Provocative growth hormone testing — insulin tolerance, arginine, clonidine & glucagon protocols to diagnose GH deficiency in short-stature children
A provocative GH stimulation test is never a first-line investigation — it is ordered only after auxology and screening labs raise real suspicion of growth hormone deficiency (GHD). Because the test is invasive, time-consuming, and carries genuine risk (especially with insulin), getting the pre-test workup right prevents both missed diagnoses and unnecessary provocative testing.
GH stimulation testing sits at the end of a diagnostic funnel, not the start of it:
• Auxology first: height velocity plotted on a growth chart over at least 6–12 months, height >2 SD below the population mean or crossing two major percentile lines, and a predicted adult height significantly below mid-parental target height • Bone age: a delayed bone age (by hand-wrist X-ray, Greulich-Pyle atlas) relative to chronological age is consistent with, though not diagnostic of, GHD • Screening labs: IGF-1 and IGFBP-3 (insulin-like growth factor binding protein-3) below age- and pubertal-stage-adjusted norms — these are GH-dependent proteins that integrate 24-hour GH secretion and are far more stable than a single random GH level • Exclude other causes first: hypothyroidism, celiac disease, chronic renal disease, skeletal dysplasia, Turner syndrome (girls), psychosocial deprivation, and constitutional delay of growth and puberty (CDGP) — the most common mimic of GHD and the hardest to distinguish clinically
Only when growth failure plus low IGF-1/IGFBP-3 persist after these are excluded does a child proceed to provocative testing, because a single random GH level is useless — physiologic pulsatile secretion means most levels drawn at any random moment are near zero even in normal children.
Because GH is secreted in pulses only 6–10 times per day (mostly during slow-wave sleep), a random daytime GH level is diagnostically worthless — provocative stimulation exists specifically to force a reproducible, maximal secretory burst that can be measured and compared to a threshold.
One of the most clinically important refinements to GH stimulation testing is priming: giving a short course of sex steroids to prepubertal children (typically girls under 8 or boys under 9 with clear evidence of impending puberty) 1–3 days before the test.
Why priming matters: • Estrogen and testosterone amplify pituitary somatotroph responsiveness to GHRH and to provocative stimuli — this is part of the normal pubertal growth spurt physiology • A prepubertal, unprimed child can show a blunted GH peak that looks like deficiency purely because the somatotrophs have not yet been sensitized by rising gonadal steroids — a false positive • Priming regimens: ethinyl estradiol 20 mcg orally twice daily for 2 days before testing in girls; a single intramuscular depot testosterone injection about a week before testing in boys • Priming is NOT used in children already in puberty (endogenous sex steroids are already present) or in adults
Fasting preparation: • Test performed after an overnight fast (typically 8–10 hours) since feeding suppresses GH secretion and blunts hypoglycemic stimuli needed for the insulin tolerance test • IV access secured before the test begins — most protocols draw baseline glucose and GH through an indwelling catheter to avoid repeated venipuncture stress, which itself can spuriously elevate GH
With the child fasted, primed if indicated, and an IV line secured, the test begins with a true time-zero blood draw for baseline GH and glucose, immediately followed by administration of a pharmacologic stimulus known to trigger a reproducible pituitary GH pulse. No single agent is perfect — each has a distinct mechanism, risk profile, and typical time-to-peak, which is why guidelines require confirmation with a second, mechanistically different agent.
The time-zero (T=0) sample is drawn immediately before the agent is given and serves two purposes: it establishes the true unstimulated GH level (usually low, 0.1–3 ng/mL) and it confirms a normal starting glucose, which is essential before proceeding to insulin-induced hypoglycemia.
A physician or trained nurse must remain at bedside for the entire test, with continuous access to IV dextrose (25–50% dextrose) and hydrocortisone in case of severe or symptomatic hypoglycemia during the insulin tolerance test. Vital signs (heart rate, blood pressure, mental status) are checked at every sampling interval.
Each agent triggers GH release through a different physiologic pathway, which is precisely why combining two mechanistically distinct agents reduces the false-positive rate of a single test:
• Arginine: a cationic amino acid infused IV that suppresses hypothalamic somatostatin (the tonic inhibitor of GH release), disinhibiting spontaneous GHRH-driven pulses. Very safe, mild nausea only, but produces a comparatively modest peak on its own — often combined with GHRH or used as the confirmatory second test. • Clonidine: an oral α2-adrenergic agonist that stimulates hypothalamic GHRH release centrally. Main risks are somnolence and symptomatic hypotension; requires blood-pressure monitoring throughout. • Glucagon: an intramuscular injection that produces a delayed rebound hypoglycemia (peaking 90–180 minutes post-injection) which secondarily triggers GH and cortisol release. Common side effects include nausea and vomiting in the first hour, which can be mistaken for a reaction to the test itself. • Insulin (insulin tolerance test, ITT): IV insulin bolus causes deliberate hypoglycemia, the single most potent physiologic stimulus to GH (and ACTH/cortisol) secretion, making the ITT the gold standard against which other agents are validated — but it is also the riskiest, carrying a real risk of seizure or loss of consciousness if glucose is not tightly monitored and rescued.
The insulin tolerance test remains the reference standard specifically because deliberate, monitored hypoglycemia is the single most reliable physiologic trigger of maximal GH secretion — but it is relatively contraindicated in young children, those with seizure disorders, and anyone with cardiovascular disease, which is why arginine, clonidine, and glucagon exist as safer alternatives.
| Product | Indication | Trial Design | Key Result |
|---|---|---|---|
| Insulin (ITT) | 0.05–0.1 U/kg IV bolus | Deliberate hypoglycemia — the most potent physiologic GH trigger; also tests ACTH/cortisol reserve | Gold standard, but highest risk (seizure, LOC) — needs continuous bedside monitoring |
| Arginine | 0.5 g/kg IV over 30 min | Suppresses hypothalamic somatostatin tone, disinhibiting endogenous GHRH pulses | Very safe, minimal side effects, often paired as the second confirmatory agent |
| Clonidine | 0.15 mg/m² oral, single dose | α2-adrenergic agonism drives central hypothalamic GHRH release | Oral, no injection needed — but causes drowsiness and can drop blood pressure |
| Glucagon | 1 mg IM (1.5 mg if >90 kg) | Delayed rebound hypoglycemia (90–180 min) secondarily triggers GH and cortisol release | Well tolerated in outpatients; nausea/vomiting common in the first hour |
After the stimulating agent is given, blood is drawn on a fixed schedule — typically at 0, 30, 60, 90, and 120 minutes — to trace out the full time course of the GH secretory burst. The shape of this curve, not just a single number, tells the clinician whether the pituitary somatotrophs mounted a normal, timely, and adequately sustained response.
A normal GH stimulation curve rises from a low baseline, reaches a clear peak somewhere in the sampling window, and gradually declines as the pulse resolves. Clinicians look at more than the single highest value:
• Time to peak: insulin and arginine typically peak around 45–60 minutes; glucagon is markedly delayed, often 120–180 minutes, which is why some glucagon protocols extend sampling to 180 or even 240 minutes • Shape of the curve: a sharp, well-defined peak is more reassuring than a flat, sluggish rise even if the peak value is borderline • Concordance with the physiologic trigger: for the ITT, the GH peak should track closely with the glucose nadir — a GH rise without adequate hypoglycemia invalidates the test and it must be repeated
Each sample is analyzed for both GH and glucose (for the ITT) so the metabolic stimulus can be verified alongside the hormonal response.
The ITT is the riskiest of the four provocative tests because it deliberately drives blood glucose low enough to guarantee a maximal counter-regulatory GH and cortisol response — typically targeting a glucose nadir below 40 mg/dL or a >50% fall from baseline.
Safety protocol during the ITT: • Continuous clinical observation for neuroglycopenic symptoms: sweating, tremor, confusion, lethargy, or seizure • Point-of-care glucose checked at every sampling timepoint, and more frequently if symptoms develop • IV dextrose (25% or 50%) immediately available at the bedside for rescue if the child becomes symptomatic or glucose falls to a dangerous level • Test is terminated early with rescue dextrose once adequate hypoglycemia and a GH sample have been obtained, even if the full time course is not completed • Physician (not nurse alone) must be present throughout given the risk of seizure or loss of consciousness
Because a valid ITT requires confirmed symptomatic hypoglycemia, roughly this is the one provocative test where the adverse effect the protocol is designed to avoid is also the marker of test validity — glucose must fall enough to prove the stimulus worked, but not so far or so uncontrolled that it becomes dangerous.
Once all timepoints are collected and assayed, the single highest GH value across the entire curve — the peak — is compared against a diagnostic cutoff. But because assay variability, pubertal status, and test-to-test biological noise all affect this number, no single stimulation test is considered sufficient on its own to diagnose GH deficiency.
For decades, a peak GH below 10 ng/mL on provocative testing was the standard cutoff for biochemical GH deficiency, based on older polyclonal radioimmunoassays (RIA). Modern immunochemiluminescent assays (ICMA) are more specific for the 22-kDa monomeric form of GH and read systematically lower than the old RIAs for the same true secretion — so many pediatric endocrine societies now use a lower cutoff, often <7 ng/mL, calibrated to the specific assay platform used.
This assay-dependence is a critical pitfall: a peak of 8 ng/mL could be labeled "deficient" on an old-style RIA cutoff and "normal" on some modern-assay cutoffs. Every laboratory result must be interpreted against the cutoff validated for that specific assay, not a single universal number.
Even in healthy, normally-growing children, a single GH stimulation test can produce a subnormal peak in up to 15–20% of cases — a false-positive rate far too high to base a lifelong diagnosis and rhGH therapy decision on one result. Sources of this noise include:
• Inherent biological variability of pulsatile GH secretion, independent of the stimulus used • Incomplete or borderline priming in prepubertal children • Differences in agent potency (arginine alone tends to produce lower peaks than insulin or clonidine) • Technical factors: IV infiltration, incorrect dose calculation, delayed or mishandled sample processing
Because of this, essentially all major pediatric endocrine guidelines require a subnormal peak GH on two separate tests using two different, mechanistically distinct stimulating agents (e.g., insulin plus arginine, or clonidine plus glucagon) before GHD is diagnosed. A child who fails one test but passes a second is not labeled GH deficient — the diagnosis requires reproducible, cross-validated evidence.
A peak GH below threshold on only one of two required stimulation tests is not diagnostic — international consensus guidelines (GRS/ESPE, 2016) explicitly require concordant subnormal peaks on two independent provocative agents before a child is labeled GH deficient and started on lifelong recombinant GH therapy.
A confirmed subnormal peak GH on two agents is necessary but still not the full picture. The final diagnostic step integrates biochemistry with structural pituitary imaging to determine whether the deficiency is isolated to GH or part of a broader combined pituitary hormone deficiency (CPHD), and to rule out a structural lesion — most importantly a craniopharyngioma — before committing a child to years of daily rhGH injections.
IGF-1 and IGFBP-3 are re-checked and interpreted alongside the confirmed subnormal peak GH result. Because these binding proteins reflect integrated, 24-hour GH exposure rather than a single provocative burst, a child with two subnormal peak GH tests plus low age-adjusted IGF-1/IGFBP-3 has strongly concordant evidence of true GHD.
Discordant results (subnormal stimulation tests but normal IGF-1) should prompt reconsideration — repeat testing, review of assay cutoffs, or reassessment for constitutional delay of growth and puberty, which can mimic biochemical GHD in a subset of prepubertal children.
Every child with confirmed GHD undergoes dedicated pituitary/hypothalamic MRI with contrast before rhGH is started, specifically to exclude or characterize structural causes:
• Craniopharyngioma: the most important lesion to exclude — a benign but locally aggressive tumor arising from Rathke pouch remnants, often presenting with growth failure, visual field defects, and headache; missing this diagnosis and treating growth failure as isolated GHD alone can delay tumor treatment • Pituitary stalk interruption syndrome (PSIS): a thin or absent pituitary stalk, ectopic posterior pituitary bright spot, and hypoplastic anterior pituitary — a congenital cause of GHD, sometimes associated with breech delivery or birth trauma history • Empty sella, septo-optic dysplasia, and other midline defects • Normal MRI: seen in the majority of isolated, idiopathic GHD cases, and does not exclude the diagnosis
A structural pituitary or hypothalamic abnormality is found on MRI in roughly 15–20% of children with biochemically confirmed GHD — high enough that pituitary imaging is considered mandatory, not optional, before a child is committed to a presumptive diagnosis of idiopathic isolated GHD.
Because the anterior pituitary produces several hormones from anatomically adjacent cell populations, a lesion or developmental defect severe enough to impair GH secretion often affects neighboring axes too. Before starting therapy, the full anterior pituitary panel is screened:
• TSH/free T4 — central hypothyroidism • ACTH/cortisol (often via the same ITT used for the GH test, since insulin-induced hypoglycemia also stimulates cortisol) • LH/FSH and sex steroids — gonadotropin deficiency, relevant at the age of expected puberty • Prolactin — elevated with stalk compression, low/absent with combined pituitary hormone deficiency
A child with GHD plus one or more additional deficient axes is diagnosed with combined pituitary hormone deficiency (CPHD) rather than isolated GHD, which changes both prognosis and the need for concurrent hormone replacement (e.g., hydrocortisone, levothyroxine) alongside rhGH.
Once isolated GHD or CPHD is confirmed, recombinant human GH (rhGH) is initiated as daily subcutaneous injections, typically starting at 0.16–0.24 mg/kg/week divided into 6–7 daily doses, titrated using growth velocity and IGF-1 levels every 3–6 months, with the goal of normalizing growth velocity and maximizing adult height.