Association of baseline steroid use with long-term rates of infection and sepsis in the REGARDS cohort

Table 3 Multivariable Cox regression model evaluating association between steroid use and first infection events

Exposure	Total N	Event N (%)	IR per 1000 py (95% CI)	Crude	Model 1	Model 2	Model 3
Exposure	Total N	Event N (%)	IR per 1000 py (95% CI)	HR (95% CI)	HR (95% CI)	HR (95% CI)	HR (95% CI)
Baseline chronic steroid use – full cohort
Steroid user	677	132 (19.5)	37.99 (32.03–45.06)	2.78 (2.33– 3.31)	2.64 (2.22–3.15)	2.60 (2.17–3.10)	2.10 (1.73–2.56)
Non-user	29,506	2,461 (8.5)	13.79 (13.26–14.35)	Ref	Ref	Ref	Ref
Baseline chronic steroid use – propensity matched^a
Steroid user	538	103 (19.2)	36.37 (29.99–44.12)	1.89 (1.38–2.60)	1.97 (1.43–2.72)	1.98 (1.44–2.73)	2.01 (1.45–2.78)
Non-user	538	60 (11.2)	19.34 (15.01–24.91)	Ref	Ref	Ref	Ref

Model 1 = adjusted for demographics (age, race, income, education, income); Model 2 = Model 1 + health behaviors (alcohol use, smoking status); Model 3 = Model 2 + chronic medical conditions (atrial fibrillation, chronic kidney disease, chronic lung disease, coronary artery disease, deep vein thrombosis, diabetes, dyslipidemia, hypertension, myocardial infarction, obesity, peripheral artery disease, stroke), Morisky Adherence Scale
IR incidence rate, PY person years
^aPropensity score includes age, race, income, education, income, alcohol use, smoking status, atrial fibrillation, chronic kidney disease, chronic lung disease, coronary artery disease, deep vein thrombosis, diabetes, dyslipidemia, hypertension, myocardial infarction, obesity, peripheral artery disease, stroke, Morisky Adherence Scale, and C-reactive protein

ISSN: 1364-8535