Fasting Insulin and Drugs That Distort This Test

What this test measures, and what it is not

Fasting insulin is the serum concentration of a person's own (endogenous) insulin after an overnight fast, usually specified as 8 to 12 hours with water only. It is a different measurement from C-peptide, from a post-meal insulin level, and from HOMA-IR, which is a calculated index that combines fasting insulin with fasting glucose rather than a separate lab draw.

Reference ranges differ by laboratory and assay, so "normal" on one lab report is not automatically normal on another. Some clinicians use a tighter working target, often cited informally as under roughly 8-10 mcIU/mL, for metabolically healthy adults not on any glucose- or insulin-altering medication, but this is a practice pattern rather than a single agreed diagnostic cutoff, and readers should confirm the specific threshold their ordering clinician is using.

The core problem with this test is that it is unusually sensitive to what a person was doing and taking in the days before the draw. A single recent dose of a corticosteroid, months of treatment with a GLP-1 receptor agonist, a hemolyzed sample, or a draw taken during acute illness can each move the number in a clinically misleading direction. Reading a fasting insulin result without knowing the medication list and pre-analytical conditions at the time of the draw is reading an incomplete number, not a definitive one.

Why this matters for insulin resistance and PCOS workups

Insulin resistance is thought to precede type 2 diabetes by years, and fasting insulin is sometimes used as an early signal of compensatory beta-cell hypersecretion, before fasting glucose itself rises. HOMA-IR (fasting insulin in mcIU/mL multiplied by fasting glucose in mmol/L, divided by 22.5) is a commonly cited research index built from this value, though the exact cutoff used to call a result "insulin resistant" varies across studies and has not been standardized into a single clinical diagnostic threshold.

In polycystic ovary syndrome (PCOS), insulin resistance is common and clinically relevant to treatment planning. The Endocrine Society's PCOS clinical practice guideline addresses insulin resistance as part of the broader metabolic risk profile in PCOS and supports metformin as a first-line pharmacologic option for insulin resistance in women not attempting conception. Fasting insulin itself is not a required Rotterdam diagnostic criterion for PCOS. (Endocrine Society, PCOS Clinical Practice Guideline: https://www.endocrine.org/clinical-practice-guidelines/polycystic-ovary-syndrome)

Drugs reported to push fasting insulin upward

The following drug classes have a well-established pharmacologic rationale for raising fasting insulin by inducing peripheral insulin resistance, which forces the pancreas to secrete more insulin to hold glucose steady. Specific percentage changes from individual trials are not reproduced here because the precise figures require verification against the primary literature; the direction and general mechanism are well supported.

Corticosteroids. Prednisone, dexamethasone, methylprednisolone, and high-dose inhaled corticosteroids increase peripheral insulin resistance in a dose-dependent way, and short courses of oral corticosteroids have been documented to raise fasting insulin meaningfully above a person's own baseline within days. This is one of the most reproducible drug effects on this test.

Atypical (second-generation) antipsychotics. Olanzapine, clozapine, and quetiapine are associated with weight gain, impaired beta-cell function, and reduced peripheral glucose uptake, all of which raise fasting insulin over time. Regulatory labeling for this drug class addresses hyperglycemia and diabetes risk; readers and clinicians should check the current FDA label for the specific agent in use rather than relying on a general statement, since labeling detail and warnings can be updated. (General FDA drug information: https://www.fda.gov/drugs)

Thiazide diuretics. Hydrochlorothiazide and chlorthalidone have long been associated with increased new-onset diabetes risk in hypertension trials, plausibly mediated in part through hypokalemia and its effect on insulin secretion and peripheral glucose disposal. A fasting insulin drawn in a patient on chronic thiazide therapy should be interpreted with this in mind.

High-dose niacin. Niacin used at lipid-modifying doses (roughly 1,000-2,000 mg/day) increases free fatty acid mobilization and has been associated with worsened glycemic control in trial populations, consistent with a compensatory rise in insulin secretion in insulin-resistant patients.

Other agents worth flagging. High-dose ethinyl estradiol-containing oral contraceptives may modestly raise fasting insulin in susceptible women, particularly those with PCOS or a family history of type 2 diabetes. Calcineurin inhibitors (tacrolimus, cyclosporine) can cause direct beta-cell toxicity, which produces a different and more concerning pattern: low endogenous insulin with high glucose, not simple elevation. Non-selective beta-blockers such as propranolol can also blunt the counterregulatory response to hypoglycemia and modestly affect insulin secretion.

Drugs reported to push fasting insulin downward

A low fasting insulin result during treatment with certain drugs usually reflects the medication working as intended, not beta-cell failure, but this distinction depends entirely on knowing the medication history.

GLP-1 receptor agonists. Semaglutide, liraglutide, tirzepatide, and dulaglutide reduce the secretory burden on the beta cell by improving peripheral insulin sensitivity and suppressing inappropriately high glucagon secretion. Large obesity and diabetes outcome trials for this drug class (for example, the STEP program for semaglutide) have reported meaningful weight loss and improvement in insulin sensitivity markers; a low fasting insulin in a patient established on one of these drugs is an expected therapeutic finding, not evidence of hypoinsulinism.

SGLT-2 inhibitors. Empagliflozin, dapagliflozin, and canagliflozin lower fasting insulin indirectly, by reducing ambient glucose (via urinary excretion) and therefore reducing the glucose-driven demand on the beta cell. This has been documented in large cardiovascular and renal outcome trials for this drug class.

Metformin. Metformin primarily reduces hepatic glucose production and improves peripheral insulin sensitivity, both of which lower the secretory demand on the beta cell. Fasting insulin in a patient on metformin is generally expected to run lower than in a metabolically similar, drug-naive comparator. This is a therapeutic effect that confounds baseline screening if metformin is started before a diagnostic fasting insulin is drawn.

Thiazolidinediones. Pioglitazone and rosiglitazone improve peripheral insulin sensitivity directly (through PPAR-gamma activation) without stimulating insulin secretion, so the net effect on fasting insulin is a genuine reduction that should be read as a pharmacologic benefit rather than a pathologic finding.

The sulfonylurea paradox

Sulfonylureas (glipizide, glimepiride, glyburide) do not fit neatly into either category. They acutely stimulate insulin secretion, which can raise an insulin level drawn soon after dosing. With years of chronic use, however, they are associated with progressive beta-cell exhaustion, which can eventually lower fasting insulin. A single fasting insulin value in a patient on long-term sulfonylurea therapy says relatively little about that person's underlying beta-cell reserve without knowing the treatment duration and trend over time.

A decision framework for interpreting fasting insulin against the medication list

Use this sequence before treating a fasting insulin result as clinically meaningful. It does not replace clinical judgment or a qualified reviewer, and it does not provide individualized dosing or diagnostic advice.

If a medication in step 2 or 3 cannot be safely stopped for a single morning, do not stop it without the prescribing clinician's guidance. The more reliable fix is to document the exact drug and dose on the lab requisition and, where clinically appropriate, plan a repeat fasting insulin 8-12 weeks after any significant medication change to separate a drug effect from a true physiologic shift.

Physiologic confounders beyond medications

Fasting insulin can also be pushed up or down by factors that have nothing to do with prescription drugs:

• Obesity and visceral adiposity are associated with higher fasting insulin in population studies, though the exact magnitude of the relationship varies by study population.
• Acute psychological or physical stress raises cortisol, which can transiently increase insulin secretion.
• Pregnancy, especially the third trimester, is associated with physiologic insulin resistance.
• Prolonged caloric restriction, very low-carbohydrate eating, and regular exercise are generally associated with lower fasting insulin.
• Puberty (Tanner stages 2-4) is associated with a physiologic rise in insulin levels independent of body composition or diet; a result that would be concerning in an adult may be expected in a young adolescent.
• Hemolysis in the collection tube can degrade insulin in vitro and produce a spuriously low result; an unexpectedly low value with an otherwise inconsistent clinical picture should prompt a repeat draw with careful technique.

One line of animal research is worth naming honestly rather than omitting: diet-induced insulin resistance and associated hepatic oxidative stress have been studied in rodent models, for example in high-fructose-fed rats, where interventions aimed at reducing oxidative stress improved insulin sensitivity markers (Beneficial Effects of Phyllanthus amarus Against High Fructose Diet Induced Insulin Resistance and Hepatic Oxidative Stress in Male Wistar Rats, 2017: https://pubmed.ncbi.nlm.nih.gov/28353042/). This is rodent, not human, evidence, and it does not establish that any particular supplement or dietary intervention will change a human fasting insulin result. It is included here only to illustrate that diet-driven insulin resistance is mechanistically plausible and actively studied, not as a treatment recommendation.

Preparing for a fasting insulin draw

• Fast 8-12 hours, water only; a 12-hour fast is preferable for consistency if the test will be repeated over time.
• Avoid a large high-carbohydrate meal and vigorous exercise in the 24 hours before the draw, since both can shift the result temporarily.
• Bring a complete list of prescription drugs, over-the-counter medications, and supplements to the visit so the ordering clinician can decide whether the result needs to be interpreted with a drug effect in mind.
• Do not stop a prescribed corticosteroid, antipsychotic, thiazide, GLP-1 agonist, SGLT-2 inhibitor, or other regularly scheduled medication before the draw without talking to the prescribing clinician; a fasting insulin drawn on a known medication and interpreted with that context is safer and more useful than one drawn after an unsupervised medication hold.
• When possible, order a baseline fasting insulin before starting a new medication known to affect insulin sensitivity, so later results can be compared against a true pre-treatment number.

Special populations

PCOS. Insulin resistance is common in PCOS across body weights, and the Endocrine Society's guideline supports addressing it as part of PCOS management even though fasting insulin is not a required diagnostic criterion. An elevated fasting insulin in a lean woman with PCOS can still be clinically meaningful even when it falls inside a lab's general reference range.

Hormone therapy. Testosterone therapy in hypogonadal men and different formulations of menopausal hormone therapy in women can each shift insulin sensitivity, with some evidence suggesting oral estrogen and androgenic progestins behave differently from transdermal estrogen and micronized progesterone. The magnitude and consistency of these effects vary across studies, and a fasting insulin drawn during a hormone therapy transition should be interpreted with the therapy timeline in mind rather than as a fixed number.

Children and adolescents. Reference ranges are age- and pubertal-stage dependent. A result that would flag as high in an adult may be an expected feature of mid-puberty.

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

Established: corticosteroids, atypical antipsychotics, thiazide diuretics, and high-dose niacin are pharmacologically expected to raise fasting insulin through induced peripheral insulin resistance, and GLP-1 receptor agonists, SGLT-2 inhibitors, metformin, and thiazolidinediones are pharmacologically expected to lower it by reducing the secretory demand on the beta cell. These mechanisms are grounded in the known pharmacology of each drug class.

Plausible but not precisely quantified here: the exact magnitude of change (for example, a specific percentage rise on a specific corticosteroid dose) cannot be stated with confidence from the sources available for this article and should be verified against the primary trial or pharmacokinetic literature before being used in a clinical or patient-facing claim.

Not established: there is no single universally agreed numeric cutoff for "insulin resistant" fasting insulin or HOMA-IR that applies across all laboratories, assays, and populations; any specific threshold cited by a clinician or lab should be understood as a practice convention, not a fixed diagnostic law.

When to seek urgent care rather than wait for interpretation

A fasting insulin result is not an emergency test. However, a fasting insulin drawn as part of a hypoglycemia workup that returns very high alongside a very low glucose can indicate an insulinoma or another cause of hyperinsulinemic hypoglycemia, which does need prompt medical evaluation. Symptoms of a low blood sugar event, such as confusion, sweating, tremor, or loss of consciousness, warrant urgent care regardless of a pending lab result.

Clinical decision checklist

Before treating a fasting insulin result as informative, confirm:

1. The fast was a true 8-12 hours with no caloric intake.
2. Insulin-elevating drugs (corticosteroids, atypical antipsychotics, thiazides, niacin) are documented, if present.
3. Insulin-lowering drugs (GLP-1 agonists, SGLT-2 inhibitors, metformin, thiazolidinediones) are documented, if present.
4. The clinical picture (BMI, waist circumference, HbA1c, fasting glucose, lipids) is consistent with the insulin value.
5. The sample was not hemolyzed.

A fasting insulin value without its accompanying medication and pre-analytical context is an incomplete piece of clinical data, not a finished one.