Relationship of glucose variability with glycated hemoglobin and mean daily glycemia: post hoc analysis of data from 5 phase 3 studies - Evidence review
The study analyzed associates smaller daily glycemic fluctuations with lower glycated hemoglobin and mean daily glycemia. The book excerpt extends this message to a lower risk of hypoglycemia: independent studies make this plausible, especially with continuous monitoring, but they do not prove that postprandial reduction alone causes such benefits.
- Published 13/11/2026
- Updated 13/11/2026
- Article history
ABSTRACT
The original study pooled data from five phase III clinical trials to verify whether glycemic variability, i.e., the magnitude of glucose fluctuations throughout the day or between different days, was linked to the main indicators of metabolic control: glycated hemoglobin and mean daily glycemia. The post hoc analysis found that lower intra-day variability was associated with better values for both indicators in the type 1 and type 2 diabetes groups examined.
In the section dedicated to postprandial glycemic variability, "The Choice to Live in Health or Illness" by Dr. A. Colonnese and R. Panzironi uses this research to argue that attenuating glycemic fluctuations is accompanied by better metabolic control and fewer episodes of hypoglycemia. The first part is directly consistent with the study; the link to a lower risk of hypoglycemia, however, requires external evidence. Meta-analysis and randomized studies with continuous monitoring show that interventions capable of reducing variability, hyperglycemia, and time spent at low values can simultaneously improve these parameters. However, they do not prove that every single reduced postprandial excursion alone produces such effects, nor that the relationship is identical in all people with diabetes.
Analyzed study
Relationship of Glucose Variability With Glycated Hemoglobin and Daily Mean Glucose: A Post Hoc Analysis of Data From 5 Phase 3 Studies
PMID: 29056082
PMCID: PMC5851228
The research question was whether glucose fluctuations were linked to the two most common indicators of overall glycemic control: glycated hemoglobin, which roughly reflects average glucose exposure over the previous months, and mean daily glycemia. The hypothesis was that greater glycemic variability could contribute to less favorable values for these parameters.
The relevance to the book's section therefore concerns a specific segment of the narrative: blood glucose excursions within the same day, including those following meals, may be associated with less stable metabolic control. The publication did not separately measure postprandial variability as a sole causal factor and did not verify whether reducing it through a specific treatment directly reduces hypoglycemia. It can document statistical associations between measures of variability and glycemic control, not an autonomous causal relationship between each post-meal spike and clinical outcomes.
Materials and methods of the original study
This is a post-hoc analysis, meaning it was conducted subsequently on data already collected from five phase III clinical trials. The data were grouped into three sets: people with type 2 diabetes treated only with basal insulin, people with type 2 diabetes treated with a basal-bolus regimen (i.e., basal insulin plus mealtime doses), and people with type 1 diabetes.
The authors evaluated variability within the day and between different days. Fluctuations were calculated primarily using standard deviation, a measure of the dispersion of values relative to their mean, and the coefficient of variation, a measure that relates dispersion to the mean value. The variability of fasting glycemia, the rate of hypoglycemia, and certain baseline characteristics of the participants were also considered.
To estimate the relative weight of the different variables on the outcomes, a generalized boosted model was used—a statistical procedure that combines many decision trees to identify which factors contribute most to predicting a result. The overall sample size, detailed inclusion and exclusion criteria, duration of individual studies, specific treatments, baseline values, and numerical estimates of the model are not reported in the available material, which consists only of the abstract.
Results
The central finding is that intra-day variability, calculated using standard deviation, had a statistically significant influence on both glycated hemoglobin and mean daily glycemia in all three groups analyzed. In simple terms, people with more fluctuating blood glucose levels throughout the day tended to have less favorable average indicators of glycemic control.
In the group with type 2 diabetes treated with basal insulin alone, the day-to-day variability of fasting glycemia, intra-day variability calculated using the coefficient of variation, and the rate of hypoglycemia also significantly influenced glycated hemoglobin. The abstract does not report the detailed direction of the effects for each variable, nor does it provide coefficients, confidence intervals, P-values, sample sizes, or percentages of explained variance. It is therefore not possible to quantify how much each factor weighed compared to the others.
The authors concluded that lower intra-day variability was associated with an improvement in mean daily glycemia and glycated hemoglobin. This supports the section of the book regarding better metabolic control. However, the abstract does not report that reducing variability directly reduced the risk of hypoglycemia; rather, it reports that the rate of hypoglycemia contributed to the prediction of glycated hemoglobin in a specific subgroup.
Interpretation
An average glucose value may appear acceptable even when it masks frequent fluctuations between high and low levels. For this reason, glycemic variability provides information distinct from glycated hemoglobin: it describes how stable glucose levels are throughout the day. The publication analyzed shows that these oscillations are closely linked to both the daily average glucose and glycated hemoglobin.
The chain of events documented by the research is: greater intra-day variability → less favorable average glycemic control. The reference to postprandial excursions is plausible because meals can contribute significantly to daily oscillations, but the study did not isolate post-meal spikes, nor did it demonstrate that they are the predominant component in every person or therapy.
The additional passage in the book, which suggests that stabilizing excursions is associated with fewer hypoglycemic events, is biologically consistent but is not directly proven by this analysis. To verify this, studies are needed that measure both variability and time spent below the hypoglycemic threshold in the same comparison, preferably with random assignment to different therapeutic strategies.
Independent supporting evidence
A meta-analysis of 15 randomized studies, totaling 2,461 participants with type 1 or type 2 diabetes, compared continuous glucose monitoring with standard care. Compared to the control, continuous monitoring was associated with an average reduction in glycated hemoglobin of 0.17 percentage points, 70.74 more minutes per day in the target glycemic range, and reductions in variability, time above range, and time below range. This evidence confirms the coexistence of improved average control, fewer fluctuations, and less exposure to hypoglycemia, but it does not separate the effect of variability alone from that of the entire sensor-based intervention. Heterogeneity between studies was high and follow-up was short, ranging from 12 to 36 weeks [1].
The GOLD-3 study, a randomized crossover trial conducted in people with type 1 diabetes on multiple daily insulin injections, offers confirmation closer to the book's formulation. During the use of continuous monitoring, the coefficient of glycemic variation fell from 0.40 to 0.37 compared to conventional therapy, with a difference of −0.03; standard deviation and mean amplitude of glycemic excursions also decreased. In the same comparison, the daily time below 70 mg/dL decreased from 4.79% to 2.79%, while the relative risk of time spent below that threshold was 0.58. This indicates a 42% reduction in time in hypoglycemia during the intervention. The result is experimental and consistent with greater glycemic stability, but continuous monitoring also changes therapeutic decisions; therefore, it does not prove that lower variability, in isolation, is the cause of the effect [2].
Conflicting evidence
Evidence does not indicate that improvements in every glycemic measure always occur in tandem. In a six-month randomized study of 224 people with type 2 diabetes on intensive insulin therapy, flash monitoring reduced time in hypoglycemia below 70 mg/dL by 0.47 hours per day, equal to 43%, and below 55 mg/dL by 0.22 hours per day, equal to 53%. However, the change in glycated hemoglobin did not differ between the sensor group and the control group: −0.29% versus −0.31%, with P=0.8222. This data shows that less time in hypoglycemia can occur even without an average benefit in glycated hemoglobin across the entire study population [3].
A small observational study of 108 people with type 2 diabetes also found that the group with glycated hemoglobin below 7% had lower glycemic variability compared to the group with values of at least 7%, but still tended to experience silent hypoglycemia more frequently and for longer durations; the differences in hypoglycemia were not statistically significant. This observation does not contradict the potential utility of reducing fluctuations, but it signals that low glycated hemoglobin and lower variability alone do not guarantee the absence of hypoglycemia [4].
These findings define the main interpretive limitation: average control, variability, and hypoglycemia are linked, but they are not synonymous. Their relationship depends on the type of diabetes, insulin treatment, monitoring technology, therapeutic behaviors, and the measure of variability used.
Summary table of evidence
The table distinguishes the publication subject to the Evidence Review from independent research that confirms or defines the narrative arc of the book.
| Reference | Study type | Population or model | Main result | Direction of evidence | Strength of evidence |
|---|---|---|---|---|---|
| Original study | Post hoc analysis of five phase III studies | Type 1 and type 2 diabetes, in three therapeutic groups | Lower intra-day variability associated with more favorable glycated hemoglobin and mean daily glycemia | direct confirmation | moderate |
| [1] | Systematic review and meta-analysis of RCTs | 2,461 people with type 1 or type 2 diabetes | Reduction in glycated hemoglobin, variability, and time below range with continuous monitoring | direct confirmation | moderate |
| [2] | Crossover RCT | Adults with type 1 diabetes on multi-injection therapy | Reduction in variability and time in hypoglycemia during continuous monitoring | direct confirmation | moderate |
| [3] | Multicenter open-label RCT | 224 adults with type 2 diabetes on intensive insulin therapy | Reduction in hypoglycemia without overall difference in glycated hemoglobin change | discordant evidence | moderate |
| [4] | Cross-sectional observational study | 108 people with type 2 diabetes monitored for 72 hours | Lower variability in the group with lower glycated hemoglobin, but a tendency toward more silent hypoglycemia | discordant evidence | low |
Confirmation of the evidence
Claim in the book
In section 3.1.1, “The Choice to Live in Health or Illness” by Dr. A. Colonnese and R. Panzironi, it is stated that the reduction of intra-day variability is associated with better metabolic control and a lower risk of hypoglycemia. The text links this concept to the stabilization of glycemic excursions after meals.
Wider intra-day glycemic excursions → greater glycemic variability → less favorable glycated hemoglobin and mean daily glycemia → higher probability of extreme glycemic values, including hypoglycemia
The publication subject to the Evidence Review was used in the book to support the following passage: lower daily glycemic variability is associated with better metabolic control, assessed by glycated hemoglobin and mean daily glycemia.
The scientific correspondence is clear for the first segment of the narrative. The original study links lower intra-day variability to more favorable values of glycated hemoglobin and mean daily glycemia in groups with type 1 and type 2 diabetes. However, it does not prove that postprandial oscillations alone are responsible for the entire variability, nor does it quantify a reduction in the risk of hypoglycemia resulting from their stabilization.
Independent evidence reinforces the general picture. A meta-analysis of randomized trials shows that strategies based on continuous monitoring can simultaneously reduce glycated hemoglobin, variability, and time below range [1]. In type 1 diabetes, a randomized study documented both lower oscillations and less time in hypoglycemia [2]. However, other studies show that a reduction in hypoglycemia can occur without better overall glycated hemoglobin [3] and that better mean values or lower variability do not necessarily eliminate the risk of silent hypoglycemia [4]. For “The Choice to Live in Health or Illness” by Dr. A. Colonnese and R. Panzironi, the most rigorous formulation is therefore that the stabilization of oscillations is a goal consistent with better glycemic control and may be accompanied by less hypoglycemia, but it does not constitute a universal causal guarantee.
Overall level of confirmation
MODERATE CONFIRMATION
The link between lower intra-day glycemic variability and average indicators of glycemic control is directly supported by the original study and independent experimental evidence. The link to fewer hypoglycemic events is supported by randomized trials, but it depends on the therapeutic context and cannot be attributed with certainty solely to the reduction of postprandial excursions.
Conclusions
Overall assessment
The narrative is moderately supported. The link between lower glycemic fluctuations and better average control is fairly well supported by clinical data. The connection to fewer hypoglycemic events is confirmed in various interventions, but remains dependent on therapy, technology, and the population; it is not demonstrated as an isolated effect of postprandial variability alone.
References
[1] Effects of Continuous Glucose Monitoring on Metrics of Glycemic Control in Diabetes: A Systematic Review With Meta-analysis of Randomized Controlled Trials
[2] A Randomized Clinical Trial of the Effect of Continuous Glucose Monitoring on Nocturnal Hypoglycemia, Daytime Hypoglycemia, Glycemic Variability, and Hypoglycemia Confidence in Persons with Type 1 Diabetes Treated with Multiple Daily Insulin Injections (GOLD-3)
[3] Flash Glucose-Sensing Technology as a Replacement for Blood Glucose Monitoring for the Management of Insulin-Treated Type 2 Diabetes: a Multicenter, Open-Label Randomized Controlled Trial
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.