Table 3.

Alternative models for evaluating the interaction term between female sex and HLA-A*03

Six alternative modelsa	n	Impact on longitudinal VLb
		Δβ ± SE	Adjusted P
a) Removing Kenyan subjects
Female sex	180	−0.29 ± 0.08	<0.001
Female sex × HLA-A*03	21	−0.79 ± 0.25	0.002
b) Removing Ugandan subjects
Female sex	136	−0.31 ± 0.09	<0.001
Female sex × HLA-A*03	13	−0.39 ± 0.28	0.175
c) Removing Zambian subjectsc
Female sex	107	−0.22 ± 0.10	0.039
Female sex × HLA-A*03	15	−0.90 ± 0.32	0.005
d) Zambian subjects onlyd
Female sex	87	−0.34 ± 0.10	0.001
Female sex × HLA-A*03	7	−0.30 ± 0.35	0.378
e) Removing rare HIV-1 subtypese
Female sex	132	−0.27 ± 0.08	0.001
Female sex × HLA-A*03	13	−0.76 ± 0.28	0.006
f) Removing HIV-1 subtype A1
Female sex	142	−0.33 ± 0.09	0.0001
Female sex × HLA-A*03	15	−0.53 ± 0.29	0.064

^aPart of the sensitivity analyses

^bRepeated measurements in the 2–36 months interval, with log₁₀-transformation before analysis. The summary statistics are adjustment for other factors shown in Table 2. β regression beta (mean deviation, Δ, from the reference group), SE standard error of the mean (Δ)

^cThe remaining subjects correspond to eastern Africans

^dCorresponding to southern Africans

^eDefined as others (not A1 and not C) in Table 1