I participated in an interesting discussion with AI (Claude Sonnet 4.6) regarding the medical mainstream embrace of glucose variability maximum CV% threshold of 36% for all patients with diabetes. I realize that that number is put in place by the professional societies based on a population target. They need to acknowledge the safety and health of the entire group. What I object to is their resistance to recognize the minority who wish to reach for optimal goals, maybe unreachable but still makes that reaching as healthy.
Here’s a helpful reminder (from another AI session with Claude) of the meaning of SD and CV:
Standard Deviation (SD) measures absolute glucose dispersion. In CGM studies, higher SD independently predicts cardiovascular events, endothelial dysfunction, and mortality even after controlling for mean glucose and HbA1c.
Coefficient of Variation (CV%) — SD divided by mean glucose, expressed as percentage — is arguably more clinically useful because it’s scale-independent. A CV% above 36% is now recognized by international CGM consensus as indicating “high variability” warranting intervention. It allows comparison across patients with different mean glucose levels.
All the usual caveats about AI are recognized. AI can make mistakes and we must remain skeptical. Having said that, here is my prompt that created the discussion:
The %CV goal adopted by the medical mainstream suffers from the same “tyranny of the population norm” that the TIR threshold targets that the 70-180 range does. There has been a move toward addressing this by recognizing that adopting a time in tight range (TITR) of 70-140 mg/dL. Has there been any movement for the professional societies to acknowledge a healthier and lower %CV limit for those patients who recognize and value this metric?
And the response:
This is a sharp and well-reasoned critique. You’re essentially arguing that the 36% CV% threshold, like the 70–180 TIR range, was calibrated against a diabetic population managing with imperfect tools, and therefore encodes the same compromise — what’s achievable, not what’s physiologically optimal. Let me address what’s actually happening in this space.
The Validity of Your Critique
The 36% CV% threshold originated from analyses of CGM data in diabetic populations on insulin, where hypoglycemia risk is real and constraining. It was chosen partly because it distinguishes “stable” from “unstable” glycemia in that context, and partly because it’s achievable without dangerous over-treatment. It was not derived by asking what CV% characterizes metabolically healthy non-diabetic individuals with optimal glucose regulation.
This is the same methodological flaw you identified in the TIR targets: the reference population is sick people being managed with blunt tools, not healthy people with well-functioning metabolism.
CGM studies in healthy non-diabetic individuals consistently show CV% values considerably lower than 36% — typically in the range of 15–25%, with genuinely metabolically healthy individuals often sitting below 20%. The 36% threshold therefore doesn’t represent health; it represents the boundary between poorly controlled and adequately controlled diabetes. Adopting it as a universal target imports the same population-norm tyranny.
What the TITR Movement Signals
The emergence of Time in Tight Range (70–140 mg/dL) as a concept is significant precisely because it implicitly acknowledges that 180 mg/dL is not a physiological ceiling of safety but a clinical convenience threshold. The TITR framework is being discussed and researched primarily in:
Type 1 diabetics using advanced closed-loop insulin systems, where tighter control is now mechanically achievable without the same hypoglycemia risk
Pregnant diabetics, where fetal harm from excursions above 140 mg/dL is well-documented and tighter targets are already standard of care
Researchers studying metabolic health in non-diabetic populations using CGM
The logical extension of TITR to variability metrics — a “tight CV%” target reflecting healthy physiology rather than managed disease — has not yet been formally proposed by major professional societies, but the conceptual groundwork is being laid.
Where the Professional Societies Currently Stand
The honest answer is that movement is nascent but not yet substantive on this specific question.
The major international CGM consensus documents — the 2019 Battelino consensus and subsequent updates — established the 36% CV% threshold as the primary variability target and have not yet stratified this by patient population, metabolic health status, or therapeutic modality. The ATTD (Advanced Technologies and Treatments for Diabetes) congress, which drives much of this consensus work, has discussed tighter metrics but has not formally proposed lower CV% targets for motivated patients seeking optimal rather than adequate control.
The ADA’s Standards of Care have progressively incorporated TIR language and CGM metrics but remain anchored to the 70–180 range and 36% CV% as primary benchmarks. The move toward TITR language is appearing more in research literature and conference presentations than in formal guideline documents.
This lag is not surprising — professional society guidelines are inherently conservative, require substantial evidence from trials powered around new endpoints, and must balance the needs of the entire diabetic population including those for whom tighter targets are dangerous.
The Emerging Two-Tier Framework
What seems to be developing organically, if not yet formally codified, is a de facto two-tier framework:
Tier 1 — Standard clinical targets: 70–180 TIR, HbA1c <7%, CV% <36%. Designed for the broad diabetic population, balancing efficacy against hypoglycemia risk and practical achievability.
Tier 2 — Optimal physiological targets: 70–140 TITR, HbA1c <5.7%, CV% <20–25%. Increasingly discussed in the context of closed-loop systems, low-carbohydrate dietary approaches, non-diabetic metabolic optimization, and longevity medicine.
The second tier exists in clinical practice — particularly among endocrinologists working with motivated type 1 patients on advanced closed-loop systems, and in the growing field of metabolic health optimization — but it lacks formal professional society endorsement.
Specific Factors That Could Drive Lower CV% Target Adoption
Building on the leverage points discussed earlier, a few specific forces could accelerate formal recognition of a lower, physiologically-grounded CV% target:
Closed-Loop Insulin System Data
The newest hybrid and fully closed-loop systems (iLet, Omnipod 5, Medtronic 780G) are achieving CV% values in the 20–28% range in type 1 diabetics — populations where such control was previously impossible without severe hypoglycemia risk. As outcomes data from these systems accumulates, it will become possible to ask whether CV% below 30% or 25% produces measurably better long-term outcomes than CV% of 33–36%. If the answer is yes, it creates an evidence-based rationale for a lower threshold that societies cannot easily ignore.
Non-Diabetic CGM Research
A growing body of research is characterizing normal glucose variability in healthy populations. If this literature consolidates around a specific CV% range — say, below 20% as the signature of genuine metabolic health — it provides a physiological reference point that is independent of diabetic population norms. This is the most intellectually clean argument for a lower target: not “we can now achieve lower variability” but “lower variability is what healthy metabolism actually looks like.”
Pregnancy Targets as a Precedent
Tighter glycemic targets in pregnancy are already standard of care — post-meal targets of below 140 mg/dL are routine in gestational diabetes management. This demonstrates that the medical mainstream already implicitly accepts tighter targets when the stakes are recognized as high enough. The argument that all patients should be treated with the urgency currently reserved for pregnant diabetics — because the stakes of decades of subclinical damage are comparably high — is coherent and could be made more forcefully.
Patient-Driven Demand Through CGM Adoption
As CGM use spreads beyond diabetics into the general population — driven by consumer health interest and direct-to-consumer marketing — a growing constituency of non-diabetic users will be generating their own CV% data and asking clinicians what it means. Clinicians currently have no good answer calibrated to that population. This creates professional pressure to develop guidance, and that guidance will need to be grounded in healthy physiology rather than diabetic population norms.
The Deeper Issue You’re Pointing To
Your critique actually touches on a fundamental tension in evidence-based medicine: guidelines optimized for population-level management are systematically misaligned with individual-level optimization. The 36% CV% threshold minimizes harm across a heterogeneous population that includes elderly patients, those with hypoglycemia unawareness, those with limited access to advanced technology, and those with multiple comorbidities. For a motivated, informed patient with access to CGM and willing to pursue dietary or technological interventions to minimize variability, that threshold is a floor being mistaken for a ceiling.
The most intellectually honest position the professional societies could take — and one that some forward-thinking clinicians are already articulating — is something like: “A CV% below 36% is the minimum acceptable target for standard diabetic management. A CV% below 20–25% reflects physiologically normal glucose regulation and should be the aspirational target for patients able and willing to pursue it.”
That distinction — between minimum acceptable and physiologically optimal — does not yet exist in formal guidelines, but it’s the direction the evidence and the logic of your argument both point.
Is anyone else interested in this? What are your thoughts? Should the primary devotion to the population demographic completely obscure the small, yet significant, group that have paid attention and shown they are willing to use diet and technology to reach near normal glycemia? And do this without sacrificing metabolic safety.
. I thank you for the feedback.