Giant cell arteritis (GCA) is complicated in 10 to 20% of cases by permanent visual ischemia (PVI). International guidelines advocate the use of intravenous pulse of methylprednisolone from 250 to 1000mg per day, for three days, followed by oral prednisone at 1mg/kg per day. The aim of this study is to assess whether this strategy significantly reduces the risk of early PVI of the second eye, compared with direct prednisone at 1mg/kg per day.

We conducted a multicentre retrospective observational study over the past 15 years in 13 French hospital centres. Inclusion criteria included: new case of GCA; strictly unilateral PVI, prednisone at dose greater than or equal to 0.9mg/kg per day; for the intravenous methylprednisolone (IV-MP) group, total dose between 900 and 5000mg, close follow-up and knowledge of visual status at 1 month of treatment, or earlier, in case of contralateral PVI. The groups were compared on demographic, clinical, biological, iconographic, and therapeutic parameters. Statistical analysis was optimised using propensity scores.

One hundred and sixteen patients were included, 86 in the IV-MP group and 30 in the direct prednisone group. One patient in the direct prednisone group and 13 in the IV-MP group bilateralised, without significant difference between the two strategies (3.3% vs 15.1%). Investigation of the association between IV-MP patients and contralateral PVI through classical logistic regression, matching or stratification on propensity score did not show a significant association. Weighting on propensity score shows a significant association between IV-MP patients and contralateral PVI (OR=12.9 [3.4; 94.3]; P<0.001). Improvement in visual acuity of the initially affected eye was not significantly associated with IV-MP (visual acuity difference 0.02 vs –0.28 LogMar), even in the case of early management, i.e., within the first 48hours after the onset of PVI (n=61; visual acuity difference –0.11 vs 0.25 LogMar). Complications attributable to corticosteroid therapy in the first month were significantly more frequent in the IV-MP group (31.8 vs 10.7%; P<0.05).

Our data do not support the routine use of pulse IV-MP for GCA complicated by unilateral PVI to avoid bilateral ophthalmologic damage. It might be safer to not give pulse IV-MP to selected patients with high risks of glucocorticoids pulse side effects. A prospective randomised multicentre study comparing pulse IV-MP and prednisone at 1mg/kg per day is desirable.