Inverse relationship between dietary fiber intake and environmental exposure to acrylamide

Dietary fiber intake was thought to decrease some environmental pollutant exposure by increasing gastrointestinal excretion. While diet is considered the major source of exposure to acrylamide (AA), the impact of dietary fiber intake on acrylamide (AA) exposure is still unknown. We analyzed the associations between dietary fiber intake and AA hemoglobin biomarkers [hemoglobin adducts of acrylamide (HbAA) and glycinamide (HbGA), and sum of HbAA and HbGA (HbAA + HbGA)] among 3448 US adults who participated in the National Health and Nutrition Examination Survey (NHANES) 2013–2016. Multivariable linear regression and cubic spline models were conducted to estimate the associations between dietary fiber intake and AA hemoglobin biomarkers. Dietary fiber intake had a strong inverse and J-shaped association with AA hemoglobin biomarkers. In the fully adjusted linear regression model, compared with participants in the lowest dietary fiber quantile, the adjusted percent change with 95% confidence intervals (CIs) in HbAA for the highest dietary fiber quantile was − 19.7% (− 26.7%, − 13.1%); for HbGA, it was − 12.2% (− 18.9%, − 4.9%), and for HbAA + HbGA, it was − 17.3% (− 23.7%, − 10.4%). Associations between higher dietary fiber intake and lower levels of environmental exposure to acrylamide hemoglobin biomarkers suggest the need to increase dietary fiber intake.


Introduction
Acrylamide (AA) exposure has been identified as a probable carcinogen to humans and has also been associated with a wide variety of health outcomes in humans, including cardiovascular disease , diabetes (Yin et al. 2021), lung disease (Liu et al. 2021a), and obesity . AA exposure can occur from tobacco smoke and a wide variety of commonly consumed dry-heated food or beverages, including fried potatoes, biscuits, potato chips, and coffee (Tareke et al. 2002). AA was measured by gas chromatography-mass spectrometry analysis of its hemoglobin (Hb) adduct to N-terminal valine following acid hydrolysis, ion-exchange chromatography, and derivatization (Bergmark et al. 1993). In humans, AA hemoglobin biomarkers including Hb adducts of AA (HbAA) and its major metabolite glycidamide (GA) (HbGA) were used to monitor exposure to AA. In the general population, diet is considered the major source of exposure to AA (Pedersen et al. 2022).
Dietary fiber intake is increasingly recognized as critical for the prevention of diseases, including cardiovascular disease (Pereira et al. 2004;Zhang et al. 2022), depression (Liu et al. 2021b), hyperuricemia (Sun et al. 2019), hypertension (Sun et al. 2018), and stroke (Larsson and Wolk 2014), and the study of dietary fiber intake has recently extended to the field of environmental toxicology (Dzierlenga et al. 2021;Shi et al. 2022). Dietary fiber intake was thought to decrease serum environmental pollutants such as perfluoroalkyl acids (Dzierlenga et al. 2021) and isopentanaldehyde ) by increasing gastrointestinal excretion. Since diet is considered the major source of exposure to AA, and dietary fiber increases the volume of feces and decreased the transit time in the intestine, dietary fiber intake may reduce the blood concentration of AA hemoglobin biomarkers. However, no epidemiological study has been reported to demonstrate the associations between daily dietary fiber intake and environmental exposure to AA. In the actual production process, there are few methods to mitigate AA exposure. High-fiber diets may be practical for populations involved in the production of AA. Therefore, our study firstly focused on the correlations between dietary fiber intake and the blood concentrations of AA hemoglobin biomarkers.

Study population
Publicly available data regarding participants who were recruited for the NHANES were used. NHANES is an ongoing cross-sectional study conducted by the National Center for Health Statistics (NCHS) of the Centers for Disease Control and Prevention (CDC). NHANES is a program of studies aiming to assess the health and nutritional status of adults and children in the USA. Data on certain aspects such as environmental exposures was also be collected. Detailed survey operation manuals, survey design, methods, available data, and consent documents are available on the NHANES website (https:// www. cdc. gov/ nchs/ nhanes/). The data were from two cycles of the NHANES (2013)(2014)(2015)(2016) and included a total of 20,146 participants. Participants with missing data on acrylamide and glycinamide were excluded (n = 15,183). A total of 466 individuals lacked data on dietary fiber intake were also excluded from this study. We selected 3448 participants aged 18 years and older who were not pregnant (Fig. 1).

Environmental AA measurements
Blood levels of HbAA and HbGA were measured in fresh or frozen erythrocytes or ethylenediaminetetraacetic acid (EDTA)-anticoagulated whole blood specimens. The measure of HbAA and HbGA includes a measurement procedure for total hemoglobin and the adduct of these chemicals at the N-terminal valine of the Hb protein chains. The measurement of total hemoglobin is performed using a commercial assay kit based on a well-established procedure. Quantitation of HbAA and HbGA is performed using Edman products by high-performance liquid chromatography/tandem mass spectrometry (HPLC-MS/MS) and processing results. Details regarding the experimental methods for HbAA and HbGA are available on the NHANES website (https:// wwwn. cdc. gov/ nchs/ data/ nhanes/ 2015-2016/ labme thods/ AMDGYD_ ETHOX_I_ MET. pdf). For concentrations of HbAA and HbGA below the limit of detection (LOD), we recorded the values as the LOD divided by the square root of 2.

Dietary information
Dietary fiber intake was obtained from two 24-h dietary recall interviews. The two interviews were conducted in person in the mobile examination center (MEC) and by telephone at an interval of 3 to 10 days. Dietary fiber intake was calculated based on an average of consumed amount of individual foods reported in the 24-h recalls according to the United States Department of Agriculture (USDA) Food and Nutrient Databases for Dietary Studies (Oza-Frank et al. 2009). Total dietary fiber intake was recoded as the fiber remaining after a food's digestion by α-amylase, protease, and amyloglucosidase and obtained from two 24-h dietary recall interviews. Dietary fiber intake was adjusted to daily energy intake. Dietary fiber intake was calculated as the residual of the regression of dietary fiber intake on total caloric intake.

Covariates
The following potential confounding factors were included in our analyses: age, sex, race, educational level, self-reported history of hypertension and diabetes, smoking (smoking at least 100 cigarettes in life or not), drinking (having at least 12 alcohol drinks per year or not), physical activity (vigorous activity, moderate activity, and never), body mass index (BMI), and total daily energy intake. Family poverty-income ratio (PIR) was an index for the ratio of monthly income to poverty and calculated by dividing family income according to the Department of Health and Human Services' (HHS) poverty guidelines. Total daily energy intake was obtained from two 24-h dietary recall interviews by trained interviewers. For each participant, daily total energy and nutrient intakes from foods and beverages were collected. The information regarding daily energy intake was extracted from the NHANES dietary data.

Statistical analysis
Measurement data were described by the mean (standard deviation, SD) or median (interquartile range, IQR) according to the normality of the data, and analysis of variance or Kruskal-Wallis H rank-sum test was used for comparison between multiple groups. Categorical variables were described by the numbers and percentage [n (%)], and the chi-square test was used to perform the comparison between groups. Since the distribution of environmental AA exposure was skewed, we log2-transformed the concentrations of HbAA and HbGA to normalize the distributions for regression analysis.
Generalized linear regression models were performed to assess the association between dietary fiber intake and the blood concentrations of HbAA, HbGA, and HbAA + HbGA. All regression analyses were adjusted by age, sex, education level, race, smoker, alcohol user, diabetes, hypertension, body mass index, poverty-income ratio, and physical activity. Cubic splines with 3 knots located at the 10th, 50th, and 90th percentiles of the distribution were conducted to explore the relationship between dietary fiber intake and the blood concentrations of HbAA, HbGA, and HbAA + HbGA. The percentage differences (% diffs) and 95% CIs in blood concentrations of HbAA, HbGA, and HbAA + HbGA across quantiles of dietary fiber intake after adjusting for confounding factors were calculated. Quartiles of dietary fiber intake were regarded as a continuous variable in the multivariable linear regression model to test for significant linear trends. Previous study demonstrated that the hemoglobin adducts of AA were strongly associated with obesity . Because fiber also lowers BMI, dietary fiber intake and the hemoglobin adducts of AA relationship confounding by obesity may be worth evaluating. Thus, we performed subgroup analysis by obesity. All statistical analyses were carried out in R (version 4.0.3). A two-sided p value < 0.05 was considered statistically significant.

Study population characteristics
The baseline characteristics of the study population by quantiles of dietary fiber intake are shown in Table 1. Participants in the lowest quantile of dietary fiber intake were more likely to be younger or male; have lower incidences of smoking, alcohol use, diabetes, and hypertension; and have a lower PIR.

Multivariable linear regression model to assess the association between dietary fiber intake and blood concentrations of HbAA, HbGA, and HbAA + HbGA
The results of the multivariable linear regression are shown in Table 2. After adjusting for all covariates, including age, sex, education level, race, smoking, alcohol use, diabetes, hypertension, body mass index, poverty-income ratio, and physical activity, dietary fiber intake was negatively associated with the concentrations of HbAA, HbGA, and HbAA + HbGA. Compared with the lowest quantile, the β values with 95% confidence intervals (CIs) across the quantiles of dietary fiber intake were − 0.13 (− 0.21 to − 0.05), − 0.24 (− 0.31 to − 0.16), and − 0.22 (− 0.31 to − 0.14) for HbAA, HbGA, and HbAA + HbGA, respectively. Dietary fiber intake was approximately a J-shaped relationship with the concentrations of HbAA, HbGA, and HbAA + HbGA, and tests for nonlinearity were significant (Fig. 2).

Subgroup analysis by obesity
In subgroup analysis by obesity, the results among obesity and non-obesity were inconsistent with those among the total population (detailed regression result data were presented in Table 3). BMI-stratified models suggest that the association between dietary fiber intake and the blood concentrations of HbAA, HbGA, and HbAA + HbGA was more significant among non-obese adults (all p value for interaction < 0.05).

Discussion
To our knowledge, our study is the first to explore the associations between dietary fiber intake and AA hemoglobin biomarkers among the general population. Our results indicated that dietary fiber intake displayed an inverse and nonlinear relationship with the blood concentration levels of AA hemoglobin biomarkers. Given our epidemiologic data suggesting that fiber might decrease blood concentration levels of AA hemoglobin biomarkers, highfiber diets are recommended for populations involved in the production of AA.  The underlying mechanism of dietary fiber intake-induced decreases in blood concentration levels of AA hemoglobin biomarkers remains unknown. Possible mechanisms are as follows. First, diet was considered the major source of exposure to AA, though insoluble dietary fiber helped to reduce the absorption of AA because it increased the volume of feces and decreased the transit time in the intestine. Second, epidemiological study showed that acrylamide exposure was positively associated with systemic inflammatory mediator plasma C-reactive protein.
Animal experiments demonstrated that dietary fiber mediated oxidation by activating the Acly/Nrf2/NF-κB signaling pathway, which exerted anti-inflammatory effects by inhibiting the differentiation of macrophages into pro-inflammatory M1 macrophages (Shao et al. 2022). Thus, it is plausible to consider that the pathogenesis of dietary fiber may be related to decreased concentrations of AA hemoglobin biomarkers through its antiinflammatory effects. Third, the liver is in the first line to help our body defense against harmful substances outside including environmental pollutants. A recent study clarified the inverse associations of dietary fiber intake with nonalcoholic fatty liver disease (NAFLD) (Zhao et al. 2020). NAFLD patients are likely to have decreased dietary fiber intake and impaired liver function, while the liver plays a crucial role in AA metabolism. Finally, intestinal microorganisms are fermented Fig. 3 Percent change in AA hemoglobin biomarkers by quartile of dietary fiber intake. Analyses were adjusted for age, sex, education level, race, smoker, alcohol user, diabetes, hypertension, body mass index, poverty-income ratio, and physical activity Table 3 Association between dietary fiber intake and serum concentration of acrylamide hemoglobin biomarkers stratified by BMI Analyses were adjusted for age, sex, education level, race, smoking, drinking, diabetes, hypertension, poverty-income ratio, and physical activity. Serum concentration of acrylamide hemoglobin biomarkers was log2-transformed to fit the regression model. HbAA, hemoglobin adducts of acrylamide; HbGA, hemoglobin adducts of glycinamide; HbAA + HbGA, sum of HbAA and HbGA; Q, quantile. * p < 0.05, ** p < 0.01, *** p < 0.001

HbAA
HbGA by dietary fiber to produce short-chain fatty acids, which can improve liver metabolism (Lundin et al. 2004), thus increasing AA metabolism. We have observed J-shaped relationship between dietary fiber intake and AA hemoglobin biomarkers. The regression splines suggested a possible threshold effect for dietary fiber intake of approximately 20 g/day on AA hemoglobin biomarkers. On the left side of threshold effect point, dietary fiber intake showed a linear and inverse relationship with the concentrations of AA hemoglobin biomarkers and the effect became flat on the right side of threshold effect point. In addition, our results showed a significant interaction between dietary fiber intake and obesity status in relation to the concentrations of AA hemoglobin biomarkers. The associations between dietary fiber intake and AA hemoglobin biomarkers were less significant among obese adults compared with non-obese adults.
Our results are consistent with previous data reporting the inverse association of dietary fiber intake with environmental pollutants such as perfluoroalkyl acids (Dzierlenga et al. 2021) and isopentanaldehyde ). In the study conducted by Dzierlenga et al., dietary fiber was associated with lower serum concentrations of perfluoroalkyl substances due to increased gastrointestinal excretion. Additionally, in the study conducted by Shi et al., dietary fiber intake was inversely associated with blood levels of isopentanaldehyde and propionaldehyde. Our results add information of interest to the available research on this topic.
There are some limitations in our study. First, we cannot infer causal relationships directly and which factor came first because of the cross-sectional nature of this study. Second, dietary fiber intakes were obtained from two 24-h dietary recall interviews at an interval of 3 to 10 days, while AA hemoglobin biomarkers can reflect the exposure level within 120 days (Bergmark et al. 1993). Dietary fiber intake evaluated over 2 days may not be long enough to estimate the cumulative effect of fiber intake on AA exposure. Third, the dietary data of fiber intake were obtained from two 24-h dietary recall interviews, and ineluctable recall bias may exist. Finally, although we try to control a series of confounding factors, unmeasured confounding factors may also play a role.

Conclusion
Dietary fiber intake was inversely and nonlinearly associated with environmental exposure to AA. The possible obesity-specific effect of dietary fiber intake on AA hemoglobin biomarkers was suggested in our study. Our results also suggested that high-fiber diets are recommended for populations exposed to elevated AA levels. However, further studies should be employed to elucidate these relationships.