Table 4. Assessing whether addition of inflammatory biomarkers to Cox regression models attenuates the association between HIV infection (stratified by HIV-1 RNA at baseline) and mortality.

	Hazard Ratio (95% Confidence Interval)
	Unadjusted for IL-6, sCD14 or D-dimer	Adjusted for IL-6, sCD14 and D-dimer	Unadjusted for IL-6, sCD14 or D-dimer	Adjusted for IL-6, sCD14 and D-dimer
	Model 1	Model 1	Model 2	Model 2
N (# deaths)	2324 (411)	2324 (411)	2324 (411)	2324 (411)
HIV uninfected	1 (ref)	1 (ref)	1 (ref)	1 (ref)
HIV+ HIV-1 RNA<500 copies/mL	1·12 (0·89-1·42)	1·20 (0·94-1·53)	1·04 (0·80-1·35)	1·09 (0·84-1·42)
HIV+ HIV-1 RNA 500-9999 copies/mL	1·55 (1·09-2·19)	1·58 (1·11-2·24)	1·34 (0·93-1·92)	1·42 (0·99-2·05)
HIV+ HIV-1 RNA ≥10000 copies/mL	2·94 (2.22-3·91)	2·29 (1·72-3·07)	2·18 (1·60-2·99)	2·00 (1·45-2·76)

Model 1 additionally adjusted for age and race-ethnicity.

Model 2 additionally adjusted for age, race-ethnicity, prevalent cardiovascular disease, diabetes, cancer, chronic obstructive pulmonary disease, hypertension, smoking, hepatitis C, BMI, total cholesterol and cholesterol lowering medication, alcohol use, cocaine use, hemoglobin, FIB-4, estimated glomerular filtration rate.

IL-6, sCD14 and D-dimer quartile thresholds were defined using quartile levels (25th, 50th, and 75th percentiles) among those who died. For, IL-6 these thresholds were 1·727, 2·91, and 5·043 pg/mL. For sCD14: 1592·84, 1883·51, and 2334·18 ug/mL. For D-dimer: 0·23, 0·39, and 0·86 ug/mL