Table 3.

Univariate and multivariate logistic modeling of three potential outcomes for emergence of resistance among patients failing antiretroviral therapy based on clinical or immunological criteria in Malawi.

Outcome of interest	At least three TAMs		K65R or K70 E tenofovir-associated mutations		Pan-nucleoside resistance Q151M and tenofovir mutations or 69 insertion
Outcome of interest	Univariate	Multivariate^a	Univariate	Multivariate^a	Univariate	Multivariate^a
CD4 cell count (cells/μl)
>100	1.0	1.0	1.0	1.0	1.0	1.0
100 or less	3.9 (1.2–12.6)	5.57 (1.41–22.02)	5.4 (1.5–19.7)	6.1 (1.47–25.0)	12.2 (1.5–97.0)	9.9 (1.2–84.0)
ZDV use
No	1.0	1.0	1.0	1.0	1.0	1.0
Yes	2.3 (0.85–6.05)	3.4 (1.07–10.97)	0.18 (0.04–0.81)	0.18 (0.04–0.94)	0.12 (0.015–0.93)	0.13 (0.02–1.07)

Data are given as odds ratio (95% confidence interval). Outcomes include having at least three TAMs, the development of tenofovir-associated mutations (K65R or K70E) or expected pan-nucleoside resistance mutation (Q151M with K65R or K70E mutations or the 69 insertion). ART, antiretroviral therapy; TAM, thymidine analog mutations; ZDV, zidovudine.

Adjusted for HIV-1 RNA (copies/ml), clinical failure, clinic site, and sex. Duration of ART was not considered because of colinearity with the use of zidovudine.