When to Start Insulin in Type 2 Diabetes

When to Start Insulin in Type 2 Diabetes: 2026 Practical Algorithm

A practical, evidence-based guide to choosing therapy by A1C, cardiovascular disease, heart failure, kidney disease, obesity, liver disease, hypoglycemia risk and cost

For MarwahMedicine.com · Physician education article · Evidence updated to ADA 2026 and ADA/EASD 2026

THE 20-SECOND ANSWER
Start or strongly consider insulin when hyperglycemia is severe or symptomatic—especially A1C >10%, glucose ≥300 mg/dL, polyuria/polydipsia with dehydration, unexpected weight loss, ketosis/catabolism, hyperglycemic crisis, or suspected insulin deficiency. If severe hyperglycemia is absent, current ADA guidance prefers GLP-1–based therapy before insulin for initial or add-on glucose lowering. When insulin is needed, basal insulin is usually the simplest starting strategy

The clinical mistake to avoid: treating a laboratory number instead of the physiology. The patient with A1C 9.2% who is stable, overweight and asymptomatic may need a very different next step from the patient with A1C 11%, weight loss, polyuria and ketones.

The insulin question has changed

For decades, type 2 diabetes treatment was often taught as a ladder: lifestyle → metformin → more tablets → insulin. That ladder is now too crude. Modern diabetes care separates two questions: “How urgently must glucose be lowered?” and “Which therapy best improves the patient’s long-term cardiovascular, kidney and weight outcomes?”

Insulin remains the fastest and most reliable glucose-lowering therapy when insulin deficiency, catabolism or marked hyperglycemia is present. But when those red flags are absent, GLP-1 receptor agonists and dual GIP/GLP-1 therapy often provide powerful glucose lowering with weight loss and less hypoglycemia, and are preferred before insulin in current guidance.

1. When should insulin be started immediately or strongly considered?

Clinical trigger	Why it matters
A1C >10% (86 mmol/mol)	Very high chronic glycemic burden; insulin should be strongly considered, especially if symptomatic.
Blood glucose ≥300 mg/dL (16.7 mmol/L)	Marked hyperglycemia where rapid, dependable correction may be needed.
Polyuria, polydipsia, dehydration	Symptoms indicate clinically important hyperglycemia rather than an isolated laboratory abnormality.
Unexpected weight loss / ketosis	Suggests catabolism and relative or absolute insulin deficiency.
Hyperglycemic crisis or acute decompensation	Insulin is central to emergency management; outpatient stepwise escalation is inappropriate.
Possible type 1 diabetes / severe insulin deficiency	Do not delay insulin while classification is being clarified.

MARWAH PEARL
A1C 10.5% is not the same patient in every case. The presence of symptoms, catabolism, ketosis, dehydration, suspected type 1 diabetes and the pace of deterioration determine urgency. Numbers guide you; physiology decides.

2. When should you NOT rush to insulin?

In adults with type 2 diabetes who do not have severe hyperglycemia or a hyperglycemic crisis, the 2026 ADA Standards prefer GLP-1–based therapy to insulin for initial or add-on glucose lowering. This is a major practical shift from the old “HbA1c high, therefore insulin” reflex.

Obesity or weight gain is a major concern: GLP-1–based therapy can improve glycemia while promoting clinically meaningful weight loss.
The patient is worried about hypoglycemia: insulin can be used safely, but GLP-1–based therapy generally carries less intrinsic hypoglycemia risk.
The patient has cardiovascular or kidney disease: choose therapies for organ protection as well as glucose lowering; insulin lowers glucose but does not replace SGLT2/GLP-1 cardiorenal strategies.
A1C is above goal but the patient is clinically stable: first ask whether adherence, therapeutic inertia, dose optimization or a more effective noninsulin combination can close the gap.

3. A practical 2026 decision algorithm

STEP 1	Severe symptoms, catabolism, ketosis, crisis, suspected type 1, A1C >10% or glucose ≥300?	Start/strongly consider insulin. Correct dehydration and acute illness where relevant.
STEP 2	No severe hyperglycemia. Is weight loss, ASCVD, HF or CKD an important priority?	Use phenotype-directed therapy. GLP-1–based therapy and/or SGLT2 inhibitor may take priority, independent of HbA1c.
STEP 3	Still above individualized glycemic target despite optimized noninsulin therapy?	Escalate potency. If insulin is required, start with basal insulin in most patients.
STEP 4	Fasting glucose reaches target but A1C remains high?	Do not keep pushing basal insulin. Look for postprandial hyperglycemia / overbasalization.
STEP 5	Need injectable intensification?	If not already used, consider GLP-1 or dual GIP/GLP-1 before prandial insulin; otherwise add prandial insulin strategically.

4. How to start basal insulin

Basal insulin alone is usually the simplest initial insulin strategy in type 2 diabetes. Its main job is to suppress hepatic glucose output overnight and between meals.

STARTING DOSE
A practical starting basal dose is approximately 0.1–0.2 units/kg/day, adjusted for the degree of hyperglycemia, age, kidney function and hypoglycemia risk. Titrate using fasting glucose patterns rather than reacting to a single reading.

Example: a 90-kg patient might begin around 9–18 units/day. The exact choice should be individualized; a frail patient with CKD and recurrent hypoglycemia risk is not titrated like a younger, highly insulin-resistant patient with marked fasting hyperglycemia.

Which basal insulin?

NPH remains effective and can be less expensive. Long-acting analogs simplify once-daily therapy for many patients and reduce nocturnal/clinically significant hypoglycemia compared with NPH in trials. Longer-acting analogs such as degludec or U-300 glargine may further reduce nocturnal hypoglycemia compared with U-100 glargine in selected patients.

5. The biggest insulin trap: overbasalization

The wrong response to an elevated A1C is to keep increasing basal insulin even after fasting glucose is controlled. Basal insulin is designed primarily to fix fasting and between-meal glucose. If fasting glucose is at target but A1C remains high, the remaining problem is often postprandial.

Clues that should trigger a review for overbasalization include:

A large bedtime-to-morning glucose fall (ADA gives ≥50 mg/dL as an example).
A large postprandial-to-preprandial glucose differential.
Recurrent or unrecognized hypoglycemia.
High glycemic variability despite a “good” fasting glucose.
A1C above goal even though fasting glucose has already been corrected.

EXAM + CLINIC PEARL
“Fasting is fine, HbA1c is high” is not a command to keep increasing basal insulin. It is a command to look at meals, postprandial glucose and the next injectable strategy.

6. What comes after basal insulin?

If basal insulin has been titrated appropriately and the A1C remains above target, intensification should address postprandial glucose. Current guidance specifically recommends considering a GLP-1 receptor agonist or dual GIP/GLP-1 agonist before adding prandial insulin if the patient is not already receiving one. This can improve glucose control while reducing the weight gain and hypoglycemia burden that often accompanies more insulin.

If prandial insulin is required, a simple first move is to add one dose at the largest meal or the meal producing the greatest excursion. The ADA suggests a starting prandial dose of about 4 units or 10% of the basal dose, then individualize and intensify as needed.

7. What happens to the other diabetes drugs when insulin starts?

Do not automatically throw away every noninsulin drug the day insulin begins. In 2026, the goal is to preserve therapies that provide useful metabolic or organ-protective benefits while reducing agents that add hypoglycemia or treatment burden without meaningful additional benefit.

Drug class	Typical approach	Clinical logic
Metformin	Usually continue if tolerated and not contraindicated.	Maintains glucose benefit; low hypoglycemia risk.
SGLT2 inhibitor	Usually continue when indicated/tolerated.	Heart and kidney protection is not replaced by insulin.
GLP-1 RA / dual GIP-GLP-1	Continue or add when appropriate.	Improves glycemia; helps weight; may reduce insulin dose/hypoglycemia burden.
Sulfonylurea / meglitinide	Often reduce or discontinue as insulin is intensified.	Adds hypoglycemia and weight gain; limited organ-protection advantage.
DPP-4 inhibitor	Often limited/discontinued during injectable intensification.	Modest glucose effect and no meaningful weight/organ-outcome advantage.

8. Monitoring in 2026: CGM is no longer “only for type 1”

The 2026 ADA Standards recommend continuous glucose monitoring for adults on insulin therapy and extend CGM use more broadly when it aids management. In a patient starting insulin, CGM can reveal nocturnal lows, fasting patterns, meal excursions and variability far more clearly than an isolated fasting value.

9. Five high-yield clinical scenarios

A1C 11.4%, glucose 340 mg/dL, polyuria and 6-kg weight loss

→ Insulin now. The patient is symptomatic and catabolic; do not waste weeks on slow stepwise escalation.

A1C 9.2%, BMI 34 kg/m², no symptoms, no ketosis

→ Do not reflexively start insulin. A potent GLP-1–based strategy may be preferable if clinically appropriate.

Basal insulin 34 units, fasting glucose 105 mg/dL, A1C 8.7%

→ Think postprandial hyperglycemia/overbasalization. Do not blindly raise basal insulin.

T2DM + CKD + albuminuria, A1C 7.2%

→ Insulin decision is not the main organ-protection question. Preserve kidney-protective therapy as indicated; HbA1c does not capture residual cardiorenal risk.

Basal insulin + recurrent nocturnal hypoglycemia

→ Reassess dose, timing, kidney function, meals and insulin type; consider CGM and simplification rather than chasing a lower fasting number.

10. The one-line rule to remember

THE RULE
Insulin is the right drug when physiology demands insulin. If severe hyperglycemia or catabolism is absent, choose the next therapy by the patient’s phenotype—not by fear of an HbA1c number.

Frequently asked questions

Is A1C >10% an absolute rule that every patient must start insulin?

No. It is a strong trigger to consider insulin, especially when symptoms or catabolic features are present. Clinical context matters.

Should GLP-1 therapy be tried before insulin?

In adults with type 2 diabetes without severe hyperglycemia or hyperglycemic crisis, ADA 2026 prefers GLP-1–based therapy before insulin for initial or add-on glucose lowering.

What is the usual basal insulin starting dose?

Approximately 0.1–0.2 units/kg/day for many adults, individualized to hyperglycemia and hypoglycemia risk.

When should prandial insulin be added?

When basal insulin has corrected fasting glucose but A1C/postprandial glucose remains high and injectable intensification is still needed. If GLP-1–based therapy has not been used, consider it before prandial insulin.

What is a practical starting prandial dose?

About 4 units or 10% of the basal insulin dose at the largest meal or greatest postprandial excursion, followed by individualized titration.

References

American Diabetes Association Professional Practice Committee. Pharmacologic Approaches to Glycemic Treatment: Standards of Care in Diabetes—2026. Diabetes Care. 2026;49(Suppl 1):S183–S215. doi:10.2337/dc26-S009. Source
Davies MJ, Aroda VR, Bajaj HS, et al. Management of Type 2 Diabetes, 2026: A Consensus Report by the ADA and EASD. Diabetes Care / Diabetologia. Published October 2, 2026. doi:10.2337/dci26-0141. Source
American Diabetes Association. Standards of Care in Diabetes—2026. Released December 8, 2025. Source

Editorial note: This article is for medical education and should be adapted to individual clinical context, local product labels and patient-specific risk. It does not replace clinician judgment.