Variations in human body water isotope composition across the United States

Accurate human provenancing using stable isotopes depends directly on solid understandings of the geographic and individual factors affecting isotope variability and incorporation into human tissues. Transfer of isotopic, and therefore spatial, information between environmental water and biological tissues is mediated by the isotopic composition of body water. Thus, there is a need to study body water isotope ratios at a population level and over a large isotopic and geographic range. We evaluated oxygen (δ¹⁸O_bw) and hydrogen (δ²H_bw) isotope values of body water from 72 volunteers in 10 different cities across the US, and over a 5-10-day period. We analyzed covariates (e.g., water intake, physical activity, biometrics, gender) that might explain individual stable isotope ratio variations and tested a predictive model that incorporates the δ-values of drinking water, food, and O₂ as well as individual variables to predict the δ-values of body water. The individual variability in body water isotope values overtime (mean 0.3‰ for δ¹⁸O_bw and 2.3‰ for δ²H_bw) was lower than the intra-city variability (mean 0.9‰ for δ¹⁸O_bw and 6.9‰ for δ²H_bw). Body water isotope values differed among cities (ANOVA: δ¹⁸O_bwF = 97.2, p < 0.001; δ²H_bwF = 176.2, p < 0.001). However, significant overlap among some cities with different drinking water was discovered. We detected significant covariation of measured drinking water and human body water isotope values (both isotope systems R² ≥ 0.89, p < 0.001) and small but significant effects of the average daily exercise and amount of fluid intake. The differences between measured and model-predicted body water values (mean 0.12 ± 1.2‰ for Δδ¹⁸O and -1.2 ± 8.2‰ for Δδ²H) were statistically indistinguishable from zero (Δδ¹⁸O t = -0.751, p = 0.45; Δδ²H t = 1.133, p = 0.26). Here we show that community level variation exists in the δ¹⁸O_bw and δ²H_bw values and the primary drivers are the regional differences in drinking water isotopes. Consistency of the body water isotope composition over the study period suggests that tissues would incorporate a stable isotope signal over time. The amount of drinking water and physical activity influence body water values, while the variation in the isotopic values of food may contribute to regional level variability, but that still remains to be assessed further. The human body water model provides accurate estimates for measured values, capturing and reproducing the main features of the body water isotope variation across space.