.12

VISUAL-1 and VISUAL-2 demonstrated that in adults with active and inactive non-infectious uveitis, fortnightly adalimumab therapy significantly reduced the risk of treatment failure relative to placebo across a range of clinical outcomes.8 9 In these trials, patients were excluded if they had isolated anterior uveitis or if they were receiving more than one concomitant immunosuppressive drug (excluding corticosteroids) within 4 weeks of the baseline visit. Furthermore, all participants underwent a mandatory prednisolone taper to 0 mg by 158 or 19 weeks.9 Given these restrictions, it cannot be presumed that a similar efficacy would be achieved in real-world clinical practice with a more heterogeneous patient population.

The purpose of this study was to evaluate the safety and efficacy of adalimumab for the management of refractory non-infectious uveitis in a real-world setting across three tertiary referral centres in Australia.


A multicentre retrospective chart review was performed of all patients with non-infectious uveitis treated with adalimumab between November 2010 and October 2016 at the Royal Victorian Eye and Ear Hospital (RVEEH), St Vincent’s Hospital Sydney and Sydney Eye Hospital. The study adhered to the tenets of the Declaration of Helsinki.

Collected data included patient demographics, clinical diagnosis, ocular examination findings at baseline and follow-up visits, prior local and systemic therapies, response to therapy and side effects.

Inflammatory status of each eye at baseline was classified as ‘active’ (based on the clinical findings of at least one active inflammatory chorioretinal or retinal vascular lesion, anterior chamber cell grade of 1+ or higher or vitreous haze grade of 1+ or higher, according to the Standardisation of Uveitis Nomenclature (SUN) Working Group13 and adapted National Eye Institute criteria14) or ‘inactive’ (corresponding to eyes without active inflammatory chorioretinal or retinal vascular lesions as well as an anterior chamber cell grade and vitreous haze grade of 0.5+ or less). For the analysis of systemic outcomes, patients who had one or both eyes active at baseline were classified as ‘active’, while those with both eyes inactive at baseline were classified as ‘inactive’.

The outcome variables included corticosteroid-sparing effect, degree of intraocular inflammation, central macular thickness (CMT), visual acuity and time-to-treatment failure, as defined by the VISUAL-1 and VISUAL-2 studies.8 9 Outcomes were collected at initiation of adalimumab, after 6 weeks and every 6 months thereafter until the final available follow-up visit. Corticosteroid-sparing success was evaluated based on the reduction of the prednisolone dose to ≤10 mg/day. Anterior chamber cell counts and vitreous haze were graded clinically as described above. Macular thickness was measured by optical coherence tomography (OCT) when available with an OCT Spectralis (Heidelberg Engineering, Franklin, Massachusetts, USA) or a Cirrus OCT (Carl Zeiss Meditec, Dublin, California, USA). Macular oedema was defined as a CMT of ≥340 µm for eyes measured with an OCT Spectralis and ≥320 µm for eyes measured with a Cirrus OCT. Best-corrected visual acuity (BCVA) was converted to logMAR (logarithm of the minimum angle of resolution) for analysis.15

Time-to-treatment failure was a composite endpoint derived from VISUAL-1 and VISUAL-2.8 9 Treatment failure after week 6 for active eyes, and at or after week 2 for inactive eyes, was defined as the presence of at least one of the following criteria: new active inflammatory chorioretinal or retinal vascular lesions relative to baseline; a two-step increase in anterior chamber cell grade relative to the last visit; a two-step increase in vitreous haze grade relative to the last visit or a worsening of BCVA by 15 letters or more on the Early Treatment Diabetic Retinopathy Study chart relative to the previous best recorded acuity. For active eyes only, anterior chamber cell grade that did not improve to 0.5+ or lower or vitreous haze grade that did not improve to 0.5+ or lower by week 6 were additional treatment failure criteria.

Adalimumab was administered at the standard dose of 40 mg subcutaneously every 2 weeks. An attempt was made to taper corticosteroids if clinically appropriate. Reasons for treatment failure were noted.

Data were analysed using Stata software V.14.2 (Stata, College Station, Texas, USA). Categorical variables were summarised as percentages. Continuous variables were summarised as means with 95% CIs (indicated by placing the upper and lower bounds in parentheses after each estimate) or medians with an IQR if they did not appear to be normally distributed. Incidence rates for clinical events were calculated at time points of interest in a per-eye or per-person analysis depending on the relevant outcome. Time-to-failure outcomes were assessed with Kaplan-Meier survivor functions using the prespecified treatment failure criteria. Total number of first treatment failure events was divided by the total time contributed by the active and inactive eyes in years to give a failure rate in eye-years.


We studied 37 eyes of 22 patients with an average age of 36.6 years (range: 11–64 years). Mean follow-up was 20.1 months (median: 13, IQR 8–30; range: 2–57 months). Clinical characteristics of the cohort are summarised in table 1. The majority of patients had posterior or panuveitis (n=12, 55%), approximately half of whom had concurrent retinal vasculitis at baseline (n=7). The remainder of patients were treated for anterior uveitis (n=10, 45%). The most common aetiological diagnoses were Behçet’s disease (n=8, 36%), followed by undifferentiated inflammation (n=5, 23%) and ankylosing spondylitis (n=4, 18%).

Table 1

Clinical characteristics of the cohort at baseline

The detailed patient demographics and treatment history are provided in online supplementary table S1. There were 19 eyes (51%) with active inflammation at baseline and 18 eyes (49%) without active inflammation. In total, 13 patients were considered ‘active’ as they had at least one eye with active inflammation at baseline and the remaining 9 patients were considered ‘inactive’ as both eyes were quiescent at baseline. The mean duration of inflammation prior to adalimumab was 83.2 months (median: 61, IQR 27–114; range: 12–266).

Supplementary file 1

Prednisolone dose at the time of adalimumab commencement ranged from none to 75 mg/day, with a mean of 18.8 mg/day (median: 15, IQR 0–25). Of the 16 patients (10 active, 6 inactive) using prednisolone at baseline, 9 (6 active, 3 inactive) were able to discontinue corticosteroids altogether. Table 2 summarises the outcomes related to systemic therapy, expressed per-person. Due to the variable follow-up period in this retrospective analysis, outcomes are reported as the incidence of a prespecified event occurring at each time point. The variation in denominators represents those subjects deemed ‘at risk’ for each specific outcome at each time point and gradually declines over time as fewer patients reached extended periods of follow-up. By 6 weeks of therapy, prednisolone was weaned to ≤10 mg/day in 5/8 patients (62.5% (30.4 to 86.5)) with active disease and in 4/4 patients (100% (45.4 to 100)) with inactive disease. In patients with active disease, the proportion maintaining a corticosteroid load of ≤10 mg/day peaked at 6 months (100% (62.8 to 100)) before falling at 12 months (85.7% (46.7 to 99.5)) and 18 months (40% (11.6 to 77.1)). In contrast, for those with inactive disease, the steroid-sparing effect was sustained in all 3/3 patients (100% (38.3 to 100)) at 12 months and in both patients (100% (29.0 to 100)) at 18 months.

Table 2

Systemic outcomes of adalimumab therapy, expressed per-person ‘at risk’

Fifteen patients (68%) had previously trialled two or more conventional immunosuppressive agents prior to adalimumab. By 6 months, a reduction in the number of concomitant immunomodulatory agents to ≤1 occurred in 2/3 patients (66.7% (20.2 to 94.4)) with active disease and 1/2 patients (50% (9.5 to 90.5)) with inactive disease (table 2).

Table 3 shows the ocular outcomes of adalimumab therapy, expressed pereye. At baseline, 63% of the active eyes had an anterior chamber cell grade of 1+ or more, whereas 53% had a vitreous haze grade of 1+ or more. Improvement in anterior chamber cells in active eyes by two or more grades (eg, from 3+ to 1+) occurred in 11/11 eyes (100% (70 to 100)) by 6 weeks, in 7/8 eyes (87.5% (50.7 to 99.9)) by 12 months and 5/5 eyes (100% (51.1 to 100)) by 18 months. During the same durations of follow-up, an equivalent improvement in vitreous haze grade in active eyes was seen in 5/10 eyes (50% (23.7 to 76.3)), 7/8 eyes (87.5% (50.7 to 99.9)) and 4/7 eyes (57.1% (25.0 to 84.2)), respectively. In contrast, ongoing clinical quiescence in inactive eyes was defined as an anterior chamber and vitreous haze grade of 0.5+ or lower at each visit and was sustained in over 88% of those ‘at risk’ throughout 24 months of follow-up.

Table 3

Therapeutic outcomes of adalimumab therapy, expressed per-eye ‘at risk’

The mean CMT of the eyes at baseline with available OCT data was 317 µm (232 to 401) for active eyes (n=11) and 450 µm (207 to 692) for inactive eyes (n=11). In active eyes, the mean change in CMT relative to baseline was −49.9 µm (−154 to 54.4) at 6 months, −73.2 µm (−245 to 99.0) at 12 months and −114 µm (−315 to 88.2) at 18 months. For initially inactive eyes, the mean change in CMT was less prominent, measuring +9.25 µm (−19.8 to 38.3) at 6 months, +14.1 µm (−67.5 to 95.8) at 12 months and −22.3 µm (−70.4 to 26.7) at 18 months. Four eyes (three with panuveitis, one with chronic anterior uveitis) had a history of epiretinal membrane in the setting of prior cataract surgery and intravitreal triamcinolone. Three of these eyes had a surgical vitrectomy and epiretinal membrane peel prior to adalimumab commencement, while the remaining eye had a slight worsening in macular thickness during the study, from 301 µm at baseline to 325 µm after 6 months of adalimumab therapy.

The mean logMAR visual acuity at baseline was 0.40 (Snellen equivalent of 20/50) in active eyes and 0.30 (Snellen equivalent of 20/40) in inactive eyes. Throughout the follow-up period most of the eyes maintained visual stability regardless of inflammatory status, as demonstrated by the negligible changes in the mean logMAR visual acuity from baseline (figure 1). The proportion of eyes that experienced any reduction in visual acuity (five letters or more) relative to the baseline examination was low, measuring 20.6% (n=7) at 6 weeks, 21.2% (n=7) at 6 months, 14.8% (n=4) at 12 months and 16.7% (n=2) at 18 months.

Figure 1

Mean change in logMAR visual acuity from baseline measurement in (A) active eyes and (B) inactive eyes. Error bars represent the  95% CI.

When determining the incidence rate of treatment failure in our cohort, the data indicated that 88% of the active eyes receiving adalimumab would have treatment failure if followed for 12 months (88.1 per 100 eye-years). In contrast, only 24% of the initially inactive eyes would have recurrence and treatment failure by the same time point (24.1 per 100 eye-years). Treatment failure occurred at the following weeks for active eyes: 6, 6, 6, 8, 14, 16, 17, 21, 54 and 107. For initially inactive eyes, treatment failure occurred at the following weeks: 6, 6, 7, 23 and 80. Per the Kaplan-Meier survivor function, after 12 months 49.8% (28.2 to 76.1) of eyes in the active group had failed treatment, while only 22.2% (9.0 to 48.9) of inactive eyes had failed at this time (figure 2). By 18 months, the figure for treatment failure in active eyes had increased to 62.3% (36.2 to 88.0) but for inactive eyes it remained at 22.2% (9.0 to 48.9). Of the prespecified failure endpoints, the most frequent reason for treatment failure in active eyes was due to vitreous haze (six eyes) (figure 3). The least frequent reasons for treatment failure in this group were new inflammatory retinal lesions and worsening anterior chamber grade, at two eyes each. Three of the active eyes had more than one reason for treatment failure: one eye developed 1+ anterior chamber and vitreous haze activity along with new retinitis, another had a two-step increase in vitreous haze associated with a significant worsening of BCVA, and the third demonstrated new vascular sheathing with 1+ vitreous haze. Another two active eyes experienced a gradual but significant decline in BCVA that could be partly explained by a posterior subcapsular cataract—both these eyes had previously received intravitreal triamcinolone prior to adalimumab. In eyes with inactive uveitis at baseline, the most frequent reason for treatment failure was an increase in anterior chamber cell grade (three eyes), while none of the eyes had failed due to a worsening of BCVA. Treatment failure by 6 months occurred more frequently in patients with active or bilateral disease and was associated with a shorter duration of inflammation prior to adalimumab and a higher number of prior immunomodulators (table 4). Tests for statistical significance were not performed due to the small sample sizes involved.

Table 4

Features associated with treatment failure by 6 months

Figure 2

Kaplan-Meier plot of the rate of treatment failure for any reason in (A) active eyes and (B) inactive eyes at baseline.

Figure 3

Reasons for treatment failure in active and inactive eyes. In three of the active eyes, there was more than one reason for treatment failure.

Adalimumab was discontinued in two patients. One patient with idiopathic ischaemic retinal vasculitis had adalimumab withdrawn after 3 months as logMAR visual acuity remained worse than 1.3 (Snellen equivalent of 20/400) in both eyes. Another patient with panuveitis and retinitis secondary to sarcoidosis self-ceased therapy after 6 months after a subjective improvement in her symptoms, despite interval development of 2+ vitreous haze. There were no serious adverse effects. The most common side effects were rash and upper respiratory tract infection, occurring in two patients each. The remainder of side effects with one event each included injection site reaction, acne and panniculitis.


In this retrospective review, we report our clinical experience with adalimumab in the treatment of non-infectious uveitis. Thirty-seven eyes (22 patients), half of which had active inflammation at baseline, were treated during an average follow-up period of 20.1 months. This case series provides additional evidence of the effectiveness of adalimumab as a glucocorticoid-sparing agent for the control of uveitis in the real-world setting.

VISUAL-1 and VISUAL-2 demonstrated the effectiveness of adalimumab relative to placebo in the treatment of non-infectious intermediate, posterior and panuveitis in a clinical trial setting.8 9 The mean duration of uveitis in those who received adalimumab was 40.2±51.2 months in VISUAL-1 and 59.5±64.5 months in VISUAL-2, whereas in our cohort there was more chronic disease, with a mean duration of uveitis of 83.2±71.9 months. Furthermore, the majority of our cohort had been pretreated with two or more immunomodulatory agents—for example, 41% were receiving concomitant methotrexate at baseline, as compared with 8% and 17% in VISUAL-1 and VISUAL-2, respectively. Despite these differences, we report similar median time to treatment failure results when using the VISUAL-1 and VISUAL-2 criteria: almost half (49.8%) the active eyes failed therapy by 21 weeks (comparable to 24 weeks in VISUAL-1), while fewer than half the inactive eyes had treatment failure after 18 months (comparable to VISUAL-2). The most frequent endpoint indicating treatment failure in the adalimumab groups in VISUAL-1 and VISUAL-2 was an increase in anterior chamber cell grade. On the contrary, in our subset of active eyes, vitreous haze was the most frequent failure endpoint, usually in association with new retinal lesions or a significant reduction in vision. This could be due to our cohort having a larger number of patients with more recalcitrant posterior uveitis, as reflected by the greater proportion who had failed two or more immunomodulatory agents prior to adalimumab.

In contrast to VISUAL-1 and VISUAL-2, we included 10 patients (16 eyes) with isolated anterior uveitis, most of whom had associated ankylosing spondylitis (40%) or juvenile idiopathic arthritis (20%). In this subset of patients, treatment failure due to a flare in anterior chamber activity occurred in four eyes (25%). The efficacy of adalimumab therapy in reducing the rate of anterior uveitis flares in ankylosing spondylitis has been previously reported in several studies.16 17 In a prospective randomised placebo-controlled trial of 90 children with active juvenile idiopathic arthritis-associated uveitis on stable doses of methotrexate, the addition of adalimumab significantly reduced the rate of treatment failures by more than half (27% vs 60%) with a longer mean duration of sustained quiescence (estimated treatment effect, 164.8 days).18

This study showed no clear improvement in visual acuity following adalimumab therapy. In a large prospective study of 131 patients on adalimumab for uveitis, visual acuity remained stable in 75.4% of eyes (n=113) after 6 months.19 Durrani et al 20 reported no significant change in visual acuity after 3–6 months of adalimumab therapy in 32 patients with ocular inflammatory disease, and Arcinue et al 21 reported similar results at 1 year in an equally heterogeneous group of 49 patients. In contrast, Bawazeer et al 22 published a series of 21 eyes with ocular Behçet’s disease, in which 81% had an improvement of visual acuity by three or more lines (average 4.33 lines) after 10 months of follow-up. In both VISUAL-1 and VISUAL-2, patients in the adalimumab group were significantly less likely than the placebo group to have treatment failure due to a worsening in the BCVA.8 9

Treatment failure in our cohort was most likely to occur between 6 weeks and 6 months, particularly in eyes with active inflammation at baseline. This corresponds to the period of most aggressive prednisolone dose reduction. At 6 months, the mean daily corticosteroid dose had decreased from 29.5 mg to 4.5 mg in patients with active disease and from 9.6 mg to 2.2 mg in patients with inactive disease. As corticosteroids are weaned, there is a risk of exacerbation of the intraocular inflammation despite biological therapy. This reflects the findings of the VISUAL-1 study8 and the experiences of other groups. In a series of 21 patients with uveitis on TNF-α blockers, Mercier et al 23 observed a reduction in mean daily prednisolone dose from 19.7 mg to 5.2 mg at 1 year, however, control of inflammation in those receiving adalimumab (n=8) dropped from 85.7% at 6 months to 66.7% at 12 months. Similarly, in a prospective trial of 31 patients with non-infectious uveitis, the clinical response rate to adalimumab dropped from 68% at 10 weeks to 39% at 50 weeks following an average prednisolone dose reduction of 90%.24

Another possible explanation for the delayed recurrence of inflammation is a loss of biological efficacy in the long-term. In a retrospective case series studying adalimumab in refractory paediatric uveitis, 12 of 14 (85.7%) patients demonstrated improvement in their intraocular inflammation within 6 weeks without the use of adjuvant corticosteroids; however, over the ensuing 15–58 months of follow-up, only 60% had effective long-term control of uveitis flares.25

Adalimumab was well tolerated in our cohort, with the most common side effects of skin rashes and respiratory tract infections being consistent with other reports in the literature.21 25–27 A number of studies have also reported more serious adverse events including but not limited to: multiple sclerosis, malignancy, lupus-like reaction and reactivation of tuberculosis.8 9 19 23 The mean follow-up period of our retrospective study was too short and sample size too small to detect rarer adverse events.

Limitations of this study include its retrospective nature, which may introduce potential biases. For example, the safety profile may be underestimated if less severe side effects were not recorded in the patients’ notes. Other limitations include the variable follow-up period and the lack of a comparative arm. One of the strengths of this study is that the data reflects routine clinical practice in a real-world patient population, where physicians are guided by their clinical judgement rather than a strict predefined protocol.

In conclusion, treatment with adalimumab seems to be efficacious in reducing the corticosteroid burden in active and inactive non-infectious uveitis. These real-world data from patients with refractory disease mirror the results from recent VISUAL-1 and VISUAL-2 clinical trials, suggesting that inflammatory activity at the time of adalimumab commencement may play a key role in determining long-term treatment success.


We thank Lauren AB Hodgson for her contribution to the study data collection.



  • Contributors All authors made significant contributions to this paper and approved the final version to be published.

  • Funding LLL is supported by a National Health & Medical Research Council Early Career Fellowship (#1109330). Centre for Eye Research Australia receives Operational Infrastructure Support from the Victorian Government in Australia.

  • Competing interests None declared.

  • Patient consent Detail has been removed from this case description/these case descriptions to ensure anonymity. The editors and reviewers have seen the detailed information available and are satisfied that the information backs up the case the authors are making.

  • Ethics approval Royal Victorian Eye and Ear Hospital Human Research Ethics Committee and the South East Sydney local health district Human Research Ethics Committee.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles

  • At a glance
    Keith Barton James Chodosh Jost B Jonas