Changing antiretrovirals whilst viral load <50 copies/ml and relationship with CD4 count changes

Table 1

		Multivariate
		RH	95% CI	P
Gender	Female versus Male	1.29	1.09-1.53	0.0038
Risk group	Heterosexual versus other	0.83	0.70-0.98	0.033
Basline	Per year later	1.32	1.27-1.38	<0.0001
Nucleoside pair	Zidovudine/lamivudine	1.00	-	-
	Didanosine/stavudine	1.94	1.49-2.53	<0.0001
	Stavudine/lamivudine	1.81	1.50-2.17	<0.0001
	Tenofovir plus 1	0.87	0.68-1.10	0.24
	Abacavir plus 1	0.62	0.46-0.83	0.0012
	Other not listed	1.06	0.80-1.40	0.68
Third drug	Single PI	1.00	-	-
	Boosted PI	0.46	0.36-0.58	<0.0001
	NNRTI	0.42	0.35-0.50	<0.0001
	Triple nucleoside	0.23	0.16-).35	<0.0001
HCV serostatus	Positive versus negative/unknown	0.80	0.67-0.96	0.017

After adjustment, changing ARVs was associated with an additional annual increase in CD4 counts of 9.3/mm³ per year (95% CI 5.7-12.9/mm³) compared to not changing ARVs. The increase was similar in patients who recycled ARVs compared to those starting an ARV to which they were naive, according to type of new ARV started (nucleoside, PI, boosted-PI or NNRTI), and number of new ARVs started (0, 1 or >2, p>0.05 all). Patients starting a new ARV class had higher increases in CD4 counts compared to those who changed ARVs but did not start a new class (8.0/mm³; 95% CI 0.2-15.8/mm³, p=0.044), although there was no differences between PI-containing or NNRTI-containing classes (p=0.54).