Normal fasting plasma glycemia and risk of type 2 diabetes - Review of the evidence
In normoglycemic adults, fasting glycemia values between 91 and 99 mg/dL were associated with a double risk of type 2 diabetes compared to lower values. Independent cohorts confirm an increasing gradient, but these observational studies do not prove that lowering such values directly prevents the disease.
- Published 12/11/2026
- Updated 12/11/2026
- Article history
ABSTRACT
The analyzed study evaluated whether differences in fasting plasma glycemia, even when below 100 mg/dL and therefore not diagnostic for diabetes, were associated with the subsequent development of type 2 diabetes. In an Italian retrospective cohort of 13,845 adults aged 40 to 69, followed for an average of approximately seven to eight years, the initial glycemia range between 91 and 99 mg/dL was associated with a higher incidence of diabetes compared to the 51-82 mg/dL range. After adjusting for available metabolic and behavioral factors, the hazard ratio—the ratio of the rate of disease onset in the compared groups—was 2.03.
The research contributes to the passage from “The Choice to Live in Health or Illness” by Dr. A. Colonnese and R. Panzironi, according to which the high range of normoglycemia can represent a prognostic indicator, i.e., associated with the probability of a future clinical event. Independent studies in the United States, Japan, and China confirm that future risk increases progressively even before the diagnostic threshold. However, observational cohorts do not prove that glycemia is the sole cause of diabetes, nor that intervening exclusively on this value necessarily changes the clinical outcome. A non-linear relationship, with elevated risk even at very low values in a Japanese cohort, also suggests not interpreting the data as a simple linear scale applicable indiscriminately to every person.
Analyzed study
Normal fasting plasma glucose and risk of type 2 diabetes
DOI: 10.2337/dc10-2263
PMID: 21498787
PMCID: PMC3114342
The research question was whether, among people without diabetes and with fasting glycemia below 100 mg/dL, values in the higher end of the normal range were associated with a higher probability of developing type 2 diabetes in subsequent years. The hypothesis was that the risk does not begin abruptly only beyond a diagnostic threshold, but increases even within the normoglycemic range.
This question is directly relevant to section 2.4.1 of “The Choice to Live in Health or Illness” by Dr. A. Colonnese and R. Panzironi. The book does not use the publication to claim that a glycemia of 91-99 mg/dL is equivalent to diabetes, but to support a more specific point: in the studied population, this range was associated with an increased future risk of diabetes compared to lower glycemic values.
As a retrospective observational study, the research can identify associations and estimate their magnitude after statistical adjustment for certain known factors. However, it cannot prove that the glycemic value, on its own, causes diabetes, nor can it establish whether pharmacologically reducing or modifying the lifestyle to lower a single glycemic value in normoglycemic individuals prevents the disease.
Materials and methods of the original study
The authors used retrospective data from the laboratory of the Desio Hospital in Lombardy. The study included 13,845 individuals—8,110 women and 5,735 men—aged between 40 and 69 years. At the time of inclusion, all participants had a fasting plasma glycemia level of less than 100 mg/dL. To be selected, they also had to have at least three additional glycemia measurements between 1992 and 2008, as well as blood lipid data.
Individuals with requests for glycated hemoglobin—a test that reflects average glucose exposure over the preceding months—were excluded with the intent of reducing the inclusion of subjects already being evaluated for possible diabetes. Type 2 diabetes was defined by two fasting glycemia measurements greater than 125 mg/dL. Participants were divided into three initial ranges: 51–82, 83–90, and 91–99 mg/dL.
Glycemia was measured using an enzymatic-colorimetric method within two hours of blood collection. The authors employed Cox models, a statistical analysis that compares the rate at which an event occurs over time, to calculate hazard ratios with 95% confidence intervals. In a random subgroup of 3,593 people, additional data obtained via questionnaire were available, including body mass index, hypertension, family history of diabetes, smoking, and alcohol consumption. The final model adjusted for these factors was therefore conducted only within this subgroup.
Results
During 108,061 person-years of observation, which is the sum of the follow-up periods for all participants, 307 incident cases of type 2 diabetes occurred. The mean follow-up was 7.9 years for women and 7.4 years for men. The incidence increased as initial glycemia values rose, even while remaining below 100 mg/dL.
In women, the percentage of new cases increased from 0.75% in the 51-82 mg/dL range to 3.37% in the 91-99 mg/dL range. In men, it increased from 0.58% to 4.08%. In other words, more new cases of diabetes were observed during follow-up in the highest range of normoglycemia compared to the lowest range.
After adjusting for age, participants in the 91-99 mg/dL range showed a hazard ratio of 2.89 for women and 2.87 for men compared to the 51-82 mg/dL range. Further adjustment for triglycerides and cholesterol had little effect on these estimates. In the final model, which also included body mass index, hypertension, family history, smoking, and alcohol in the subgroup with complete data, the 91-99 mg/dL range maintained a hazard ratio of 2.03, with a 95% confidence interval from 1.18 to 3.50.
This indicates that, in the available data, the 91-99 mg/dL range was associated with a rate of diabetes onset approximately double that of the lowest glycemic range. For the intermediate range, the final model reported a hazard ratio of 1.42 with a confidence interval from 0.42 to 4.74: a very wide interval that includes 1, and is therefore also compatible with the absence of a statistically clear difference. The text also contains a discrepancy between the intermediate range initially defined as 83-90 mg/dL and the one reported in the final model as 86-90 mg/dL.
Interpretation
In simple terms, the study suggests that fasting glycemia still below the diagnostic threshold for prediabetes may contain information about future metabolic risk. The result does not imply that every person with 91-99 mg/dL will develop diabetes. Instead, it indicates that, in the observed cohort, this range was associated with a higher future frequency of diabetes compared to lower values.
The segment of the book's narrative that is effectively documented is therefore: fasting glycemia in the high range of normoglycemia → higher future risk of type 2 diabetes. The study does not directly measure insulin sensitivity—that is, the ability of tissues to respond to insulin—nor the function of pancreatic beta cells, which produce insulin. It may therefore be consistent with early metabolic alterations, but it does not demonstrate which biological mechanism produced the association.
Adjustment for certain risk factors strengthens the prognostic interpretation of the glycemic value, but does not completely eliminate the possible effect of unmeasured or incompletely measured factors. The limitations of the selection are also relevant: these are people who had repeated laboratory tests, came from a specific geographic area, and had detailed clinical data available for only a portion of the sample. The result should therefore be read as observational evidence of risk stratification, not as an independent diagnostic or therapeutic indication.
Independent supporting evidence
Direct confirmation of this same narrative comes from a U.S. cohort of 46,578 adults without diabetes and with an initial glycemia of less than 100 mg/dL. After adjusting for age, sex, body mass index, blood pressure, lipids, smoking, and other conditions, each 1 mg/dL increase in fasting glycemia was associated with a 6% increase in the rate of diabetes diagnosis. Compared to values below 85 mg/dL, the 95-99 mg/dL range was associated with a hazard ratio of 2.33, with a 95% confidence interval from 1.95 to 2.79. This study directly confirms the passage in the book stating that the high end of normoglycemia signals a higher future risk [1].
Another cohort, consisting of 13,163 Israeli men aged between 26 and 45 years and followed from 1992 to 2004, observed a progression of risk starting as early as glycemia levels of 87 mg/dL or higher, compared to the quintile with values below 81 mg/dL. The model was adjusted for age, family history, body mass index, physical activity, smoking, and triglycerides. This evidence is particularly useful because it shows the same direction of association in younger adults and in a different geographical context; however, it remains limited to the male population [2].
In the Panasonic cohort, 37,148 Japanese workers who were initially normoglycemic were followed from 2008 to 2018, and 1,028 developed diabetes. Between 90 and 99 mg/dL, each 1 mg/dL increase was associated with a 9.0% increase in the risk of developing diabetes. Compared to a glycemia of 89 mg/dL or lower, the hazard ratios increased progressively up to 3.41 for the 98-99 mg/dL range. The gradient was also present in people with obesity, suggesting that the association did not depend exclusively on excess weight [3].
A more recent analysis examined two Asian cohorts: 204,640 people in China and 15,464 in Japan. The relationship between initial normal glycemia and future diabetes did not appear perfectly linear, but rather J-shaped. The risk increased more rapidly beyond inflection points estimated at around 81 mg/dL in the Chinese cohort and 94 mg/dL in the Japanese one. This result confirms the book's central idea—that risk can increase before diagnostic thresholds—but clarifies that the shape of the relationship may vary between populations [4].
Conflicting evidence
No sufficiently relevant directly conflicting evidence has been identified that denies the association between fasting glycemia in the high-normal range and an increased future risk of diabetes. However, there is data that complicates the idea of a linear trend in which every lower value is necessarily associated with lower risk.
In a Japanese national cohort of 186,749 people without diabetes, followed for three years, the risk of new-onset diabetes was found to be increased both with glycemia levels between 90 and 125 mg/dL and with values below 70 mg/dL, compared to the 85-89 mg/dL range. This finding does not contradict the increased risk in the high-normal range, but it shows a J-shaped relationship and suggests that very low values may identify different biological or clinical conditions [5].
This observation requires caution. Fasting glycemia is an indicator influenced by actual fasting, acute conditions, liver function, medications, body composition, and analytical variability. Furthermore, differences between studies in the definition of diabetes, participants, duration of follow-up, and variables considered can alter risk estimation. Therefore, the glycemic value must be interpreted together with the overall clinical picture, not as an isolated measure capable of defining an individual's metabolic fate on its own.
Summary table of evidence
The table distinguishes the study under review from independent research that confirms or provides context for the passage in the book.
| Reference | Study type | Population or model | Main result | Direction of evidence | Strength of evidence |
|---|---|---|---|---|---|
| Original study | Retrospective cohort | 13,845 Italian adults, 40-69 years old | Glycemia of 91-99 mg/dL associated with an HR of 2.03 for diabetes compared to 51-82 mg/dL in the subgroup with complete data | direct confirmation | moderate |
| [1] | Retrospective observational cohort | 46,578 normoglycemic US adults | Each 1 mg/dL associated with an HR of 1.06; glycemia of 95-99 mg/dL associated with an HR of 2.33 compared to less than 85 mg/dL | direct confirmation | moderate |
| [2] | Observational cohort | 13,163 Israeli men, 26-45 years old | Progressively higher risk from 87 mg/dL compared to values below 81 mg/dL | direct confirmation | moderate |
| [3] | Retrospective cohort | 37,148 normoglycemic Japanese workers | Between 90 and 99 mg/dL, each 1 mg/dL increase associated with a 9% increase in diabetes risk | direct confirmation | moderate |
| [4] | Two observational cohorts | 204,640 people in China and 15,464 in Japan | J-shaped relationship; more rapid risk increase above approximately 81 mg/dL in China and 94 mg/dL in Japan | indirect confirmation | moderate |
| [5] | Observational cohort | 186,749 Japanese adults without diabetes | Increased risk both at 90-125 mg/dL and below 70 mg/dL compared to 85-89 mg/dL | discordant evidence | moderate |
Confirmation of the evidence
Claim in the book
In section 2.4.1, “The Choice to Live in Health or Illness” by Dr. A. Colonnese and R. Panzironi, it is stated that small differences in fasting glycemia, even within the range commonly considered physiological, can be associated with metabolic alterations and clinically relevant consequences. Specifically, it reports that values of 91-99 mg/dL, in a Mediterranean population of nearly 14,000 people, were associated with an approximately double risk of diabetes compared to lower glycemia levels and constituted a significant prognostic indicator.
Fasting glycemia in the high-normal range → observed higher risk of type 2 diabetes → prognostic indicator of future metabolic risk
The publication subject to the Evidence Review was used in the book to support the following passage: in initially normoglycemic Mediterranean adults, a fasting glycemia of 91-99 mg/dL was associated with an approximately double subsequent risk of type 2 diabetes compared to lower glycemic values.
The correspondence between the original study and the narrative in “The Choice to Live in Health or Illness” by Dr. A. Colonnese and R. Panzironi is specific and direct. The publication analyzed does not demonstrate the entire biological sequence leading to diabetes, nor does it identify with certainty the individual metabolic alterations already present. However, it documents the fundamental prognostic shift: in the Italian cohort, the 91-99 mg/dL range was associated with a higher future incidence of diabetes.
Independent evidence converges on the same point. The U.S. cohort by Nichols and colleagues observed an increased risk for every 1 mg/dL increment and a hazard ratio of 2.33 for the 95-99 mg/dL range compared to values below 85 mg/dL [1]. The Israeli cohort detected a progressive increase in risk starting from 87 mg/dL in young men [2]. The Panasonic cohort confirmed an increasing gradient between 90 and 99 mg/dL, with a 9% increase in risk for every mg/dL and maximum hazard ratio values exceeding three in the 98-99 mg/dL range [3].
The two cohorts analyzed by He and colleagues expand the narrative, showing that the relationship may be non-linear and differ between populations, while still maintaining an increased risk before diagnostic thresholds [4]. Finally, the cohort by Ogata and colleagues reminds us that the relationship should not be reduced to a straight line: values below 70 mg/dL were also associated with a higher risk compared to the 85-89 mg/dL range [5]. Overall, the literature confirms that the high end of normoglycemia can function as a prognostic indicator, but it does not automatically transform a single value into a disease diagnosis or proof of causality.
Overall level of confirmation
STRONG CONFIRMATION
This finding is directly supported by the original study and by multiple large, independent cohorts conducted in different populations. The strength of the evidence is limited by its observational nature, the potential presence of confounding factors, and the variability in the shape of the relationship between glycemia and risk across different populations.
Conclusions
Overall assessment
The narrative is well-supported as a prognostic association: fasting glycemia values in the high-normal range are consistently associated with a higher future risk of type 2 diabetes. The confirmation is strong due to the consistency across independent cohorts, but it remains observational; it does not prove causality nor does it justify diagnostic interpretations based on a single value.
References
[3] Fasting plasma glucose level in the range of 90–99 mg/dL and the risk of the onset of type 2 diabetes: Population‐based Panasonic cohort study 2
[4] J-shape relationship between normal fasting plasma glucose and risk of type 2 diabetes in the general population: results from two cohort studies
[5] Low fasting plasma glucose level as a predictor of new-onset diabetes mellitus on a large cohort from a Japanese general population
Editorial note
Artificial Intelligence was used solely as a tool to support editorial and documentary work. All content, interpretations, bibliographic citations and conclusions have been verified, reviewed and approved by the author, who assumes full scientific responsibility for them.
The manuscript was also subject to independent peer review prior to publication.
To ensure maximum transparency, the article's complete history — including the dates of submission, review, acceptance and publication, as well as information on the peer review process — is available on the dedicated Article History page.