Coefficient of Variation for Glucose

Medical Disclaimer

This article is for informational purposes only and does not constitute medical advice. Always consult your physician or a qualified healthcare provider regarding any medical condition or treatment.

Key Takeaways

  • Coefficient of variation (CV) for glucose is standard deviation divided by mean glucose times 100 — a normalized measure of how much your glucose swings relative to its average.
  • The Battelino 2019 international consensus target is CV less than or equal to 36 percent. CV above 36 percent is associated with significantly higher hypoglycemia risk.
  • CV is a better predictor of hypoglycemia than A1C alone — two people with the same A1C can have very different hypoglycemia exposure.
  • A1C captures average glucose; CV captures variability. Together they describe glycemic exposure better than either alone.
  • Strategies to lower CV include consistent meal timing, accurate carb counting, basal-bolus precision, post-meal walks, and automated insulin delivery (AID) systems.

Coefficient of variation (CV) for glucose is standard deviation divided by mean glucose, expressed as a percent. The international consensus target is CV less than or equal to 36 percent. CV captures glycemic variability that A1C and mean glucose miss — and high CV is a strong predictor of hypoglycemia risk.

How CV Is Calculated

  • Collect all CGM glucose readings over a window (usually 14 days).
  • Compute the mean glucose.
  • Compute the standard deviation (SD) of the readings.
  • CV (%) = (SD / mean) × 100.
  • Most CGM software (Dexcom Clarity, FreeStyle LibreView, etc.) calculates CV automatically.

Worked Example

14-Day Mean Glucose (mg/dL) SD (mg/dL) CV (%) Interpretation
140 30 21.4 Excellent — very stable
150 45 30.0 Good — within target
160 58 36.3 At threshold — hypoglycemia risk increasing
170 72 42.4 High variability — hypoglycemia risk significant
180 90 50.0 Very high — frequent hypoglycemia likely
200 120 60.0 Severe variability — review therapy urgently

Why 36 Percent Is the Target

The Monnier et al. 2017 analysis of CGM data found that hypoglycemia (less than 70 mg/dL) frequency rose sharply above CV of about 36 percent. Below 36 percent, lows were rare even with moderate mean glucose. Above 36 percent, lows became common — and increased disproportionately with each percentage point higher. The Battelino 2019 international consensus adopted CV less than or equal to 36 percent as the variability target.

CV vs Other Variability Metrics

Metric What It Captures Typical Target
SD (standard deviation) Absolute spread in mg/dL <50 mg/dL (rough guidance)
CV (coefficient of variation) Spread normalized to mean ≤36%
MAGE Mean amplitude of large excursions <60 mg/dL (rough guidance)
MODD Mean of daily differences (day-to-day stability) <40 mg/dL
iAUC Incremental area under curve (post-meal exposure) Lower is better

CV is the most commonly reported in clinical CGM software because it is easy to calculate and normalize. For an explanation of MAGE and how it differs from CV, see our companion guide.

Why Variability Matters Independently

  • Hypoglycemia risk: the dominant clinical reason. High CV means more frequent lows at the same average glucose.
  • Oxidative stress: glucose swings generate more reactive oxygen species than steady hyperglycemia in lab and animal models.
  • Endothelial dysfunction: rapid glucose changes injure the vascular lining.
  • Inflammation: markers like IL-6 and TNF-alpha rise with variability.
  • Patient-perceived burden: people with high CV report more fatigue, cognitive symptoms, and quality-of-life impact.
  • Possible independent contribution to complications: evidence is mixed; some trials suggest variability adds risk beyond mean glucose, others do not.

What Drives High CV

  • Insulin therapy with imprecise dosing or timing.
  • Inconsistent meals — variable carb amounts and timing.
  • Gastroparesis (variable absorption).
  • Alcohol — delays liver glucose output, causes late lows.
  • Exercise without dose adjustment.
  • Stress and acute illness.
  • Overcorrection — large insulin doses for hyperglycemia followed by overshoot and hypoglycemia.
  • Dawn phenomenon and Somogyi effect — see our dawn phenomenon vs Somogyi guide.
  • Hormonal variability (menstrual cycle, thyroid disorders).
  • Pump malfunction or insulin spoilage.

Strategies to Lower CV

Strategy Typical CV Reduction
Automated insulin delivery (AID) 5 to 10 percentage points
Accurate carb counting 2 to 5 points
Pre-meal insulin timing (15 min before) 2 to 4 points
Consistent meal timing and content 3 to 6 points
Lower-carb / lower-GI diet 3 to 5 points
Post-meal walking 2 to 4 points
Stop overcorrecting highs 2 to 5 points
Address gastroparesis 3 to 8 points
Switch from NPH to glargine or degludec 3 to 6 points

The biggest single intervention for most people is reducing hypoglycemia. Eliminating lows compresses the distribution and lowers CV mechanically.

CV in Different Populations

  • Type 1 diabetes: CV typically 30 to 45 percent on multiple daily injections; 25 to 35 percent on AID.
  • Type 2 on insulin: CV typically 25 to 40 percent.
  • Type 2 on oral agents: CV typically 18 to 28 percent.
  • People without diabetes: CV typically less than 17 percent.
  • Pregnancy: tighter targets — CV less than 30 percent ideal.
  • Older adults: CV less critical than absolute hypoglycemia avoidance.

When to Pay Closer Attention to CV

  • Recurrent unexplained hypoglycemia.
  • Hypoglycemia unawareness.
  • Frequent severe hypoglycemia requiring assistance.
  • Wide swings reported by patient (“rollercoaster”).
  • Considering CGM upgrade or AID system.
  • Pregnancy planning or active pregnancy.
  • Switching insulin regimens.
  • Workup for brittle diabetes.

Limitations of CV

  • Requires CGM data — not available from fingerstick alone.
  • Sensor noise affects calculated SD (and therefore CV).
  • Does not show the timing of variability (e.g., postprandial vs nocturnal).
  • Mean glucose context still matters — CV of 40 percent with mean 130 is different from CV of 40 percent with mean 200.
  • Tight CV achieved with very low mean glucose may hide significant hypoglycemia.
  • Different 14-day windows can give noticeably different CV values during therapy changes.

How to Use CV in Practice

  1. Find CV on your AGP report or in the CGM app summary.
  2. Note whether CV is above or below 36 percent.
  3. If above 36 percent, look at TBR — eliminating lows is the highest-yield action.
  4. Identify daily patterns: morning highs followed by post-meal lows? Overnight swings?
  5. Adjust one thing at a time: meal carbs, insulin timing, basal rate, or activity timing.
  6. Recheck CV after 2 weeks of stable behavior.
  7. Combine CV with TIR and TBR for a complete picture.

See Time in Range, GMI vs A1C, MAGE and glucose variability, and Ambulatory Glucose Profile.

The Bottom Line

Coefficient of variation for glucose normalizes glycemic variability — standard deviation divided by mean, times 100. Target CV less than or equal to 36 percent. Above 36 percent, hypoglycemia becomes frequent and unpredictable, regardless of A1C. CV captures what averages miss: how much your glucose swings. Lower CV through consistent meals, accurate insulin timing, AID systems, and elimination of lows. Review CV alongside TIR, TBR, and GMI to get the complete picture of glycemic control.

Frequently Asked Questions

What is a good coefficient of variation for blood glucose?

The international consensus target is CV less than or equal to 36 percent. CV at or below 36 percent indicates relatively stable glucose; CV above 36 percent is associated with substantially increased hypoglycemia risk. Tighter targets (CV less than 30 percent) are aspirational but harder to maintain, especially with insulin therapy. CV varies more in type 1 diabetes than type 2.

How is CV different from standard deviation for glucose?

Standard deviation (SD) is the absolute spread of glucose values in mg/dL. CV is SD divided by the mean glucose, expressed as a percent. This normalization makes CV comparable across people with different mean glucose levels. Two people with the same SD of 60 mg/dL but means of 120 vs 200 have CVs of 50 percent vs 30 percent — very different variability profiles.

Why does glucose variability matter if my A1C is good?

Variability matters for two reasons. First, hypoglycemia risk rises sharply with high CV — you can have a good A1C while still spending dangerous time below 70 mg/dL. Second, variability itself appears to contribute to complications via oxidative stress, endothelial dysfunction, and inflammation, independent of mean glucose. Capturing variability with CV (or MAGE) complements the average information in A1C and mean glucose.

How do I lower my CV for glucose?

Strategies include consistent meal timing and carb amounts; accurate carb counting; pre-meal insulin timing (15 minutes before eating); basal rate or basal insulin precision; post-meal walking (10 to 20 minutes); avoiding rapid corrections that overshoot; reducing high-glycemic-index carbs; and automated insulin delivery (AID) systems, which often drop CV by 5 to 10 percentage points. Address hypoglycemia first — most CV reduction comes from eliminating the lows.

Sources

  1. Battelino T, et al. Clinical Targets for Continuous Glucose Monitoring Data Interpretation. Diabetes Care 42(8):1593-1603, 2019.
  2. Monnier L, et al. Toward Defining the Threshold Between Low and High Glucose Variability. Diabetes Care 40(7):832-838, 2017.
  3. American Diabetes Association. Standards of Care in Diabetes 2024. Diabetes Care 47(Suppl 1).