Insulin to Carb Ratio Determination: How to Find Your ICR
The direct answer
An insulin-to-carb ratio expresses how many grams of dietary carbohydrate one unit of a rapid-acting insulin analog (lispro, aspart, glulisine, or faster aspart) covers at a meal, applied only to mealtime bolus dosing, not to background basal insulin. A common way to generate a starting number is the 500 Rule: divide 500 by total daily insulin dose (TDD, basal plus all bolus units). That number is an approximation drawn from clinical convention, not a lab-measured value, and it needs confirmation through structured meal testing before a person relies on it. Adult ratios commonly fall somewhere between roughly 1 unit per 8 grams and 1 unit per 20 grams of carbohydrate, with children often needing higher gram-per-unit ratios and adolescents in puberty sometimes needing much lower ones; the only way to know an individual's actual ratio is to test it.
ICR, ISF, basal rate, and sliding scale are not interchangeable terms
These four terms get confused often enough that mixing them up is a common source of dosing error.
- ICR (insulin-to-carb ratio): grams of carbohydrate covered by one unit of rapid-acting insulin at a meal.
- ISF or correction factor (insulin sensitivity factor): how many mg/dL one unit of rapid-acting insulin is expected to lower blood glucose, used to correct a pre-meal glucose that is already above target.
- Basal rate or basal dose: the background long-acting or pump-delivered insulin that covers glucose output between meals and overnight, independent of what a person eats.
- Sliding scale: a fixed dose keyed only to the current glucose reading, with no adjustment for the carbohydrate about to be eaten. The American Diabetes Association does not recommend sliding-scale-only regimens for outpatient management of type 1 diabetes because they ignore carbohydrate intake entirely.
A dose that goes wrong at a meal can be a basal problem, an ICR problem, a timing problem, or a carb-counting problem. Confusing these categories is the single most common reason people adjust the wrong variable, discussed further below.
The useful question is not "what is my ICR" but "which variable actually failed"
A glucose reading that is too high or too low after a meal has more than one possible cause: an inaccurate ICR, an inaccurate basal rate, an inaccurate carbohydrate count, insulin timing that does not match the food's glycemic profile, or a temporary physiologic state such as illness or exercise. Treating every out-of-range reading as an ICR problem, and adjusting the ratio repeatedly without isolating the actual cause, is a pattern that tends to make control worse rather than better. The sections and framework below are organized around identifying which variable is actually responsible before changing anything.
The 500 Rule: a starting calculation, not a final answer
The 500 Rule is the standard formula used clinically to generate an initial ICR estimate for someone using rapid-acting insulin analogs.
ICR (grams per unit) = 500 / TDD
A person with a total daily dose of 40 units gets a starting estimate of 500 / 40 = 12.5, rounded to roughly 1 unit per 12 or 13 grams. This approach is widely taught in diabetes education and is consistent with the general framework in the American Diabetes Association's Standards of Care for pharmacologic glycemic management (ADA Standards of Care, Diabetes Care), though the Standards of Care describe general principles of insulin therapy rather than a single mandated formula, and the exact wording should be checked against the current edition before it is cited as an explicit endorsement.
Some clinicians use a 450 Rule (450 divided by TDD) for people using regular human insulin rather than a rapid-acting analog, because regular insulin has a slower onset and longer duration, and a more conservative starting ratio reduces the risk of late hypoglycemia. This is a clinical convention, not a formally validated formula with its own trial evidence.
Practical cautions before running the calculation:
- TDD should reflect a stable, representative period of at least three to five days, not a day affected by illness, a growth spurt, or unusual correction dosing. Using an atypical day inflates TDD and produces a ratio that delivers too little insulin per gram.
- Someone newly diagnosed with type 1 diabetes and still in a partial remission ("honeymoon") period should expect insulin needs, and therefore the ICR, to shift over the following months and should plan for more frequent review, not a one-time calculation.
Verifying the ratio with a real-meal test
The 500 Rule produces an estimate. A structured meal test is what turns that estimate into something a person can actually rely on for dosing.
Choose a meal with carbohydrate content you can measure precisely, weighed rather than guessed. Check glucose before eating and confirm it is close to target. Dose the calculated bolus, eat, avoid exercise for at least two hours, and check glucose again at 2 hours and 4 hours after the meal.
Three general outcomes:
- Glucose returns to within about 30 mg/dL of the pre-meal value by 2 to 4 hours: the ratio is likely accurate for that meal type.
- Glucose rises well above the pre-meal value (commonly described as 40 to 50 mg/dL or more) and stays elevated at 4 hours: the ratio may be too weak, meaning it is assigning too few units per gram. This should prompt a review of carbohydrate counting accuracy first, since undercounted carbs look identical to a weak ratio on a glucose curve.
- Glucose falls well below the pre-meal value or drops under 70 mg/dL: the ratio may be too strong, or the carbohydrate content was overestimated.
Repeating the test two or three times on different days, rather than acting on a single reading, matters because stress, minor illness, and imprecise carb counting can all produce a one-off anomaly that has nothing to do with the ratio itself. A continuous glucose monitor gives a full postprandial curve instead of two point-in-time checks, which makes distinguishing a timing mismatch from a true ratio error considerably easier.
The correction factor and how it combines with the ICR
The ICR does not account for a glucose level that is already elevated before the meal starts. The correction factor, also called the insulin sensitivity factor (ISF), fills that gap.
ISF (mg/dL per unit) = 1,800 / TDD (for rapid-acting analogs; some clinicians use 1,500 as the divisor for regular human insulin, reflecting its slower, more prolonged action)
For someone with a TDD of 40 units, ISF is approximately 1,800 / 40 = 45 mg/dL per unit. If pre-meal glucose is 210 mg/dL against a target of 120 mg/dL, the correction dose is (210 − 120) / 45 ≈ 2 units.
Total mealtime bolus = (grams of carbohydrate / ICR) + ((current glucose − target glucose) / ISF)
If pre-meal glucose is already at target, the correction term is zero. If it is below target, the correction term is negative and units are subtracted from the meal dose.
Like the 500 Rule, the 1,800 divisor is a clinical convention used to generate a starting estimate, not a value validated in a way that guarantees accuracy for a given individual. Insulin sensitivity varies enough between people, and within the same person over time, that the formula's output should be treated the same way as the ICR estimate: a number to verify with real data, not a final answer.
How ICR fits into a basal-bolus regimen
In basal-bolus therapy, a long-acting basal insulin (glargine, detemir, or degludec) covers background glucose between meals and overnight, while a rapid-acting analog covers each meal using the ICR. A common starting split allocates roughly 40 to 50 percent of TDD to basal insulin and the remainder to mealtime bolus doses, though this varies by individual.
Getting the attribution right matters. If fasting glucose starts in range but drifts upward through the morning without any food intake, that pattern points to an underdosed basal rate, not an inaccurate ICR. Increasing the mealtime dose to compensate for a basal problem tends to produce hypoglycemia later, once the basal gap is eventually corrected. Professional guidance from groups such as the American Diabetes Association and the American Association of Clinical Endocrinology generally frames basal insulin titration as the step to get right before fine-tuning prandial ratios, though the exact wording of any specific consensus statement should be verified against the current published document rather than quoted from memory.
Rapid-acting analogs differ somewhat in onset and peak timing (for example, faster-acting formulations of aspart are marketed with a shorter onset than standard aspart), and switching between formulations without reassessing timing or the ratio can shift the postprandial curve even if the ICR number itself has not changed. Specific onset and peak figures vary by product label and study design, and a person switching insulins should re-verify their ratio with meal testing rather than assume it carries over unchanged.
ICR settings on an insulin pump
Insulin pumps allow multiple ICRs across the day, which is useful because insulin sensitivity is not constant around the clock. A common pattern uses a lower, stronger ratio at breakfast (reflecting the dawn phenomenon's tendency to reduce morning insulin sensitivity) and higher, more conservative ratios later in the day, though the actual pattern for a given person can only be established through testing at each meal time.
A pump's bolus calculator combines the programmed ICR, the correction factor, active insulin already on board, and the current glucose value to suggest a dose. Automated insulin delivery and closed-loop systems can adjust some dosing parameters dynamically based on continuous glucose data; this is an active area of clinical trial research, and specific outcome figures from any individual trial should be checked against the published study before being cited, since headline numbers vary considerably by device, population, and trial duration.
For someone starting pump therapy, a reasonable approach is to begin with a single 24-hour ICR from the 500 Rule, verify it across at least a week of meals, and only split it into time-of-day-specific ratios once the single ratio has been confirmed as reasonably accurate. Splitting too early makes it much harder to tell which time period's ratio is actually off.
What changes an ICR over time
An ICR is not a permanent number. Several situations warrant reassessment:
- Weight change. A gain or loss of roughly 5 pounds or more shifts TDD and is a reasonable trigger to recalculate.
- Exercise. Aerobic activity increases insulin sensitivity for roughly 24 to 48 hours afterward. Some people use a more conservative, higher-gram ratio on and after training days to reduce delayed hypoglycemia risk. Resistance training has a more variable and less predictable effect on glucose during the session itself.
- Illness. Counter-regulatory hormones released during infection reduce insulin sensitivity, sometimes sharply. A temporary, more aggressive ratio may be needed during a fever, reviewed daily rather than made permanent.
- Puberty. Adolescents commonly need substantially more insulin during the growth hormone surge of puberty, and pediatric guidelines such as those from the International Society for Pediatric and Adolescent Diabetes (ISPAD) recommend frequent review during this period. Exact percentage increases vary by individual and should not be assumed from a general figure.
- Pregnancy. Insulin resistance increases progressively through the second and third trimesters in people with pre-existing diabetes. Guidance from the American College of Obstetricians and Gynecologists on pregestational diabetes supports reassessing insulin doses at each prenatal visit rather than assuming a fixed ratio carries through pregnancy.
- Glucocorticoid therapy. Systemic steroids such as prednisone commonly cause pronounced postprandial hyperglycemia, and a steroid-specific, more aggressive ratio for affected meals is often needed for the duration of the course, reviewed daily.
Carbohydrate counting accuracy: the prerequisite the ratio depends on
No ICR works if the carbohydrate estimate feeding into it is wrong. A ratio of 1 unit per 10 grams produces very different outcomes depending on whether an actual 90-gram meal was logged as 60 grams.
Weighing food and using nutrition labels or a food database is generally more accurate than estimating portion size by eye, though the exact degree of improvement varies by study and by how experienced the person is at counting. Anyone troubleshooting an apparent ICR problem should rule out a carb-counting error before changing the ratio itself, since the two produce identical glucose patterns.
High-fat, high-protein meals add a separate complication: fat and protein slow gastric emptying and can produce a delayed glucose rise several hours after a meal that otherwise looked normal at the 2-hour check. Some clinicians use an extended or dual-wave bolus for these meals rather than changing the ICR itself, since the mismatch is about timing, not the ratio.
Decision framework: what a postprandial glucose pattern actually tells you
Most self-adjustment mistakes come from treating every abnormal post-meal reading as an ICR problem. This framework separates the patterns that actually indicate a ratio problem from the ones that indicate something else, so the correct variable gets adjusted.
| Pattern on 2-hour post-meal check | Most likely cause | What to check first | Action |
|---|---|---|---|
| Consistently 40+ mg/dL above pre-meal, stays high at 4 hours | ICR too weak, or carbohydrate undercounted | Re-weigh the meal on a separate day with the same ratio | If carb count was accurate, consider a modest decrease in the ICR gram number, then re-verify |
| Returns to within roughly 20-30 mg/dL of pre-meal | Ratio working as expected for this meal | Nothing needed | No change |
| Drops 30+ mg/dL below pre-meal, or below 70 mg/dL | ICR too strong, or carbohydrate overcounted | Re-check the carb count and confirm no unusual exercise occurred | If carb count was accurate, consider a modest increase in the ICR gram number, then re-verify |
| High fasting or pre-lunch reading despite no food eaten since the last meal | Basal insulin underdosed, not an ICR problem | Track glucose overnight and between meals with no bolus involved | Discuss basal titration with the care team before touching the ICR |
| Sharp peak at 1 hour, back near pre-meal by 2 hours | Insulin injected too late relative to a fast-digesting meal | Injection timing relative to eating | Inject 10 to 15 minutes earlier for similar meals; do not change the ICR |
| Normal at 2 hours, rises at 3 to 4 hours | High-fat or high-protein meal with delayed absorption | Fat and protein content of the meal | Consider an extended or dual-wave bolus for that meal type; do not permanently lower the ICR |
| Pattern only appears during illness, steroid use, or a training block | Temporary physiologic change | Timing relative to the trigger | Use a temporary, clearly time-limited adjustment; revert once the trigger resolves |
The rule underneath this table: change the ICR only after carbohydrate counting, injection timing, and basal dosing have each been checked and ruled out as the actual cause, and only after the pattern has repeated across more than one meal.
When self-adjustment is reasonable and when to involve a clinician
Small self-adjustments, on the order of 10 to 15 percent, based on repeated structured meal-test data, are a reasonable and commonly taught approach for people already educated in carb counting and dose calculation. Larger swings, or needing to adjust more than about once every two weeks, are a signal to bring in the diabetes care team rather than keep self-adjusting.
Contact a clinician when:
- Pre-meal glucose is consistently out of range despite basal insulin that otherwise appears well-titrated, since the cause may be basal dosing rather than the ICR.
- Hypoglycemia below 70 mg/dL is occurring more than twice a week.
- HbA1c remains elevated despite doses that appear to follow the calculated ratios correctly.
- A new medication that affects insulin sensitivity is starting, such as a GLP-1 receptor agonist, an SGLT2 inhibitor, or a systemic glucocorticoid. GLP-1 receptor agonists slow gastric emptying and commonly reduce postprandial glucose excursions, which often means mealtime insulin needs drop and the ICR gram number may need to increase; this should be adjusted with clinical input rather than assumed.
When people using insulin receive structured insulin-to-carbohydrate ratio (ICR) training as part of diabetes self-management education, they tend to achieve better blood sugar control, though the magnitude of improvement differs across study populations and should be verified in the original research rather than assumed to be consistent across all groups. Endocrinology and diabetes technology organizations generally recommend that people using insulin have regular contact with a diabetes care and education specialist to adjust insulin doses and review carbohydrate ratios, though different organizations suggest different visit frequencies and current guidelines should be consulted for specific recommendations rather than assuming one approach fits all patients.
Pediatric and adolescent considerations
Children typically need higher gram-per-unit ratios than adults because of lower body mass and, in younger children, often greater baseline insulin sensitivity. Adolescents in mid-puberty can need considerably lower, stronger ratios during the growth hormone surge. The 500 Rule still applies mathematically, using the child's actual TDD, but the resulting number moves in both directions more than it typically does in adults, which is why pediatric diabetes guidelines such as ISPAD's recommend more frequent review, often at each clinic visit for younger children.
Unpredictable appetite in young children creates a practical wrinkle: dosing after a meal, once it is clear how much the child actually ate, is sometimes preferred over pre-meal dosing to avoid hypoglycemia when a child leaves food unfinished. Faster-onset insulin formulations make this approach more workable by narrowing the gap between injection and peak action, though the appropriateness of post-meal dosing for a specific child is a decision for the family and care team, not a general recommendation.
What is established, what is plausible, and what is not established
Established: Carbohydrate intake is a primary driver of postprandial glucose excursions, and matching mealtime insulin to carbohydrate content is standard practice in intensive insulin therapy for type 1 diabetes. The Diabetes Control and Complications Trial demonstrated that intensive insulin therapy, which included carbohydrate-matched bolus dosing, produced substantially lower HbA1c and reduced the risk of retinopathy progression compared with conventional therapy over years of follow-up (DCCT, N Engl J Med 1993). The 500 Rule and the 1,800 Rule are widely used, clinically accepted starting-point formulas.
Plausible but not rigorously quantified for an individual: The exact numeric ranges commonly cited for how much a ratio should change with exercise, illness, puberty, or medication changes are useful clinical heuristics rather than precisely validated figures that apply the same way to everyone. Individual variability is large enough that these ranges should guide a starting adjustment, not replace verification.
Not established from the material available here: Precise statistics quoting a specific percentage improvement in time-in-range, a specific correlation coefficient between a TDD-based formula and measured insulin sensitivity, or a specific percentage reduction in HbA1c attributable to ICR training specifically (as opposed to structured diabetes education generally) require verification against the primary published study before they should be presented as a fixed number. Readers and clinicians relying on this article for a specific cited statistic should confirm it against the original trial or guideline rather than this summary.
Frequently asked questions
What is an insulin-to-carb ratio?
How do I calculate a starting insulin-to-carb ratio?
What is the 500 Rule?
How is the correction factor different from the ICR?
How do I know if my ICR needs to change?
Can I use different ICRs at different times of day?
Does exercise change my ICR?
What is the difference between an ICR and a sliding scale?
Do GLP-1 medications change insulin needs?
References
- Diabetes Control and Complications Trial Research Group. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med. 1993;329(14):977-986. https://www.nejm.org/doi/10.1056/NEJM199309303291401
- American Diabetes Association Professional Practice Committee. Standards of Care in Diabetes. Diabetes Care. https://diabetesjournals.org/care/article/47/Supplement_1/S158/153956
Other guideline bodies referenced by name in this article, including ISPAD, ACOG, AACE, and the Academy of Nutrition and Dietetics, publish current recommendations relevant to insulin dosing in children, pregnancy, and clinical practice generally. Specific document titles, publication years, and any quoted statistics attributed to these bodies should be verified against the current published guideline before being cited as an exact figure or direct quotation.