Temporal artery arteritis (TA), also known as cranial arteritis or Horton's arteritis, belongs to the family of giant cell arteritis, which are granulomatous vasculitides affecting medium and large calibre arteries. It occurs mainly in individuals over 50 years of age, with a peak incidence in the eighth decade. It is more prevalent in females (3:1).1,2 The incidence in Spain is around 10 cases per 100,000 population.3

Diagnosis is based on the criteria proposed in 1990 by the American College of Rheumatology (ACR-SCORE),4 which include clinical signs and symptoms typical of the disease: age >50 years, localised headache of recent onset, pain on palpation of the temporal artery or decreased pulse, sedimentation rate >50 mm/hour in the first hour, histological abnormalities (necrotising arteritis or granulomatous process with multinucleated giant cells). The presence of 3 or more of the 5 parameters establishes a diagnosis of TA with a sensitivity of 93.5% and a specificity of 91.2%.4

TA biopsy is usually performed on an outpatient basis with a surgical time of 20−50 min. It has a complication rate of between .5% and 1%; the main complication is bleeding (haemorrhage or haematoma), followed by surgical wound infection, nerve damage, vascular damage (arterial or venous), skin necrosis, or stroke. We did not assess complications associated with biopsy in this study, because we currently use minimally invasive techniques and biopsy is an invasive procedure.5 Biopsy remains the gold standard for the diagnosis of TA.2

Treatment with high-dose glucocorticoids is indicated for these patients, which have many side effects such as bone fractures, avascular necrosis of the hip, diabetes mellitus, infections, gastrointestinal bleeding, cataracts, and hypertension.2,6

Cost-effectiveness analyses are not used to find the cheapest or most effective technique, but to optimise health system resources to achieve the maximum health benefit with the available resources.7,8

In our hospital, TA is diagnosed by the internal medicine department, which requests the analyses, ultrasound, and biopsies. The aim of this study is to determine the costs generated by this diagnosis and, secondly, to conduct a cost-effectiveness study of different diagnostic strategies (clinical, biopsy, and/or Doppler ultrasound) and treatment (whether a negative diagnosis of TA implies discontinuation of corticosteroids).

We conducted a single-centre, retrospective observational study, according to clinical practice, of patients with suspected TA from 2012−2021. Patients with a primary and secondary diagnosis of temporal artery arteritis were requested from the hospital coding service. Approval for this study was obtained from the hospital’s ethics committee (PI 21-2443).

Population included in the study and descriptive analysis of the variables

Patients with a diagnosis of TA made in our hospital after 2012. In this period there were 107 patients with a diagnosis of TA, of which 32 patients were excluded (26 diagnosed before 2012, 6 diagnosed in a different hospital) (Fig. 1).

Fig. 1.

Graphic summary of the study.

(0.12MB).

Demographic data and comorbidities were collected, as well as signs and symptoms suggestive of TA, analytical data such as ESR > 20.00 mm and CRP > 5.00 mg/dl, Doppler ultrasound assessing for the presence of halo sign, and biopsy with the presence of multinucleated giant cells. A diagnosis of TA was considered with an ACR-SCORE greater than or equal to 3.

TA is a disease that is of concern to physicians and patients in terms of diagnosis and potential relapses.11 Thorough history taking and clinical examination are still of paramount importance in its diagnosis.12 The symptoms and signs with the highest predictive value are visual symptoms (loss of vision or diplopia), new-onset headache, preauricular pain, jaw claudication, ESR greater than 50, and age over 50 years.2,11,13,14 In our series, headache, temporal pain, and jaw claudication were predictors of giant cell arteritis.

Smith et al.15 in 1997 were the first to use colour Doppler ultrasound to establish a diagnosis of TA. They concluded that there are some characteristic signs detectable with Doppler ultrasound, the most specific being the presence of a dark halo sign, secondary to oedema of the arterial wall. Since then, several studies have evaluated this technique, obtaining a sensitivity and specificity for the halo sign of 68%–69% and 81%–91% in the diagnosis of TA.16–18 In our study, the presence of halo sign has a sensitivity and specificity of 58.3% and 63.6%, respectively, for diagnosis of TA, which differs from the published results; this is because the presence of a halo sign is compared with positive pathological anatomy and not with diagnosis according to an ACR-SCORE greater than 3. However, for González Porto et al.19 who compare ultrasound with clinical findings, sensitivity is 42.6% and specificity is 65.7%, which are results more similar to those obtained in our study. However, this was not the aim of our study.

Some authors suggest that in the case of a high pretest probability (clinical and analytical), a pathological Doppler may be sufficient to diagnose TA, without the need for a biopsy.11,12,20,21 In this study, elevated ESR and CRP are predictors of a positive biopsy or presence of a halo sign on ultrasound, and therefore diagnosis of TA.

The sensitivity of temporal artery biopsy can be as low as 39%, and about 7% of all temporal artery biopsies may not actually consist of arterial tissue.22 In our work the biopsy yielded a sensitivity of 33% and only 4% of the samples were insufficient. The positivity of temporal artery biopsy depends on the size, with biopsies smaller than .5 cm the yield is low and those larger than 2 cm in length tend to be high yield.2 It was not the aim of this study to determine the length of the temporal artery analysed. In the study by González-Gay et al.,23 190 patients were diagnosed with TA; the best predictive model for biopsy-proven TA included a history of constitutional syndrome (OR = 6.1), an abnormal temporal artery on physical examination (OR = 3.2), and the presence of visual complications (OR = 4.9). A subset of patients had a high probability of a negative temporal artery biopsy, but met the 1990 ACR classification criteria for TA when they were applied. This subset with negative temporal artery biopsy had less severe ischaemic complications than those with biopsy-proven GCA. Thus, in patients without visual manifestations, with a normal temporal artery on physical examination and without constitutional syndrome, the risk of having an abnormal temporal artery biopsy was low.23

Many temporal artery biopsies are negative, less than 60% of patients with clinical temporal arteritis have a positive biopsy,11,13,24,25 in our work only 50% of biopsies were positive, we consider it essential like Gajree et al.13 and Lyons et al.,11 to find ways to reduce the number of patients undergoing unnecessary biopsy, to reduce workload and optimise services. It is no longer standard clinical practice, due to the use of temporal artery ultrasound, to perform a second contralateral biopsy in patients with a high suspicion of GCA, as it increases the biopsy yield by 10%.26 This study demonstrates that there are clinical factors (headache, temporal pain, and jaw claudication) and analytical factors (elevated ESR and CRP) already described by other authors, such as González-López12 who present a tool such as the calculator (logistic regression analysis) that can help predict the outcome of the biopsy, help improve clinical decision-making, and avoid unnecessary biopsies.10

The cost-effectiveness study by Luqmani et al.21 indicates that ultrasound alone is more cost-effective than biopsy alone, largely due to its much lower cost (58 versus £514) and higher sensitivity (54% versus 39%), in our work there is also an ostensible difference between the 167.26 for Doppler ultrasound and 339.75 for biopsy and we have similar results for sensitivity (58% versus 33%). If we compare this with the cost incurred in our hospital (414/patient) (many patients undergo both biopsy and Doppler ultrasound), we are talking about a saving ranging from 59% to 18%.

If we follow the most cost-effective strategy, we would rely on ultrasound and clinical judgement alone as a means of diagnosis.21 The rationale behind the additional use of biopsy, despite a negative scan, would be to provide further evidence to rule out disease, and to support discontinuation of therapy in the event that both tests are negative.2,20,21 Brabyn et al.14 report corticosteroid maintenance in 70%–75% of patients, which is comparable to our study, where only 6% of our patients were withdrawn from corticosteroid therapy. In this study, 29.3% of patients were kept on corticosteroids despite a negative diagnosis. A reason for this could be that 28% of the patients had a diagnosis of polymyalgia rheumatica. Also, in our study there were 2 deaths, and both were taking corticosteroids at high doses, although mortality and potential complications were not the aim of this study.

It has already been published that the use of a pre-test clinical probability pathway and non-crucial diagnostic tests to initiate early treatment and prevent vision loss are cost-effective strategies,11,12,21 which we fully agree with. We must be aware of the costs generated by our actions and even potential complications for patients. With this study, we can conclude that when clinical and laboratory data are available, a diagnosis of TA can be made, leaving biopsy for those patients in whom we wish to discontinue corticosteroids, i.e., not performing a biopsy when we know that we are not going to discontinue corticosteroids.

Work is needed to validate the predictive calculator.

In our population, the presence of headache, temporal pain, and jaw claudication are predictors of giant cell arteritis. Elevated ESR and CRP are predictors of positive biopsy and presence of halo sign on ultrasound.

ACR-SCORE ≥ 3 with Doppler ultrasound or biopsy, as well as diagnosis with discontinuation of corticosteroids are cost-effective.

The present research received no specific support from public sector agencies, commercial sector, or non-profit entities.

The authors have no conflict of interests to declare.