Methods

Top Methods Statistical Methods Results Discussion Author & Article Info References

Patients

Sixty patients with Takayasu arteritis were seen at the National Institutes of Health from 1970 to 1990. Patients were recruited as part of vasculitis protocols and without bias to economic, social, or racial status. Medical care was provided without cost to the patients. The diagnosis was based on the presence of symptoms and signs of ischemic, inflammatory large-vessel disease, as well as supportive arteriographic findings. All patients had periodic arteriographic studies and had abnormalities that included multifocal areas of stenosis or aneurysm formation (or both) of the aorta or its primary branches [or both]. Other causes of large-vessel abnormalities were excluded based on clinical criteria and, where appropriate, on serologic studies. These include 1) inflammatory aortitis [such as syphilis; tuberculosis; systemic lupus erythematosus; rheumatoid arthritis; spondyloarthropathies; Buerger, Behcet, Cogan, and Kawasaki diseases; and giant cell arteritis]; 2) developmental anomalies [such as the Ehlers-Danlos syndrome and the Marfan syndrome]; and 3) other aortic abnormalities (such as neurofibromatosis, ergotism, and radiation fibrosis).

Once enrolled, patients were prospectively followed and assigned to treatment protocols if their disease was active. Frequency of follow-up depended on disease activity and treatment and varied from monthly to yearly visits. Each visit included a complete history and physical examination, routine laboratory tests, blood pressure measurements in all four limbs (cuff or Dynamapp^Doppler), electrocardiogram, and chest radiograph. A complete aortogram was done if active, new disease was suspected or if required by specific treatment protocols every 4 to 6 months. Patients enrolled in the methotrexate protocol had angiographic studies at study entry and 6 months and 12 months after enrollment. Angiography was done by the intra-arterial injection of approximately 100 mL of iopamidol (61%; Isovue-300, Squibbs Diagnostic, New Jersey) using femoral puncture.

Criteria for Disease Activity

New onset or worsening of two or more features of Takayasu arteritis Table 1 defined "active disease". Typical angiographic findings of disease are shown in Figure 1. A clear-cut decrease in symptoms or improved clinical findings or both indicated partial remission or smoldering disease. Complete resolution of all clinical features or their stabilization, in the setting of fixed vascular lesions, was indicative of remission.

View larger version (103K): [in this window] [in a new window]   Figure 1. Angiogram showing critical stenoses. Long segments of critical stenoses (>70%) () of both common carotid arteries, the subclavian artery, and the proximal right vertebral arteries.

Treatment

Patients with active disease were initially treated with oral glucocorticoids at a dose of 1 mg/kg per day for 3 months. If patients could not then be tapered to an alternate-day regimen without disease exacerbation, a cytotoxic agent was added as previously described [18, 19]. Briefly, such patients received either cyclophosphamide, azathioprine, or methotrexate. Cyclophosphamide or azathioprine was administered as a daily oral dose of 1 to 2 mg/kg per day in conjunction with daily glucocorticoids. Methotrexate was given once weekly, the oral dose ranging from 0.15 to 0.35 mg/kg. The usual starting dose for methotrexate was about 15 mg/wk, with 2.5-mg increments every 1 to 2 weeks to achieve a maximum tolerated weekly dose of up to 25 mg.

If active disease had improved, daily glucocorticoid therapy was tapered to an alternate-day regimen. If disease remained inactive, glucocorticoid therapy was further tapered and stopped. Patients who also received a cytotoxic agent continued that agent alone for 1 year after remission. Then, the cytotoxic agent was tapered; cyclophosphamide and azathioprine were tapered by decrements of 25 mg/mo, and methotrexate was tapered by 2.5 mg/mo until discontinued. If disease recurred, daily glucocorticoids or a cytotoxic agent (or both) were increased or restarted. If the glucocorticoids could not be tapered to an alternate-day regimen within 6 months or could not be discontinued completely within 12 months of starting the cytotoxic agent, then that cytotoxic agent was considered a failure. Such patients were treated with the minimum daily glucocorticoid dose that could control their disease.

Human Lymphocyte Antigen Typing

Lymphocytes were separated and typed for HLA-A, B, and C. B lymphocytes were obtained for HLA DR, DQ, and DRw typing. Sera or cells were typed by the fluorescent technique (from the National Institutes of Health) using a microtoxicity assay with eosin staining [20]. Antigenic typing complied with guidelines from the International Histocompatibility Workshop.

Statistical Methods

Top Methods Statistical Methods Results Discussion Author & Article Info References

The frequencies of symptoms and clinical findings were compared for association with and without specific arterial involvement by the Fisher Exact test, with adjustments made for multiple comparisons using the modified Bonferroni method [21]. Laboratory results observed during active disease and remission were compared using the two-tailed paired t-test. Medians for diagnostic delay and duration of follow-up for pediatric-onset and adult-onset Takayasu arteritis were compared by the Wilcoxon two-sample test. The estimated cumulative occurrences of patient remission and relapse over time were determined by the Kaplan-Meier method [22]. The distributions of time to response of the two age groups were compared by the log-rank method. The probability that observed results were consistent with reported demographic and HLA markers in the general population [23, 24] was tested using binomial confidence intervals. Ninety-five percent confidence intervals (CIs) were calculated using observed frequencies in the study group; CIs that did not include the national percentage were statistically significant (P < 0.05).

Results

Top Methods Statistical Methods Results Discussion Author & Article Info References

Sixty unrelated patients with Takayasu arteritis were followed during a median period of 5.3 years (range, 0.5 to 20 years). Fifteen patients were lost to follow-up; 11 of these were followed for 3 to 19 years. The cohort included 58 females and 2 males. The racial background of the patients was as follows: white, 75%; African-American, 11.7%; Asian, 10%; and American Indian, 3.3%; Hispanics made up 7% of the population. The racial distribution was similar to that reported for the United States in the 1990 Census, except for Asians, who comprised 2.9% of the normal population but accounted for 10% of our study group (P = 0.05) [21].

The median age at onset of first symptoms of the disease was 25 years (range, 7 to 64 years), and 21 patients (35%) presented in the third decade. Nineteen patients (32%) were juveniles (<20 years) at disease onset. Eight patients (13%) had onset of their disease after 40 years of age. The median duration of follow-up for pediatric-and adult-onset patients was similar (4.8 compared with 5.4 years, respectively, P > 0.20).

Patients often had symptoms for several months before the diagnosis of Takayasu arteritis was made. The median delay between the onset of first symptoms and the diagnosis of disease was 10 months (range, 0 to 13 years). The pediatric patients had a median delay in diagnosis of 19 months, but adults had a delay of only 5 months (P = 0.01). The diagnosis in 12 patients was not made for at least 3 years, even though 8 had either limb claudication or absent or diminished pulses. The remaining 4 patients had features (such as headache, hypertension, dizziness, arthritis, and erythema nodosum) for which an inflammatory or vascular cause was not considered. Six patients (10%) lacked symptoms at disease onset, but the diagnosis was pursued because of clinical findings of asymmetric blood pressures, absent or diminished peripheral pulses, or a bruit.

Clinical Features

The clinical features of all 60 patients are shown in Figure 2.

View larger version (27K): [in this window] [in a new window]   Figure 2. Frequency of clinical features of Takayasu arteritis at presentation and during the course of disease. Abdom = abdominal; Aberr = aberration; Asym = asymmetric; CHF = congestive heart failure; CNS = central nervous system; Dim = diminished; HBP = high blood pressure; Lt = light; MI = myocardial infarction; Regurg = regurgitation; Subclav = subclavian; TIA = transient ischemic attack; Wt = weight.

Vascular Findings

The most common clinical finding was a bruit (80%), and the most common site was in the carotid vessels (70%). Thirty-three percent of patients had multiple bruits. Femoral (3.3%) and renal (1.7%) bruits were uncommon. Claudication and a diminished or absent pulse were more common in the upper limbs (62% and 53%, respectively) than in the lower limbs (32% and 15%, respectively). Carotodynia was present in 32% of patients. Claudication at disease onset was less common in pediatric patients than in adult patients (11% compared with 46%), and this continued throughout the course of disease (47% compared with 81%, P = 0.02).

Lightheadedness or dizziness occurred in one third of the patients. Fifty-seven percent of patients with some degree of vertebral artery stenosis experienced lightheadedness. In contrast, only 23% of those without vertebral artery disease had lightheadedness (P = 0.04). Lightheadedness was not associated with the presence of lesions in the common carotid artery (P > 0.2).

Hypertension occurred in 20 patients (33%) during the course of their disease. Six of these 20 patients had unilateral (all right-sided) renal artery stenosis and 11 had bilateral renal artery stenosis. One of these patients, who had a nonfunctional right kidney, developed severe hypertension and required a nephrectomy for blood pressure control. Seventy-four percent of patients with renal artery stenosis had hypertension, whereas only 3 (8%) patients without renal artery stenosis had hypertension (P < 0.001). Two of these 3 patients had stenosis of the suprarenal, mid-abdominal aorta. Extremity blood pressure readings corresponded with angiographic progression of stenotic lesions affecting subclavian, iliac, or femoral vessels. When systemic hypertension was present in the setting of 4-limb ischemia, blood pressure could not be accurately monitored by extremity cuff readings. Only central, systemic pressures obtained during angiography were reliable, making treatment of hypertension extremely difficult.

Musculoskeletal and Constitutional Findings

Muscle and joint symptoms occurred in approximately half of all patients. One patient presented with symmetrical, bilateral, peripheral polyarthritis, erythema nodosum, and anterior uveitis at 4 years of age. She was initially thought to have juvenile rheumatoid arthritis, but because of a poor response to conventional therapy, she had a synovial biopsy of the wrist that showed a noncaseating granuloma consistent with sarcoidosis. Levels of serum angiotensin-converting enzyme were elevated. Nine years later, she had stenotic lesions of the main branches of the aortic arch that were typical of Takayasu arteritis.

A 32-year-old woman developed sarcoidosis that presented as bilateral uveitis and hilar adenopathy, concurrent with ischemic symptoms of Takayasu arteritis. A transbronchial biopsy specimen showed a noncaseating granuloma. Special stains and cultures for infectious agents were negative. Her angiogram showed stenotic lesions of the subclavian and splenic arteries. Both patients were African-Americans.

Systemic symptoms were reported in fewer than 50% of patients. In contrast to other clinical features, systemic complaints were more common at disease onset than later in the course of the disease. A vague feeling of ill health was noted in 33% of patients. Fever of 100 °C or more occurred in 27% of patients.

Neurologic and Cardiac Findings

Visual disturbances were most often bilateral. The relations between both common carotid and vertebral artery disease and the presence of visual aberration were statistically significant. Forty-eight percent of patients with and 18% of patients without vertebral artery involvement experienced visual aberration (P = 0.04); 40% of patients with and 12% of patients without common carotid artery involvement had visual aberration (P = 0.05). One patient developed permanent monocular blindness. No other cause was identified and her blindness was attributed to Takayasu arteritis, although the mechanism was unclear. Eight of the 10 patients who had either transient ischemic attacks or cerebrovascular accidents had carotid or vertebral artery disease that anatomically corresponded to neurologic deficits. One of the remaining 2 patients had hypertension, whereas the other had isolated stenosis of the innominate artery.

Nearly 40% of patients had cardiac disease. Aortic regurgitation resulting from a dilated aortic root and separation of valve leaflets was the most common cardiac finding (20%).

Miscellaneous Findings

Headache (42%) was not associated with either common carotid or vertebral artery disease. Four patients had migraine headaches. Shortness of breath, not clinically attributable to cardiac or pulmonary disease, affected 18% of all patients at some time during their disease. Four other patients (7%) had inflammatory bowel disease, two had Crohn disease, and two had ulcerative colitis. Activity of inflammatory bowel disease was independent of Takayasu arteritis disease activity and either preceded (two patients) or followed (two patients) the onset of Takayasu arteritis. Erythema nodosum in five patients was associated with active vascular disease and was independent of inflammatory bowel disease. Skin lesions resolved with remission of Takayasu arteritis.

Angiographic Findings

All patients had aortic angiograms that involved visualization of at least primary tributaries. Two thirds of patients (65%) had aortic lesions (Table 2); 32% of the lesions involved the aortic arch and its branches (type I) [2], and 68% of them involved the aorta and its branches above and below the diaphragm (type III). No patient had involvement of the abdominal aorta and its branches alone (type II). Type IV angiographic findings refer to pulmonary artery involvement. Pulmonary arteriography was done in four patients because of clinical features of pulmonary hypertension. All four patients had stenotic lesions. Both pulmonary arteries were involved in two patients.

Long, stenotic lesions occurred in almost all patients (98%), whereas irregularity of the vessel wall and post-stenotic dilatation were present in one third of the patients and aneurysms were present in 27%. Vascular calcification was rarely seen and, when present, was confined to the abdominal aorta. No racial or age-group differences for site, type of lesion, or extent of disease were found.

Human Lymphocyte Antigen Studies

Twenty-one of the 60 patients had HLA-typing. Fifteen were white, 3 were African-American, 2 were Asian, and 1 was American Indian. All but 1 were women. Thirteen of these 21 patients have been previously described [25]. Eight additional patients were typed. Thirty-three percent of white patients expressed DR4 antigen, 53% had DQ3, and 20% had DR1 (all of these antigens occur in similar frequencies in the normal white population). Forty percent of white patients had Cw3 compared with 23% in the normal population (P = 0.13). The frequencies for Asian, African-American, and American Indian patients typed were too small for statistical analysis. Twelve patients were typed for Bw antigen; none had Bw52. None of the 21 patients had DR12.

Laboratory and Histologic Findings

All patients had routine laboratory tests done at each visit. The erythrocyte sedimentation rate was elevated in 72% of patients during active disease. Erythrocyte sedimentation rates were normal in only 56% of patients with disease remission. Arterial biopsy specimens were obtained from nine patients during arterial bypass procedures when their disease was clinically inactive (see Methods for the criteria for disease activity). Vasculitis was present in four specimens (44%) that were obtained from the aortic arch and the renal and femoral vessels. Although two of the nine patients who had bypass surgery had elevated erythrocyte sedimentation rates at the time of biopsy, only one of the four patients with histologically active disease had an elevated erythrocyte sedimentation rate.

Medical Treatment

Forty-eight patients received glucocorticoids alone or in combination with a cytotoxic agent for active disease. These 48 patients had 70 courses of glucocorticoid therapy (Table 3). Approximately half of the courses of therapy (trials) resulted in remission. Remission was achieved at least once in 60% of patients (95% CI, 45% to 74%) treated with glucocorticoids alone. First-time therapy with glucocorticoids resulted in 52% of patients (CI, 37% to 67%) achieving remission. The estimated median time to remission was 22 months (CI, 11 to 35 months). Although the percentages of pediatric-and adult-onset patients who achieved remission in the first trial were similar (60% and 40%, respectively; P > 0.2), the estimated cumulative time to remission in these two groups differed in the first trial, with medians of 11 months (CI, 9 to 16 months) and 33 months (CI, 14 to 46 months), respectively (P = 0.04).

In 25 patients, a cytotoxic agent was added because of failure to induce remission with glucocorticoids alone or the inability to taper glucocorticoid dosages once the disease was controlled. Although 10 patients received more than one type of cytotoxic agent at different periods of disease activity, no patient received combination cytotoxic therapy. One third of the 39 trials with cytotoxic therapy plus glucocorticoids resulted in remission (Table 3). This therapy led to remission at least once in 40% of patients in this group (CI, 21% to 61%). On the first trial, 32% of patients (CI, 15% to 54%) achieved remission. The estimated median time to remission was 20 months (CI, 15 to >22 months). The response to glucocorticoids in combination with a cytotoxic agent was similar for pediatric- and adult-onset patients. Twenty-three percent of all treated patients, more than 50% of whom were followed for 3 or more years, never achieved a remission.

Remission or Relapse or Both

For the 60 patients studied, 63% of 382 patient-years of follow-up were spent in remission. Forty-five percent (CI, 30% to 59%) of all patients in remission had at least 1 relapse. Thirty-one patients had 5 or more years of follow-up. For these patients, the estimated cumulative percentage for patients who relapsed at 5 years was 55% (CI, 38% to 78%). Eleven of 12 treated patients relapsed (92%; CI, 62% to 100%) who sustained a remission of less than 5 years, whereas only 3 of 12 patients who maintained a remission for more than 5 years (25%; CI, 6% to 57%). None of the 12 patients who presented with inactive disease and were never treated relapsed. Six of these 12 patients had follow-up in excess of 5 years (median, 6.8 years).

Surgical Treatment

Seventy-nine vascular procedures were done in 30 patients [mean, 2.6 per patient; range, 1 to 8 per patient] (Table 4). Indications for surgery were as follows: 1) hypertension associated with critical stenosis of the renal vessel[s]; 2) extremity ischemia limiting activities of daily living; 3) clinical features of cerebrovascular ischemia or critical (>70%) stenosis [or both] of at least three cerebral vessels; 4) moderate [grade II, New York Heart Association] aortic regurgitation; and 5) cardiac ischemia in the setting of proven coronary artery stenosis.

Arterial Reconstruction

Fifty bypass procedures were done in 23 patients. Dacron was the most common synthetic graft and was used for 25 bypass procedures. Autologous vessel grafts were most often selected for bypass of stenotic lesions of the renal and coronary arteries and, in only one patient, the subclavian artery. The saphenous vein was used in 10 of 11 autologous grafts. Thirty percent of the bypass procedures were followed by one or more complications (Table 4). These complications were usually restenosis (24%), thrombosis (4%), hemorrhage (2%), and infection (2%). Complications occurred in 36% of synthetic grafts. Only 1 of 11 (9%) autologous grafts was followed by anastomotic stenosis.

Nine patients had 12 bypass grafts placed for critical stenosis of carotid vessels and for prophylaxis against cerebrovascular accidents. The ascending aorta was the most common proximal anastomotic site. These 9 patients were followed for 8 months to 12 years (median follow-up, 44 months). Only 1 patient has had a cerebrovascular accident since the procedure, and her neurologic deficits resolved. Nineteen other patients had varying degrees of stenosis of at least 2 of 4 cerebral vessels. Either these stenotic lesions were deemed not critical or bypass was considered a substantial risk. These patients have been followed for up to 13.5 years. Only 1 patient in this group has since had a transient ischemic attack (median follow-up, 35 months). Three patients with coronary artery stenosis and angina had coronary bypass surgery that resulted in alleviation of symptoms. Two of 12 patients with aortic regurgitation required aortic valve replacement. Prosthetic valve endocarditis occurred in 1 patient. Aortic root aneurysmal repair was done in 3 patients and was uncomplicated.

Transluminal Angioplasty

Twenty percutaneous transluminal angioplasty procedures were done in 11 patients (Table 4). Percutaneous transluminal angioplasty procedures were most often done on subclavian and renal vessels. Only 56% of angioplasty procedures were successful on the first attempt, whereas only 33% succeeded on a second attempt. Restenosis occurred within 3.5 to 13.6 months. Three patients eventually required a bypass procedure. Three of 7 percutaneous transluminal angioplasty procedures that were done for recanalization of the renal arteries were successful, although in two cases a subsequent bypass procedure was required. Seven patients had percutaneous transluminal angioplasty or bypass graft placement (or both) for blood pressure control. Five of these 7 patients have improved blood pressure control. Three have discontinued all antihypertensive medications, and 2 require less medication.

Morbidity

Disease

Thirty-four patients were interviewed about the long-term effects of the disease on their activities of daily living. They showed three levels of activity: 1) some were unaffected by their disease; 2) some either changed their job description or hours or were unable to work at times of disease activity but were able to resume full function when in remission; and 3) some were placed on permanent disability. Nine patients (26%; CI, 13% to 44%) were not functionally affected by their disease, including 3 who were in continuous remission throughout the study period. The remaining 25 patients (74%; CI, 56% to 87%) had some degree of morbidity. Sixteen patients (47%; CI, 30% to 65%) were placed on permanent disability because they were unable to consistently do or maintain full daily functions. Nine other patients (26%; CI, 13% to 44%) had interruptions in their daily activities coinciding with periods of active disease.

Two of the five patients who had a cerebrovascular accident had residual hemiparesis. One patient developed unilateral blindness. One patient had a nephrectomy for a nonfunctional kidney that was associated with severe hypertension and renal artery stenosis.

Treatment

Of all 48 patients who received therapy, 7 developed 1 or more serious infections requiring hospitalization. The most common infection was herpes zoster (4 patients). Five patients treated with cyclophosphamide developed hemorrhagic cystitis; 1 required transfusion. No patient treated with cyclophosphamide has thus far developed a malignancy or myelodysplasia. All pa tients treated with glucocorticoids became cushingoid, whereas 4 developed cataracts and 3, osteoporotic fractures. Of the 17 patients who received methotrexate, an increase in serum transaminase levels or gastrointestinal intolerance was noted in 24%. Dose reduction was successful in diminishing both side effects, and patients were able to continue methotrexate therapy. Results of our study of methotrexate therapy for Takayasu arteritis have been previously reported [26].

Pregnancy

Five patients with Takayasu arteritis became pregnant, and all had spontaneous vaginal delivery of a normal, live fetus. Three of these patients had aortic lesions above and below the diaphragm. None of these patients, however, had hypertension, renal artery stenosis, aneurysms, or cardiac failure. Only one patient had exacerbation of symptoms during pregnancy.

Angiographic Follow-up

Thirty-four patients had serial angiograms (at least two) done during a single period of active disease (median time between angiograms, 9 months). Eighty-eight percent developed new lesions. Of 18 patients who had serial angiograms during a single period of apparent clinical remission (median time between angiograms, 17.5 months), 61% developed abnormalities at sites previously unaffected by disease. Fifteen of these 18 patients continued in clinical remission, whereas 3 had overt relapse of disease (the earliest occurring 10 months after a new lesion was shown).

Serial angiographic findings were also evaluated in patients with active disease during a single, continuous course of treatment. Sixteen of 20 patients treated with glucocorticoids alone and all 5 patients treated with glucocorticoids plus a cytotoxic agent developed new arterial lesions (median time between angiograms, 10.0 and 6.4 months, respectively). Therefore, whether patients were in apparent remission or were treated for active disease, new lesions frequently developed.

Mortality

Two patients died during the study period. One patient who had disease for 9 years and was in remission died suddenly while asleep. She had extensive disease, with aneurysms of the innominate artery and the aortic arch. No autopsy was done. The other patient committed suicide.

Discussion

Top Methods Statistical Methods Results Discussion Author & Article Info References

We have prospectively studied 60 patients with Takayasu arteritis to evaluate their clinical features and the outcome of therapeutic interventions. This study is unique: 1) Angiographic evidence of disease was required in all patients; 2) sequential angiography was used to follow disease progression; 3) defined criteria for disease activity were used; and 4) established guidelines for the treatment of active disease were applied. The median period of follow-up per patient was 5.3 years (range, 0.5 to 20 years).

The median age at disease onset of 25 years is similar to that seen in Asian countries [1-4]. This finding is in contrast to the later age of onset (41 years) reported in Europe [12, 13]. Females predominated, as previously reported in most other series of patients. Males are not exempt, however, and some authors report almost equal occurrence in both sexes [27, 28]. Asians were disproportionately over-represented in our cohort compared with their presence in the U.S. population [23]. Such ethnic predilection has raised questions about the possible role of HLA antigens in predisposing persons to Takayasu arteritis. Fifty percent of Japanese patients with Takayasu arteritis have statistically significant associations with Bw52 and DR12 antigens when compared with healthy Japanese controls [29, 30]. An earlier study of 11 patients with Takayasu arteritis by our Institute found HLA-DR4 in 7 of 10 patients and MB3 in 10 patients [25]. However, similar associations were lacking in our larger updated cohort of patients. Other North American studies of Takayasu arteritis have also failed to identify HLA associations with disease [16, 31].

The spectrum of presentation, disease severity, and pace of disease progression can often lead to an inaccurate assessment and a delay in diagnosis. In this study, the delay in diagnosis in children may reflect failure to recognize the disease and, therefore, the hesitancy of pediatricians to request diagnostic angiographic studies. The limited reliability of clinical, laboratory, and angiographic markers of disease activity was also evident. Systemic symptoms occurred in only a third of patients at the onset of disease and in an additional 10% throughout their disease course. These findings fail to support the notion that most patients with Takayasu arteritis have a "triphasic" illness characterized by prepulseless inflammation, followed by painful ischemic vessels, and ending in "burnt-out disease" [14, 16]. Rather, patients with Takayasu arteritis had a diverse course. Twenty percent of our patients had a monophasic disease that did not require therapy, and 57% of all patients never had constitutional symptoms.

Vascular ischemic symptoms were the hallmark of the disease and were apparent as a bruit, claudication, or diminished pulses. Most of the patients we studied had extensive disease of the aorta and its branches, and stenotic lesions were 3.6-fold more common than were aneurysms. Calcification was uncommon but has been reported in 10% to 25% of patients [32]. No patient had involvement of the abdominal aorta alone. Similar angiographic features of Takayasu arteritis have been reported in India, Thailand, and Mexico [4, 14, 33]. In contrast, aortic arch disease has been more commonly observed among the Japanese, whereas mid-aortic disease, aneurysms, and renal artery stenosis have been more common in children from India, South Africa, Korea, and Singapore [10, 34-36]. Regardless of nationality, a young patient with hypertension should be further evaluated for extremity claudication, cerebral ischemia, and renal artery stenosis. The differential diagnosis should include Takayasu arteritis, an illness in which most patients have multiple abnormalities of the aorta or its branches (or both). These vessels should be clinically and angiographically evaluated.

Hypertension was less common in our series of patients compared with others (33% compared with 60%) but correlated significantly with the presence of renal artery stenosis in 74% of patients with hypertension [1, 14]. Abnormal vascular compliance and dysfunctional baroreceptors are alternative pathophysiologic mechanisms that may account for hypertension in the absence of renal artery stenosis [37, 38]. Surgical correction of renal artery stenosis is usually an effective means of decreasing or eliminating hypertension. In the presence of extremity vascular stenosis, peripheral blood pressure readings were often misleading. A further dilemma occurs when central hypertension and cerebral vessel stenosis occur simultaneously. In these patients, aggressive blood pressure control may result in cerebral ischemia. Central blood pressure determinations should always be obtained in patients with extremity vessel stenosis to assess the reliability of conventionally obtained blood pressure measurements. The treatment of central hypertension in the setting of carotid or vertebral stenosis (or both) remains an unsettled issue.

The associations of lightheadedness and dizziness with vertebral artery stenosis were statistically significant. Eighty percent of patients who had transient ischemic attacks or a cerebrovascular accident had carotid or vertebral artery disease (or both). These findings would imply that bypass procedures be considered in patients with symptomatic, surgically approachable carotid disease [39]. However, it is controversial whether clinicians should treat asymptomatic carotid lesions. Symptomatic coronary artery disease occurred in 13% of our patients, a percentage similar to that reported in other studies (6% to 16%), including those where coronary angiography was done on all patients [1, 4, 16, 40].

Studies in which pulmonary angiography was done on unselected patients found pulmonary involvement in 15% to 70% of patients [41-44]. In our study, all four patients with clinical features of pulmonary hypertension had pulmonary artery stenosis. Other patients may have had subclinical pulmonary hypertension. Nonetheless, we would not recommend routine pulmonary angiography in all patients with Takayasu arteritis.

The current laboratory and angiographic parameters from patients with Takayasu arteritis lack sensitivity; thus, prospective studies that evaluate treatment are difficult to do. In our patients, the erythrocyte sedimentation rate was elevated in 72% of patients with active disease and in 56% of patients in remission. Therefore, the erythrocyte sedimentation rate was an unreliable marker of disease course. Other reports support this evaluation, although others have found the erythrocyte sedimentation rate to be a reliable and useful marker to determine therapy [14, 45]. To date, other surrogate markers of active disease (endothelins, von Willebrand factor antigen, factor VIII), have either not been adequately studied or have not been found to be superior to the erythrocyte sedimentation rate in monitoring patients [46-48]. Angiography is considered the gold standard in delineating abnormal vessels in patients with Takayasu arteritis.

In this study, angiograms were done in some patients on a scheduled protocol regardless of their clinical disease course or were done in other patients based on presumed disease exacerbations. Our data established that although angiography was sensitive in detecting abnormal anatomy in patients with Takayasu arteritis, it could not determine the cause of vascular narrowing (for example, active inflammation as opposed to intimal and adventitial fibrosis) of the vessel wall. Therefore, angiographic findings in patients with Takayasu arteritis, like the erythrocyte sedimentation rate, have to be interpreted within the clinical context.

Glucocorticoids still remain an effective palliative agent for most patients with active Takayasu arteritis. Sixty percent of our patients treated with glucocorticoids alone responded at least once. Other series of patients report a response to glucocorticoids ranging from 20% to 100% [5, 14, 15]. The addition of cytotoxic agents induced remission in 40% of our steroid-resistant patients. The relapsing nature of the disease often required repeated courses of therapy. Twenty-three percent of all patients had continuous disease regardless of therapy. The morbidity associated with glucocorticoids and cyclophosphamide therapy was similar to our experience with the same drug regimen in patients with Wegener granulomatosis [17].

Although cytotoxic therapies in this study were not randomly assigned, methotrexate appeared to be a viable alternative immunosuppressive agent compared with cyclophosphamide in the treatment of patients with Takayasu arteritis and has substantially less toxicity [26, 49]. We now realize that control rather than eradication of disease is a more realistic therapeutic goal for patients with Takayasu arteritis.

Whenever possible, surgery was done when the disease was quiescent. Previous studies have shown the increased risk for failure of surgical procedures during active disease [32, 50]. The finding that 44% of arterial biopsy specimens from our patients showed active vasculitis underscores the poor correlation between clinical assessment and disease activity. The histologic finding of active inflammation in as many as 40% of biopsy specimens has been previously reported [51]. This histologic evidence of covert active inflammation argues for a prospective trial of perioperative steroids in patients with Takayasu arteritis, in whom continued treatment would depend on histopathologic results.

For bypass surgery, areas of the arterial tree that were unaffected by disease were the preferred anastomotic sites. Autologous grafts were more successful than synthetic grafts, but their application is often limited by the volume of flow being diverted in large-vessel disease. Twenty-four percent of bypass procedures were complicated by stenoses. Our bypass patency rate (70%) was comparable to data from other studies [52, 53]. The success of percutaneous transluminal angioplasty in patients with Takayasu arteritis is variable [54, 55]. Percutaneous transluminal angioplasty was most applicable for patients with renal artery stenosis but had a success rate of only 56% within 1 year. However, percutaneous transluminal angioplasty is associated with a relatively low risk; in patients with short, proximal stenotic lesions who require urgent relief, this procedure may be palliative.

We were intrigued by the finding of sarcoidosis in 2 of our 60 patients. The distribution of lesions described in patients with sarcoid vasculopathy is similar to that seen in patients with Takayasu arteritis. The association of sarcoidosis with features of Takayasu arteritis has led to the concept of "the Takayasu syndrome" [56, 57]. Inflammatory bowel disease has also been reported in association with Takayasu arteritis [58, 59]. The occurrence of inflammatory bowel disease in 7% of our patients exceeds that expected for the incidence of either disease in the normal population (Takayasu arteritis: 1 to 2.6/million persons; ulcerative colitis: 6 to 8/100 000 persons; Crohn disease: 2/100 000) [16, 60]. The occurrence of two such rare diseases in the same person raises questions of a common cause or mutual susceptibilities.

Our study is the first to evaluate disability in patients with Takayasu arteritis. Seventy-four percent of patients had a substantial compromise in activities of daily living. In our series of patients, mortality attributed to Takayasu arteritis was low (2%). Five-year and 10-year survival rates of 80% to 90% have been reported from large series of patients [2, 4, 16]. Hypertension, cardiac involvement, aortic or arterial aneurysms (or both), and severe functional disability predict greater morbidity and mortality. Death may occur from cardiac failure, myocardial infarction, stroke, aneurysmal rupture, and renal failure.

Only five of our patients became pregnant during the study period, and all had a favorable outcome. Data from large series of patients indicate that hypertension, aneurysms, and extent of disease are associated with increased risks for maternal and fetal death [61, 62]. Increases in intravascular volume during pregnancy may exacerbate hypertension, aortic insufficiency, and congestive heart failure [63, 64]. Immunoinflammatory features of disease activity appear to be unaffected by pregnancy. Nonetheless, if pregnancy is contemplated, it may be safer during periods of remission. Moderate dosages of glucocorticoids have generally not had an adverse effect on fetal development. Indications for cesarean section are the same as for any other pregnancy and bear no relation to the coexistence of Takayasu arteritis. Pregnancy in patients with Takayasu arteritis is best managed by high-risk obstetricians and neonatologists in conjunction with the internist.

Since our earlier studies of Takayasu arteritis in 1985 [17], we have a better appreciation of the limitations in the evaluation of disease activity. The current therapeutic options, although palliative in most patients, cause substantial toxic reactions and fail to prevent relapse. Improved, less costly vascular imaging techniques, which obviate the risks for repeated radiation exposure and have a resolution similar to that of angiography, are needed to assess the nature of vessel involvement and disease progression. Sensitive and specific marker(s) of disease activity are needed to guide medical and surgical intervention. New therapeutic modalities aimed at abating the inflammatory and ensuing fibrotic responses would decrease the statistically significant morbidity experienced by most patients. At present, these modalities are not available.