Table 3.
Model-Adjusted Changes Over Time in Weight, Stratified by Clinical Characteristics
| Clinical Characteristica | Mean (95% CI) Change in Weight in SWAD,b kg | Mean (95% CI) Change in Weight in STAY,b kg | Difference Between Means, SWAD-STAY (95% CI)b |
|---|---|---|---|
| Baseline age, years | |||
| Age <50, n = 595 | 2.4 (1.1–3.7) | 0.7 (−.2 to 1.5) | 1.7 (.3–3.1)* |
| Age ≥50, n = 523 | 2.3 (1.0–3.5) | −0.3 (−1.1 to .6) | 2.5(1.3–3.8)*** |
| Race/ethnicity | |||
| White/other, non-Hispanic, n = 172 | 1.8 (−.1 to 3.7) | 0.2 (−1.1 to 1.5) | 1.6 (−.6 to 3.7) |
| Black, non-Hispanic, n = 680 | 3.0 (1.8–4.2) | 0.6 (−.7 to 1.3) | 2.4 (1.1–3.7)*** |
| Hispanic, n = 266 | 1.8 (.1–3.5) | −0.6 (−2.0 to .7) | 2.4 (.7–4.1)** |
| Baseline BMI,b kg/m2 | |||
| BMI <30, n = 584 | 2.7 (1.8–3.7) | 0.4 (−.2 to 1.1) | 2.3 (1.3–3.3)**** |
| BMI ≥30, n = 507 | 2.1 (.5–3.8) | 0.4 (−.7 to 1.5) | 1.7 (.1–3.4)* |
| Baseline viral load,c copies/ml | |||
| Detectable, n = 208 | 1.0 (−1.1 to 3.0) | 0.3 (−1.3 to 2.0) | 0.6 (−1.5 to 2.7) |
| Undetectable, n = 910 | 2.7 (1.7–3.8) | 0.1 (−.5 to .8) | 2.6 (1.5–3.6)**** |
| Baseline CD4, cells/mm3 | |||
| CD4 <350, n = 153 | 1.6 (−1.1 to 4.4) | −0.1 (−1.7 to 1.5) | 1.7 (−1.1 to 4.6) |
| CD4 ≥350, n = 957 | 2.4 (1.5–3.4) | 0.3 (−.3 to .9) | 2.1 (1.1–3.1)**** |
| Baseline ART regimend | |||
| NNRTI, n = 526 | 3.1 (1.4–4.8) | 0.5 (−.4 to 1.3) | 2.7 (1.0–4.3)** |
| PI, n = 543 | 2.1 (1.0–3.2) | 0.04 (−.8 to .9) | 2.1 (.9–3.3)*** |
| INSTI type, SWAD group only | |||
| Dolutegravir, n = 97 | 1.7 (−.3 to 3.7) | … | … |
| Raltegravir, n = 85 | 1.3 (−2.0 to 4.5) | … | … |
| Elvitegravir, n = 52 | 2.7 (−.7 to 6.1) | … | … |
*P < .05, **P < .01, ***P < .001, ****P < .0001.Abbreviations: ART, antiretroviral therapy; BMI, body mass index; CI, confidence interval; INSTI, integrase strand transfer inhibitor; NNRTI, non-nucleotide reverse transcriptase inhibitor; PI, protease inhibitor; STAY, women who remained on non-INSTI ART during follow-up; SWAD, women who switched to or added an INSTI to ART.
a Clinical characteristics from the first visit were included during the follow-up period.
b From models adjusted for age (except age-stratified models), site, race/ethnicity (expect race/ethnicity-stratified models), income, smoking status, education, and baseline ART (except regimen-stratified models).
c The limits of detection for viral load assays were ≤80 copies/mL for 70 participants (8%), ≤48 copies/mL for 187 (21%), and ≤20 copies/mL for 653 (72%). Those with detectable viral loads were above the limit of detection but had <1000 copies/mL.
d Categories are mutually exclusive; women who received both an NNRTI and PI at baseline were not included in either group.