Basal Bolus Insulin Regimen - Carb Ratios & Correction Doses

The direct answer

Basal bolus therapy separates two jobs that a healthy pancreas does automatically: keeping glucose stable between meals (basal insulin) and covering the glucose rise after eating (bolus insulin, dosed against carbohydrate grams and a correction factor). Starting doses are commonly estimated from body weight and total daily dose using teaching formulas such as the 500 rule and 1,700 rule, but these are population averages, not individualized prescriptions, and every source that publishes them also states they must be refined against a patient's actual glucose response over roughly two to four weeks. The American Diabetes Association's Standards of Care in Diabetes is the current authoritative source for glucose targets and regimen selection, and it should be checked for the year's updated numbers before any target is applied to a specific patient. This article explains the math, where it can go wrong, and when a reading is a signal to call a clinician rather than adjust a dose alone.

What a basal bolus regimen actually does

The basal component is a long-acting insulin analog, most often glargine (available as U-100 or U-300 formulations, brand names Lantus and Toujeo), detemir (Levemir), or degludec (Tresiba). Its job is to suppress the liver's background glucose output so fasting and between-meal glucose stays roughly flat. The bolus component is a rapid-acting analog, such as aspart (NovoLog), lispro (Humalog), or glulisine (Apidra), given at meals and calculated from carbohydrate content plus a correction dose if glucose is already above target.

Multiple daily injection (MDI) therapy with separate basal and bolus components, or an insulin pump delivering only rapid-acting insulin on a programmed schedule, is the regimen structure most guideline bodies recommend for type 1 diabetes and for people with type 2 diabetes who need more than basal insulin alone to reach their individualized glucose target. The exact current wording and A1C targets should be confirmed against the ADA's yearly Standards of Care, since target ranges and preferred regimens are revised annually.

Evidence anchor: the claim that basal bolus therapy outperforms older premixed (biphasic) insulin regimens on A1C is well established in the clinical literature and in guideline text, but a specific numeric size of that advantage (for example, "1.0 to 1.5 percentage points") could not be verified against a specific study in the source material for this article and should not be quoted as an exact figure without checking a current systematic review. If a precise number is needed for a patient conversation, verify it against a current meta-analysis rather than repeating an inherited figure.

Setting a starting basal dose

Clinicians commonly start basal insulin using a weight-based estimate, often cited in the range of 0.1 to 0.2 units per kg per day for people starting insulin who are not already on any insulin, with higher starting doses sometimes used in type 1 diabetes depending on residual insulin production. This is a starting point set and adjusted by the prescribing clinician, not a formula patients should use to self-select a dose.

Titration protocols vary by insulin and by practice, but a common pattern taught in diabetes education is a small fixed increase (for example, 1-2 units) every few days if fasting glucose remains above target on repeated mornings, done only after ruling out unrecognized nighttime hypoglycemia. Degludec has a long duration of action and takes several days to reach a steady blood level, which is a specific reason not to re-adjust degludec doses more often than every 3-4 days: doing so risks "stacking" doses before the prior increase has fully taken effect.

As a rough rule of thumb once a regimen is optimized, basal insulin often makes up somewhere near 40-50% of total daily dose, with the rest split among mealtime bolus doses. If basal is well above that share and after-meal glucose is still high, that pattern points toward adding or increasing bolus coverage rather than increasing basal further, but this pattern should be reviewed with a clinician rather than acted on unilaterally, since it can also reflect a carb-counting problem.

Insulin-to-carb ratio: the 500 rule and its limits

The insulin-to-carb ratio (ICR) describes how many grams of carbohydrate one unit of rapid-acting insulin covers. An ICR of 1:15 means one unit covers 15 grams of carbohydrate.

The 500 rule is a widely taught starting estimate: divide 500 by total daily insulin dose (TDD). A patient using 50 units per day gets a starting ICR of 500/50 = 1:10. This is a teaching heuristic derived from clinical experience and retrospective analyses of pump users, not a fixed physiological constant, and different sources present slightly different versions of the underlying constant. That variation itself is worth knowing: if two clinicians or two apps give a slightly different starting ratio for the same patient, that is expected, not an error.

Carbohydrate sensitivity commonly varies by time of day, with many people needing a more aggressive (lower-carb-per-unit) ratio at breakfast than at dinner, a pattern often attributed to morning cortisol and growth hormone effects on insulin resistance. This is a plausible and commonly taught pattern, but the degree of variation is individual and should be confirmed with logged data rather than assumed from a formula.

To refine an ICR after starting therapy, a structured log works better than memory:

1. Pre-meal glucose
2. Grams of carbohydrate eaten
3. Bolus units given
4. Glucose roughly 2 hours after the meal

If 2-hour glucose rises more than about 40-50 mg/dL above the pre-meal value on repeated similar meals, the ratio is likely too generous (too many carbs allowed per unit). If it drops sharply, the ratio may be too aggressive. Structured carbohydrate counting education, as opposed to informal estimation, is consistently associated with better bolus accuracy in the diabetes education literature, and a common real-world source of error is underestimating restaurant portion sizes, particularly for pasta and rice dishes, which frequently contain more carbohydrate than a home-cooked equivalent portion.

Correction factor: the 1,700 (or 1,800) rule

The correction factor (also called insulin sensitivity factor, ISF) estimates how many mg/dL one unit of rapid-acting insulin will lower blood glucose. The commonly taught formula is:

CF = constant ÷ TDD

Different sources use different constants for this formula, most often 1,700 or 1,800 for mg/dL units (some older material uses 1,500 for regular insulin, which is no longer the mainstay of mealtime dosing). Using 1,700 as an example: a patient with a TDD of 50 units gets CF = 1,700/50 = 34 mg/dL per unit. If target glucose is 100 mg/dL and current glucose is 202 mg/dL, the estimated correction dose is (202-100)/34 ≈ 3 units.

For patients using mmol/L, some sources use a constant of 100 divided by TDD as a rough mmol/L equivalent; this conversion should be checked against the patient's own care team or pump manufacturer documentation rather than assumed, since conventions differ.

Correction doses carry a specific risk that ICR-only dosing does not: stacking, where a second correction is given before the first has finished acting. Rapid-acting analogs remain pharmacologically active for roughly 3-4 hours in most people. Many modern insulin pumps and connected bolus-calculator apps track "insulin on board" (IOB) and subtract it automatically before suggesting a correction dose. People using injections without such a tool need to track this manually or with an app, and should ask their prescribing clinician which tool, if any, is appropriate for their regimen.

Mealtime bolus timing and carb counting

Rapid-acting insulin analogs generally begin lowering glucose within roughly 10-20 minutes of subcutaneous injection, while glucose from a meal can start rising within about 15 minutes of the first bite. A commonly taught practical guideline is to inject 10-15 minutes before eating for most meals, though this varies with the specific insulin, the meal's glycemic index, and individual absorption.

People with gastroparesis or unpredictable oral intake are often advised to dose after eating, or to use a pump's extended/dual-wave bolus feature that spreads insulin delivery over a longer window, matched to slower or delayed carbohydrate absorption. This kind of adjustment should be set up with a clinician familiar with the individual's gastric emptying pattern.

Practical carb-counting steps:

1. Read the nutrition label and use total carbohydrate, not "net carbs," unless a clinician has specifically advised a net-carb approach.
2. Measure or weigh the portion actually eaten rather than estimating.
3. Divide grams eaten by ICR to get bolus units.
4. Add a correction dose if pre-meal glucose is above target.
5. Subtract any insulin on board from a correction given within the past several hours.

The ADA has previously allowed subtracting fiber above a threshold (commonly cited as 5 grams per serving) from total carbohydrate before calculating a bolus, on the reasoning that high-fiber foods blunt the after-meal glucose rise; this detail should be checked against the current year's ADA nutrition guidance, since specific numeric thresholds in nutrition guidelines are periodically revised.

Translating an injection regimen to pump settings

An insulin pump (continuous subcutaneous insulin infusion, CSII) delivers only rapid-acting insulin, programmed as a continuous basal rate that replaces the long-acting injection entirely. Converting an MDI regimen to pump settings starts from the same three numbers already discussed: TDD, basal share, and ICR/CF.

A commonly taught starting approach, always done under clinician supervision rather than self-directed:

Step 1. Some clinicians reduce the prior MDI TDD by roughly 10-20% when starting pump therapy, reflecting that continuous low-dose subcutaneous delivery can be somewhat more efficient than once- or twice-daily injections for some patients. This reduction is not universal and should be set by the prescribing clinician.

Step 2. Allocate roughly 50% of the (possibly reduced) pump TDD to basal, spread evenly across 24 hours as an initial flat rate.

Step 3. After 2-4 weeks of continuous glucose monitor (CGM) data, basal rates are typically broken into time segments rather than left flat. Many patients need a higher rate in the early morning hours (dawn phenomenon) and a lower rate before planned exercise.

Step 4. Enter ICR and CF into the pump's bolus calculator, often as separate values for morning, afternoon, and evening, refined the same way described above.

Worked example and a decision framework for trusting the numbers

The table below shows the arithmetic for a hypothetical 70 kg adult with type 1 diabetes on a TDD of 42 units. These are illustrative calculations only, not a recommendation for any specific reader's dose.

When to trust a formula-derived number, and when to override it:

• Trust it as a starting point only during the first 1-2 weeks after a regimen change, paired with frequent glucose checks or CGM review.
• Override toward more insulin per gram of carb (lower ICR number) if 2-hour post-meal glucose is repeatedly more than 40-50 mg/dL above pre-meal glucose on similar meals, after confirming the carb count itself was accurate.
• Override toward less insulin per gram of carb (higher ICR number) if glucose repeatedly drops sharply after meals, especially if it drops below 70 mg/dL.
• Do not re-titrate basal more than once every 3 days, and never before ruling out unrecognized nighttime hypoglycemia as the cause of a high morning reading.
• Treat any inherited "standard" constant (500, 1,700, 1,800) as a convention, not a physiological law. If an app, pump, or clinician gives a different starting number than this article's formula, that is expected variation between conventions, not necessarily an error, and the clinician's number should generally take precedence.
• Stop self-adjusting and call the care team if fasting glucose stays above the individualized target for more than 3 days despite following the titration plan, or if hypoglycemia below 54 mg/dL happens more than once in a week.

Hypoglycemia risk on an intensified regimen

Hypoglycemia is the primary safety tradeoff of basal bolus therapy compared with simpler regimens, because tighter glucose control mechanically increases the number of moments where a dosing error can push glucose too low. This tradeoff is well established across decades of intensive insulin therapy research, though the exact size of the increased risk depends heavily on the specific regimen, insulin types used, and patient population, and a precise historical statistic should be verified against a current source rather than assumed to still apply to modern insulin analogs and CGM-supported regimens.

Glucose below 70 mg/dL is commonly used as a first-level hypoglycemia alert threshold. The standard first-response approach taught in diabetes education is the "15-15 rule": consume about 15 grams of fast-acting carbohydrate, recheck glucose after about 15 minutes, and repeat if still below threshold.

Newer basal insulins with flatter, more predictable action profiles (glargine U-300, degludec) are commonly described as lowering nighttime hypoglycemia risk compared with older glargine U-100, a finding that has been reported in randomized comparisons, though a specific percentage reduction should be checked against a current trial report rather than repeated as a fixed figure.

Anyone on insulin, and their household members, should know how to use a glucagon rescue product for severe hypoglycemia (loss of consciousness or inability to safely swallow). Both injectable and intranasal glucagon formulations are FDA-approved and available; the specific product, dose, and administration steps should come from the prescribing clinician or the product's current FDA label rather than from a general description in an article like this one.

Monitoring and when a number means "call the clinician"

Dose changes should be based on data, not a single reading: pre-meal and roughly 2-hour post-meal checks, or CGM data covering at least two weeks, are the minimum most clinicians want before adjusting a regimen. Current ADA glucose and A1C targets, including individualized targets for pregnancy, older adults, and those with a history of severe hypoglycemia, change periodically and should be pulled from the current year's ADA Standards of Care rather than treated as fixed.

CGM "time in range" (TIR), typically defined as the percentage of readings between 70-180 mg/dL, is increasingly used alongside A1C to describe glucose control, since two people with the same A1C can have very different amounts of time spent low or high. A commonly cited rough correspondence exists between certain TIR percentages and A1C values, but this correspondence is approximate and population-level, not a substitute for an individual's actual A1C measurement.

Contact a clinician promptly, rather than continuing to self-adjust, if:

• Fasting glucose stays above the individualized target for more than 3 days despite following an agreed titration plan
• Glucose below 54 mg/dL (a more severe hypoglycemia threshold) happens more than once in a week
• A dose that previously worked stops working, without an obvious explanation (illness, activity change, new medication)
• Confusion, loss of consciousness, or inability to safely treat a low without help from another person occurs at any point, this is urgent and may need emergency care

Special situations that change the numbers

Pregnancy requires tighter glucose targets and much more frequent dose revision, since insulin resistance rises through gestation. Regimen decisions in pregnancy should be made with an obstetric and endocrine team using current professional-body guidance rather than general formulas.

Kidney impairment reduces insulin clearance, and doses that were correct at normal kidney function can become too high as kidney function declines, increasing hypoglycemia risk. This requires clinician-directed dose reduction and closer monitoring, not a fixed percentage adjustment applied without supervision.

Older adults, especially those with multiple health conditions, are at higher hypoglycemia risk partly because the body's warning signs of low glucose become less reliable with age. Guideline bodies generally support a somewhat less strict A1C target and simpler regimens (basal insulin alone, or basal plus a single bolus dose) in this group when appropriate, individualized to the person's overall health and goals of care.

What is established, what is plausible, and what is not established

Established: Basal bolus therapy, structured around separate long-acting and mealtime rapid-acting insulin, is a standard and guideline-supported approach for type 1 diabetes and for type 2 diabetes needing more than basal insulin alone. Carbohydrate counting improves bolus dosing accuracy compared with unstructured estimation. Insulin stacking from repeated correction doses is a real and preventable cause of severe hypoglycemia. Newer long-acting analogs have more predictable pharmacokinetic profiles than older formulations.

Plausible but not fully quantifiable from general sources: The exact size of A1C benefit from basal bolus versus premixed regimens, the exact percentage reduction in nighttime hypoglycemia with newer basal analogs, and exact correspondence between specific TIR percentages and A1C values. These directions are consistent with the literature, but a specific number should be pulled from a current systematic review or the ADA's current Standards of Care before being quoted to a patient as a precise figure.

Not established as a fixed rule for any individual: That the 500 rule, 1,700/1,800 rule, or any weight-based starting basal dose gives the correct number for a specific person without refinement. These are population-derived starting estimates explicitly intended to be adjusted using an individual's own glucose data.

Working with a clinician or diabetes educator

Starting numbers from these formulas are a beginning, not an endpoint. Most people need several weeks of systematic glucose logging, ideally with CGM, before their real ICR and correction factor are known with any confidence. A certified diabetes care and education specialist (CDCES) can often shorten that process and troubleshoot patterns that a formula cannot capture, such as delayed absorption from high-fat meals or exercise-related glucose swings. Bringing at least two weeks of glucose or CGM data to every visit, rather than a single recent reading, lets a clinician distinguish a true pattern problem from normal day-to-day variation.