A large-scale international study has found that higher consumption of meat, particularly red and processed meat, may be associated with an increased risk of developing type 2 diabetes.
In the study, published in The Lancet Diabetes & Endocrinology, investigators analyzed the data of nearly 2 million adults across 31 cohorts in 20 countries. The investigators utilized a federated meta-analysis approach to harmonize data across diverse populations. Dietary information was collected primarily through food frequency questionnaires (26 cohorts), with some using dietary history (3 cohorts) or dietary records (2 cohorts).
Hazard ratios (HR) were estimated using Cox regression models, adjusted for potential confounders including age, sex, education, smoking, physical activity, alcohol intake, total energy intake, body mass index (BMI), and other dietary factors. Prentice-weighted Cox regression was used for the EPIC-InterAct case-cohort study.
Among the key findings were:
- Consumption of unprocessed red meat was associated with a 10% higher risk of type 2 diabetes per 100 g/day intake (HR = 1.10, 95% confidence interval [CI] = 1.06–1.15, I² = 61%).
- Processed meat consumption showed a 15% higher risk per 50 g/day intake (HR = 1.15, 95% CI = 1.11–1.20, I² = 59%).
- Poultry consumption was linked to an 8% higher risk per 100 g/day intake (HR = 1.08, 95% CI = 1.02–1.14, I² = 68%).
Over a median follow-up of 10 years (interquartile range [IQR] = 7–15), 107,271 incident cases of type 2 diabetes were identified. Median meat consumption varied widely across cohorts:
- Unprocessed red meat: 0–110 g/day
- Processed meat: 0–49 g/day
- Poultry: 0–72 g/day.
Regional variations:
- Americas (12 cohorts): HRs = 1.13 (95% CI = 1.06–1.20) for unprocessed red meat, 1.17 (95% CI = 1.10–1.24) for processed meat
- European region (9 cohorts): HRs = 1.06 (95% CI = 1.04–1.09) for unprocessed red meat, 1.13 (95% CI = 1.07–1.19) for processed meat
- Western Pacific region (7 cohorts): HRs = 1.17 (95% CI = 1.01–1.36) for unprocessed red meat, 1.15 (95% CI = 1.01–1.32) for processed meat
- Eastern Mediterranean region (2 cohorts) and South Asia (1 cohort): Associations not evident.
Sensitivity analyses:
- Models without BMI adjustment showed stronger associations: HRs = 1.18 (95% CI = 1.07–1.29) for unprocessed red meat, 1.23 (95% CI = 1.14–1.34) for processed meat, and 1.21 (95% CI = 1.12–1.31) for poultry
- Using a fixed-effect meta-analysis, the association for poultry was weaker: HR 1.02 (95% CI = 1.00–1.04) per 100 g/day.
Substitution analyses:
- Replacing 50 g/day of processed meat with 100 g/day of unprocessed red meat: HR = 0.93 (95% CI = 0.90–0.97)
- Replacing 50 g/day of processed meat with 100 g/day of poultry: HR = 0.90 (95% CI = 0.82–0.97)
- No significant difference when replacing unprocessed red meat with poultry: HR = 0.98 (95% CI = 0.90–1.97).
Additional findings included:
- The association of unprocessed red meat with type 2 diabetes was 8% weaker in European cohorts compared with American cohorts (P = .022).
- Stronger associations were observed in cohorts with larger numbers of type 2 diabetes cases.
- Log-linear dose-response relationships were found for each meat type, without obvious threshold or ceiling effects.
Methodological considerations:
- 18 of the 31 cohorts had not previously published on this topic, potentially reducing publication bias.
- The NutriGrade scoring system rated the certainty of evidence as high (≥ 8 points) for unprocessed red meat and processed meat, and moderate (7 points) for poultry.
Limitations included potential residual confounding, use of baseline-only dietary data in most cohorts, heterogeneity in data collection methods, and limited representation from certain global regions. The weaker and more heterogeneous association for poultry warrants further investigation.
Declaration of interests can be found in the study.