To evaluate the incidence of optic neuritis (ON) in patients using anti–tumor necrosis factor (TNF) alpha therapy.
Retrospective, population-based cohort study.
We identified new users of anti-TNF therapy (etanercept, infliximab, or adalimumab) or nonbiologic disease-modifying antirheumatic drugs (DMARDs) during 2000–2007 from the following data sources: Kaiser Permanente Northern California, Pharmaceutical Assistance Contract for the Elderly, Tennessee Medicaid, and National Medicaid/Medicare. Within this cohort, we used validated algorithms to identify ON cases occurring after onset of new drug exposure. We then calculated and compared ON incidence rates between exposure groups.
We identified 61 227 eligible inflammatory disease patients with either new anti-TNF or new nonbiologic DMARD use. Among this cohort, we found 3 ON cases among anti-TNF new users, occurring a median of 123 days (range, 37–221 days) after anti-TNF start. The crude incidence rate of ON across all disease indications among anti-TNF new users was 10.4 (95% CI 3.3–32.2) cases per 100 000 person-years. In a sensitivity analysis considering current or past anti-TNF or DMARD use, we identified a total of 6 ON cases: 3 among anti-TNF users and 3 among DMARD users. Crude ON rates were similar among anti-TNF and DMARD groups: 4.5 (95% CI 1.4–13.8) and 5.4 (95% CI 1.7–16.6) per 100 000 person-years, respectively.
Optic neuritis is rare among those who initiate anti-TNF therapy and occurs with similar frequency among those with nonbiologic DMARD exposure.
Optic neuritis (ON) is a heterogeneous condition with a number of potential etiologies including infectious, autoimmune, toxic, and demyelinating conditions. The incidence of optic neuritis is not well established, nor is the proportion of optic neuritis caused by various etiologies well documented. Modern estimates of disease rates are lacking, but population-based data from Minnesota in the late 1980s suggest idiopathic optic neuritis (ie, no identifiable cause) occurs at annual rates of 5 per 100 000 persons. More recently, certain biologic immunosuppressive therapies have been linked to triggering acute demyelinating disease, including optic neuritis. These therapies inhibit tumor necrosis factor (TNF) alpha and are now widely employed against rheumatoid arthritis, inflammatory bowel disease, psoriasis, and other inflammatory conditions, including noninfectious uveitis. Case reports of patients developing optic neuritis during anti-TNF use exist, although to date, no population-based, formal analytic studies have been conducted to explore the risk of this presumed complication.
To evaluate the association of optic neuritis and anti-TNF therapy, we first reviewed all spontaneous optic neuritis reports from the National Registry of Drug-Induced Ocular Side Effects (Casey Eye Institute, Portland, Oregon, USA). After finding a sufficiently large number of cases within this database to prompt further study, we then proceeded to evaluate this reported association in the context of a large, population-based collaboration called “Safety Assessment of Biologic ThERapy (SABER),” in which the rate of optic neuritis could be calculated and compared between patients starting biologic disease-modifying drugs (DMARDs) (ie, anti-TNF therapy) and similar patients starting nonbiologic disease-modifying antirheumatic drugs (DMARDs; eg, methotrexate, others).
National Registry of Drug-Induced Ocular Side Effects
The National Registry of Drug-Induced Ocular Side Effects passively collects reports of ocular toxic drug events from physicians within the United States and abroad. In addition, the registry is linked with the Food and Drug Administration’s (FDA) Medwatch system (Rockville, Maryland, USA) and the World Health Organization’s (WHO) Spontaneous Event Reporting Systems (Uppsala, Sweden) such that events reported to all 3 systems are retrievable within the National Registry of Drug-Induced Ocular Side Effects. To search this registry to identify anti-TNF-associated cases of optic neuritis reported from January 1, 1999 to September 22, 2011, we used the following search terms: “optic neuritis,” “optic neuropathy,” “etanercept,” “infliximab,” “adalimumab,” “golimumab,” “certolizumab,” and “tumor necrosis factor alpha antagonist.” For each reported case, we extracted descriptive data with regard to timing of optic neuritis onset after drug start, resolution of optic neuritis after drug cessation, patient demographics, and outcome information, where reported. These reports generally contained very little clinical information, making validation of optic neuritis cases not possible.
SABER Data Sources and Cohort Formation
We used data from 4 large automated US databases from 1998 through 2007: (1) National Medicaid and Medicare databases (Medicaid Analytic eXtract, 2000–2005; Medicare, 2000–2006; and Medicare Part D, 2006); (2) Tennessee Medicaid (TennCare, 1998–2005); (3) New Jersey Pharmaceutical Assistance to the Aged and Disabled and Pennsylvania Pharmaceutical Assistance Contract for the Elderly (PAAD/PACE, 1998–2006); and (4) Kaiser Permanente Northern California (KPNC, 1998–2007). We used validated algorithms to identify patients with immune-mediated inflammatory diseases of interest (rheumatoid arthritis, psoriasis, ankylosing spondylitis, and inflammatory bowel disease). Patients were eligible for inclusion if they had a baseline period of 365 days with continuous enrollment in the respective database preceding the first biologic or nonbiologic DMARD prescription fill. Patients with diagnoses for ≥2 autoimmune diseases, or history of any ON ICD-9 codes given prior to first DMARD prescription fill (T 0 ) (T 0 was defined as the “index date”), were excluded. Among potential cohort members, we identified new users of study DMARDs, defined as having filled a prescription for a study DMARD after 365 baseline days without prescriptions filled for the specific study medication or others in the same group. Study DMARDs were classified into 2 groups: TNF-α antagonists (including infliximab, adalimumab, and etanercept) and alternate DMARD regimens. For rheumatoid arthritis (RA), the alternate regimens were initiation of leflunomide, sulfasalazine, or hydroxychloroquine after use of methotrexate in the previous year (“methotrexate failures”). For inflammatory bowel disease (IBD) the comparison group was initiation of azathioprine or 6-mercaptopurine (AZA/6-MP). For the combined group of patients with psoriasis (PsO), psoriatic arthritis (PsA), or ankylosing spondylitis (AS) (PsO-PsA-AS), the comparison was initiation of nonbiologic DMARDs (including methotrexate, hydroxychloroquine, sulfasalazine, and leflunomide). Follow-up continued through the earliest of the following dates: death, loss of enrollment, development of optic neuritis, switch to another DMARD regimen or the discontinuation of the current regimen (defined as 30 days without medication), or study end, whichever came first. Patients who left the cohort could subsequently contribute new episodes of medication use if selection criteria were re-fulfilled. Patients could contribute episodes to more than 1 exposure group. A detailed description of SABER data development has been reported elsewhere.
Because it is unknown whether the use of these therapies could impart long-term risk for ON for some time after discontinuation, a secondary analysis (termed “current and past users”) was conducted in which new users were considered exposed during drug use and for up to 365 days after drug discontinuation, with such exposure censored at time of death, optic neuritis development, or initiation of an alternative DMARD regimen.
Optic Neuritis Case Finding
To identify optic neuritis cases, we used a combination of ICD-9 codes for optic neuritis and laboratory tests (eg, rapid plasmin reagin [RPR], and serum angiotensin converting enzyme [ACE] level, orbital magnetic resonance imaging, others) typically ordered during an acute optic neuritis examination. We identified optic neuritis cases using the following algorithm: any patient given 1 inpatient or outpatient ICD-9 code for optic neuritis (377.30 or 377.32) with evidence of serum ACE testing within 90 days of the code OR any patient given an optic neuritis ICD-9 code 3 times within 60 days. Prior to its use, we determined that this case-finding algorithm had a positive predictive value of 100% by conducting a manual review of all cases with ≥1 optic neuritis diagnostic codes (n = 135) during 2008 at Oregon Health & Science University (OHSU) and the Portland Veteran’s Administration Medical Center in Portland, Oregon (data not shown).
Because the above case-finding algorithms were of high positive predictive value (PPV), as part of additional sensitivity analysis, we constructed optic neuritis case-finding algorithms that were less stringent and presumably more sensitive. For this analysis, we used ICD-9 codes only (ie, we did not require serum ACE testing) and defined optic neuritis as a physician-given inpatient or outpatient ICD-9 code (377.30 or 377.32) given ≥2 times within 90 days.
Covariates measured within the baseline period prior to drug start included demographics (age, sex, race, residence [urban/rural], nursing home/community dwelling, calendar year); generic markers of comorbidity, including number of hospitalizations, outpatient and emergency room visits, and number of different medication classes filled; and surrogate markers of inflammatory disease severity, including extraarticular disease manifestations, number of intraarticular and orthopedic procedures, number of laboratory tests ordered for inflammatory markers, and use of corticosteroids.
We calculated crude incidence rates of optic neuritis by underlying disease group (RA, IBD, PsO-PsA-AS) as well as by first DMARD exposure (new biologic users vs nonbiologic DMARD). We compared crude incidence rates between treatment groups. All analyses were done in SAS (SAS Institute, Cary, North Carolina, USA); there were insufficient data to adjust for potentially confounding factors.
National Registry of Drug-Induced Ocular Side Effects
From the national registry collecting passively reported data, we identified 358 reports of optic neuritis occurring in association with anti-TNF therapy. Seventy-five percent were female, median age 44 years (range, 6–78 years), with median time between anti-TNF start and diagnosis of 182 days (range, 2–388 days). Cases were most numerous with etanercept (n = 169), with fewer reported with monoclonal antibodies infliximab (n = 122), adalimumab (n = 55), golimumab or certolizumab (n = 5), or multiple drugs (n = 7). Re-challenge (n = 5) and de-challenge (n = 33) information was available for very few reported cases. One of 5 patients re-challenged with drug had recurrent symptoms, whereas 13 of 33 patients stopping drug had continued symptoms despite drug stoppage.
“Safety of Biologic Therapy (SABER) Cohort Study:”
Within SABER, we identified 61 227 eligible inflammatory disease patients with either new anti-TNF or new nonbiologic DMARD use. Baseline characteristics between those treated with anti-TNF therapy and those in comparison nonbiologic groups were similar within all disease cohorts ( Table 1 ).
|Variable||Rheumatoid Arthritis N (%)||Inflammatory Bowel Disease N (%)||Psoriasis, Psoriatic Arthritis, Ankylosing Spondylitis N (%)|
|Anti-TNF||Nonbiologic DMARD||Anti-TNF||Nonbiologic DMARD||Anti-TNF||Nonbiologic DMARD|
|Age, mean (SD), y||57.73 (14.53) a||58.47 (14.27)||40.39 (16.13) a||40.38 (17.80)||48.82 (15.33) a||52.19 (16.82)|
|Female||20 955 (85.9) a||10 205 (86.3)||2559 (66.5)||4330 (63.1)||2854 (56.1)||4331 (61.4)|
|White||15 244 (62.5) a||7340 (62.0)||3010 (78.2) a||5075 (73.9)||3716 (73.0) a||4986 (70.7)|
|Black||3927 (16.1)||1831 (15.5)||586 (15.2)||993 (14.5)||357 (7.0)||576 (8.2)|
|Other||5212 (21.4)||2659 (22.5)||254 (6.6)||799 (11.6)||1017 (20.0)||1489 (21.1)|
|Nursing home resident||992 (4.1) a||493 (4.2)||99 (2.6)||167 (2.4)||146 (2.9) a||334 (4.7)|
|≥1 hospitalization during baseline||6995 (28.7) a||3305 (27.9)||2133 (55.4) a||3387 (49.3) a||1042 (20.5) a||1613 (22.9)|
|Charlson-Deyo comorbidity score mean (SD)||1.72 (1.13) a||1.73 (1.17)||0.51 (0.95)||0.47 (0.90)||0.74 (1.13) a||0.79 (1.18)|
|≥1 inflammatory marker tested||8955 (36.7) a||4380 (37.0)||1094 (28.4) a||2058 (30.0)||1045 (20.5) a||1370 (19.4)|
|Mean glucocorticoid use, prednisone equivalents|
|None||9732 (39.9) a||5079 (42.9)||1714 (44.5) a||2773 (40.4)||4038 (79.3) a||5461 (77.5)|
|<5 mg||7552 (31.0)||3650 (30.9)||609 (15.8)||973 (14.2)||700 (13.8)||1162 (16.5)|
|5–10 mg||2495 (10.2)||1056 (8.9)||933 (24.2)||1954 (28.5)||156 (3.1)||153 (2.2)|
|>10 mg||4604 (18.9)||2045 (17.3)||594 (15.4)||1167 (17.0)||196 (3.9)||275 (3.9)|
|Any orthopedic surgery||1752 (7.2) a||633 (5.4)||66 (1.7)||85 (1.2)||189 (3.7) a||177 (2.5)|
|Any intraarticular injection||8607 (35.3) a||3596 (30.4)||198 (5.1)||259 (3.8)||695 (13.7) a||817 (11.6)|
|Chronic obstructive lung disease||3241 (13.3) a||1584 (13.4)||311 (8.1)||484 (7.0)||502 (9.9) a||870 (12.3)|
|Cerebrovascular disease||947 (3.9) a||419 (3.5)||82 (2.1)||118 (1.7)||116 (2.3) a||248 (3.5)|
|Diabetes||4618 (18.9) a||2266 (19.2)||336 (8.7)||541 (7.9)||1021 (20.1)||1337 (19.0)|
|Obesity||2153 (8.8) a||1227 (10.4)||276 (7.2)||676 (9.8)||697 (13.7) a||953 (13.5)|
|Adalimumab||5888 (24.1)||—||118 (3.1)||—||294 (5.8)||—|
|Etanercept||10 283 (42.2)||—||—||—||4270 (83.9)||—|
|Infliximab||8212 (33.7)||—||3732 (96.9)||—||526 (10.3)||—|
|Hydroxychloroquine||—||5730 (48.4)||—||—||—||569 (8.1)|
|Leflunomide||4569 (38.6)||133 (1.9)|
|Sulfasalazine||—||1531 (12.9)||—||—||—||858 (12.2)|
Primary Analysis (“Current User” Analysis)
In primary current user analysis, we identified 3 optic neuritis cases in 28 898 person-years of exposure among anti-TNF new users with RA (n = 2) or IBD (n = 1), occurring a median of 123 days (range, 37–221 days) after anti-TNF start. These 3 cases occurred in female subjects, median age 60 years (range 40–65 years), with infliximab (n = 2) and etanercept (n = 1). No case occurred in 15 799 person-years of exposure within the nonbiologic DMARD comparison group. Crude incidence rates of optic neuritis across all disease indications among anti-TNF new users was 10.4 (95% CI 3.3–32.2) per 100 000 person-years ( Table 2 ).