Reassessing classification criteria for Takayasu arteritis: Comment on the article by Lee et al


Reassessing Classification Criteria for Takayasu Arteritis: Comment on the Article by Lee et al

To the Editor:

We read with interest the recent article by Lee et al (1) describing a retrospective study of 38 patients who met the American College of Rheumatology (ACR) 1990 classification criteria for Takayasu arteritis (TA) (2). We agree with Lee and colleagues' conclusion that 18F-fluorodeoxyglucose–positron emission tomography (FDG-PET) may be a useful tool for aiding in the assessment of disease activity in patients with TA; however, the description and definition of the TA patient population warrants some balanced comments.

Vasculitides are a heterogeneous group of syndromes, and the 1990 criteria established by the ACR were designed to differentiate among patients with 7 types of vasculitis (3). In 1998 it was concluded that the ACR 1990 classification criteria function poorly in the diagnosis of specific vasculitides (4). Patients who do not have a vasculitis syndrome may meet ACR criteria, and patients who have a specific type of vasculitis may meet criteria for >1 disease (4). TA and giant cell arteritis (GCA) are the 2 most common types of large vessel vasculitis. Historically, TA and GCA have been considered distinct diseases based on differences in age at onset, ethnic distribution, and clinical features, including predilection for involvement of certain arterial territories. As recent observations have shown that the histopathology of arterial lesions in these 2 diseases is difficult to distinguish, it is speculated whether TA and GCA are part of a spectrum of manifestations within a single disease (5). On angiography, arteriographic lesions are recognized as the sequelae of inflammation; strong similarities and subtle differences in these lesions were observed between GCA and TA (6).

GCA is associated with polymyalgia rheumatica (PMR) in ∼40% of patients (7). 18F-FDG–PET or hybrid 18F-FDG–PET/computed tomography (CT) images of patients with PMR reveal a characteristic pattern of pathologic 18F-FDG uptake in the soft tissue and ligaments (perisynovitis or enthesitis) around the shoulders, lumbar spinous processes, and ischial tuberosities (8). 18F-FDG–PET or PET/CT images of patients with GCA (with and without PMR) show a homogeneous/smooth, linear or long segmental pattern of 18F-FDG uptake in the thoracic aorta and its main branches. The maximal standard uptake value of FDG rarely exceeds 5.0 (9, 10).

TA has an incidence of only 2 cases per 1 million persons. The mean age at onset is 35 years, and prevalence of the disease in women is 2–25 times higher than that in men. During the course of the disease, stenoses, occlusions, and aneurysms may occur; the mortality rate of TA is similar to that found in malignancies and is reported to be as high as 35% 5 years after diagnosis. Both 18F-FDG–PET or PET/CT and pathology studies have revealed that inflammation in TA may be most active in the aorta and its branches, but in a more focal, localized, inhomogeneous pattern (11, 12). TA has also been reported to manifest as isolated involvement of renal arteries (for which renal revascularization was required) (13), of pulmonary arteries resulting in occlusion (14), of vertebral arteries resulting in neurologic symptoms (15), and of coronary arteries, requiring bypass surgery (16). Interestingly, the large number of recent publications on 18F-FDG-PET or PET/CT imaging of large vessel vasculitis may lead to one predominant conclusion: that 18F-FDG-PET images reflect the more aggressive clinical course of TA compared to GCA.

It is confusing, however, that a patient with isolated intense pathologic 18F-FDG uptake in the vertebral arteries and with neurologic symptoms was diagnosed as having TA in one case report (15) and as having GCA in another (17), given the age at onset of disease (which is a required component of the ACR 1990 criteria). The idea that GCA and TA are part of a spectrum of conditions that comprise a single disease was first proposed by Hall (18), who suggested that PMR, GCA, and TA constitute an “unholy trinity” of a single disease.

It is necessary to avoid a significant bias in how data are gathered for classifying each disease. The criteria for GCA were developed at a time when involvement of the aorta and its main branches was not a well-recognized feature of GCA, and instead there was a focus on cranial features of the disease. Through 18F-FDG-PET or PET/CT imaging it has become evident that large vessel involvement is much more prevalent in GCA than previously recognized. The notion that there is substantial underdiagnosis of GCA is supported by several observations. First, a retrospective study of arterial changes in 20,591 autopsy subjects (in 1973) showed that PMR with signs of aortic GCA is far more common than is diagnosed clinically. Arteritis was found in 79 subjects (0.4%), and 38 of those cases had temporal arteritis. It was concluded that the incidence of 20–30 cases in GCA per 100,000 persons is probably an underestimation (19). Second, large vessel involvement in GCA is not commonly recognized in patients with positive findings on temporal artery biopsy. Third, large vessel involvement in GCA may occur without involvement of the temporal arteries. In a recent series it was found that half of patients with increased 18F-FDG uptake in the thoracic aorta and its main branches, which was suggestive of GCA, had positive temporal artery biopsy results (9).

Based on the epidemiology of both GCA and TA, it is highly unlikely that all 38 patients in the retrospective study by Lee et al had TA. As the median standard uptake value intensity in 24 patients with clinically active disease was 1.01 (range 0.71–2.36), it is more probable that this report describes 37 patients with GCA and 1 patient with TA. Now that imaging techniques, such as arteriography and 18F-FDG–PET or PET/CT, are more widely used, the inappropriateness of the ACR 1990 criteria in the nomenclature of large vessel vasculitis is becoming more evident. Therefore, the time may very well have come to regard the ACR 1990 criteria as obsolete and to develop an updated version incorporating the above-mentioned considerations.

H. Balink MD*, P. M. Houtman MD, PhD*, H. J. Verberne MD, PhD†, * Medical Center Leeuwarden, Leeuwarden, The Netherlands, † Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands.