Although compelling, no clear-cut links between dietary factors and the risk of developing rheumatoid arthritis (RA) have been confirmed. Recent investigations have implicated increased consumption of red meat as a risk factor for incident inflammatory arthritis. Here, Benito-Garcia et al (Arthritis Research & Therapy 2007; 9: R16) investigate the risk of dietary protein intake and iron consumption on incident RA in women enrolled in the Nurses’ Health Study (NHS).
Methods
Subjects participating in the NHS (a geographically diverse cohort study of 121,700 female nurses begun in 1976) had regular assessments of diet using the Food Frequency Questionnaire (FFQ), a self-reported inventory of the types and amounts of typical dietary intake. Those without RA at enrollment were followed longitudinally to assess the putative effect of dietary intake of proteins (including animal and vegetable protein) and iron (dietary and supplemental) on the risk of developing RA.
Results
Data from 79,193 women were included with a mean age of 56 years. High consumption of protein and heme iron (from animal sources) was associated with an elevated body mass index (BMI). Women with low intake of protein and high intake of heme iron were more likely to smoke and less likely to have breastfed.
In analyses adjusted only for age, higher protein intake was associated with RA risk; however, the risk was not significant when adjusting for established or potential risk factors for incident RA (BMI, smoking, and total lifetime breastfeeding history). Likewise, no significant increased RA risk was identified for high consumers of total meat, red meat, poultry, fish, or dietary or supplemental iron compared to low consumers in age-adjusted or fully adjusted multivariable models.
Conclusions
Incident RA risk was not increased in high consumers of protein, iron, or meat compared to low consumers.
Editorial Comment
These results contrast with a smaller prospective investigation (Arthritis Rheum 2004; 50(12):3804) from the European Prospective Investigation of Cancer in Norfolk (UK), in which higher consumers of protein and meat demonstrated an increased risk of developing inflammatory polyarthritis (including RA, psoriatic arthritis, or post-infectious inflammatory arthritidies) compared to lower consumers. There are differences in the two investigations that may account for the contrasting findings. For one, the present study involved only female nurses from geographically diverse regions of the United States, while the former involved men and women from the same region of the UK, a population in which the association between protein consumption and polyarthritis may be elevated for unidentified reasons. In addition, female nurses may be a more health conscious group, and extremes of dietary intake may not be comparable between the two populations. Finally, the present study includes only incident RA, while a variety of inflammatory polyarthritides were included in the former study. Methodologically, the present study may be more convincing due to its larger sample size, accounting for an important confounder of incident RA not adjusted for in the earlier study (BMI), and longer follow-up.
