Giant-cell arteritis related strokes: scoping review of mechanisms and rethinking treatment strategy?

Stroke is a rare and severe complication of giant cell arteritis (GCA). Although early diagnosis and treatment initiation are essential, the mechanism of stroke is often related to vasculitis complicated by arterial stenosis and occlusion. Its recurrence is often attributed to early steroid resistance or late GCA relapse, so immunosuppressive treatment is often reinforced. However, many questions concerning the mechanisms of stroke remain elusive, and no review to date has examined the whole data set concerning GCA-related stroke. We therefore undertook this scoping review. GCA-related stroke does not necessarily display general signs and inflammatory parameters are sometimes normal, so clinicians should observe caution. Ischemic lesions often show patterns predating watershed areas and are associated with stenosis or thrombosis of the respective arteries, which are often bilateral. Lesions predominate in the siphon in the internal carotid arteries, whereas all the vertebral arteries may be involved with a predominance in the V3-V4 segments. Ultrasonography of the cervical arteries may reveal edema of the intima (halo sign), which is highly sensitive and specific of GCA, and precedes stenosis. The brain arteries are spared although very proximal arteritis may rarely occur, if the patient has microstructural anatomical variants. Temporal artery biopsy reveals the combination of mechanisms leading to slit-like stenosis, which involves granulomatous inflammation and intimal hyperplasia. The lumen is sometimes occluded by thrombi (<15%), suggesting that embolic lesions may also occur, although imaging studies have not provided strong evidence for this. Moreover, persistence of intimal hyperplasia might explain persisting arterial stenosis, which may account for delayed stroke occurring in watershed areas. Other possible mechanisms of stroke are also discussed. Overall, GCA-related stroke mainly involves hemodynamic mechanisms. Besides early diagnosis and treatment initiation, future studies could seek to establish specific preventive or curative treatments using angioplasty or targeting intimal proliferation.


Estimation of biases
Since included studies were based on consecutive GCA patients split by ischemic events and since most the data were common clinical and biological parameters, bias in the selection of patient participant to the study, bias due to missing data and bias in selection of the reported patient were low.Bias in the measurement of outcome may have occurred in ill or subjectively defined clinical signs, like scalp tenderness or jaw claudication.
Bias due to confounding effect may occurred for transient ischemic signs, which were clinical grounded, and may have influenced the risk of persisting ischemic event also defined by radiology.
Definition of ischemic event was variable among studies especially by the variable inclusion, besides strokes, of ophthalmic or other cranial ischemic events, potentially generating a bias in classification of patients.Admitting reasonably similar mechanisms of ischemic events, the direction of the bias is likely to increase the effect estimate.

Statistical analysis
The estimated pooled effect of each parameter is calculated using the mean of standardized effect of each study weighted by the inverse of variance.Statistic heterogeneity is evaluated by the Cochran's Q test, which is complemented with the I 2 statistic: heterogeneity values of I 2 of 25%, 50%, and 75% was designated as low, moderate and high.Assuming a high level of heterogeneity, random effect model was used to assess the pooled effect and 95% confidence interval (CI).The presence of publication bias was evaluated by funnel plots.For each risk factor, results were given as forest plots.P-value below 0.05 was considered significant.All statistical analyses are performed by SAS software version 9.4 (SAS Institute, Cary, NC, USA) and Review Manager (RevMan) 5.4.1 software from the Cochrane Collaboration.Results are given in odds ratio, 95% confidence interval (OR [95% CI]).

Results
Twenty-two studies published between 1991 and 2023 were extracted from the literature, and one was excluded due to absence of stroke from cranial ischemic events.
Forest plots and odds ratio for risk factors of ischemic events associated GCA are given below in Suppl.Table S2 and Forest plots (for discussion: see Main document).

2
Supplementary Tables 2.1.1Supplementary Table S1.Inclusion criteria of ischemic events in selected studies.

Authors
S18. Forest plot showing the risk of ischemic complication associated with abnormal temporal artery on physical examination.3.19 Figure S19.Forest plot showing the risk of ischemic complication associated with erythrocyte sedimentation rate (ESR; in mm by hours).3.20 Figure S20.Forest plot showing the risk of ischemic complication associated with age at diagnosis.3.21 Figure S21.Forest plot showing the risk of ischemic complication associated with C-reactive protein (CRP) levels (mg/L).3.22 Figure S22.Forest plot showing the risk of ischemic complication associated with hemoglobin level (g/dL)S24 Forest plot showing the risk of ischemic complication associated with albumin (g/L).3.26 Figure S26 Forest plot showing the risk of ischemic complication associated with atrial fibrillation.