Summary Risk of fracture due to glucocorticoid-induced osteoporosis (GIO) can be reduced by anti-osteoporosis (OP) medications. The proportion of patients on long-term glucocorticoid therapy who received anti-OP medications according to the GIO management guidelines has increased in recent years, but is still suboptimal. Introduction Adherence of physicians to guidelines for glucocorticoid (GC)-induced osteoporosis (GIO) management is currently unclear. This study aimed to clarify the state of guideline adherence by physicians in Japan and identify factors associated with guideline adherence using a nationwide health insurance claims database (NDBJ). Methods Patients aged ≥ 50 years who were prescribed GC for ≥ 90 days after 180 days without a GC prescription and who were followed up for osteoporosis (OP) management for the subsequent 360 days during the period spanning 2012–2018 were selected from the NDBJ. Guideline adherence was evaluated with the proportion of patients who received OP management as recommended by the Japanese guidelines. Information on previous vertebral and hip fractures, dementia, and polypharmacy was obtained. Factors associated with OP management were evaluated by logistic regression analysis. Results A total of 512,296 patients were considered to be at high risk of fracture according to the guidelines. Proportions of patients receiving OP management (BMD testing or anti-OP medications) have increased in recent years. In 2017, 33.7% of men and 55.3% of women received OP management in the initial 90 days of GC therapy. Female sex, previous anti-OP medications, polypharmacy, and higher GC dose were significantly associated with receiving OP management, while dementia showed an inverse association. A prior history of hip fracture, a strong risk factor for future fracture, was not significantly associated with receiving OP management. Conclusions Although guideline adherence by physicians has increased in recent years, it remains suboptimal. Further efforts to improve guideline adherence are necessary. Trial registration number The present study is not registered.