. Author manuscript; available in PMC: 2012 Apr 26.

Published in final edited form as: J Int Assoc Physicians AIDS Care (Chic). 2011 Apr 20;10(3):160–170. doi: 10.1177/1545109711402213

Table 3.

Adjusted Analyses for All-Cause Mortality

Parameters		Adjusted Analysis^a

		HR	95% CI	P
Latest CD4 count per mm³ (lagged by 3-months)	>500	Reference
	350–500	1.36	0.83–2.19	.212
	200–350	2.44	1.58– 3.77	<.001
	100–200	3.81	2.30–6.31	<.001
	50–100	4.11	1.94–8.69	<.001
	<50	21.56	13.12–35.45	<.001
Cohort	AHOD	Reference
	TAHOD	0.48	0.18–1.32	.157
Interaction term^b				.254^b

Abbreviations: HR, Hazard ratio; TAHOD, TREAT Asia HIV Observational Database; AHOD, Australian HIV Observational Database; CD4 count refers to CD4 T-cell count.

Models were a priori adjusted for fixed covariates, including cohort (TAHOD/AHOD), gender, hepatitis B and C coinfection, HIV exposure category, and prior AIDS; and time-updated covariates including HIV RNA category lagged by 3 months (categorized as <500, 500–10,000, and >10,000 copies/mL), age, on or off cART, calendar year (categorized as earlier than 2000, 2000–02, 2002–04, 2004–06, 2006–08, and later than 2008) and new AIDS (for mortality endpoint only).

P value is for the interaction term (product of cohort (TAHOD/AHOD) and categorized CD4 count lagged by 3-months), obtained in the adjusted model predicting respective endpoint. Excluding the interaction term from the multivariable model changes the HR for cohort (TAHOD vs. AHOD) to 1.43 (95% CI: 1.02–2.00) for AIDS or death and 0.84 (95% CI: 0.53–1.33) for all-cause mortality.