**Background:** Graves' orbitopathy (GO) is an autoimmune orbital disorder and the main extrathyroidal manifestation of Graves' disease. For moderate-to-severe and active GO, intravenous glucocorticoids (GCs) are first-line treatment. The standard EUGOGO protocol uses 12 weekly infusions (total 4.5 g methylprednisolone), requiring frequent hospital visits. This study evaluated the safety of a non-standard regimen with three consecutive daily 1 g methylprednisolone pulses followed by intramuscular doses (total 3.6 g), aiming to reduce hospitalization burden.
**Methods:** This retrospective single-centre study reviewed 161 medical records of patients with active, moderate-to-severe (n=110) or sight-threatening (n=51) GO treated between 2014 and 2021. The regimen consisted of intravenous methylprednisolone 1 g for three consecutive days, then intramuscular methylprednisolone 600 mg in divided doses. Most patients also received l-ornithine l-aspartate (LOLA) and a proton pump inhibitor. Data included demographics, smoking status, disease duration, clinical activity score (CAS), laboratory results (glucose, ALT, AST, TRAb, etc.), and side effects. Hyperglycemia was defined as fasting glucose ≥100 mg/dL or postprandial ≥140 mg/dL; clinically significant aminotransferase increase was >3× upper limit of normal. Statistical analyses included Wilcoxon test, Mann–Whitney test, Pearson's Chi-squared test, logistic regression, and multidimensional correspondence analysis (MPA).
**Key Results:** The cohort was 70.8% female, median age 56 years, median BMI 26.7 kg/m², 36.6% smokers, median disease duration 12 months, median CAS 4. Side effects occurred in 116 hospitalizations (71%). The most common were hyperglycemia (n=95, 59% of patients) and elevated aminotransferases (n=31, 19%). Other side effects included headache (n=11), facial erythema (n=11), abdominal pain (n=5), blood pressure increase (n=3), insomnia (n=3), resting tremor (n=2), peripheral oedema (n=2), and delirium (n=1). No severe complications (death, liver failure, thrombosis) were reported. After therapy, median glucose increased from 96 to 119 mg/dL (p<0.001), ALT from 19 to 23 U/L (p<0.001), and AST decreased from 19 to 17 U/L (p=0.010). Only two patients had aminotransferases >3× ULN. In multivariate logistic regression, higher TRAb (OR 0.951, 95% CI 0.923–0.98, p=0.001) and higher CAS (OR 0.734, 95% CI 0.571–0.943, p=0.016) were associated with lower odds of hyperglycemia. For elevated aminotransferases, higher baseline ALT (OR 1.161, 95% CI 1.077–1.251, p<0.001), male sex (OR 0.116, 95% CI 0.023–0.58, p=0.009), and GO duration >1 year (OR 3.608, 95% CI 1.035–12.578, p=0.044) were significant predictors. MPA showed that smokers and patients not receiving LOLA had higher probability of elevated aminotransferases.
**Clinical Implications:** This non-standard regimen appears safe for active GO, with no serious adverse events. Hyperglycemia is common but manageable. The regimen reduces the need for weekly hospital visits, which is particularly beneficial for patients in regions with limited access to specialized centres. However, careful patient selection (including virological and metabolic screening) and monitoring of glucose and liver enzymes are essential. The study's limitations include its retrospective design, missing data, and lack of a control group. Prospective comparative studies are needed to confirm these findings.