Table 1:

Effect of adherence interruptions on inflammatory biomarkers. Both sets of models adjust for potential confounders (baseline inflammation level, use of a nevirapine versus efavirenz-based regimen, duration of viral suppression, age, gender, baseline log viral load, depression, and heavy alcohol use); the second model also adjusts for average adherence.

	Effect of treatment interruption on biomarkers of inflammation
Biomarker	Without adjustment for average adherence		With adjustment for average adherence
	Effect (95% CI)	P	Effect (95% CI)	P
sCD14	2.6 (0.7, 4.6)	0.008	4.4 (0.9, 8.0)	0.014
sCD163	4.6 (1.7, 7.6)	0.002	9.8 (5.0, 14.9)	<0.001
D-dimer	10.1 (2.6, 18.0)	0.007	2.4 (−9.4, 15.8)	0.70
HLA−DR+/CD8+	2.7 (0.3, 5.1)	0.025	0.7 (−3.2, 4.7)	0.74
IL-6	14.1 (3.5, 25.8)	0.008	5.9 (−8.5, 22.6)	0.44
K/T ratio	4.9 (1.9, 8.1)	0.002	8.3 (2.3, 14.6)	0.006