Background A comprehensive description of disease biomarker levels in people of different habitual diet groups is lacking. We conducted cross-sectional analyses of mean biomarker concentrations by diet group in a large cohort.

Methods The UK Biobank recruited around 500,000 middle aged participants throughout the United Kingdom in 2006–2010. Blood and urine samples were collected from the majority of participants, and assayed for a range of serum and urinary biomarkers related to disease status of six outcomes (cardiovascular diseases, bone and joint health, cancer, diabetes, renal disease, and liver). Using multivariable linear regression adjusted for age, sex, fasting status, body mass index and lifestyle confounders, we estimated geometric mean biomarkers concentrations by six diet groups (221,295 regular meat-eaters, 222,038 low meat-eaters, 5,053 poultry eaters, 10,470 fish eaters, 6,804 vegetarians, 416 vegans) in white British participants, and two diet groups in British Indian participants (4091 meat eaters, 1444 vegetarians).

Results We observed differences in the concentrations of many biomarkers by diet group. The biomarkers with the largest percentage difference by extreme diet groups (i.e. vegans versus regular meat-eaters) within each disease outcome group in the white British population are reported below. Compared with white British regular meat-eaters, white British vegans had lower C-reactive protein (1.10, 1.00–1.21 versus 1.43, 1.43–1.44 mg/L) and low density lipoprotein cholesterol (3.12, 3.06–3.19 versus 3.65, 3.65–3.65 mmol/L); lower vitamin D (34.3, 33.0–35.8 versus 44.5, 44.4–44.5 nmol/L); higher sex hormone-binding globulin (51.1, 48.9–53.3 versus 45.0, 44.9–45.0 nmol/L); lower haemoglobin A1C (HbA1C, 33.8, 33.4–34.2 versus 35.2, 35.2–35.2 mmol/mol); lower urinary creatinine (5389, 5076–5723 versus 7289, 7269–7308 µmol/L); and lower gamma glutamyltransferase (23.4, 22.1–24.7 versus 29.7, 29.6–29.8 U/L). Patterns were similar in British Indians, with the exception of HbA1C which was not significantly different between meat-eaters and vegetarians. In both ethnicities, the differences in biomarker levels by diet group were consistent between men and women.

Conclusion In this large population cohort, participants of different diet groups exhibited differences in many biomarkers. These biomarkers are associated with disease risk, and therefore the observed differences may be suggestive of differences in future disease risks by diet group, which should be further investigated.