Anti–tumor necrosis factor therapy in patients with difficult to treat Takayasu arteritis

Abstract

Objective

Granulomatous inflammation is a typical feature of Takayasu arteritis (TA), and tumor necrosis factor (TNF) is important in the formation of granulomas. In this study, we assessed therapy with anti-TNF agents in patients with TA that was not controlled by glucocorticoid therapy or other immunosuppressants.

Methods

We conducted an open-label trial of anti-TNF therapy at 3 academic medical centers over a period of 4.25 years. Fifteen patients with active, relapsing TA (median 6 years) were selected. Seven received etanercept (later changed to infliximab in 3 patients), and 8 received infliximab. Relapses had occurred in all patients while they were receiving glucocorticoids and, in 13 patients, additional immunosuppressive drugs. No other agents were added to the treatment regimen concurrently with anti-TNF. If patients were receiving cytotoxic agents, the dosage was not increased. Clinical symptoms were recorded, and physical examinations, laboratory studies, and serial magnetic resonance imaging were performed.

Results

The median daily dose of prednisone required to maintain remission prior to anti-TNF therapy was 20 mg. Ten of the 15 patients achieved complete remission that was sustained for 1–3.3 years without glucocorticoid therapy. Four patients achieved partial remission, with a >50% reduction in the glucocorticoid requirement. At a median of 12 months of followup, the median dose of prednisone was 0. Therapy failed in 1 patient. In 9 of the 14 responders, an increase in the anti-TNF dosage was required to sustain remission. Two relapses occurred during periods when anti-TNF therapy (etanercept) was interrupted, but remission was reestablished upon reinstitution of therapy.

Conclusion

In this pilot study of relapsing TA, addition of anti-TNF therapy resulted in improvement in 14 of 15 patients and sustained remission in 10 of 15 patients, who were able to discontinue glucocorticoid therapy. Anti-TNF may be a useful adjunct to glucocorticoids in the treatment of TA. Our results justify a randomized, controlled clinical trial of anti-TNF therapy for TA.

Takayasu arteritis (TA) is an idiopathic systemic granulomatous disease of the large and medium-sized vessels that may lead to segmental stenosis, occlusion, dilatation, and/or aneurysm formation in the aorta and/or its main branches. Coronary and/or pulmonary arteries may also be affected (1). Granulomatous inflammation is dependent in part on tumor necrosis factor (TNF). TNF production occurs primarily in macrophages, T cells, and natural killer cells. TNF induces macrophage production of interleukin-12 (IL-12) and IL-18, which are potent cytokines that enhance CD4 T lymphocyte differentiation to Th1 cells and activate natural killer cells. IL-18–influenced Th1 lymphocyte production of interferon-γ leads to recruitment and activation of macrophages, a critical feature of granuloma formation. The pathogenesis of TA includes vessel injury due to products from activated T cells, natural killer cells, γ/δ cells, and macrophages (2). It is therefore reasonable to consider that TNF inhibition might be useful in abrogating the immunoinflammatory features of TA.

At the time patients were enrolled in this study, 2 TNF antagonists, etanercept and infliximab, were commercially available. Etanercept is a fusion protein of two p75 subunits of the TNF receptor linked to the Fc portion of human IgG1. Infliximab is a murine–human chimeric monoclonal IgG1 antibody that binds to human TNF, causing its inactivation.

The efficacy of TNF inhibitors in rheumatoid arthritis, juvenile rheumatoid arthritis, psoriasis, psoriatic arthritis, Crohn's disease, and ankylosing spondylitis has been established. Cantini et al (3) have recently reported encouraging results with the use of infliximab in 3 of 4 patients with longstanding active giant cell arteritis, a form of granulomatous inflammation affecting large and medium-sized vessels. In another report of 2 patients in whom infliximab was provided as the initial and sole therapy for giant cell arteritis, an initial response was noted, but relapse occurred during followup (4). In each case, remission was achieved with corticosteroid therapy.

The granulomatous nature of the histopathologic lesions in TA led us to consider an evaluation of the therapeutic benefits of anti-TNF therapy in patients in whom disease was particularly resistant to conventional therapies.

PATIENTS AND METHODS

Patient selection.

Patients were included in the study if they had 1) previous clinical and imaging (invasive angiography and magnetic resonance imaging [MRI]) data that supported the American College of Rheumatology classification criteria for TA (5), 2) required toxic doses of glucocorticoids to maintain remission, and 3) either experienced multiple relapses while receiving conventional and experimental therapy (13 of the 15 patients) or refused re-treatment with glucocorticoids following relapses (2 of the 15 patients). All patients were skin tested for tuberculosis by placement of an intermediate-strength purified protein derivative of tuberculin. Results of all the skin tests were negative, and none of the patients had chest roentgenogram evidence suggestive of tuberculosis.

Exclusions.

Patients were excluded from study if they had any of the following: active systemic infections, neutropenia (<4,000 white blood cells/mm³), thrombocytopenia (<120,000 platelets/mm³), acute or chronic liver dysfunction, malignancy, excessive alcohol use, pregnancy (and/or breastfeeding), or a history of noncompliance with medical care.

Previous therapy.

In all but 1 patient, TA had previously responded to high-dose glucocorticoid therapy, but relapses occurred upon dosage reduction. In 3 patients, complete remission was never achieved. In 2 patients, unacceptable glucocorticoid toxicity precluded further use of this therapy. Glucocorticoid therapy had induced psychosis in one of them; the other patient refused re-treatment with glucocorticoids because of cosmetic and less-severe behavioral side effects. In 13 patients, 1 or more of the following agents was, at some point, added to the glucocorticoid regimen: methotrexate (in 13), cyclophosphamide (in 6), mycophenolate mofetil (in 3), azathioprine (in 3), cyclosporin A (in 2), or tacrolimus (in 2) (Table 1). Eight patients received at least 2 agents in addition to glucocorticoids. None of these 6 immunosuppressive agents was successful in providing sustained remission and allowing the dosage of glucocorticoids to be tapered and discontinued.

Table 1. Demographic and clinical characteristics of the patients with TA treated with anti-TNF therapy

Patient	Anti-TNF treatment*		Age/sex/race, year of disease onset†	Vascular lesions before anti-TNF therapy‡	Before anti-TNF treatment			After anti-TNF treatment		Followup, no. of months of anti-TNF therapy#	New lesions on followup imaging studies**	Comments††
	TNF agent, initial stable dosage	Final stable dosage (if changed)			Pred., lowest dosage without relapse§	Other¶		Pred., lowest dosage§	Other
						Remote past	At initiation of TNF
* Ten of the 15 patients (9 of the 14 responders) required a dosage change or a change in the anti–tumor necrosis factor (anti-TNF) agent. Inflix. = infliximab; etan. = etanercept. † The mean age of the patients was 27.5 years (median 26.0 years). Fourteen were female; 1 was male. One patient was Indian (I); 14 (93%) were white (W). Their mean duration of Takayasu arteritis (TA) was 6.5 years (median 6.0 years; range 2–16 years). ‡ AR = aortic regurgitation; IC = internal carotid; EC = external carotid; SC = subclavian; CC = common carotid; CABG = coronary artery bypass graft (number of arteries); SM = superior mesenteric; IM = inferior mesenteric. § Before anti-TNF therapy, the mean prednisone (pred.) dosage was 21 mg/day (median 20 mg/day; range 0–40 mg/day). After anti-TNF therapy, the mean prednisone dosage was 1.7 mg/day (median 0; range 0–20 mg/day). After anti-TNF therapy, 10 of the 15 patients did not require prednisone. ¶ In no case was a new agent added at the same time anti-TNF therapy was started. MTX = methotrexate; MMF = mycophenolate mofetil; tacro. = tacrolimus; CSA = cyclosporin A; CYC = cyclophosphamide; AZA = azathioprine. # The median followup period was 12 months (mean 21.7 months). ** All patients had at least 1 invasive angiogram. Followup studies were by magnetic resonance imaging unless a surgeon required a preoperative invasive angiogram prior to vascular bypass surgery. Five patients had progressive lesions after anti-TNF therapy (1 was noncompliant, and 4 improved, apart from the development of new lesions). †† Four patients were in remission and required <10 mg of prednisone per day (representing a >50% dosage reduction). Ten of the 15 patients did not require prednisone.
1	Inflix. 3 mg/kg every 8 weeks	7 mg/kg every 8 weeks	27/F/W, 1998	Aortic root aneurysm; AR; stenoses in the right IC and EC; bilateral SC and celiac aneurysms	20 mg/day	MTX, MMF	Tacro., 6 mg/day	0	0	40	None	Crohn's disease also in remission (frequency of Crohn's disease is increased in TA [7, 8])
2	Etan. 25 mg twice a week	50 mg twice a week	42/M/W, 1987	Pericarditis; stenoses in the left SC, left vertebral, abdominal aorta, and bilateral iliac arteries	12.5 mg/day	MTX, CSA, tacro., CYC	MMF 1,500 mg twice a day	0	0	31	None	Episodes of pericarditis treated with short courses of prednisone and increase in etan. to 50 mg twice a week
3	Etan. 25 mg twice a week	–	19/F/I, 1999	Aortic root dilatation; AR; stenoses in the left SC and left vertebral arteries; occluded left CC artery	20 mg/day (refused pred. after relapse; treated only with etan.)	None	None	0	0	35	None	Remission with anti-TNF monotherapy
4	Etan. 25 mg twice a week	50 mg twice a week	27/F/W, 1997	Aortic root dilatation; stenoses in the left CC and bilateral SCs	40 mg every other day (refused daily pred.)	MMF	MTX 25 mg/week	0	0	51	Stenosis in bilateral axillary arteries at 23 months of followup	Etan. increased to 50 mg twice a week at 23 months; stable over ensuing 28 months
5	Etan. 25 mg twice a week	50 mg twice a week. Then inflix. 5 mg/kg every 8 weeks	19/F/W, 1997	Left main coronary artery (CABG ×2); stenoses in bilateral CCs and pulmonary arteries; left SC occlusion	25 mg/day	MTX, CYC	AZA 150 mg/day	0	AZA 150 mg/day	48	Coronary restenosis during period of noncompliance	Etan. increased to 50 mg twice a week and switched to inflix. 5 mg/kg every 8 weeks when etan. not available; stable remission for 34 months
6	Inflix. 5 mg/kg every 8 weeks	5 mg/kg every 6 weeks	18/F/W, 2001	Aortic arch wall thickening; stenoses in the left CC and celiac arteries; SM artery occlusion	20 mg/day	–	MTX 25 mg/week	2.5 mg/day	MTX 25 mg/week	15	None	Mild relapse at 2.5 mg/day of pred.; increased to 10 mg/day. Increase in inflix. to 7 mg/kg every 8 weeks
7	Inflix. 5 mg/kg every 8 weeks	–	30/F/W, 2001	Stenoses in bilateral CCs and innominate; occlusion left SC	15 mg/day	CYC	MTX 20 mg/week	2.5 mg/day	MTX 20 mg/week	3	None	Aorta to bilateral CC artery bypass. Sepsis (histoplasmosis) after 3 months of treatment. Inflix. stopped and amphotericin initiated. Patient recovered
8	Etan. 25 mg twice a week	50 mg twice a week. Then inflix. 5 mg/kg every 8 weeks	28/F/W, 1999	Aortic arch wall thickening; stenoses in the left SC, axillary, celiac, and pulmonary arteries	20 mg/day	–	MTX 25 mg/week	0	MTX 25 mg/week	17	None	Relapse with etan. therapy at 4 months. Remission with inflix. 5 mg/kg every 8 weeks
9	Inflix. 3 mg/kg every 8 weeks	7 mg/kg every 8 weeks	22/F/W, 1994	Dilated thoracic aorta; stenoses in the left CC, right renal, and right SC arteries; occlusion in the left SC and SM arteries	15 mg/day	–	MTX 20 mg/week	5 mg/day	MTX 20 mg/week	8	Stenoses in the right CC and iliac arteries	Left CC to left SC artery bypass before inflix. Inflix. increased to 7 mg/kg every 8 weeks
10	Inflix. 3 mg/kg every 8 weeks	–	41/F/W, 1994	Occlusions in bilateral SCs and left CC; stenoses in bilateral iliacs, celiac, and SM arteries	20 mg/day	–	MTX 15 mg/week	0	MTX 7.5 mg/week	11	None	–
11	Inflix. 3 mg/kg every 8 weeks	10 mg/kg every 4 weeks	17/F/W, 2000	Aortic root dilatation; AR; stenoses in bilateral SCs, IM artery, celiac, bilateral renal arteries; bilateral SC aneurysms	30 mg/day	CYC	MTX 20 mg/week	10 mg/day	MTX 20 mg/week (MMF 1 gm twice a day)	8	Stenosis in the infrarenal abdominal aorta	Malignant hypertension and pulmonary edema at presentation. Relapse with 3 mg/kg of inflix.; increased to 5, 7, and 20 mg/kg every 4 weeks
12	Etan. 25 mg twice a week	Inflix. infusion reaction. Then etan. 25 mg twice a week	24/F/W, 1999	Thoracic aorta dilatation; left axillary artery stenosis	0 (never achieved remission)	–	MTX 15 mg/week	0	0	12	None	MTX intolerance. Steroid psychosis (thus not started for flare). Inflix. infusion reaction led to discontinuation, which led to relapse; resolved with etan. monotherapy
13	Etan. 25 mg twice a week	–	25/F/W, 1990	Stenosis in the abdominal aorta plus dissection; stenoses in bilateral SCs, bilateral CCs, left renal, and SM arteries	30 mg/day (never achieved remission)	CYC, MTX, AZA	–	0	0	11	None	Arthralgias during 2 periods of etan. unavailability; each promptly resolved with resumption of etan. Herpes zoster resolved with treatment
14	Inflix. 5 mg/kg every 8 weeks	–	48/F/W, 1996	Stenoses in the innominate, left SC, left CC, and left vertebral arteries	40 mg/day (never achieved remission)	Refused cytotoxic agents	–	0	0	11	None	–
15	Inflix. 5 mg/kg every 8 weeks	10 mg/kg every 4 weeks	26/F/W, 1993	Coronary arteries (CABG ×3); stenoses in bilateral SCs	30 mg/day	CSA, MTX, CYC, AZA	–	20 mg/day	CYC	9	New stenosis in the right SC artery	Treatment failure, all modalities. Aggressive disease. Inflix. increased to 10 mg/kg for 4 weeks (×3 doses)

Anti-TNF treatment schedules.

This was an open-label study. Seven patients received etanercept and 8 received infliximab. In 3 patients, etanercept was later changed to infliximab because of limitations in drug availability (2 patients) or relapse while receiving etanercept (1 patient). In 1 patient, return to etanercept treatment later occurred because of an infliximab infusion reaction. Etanercept was initially administered subcutaneously at a dosage of 25 mg twice a week. Among the 11 infliximab-treated patients, the starting doses varied from 3 mg/kg to 5 mg/kg intravenously, and the same sequential regimen as used in rheumatoid arthritis trials was followed (initial dose, second dose 2 weeks later, third dose 6 weeks later, and doses every 4–8 weeks thereafter).

All patients who received high-dose glucocorticoids plus a cytotoxic agent received 1 double-strength tablet of trimethoprim/sulfamethoxazole 3 times a week for prophylaxis against Pneumocystiscarinii pneumonia.

Study end points.

The success of therapy was evaluated according to the resolution of symptoms of active disease, the ability to taper and/or discontinue prednisone and other immunosuppressive agents without relapse, and the absence of new vascular lesions. During the initial phases of disease, all patients had been studied by invasive, catheter-directed angiography of the entire aorta and its primary branch vessels. Active or relapsing disease was considered to be present in the setting of new vascular lesions, as visualized by MRI or catheter-directed angiography, at sites that were previously unaffected. New angiographic lesions were accepted as definite evidence of relapse. The utility and safety of MRI for following large-vessel anatomy is well established. However, the role of MRI, with edema-weighted image generation, as a surrogate of vascular inflammation is less reliable and is not endorsed as a guide to treatment (6). Active disease was also assumed if at least 2 of the following 5 features were present: 1) new onset of carotodynia or pain over other large vessels, 2) transient ischemic episodes that could not be attributed to other factors, 3) new bruit or new asymmetry in pulses or blood pressure, 4) fevers in the absence of infection, or 5) a reproducible increase in the Westergren erythrocyte sedimentation rate (ESR) (1).

Complete and sustained remission was defined as the absence of features of active disease, the absence of new lesions on sequential imaging studies, and no glucocorticoid therapy for at least 6 months. Improvement or partial remission was considered to have occurred if the glucocorticoid dosage could be reduced by at least 50%, compared with the prestudy dosage, before the disease recurred.

Dosages of medications were modified according to the patient's clinical status at each visit (Table 1).

RESULTS

The demographic features of patients with TA in this study were similar to those of a previously reported series of patients in the US. Ninety-three percent of the patients were women. All but 1 patient was white. One female patient was Indian. The mean age of the patients was 27.5 years (median 26.0 years, range 17–48 years), and the mean and median duration of illness prior to anti-TNF therapy was 6.5 years and 6 years, respectively (range 2–16 years). Invasive angiography was performed in all patients initially for purposes of confirming the initial diagnosis and in some patients at a later time for planning surgical procedures. MRI was used for periodic followup, including in the absence of new symptoms or findings, to determine the presence and distribution of lesions. Particular emphasis was placed on identifying new lesions in previously unaffected regions, which was interpreted as a sign of intervening disease activity, progression, or relapse. Each study included the entire aorta and all primary branch vessels.

Ten of the 15 patients treated with anti-TNF therapies (67%) achieved sustained remission of disease (no new vascular lesions and ability to discontinue glucocorticoid therapy) that lasted 1–3.3 years. Four patients (27%) achieved partial remissions and were able to reduce their glucocorticoid requirements by at least 50%. At 12 months of followup (median and mean 21.7 months, range 3–51 months), the median daily dose of prednisone was 0 (range 0–20 mg/day), compared with the effective median dose of 20 mg/day (range 12.5–40 mg/day) before anti-TNF therapy. In 1 patient (patient 15), anti-TNF therapy failed, new arterial lesions occurred, and the glucocorticoid dosage was never significantly tapered. This patient had previously failed to achieve remission with numerous other adjunctive agents.

Repeated imaging studies revealed that 5 patients (2 complete responders, 2 partial responders, and 1 treatment failure) had new lesions over a median followup period of 12 months (range 8–51 months) of anti-TNF therapy. Overall, 9 of 14 responders to anti-TNF therapy (64%) required an increase in the dosage of the biologic agent to achieve a sustained remission. In 1 patient who became a complete responder, new axillary artery stenoses occurred 23 months after initiation of etanercept therapy (25 mg twice a week). The dosage was increased to 50 mg twice a week, and new lesions did not occur in the ensuing 28 months.

Two relapses occurred during periods of anti-TNF therapy interruptions. In one of the patients, a new coronary lesion occurred during a 3-month period of noncompliance with etanercept therapy. Restarting therapy was followed by sustained remission over the ensuing 34 months. In the second patient, relapse occurred during a period of anti-TNF therapy interruption because etanercept was unavailable. Disease remission was promptly reestablished when anti-TNF therapy was reinstituted (the only variable that changed).

Adverse events.

Among the 15 patients in this cohort, anti-TNF therapy was definitely associated with 1 adverse event and was possibly associated with 2 other adverse events. One patient (patient 12) had an infusion reaction to infliximab and was later treated with etanercept. Three months after starting infliximab, 1 patient (patient 7) developed disseminated histoplasmosis after travel to an endemic area, an event considered to be possibly related to the experimental therapy. Infliximab was discontinued, and complete recovery occurred following parenteral treatment with amphotericin and, later, orally administered itraconazole. Patient 13 developed herpes zoster limited to 1 dermatome after receiving etanercept therapy for 8 months, another possible anti-TNF–related toxicity. Etanercept was temporarily discontinued in this patient, and treatment with valacyclovir led to full clinical resolution.

Case summaries.

For more in-depth information, we provide the case summaries of 2 of the study patients, patients 1 and 5 in Table 1.

Patient 1.

This patient, a 27-year-old woman, experienced fever, fatigue, new-onset headache, syncope, seizures, bilateral arm claudication, and transient left hemiparesis (Figure 1). Her ESR was 105 mm/hour. Angiography revealed fusiform dilatation of the aortic root and aortic valve annulus, aneurysmal dilatation of the left internal carotid artery, and alternating areas of dilatation and stenosis of the subclavian arteries bilaterally. A celiac artery aneurysm was also present (Figure 2). There were also stenoses of the right internal and external carotid arteries. A diagnosis of TA was made. Because of the concurrent recent onset of frequent diarrhea, endoscopy and biopsy were performed, and the findings were compatible with Crohn's disease. Treatment with prednisone, 30 mg/day, and mesalamine led to improvement in the fatigue, fevers, and headache. However, repeated attempts to taper the prednisone dosage led to recurrence of fever, fatigue, headaches, light-headedness, arm claudication, and watery/bloody diarrhea. Improvement occurred when higher doses of prednisone were restarted. A 100-pound weight gain occurred over a period of 1 year.

Figure 1. Graphic representation of the course of illness and treatment in patient 1, a woman with Takayasu arteritis and Crohn's disease who first presented at age 27 years (in 1998) with transient left hemiparesis, syncope, headache, and seizure. ESR = erythrocyte sedimentation rate (Westergren); CRP = C-reactive protein; bid = twice daily; qd = every day; anti-TNF = anti–tumor necrosis factor.

Figure 2. Magnetic resonance angiography of patient 1 obtained before initiation of anti–tumor necrosis factor therapy, demonstrating dilatation of the aortic root, multiple alternating areas of dilatation and stenosis of the subclavian arteries bilaterally, and a celiac artery aneurysm. Stenoses of the right internal and external carotid arteries were also present (not shown). H = head.

After referral to one of the authors, methotrexate (20 mg/week orally), mycophenolate mofetil (1.5 gm twice a day), and tacrolimus (6 mg/day) were added sequentially to the prednisone therapy. However, each immunosuppressive medication in turn was incapable of sustaining remission. Multiple attempts to taper the prednisone dosage below 20 or 30 mg/day failed.

Following 2 years of waxing and waning symptoms, infliximab therapy was begun for both the active TA and the active Crohn's disease. She received an initial intravenous dose of 3 mg/kg, which was repeated 2 weeks and 6 weeks later, and then every 8 weeks thereafter. After 2 doses of infliximab, the diarrhea decreased, and she was able to perform modest exercises. Three months after starting infliximab, she reported that she felt the best she had felt in many years and was able to perform low-impact aerobics for 30 minutes once or twice daily. MR vascular imaging failed to reveal any new lesions. The prednisone dosage was successfully tapered and discontinued within 4 months of starting infliximab, and tacrolimus was discontinued 4 months later.

After 6 months and 10 months of infliximab therapy, mild relapses (Figure 1) led to increases in the intravenous dosage of infliximab to 5 mg/kg and 7 mg/kg every 8 weeks, respectively. Prednisone was not instituted at these times. Thirty-eight months since initiation of infliximab therapy and 2 dosage increases, she remains clinically asymptomatic for both the TA and the Crohn's disease. She has not required glucocorticoid therapy for 44 months. Sequential MR vascular imaging studies have not revealed new lesions.

Patient 5.

This patient, a 19-year-old woman, experienced recurrent light-headedness for several years. In 1997, TA was diagnosed based on the presence of bilateral carotid bruits, an aortic murmur, and asymmetric blood pressures, as well as angiographic findings (Figure 3). Blood pressure measurements were as follows: undetectable in the left arm, 140/64 mm Hg in the right arm, 120/60 mm Hg in the right leg, and 135/70 mm Hg in the left leg. Conventional angiography revealed mild aortic valve insufficiency, bilateral common carotid stenoses, left subclavian artery occlusion, and branch artery occlusions to the right upper and lower lobes of the lung.

Figure 3. Graphic representation of the course of illness and treatment in patient 5, a woman with Takayasu arteritis who first presented at age 19 years (in 1997) with symptoms of cerebral ischemia. ESR = erythrocyte sedimentation rate (Westergren); CRP = C-reactive protein; MR = magnetic resonance imaging; anti-TNF = anti–tumor necrosis factor; b/w = biweekly.

Prednisone, 60 mg/day, was begun, and her light-headedness diminished. She was referred to one of the authors in May 1997. Methotrexate was added, and the dosage increased gradually to 25 mg/week orally. However, 4 relapses followed, each requiring increased glucocorticoid therapy. Apparent remission, without the need for glucocorticoids, was briefly achieved in May 1998. However, within 4 months, she experienced angina and increases in levels of acute-phase reactants. Angiography revealed >50% stenosis of the left main, left circumflex, and left anterior descending coronary arteries. Prednisone was resumed at a dosage of 60 mg/day, and cyclophosphamide 125 mg/day was initiated. Within 3 weeks, coronary revascularization was performed. Cyclophosphamide-induced cystitis led to cessation of that agent and initiation of azathioprine, 150 mg/day. By mid-July 1999, the prednisone dosage had been gradually tapered and stopped. Two weeks later, another flare occurred, which was characterized by elevations in levels of acute-phase reactants and MRI findings of increased aortic wall thickness and edema.

Because of significant cumulative glucocorticoid toxicity, her reluctance to restart prednisone, and failure to sustain remission or to tolerate methotrexate, azathioprine, or cyclophosphamide, etanercept was begun at a dosage of 25 mg subcutaneously twice a week. Four months after starting etanercept, the aortic wall thickening had diminished and edema had resolved, as demonstrated by MRI. The patient was asymptomatic, was exercising 2–3 times a week, and her ESR was 19 mm/hour. In November 2000, she experienced left-sided chest pain, and she reported that she had not taken the etanercept over the preceding 3 months. Angiography revealed 90% stenosis of the left main coronary artery and occlusion of the right internal mammary artery and the bypass grafts. Prednisone was resumed at a dosage of 60 mg/day, and a second myocardial revascularization procedure was performed.

In an attempt to achieve better disease control, the etanercept dosage was increased to 50 mg twice a week. As of November 2001, she had been well and had successfully discontinued the prednisone. Unavailability of etanercept in December 2001 required a switch to infliximab (5 mg/kg intravenously every 8 weeks), which has been continued to the present. She also continues to receive azathioprine (150 mg/day). Glucocorticoid therapy has not been required over the last 21 months.

DISCUSSION

This is the first long-term study of anti-TNF therapy in Takayasu arteritis. The results suggest that this new form of therapy may be efficacious in this disorder. This study cohort is unique in that it focuses on patients who were particularly resistant to the usual therapy for TA. Prior to anti-TNF therapy, none of the 15 patients in this series had experienced sustained remission. For each of the 14 patients who responded to anti-TNF therapy, subjective and objective signs of improvement were apparent within 2 weeks to 2 months. Glucocorticoid-free remissions were achieved in 10 of the 15 patients (67%). Conventional treatment of relapse in TA consists of reinstitution of high-dose glucocorticoid therapy, which 2 patients refused. Both of these patients subsequently achieved remission with only the addition of anti-TNF therapy. In 2 other patients who had flares of TA during lapses in anti-TNF therapy, prompt restoration of remission occurred when the anti-TNF agents were restarted. Overall, the addition of anti-TNF therapy resulted in improvement in 93% of the patients and sustained remission in 67%. Only 1 patient did not respond at all. Anti-TNF treatment also allowed for the reduction or elimination of glucocorticoid therapy in almost all of the patients.

Because this was an open-label trial, it must be considered a pilot study. The relatively small number of patients does not allow for conclusions to be made about the relative efficacy of the two TNF inhibitors used, nor does the design of this study allow definitive conclusions to be made about the role of anti-TNF therapy in TA. It is not known whether increases in anti-TNF therapy will be required with more-prolonged followup or whether anti-TNF therapy will be required to sustain remission in the future. It could be argued that some or all of the responding patients in this study may have achieved spontaneous remissions, independently of the experimental therapy. Evidence against this interpretation is 1) the rapid response (2 weeks to 2 months) to anti-TNF therapy in 14 of the 15 patients, 2) the relapses when etanercept treatment was interrupted, followed by remission after treatment was restarted, and 3) the remission achieved with anti-TNF therapy without glucocorticoids in the 2 patients with chronically active TA who refused re-treatment with glucocorticoids.

It is important to note that anti-TNF therapy does have known toxicities, and the full extent of the risk of these new medications in patients with TA was not measured in this study. Two opportunistic infections were seen among the 15 study patients: one case of histoplasmosis and one case of herpes zoster. Whether these infections may have occurred in the absence of anti-TNF therapy is uncertain. Given that many patients with TA who might theoretically benefit from anti-TNF therapy will be receiving concomitant glucocorticoid and/or other immunosuppressive therapy, caution must be exercised, and vigilance to detect infections must be maintained by the treating clinicians.

These preliminary results suggest that anti-TNF therapy may be a useful adjunct to glucocorticoids in the treatment of patients with TA. However, limitations in the study size and design preclude definitive conclusions about therapeutic efficacy. Our results justify the performance of an adequately powered, randomized controlled trial that would determine whether anti-TNF treatment shortens the duration of conventional glucocorticoid use, with an attendant reduction in both disease-induced and treatment-induced morbidity and mortality.

Acknowledgements

The authors are grateful to Drs. Eugene Y. Kissin and Millicent Tan-Ong for their assistance with the data collection and Debora Bork for her graphics expertise. We are particularly indebted to our referring physicians, Drs. John Tesser, Hal Mitnick, Nancy Nowlin, and J. Alan MacFarlane.