Tumor necrosis factor inhibitors in patients with Takayasu arteritis: Experience from a referral center with long-term followup




To report a single-center experience with the use of tumor necrosis factor (TNF) inhibitors in patients with Takayasu arteritis (TA).


We retrospectively studied a cohort of patients with refractory TA evaluated at our institution and treated with TNF inhibitors. American College of Rheumatology criteria for TA were used for inclusion. Disease activity was assessed according to the National Institutes of Health criteria.


We included 20 patients (19 women, 17 white) with a mean ± SD age of 33 ± 10.2 years and a median disease duration of 15.9 months (interquartile range [IRQ] 2–32.7 months) prior to the use of TNF inhibitors. Before the use of TNF inhibitors, all 20 patients received prednisone. Other medication use included methotrexate (18 patients), azathioprine (5 patients), mycophenolate mofetil (3 patients), and cyclophosphamide (3 patients). Seventeen patients (85%) received infliximab, 2 patients (10%) received adalimumab, and 1 patient (5%) received etanercept. The median duration of treatment with TNF inhibitors was 23.0 months (IQR 8.7–38.9 months). Treatment with TNF inhibitors resulted in disease remission in 18 (90%) of 20 patients and sustained remission in 10 patients (50%). Ten (83%) of 12 patients were able to taper prednisone below 10 mg and 7 patients discontinued prednisone. However, 6 of the 18 patients achieving remission experienced relapse while receiving TNF inhibitors. Eleven patients (55%) discontinued TNF inhibitors for the following reasons: relapse, persistently active disease, lack of corticosteroid-sparing effect, adverse effects (4 patients), and other reasons (4 patients).


In this study, treatment with TNF inhibitors induced remission, including sustained remission in patients with refractory TA. However, 33% of patients experienced disease relapse while receiving TNF inhibitors and 20% discontinued treatment because of adverse events.


Takayasu arteritis (TA) is a rare granulomatous inflammatory vasculitis affecting the aorta and its branches. The incidence varies in different countries and TA is considered more frequent in Asia and is less common in North America, with an estimated incidence of 2.6 per million per year based on a population-based study from Olmsted County, Minnesota (1, 2). It is more frequently observed in women. Although the age at onset is primarily between 10 and 40 years, TA may present at a later age.

The vascular inflammation in TA causes damage leading to arterial stenosis, occlusion, dilatation, and/or aneurysm formation. End-organ ischemia due to these vascular changes is responsible for most of the clinical manifestations of the disease (3). The morbidity related to TA is significant despite current therapeutic regimens (4). To date, there have been no published randomized controlled clinical trials in TA that might guide optimal therapy. Typically, high-dose corticosteroids (CS) are used initially to decrease the inflammatory burden and induce remission of the disease. However, a large proportion of patients relapse as CS are tapered and immunosuppressive medications are required to treat relapses and avoid excessive use of CS. Methotrexate, azathioprine, and/or mycophenolate mofetil are the most commonly used immunosuppressant medications (5).

Despite intensive medical treatment, approximately one-half of the patients with TA require vascular intervention (vascular surgery or percutaneous angioplasty) (4). In many patients, conventional therapies are unable to induce sustained remission of TA and progression of vascular damage occurs in spite of treatment (5). Therefore, there remains an unmet need for more effective therapy for TA.

It has been speculated that since tumor necrosis factor (TNF) plays a role in granuloma formation, it is plausible that TNF inhibitors may be effective in this condition. Apart from isolated case reports (6–10), data regarding the effect of TNF inhibitors in TA are derived from an open-label study of 15 patients (11) and a retrospective single-center study of 25 patients (12). These previous studies have suggested that TNF inhibitors may be effective in patients with refractory TA.

The aim of this study was to report our experience with the use of TNF inhibitors (infliximab, adalimumab, or etanercept) for TA from a single referral center. We describe the initial therapeutic effect of TNF inhibitors and also report on extended followup of these patients.

Significance & Innovations

  • We report our experience with the use of tumor necrosis factor (TNF) inhibitors for Takayasu arteritis (TA) in a large series of patients compared to 2 previous pilot studies.

  • Treatment with TNF inhibitors can induce remission, including sustained remission in patients with refractory TA.

  • Thirty-three percent of the patients experienced disease relapse while receiving TNF inhibitors and 20% discontinued treatment because of adverse events.


We conducted a retrospective single-center study of patients with TA treated with TNF inhibitors seen at Mayo Clinic, Rochester, Minnesota, between January 1, 2004, and December 31, 2009. The study was approved by the Institutional Review Board at Mayo Clinic. Patients who denied authorization of the use of their medical records for research purposes were excluded.

Using International Classification of Diseases, Ninth Revision codes for TA, we identified all patients evaluated at our institution between 2004 and 2009. All medical records were reviewed to confirm the diagnosis. We included patients who met the American College of Rheumatology criteria for TA (13). Additionally, we only included patients who were evaluated within 1 year of diagnosis. We identified all TA patients treated with infliximab, etanercept, or adalimumab by review of the medical records.

Data were abstracted using a standardized data collection form. This included details on patient demographics, treatment received before TNF inhibitors, indication for TNF inhibitors, type of TNF inhibitors used, laboratory evaluations, and patient outcomes. Choice of treatment and frequency of followup were based on the treating rheumatologists' discretion. However, physical examination, laboratory evaluation, and periodic vascular imaging study were part of the evaluations used in the long-term followup of these patients.

The effect of TNF inhibitors during followup was categorized based on disease activity as remission, sustained remission, or relapse. Additionally, we also evaluated the CS-sparing effect of TNF inhibitors and adverse effects of therapy. Disease activity was assessed using the definition coming from a previous National Institutes of Health study (14). Active disease was defined as new onset or worsening of at least 2 of the following disease features: 1) presence of systemic signs or symptoms (e.g., fever, arthralgia) that were not attributable to another condition, 2) elevation of acute-phase reactants (i.e., erythrocyte sedimentation rate [ESR], C-reactive protein [CRP]) in the absence of infection or malignancy, 3) onset of signs or symptoms of vascular insufficiency (unequal or absent pulses or blood pressure, limb claudication), or 4) new vascular lesion(s) on imaging studies, i.e., new stenosis or new dilatation (not previously diagnosed). Isolated arterial wall thickening was not considered significant in this study. Disease remission was defined as resolution of clinical and laboratory features of active disease and the absence of new vascular lesions on sequential imaging studies. Sustained remission was defined in those who met these conditions for at least 6 months while receiving a treatment regimen including prednisone <10 mg/day. Relapse was assessed by the treating physician. Features of relapse included: 1) onset of systemic signs or symptoms (e.g., fever, arthralgia), 2) elevation of acute-phase reactants, 3) onset of signs or symptoms of vascular insufficiency, or 4) new vascular lesion(s) on imaging studies. One or more of these items was sufficient to define the relapse after a period of remission (defined as resolution of clinical and laboratory features of active disease and the absence of new vascular lesions on sequential imaging studies). Of course, relapse was considered if signs and symptoms were not attributable to another condition.

The data are reported using descriptive statistics, including the percentage, and means with the SD for the normally distributed variables. The majority of continuous variables were summarized using medians with first and third quartiles, which we will refer to as the interquartile range (IQR) due to the limited sample size in our study. Statistical analysis was performed using SAS software, version 9.1.


Twenty-three patients received TNF inhibitors for TA while followed at our institution during the study period. We excluded 2 patients with insufficient followup data (less than 3 months) and 1 patient who received only 2 infusions of infliximab (given concurrently with cyclophosphamide for life-threatening bilateral carotid artery stenosis). The final study cohort included 20 patients with TA. The median total followup for these patients was 54 months (IQR 34–82 months).

The patients were predominantly women (19 patients [95%]) and white (17 patients [85%]). Other ethnicities were Asian, African American, and Native American (1 patient each). The mean ± SD age at diagnosis of TA was 29.8 ± 9.6 years. The median disease duration at the time of TNF inhibitor initiation was 15.9 months (IQR 2–32.7 months). The indication for TNF inhibitors was active and/or relapsing disease despite prior treatment in 19 patients (95%) and change in therapy due to an adverse reaction to mycophenolate mofetil (1 patient [5%]). Among the 19 patients with active and/or relapsing disease, 6 patients (32%) had developed new arterial lesions while receiving conventional immunosuppressive agents, prompting the switch to TNF inhibitors.

Before starting TNF inhibitors, all patients had received CS at a median initial dosage of 60 mg of prednisone/day (IQR 60–60, range 40–100). Methotrexate was used in 18 patients (90%) at a median dosage of 20 mg/week (IQR 20–23.75). Five patients (25%) received azathioprine (median dosage 100 mg/day [IQR 100–250]), 3 patients (15%) had received mycophenolate mofetil (2,000 mg/day), and 3 patients (15%) were treated with cyclophosphamide (100 mg/day). The median number of immunosuppressive treatments (other than prednisone) before starting the TNF inhibitors was 1 (IQR 1–2, range 0–3).

Prior to treatment with a TNF inhibitor, the median ESR was 33 mm/hour (IQR 13–65) and the median CRP level was 10 mg/liter (IQR 3–33). CRP level and/or ESR were elevated in 62% of the patients (i.e., CRP level >8 mg/liter and ESR >30 mm/hour). The median dosage of prednisone at the time TNF inhibitors were started was 12.25 mg/day (IQR 3–25), and 11 of 20 patients were receiving more than 10 mg of prednisone/day.

Infliximab was prescribed in 17 patients (85%), in combination with methotrexate (14 patients) and with azathioprine (2 patients), and as monotherapy (1 patient). Adalimumab and etanercept were used as monotherapy in 2 patients and 1 patient, respectively. The dosage of adalimumab was 40 mg/2 weeks subcutaneously and the dosage of etanercept was 50 mg/week subcutaneously. The initial dose of infliximab was 3 mg/kg intravenously every 8 weeks. However, this dose was increased in 16 patients and the final infliximab doses ranged between 3 and 7 mg/kg administered every 4 to 6 weeks.

After initiation of TNF inhibitors, 18 patients (90%) went into remission. One patient (5%) with persistently active disease was lost to followup after 4 months of treatment. Another patient discontinued TNF inhibitors due to intolerance without achieving remission. The median time to remission was 3.5 months (IQR 1.5–6.0 months). Sustained remission was achieved in 10 patients (50%) after a median of 17.2 months (IQR 12.0–26.5 months) from initiation of therapy. Relapse occurred in 6 of 18 patients while receiving TNF inhibitors after a median of 3.6 months (IQR 2.9–16.8 months) after remission. None of the patients developed a new arterial lesion while receiving TNF inhibitors.

Of the 14 patients who were receiving prednisone at initiation of therapy, 7 (50%) were able to discontinue CS. Additionally, 10 of the 12 subjects receiving prednisone doses ≥10 mg daily were able to taper prednisone below 10 mg daily.

The median duration of treatment with TNF inhibitors was 23.0 months (IQR 8.7–38.9 months), with 10 patients still receiving TNF inhibitors at last followup. The persistence with the first course of TNF inhibitor therapy was 73.8% at 1 year and 55.4% at 2 years. During followup, 11 patients (55%) discontinued treatment with TNF inhibitors. Three patients discontinued therapy because of relapse, persistently active disease, and lack of CS-sparing effect (1 patient each). Four patients discontinued TNF inhibitors (infliximab) because of adverse effects, including postoperative infection after hand surgery, serum sickness–like reaction, intolerance to the treatment, and pancreatic adenocarcinoma. All 4 of these patients were concurrently treated with methotrexate, and 3 were receiving prednisone. The patient who developed a pancreatic adenocarcinoma had also previously received azathioprine. The 4 remaining patients stopped their treatment for other reasons: loss of insurance coverage, preoperatively (was not resumed after intervention), patient preference, and unknown reason. TNF inhibitors were discontinued in 2 patients after a period of remission. Unfortunately, both of these patients relapsed 5 months and 11 months after discontinuation of TNF inhibitors. In both cases, new arterial stenoses were noted. Remission was later achieved in both patients after resuming TNF inhibitors. A summary schematic representation of the outcomes of the 20 patients receiving TNF inhibitors is shown in Figure 1.

Figure 1.

Disease activity in patients with Takayasu arteritis treated with tumor necrosis factor (TNF) inhibitors. Red bar = active disease; green bar = remission; blue bar = sustained remission; vertical black line = discontinuation of TNF inhibitors; * = temporarily discontinued or changed TNF inhibitor treatment during followup.

We recorded 6 infections in 5 patients while receiving TNF inhibitors, including 3 pneumonias, 1 herpes zoster reactivation, 1 pyelonephritis, and 1 postoperative infection. Only the patient with infection after hand surgery required hospitalization and TNF inhibitor withdrawal. In the remaining patients, the TNF inhibitor was only held temporarily. One patient receiving infliximab experienced infusion-related reactions, following which treatment was switched to etanercept, resulting in sustained remission.


In our single-center retrospective study, TNF inhibitors were effective in inducing remission in 90% of patients with refractory TA. Sustained remission with concomitant prednisone use of less than 10 mg/day was achieved in 50% of the cases. However, relapses occurred in 33% of patients. Persistence with the first course of TNF inhibitor therapy was only 55% at 2 years. Twenty percent of the patients discontinued TNF inhibitors because of adverse events.

Our study included a relatively large cohort of patients with a very rare disease, with a maximum followup time of 81 months. Our experience with TNF inhibitors is consistent with 2 other studies conducted in comparable clinical settings (refractory TA) (11, 12). The rates of remission and sustained remission are very similar between the 3 studies (Table 1). The relapse rate in our study is lower than previously reported. This may in part be related to the differences in frequency of imaging used in the studies. The patients reported by Molloy et al also may have had more severe disease, since 13 of 25 patients had not achieved remission at any time since their initial diagnosis (12). Additionally, the median disease duration before TNF inhibitors was considerably longer in the studies by Molloy et al and Hoffman et al (116 and 72 months, respectively) compared to only 16 months in our study. The observations raise the question of whether TNF inhibitor therapy might be considered earlier in the course of the disease.

Table 1. Effect of TNF inhibitors in difficult-to-treat Takayasu arteritis: comparison of our study with 2 previous studies*
Author, year (ref.)No. of patientsType of TNF inhibitor: no.RemissionSustained remissionRelapseNew arterial lesion during TNF inhibitor useCS cessation
  • *

    Values are the number/total (percentage) unless otherwise indicated. TNF = tumor necrosis factor; CS = corticosteroids; inflix = infliximab; etan = etanercept; ada = adalimumab.

Hoffman et al, 2004 (11)15Inflix: 814/15 (93)10/15 (66) 4/14 (28)10/15 (66)
  Etan: 7     
Molloy et al, 2008 (12)25Inflix: 1622/25 (88)15/25 (60)15/24 (62)4/24 (16)15/25 (60)
  Etan: 9     
Present study20Inflix: 1718/20 (90)10/20 (50)6/18 (33)07/12 (58)
  Etan: 1     
  Ada: 2     

Adverse effects, including infections, were significant while receiving TNF inhibitors, affecting drug persistence with TNF inhibitors (73.8% still receiving TNF inhibitors at 1 year and 55.4% at 2 years for the first course of TNF inhibitor therapy). In comparison, persistence with TNF inhibitors for the treatment of rheumatoid arthritis is 76% at 1 year and 65% at 2 years with infliximab (15). Previous exposure to other immunosuppressants and higher doses of CS may contribute to a lower persistence with TNF inhibitors in TA.

The observation of relapse after discontinuation of TNF inhibitors, and remission when reintroduced, suggests that in many patients, a prolonged course of TNF inhibitors might be considered to improve long-term disease control. Relapse of the disease after TNF inhibitor discontinuation was described in a previous study by Molloy et al, where all 3 patients who discontinued etanercept and 8 of 9 patients who discontinued infliximab experienced relapses (12). Two of these 9 patients resumed infliximab and achieved remission again (12).

Most often, TNF inhibitors were used after other immunosuppressants have failed (methotrexate, mycophenolate mofetil, azathioprine, and even cyclophosphamide). Because of the widespread use of TNF inhibitors for other diseases (e.g., rheumatoid arthritis, Crohn's disease, spondylarthropathy, etc.), the safety profile of TNF inhibitor agents is well understood (16). The cost of TNF inhibitors may be of concern, but this has to be considered relative to the costs associated with ongoing medical and surgical costs from a chronic vascular disease (2). Therefore, it may be worthwhile to consider using TNF inhibitors earlier in the course of TA to reduce the probability of progressive vascular damage.

This study reports our experience with TNF inhibitors in the treatment of TA. In our center, TNF inhibitors were used as a third line of treatment for cases that were refractory after treatment with CS and other immunosuppressants. The study is not a controlled clinical trial and has the limitations of a retrospective review. The level of evidence for using TNF inhibitors in TA provided by such a study should be considered with care.

The growing reported experience regarding the use of TNF inhibitors in TA stresses the need for clinical trials in this disease, which will need to be carried out as collaborative multicenter international efforts.


All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Schmidt had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Schmidt, Bacani, Matteson, Warrington.

Acquisition of data. Schmidt, Bacani, Matteson, Warrington.

Analysis and interpretation of data. Schmidt, Kermani, Crowson, Matteson, Warrington.