Antioxidants Associated With Oncogenic Human Papillomavirus Infection in Women

Abstract

Background

Human papillomavirus (HPV) infection is a major cause of cervical cancer. Studies showed the onset of HPV carcinogenesis may be induced by oxidative stress affecting the host immune system. The association between antioxidants and oncogenic HPV remains unclear. In this study, we aim to identify antioxidants associated with vaginal HPV infection in women.

Methods

The associations between the 15 antioxidants and vaginal HPV infection status (no, low-risk [LR], and high-risk [HR] HPV) were evaluated using 11 070 women who participated in the 2003–2016 National Health and Nutrition Examination Survey (NHANES).

Results

We identified serum albumin and 4 dietary antioxidants (vitamin A, B2, E, and folate) inversely associated with HR-HPV infection. Women with a low level of albumin (≤39 g/L) have a significantly higher risk of HR-HPV (odds ratio [OR] = 1.4, P = .009 vs >44 g/L). A Nutritional Antioxidant Score (NAS) was developed based on these 4 dietary antioxidants. The women with the lowest quartile NAS had a higher chance of HR-HPV (OR = 1.3, P = .030) and LR-HPV (OR = 1.4, P = .002) compared with the women with the highest quartile NAS.

Conclusions

We identified 5 antioxidants negatively associated with vaginal HR-HPV infection in women. Our findings provide valuable insights into understanding antioxidants’ impact on HPV carcinogenesis.

Serum albumin and 4 dietary antioxidants (vitamin A, B2, E, and folate) were inversely associated with high-risk HPV vaginal infection. The nutritional antioxidant score based on these 4 dietary antioxidants was also inversely associated with high-risk and low-risk HPV infection.

Human papillomavirus (HPV) is a well known risk factor for cervical cancer, which is the fourth most common female cancer and contributes to 7.5% of cancer deaths for women worldwide in 2018 [1]. Certain HPV strains are more likely to incite precursor events leading to carcinogenesis. These strains are called oncogenic or high-risk [HR] HPV strains. The remaining strains of HPV are classified as non-oncogenic (or low risk [LR]). Almost all cervical cancers (99.7%) are directly linked to the previous infection with 1 or more HR-HPV infections [2]. Many women in the United States have been exposed to HR-HPV. The prevalence of genital HR-HPV is 20.4%, and the prevalence of any genital HPV infection is 39.9% among US women aged 18–59 [3]. Although most HR-HPV infections will clear without treatment, some will persist. Studies showed that approximately 43% of women without high-grade cervical lesions have a 12-month persistent infection of HR-HPVs [4, 5]. Currently, there is no effective antiviral therapy to clear genital HPV infection. Thus, it is important to identify modifiable factors, such as antioxidants, associated with HR-HPV infection to prevent HPV carcinogenesis onset.

In the natural history of HPV infection, carcinogenesis is a rare event and is usually triggered by several cofactors that induce HPV infection into tumor transformation [6]. In addition to the HPV’s viral characteristics, host factors also play an essential role in carcinogenesis. These host risk factors are often classified into either behavioral or biological categories [7]. Behavioral risk factors (such as sexual behaviors) primarily affect HPV acquisition, and biological risk factors primarily impact downstream transitions in the oncogenesis pathway [8]. For biological risk factors for HPV infection, intrinsic host factors, including antioxidants, affecting immune response play an important role in HPV infection in addition to characteristics of the HPV virus itself.

Oxidative stress (OS) has been found to affect the immunity system and is a promising cofactor in HPV-related carcinogenesis. Oxidative stress or excessive production of reactive oxygen species (ROS) is an imbalance of oxidants and antioxidants, usually with a surplus of oxidants. Oxidative stress can change various cell components and cellular signaling mechanisms [9], creating an environment that promotes cancer initiation and progression that may interact with HPV infection [10]. In addition, OS is a suggested cofactor to work synergistically or independently to HR-HPV infection. Several biochemical studies and clinical epidemiology evidence support the hypothesis that OS may impact HPV acquisition and persistence [6]. Another study showed that HPV-positive cervical lesions had a higher level of oxidative deoxyribonucleic acid (DNA) damage than the HPV-negative with cervical cytology normal control samples [11].

The direct measure of ROS is a complex task. It is challenging to reach high accuracy and precision because of their short lifespan and rapid reactivity with redox state regulating components [12, 13]. The imbalance between ROS that causes OS is due to limitations of antioxidants available and exposure to harmful agents [12]. Therefore, assessment of antioxidant status is the most common method to evaluate OS’s status [12, 13]. The antioxidant defense system includes endogenous antioxidants and dietary antioxidants, which work together to retain redox homeostasis [14] interactively. Previous studies have generally reported that the consumption of specific nutrients may be more effective in reducing OS, and nutrient deficits could link to cervical cancer development [14]. In addition, women who have a healthy dietary pattern such as high fruit and vegetable consumption have been observed to be associated with a lower risk of HPV infection and cervical cancer [15]. High risk-HPV clearance was associated with high fruit and seafood/plant protein intake for women with low-grade cervical cytology and/or positive HR-HPV test based on a prospective study [16]. However, the associations between specific antioxidants and genital HPV infection are still understudied.

This study aims to assess levels of 15 antioxidant measures (4 endogenous antioxidants and 11 dietary antioxidants) associated with vaginal HPV infection, especially HR-HPV infection. In this study, we used vaginal HPV infection collected as a proxy for cervical HPV infection. Studies showed good concordance of HPV DNA detection between clinician-collected cervical samples and self-collected vaginal samples. The concordance of HPV16/18 and any HR-HPV detection using vaginal samples compared with cervical samples was 91%–94% and 73–89%, respectively, based on 3 different assays [17]. Another study tested 6 assays and showed (1) an almost perfect agreement of HPV16/18 between self-collected vaginal samples and clinician-collected cervical samples on all assays (94%–99%) and (2) good agreement for non-HPV16/18 oncogenic HPV types (64%–73%) [18].

METHODS

Study Population

This study used the population-based National Health and Nutrition Examination Survey (NHANES, n = 11 070), a US national representation, and cross-sectional survey. Our study population’s inclusion criteria are women aged 18–59 years, with valid information of vaginal HPV infection and the selected antioxidants. Those with cancer and any dose of HPV vaccination were excluded from this study. In NHANES, the eligible age for detecting vaginal HPV infection is 18–59 years. We applied the 14-year NHANES data (years 2003–2016) by integrating data from the seven 2-year cycles for this study. This study is based on a secondary data analysis using open-access NHANES data, so informed consent is not applicable.

Detection and Classification of Human Papillomavirus Infections

The measures of HPV infection were based on HPV genotyping using DNA extracted from self-collected vaginal swabs. The DNA extracts used for the Linear Array HPV test are stored at −20°C for temporary storage and a −80°C freezer for long-term storage. The Roche Linear Array HPV Genotyping test was applied to test the 37 HPV genotypes: 6, 11, 16, 18, 26, 31, 33, 35, 39, 40, 42, 45, 51, 52, 53, 54, 55, 56, 58, 59, 61, 62, 64, 66, 67, 68, 69, 70, 71, 72, 73, 81, 82, 83, 84, IS39, and CP6108. The HR-HPV or oncogenic HPV types, which have a high risk of causing cancer, was defined as the established 12 HPV types (16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, and 59) based on the International Agency for Research on Cancer [19, 20]. The other 25 HPV types are considered low-risk types. This assay is based on HPV L1 consensus polymerase chain reaction with biotinylated PGMY09/11 primer sets. The details of detecting the HPV measures are included in the NHANES website [21].

Selection of Antioxidants

This study tested 15 antioxidants: 4 endogenous antioxidants collected from serum samples and 11 dietary antioxidants based on the self-reported dietary intake questionnaires. These 4 endogenous antioxidants were bilirubin, ferritin, albumin, and uric acid. Detailed procedures and analytic methods are described in the NHANES Description of Laboratory Methodology section. For the dietary antioxidants, the two 24-hour dietary intake reports were applied. The 11 dietary antioxidants included vitamin A, B2, C, D, and E, alpha-carotene, selenium, lycopene, lutein + zeaxanthin, beta-cryptoxanthin, and folate.

Statistical Analyses

The participant’s demographic factors, health-related behaviors, other potential risk factors for HPV infection, and antioxidants by the vaginal HPV status were summarized using descriptive statistics. The outcome is the grouped vaginal HPV status (ie, no HPV, LR-HPV, and HR-HPV). The HR-HPV group was defined as any occurrence of the 12 HR-HPV types. The LR-HPV group was defined as no HR-HPV infection and infection with at least 1 of the 25 LR-HPV types. The selected demographic factors (such as age, race, and ratio of family income to poverty) and behavioral factors (such as cigarette smoking, alcohol intake, lifetime illegal substance use, and the number of sexual partners in the past year) are shown in Table 1.

Table 1.

Demographic and Behavioral Factors Associated With Vaginal HPV Status in Women, NHANES Year 2003–2016 (n = 11 070)

		Vaginal HPV, wt%a
Variable	Total (wt%)a	No	Low Risk	High Risk	P Valueb
Total	11 070 (100)	59.7	21.2	19.1	-
Age
18–26	2510 (17.6)	50.6	18.4	31.1	<.001
27–35	2343 (20.8)	57.8	20.6	21.7
36–45	2845 (27.2)	60.3	23.2	16.5
46–59	3372 (34.5)	64.9	21.5	13.6
Race
Non-Hispanic white	4180 (63.3)	63.2	19.2	17.6	<.001
Non-Hispanic black	2589 (13.5)	41.1	32.0	26.9
Mexican American and other Hispanic	3241 (15.9)	57.6	22.6	19.8
Others	1060 (7.2)	67.2	16.3	16.6
Education
<High school	791 (4.3)	59.1	22.4	18.5	<.001
High school	3611 (31.3)	53.0	24.4	22.7
>High school	5793 (64.4)	63.2	19.8	17.0
Marital Status
Married/live with partner	6273 (64.0)	67.6	18.5	13.9	<.001
Never married	2593 (19.5)	48.4	20.8	30.8
Widowed/divorced/separated	1830 (16.5)	42.8	32.3	24.9
Family Monthly Poverty Level Indexc
≤1	2689 (17.2)	47.0	24.9	28.2	<.001
1.1–2	2529 (19.5)	53.4	25.3	21.3
2.1–4	2550 (28.0)	62.6	19.4	18.0
>4	2530 (35.3)	67.8	18.2	14.0
Obesity Status (Body Mass Index)
Underweight or normal (<25)	4179 (41.6)	59.8	19.5	20.7	.005
Overweight (25–29.9)	2774 (26.3)	60.3	21.7	18.0
Obese (≥30)	3671 (32.2)	59.3	22.9	17.9
Smokingd
Never	6703 (61.5)	65.0	18.2	16.7	<.001
Former	1487 (16.9)	61.7	22.6	15.7
Current	2147 (21.6)	43.5	29.1	27.4
Number of Days Use Alcohol Past Year
0	3366 (30.3)	67.5	18.5	14.0	<.001
1–25	3035 (35.8)	58.8	21.7	19.5
>25	2417 (33.9)	55.4	22.5	22.2
Lifetime Illegal Substance Use
No	5327 (51.4)	66.2	18.2	15.6	<.001
Yes	4034 (48.6)	53.3	24.7	22.0
Number of Sex Partners Past Year
0	1546 (15.9)	65.7	21.5	12.8	<.001
1	6422 (72.7)	63.5	20.1	16.4
≥2	1144 (11.5)	27.5	29.7	42.9

Abbreviations: HPV, human papillomavirus; NHANES, National Health and Nutrition Examination Survey.

^aWeighted% based on the NHANES sampling weights.

^bThe Rao-Scott χ ² test.

^cA ratio of monthly family income to the US Federal Department of Health and Human Services poverty guidelines specific to family size.

^dNever smokers are defined as those who smoked less than 100 cigarettes in life, and former smokers are defined as those who smoked at least 100 cigarettes in life but did not smoke currently.

For the dietary antioxidants, the mean of the two 24-hour dietary intake results was used. If dietary intake information was collected in only 1 day, we applied the 1-day measures. Due to the skewness, all markers were natural log-transformed for further analyses. For evaluating associations between categorical factors and the 3-group HPV status, we applied the Rao-Scott χ ² test. The differences in biomarker distributions between the 3 HPV infection statuses were compared using univariate multinomial logistic model.

We applied multinomial logistic models to evaluate antioxidants associated with vaginal HPV infection status adjusting for potential confounding factors because the outcome variable had 3 categories (no HPV, LR-HPV, and HR-HPV). The 9 factors used for adjusting in the multivariable models were age, race, education, income, marital status, smoking status, lifetime illegal substance use, number of days with alcohol intake in the past year, and number of sexual partners in the past year. For income, the family monthly poverty level index was defined as a ratio of monthly family income to the US Federal Department of Health and Human Services poverty guidelines specific to family size. For smoking status, never smokers were defined as those who smoked less than 100 cigarettes in life, and former smokers were defined as those who smoked at least 100 cigarettes in life but did not smoke currently. The adjusted odds ratios (ORs) and the corresponding 95% confidence intervals (CIs) of these factors were calculated. For the significant antioxidants, we built the Nutritional Antioxidant Score (NAS) using the principal component analysis. The principal component analysis can be used to reduce data dimension and make results easy for interpretation by reducing the numbers of variables, create new uncorrelated variables (or principal components), and minimize information loss [22]. In this principal component analysis, factors were rotated by orthogonal transformation; thus, the components were uncorrelated. The number of components was selected based on the components, which could explain a large proportion of data variance (eigenvalues > 1). The statistically significant level is 0.05. All analyses were weighted to account for the complex sampling design applied in NHANES using SAS 9.4 (SAS Institute, Inc., Cary, NC). The weighting analytical procedures using SAS followed the suggested instructions [23].

RESULTS

A total of 11 070 eligible women in the NHANES 2003–2016 cohort were used for this study. The overall prevalence of HR-HPV was 19.1%, and the prevalence of only LR-HPV (without any HR HPV) was 21.2%. The demographics and behavioral factors by the 3 vaginal HPV statuses were shown in Table 1. For the overall study population, 34.5% were aged 46–59 years, and 63.3% were non-Hispanic white. Furthermore, 64.4% of the population had an above high school degree, 35.3% of the population reported high income with the ratio of family income to poverty >4, and 64% of the population were married or lived with a partner. Regarding health-related behaviors, 72.7% of people reported they had 1 sex partner in the past year. More than half of the study population were never smokers (61.5%) and did not use illegal substances in their lifetime (51.4%). Approximately 70% of people used alcohol at least 1 day in the past year. The vaginal HP-HPV rates were higher for women between the ages of 18 and 26, non-Hispanic Black, low-income, never married, received high school education, and are either underweight or normal weight. Although behaviorally, the highest vaginal HR-HPV rates were found in women who had more than 1 sex partner in the past year (42.9%), are current smokers (27.4%), used illegal substances in their lifetime (22.0%), and consumed alcohol over 25 days in the past year (22.2%).

The majority of the antioxidants had an inverse association between levels of antioxidants and vaginal HR-HPV infection without controlling for other factors, except ferritin, uric acid, and vitamin D (Table 2). After adjusting for the 9 potential confounding factors listed under Methods, we tested 1 antioxidant at a time in the weighted logistic regressions. As shown in Table 3, 5 antioxidants showed significant negative associations with HR-HPV status. The 5 antioxidants significantly associated with HR-HPV infection after adjusting for the selected factors are albumin (OR of HR-HPV = 0.29 for natural log-transformed albumin, P = .011), vitamin A (OR = 0.89, P = .023), vitamin B2 (OR of HR-HPV = 0.84, P = .024), vitamin E (OR of HR-HPV = 0.83, P = .021), and folate (OR of HR-HPV = 0.79, P = .003). Among them, vitamin E and folate also showed a significant negative effect on LR HPV. Our results demonstrated the negative effects of albumin, vitamin A, vitamin B2, vitamin E, and folate on HR-HPV infection with and without adjusting for the confounding factors.

Table 2.

Levels of Antioxidants by Vaginal HPV Infection Status

Biomarker (Unit)	Total Wt Mean ± SEa (Sample Size)	No HPV Wt Mean ± SEa (Sample Size)	Low-Risk HPV Wt Mean ± SEa (Sample Size)	High-Risk HPV Wt Mean ± SEa (Sample Size)	P Valueb
Endogenous Antioxidants
Bilirubin (µmol/L)	10.72 ± 0.08 (10 524)	10.93 ± 0.10 (5914)	10.33 ± 0.12 (2402)	10.49 ± 0.12 (2208)	.025
Ferritin (µg/L)c	56.90 ± 0.94 (6088)	57.10 ± 1.32 (3304)	58.41 ± 1.75 (1375)	54.82 ± 1.71 (1409)	.698
Albumin (g/L)	41.94 ± 0.06 (10 533)	42.12 ± 0.07 (5918)	41.72 ± 0.09 (2406)	41.58 ± 0.11 (2209)	<.001
Uric acid (µmol/L)	273.95 ± 0.87 (10 531)	272.85 ± 1.14 (5918)	276.47 ± 1.74 (2405)	274.63 ± 1.49 (2208)	.243
Dietary antioxidants	(n = 10 543) d	(n = 5892) d	(n = 2409) d	(n = 2242) d
Vitamin A, RAE (mcg)	589.74 ± 9.40	614.42 ± 11.64	574.79 ± 17.46	530.23 ± 13.43	<.001
Vitamin B2 (mg)	1.89 ± 0.02	1.93 ± 0.02	1.86 ± 0.03	1.84 ± 0.03	<.001
Vitamin C (mg)	78.44 ± 1.16	80.77 ± 1.44	73.94 ± 1.99	76.29 ± 2.33	<.001
Vitamin D (mcg)	4.02 ± 0.05 (7509)	4.11 ± 0.07 (4315)	3.83 ± 0.13 (1720)	3.93 ± 0.11 (1474)	.099
Vitamin E (mg)	7.48 ± 0.10	7.76 ± 0.11	7.17 ± 0.18	6.95 ± 0.16	<.001
Alpha-carotene (mcg)	415.96 ± 20.96	451.14 ± 29.17	398.98 ± 29.08	326.22 ± 27.30	<.001
Selenium (mcg)	97.29 ± 0.64	98.26 ± 0.87	95.99 ± 1.29	95.77 ± 1.22	.023
Lycopene (mcg)	4885.16 ± 104.44	4979.59 ± 132.45	5019.45 ± 242.44	4442.64 ± 186.86	.037
Lutein + zeaxanthin (mcg)	1566.30 ± 39.80	1631.99 ± 51.06	1511.59 ± 79.11	1424.55 ± 93.74	<.001
Beta-cryptoxanthin (mcg)	92.05 ± 2.93	96.30 ± 4.06	82.90 ± 4.23	89.16 ± 5.71	<.001
Folate, DFE (mcg)	479.46 ± 4.87	494.24 ± 5.84	462.04 ± 9.84	453.28 ± 8.09	<.001

Abbreviations: DFE, dietary folate equivalents; HPV, human papillomavirus; RAE, retinol activity equivalents; SE, standard error.

^aWeighted mean ± SE.

^bCompared natural log-transformed measures among the 3 HPV groups using univariate weighted multinomial logistic model.

^cMissing 2011–2012 and 2013–2014 cycle data.

^dSample size for dietary antioxidants, except vitamin D.

Table 3.

Antioxidants Associated With Vaginal HPV Infection Status

	Unadjusted		Adjusteda
Biomarker (Unit)	Low-Risk HPV vs No HPV OR (95% CI)	High-Risk HPV vs No HPV OR (95% CI)	Low-Risk HPV vs No HPV OR (95% CI)	High-Risk HPV vs No HPV OR (95% CI)	Sample Size
Endogenous Antioxidants b
Bilirubin (µmol/L)	0.79 (0.67–0.94)**	0.84 (0.70–1.00)*	0.90 (0.79–1.02)	0.95 (0.81–1.11)	7400
Ferritin (µg/L)	1.04 (0.97–1.11)	1.01 (0.93–1.09)	1.04 (0.94–1.15)	1.06 (0.95–1.17)	4188
Albumin (g/L)	0.25 (0.13–0.49)***	0.16 (0.07–0.34)***	0.68 (0.29–1.63)	0.29 (0.11–0.76)*	7405
Uric acid (µmol/L)	1.24 (0.94–1.63)	1.15 (0.93–1.42)	1.33 (0.92–1.92)	1.22 (0.92–1.61)	7404
Dietary Antioxidants b
Vitamin A, RAE (mcg)	0.84 (0.76–0.93)**	0.75 (0.68–0.82)***	0.94 (0.84–1.05)	0.89 (0.8–0.98)*	7548
Vitamin B2 (mg)	0.77 (0.67–0.89)**	0.72 (0.63–0.83)***	0.88 (0.75–1.04)	0.84 (0.72–0.98)*	7548
Vitamin C (mg)	0.87 (0.82–0.92)***	0.85 (0.80–0.91)***	0.93 (0.87–1.00)*	0.94 (0.88–1.02)	7548
Vitamin D (mcg)	0.93 (0.87–1.00)*	0.96 (0.90–1.02)	0.99 (0.90–1.07)	1.03 (0.95–1.07)	5512
Vitamin E (mg)	0.77 (0.68–0.86)***	0.7 (0.62–0.79)***	0.84 (0.70–1.00)*	0.83 (0.7–0.97)*	7548
Alpha-carotene (mcg)	0.98 (0.96–1.00)*	0.94 (0.92–0.96)***	1.01 (0.98–1.03)	0.98 (0.96–1.01)	7548
Selenium (mcg)	0.87 (0.76–0.99)*	0.83 (0.71–0.96)*	0.91 (0.76–1.09)	0.86 (0.69–1.06)	7548
Lycopene (mcg)	0.99 (0.98–1.00)	0.99 (0.97–1.00)*	1.00 (0.98–1.01)	0.99 (0.97–1.00)	7548
Lutein + zeaxanthin (mcg)	0.89 (0.84–0.95)**	0.82 (0.77–0.88)***	0.97 (0.90–1.04)	0.93 (0.85–1.01)	7548
Beta-cryptoxanthin (mcg)	0.97 (0.94–1.00)*	0.94 (0.91–0.96)***	1.01 (0.97–1.04)	0.97 (0.93–1.01)	7548
Folate, DFE (mcg)	0.73 (0.64–0.84)***	0.72 (0.63–0.82)***	0.81 (0.69–0.95)**	0.79 (0.69–0.92)**	7548

Abbreviations: CI, confidence interval; DFE, dietary folate equivalents; HPV, human papillomavirus; OR, odds ratio; RAE, retinol activity equivalents.

*P < .05, **P < .01, and ***P < .001 based on multinomial logistic models; bold: adjusted P < .05 for high-risk HPV.

^aMultinomial logistic model adjusted for 9 factors (age, race, education, income, marital status, smoking status, lifetime illegal substance use, past 12-month alcohol intake, number of sexual partners in past 12 months).

^bBased on log-transformed measures.

For building an integrated score of the significant antioxidants, we used the principal component analysis based on the 5 significant antioxidants (albumin, vitamin A, vitamin B2, vitamin E, and folate). As shown in Supplement Figure 1, the first 2 principal components had an eigenvalue of >1, which explained 75% of the total variance among these 5 antioxidants. The first component had similar factor loadings for the 4 dietary antioxidants (vitamin A, vitamin B2, vitamin E, and folate) except a small factor loading for albumin (−0.015), and the primary factor loading for the second component was for albumin (loading = 0.994). This indicated that albumin provided an independent contribution to HPV infection compared with the 4 dietary antioxidants. As shown in Supplement Figure 2, the NAS’s factor loadings were in a range of 0.466–0.528, and this NAS explained 69% of the total variance among these 4 dietary antioxidants. Thus, we built the NAS based on the first principal component with these 4 dietary antioxidants. The calculation and distribution of NAS are listed in Supplement Figure 3. The higher NAS value indicates higher levels of dietary antioxidants based on vitamin A, B2, E, and folate.

Based on the quartiles of the NAS, we classified participants into 4 groups: low (NAS ≤ −0.46), medium-low (NAS = −0.45 to 0.15), medium-high (NAS = 0.16 to 0.72), and high (NAS > 0.72). As shown in Figure 1, the prevalence of HR-HPV is 23.8%, 19.5%, 17.3%, and 16.2% for the low, medium-low, medium-high, and high NAS groups. There was a significant decrease in HR-HPV prevalence in the subgroups from low to medium-high NAS group; however, we observed similarities between medium-high and high NAS group (the between-group comparison P = .006, P = .023, and P = .464, respectively). The LR-HPV prevalence for the low, medium-low, medium-high, and high groups were 24.3%, 23.6%, 19.0%, and 19.0%, respectively. The LR-HPV prevalence was similar between the 2 low NAS groups and between the 2 high NAS groups; however, it was significantly lower in the medium-low than the medium-high NAS group (P = .204, P = .0015, and P = .907, respectively). For easy interpretation, we also classified albumin using 39, 41, and 44 g/L as the cut-points based on its weighted quartiles. As shown in Table 4, the women with a low level of albumin (≤39 g/L) have a significantly higher risk of HR-HPV (OR = 1.4, P = .009 vs >44 g/L). The women with the lowest quartile NAS (NSA ≤ −0.46) had a higher chance of both HR-HPV (OR = 1.3, P = .030) and LR-HPV (OR = 1.4, P = .002) compared with the women with the highest quartile NAS.

Figure 1.

Vaginal human papillomavirus (HPV) status by the levels of the Nutritional Antioxidant Score (NAS) based on 4 dietary antioxidants (vitamin A, B2, E, and folate). Note: HPV% were calculated by considering the sampling weights. Low: less than the lowest 25% of the NAS (NAS ≤ −0.46, n = 2936), medium-low: 25.1%–50% (NAS = −0.46 to 0.15, n = 2720), medium-high: 50.1%–75% (NAS = 0.16 to 0.72, n = 2559), high: >75% (NAS > 0.72, n = 2328) of the NAS. For high-risk HPV prevalence: low < medium-low < medium-high = high with P = .006, P = .023, and P = .464, respectively. For low-risk HPV prevalence (24.3%, 23.6%, 19.0%, and 19.0%): low = medium low < medium high = high with P = .204, P = .0015 and P = .907, respectively.

Table 4.

Impact of albumin and the NAS Associated With Vaginal HPV Status

		Model 1 (n = 10 073)a		Model 2 (n = 7253)b
Biomarker (Unit)	Sample Size	Low-Risk HPV vs No HPV OR (95% CI)	High-Risk HPV vs No HPV OR (95% CI)	Low-Risk HPV vs No HPV OR (95% CI)	High-Risk HPV vs No HPV OR (95% CI)
Albumin (g/L)
≤39	2687	1.3 (1.1–1.6)**	1.6 (1.3–2)***	1.0 (0.8–1.3)	1.4 (1.1–1.9)**
39.1–41	2349	1.2 (1–1.4)*	1.3 (1–1.6)*	1.0 (0.8–1.2)	1.2 (0.9–1.6)
41.1–44	3506	1.0 (0.9–1.2)	1.2 (1–1.4)	0.9 (0.7–1.1)	1.1 (0.8–1.4)
>44	1991	1	1	1	1
Nutritional Antioxidant Score
Low (≤25%)	2936	1.7 (1.4–2.1)***	1.8 (1.4–2.1)***	1.4 (1.1–1.8)**	1.3 (1.0–1.7)*
Median low (25.1–50%)	2720	1.4 (1.1–1.6)***	1.3 (1.1–1.5)**	1.2 (1.0–1.5)	1.2 (0.9–1.4)
Median high (50.1–75%)	2559	1.1 (0.9–1.3)	1.1 (0.9–1.3)	1.1 (0.9–1.5)	1.1 (0.9–1.4)
High (>75%)	2328	1	1	1	1

Abbreviations: CI, confidence interval; HPV, human papillomavirus; NAS, Nutritional Antioxidant Score; OR, odds ratio; RAE, retinol activity equivalents.

*P < .05, **P < .01, and ***P < .001 based on multinomial logistic models.

^aModel 1: model with albumin and NAS (based on vitamin A, B2, E, and folate).

^bModel 2: albumin and NAS + 9 covariates (age, race, education, income, marital status, smoking status, lifetime illegal substance use, past 12-month alcohol intake, number of sexual partners in past 12 months).

DISCUSSIONS

Our study demonstrates that elevated levels of the 5 antioxidants (albumin, vitamin A, vitamin B2, vitamin E, and folate) were significantly associated with a lower risk of HR-HPV infection. Our findings can be generalized to general US adult women aged because the NHANES, the US national survey, was applied. For our study participants, the averages of albumin levels (41.9 g/L) for all female participants and by vaginal HPV infection status (no HPV, LR-HPV, and HR-HPV) were in the normal range, which is 35–50 g/L [24] (see Supplement Table S1). The means of vitamin B2 (1.9 mg) and folate (479.5 mcg, dietary folate equivalents [DFE)] intake for all female participants and by vaginal HPV infection status were higher than the recommended dietary allowances, 1.1 mg and 400 mcg DFE, respectively [25]. However, the average levels of the participants’ vitamin A (589.7 mcg retinol activity equivalents [RAE]) and vitamin E (7.5 mg) were lower than the recommended dietary allowances of 700 mcg RAE and 15 mg, respectively [26, 27]. Prior research has shown that a healthy diet and frequent consumption of antioxidant nutrients could decrease the risk of HPV infection and cervical cancer. Frequent intake of fruits with high antioxidant nutrients has also shown inverse associations with precancer cervical lesions and cervical cancer [28]. Women who consumed more fruits and vegetables were shown to have a lower risk of HR-HPV infection [29]. In addition, women with a low dietary antioxidant intake had a higher risk of being infected by HR-HPV [30]. Our study showed evidence that a lower level of 5 specific antioxidants was associated with a higher risk of HR-HPV infection.

Albumin, the most bountiful circulating protein in the plasma, uses important antioxidant activities [31]. Serum albumin is viewed as the primary extracellular molecule for balancing the plasma redox state[32]. Serum albumin can be reduced by inflammation and malnutrition of inadequate protein intake because both of these mechanisms lower their synthesis rate [33]. Serum albumin has been used in building several indicators for measuring systemic inflammation, such as the Inflammatory-Nutritional Index, the Glasgow Prognostic Score, and cervical cancer systemic inflammation score [34–36]. The low serum albumin level was associated with short overall survival and disease-free survival for cervical cancer patients [36, 37]. Decreased serum albumin was found to be associated with the increased systemic inflammation and impaired immune response. Prior research observed that a low serum albumin level is associated with cancer prognosis in various types of cancers, including cervical cancer [38]. Thus, lower serum albumin levels may also be associated with an increased risk of cancer patients’ mortality [39]. These findings of serum albumin from the literature are consistent with the results of our study.

Previous studies also observed that women with HR-HPV infection had a lower vitamin A intake than those without HPV infection. One study showed that insufficient intake of vitamin A was associated with HPV infection [40], HR-HPV infection [30], and cervical cancer [41]. Vitamin A impacts cellular activities and plays an important role in the malignant transformation of tissues [42]. In addition, vitamin E has been shown to reduce the risk of cervical cancer because it could reduce inflammatory response and suppress the expression of HPV oncogenes [43]. A systematic review showed that vitamin E and folate might protect women against HPV infection [43]. A meta-analysis based on 15 studies showed that vitamin E intake and serum vitamin E levels had a significant inverse association with the risk of cervical neoplasm [44].

Folate is involved in red blood cells, DNA synthesis, DNA repair, DNA methylation, and cell proliferation [45]. A meta-analysis showed that a lower serum folate level was associated with the elevated risk of cervical cancer [14]. In addition, low serum folate level was shown to be associated with cervical cancer progression [46]. High folate level enhanced HR-HPV clearance based on a prospective study with 345 women with a risk of developing cervical intraepithelial neoplasia [47]. Vitamin B2 (or riboflavin) plays a critical role in carbon metabolism for normal cellular processes [48]. Vitamin B2 is also involved in metabolic redox reactions and 1-carbon metabolism, which can interfere with DNA functions [49]. Thus, deficiency in vitamin B2 may induce carcinogenesis from impairments in DNA replication, repair, or epigenetically through methylation to control gene expression [48]. Levels of vitamin B2 were significantly lower in the cervical tissues with HPV 16/18 than HPV-negative tissues among cervical cancer patients [48]. Women with cervical cancer or cervical intraepithelial neoplasia have decreased vitamin B2 levels in the plasma and cervical tissue samples compared with women with normal cervical tissues [48]. In addition, deficiency of vitamin B2 has been shown to involve progression of various cancers, such as cervical cancer and colorectal cancer [48].

There are several strengths of this study. The NHANES data includes a sound sampling design; therefore, the study findings with appropriate sampling weighting can represent the US population. Moreover, the national survey data provide a large sample size and abundance of HPV health-related information. This allowed us to control many potential confounders, such as smoking, alcohol intake, and illegal substance use, when evaluating the impact of antioxidants associated with HR-HPV. Previous literature showed that exogenous OS sources are environmental pollution, adverse health behaviors (such as smoking, alcohol intake, and unhealthy diet), drugs, and lifetime stress, which may be associated with health outcomes [50]. Our findings support the association between several antioxidants and HR-HPV infection, even when controlling for demographic characteristics and health behaviors. While having these strengths, this study has the following limitations. First, this study cannot identify or infer a causal relationship between antioxidants and HR-HPV infection due to the usage of cross-sectional NHANES data. Second, this study cannot evaluate the impact of antioxidants associated with HPV acquisition and persistence because only 1-time HPV infection was measured. Furthermore, large and prospective studies are needed to evaluate the effect of antioxidants and HR-HPV acquisition and persistence. Third, there may be a recall bias for the self-reported health behaviors such as sex partners and alcohol intake during the interview.

CONCLUSIONS

In summary, our findings demonstrated inverse associations between the 5 antioxidants and HR-HPV infection. The higher serum albumin and higher intake of the 4 dietary antioxidants (vitamins A, B2, and E, and folate) were associated with a lower risk of HR-HPV infection. Antioxidants may provide a protective effect on HR-HPV infection, decreasing the risk of cervical cancer development. However, future longitudinal studies will still be needed to verify the role of antioxidants on HPV carcinogenesis.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Financial support. This study was funded by the National Institutes of Health (NIH)/ National Institute of General Medical Sciences (NIGMS) (P20 GM121288, Principal Investigator [PI] K. R. and subproject PI H.-Y. L.).

Potential conflicts of interest. All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest.