TABLE 1.
Comparison of specimen collection methods for genital HIV viral load (VL) detection by type of specimen and visit among HIV-positive women not using ART in Cape Town, South Africa
| Menstrual Cup (MC) | Endocervical swab (ECS) | Swab-enriched cervicovaginal lavage (eCVL) |
P-values* |
|||
|---|---|---|---|---|---|---|
| MC vs ECS | MC vs eCVL | eCVL vs ECS | ||||
| Enrollment visit (n=51) | ||||||
| Order of specimen collection | First | Second | Third | |||
| Insufficient specimen, n (%) | 4 (8%) | 3 (6%) | 0 (0%) | 1.000 | 0.125 | 0.250 |
| Among those with sufficient specimen, n (%) with quantifiable HIV VL | 28 (60%) | 18 (38%) | 21 (41%) | 0.004 | 0.004 | 1.000 |
| Among those with quantifiable HIV VL, median log 10 copies/mL of HIV RNA (IQR)a | 3.55 (3.07, 4.14) | 2.97 (2.42, 3.47) | 3.05 (2.84, 3.37) | 0.002 | <0.001 | 0.820 |
| Median log 10 copies/mL of plasma HIV VL (IQR) | 3.78 (3.16, 4.30) | |||||
| Spearman’s correlation (95% CIs) between genital and plasma HIV VL (including ND and NQ) | 0.44 (0.18, 0.64) | 0.47 (0.22, 0.66) | 0.45 (0.20, 0.64) | |||
| Month 3 visit (n=48) | ||||||
| Order of specimen collection | Third | First | Second | |||
| Insufficient specimen, n (%) | 2 (4%) | 0 (0%) | 0 (0%) | 0.500 | 0.500 | 1.000 |
| Among those with sufficient specimen, n (%) with quantifiable HIV VL | 26 (57%) | 14 (29%) | 21 (44%) | 0.001 | 0.062 | 0.016 |
| Among those with quantifiable HIV VL, median log 10 copies/mL of HIV RNA (IQR)a | 3.70 (3.23, 4.14) | 2.76 (2.50, 3.37) | 2.96 (2.77, 3.64) | 0.001 | <0.001 | 0.064 |
| Median log 10 copies/mL of plasma HIV VL (IQR) | 3.77 (2.99, 4.45) | |||||
| Spearman’s correlation (95% Cis) between genital and plasma HIV VL (including ND and NQ) | 0.62 (0.40, 0.77) | 0.46 (0.20, 0.66) | 0.64 (0.44, 0.78) | |||
| Month 6 visit (n=49) | ||||||
| Order of specimen collection | Second | Third | First | |||
| Insufficient specimen, n (%) | 6 (12%) | 0 (0%) | 0 (0%) | 0.031 | 0.031 | 1.000 |
| Among those with sufficient specimen, n (%) with quantifiable HIV VL | 19 (44%) | 16 (33%) | 23 (47%) | 0.146 | 1.000 | 0.065 |
| Among those with quantifiable HIV VL, median log 10 copies/mL of HIV RNA (IQR)a | 3.71 (3.29, 4.36) | 2.99 (2.45, 3.31) | 3.09 (2.71, 3.76) | 0.005 | 0.002 | 0.140 |
| Median log 10 copies/mL of plasma HIV VL (IQR) | 3.74 (3.23, 4.04) | |||||
| Spearman’s correlation (95% Cis) between genital and plasma HIV VL (including ND and NQ) | 0.52 (0.26, 0.71) | 0.48 (0.23, 0.67) | 0.59 (0.37, 0.74) | |||
IQR, interquartile range; ND, non-detectable; NQ, non-quantifiable
*
Matched analyses using McNemar’s exact for proportion and Wilcoxon rank-sum for medians.
a
Medians are presented for all of that type of specimen with quantifiable VL; P-values are computed for those in which both types of specimens had quantifiable VL.