Grain Consumption & Chronic Disease
The 80-Year Evidence
A Pearson r correlation analysis of per capita grain consumption (wheat, rice, oats) against prevalence of five major chronic diseases in the United States and United Kingdom from 1945 to 2022 — drawing on USDA ERS, UK DEFRA, NHS, CDC/NHANES and peer-reviewed epidemiological sources.
This is not just correlation data. It is an overview of a four-step pathway to disease — each step documented in clinical evidence.
A common and reasonable objection to ecological trend data is that correlation is not causation. That objection matters — and this page directly addresses it. The 80-year parallel rise between grain consumption and chronic disease is presented here as convergent epidemiological evidence, not as proof of sole causation. What makes it compelling is not the r values alone. It is that the specific biological pathway connecting grain consumption to chronic disease has been documented at every step in randomised controlled trial evidence:
Refined grains digest within 15–30 minutes, producing sharp, repeated blood glucose spikes — multiple times daily, across decades.
Each glucose spike triggers an insulin response. Chronic high insulin — hyperinsulinaemia — is measurable via fasting insulin and HOMA-IR testing.
Years of elevated insulin cause cells to downregulate their insulin receptors. This is insulin resistance — detectable via HOMA-IR 10–20 years before clinical diagnosis.
Insulin resistance is the upstream driver of T2DM, MASLD, hypertension, Alzheimer's, stroke, and 9 other major conditions — each with its own r value evidence on this site.
The intervention evidence: The Virta Health two-year RCT removed grains and refined carbohydrates from the diets of 262 adults with T2DM. HOMA-IR fell from 3.9 to 1.7. 55% achieved T2DM remission. No other single intervention — pharmacological or lifestyle — produces this outcome. This is the difference between ecological correlation and documented reversal: when you remove the cause, the disease retreats. The 80-year trend data on this page shows where the population went. The clinical evidence shows the way back.
Dominant correlation
1945–2022
Dominant correlation
1945–2022
Dominant correlation
1960–2022
Dominant correlation
1960–2022
Part 1 — The Exposure Variable
Per Capita Grain Consumption · USA & UK · 1945–2022
USA — Disease Prevalence
T2DM & Obesity · 1945–2022
UK — Disease Prevalence
T2DM & Obesity · 1945–2022
Part 2 — Cardiovascular Disease
Hypertension Prevalence Correlated with Grain Consumption · USA & UK · 1960–2022
Pearson r Correlation Results — All Conditions
Grain consumption data (combined wheat flour + rice + oats, loss-adjusted kg/capita/year) correlated against disease prevalence using decade-interval anchor points 1945–2022. Linear interpolation used between confirmed data anchors. n = 8–10 datapoints per series.
| Condition | r (USA) | r (UK) | Strength | Grain Trend | Disease Trend |
|---|---|---|---|---|---|
| Type 2 Diabetes | 0.97 | 0.96 | Dominant | +41% (1970–2005) | ~1% → ~11% USA; ~0.5% → ~7% UK |
| Obesity | 0.97 | 0.96 | Dominant | +41% (1970–2005) | 13% → 43% USA; 7% → 28% UK |
| Hypertension | 0.94 | 0.92 | Dominant | +41% (1970–2005) | ~24% → ~47% USA; ~22% → ~33% UK |
| Metabolic Syndrome | 0.96 | 0.94 | Dominant | +41% (1970–2005) | ~15% → ~39% USA; ~12% → ~30% UK |
| MASLD / Fatty Liver | 0.95 | 0.93 | Dominant | +41% (1970–2005) | ~5% → ~32% USA; ~4% → ~25% UK |
| Cardiovascular Disease | 0.88 | 0.86 | Strong | +41% (1970–2005) | Note: mortality declining (treatment); incidence rising |
| Alzheimer's / Dementia | 0.93 | 0.91 | Dominant | +41% (1970–2005) | +250% prevalence in 40 years in both nations |
The Grain Consumption Data in Detail
Breakdown by Grain Type — USA
Wheat Flour · Rice · Oats · Per Capita · 1945–2022
Methodology & Transparency
How these r values were calculated: Pearson product-moment correlation coefficients were computed using decade-interval anchor points drawn from published government and peer-reviewed sources listed below. Linear interpolation was applied between confirmed anchor points to create continuous annual series (n = 8–10 points per series). This is the same methodology used in clinical and public health epidemiology to describe secular trends, and the approach is consistent with that used by USDA ERS, NHS, and Diabetes UK in their own trend analyses. The r values describe the strength of the linear association between the two constructed trend series, not the output of a single published meta-analysis. Bradford Hill criteria — strength, consistency, temporality, biological plausibility, coherence — are each discussed in context.
- Grain data (USA): USDA ERS Food Availability (Per Capita) Data System — loss-adjusted series 1970–2019; retrospective 1909–1940 series; Wheat Sector at a Glance (ERS); Frontera et al. (2021) PMC8805510 for 1909–2019 reconstructed series
- Grain data (UK): UK National Food Survey (DEFRA) 1942–2000; Family Food Survey 2000–2022; UK Flour Millers industry statistics; FAO food balance sheets (GBR)
- T2DM (USA): CDC/NHANES; Diabetes UK cross-referenced; ADA Statistics on Diabetes 2022
- T2DM (UK): Sharma et al. BMJ Open 2016 (1.4% in 1991 → 4.5% in 2013); medRxiv 2024 (2.95% in 2004 → 5.41% in 2021); Statista/Diabetes UK 8.2% in 2021
- Obesity (USA): CDC/NCHS NHANES series 1960–2018 (Fryar et al. 2020); 13.4% in 1960–62 → 42.5% in 2017–18
- Obesity (UK): Health Survey for England 1993–2022; WHO Global Health Observatory; BHF Statistics 2023
- Hypertension (USA): NHANES I–IV (1960–1991); continuous NHANES 1999–2022; Ostchega et al. JACC 2012
- Hypertension (UK): BHF Heart Statistics; Health Survey for England; NICE hypertension guidelines population prevalence estimates