Detection and Quantitation of Human Immunodeficiency Virus Type 1 in the Female Genital Tract

Abstract

Human immunodeficiency virus type 1 (HIV-1) was detected in the genital tracts of 59% of 225 women by RNA PCR and in 7% of the women by culture. In a comparison of two sampling methods, endocervical swabs were more sensitive than cervicovaginal lavage for HIV-1 RNA detection by PCR but not by culture and their sensitivity was independent of the concentration of HIV-1 RNA.

The measurement of human immunodeficiency virus type 1 (HIV-1) in the genital compartment has important implications for studies of transmission. Few studies have simultaneously assessed multiple measurement parameters, as well as sampling methods, and the various laboratory factors that may affect those measurements. HIV-1 has been cultured from the cervix and the vagina as cell-free and cell-associated virus using various sampling methods (1, 7, 10, 14, 17–19; M. A. Wainberg, R. Beaulieu, C. Tsoukas, and R. Thomas, Letter, AIDS 7:433–434, 1993). Positive-culture rates range from 11 to 71%. PCR methods have been used to quantitate HIV-1 RNA in the genital tract, including cervicovaginal lavage (CVL) (2, 16), vaginal lavage (6), vaginal (8, 9) and cervical S. Cu Uvin, A. M. Caliendo, S. E. Reinert, K. H. Mayer, T. P. Flanigan, and C. C. Carpenter, Letter, AIDS 12:826–827, 1998) swabs, and the endocervical cytobrush (6). The rates of detection ranged from 29 to 64%.

Clinical specimens were obtained from a cross-sectional study of 225 HIV-infected women, from five sites, enrolled in a substudy of the Women's Interagency HIV Study group (2). CVL was performed using a 10-ml syringe filled with sterile, nonbacteriostatic saline and aimed directly at and into the os. The pooled fluid was aspirated, and the entire volume was collected, stored on ice, and processed within 4 to 6 h. Unfractionated aliquots were stored at −70°C for RNA quantitation and were checked for the presence of blood, leukocytes, and semen. The remaining sample was centrifuged, the supernatant was collected and stored, and the pellet was resuspended in 1 ml of cell culture medium.

Two swabs for peripheral blood mononuclear cell culture and RNA determinations (Puritan Sterile Dacron Polyester Tip Applicators and Puritan Dacron Polyester, respectively) were inserted together into the cervical os and rotated 720°. The RNA swab was placed in a vial containing 4 M guanidine isothiocyanate solution stored at −70°C, and the culture swab was placed in a vial containing culture medium and stored at room temperature. Specimens were processed within 4 to 6 h. Peripheral blood mononuclear cell cultures were performed in accordance with standard procedures (5) with the addition of nystatin at 160 U/ml. A pilot study determined the optimal concentration of nystatin to address the problem of fungal contamination in genital tract cultures. RNA was quantitated using the Roche Amplicor HIV-1 Monitor test; specimens were processed using standard Roche procedures (12), and results were calculated in accordance with the manufacturer's recommendations, with a lower limit of detection of 400 copies/ml.

Only 7% of the women were culture positive in the genital tract; RNA was detectable in the genital tracts of 59% of the women. Both genital tract culture and the RNA assay were positive for 5% of the women, and one woman was positive by culture only. In sum, 64% of the women were positive by one or both methods. There was a significant association between the presence of microscopic blood and genital tract RNA. Of 74% of the women who were positive for both microscopic blood and RNA, 49% had CVL fluid collected first (Fisher's exact test, two tailed [FET], P = 0.003) and 51% had a swab collected first (FET, P = 0.02).

Table 1 shows the lack of agreement between positive swab and CVL samples. This lack of concordance was significant for RNA detection but not for culture, presumably because of the small number of positives. The effect of specimen collection order on the detection of HIV by either culture or RNA assay was assessed by alternating collection methods. When CVL was performed first (n = 102), the rate of swab positivity for both culture and the RNA assay increased from 2.6 to 4.1% and from 47.3 to 51.8%, respectively. Similarly, when swabs were taken first (n = 123) the rate of CVL positivity increased from 3.1 to 5.3% for culture and from 22.0 to 52.4% for the HIV-1 RNA assay. A similar quantitative effect on HIV-1 RNA levels was seen. None of the effects reached significance (FET or Wilcoxon's test).

TABLE 1.

Lack of agreement between positive swab and CVL samples

Test results	Total no. (%) positive by:
	HIV-1 cultureb	HIV-1 RNA detectionc
CVL+/swab+	1 (0.5)	31 (17.8)
CVL+/swab−	8 (3.7)	11 (6.3)
CVL−/swab+	6 (2.9)	54 (31.0)
CVL−/swab−	192 (92.8)	78 (44.8)
P valuea	0.27	<0.001

^aKappa test. 

^bn = 207. 

^cn = 174. 

The standard female genital tract sampling method has been swabbing or scraping for culture (1, 10, 18, 19; Wainberg et al., Letter) or CVL for RNA assay (2, 6, 16). Although virus could be cultured equally well from swab and CVL specimens, the ability to detect RNA and absolute levels of RNA were highest for swab specimens (Fig. 1), no other studies have made a direct comparison. The approximately 10-fold dilution factor associated with CVL may, in part, explain the lower detection rate and level of RNA for CVL. Lavage cell pellets were used for culture, and there was no difference between the ability to culture virus from CVL samples or swab samples, also suggesting that those RNA differences are due to dilution. The median difference in RNA level was approximately 0.5 log higher for swab specimens and was independent of the overall level of virus in the genital tract.

FIG. 1.

Differences between log₁₀ estimated RNA concentrations from swab collections and CVL collections plotted against the log-transformed mean of the two estimates. A positive difference indicates that the swab collection RNA copy number was higher than the CVL collection RNA copy number.

We confirmed the findings for women on antiretroviral therapy that have been shown previously for women not on antiretroviral therapy that multiple samplings are necessary to rule out the absence of infectious virus (10; A. Kovacs, P. Reichelderfer, and D. Wright for the DATRI 009 Study Group, Proc. 12th World AIDS Conf., abstr. 23488, p. 439, 1998). The level of infectious virus detected by culture is lower in the female genital tract than in the male genital tract (4). This may be, in part, a biologic explanation for the lower rate of HIV-1 transmission from females to males. No studies have simultaneously compared swab and CVL samples for HIV-1 RNA. We found that the percentage of RNA-positive samples was greater for swabs than for CVL, although some women had detectable RNA in CVL fluid but not in swab specimens. This lack of agreement between sampling methods again suggests that multiple sampling may be as important as the sampling method used. Other longitudinal studies using additional specimen collection methods have confirmed the necessity of multiple sampling (P. S. Reichelderfer, R. Coombs, D. Wright, D. Burns, and A. Kovacs, for The WHS 001 Study Group, Proc. 38th Intersci. Conf. Antimicrob. Agents Chemother. abstr. I-251, p. 443, 1998).

In summary, we have shown that it is possible to detect HIV-1 in the genital tracts of most infected women by using a combination of assays and sampling methods. In this study, RNA was best measured by endocervical swab. Regardless of the sampling method used, multiple samples appear to be needed to rule out the absence of HIV-1 in the genital tract.