🟠 Moderate Evidence
In 2015, researchers at the Weizmann Institute of Science in Israel recruited 800 adults and monitored their blood glucose responses using continuous glucose monitors while they consumed identical meals. The results challenged a fundamental assumption in nutritional science: that the same food produces the same metabolic effect in all people. Despite eating the same 50-gram portion of white bread under identical conditions, participants showed a more than fivefold variation in their glucose response, with the top 10% of responders averaging over 79 mg/dL·h compared to under 15 mg/dL·h for the bottom 10%, measured as incremental area under the glucose curve.
Key takeaways
- Identical carbohydrate meals produce glucose responses that vary more than fivefold between individuals, even among people with similar clinical profiles
- Standard clinical variables like fasting glucose and body mass index explain only part of the variation; microbiome composition, sleep duration, and prior meal content are independently predictive
- These findings suggest that one-size-fits-all glycemic index labels may not accurately predict individual postprandial glucose responses
Study at a Glance
| Source | Weizmann Institute of Science |
| Study type | Observational cohort study |
| Sample size | N = 800 adults |
| Population | Western adults, 54% overweight, 22% obese, 24% prediabetic; none with diagnosed type 2 diabetes |
| Country | Israel |
Glucose Response Variability in 800 Adults Eating Identical Bread
Incremental area under glucose curve (mg/dL·h) for 50g white bread, Weizmann cohort
Source: Weizmann Institute of Science, 2015 | Georgian Medical Journal News
Identical meals, divergent responses
The Weizmann study enrolled 800 participants who were broadly representative of a Western adult population, including 54% who were overweight and 22% who were obese. The researchers gave each participant an identical standardized meal—50 grams of carbohydrate from white bread, same portion, same timing—and measured the postprandial glucose response. The cohort also included 24% with HbA1c levels in the prediabetic range, though none had been formally diagnosed with type 2 diabetes.
The glucose curves that emerged were strikingly different. The average 2-hour incremental area under the glucose curve across the entire cohort was 44 mg/dL·h, but this aggregate figure masked enormous individual variation. The bottom 10% of responders had an average curve area under 15 mg/dL·h, while the top 10% exceeded 79 mg/dL·h—a difference of more than fivefold from the same meal. This wasn’t a matter of measurement error or fringe cases; the variation was substantial and systematic across a representative population.
Standard clinical markers explain only part of the story
Initially, the large variation appeared to align with established metabolic risk factors. People with higher body mass index, higher baseline HbA1c, and higher fasting glucose tended to produce larger glucose spikes in response to the bread. This pattern was predictable and consistent with insulin resistance and reduced beta cell function—metabolic phenomena well documented in diabetes research.
However, this accounted for only a portion of the variation. Even within the subgroup of participants with normal fasting glucose and no obvious metabolic disease, substantial differences in glucose response persisted. Two adults with identical age, fasting glucose, and body composition could produce postprandial glucose curves that bore little resemblance to one another. The researchers then examined variables beyond conventional clinical parameters and found that microbiome composition, sleep duration the night before the test, physical activity around the meal, and the composition of the previous meal were independently predictive of individual glucose response after accounting for standard clinical variables.
The same 50-gram carbohydrate portion produced glucose responses ranging from under 15 to over 79 mg/dL·h—a fivefold spread—even among individuals with similar age, weight, and fasting glucose levels. Microbiome composition and behavioral factors like sleep and prior food intake independently predicted individual variation.
— Zeevi et al., Weizmann Institute of Science (2015)
Implications for nutrition labeling and personalized medicine
These findings have direct relevance to how nutritional information is communicated to the public and to emerging approaches in personalized metabolic health. The glycemic index—a standardized measure that ranks foods by their effect on blood glucose in a reference population—assumes a relatively consistent effect across individuals. The Weizmann cohort demonstrates that this assumption may be significantly oversimplified. Two people following the same meal plan based on glycemic index values might experience vastly different glucose trajectories, potentially affecting satiety, energy, and long-term metabolic health.
The identification of modifiable factors—sleep quality, physical activity timing, and prior dietary composition—suggests that personalized nutrition guidance might benefit from assessment of individual metabolic signatures, potentially including microbiome analysis where relevant. This aligns with the emerging field of precision nutrition, though the cost-effectiveness and clinical utility of such approaches in routine care remain subjects for further research. The findings also highlight a gap between population-level nutritional guidance and individual metabolic reality, a distinction that becomes increasingly important as continuous monitoring technologies become more accessible.
What this means
Frequently asked questions
Does this mean glycemic index is useless?
No. Glycemic index remains a useful population-level tool for comparing foods and making general dietary recommendations. However, the Weizmann findings suggest that individual responses can vary considerably from the population average, and one person’s glucose response to white bread may be quite different from another’s. For people with prediabetes or diabetes, this variability may justify more personalized assessment rather than relying solely on published glycemic index values.
Can I change my microbiome to lower my glucose spikes?
The Weizmann study shows that microbiome composition is associated with glucose response, but the direction of causality (does microbiome composition determine glucose response, or does glucose response reflect microbiome composition?) remains unclear from this observational data. Interventional studies examining whether specific microbiota changes produce measurable shifts in glucose handling are ongoing. In the meantime, established microbiota-supporting practices—adequate fiber intake, diverse plant foods, and adequate sleep—may have benefits independent of glucose response.
What should I do differently based on this research?
The most immediately actionable findings from the Weizmann cohort relate to modifiable factors: prioritizing adequate sleep, timing physical activity around meals, and paying attention to the composition of preceding meals. For individuals concerned about glucose control or at risk for type 2 diabetes, discussing personalized metabolic assessment with a healthcare provider—potentially including short-term continuous glucose monitoring—may provide more useful data than generic nutritional advice. Work with a registered dietitian familiar with personalized nutrition approaches if this level of detail is relevant to your health goals.
As continuous glucose monitoring becomes more widely available and affordable, the gap between population-level nutritional science and individual metabolic reality is likely to narrow. Future research should focus on understanding the mechanisms linking microbiota, sleep, activity, and prior dietary composition to individual glucose responses, and on evaluating whether personalized dietary guidance based on metabolic profiling produces better long-term health outcomes than population-based approaches. Until then, the Weizmann findings serve as a reminder that metabolic responses are highly individual and that nutrition guidance may be most effective when it acknowledges this variability.
Source: Weizmann Institute of Science glucose response study, 2015
Was this article helpful?
Disclaimer. This article is health journalism intended for general information and education. It is not medical advice and is not a substitute for professional diagnosis or treatment. Always consult a qualified healthcare provider about your individual circumstances. Full disclaimer →
Related Coverage
Medically reviewed by Prof. Giorgi Pkhakadze, MD, MPH, PhD. Spotted an error? Contact the editorial team.
