CVD is a leading cause of morbidity and mortality worldwide, partly driven by modifiable risk factors such as poor diet and physical inactivity which contribute to hypertension and dyslipidaemia(1). Inflammation plays a key role in the development and progression of CVD, by promoting endothelial dysfunction through the recruitment of immune cells and release of pro-inflammatory cytokines. These processes contribute to the accumulation of lipids and immune cells in the arterial wall, forming plaque that are prone to instability(Reference Henein, Vancheri and Longo2). Unresolved inflammation weakens the fibrous cap of these plaques, increasing risk of rupture and thrombosis. Reduced nitric oxide (NO) bioavailability exacerbates this process by impairing platelet regulation, promoting oxidation and inflammation and contributing to vascular damage(Reference Jebari-Benslaiman, Galicia-García and Larrea-Sebal3). Nitric oxide is essential for vascular homoeostasis(Reference DeMartino, Kim-Shapiro and Patel4) and is derived from endogenous production via the L-arginine-NO pathway and exogenous sources like dietary nitrate in vegetables, meat and water, with vegetables being the primary contributor (∼80 % of exogenous intake)(Reference Andrabi, Sharma and Karan5). Dietary nitrate is converted to NO through nitrate-reducing bacteria in saliva and in the stomach forms nitrous acid breaking down into NO. This process is enhanced by reducing agents such as vitamin C and polyphenols, such as epicatechin, catechin, quercetin and oleuropein(Reference Bowles, Burleigh and Mira6,Reference Rocha, Gago and Barbosa7) . Alternatively, absorbed nitrite, a metabolite of nitrate, can be converted to NO by various enzymes.
Two major dietary sources of nitrate (vegetables and meat) appear to play a distinct role in human health(Reference Bondonno, Zhong and Bondonno8). Vegetable-derived nitrate is considered cardioprotective, as shown by meta-analyses linking habitual intake to reduced CVD incidence and mortality, and supplementation (e.g. beetroot juice) to improved blood pressure, arterial stiffness and endothelial function(Reference Tan, Stagg and Hanlon9,Reference Jackson, Patterson and MacDonald-Wicks10) . Leafy greens and beetroot are particularly cardioprotective due to their high nitrate content, and micronutrients such as vitamin C and polyphenols inhibit the formation of harmful N-nitrosamines produced when nitrate reacts with amines under acidic gastric conditions(Reference Ahluwalia, Gladwin and Coleman11). Conversely, nitrate in processed meats is often linked to adverse cardiovascular outcomes (likely due to concomitant high Na content), including increased blood pressure, oxidative stress and endothelial dysfunction(Reference Kotopoulou, Zampelas and Magriplis12). Without protective components such as vitamin C and polyphenols, nitrate in processed meats can form N-nitrosamines that are linked to cancer and oxidative stress(Reference Bowles, Burleigh and Mira6). Although few studies have directly compared nitrosamine formation from different nitrate sources, one dietary intervention randomised crossover trial demonstrated that a diet high in red or processed meat markedly increased faecal N-nitroso compounds compared with a vegetarian diet, with the vegetarian diet associated with very low endogenous nitrosation(Reference Joosen, Kuhnle and Aspinall13).
The health effects of dietary nitrate may go beyond traditional cardiovascular outcomes, with growing evidence suggesting it modulates inflammation through the reduction of leukocyte recruitment, inhibition of the expression of adhesion molecules and reduction of vascular permeability and leukocyte transmigration(Reference Raubenheimer, Bondonno and Blekkenhorst14). Given inflammation’s role in CVD progression and dietary nitrates’ potential to influence it, exploring vascular markers like lipoprotein-associated phospholipase A2 (Lp-PLA2) is important. Lp-PLA2 is a vascular-specific inflammatory marker linked to plaque instability and CVD (e.g. coronary heart disease, stroke and type 2 diabetes)(Reference Burke and Dennis15). Unlike C-reactive protein (CRP), which reflects systemic inflammation, Lp-PLA2 is more directly associated with vascular inflammation and atherosclerosis. It catalyses the breakdown of oxidised phospholipids on LDL, producing proinflammatory by-products, such as lysophosphatidylcholine (lysoPC) and oxidised nonesterified fatty acids (oxNEFA), which contribute to endothelial dysfunction and plaque instability(Reference Corson, Jones and Davidson16).
While dietary nitrate is linked to improved cardiovascular health, the effects of different sources (plants, animal and processed meat) on inflammation, lipids and other markers of CVD risk remain unclear. This study investigates source-specific nitrate intake and its associations with inflammatory markers CRP and Lp-PLA2, lipid metabolism and CVD risk across varying cardiovascular risk levels.
Methods
Participants
This cross-sectional study was conducted on the Gold Coast, Queensland, Australia, between February 2021 and April 2022. Full details of the study methods have been published previously(Reference English, Lohning and Mayr17). One hundred participants, adults aged 18–70 years, were purposively recruited to achieve a sample of healthy adults at varying risk of CVD (according to levels of modifiable risk factors, medication use and family history). Smokers and those with a history of myocardial infarction, stroke or peripheral vascular disease or taking medications affecting Lp-PLA2 and lipids were excluded. This study protocol was approved by the Bond University Human Research Ethics Committee (approval DR03194) and conforms to the ethical guidelines of the 1975 (revised in 1983) Declaration of Helsinki. All participants provided written informed consent.
Data collection
Data were collected through face-to-face visits at the Bond Institute of Health and Sport encompassing anthropometric, biochemical and clinical measurements. Habitual dietary intake was assessed at the visit via the European Prospective Investigation into Cancer FFQ, with minor modifications applied to reflect Australian foods(Reference Riboli18). A subset of participants (n 89) completed an estimated 3-d food diary (consecutive days with one weekend day) in household measures the week following their study visit. For each participant, the reported daily quantity (g/d) of each food item reported in the FFQ and diary was calculated by averaging intakes over the respective periods reported. Average daily food item intakes were multiplied by the designated median nitrate value (mg/g fresh weight) for each food item, using values obtained from national, comprehensive nitrate food composition databases(Reference Zhong, Blekkenhorst and Bondonno19,Reference Zhong, Liu and Blekkenhorst20) .
Nitrate intake estimates from the FFQ (n 100) were the primary source of data for the analysis. These estimates were calibrated using the 3-d food diary data (n 89) as these data are more likely to be representative of usual intakes. Calibration was performed using simple linear regression, with the 3-d diary intake as the dependent variable and FFQ-derived intake as the predictor, as described by Kippnis et al. (Reference Kipnis, Midthune and Freedman21) The slope (β) of the regression line (representing the unstandardised regression coefficient) was used to calibrate each FFQ-derived nitrate estimate (mg/d) using the formula:
Calibrated nitrate intake = α + (β × FFQ)
where α is the intercept and β is the slope.
Agreement between FFQ and 3-d diary estimates was assessed by nitrate source using Spearman’s correlation coefficients and Bland–Altman analysis. Correlation values ≥ 0·3 were considered a medium effect size according to Cohen (1988)(Reference Cohen22), and mean differences close to zero with narrow limits of agreement were interpreted as indicators of acceptable alignment. The calibrated nitrate values were used as independent variables in all regression analyses.
Foods were grouped into three main categories:(Reference Erichsen, Pokharel and Kyrø23)
-
1. Plant sources: fruits, vegetables, legumes, wholegrains, nuts, seeds, tea and coffee
-
2. Naturally occurring animal sources: red meat, poultry, offal, dairy, eggs, fish and other seafood
-
3. Additive-permitted meat sources: sausage, bacon, corned beef and other processed meats.
Animal-sourced nitrate was differentiated as ‘naturally occurring’ (e.g. red meat, poultry) and ‘additive-permitted’ (e.g. processed meats such as bacon and sausage) to better understand the distinct impacts of naturally present v. added nitrates on inflammatory markers and lipid profiles, given their differing metabolic pathways and potential health implications. A separate analysis was conducted to evaluate the specific impact of total vegetable-sourced nitrate as well as nitrate-rich vegetables such as leafy greens and beetroot on CVD risk factors.
Outcomes included inflammatory markers (Lp-PLA2 and hs-CRP) and CVD risk factors (waist circumference, HDL- and LDL-cholesterol, TAG, systolic and diastolic blood pressure). Anthropometric data were measured in the fasting state with light clothing but without shoes. Standing height was measured to the nearest 0·1 cm using a wall mounted stadiometer. Weight was measured with a calibrated digital scale to the nearest 0·1 kg. Waist circumference was measured six times, three times at the umbilicus and three at minimum waist(Reference Brown, Randhawa and Canning24). BMI was calculated as weight (kg)/height (m2).
Blood pressure was measured seated, with a Creative Medical PC-900 Pro Vital Signs monitor in the non-dominant arm, with a clinical cuff and measured in triplicate, two minutes apart. The first measurement was disregarded, and the second and third measurement were averaged(Reference O’Brien, Coats and Owens25).
Information on age, sex, medical history, menopausal status, medication and supplement intake, smoking status and alcohol consumption was self-reported from questionnaires completed during the study visit.
Physical activity was measured using the WHO Global Physical Activity Questionnaire(Reference Armstrong and Bull26). Participant’s metabolic equivalent minutes per week were calculated based on the participant responses from sixteen questions assessing time spent physically active during work, travel and recreation in addition to sedentary time. Physical activity levels were categorised into tertiles based on WHO’s physical activity recommendation where 0 = low MET < 600 min/week, 1 = moderate metabolic equivalent ≥ 600 to < 1500 min/week and 2 = high metabolic equivalent ≥ 1500 min/week.
Assessment of biomarkers
Fasting blood samples were collected in EDTA tubes and centrifuged within 30 min of collection at 4°C and 1·3 relative centrifugal force (rcf) for 15 min. Plasma was aliquoted and immediately frozen at −80° until assays were performed. Lipids were measured with Afinion point of care machine within 10 min of blood collection. Measurements included LDL, HDL, TAG, non-HDL-cholesterol and Total:HDL ratio. All values that exceeded detection range of the instrument were recorded as the value at the top of detection range. Lp-PLA2 activity was determined using a commercial colorimetric assay Cayman Chemical Co, USA with 2-thio-PAF as substrate. All samples were analysed in triplicate. Briefly, Lp-PLA2 hydrolyses the acetyl thioester bond at the sn-2 position of the substrate, 2-thioPAF, creating free thiols which are detected using 5,5’dithio- bis-(2-nitrobenzoic acid, DTNB). Absorbance was measured at 412 nm using an OMEGA Fluostar microplate plate reader. High-sensitivity CRP was assayed by Queensland Health, Pathology Queensland.
Data analysis
All data were analysed using SPSS version 29.0.1.0 (171) (SPSS Inc.). Data were assessed for normality via Q–Q plots. Variables not normally distributed were log transformed before data analysis where possible (hsCRP). Independent t tests were performed on normally distributed variables to test for differences between males and females. Mann–Whitney U tests were performed for non-normally distributed variables. Multiple regression models were conducted using pairwise deletion (exclude cases pairwise), allowing participants with non-missing values for the variables included in each model to contribute to that analysis. Checks for multicollinearity were conducted using variance inflation factor and tolerance indices and revealed no variables were highly correlated.
Multiple linear regression examined associations between calibrated source-specific nitrate and inflammatory markers CRP and Lp-PLA2 and CVD risk factors. Exposure variables (plant-sourced, naturally occurring animal-sourced and additive-permitted meat-sourced) were standardised, and beta coefficients in regression models represent the effect per 1 sd increase in nitrate intake. Results of multiple linear regression are reported as standardised coefficients β and 95 % CI, and significance is reported as P values where P < 0·05 was considered statistically significant.
Covariates included age, sex, waist circumference, physical activity levels, alcohol intake and socio-economic index. Socio-economic status was determined using the Australian Socio-Economic Indexes for Areas database, based on participant’s postcode at the time of the study(27).
Three models of adjustments were used. Model 1 adjusted for age and sex. Model 2 included age, sex, physical activity, waist circumference, socio-economic index and alcohol consumption. Model 3, based on methods from a similar study(Reference Bondonno, Pokharel and Bondonno28) included further adjustments for dietary intake from other nitrate sources:
(a) plant-sourced nitrate – adjusted for red meat, poultry, offal, dairy, eggs, fish and other seafood, sugar and confectionary, soft drinks and refined grains
(b) naturally occurring animal and additive-permitted meat sources – adjusted for fruits, vegetables, legumes, wholegrains, nuts, seeds, tea and coffee, sugar and confectionary, soft drinks and refined grains.
The ‘all-components model’ was used in model 3 where all dietary covariates, including nitrate sources other than the exposure of interest, were included in the model. This method provides unbiased estimates compared to other energy adjustment techniques(Reference Tomova, Arnold and Gilthorpe29) and accounts for underlying dietary patterns.
Regression results were translated into practical terms by multiplying β-coefficients by the sd of the outcome variable to estimate the absolute change in the outcome (e.g. LDL-cholesterol) associated with a one sd increase in predictor variable (e.g. plant-sourced nitrate intake). The sd for additive-permitted meat nitrate was estimated using the interquartile range divided by 1·35, to approximate the sd for skewed data(Reference Wan, Wang and Liu30).
As a sensitivity analysis, models were re-run with adjustment for total energy intake instead of dietary food-group covariates to assess whether associations were driven by overall energy intake.
Results
Clinical characteristics
A total of 132 adults were recruited; four did not meet inclusion criteria and twenty eight declined to participate, leaving 100 participants for analysis. Demographic and clinical characteristics for the total cohort, and males and females, are shown in Table 1. The mean age was 49 (range 20–69) years, and 92 % of participants were Caucasian.
Demographic and clinical characteristics of cross-sectional study participants
DBP, diastolic blood pressure; hsCRP, high sensitivity C-reactive protein; Lp-PLA2, lipoprotein-associated phospholipase A2; mg/l, milligrams per litre; mm Hg, millimetres of mercury; mmol/l, millimoles per litre; nmol/min/ml, nanomoles per min per millilitre; PA, physical activity; SBP, systolic blood pressure. Bolded P values denotes statistical significant difference between the groups.
‡ n 99.
Females consumed more plant-sourced nitrate, while males consumed more additive-permitted meat-sourced nitrate. Naturally occurring animal-sourced nitrate did not differ by sex (Table 1).
Dietary data calibration and agreement
Spearman’s correlations showed moderate agreement between FFQ and 3-d diary estimates for plant (ρ = 0·412, P < 0·001), naturally occurring animal (ρ = 0·460, P < 0·001) and additive-permitted meat nitrate sources (ρ = 0·335, P = 0·001). Bland–Altman analyses demonstrated small mean differences (plant: 10·29 mg, naturally occurring animal: 1·47 mg and additive-permitted meat: 0·085 mg), with most values falling within the limits of agreement. The observed agreement between methods supports the use of the calibrated nitrate variables as adjusted exposure estimates in subsequent analyses.
Table 2 presents multiple linear regression results of calibrated nitrate intake from plant, naturally occurring animal and additive-permitted meat sources with inflammatory markers CRP and Lp-PLA2 and CVD risk factors.
Multiple linear regression between nitrate from plant and animal sources and CRP and Lp-PLA2 and cardiovascular risk factors
DBP, diastolic blood pressure; hsCRP, high sensitivity C-reactive protein; Lp-PLA2, lipoprotein-associated phospholipase A2; SBP, systolic blood pressure.
Beta coefficients represent the change in the outcome variable per 1 standard deviation increase in nitrate intake.
Model 1 adjusted for sex and age. Model 2 adjusted for sex, age, physical activity, waist circumference, socio-economic index and alcohol consumption. *Model 2 adjusted for sex, age, physical activity, socio-economic index and alcohol consumption †Model 3 adjusted for the covariates in Model 2 plus intakes of red meat, processed meat, poultry, dairy, eggs, seafood, sugar and confectionary, soft drinks and refined grains. ‡Model 3 adjusted for covariates in Model 2 plus intakes of wholegrains, refined grains, vegetables, nuts, seeds, fruits, sugar and confectionary, soft drinks and tea and coffee.
Bolded P-values denotes statistical significance.
Plant-sourced nitrate was inversely associated with LDL in model 1 (β = −0·215, 95 % CI −0·398, −0·023, P = 0·029), model 2 (β = −0·191, 95 % CI −0·369, −0·004, P = 0·045) and model 3 (β = −0·245, 95 % CI −0·435, −0·044, P = 0·017). For model 2, this corresponds to a 0·21 mmol/l decrease in LDL per 1 sd increase in nitrate intake (56·75 mg/d, ∼40 g spinach). However, in sensitivity analyses adjusting for total energy intake (rather than controlling for food groups), the association between plant-sourced nitrate and LDL-cholesterol was attenuated and no longer statistically significant (online Supplementary Table 2).
In analysis by vegetable type (online Supplementary Table 1), total vegetable-sourced nitrate was inversely associated with LDL-cholesterol in model 1 (β = −0·217, 95 % CI 0·409, −0·204; P = 0·040), model 2 (β = −0·196, 95 % CI −0·382, −0·009; P = 0·040) and model 3 (β = −0·248, 95 % CI −0·450, −0·044; P = 0·018). Leafy greens and beetroot-sourced nitrate were inversely associated with LDL-cholesterol in model 1 (β = −0·210 95 % CI −0·402, −0·016; P = 0·034) and model 3 (β = −0·246, 95 % CI −0·450, −0·040; P = 0·020) and with TAG in model 1 only (β = −0·214, 95 % CI −0·414, −0·014; P = 0·036). Consumption of plant-sourced nitrate was not associated with waist circumference, HDL, TAG, blood pressure, Lp-PLA2 or hsCRP.
There were no associations between naturally occurring animal-sourced nitrate and any CVD risk factor or Lp-PLA2 or hsCRP (e.g. in model 2, all regression coefficients were small in magnitude, including waist circumference (β = −0·039 (–0·230, 0·151)), HDL (β = −0·107 (–0·261, 0·048)), LDL (β = 0·064 (–0·119, 0·246)), TG (β = 0·074 (–0·105, 0·253)), SBP (β = 0·067 (–0·096, 0·229)), DBP (β = 0·130 (–0·043, 0·302)), Lp-PLA₂ (β = −0·058 (–0·251, 0·135), and hsCRP (β = −0·019 (–0·085, 0·065)), with all 95 % CIs crossing the point of null effect (Table 2)).
Additive-permitted meat-sourced nitrate was inversely associated with HDL in model 1 (β = −0·323, 95 % CI −0·501, −0·145, P = 0·001), model 2 (β = −0·208, 95 % CI −0·362, −0·054, P = 0·009) and model 3 (β = −0·205, 95 % CI −0·363, −0·047; P = 0·012) and positively associated with waist circumference in model 2 (β = 0·192, 95 % CI 0·005, 0·380; P = 0·042). For model 2, this corresponded to a 0·10 mmol/l decrease in HDL-cholesterol and a 1·29 cm increase in waist circumference per 1 sd increase in nitrate (∼0·08 mg/d, ∼2·4 g bacon). These associations remained statistically significant in sensitivity analysis controlling for energy intake (online Supplementary Table 2).
Additive-permitted meat-sourced nitrate was positively associated with TAG (β = 0·240, 95 % CI 0·051, 0·429; P = 0·013) and diastolic blood pressure (β = 0·259, 95 % CI 0·076, 0·442; P = 0·006) in model 1 only and was not associated with LDL-cholesterol, systolic blood pressure, Lp-PLA2 or hsCRP.
Discussion
This study examined the relationship between dietary nitrate intake and Lp-PLA2 and hsCRP and CVD risk factors in Australian adults. The results suggest that the health effects of nitrate may depend on its source, with plant-sourced nitrate associated with beneficial effects, while additive-permitted meat-sourced nitrate was linked to adverse outcomes. Specifically, the results showed lower LDL-cholesterol levels to be associated with plant sources and lower HDL-cholesterol and higher waist circumference with additive-permitted meat, but not naturally occurring animal sources. In sensitivity analyses adjusting for total energy intake, rather than controlling for dietary food groups, the association between plant-sourced nitrate and LDL-cholesterol was attenuated and no longer statistically significant, suggesting that this relationship may be partly explained by overall dietary intake rather than nitrate alone. In contrast, associations between additive-permitted meat-sourced nitrate and both lower HDL-cholesterol and higher waist circumference remained statistically significant after energy adjustment, indicating that these findings were not explained by overall energy intake alone. Contrary to expectations, nitrate from any source was not associated with either inflammatory biomarker, with no relationships observed for hsCRP or Lp-PLA₂.
This exploratory study found that ∼57 mg/d of plant-sourced nitrate (equivalent to∼40 g serve of spinach) was associated with 0·21 mmol/l lower LDL-cholesterol in the primary models. However, several important limitations temper this interpretation. First, our relatively small sample size (n 100) limits statistical power and generalisability. Second, the association was marginally significant (P = 0·045) and, critically, was attenuated and no longer statistically significant in sensitivity analyses adjusting for total energy intake rather than food groups. This suggests the finding may not be robust to alternative modeling approaches and should be considered hypothesis-generating rather than conclusive. Nevertheless, our findings align with broader research reporting links between plant-sourced nitrate and improved cardiovascular outcomes. A meta-analysis reported that nitrate consumption, primarily from beetroot juice or leafy greens has been associated with improvements in blood pressure, endothelial function, arterial stiffness and platelet aggregation(Reference Pinaffi-Langley, Dajani and Prater31) and higher habitual vegetable-sourced nitrate intake has been associated with lower CVD incidence and mortality risk(Reference Bondonno, Pokharel and Bondonno28,Reference Bondonno, Dalgaard and Blekkenhorst32) . However, evidence specifically examining plant-sourced nitrate effects on LDL-cholesterol remains limited and mixed(Reference Darabi, Siervo and Webb33–Reference Basaqr, Skleres and Jayswal35).
The observed associations in our study could reflect the effects of other bioactive components in vegetables (e.g. vitamin C, fibre, antioxidants and/or polyphenols) or broader dietary characteristics of individuals consuming plant-rich diets, including lower saturated fat intake and higher overall diet quality, rather than nitrate alone. However, experimental evidence from trials showing that vascular benefits of nitrate-rich vegetables are attenuated when the nitrate–nitrite–NO pathway is disrupted(Reference Bondonno, Liu and Croft36,Reference Kapil, Haydar and Pearl37) suggests nitrate itself may play a contributory role. The relative contribution of nitrate compared with other bioactive components cannot be determined from the present observational study and requires further investigation in adequately powered studies with robust methodology.
The finding that there were no significant associations between plant-sourced nitrate and Lp-PLA2 or CRP in the current study should be interpreted cautiously given our small sample size. With only 100 participants, we were likely underpowered to detect modest associations with these inflammatory markers, even if they exist. A stronger relationship with Lp-PLA2 was anticipated, given its role as a vascular-specific marker of inflammation and evidence suggesting plant-sourced nitrate modulates inflammatory pathways(Reference Raubenheimer, Bondonno and Blekkenhorst14). Plant-sourced nitrate improves endothelial function by increasing NO, which inhibits platelet function and reduces oxidative stress and inflammation, processes that may lower Lp-PLA2 levels and increase plaque stability(Reference Pinaffi-Langley, Dajani and Prater31). In contrast, Lp-PLA2 promotes vascular inflammation by activating platelets and contributing to plaque instability through the production of oxNEFA and lysoPC, two proinflammatory bioactive products(Reference English, Mayr and Lohning38). In a related study with overlapping authors, Lp-PLA2 was inversely associated with cruciferous vegetables, known to be rich sources of dietary nitrate, in unadjusted models, however, associations disappeared after controlling for confounders(Reference English, Jones and Lohning39).
The current study found no associations between naturally occurring animal-sourced nitrate and inflammatory markers or CVD risk factors, despite considerable evidence linking red meat to adverse cardiovascular outcomes. A systematic review reported a 23 % higher cardiovascular mortality risk for adults with the highest red and processed meat consumption and a 1·8 % increased risk per additional 10 grams per day(Reference Bhandari, Liu and Lin40). However, this review did not analyse red and processed meats separately nor investigate nitrate-specific effects. Participants in the current study reported red meat intakes that were within the dietary guideline limit of 455 grams per week(41) (mean 0·97 (sd 0·92) serves/d). The lack of observed associations aligns with expectations given that participants were consuming red meat within recommended amounts. This finding suggests that adherence to dietary guidelines for red meat intake may mitigate potential risk associated with red meat consumption. Additionally, the inclusion of other animal sources, such as fish and poultry, within the naturally occurring animal-sourced nitrate group, may have further moderated the associations specific to red meat.
A Danish study found higher intakes of both naturally occurring animal and additive permitted meat-sourced nitrate to be independently associated with increased CVD-related mortality(Reference Bondonno, Pokharel and Bondonno28), highlighting the potential cardiovascular risk of nitrate sourced from processed meat. Similarly, a study in Greece found that nitrate from processed meat was associated with increased diastolic blood pressure(Reference Kotopoulou, Zampelas and Magriplis12), suggesting its potential adverse effects on cardiovascular health. Potential evidence of processed meat nitrate affecting lipid levels can be seen in a recent cross-sectional study, which found that processed meat consumption was significantly associated with lower HDL-cholesterol in overweight and obese women(Reference Zandvakili, Shiraseb and Hosseininasab42).
The current study adds to this body of evidence, with a key finding that ∼0·08 mg/d of additive permitted nitrate (equivalent to ∼2·4 g bacon) was associated with 0·10 mmol/l lower HDL-cholesterol. While these results suggest a potential role for nitrate in the modulation of HDL metabolism, other components in processed meats, such as saturated fat, polycyclic aromatic hydrocarbons or advanced glycation end products, may have also contributed to the observed associations. Further research is needed to elucidate the exact mechanisms linking processed meat consumption with lower HDL-cholesterol.
Strengths of this study include the strict exclusion criteria to prevent confounding from medication and supplement intake, smoking and existing CVD on the reported vascular and lipid and inflammatory biomarker levels. In addition, certain ethnicities known to have lower levels of Lp-PLA2 due to genetic polymorphisms were excluded to allow a more uniform sample for analysis. Diet was assessed using a validated FFQ calibrated against 3-d diet diaries, and nitrate intake from food was sourced from a national database.
There were several limitations to this study. The favourable associations observed with plant-sourced nitrate may reflect the effects of other bioactive components in vegetables, such as fibre, micronutrients, antioxidants or polyphenols, rather than nitrate alone. Similarly, observed associations between additive permitted meat-sourced nitrate intake and adverse cardiovascular risk factors may not be solely attributable to nitrate content. Other components of processed meats, such as Na, saturated fat or pro-inflammatory by-products of food processing such as polycyclic aromatic hydrocarbons and advanced glycation end products, may have contributed to these findings. Future research is needed to disentangle the specific effects of nitrate from the broader dietary matrix.
The cross-sectional design limits causal inferences, and the sample size may reduce power to detect associations with certain nitrate groups. Additionally, FFQ may overestimate intake of nitrate-rich foods, such as vegetables(Reference Michels, Welch and Luben43). Moreover, the FFQ did not capture cooking methods, which could influence nitrate levels and potentially impact the accuracy of dietary nitrate estimates(Reference Alexander, Benford and Cockburn44). The calibration of FFQ nitrate values using 3-d diet diaries helped reduce potential bias in intake estimates, partially addressing the known limitations of FFQ-based dietary assessment.
We did not measure plasma or urinary nitrate/nitrite. These biomarkers primarily reflect recent intake and endogenous NO metabolism and are not considered robust markers of habitual nitrate intake in free-living populations(Reference Pannala, Mani and Spencer45). We prioritised detailed, source-specific dietary assessment using food frequency derived nitrate estimates calibrated against 3-d food diaries. In addition, nitrite intake was not calculated due to the much-limited data availability as compared with nitrate values, and the limited data coverage would have introduced substantial missing data and reduced the reliability of the estimates in this cohort.
An important consideration is the potential impact of nitrate-reducing bacteria on nitrate metabolism. Nitrate-reducing bacteria in the enterosalivary pathway are crucial for converting dietary nitrate into nitrite, a NO precursor. The use of broad-spectrum antibiotics or antibacterial mouthwash(Reference Demmer, Jacobs and Singh46), which were not recorded or controlled for in this study, can significantly reduce nitrate-reducing bacteria, potentially impairing NO production and influencing cardiovascular outcomes. Future research should address these factors to clarify their impact.
Conclusion
This study supports emerging evidence that the health effects of dietary nitrate may be dependent on their source. Plant-based sources of nitrate were significantly associated with lower LDL-cholesterol, although this association was attenuated in sensitivity analyses adjusting for total energy intake. Naturally occurring animal-sourced nitrate was not associated with inflammatory markers or CVD risk factors. In contrast, additive-permitted meat-sourced nitrate was associated with lower HDL-cholesterol and a higher waist circumference. Future well-designed studies are needed to confirm these findings and further explore the relationship between nitrate source and lipid outcomes.
Supplementary material
BJN- For supplementary material/s referred to in this article, please visit https://doi.org/10.1017/S0007114526106680
Acknowledgements
We acknowledge Hiu Yee Liu who provided research assistant support to analyse the 3-d diet diaries, and Courtney Brackenrig and Aubane Ville who were involved in the preliminary calculation of nitrate food groups and Mark Jones for providing statistical guidance on the analytical approach used in this study.
C. E. was supported by an Australian Government Research Training Program Scholarship.
C. E., C. P. B. and D. P. R. conceived the study, C. E. and D. P. R. collected the data, C. E. and A. L. performed the laboratory analyses, L. Z. created the nitrate database and assigned nitrate content to foods, C. E. performed the data analysis, C. E., N. P. B., C. P. B. and D. P. R. analysed the dietary data and C. E. wrote the initial draft of the manuscript. All authors interpreted the data and critically reviewed and approved the final manuscript.
The authors declare no conflicts of interest.
Data described in the manuscript, code book and analytic code will be made available upon request pending application and approval. Requests should be emailed to the corresponding author.
Open access
