Table 3 Results of verifying the influence of glucocorticoid dose on the incidence of adverse events in multivariate logistic regression models* based on the data from the patients enrolled in the OM

Pattern/adverse event	p Value of glucocorticoid dose	Hazard ratio† for glucocorticoid intake >6 months
Pattern/adverse event	p Value of glucocorticoid dose	<5 mg/day	5–7.5 mg/day	>7.5 mg/day
“Linear” rising
Cushingoid phenotype	<0.001	–	4.6‡	10.4‡
Ecchymosis	<0.001	5.2	7.2	9.4
Leg oedema	0.023	–	1.8‡	2.8‡
Mycosis	ns
Parchment-like skin	0.037	3.2	3.7	4.8
Shortness of breath	ns
Sleep disturbance	0.003	1.9	2.1	2.6
Threshold at
<5 mg/day
Eye cataract	ns
5–7.5 mg/day
Epistaxis	0.027	–	5.8‡	2.9‡
Weight gain	0.004	–	2.4‡	2.4‡
>7.5 mg/day
Depression, listlessness	ns
Glaucoma	ns
Increase in blood pressure	ns

*Stepwise forward method of logistic regression incorporating potential impact factors age, sex, disease duration, disease severity and 20 comorbidities (hypertension, coronary disease, heart insufficiency, stroke, high cholesterol, diabetes, cancer, leukaemia/lymphoma, chronic lung disease, chronic kidney disease, chronic liver disease, chronic gastrointestinal disease, degenerative spine disease, hip or knee arthrosis, other arthrosis, psoriasis, circulatory disorder of leg arteries, thrombosis, osteoporosis, other comorbidities). †Reference group is “no glucocorticoid in past 12 months”. ‡Reference group is pooled “no glucocorticoid in past 12 months” and “glucocorticoid intake for more than 6 months less than 5 mg/day”, as the partial hazard ratio between “glucocorticoid intake for more than 6 months less than 5 mg/day” and “no glucocorticoid in past 12 months” was not significant. ns, not significant; OM, osteoporosis module.