Racial differences in cervical cytokine concentrations between pregnant women with and without bacterial vaginosis

2.1. Study participants 

A nested case control analysis was performed from a prospective cohort study done at Magee-Womens Hospital in Pittsburgh, USA. Inclusion criteria for the cohort study were singleton intrauterine gestation prior to 13 weeks and a self-reported race of either black or white. Exclusion criteria included vaginal bleeding, fetal anomalies, known thrombophilias, pre-gestational diabetes mellitus, chronic hypertension requiring medication, current or planned cervical cerclage, immune compromise (HIV-positive, use of systemic steroids within 6 months, use of post-transplant immunosuppressive medication) and autoimmune disease (inflammatory bowel disease, systemic lupus erythematosus, rheumatoid arthritis, scleroderma). These exclusions were developed prior to study enrollment because they are believed to be associated with preterm delivery or an alteration in the immune status, which would confound the associations we proposed to examine. All women provided demographic, medical and clinical information through standardized, closed question, interviews administered by research personnel. This study was approved by the University of Pittsburgh and Vanderbilt University Institutional Review Boards.

A total of 372 women were initially enrolled in this study. There were 208 (55.9%) women excluded from this analysis. Exclusion criteria included: absence of a BV score (1.1%), no cytokine measurements (16.1%), presence of Trichomonas vaginalis (10.2%), presence of Neisseria gonorrhoeae (0.3%), presence of Chlamydia trachomatis (4.0%), antibiotic use 3 months prior to pregnancy (18.0%) and race other than self-identified white or black (14.0%). Women with an intermediate BV score (Nugent score of 4–6, 15.3%) were also excluded because of small numbers. These exclusion criteria were chosen before data analysis because these variables could bias the cytokine measurements being examined. A total of 164 women were included for analysis. There were 28 white women with BV, 42 black women with BV, 55 white women with normal flora and 39 black women with normal flora.

2.2. Microbiologic assessment 

Two vaginal swabs were collected for culture and identification of vaginal flora. BV was diagnosed by vaginal pH≥4.7 and a score of 7 through 10 from a Gram-stained vaginal smear interpreted using the Nugent method (Nugent et al., 1991). The identification of T. vaginalis was by culture using Diamonds media, incubated at 37°C in 5% CO2 for up to 5 days. Each day, microscopic identification by direct observation of motile forms was performed. If the culture media was negative for 5 days, the results were considered negative. C. trachomatis and N. gonorrhoeae were identified using nucleic acid amplification tests and culture, respectively.

2.3. Cytokine measurements 

At a first trimester study visit (median gestation 6.5 weeks), in accordance with a standardized protocol, a pelvic examination was performed using a clean, non-lubricated speculum. Two Dacron swabs were placed in the cervix and left there for 10s to achieve saturation for the assay of cytokines. These swabs were placed in a plastic tube containing 4ml of purified bovine serum (final dilution of 1:5) and stored at −80°C until assay. The sample was thawed at room temperature, placed in a spin-X centrifuge filter unit and centrifuged at 12,000rpm for 20min. Multiplex-based platforms quantify cytokine concentrations in plasma with a high degree of agreement and correlation to ELISA, while providing improved sensitivity (Prabhakar et al., 2002, Pickering et al., 2002, Nelson et al., 2003, Biagini et al., 2004). We used the Luminex LabMAP™ and a Beadlyte® analyte kit (Upstate, Charlottesville, Virginia) designed to assay the following 28 cytokines; eotaxin (chemokine CC motif ligand 11), granulocyte–macrophage colony-stimulating factor (GMCSF), interferon-gamma (IFN-γ), interferon-gamma inducible protein 10 (IP10), interleukins (1α, 1β, 2, 3, 4, 5, 6, 7, 8, 10, 12 subunit p40, 12 subunit p70, 13, 15), monocyte chemotatic protein 1 (MCP1), platelet-derived growth factor (PDGF-AA and PDGF-BB), fms-related tyrosine kinase 3 (FLT3), macrophage inflammatory protein 1-alpha (MIP1a), regulated upon activation, normally T-expressed and presumably secreted (RANTES), tumor necrosis factor (TNF), vascular endothelial growth factor (VEGF), epidermal growth factor (EGF) and fibroblast growth factor 2 (FGF2) from a single aliquot of 50μl of specimen. These factors were chosen because they are early response cytokines as well as molecules important in the downstream cascade of inflammatory events. A monoclonal antibody specific for a cytokine is covalently linked to a fluorescent bead set, which captures the cytokine. A complementary biotinylated monoclonal cytokine antibody then completes the immunological sandwich and the reaction is detected with streptavidin–phycoerythrin. Each sample was frozen at −80°C after initial collection and never thawed until this study. After thawing, multiplex assays were performed. These samples were previously analyzed in duplicate using ELISA for IL-1β, IL-6 and IL-8. We found a very high degree of correlation between concentrations determined by ELISA and by multiplex assay (r2=0.95 to 0.97) for each of the cytokines. The cervical fluid collection protocol used in this study has served as the validated methodology for several large multi-center trials (Carey et al., 1993, Carey et al., 2000, Goldenberg et al., 1996). The inter-swab change in weight (pre- vs. post-collection) is ±1.2% and the inter-assay variation for cervical cytokines has been consistently <8% (Simhan et al., 2003b, Simhan et al., 2005a). Microbiologic and cytokine assays were performed in the Research Microbiology Laboratory at the Magee-Womens Research Institute by laboratory technologists blinded to any clinical or identifying data from the study subjects.

2.4. Statistical analysis 

Statistical analyses were performed using Stata® (StataCorp., 2007). Differences between women with BV and those with normal flora were determined by Fisher's exact test for several baseline characteristics, such as marital status, income, education, smoking and medical history. Cytokine distributions were not normally distributed; therefore, a Mann–Whitney U-test was used to examine differences between BV statuses within each population. A total of 56 comparisons were made and false discovery rate (FDR) (Benjamini and Hochberg, 1995) was used to correct for multiple testing. An alpha of 0.2 was used to determine the FDR threshold value.

To examine the possibility of differences between black women and white women, a Mann–Whitney U-test was used to compare black women with normal flora to white women with normal flora, and black women with BV to white women with BV. The results of these 56 comparisons were corrected also for multiple tests with FDR.

3.1. Baseline characteristics 

Among white women, those with BV were more likely to report prior episodes of BV than those with normal flora (P=0.004) (Supplemental Table 1). Among black women, cigarette use prior to pregnancy (P=0.031) and number of sexual partners prior to pregnancy (P=0.020) were greater in women with BV compared to those with normal flora. Black women were heavier (P=0.011), had a higher BMI (P=0.008), were more likely to have had a previous pregnancy (P=0.002) and report previous episodes of BV (P<0.001) than white women. Black women were less likely to smoke both prior to (P=0.001) and after pregnancy (P=0.038) than white women.

3.2. Differences in cytokine concentrations by BV status 

In whites, IP10 (medians 1807.8 and 3972.8ng/ml) and MCP1 (medians of 348.0 and 859.0ng/ml) were lower in women with BV compared to those with normal flora (Table 1A; Supplemental Table 2). In black women, IL-1α (medians of 1165.2 and 558.4ng/ml) concentrations were elevated among women with BV compared to those with normal flora (Table 1B; Supplemental Table 3).

Table 1. Significant concentration differences after correction for multiple testing between women with and without BV for whites (A) and for blacks (B)

	Cytokine (ng/ml)	Normal flora	BV	P value
	(A)
	IP10	3972.8 (5.0–30181.9)	1807.8 (146.6–25000.0)	0.001
	MCP1	859.0 (40.0–30631.9)	348.0 (48.0–22146.0)	0.006
	(B)
	IL-1α	558.4 (104.0–5554.6)	1165.2 (163.8–33876.5)	<0.001

The median cytokine concentration (range) is presented.

3.3. Cytokine concentrations by race 

White women with normal flora had higher levels of IL-1α (medians of 904.9 and 558.4ng/ml), IL-6 (3967.1 and 2162.2ng/ml), IL-10 (medians of 73.3 and 43.0ng/ml) and PDGF-BB (medians of 2362.2 and 1204.7ng/ml) compared to black women with normal flora (Table 2; Supplemental Table 4). However, these results did not hold up to correction for multiple testing. There were no significant differences between races for women with BV (Supplemental Table 5).

Table 2. Comparison of cytokine concentrations between races in women with normal flora

	Cytokine (ng/ml)	White women	Black women	P value
	IL-1α	904.9 (166.0–11098.1)	558.4 (104.0–5554.6)	0.047
	IL-6	3967.1 (62.2–25000.0)	2162.2 (5.0–14744.0)	0.010
	IL-10	73.3 (0.0–497.4)	43.0 (1.3–391.0)	0.016
	PDGF-BB	2362.2 (236.2–33201.0)	1204.7 (12.0–9651.0)	0.010

The median cytokine concentration (range) is presented.

3.4. Cytokine concentrations excluding women with C. albicans 

In order to ensure that the cytokine concentrations differences observed were not driven by infection with C. albicans, all women with this infection were excluded from the analysis (38 total women). After correcting for multiple testing, white women with BV had significantly lower concentrations of IP10 (P=0.003) compared to women with normal flora. Black women with BV had significantly higher concentrations of IL-1α (P=0.002) compared to women with normal flora. When comparing concentrations by race, white women with normal flora had higher levels of IL-10, IL-12(p40) and PDGF-BB than black women with normal flora. There were no significant differences between races in women with BV.