The association between homocysteine and bacterial vaginosis: results from NHANES 2001–2004

Data source

We used two phases, 2001–2002 and 2003–2004, from The National Health and Nutrition Examination Survey (NHANES) for this study. The health and nutrition status of diverse American populations is assessed through the NHANES, a cross-sectional study. Among them, demographic, socioeconomic, educational, age, dietary, and health-related issues were obtained using questionnaire forms, whereas most of the physical examinations, including laboratory results such as HCY levels in this study, were obtained at the Mobile Inspection Center (MEC). Written informed consent was obtained from all participating individuals in this study, which was approved by the research ethics review board of the NCHS (https://wwwn.cdc.gov/nchs/nhanes/default.aspx). We included a total of 21,161 data points, 16,360 after deleting 4801 data points without specific Hcy values, 2557 after deleting 13,803 data points with unknown BVs, and 2398 after excluding 159 data points with unknown covariates.

Measurement of homocysteine

The Abbott HCY assay was used on the Abbott AxSYM analyzer to measure total HCY (THCY) in plasma using the Fluorescence Polarization Immunoassay (FPIA) from Abbott Diagnostics. Dithiothreitol (DTT) with albumin and other small molecules for free thiols, in contrast to the FPIA detection system, which is made up of certain monoclonal antibodies and fluoresceinated SAH analog tracers, the injection of S-adenosylhomocysteine (SAH) hydrolase catalyzes the conversion of HCY to SAH in the presence of adenosine. THCY was calculated by the Abbott AxSYM immune analyzer using a machine-stored calibration curve. Since the FPIA method is equivalent to others, it was used as the primary method for THCY determination in nhanes 2003–2004. HHCY was defined as plasma total HCY > 15 μmol/L.

Measurement of bacterial vaginosis

Vaginal swabs were self-collected in a private bathroom at the Mobile Inspection Center (MEC) after participant consent. NHANES personnel collected samples onto slides for Gram staining, after which slides were scanned at low magnification to locate epithelial cell populations. The average number of Lactobacillus morphotypes, Gardnerella spp., anaerobic gram-negative rods, and Mobil morphotypes was quantified. The BV fraction was measured by Nugent’s method. There are three categories of BV outcome, with Nugent’s scores of 0–3 suggestive of normal vaginal flora, 4–6 suggestive of intermediate, and 7–10 suggestive of BV positivity.

Covariates

These factors may have influenced the results, such as demographic variables including age, race, education level, body mass index (BMI), and other laboratory indicators. Serum vitamin B12, ferritin, percentage of segmented centrioles, and number of segmented centrioles were selected as potential covariates in our study. Participant age was obtained by questionnaire according to certain standards. Race/ethnicity was categorized as non-Hispanic white, non-Hispanic black, Mexican American, other Hispanic, and other races. Educational attainment was categorized into three categories: less than high school, high school, college and above. BMI was determined according to the NHANES III anthropometric procedures standards with the correct technique for participant height and weight, obtained as weight in kilograms divided by height in meters (low BMI, underweight and healthy weight ≤ 24.9, high BMI, overweight and obese > 25). The Quantaphase II Folate/Vitamin B12 radiometric test kit from Bio-Rad laboratory was used to obtain serum vitamin B12. To inactivate endogenous folate binding proteins, serum or whole blood hemolytic samples were combined with 125 folic acid and 57 vitamin B12 in a solution containing dithiols (DTT) and cyanide. The combination was then heated, converting various forms of vitamin B12 to cyanocobalamin. Binding to the immobilized affinity purified porcine intrinsic factor and folate binding protein after waiting for the mixture to cool. The addition of these substances allows the pH of the reaction mixture to be adjusted and buffered to 9.2. The reaction mixture was then incubated at room temperature for one hour. The reaction mixture was centrifuged, decanted and labeled, and unlabeled vitamin B12 was precipitated at the bottom of the tube after binding to immobilized binding proteins. Supernatants containing unbound vitamin B12 were discarded, and the radioactivity associated with the pellets was measured. A standard curve (created from a precalibrated folate/B12 standard in a human serum albumin base) was used to determine the amount of vitamin B12 in each participant’s serum. Using a Roche/Hitachi 912 clinical analyzer, ferritin was quantified by immunoturbidimetry. Antigen/antibody complexes are produced when latex-bound ferritin antibodies interact with antigens in the sample. Measurements were made turbidimetrically after agglutination. Measured at 700 nm (primary wavelength), the complexes formed were directly proportional to ferritin concentration. Segmented neutrophil percentages and segmented neutrophil numbers were obtained using a Beckman Coulter maximum instrument in flow check centers (MECs), performing complete blood counts on blood specimens and providing blood cell distribution and differential analysis of white blood cells using VCS technology. These covariates are all available in NHANES with relevant access methods and laboratory test data.

Statistical analysis

All correlation analyses were performed using the statistical software R (http://www.R-project.org , R Foundation) and using freeware version 4.0 (licensed). We used the weights recommended by the NHANES database, taking into account the oversampling of minority groups, and all included data were analyzed after weighting to achieve as unbiased and accurate effect estimates as possible. In the study population description baseline tables, continuous variables are presented using survey weighted means (95% CI), and P values are measured using survey weighted linear regression; Categorical variables used survey weighted percentages (95% CI), and P values were measured using survey weighted chi square tests. The relationship between HCY and BV was assessed by input type linear regression models. The standardized β values were used to compare the relative predictive strength of different covariates in the regression model, and the variance inflation factor (VIF) was used to assess multicollinearity for all covariates in the regression model. Covariates were included in the final adjusted model as potential confounders if they changed the estimate of HCY with BV by more than 10% or had a clear association with BV. The log likelihood ratio test was similarly used in threshold effect analysis to assess the linearity of the HCY and BV models. The linear relationship of HCY and BV was further explored using smooth curve fitting. To investigate the quantitative relationship between HCY levels with BV, we performed a linear regression relationship and constructed two adjusted models according to covariates. Hierarchical regression analysis was used to account for differences between age, race, education, and BMI. A two-tailed P < 0.05 was considered statistically significant.

Ethics statement

The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013). All information from the NHANES program is available and free for public, so the agreement of the medical ethics committee board was not necessary.