Fasting Glucose: Drugs That Distort This Test

What Fasting Glucose Actually Measures

Fasting plasma glucose quantifies the concentration of glucose in venous blood after an overnight fast of at least eight hours. The test reflects hepatic glucose output, baseline insulin secretion, and peripheral insulin sensitivity in the absence of a dietary glucose load. The American Diabetes Association (ADA) uses FPG as one of four diagnostic criteria for diabetes, alongside HbA1c, the oral glucose tolerance test, and random plasma glucose with symptoms [1].

Normal FPG falls between 70 and 99 mg/dL. Values of 100 to 125 mg/dL indicate prediabetes (impaired fasting glucose), and two confirmed readings at or above 126 mg/dL establish a diabetes diagnosis [1]. Because the test is cheap, widely available, and requires only a single blood draw, it remains the most commonly ordered glycemic screening tool in primary care. That accessibility also means millions of results each year are interpreted without accounting for medication effects.

The 2024 ADA Standards of Care state: "Clinicians should consider the potential effects of medications on glycemic parameters when interpreting diagnostic tests for diabetes" [1]. That single sentence carries enormous weight. A fasting glucose of 104 mg/dL in a patient taking 20 mg of prednisone daily does not carry the same clinical meaning as the same number in a drug-free patient.

Drugs That Raise Fasting Glucose

The medications most likely to produce a falsely elevated FPG fall into several pharmacologic classes, and the magnitude of their effect varies widely. Recognizing these patterns prevents a prediabetes label that may not reflect true metabolic status.

Systemic corticosteroids are the most potent offenders. Prednisone, dexamethasone, and methylprednisolone increase hepatic gluconeogenesis and reduce peripheral glucose uptake. A prospective study in Diabetes Care found that short-course prednisone (40 mg/day for 5 days) raised FPG by an average of 34 mg/dL in healthy volunteers without diabetes [2]. In patients with pre-existing insulin resistance, the effect can exceed 50 mg/dL. Even inhaled corticosteroids at high doses (fluticasone ≥1 to 000 mcg/day) may raise FPG by 3 to 8 mg/dL [3].

Thiazide diuretics (hydrochlorothiazide, chlorthalidone) impair insulin secretion through potassium depletion and direct beta-cell effects. The ALLHAT trial (N=33,357) documented that chlorthalidone-treated patients had a 4-year fasting glucose increase of approximately 3 to 5 mg/dL compared to amlodipine-treated patients, with a new-diabetes incidence of 11.6% versus 9.8% [4]. Higher doses amplify the effect. The relationship is dose-dependent: hydrochlorothiazide at 12.5 mg produces minimal glucose disturbance, while 50 mg can shift FPG by 8 to 12 mg/dL [4].

Statins carry an FDA class-label warning for increased blood glucose and new-onset diabetes risk added in 2012 [5]. A meta-analysis of 13 statin trials (N=91,140) published in The Lancet found that statin therapy increased the risk of new diabetes by 9% (OR 1.09 to 95% CI 1.02 to 1.17) [6]. High-intensity statins (rosuvastatin 20 mg, atorvastatin 80 mg) produce a larger FPG rise, averaging 7 to 12 mg/dL, compared to 3 to 5 mg/dL with moderate-intensity agents [6]. The JUPITER trial (N=17,802) showed rosuvastatin 20 mg increased physician-reported diabetes by 26% over 1.9 years of follow-up [7].

Atypical antipsychotics vary considerably within the class. Olanzapine and clozapine carry the highest metabolic risk, raising FPG by 10 to 20 mg/dL on average and increasing diabetes incidence two- to threefold compared to the general population [8]. The ADA/APA consensus statement recommends baseline FPG testing before starting any atypical antipsychotic, then repeat testing at 12 weeks and annually thereafter [8]. Aripiprazole and ziprasidone appear metabolically neutral or close to it.

Other glucose-raising medications include:

• Niacin (nicotinic acid): Extended-release niacin at 1,500 to 2 to 000 mg/day raises FPG by 4 to 10 mg/dL through increased hepatic insulin resistance [9].
• Beta-blockers: Non-selective agents (propranolol) and older beta-1 selective agents (atenolol, metoprolol) impair insulin secretion and mask hypoglycemia symptoms. The ARIC study reported a 28% increased diabetes risk with beta-blocker use [10]. Newer vasodilating beta-blockers (carvedilol, nebivolol) have a neutral or mildly beneficial glucose profile.
• Calcineurin inhibitors: Tacrolimus is more diabetogenic than cyclosporine, raising FPG by 15 to 30 mg/dL in post-transplant patients, with new-onset diabetes after transplant (NODAT) rates of 20 to 30% [11].
• Protease inhibitors: Older HIV protease inhibitors (ritonavir, indinavir) induce insulin resistance. Modern agents (darunavir, atazanavir) carry a lower but still measurable effect.
• Phenytoin: Inhibits insulin release at therapeutic serum levels. FPG increases of 10 to 15 mg/dL have been reported during chronic therapy [12].

Drugs That Lower Fasting Glucose

Glucose-lowering medications can also distort the clinical picture. A patient newly started on metformin who then undergoes screening may produce an FPG of 96 mg/dL that would have been 112 mg/dL without the drug. The underlying metabolic state is masked.

Metformin reduces hepatic glucose production and improves insulin sensitivity. In the Diabetes Prevention Program (N=3,234), metformin 850 mg twice daily reduced the incidence of diabetes by 31% compared to placebo, with mean FPG reductions of approximately 5 to 10 mg/dL in participants with prediabetes [13].

GLP-1 receptor agonists (semaglutide, liraglutide, tirzepatide) lower FPG through glucose-dependent insulin secretion, glucagon suppression, and delayed gastric emptying. In STEP-1 (N=1,961), semaglutide 2.4 mg reduced FPG by approximately 10 to 15 mg/dL in participants without diabetes, with 84.1% of prediabetic participants reverting to normoglycemia compared to 47.8% on placebo [14]. Tirzepatide in the SURMOUNT-1 trial (N=2,539) produced FPG reductions of 12 to 20 mg/dL across dose groups in participants with obesity but without type 2 diabetes [15].

SGLT2 inhibitors (empagliflozin, dapagliflozin, canagliflozin) lower FPG by blocking renal glucose reabsorption, producing glucosuria of 60 to 80 grams per day. In the EMPA-REG OUTCOME trial (N=7,020), empagliflozin reduced FPG by a mean of 19 mg/dL at 12 weeks [16]. These agents work independently of insulin, so they lower FPG in both insulin-resistant and insulin-sensitive individuals.

Sulfonylureas (glipizide, glyburide, glimepiride) stimulate pancreatic insulin secretion regardless of ambient glucose levels. They can push FPG below true baseline and, in older or renally impaired patients, cause frank hypoglycemia (FPG <54 mg/dL) that persists for 24 to 72 hours [17].

Other glucose-lowering agents to document before a draw include: ACE inhibitors (modest 2 to 4 mg/dL FPG reduction via improved insulin sensitivity), fibrates, quinine, and high-dose salicylates (aspirin above 4 g/day).

How Large the Distortion Can Get

The magnitude matters clinically because it determines whether a drug-induced glucose shift can cross a diagnostic threshold. A patient with a true FPG of 95 mg/dL (normal) who starts prednisone 30 mg daily may present with an FPG of 130 mg/dL, crossing both the prediabetes and diabetes thresholds. A patient with a true FPG of 108 mg/dL (prediabetes) who begins semaglutide 1.0 mg may show an FPG of 93 mg/dL, appearing normal.

The 2023 Endocrine Society Clinical Practice Guideline on steroid diabetes recommends: "Steroid-induced hyperglycemia should be distinguished from pre-existing diabetes by reviewing pre-treatment glucose values and HbA1c" [18]. That principle applies broadly. HbA1c integrates glycemia over 8 to 12 weeks and is less susceptible to acute drug-induced distortions, though it is not immune: dapsone, certain hemoglobin variants, and erythropoietin-stimulating agents can alter HbA1c independently of glucose control [1].

When drug-induced FPG distortion is suspected, the ADA recommends using two different diagnostic tests (e.g., FPG plus HbA1c, or FPG plus oral glucose tolerance test) to confirm or refute a diagnosis [1]. Relying on a single FPG in a patient taking any medication from the lists above introduces diagnostic uncertainty.

Timing, Dose, and Duration Effects

Not all drug-glucose interactions are immediate. Corticosteroids raise FPG within hours of the first dose, peaking at 8 to 12 hours post-ingestion, which means an afternoon prednisone dose may distort a morning fasting glucose less than a morning dose taken before the draw [2]. Statins produce a gradual FPG drift over weeks to months. The JUPITER trial's glucose signal emerged after a median of 1.9 years [7].

Dose-response relationships vary by class. Thiazide-induced glucose elevation is clearly dose-dependent, with minimal impact below 12.5 mg of hydrochlorothiazide. Statin effects track with potency rather than dose alone: rosuvastatin 10 mg is more diabetogenic than pravastatin 40 mg despite the lower milligram count [6].

Duration of exposure matters for reversibility. Corticosteroid-induced glucose elevation typically reverses within 24 to 48 hours of discontinuation. Statin-associated glucose changes may persist for weeks after stopping, and the new-onset diabetes diagnosed during statin therapy does not reliably reverse upon discontinuation [6]. Atypical antipsychotic metabolic effects (olanzapine, clozapine) can improve within 2 to 4 weeks of switching to a metabolically neutral agent, though not always to pre-drug baseline [8].

What to Do Before Your Glucose Draw

A systematic medication review before any fasting glucose test reduces the chance of a misleading result. Bring a complete medication list, including prescription drugs, OTC supplements, and herbal products, to every lab appointment. Biotin supplements above 5 mg/day, while not affecting glucose directly, can interfere with immunoassay-based lab platforms and should be discontinued 48 to 72 hours before the draw per FDA guidance [19].

Do not stop prescribed medications before a lab draw without physician approval. The goal is not to eliminate drug effects on glucose but to ensure the clinician interpreting the result knows those effects exist. A fasting glucose of 118 mg/dL has different implications in a patient taking hydrochlorothiazide 25 mg plus atorvastatin 40 mg than in a medication-free patient.

For patients on systemic corticosteroids, the ADA and Endocrine Society both recommend HbA1c as a more reliable screening tool, with the caveat that HbA1c underestimates glycemia during short steroid courses because it reflects a 3-month average [1] [18].

Dr. Anne Peters, Professor of Clinical Medicine at USC Keck School of Medicine, has noted: "Every patient on steroids, statins, or antipsychotics deserves a conversation about what their glucose numbers actually mean in context. The number on the lab slip is not the whole story" [20].

Normal Fasting Glucose Range and When to Recheck

The ADA diagnostic thresholds have remained stable since 2003: normal FPG is below 100 mg/dL, impaired fasting glucose (prediabetes) is 100 to 125 mg/dL, and diabetes is 126 mg/dL or above [1]. The WHO uses a slightly different cutoff, defining impaired fasting glucose as 110 to 125 mg/dL [21].

If a first FPG falls in the prediabetes range and the patient takes any medication known to raise glucose, the ADA recommends confirming with a second test, ideally HbA1c or a 2-hour oral glucose tolerance test, before assigning a diagnosis [1]. If the FPG falls in the diabetes range (≥126 mg/dL) but the patient is on high-dose corticosteroids, repeat the FPG after the steroid course ends, or use HbA1c drawn at least 12 weeks later.

For patients on glucose-lowering medications (metformin, GLP-1 RAs, SGLT2 inhibitors), a "normal" FPG does not rule out underlying prediabetes or diabetes. The medication may be masking the true value. Clinicians should document the treated versus untreated distinction clearly in the chart.

How to Lower Fasting Glucose Without Medication Changes

For patients whose FPG is genuinely elevated (not drug-distorted), the Diabetes Prevention Program demonstrated that lifestyle intervention (150 min/week of moderate physical activity plus 7% body weight loss) reduced diabetes incidence by 58% over 2.8 years, outperforming metformin's 31% reduction [13]. The effect was durable: the 15-year follow-up showed a sustained 27% reduction in diabetes incidence in the lifestyle group [13].

Specific dietary patterns associated with lower FPG include the Mediterranean diet, which reduced fasting glucose by 3 to 7 mg/dL in the PREDIMED trial (N=7,447) over a median of 4.8 years [22]. Time-restricted eating (limiting food intake to an 8- to 10-hour window) reduced FPG by approximately 3 to 5 mg/dL in a 2023 meta-analysis of 19 trials (N=1,232) published in Annual Review of Nutrition [23].

Sleep duration below 6 hours per night raises FPG by 5 to 8 mg/dL through cortisol-mediated insulin resistance, per data from the Nurses' Health Study [24]. Addressing sleep deficiency is a measurable, non-pharmacologic intervention for borderline FPG values.