A new method for investigating the relationship between diet and mortality: hazard analysis using dietary isotopes

Abstract

Background: The population of Roman Britain are renowned for having elevated nitrogen (δ¹⁵) stable isotope values, which have been interpreted as evidence for the increased consumption of marine products. However, such results are now understood to also reflect episodes of stress and disease, suggesting that new interpretations are warranted.

Aim: To test a novel approach which combines hazard mortality analysis and stable isotope data to determine whether there is a relationship between age-at-death, elevated δ¹⁵N values and mortality risk.

Subjects and methods: This study used published osteological and dietary stable isotope data for nitrogen (δ¹⁵N) and carbon (δ¹³C) of 659 1st–5th century AD individuals aged >12 years old excavated from Roman cemeteries in Britain. The relationship between diet and mortality risk was assessed using the Gompertz hazard model, and differences in median reported isotope values between the sexes was determined using a Mann Whitney test.

Results: It was discovered that higher δ¹⁵N levels are associated with elevated risks of mortality, whereas the opposite pattern was observed for δ¹³C, and males had higher median δ¹³C and δ¹⁵N values.

Conclusion: This study successfully demonstrated that stable isotope data can be integrated into hazard models, allowing one to connect diet and mortality in past populations. It supports the findings of other isotope studies which have established that individuals with childhood stress/trauma will have different isotope patterns.

Keywords:

• Mortality
• diet
• hazard analysis
• stable isotope analysis
• Roman Britain

Acknowledgements

We are most grateful to Sally Brooks at MoL for sourcing many of the reports used in this study, and to Colleen Cummings for allowing us to use her unpublished PhD data. An Audrey Barrie Brown Memorial Fund award made to RCR and CH by the Roman Research Trust enabled the stable isotope research on the Dorset material to be conducted, and we are most grateful to Nancy Beavan for all her work on the Dorset samples. Express thanks are given to colleagues at Dorset County Museum, the Dorset Natural History and Archaeological Society, the Natural History Museum (London), and Pre-Construction Archaeology for permission to sample the collections in their care.

Disclosure statement

No potential conflict of interest was reported by the authors.