THE NATIONAL MEDICAL JOURNAL OF INDIA VOL. 5, NO.3

How to do it129

Assessment of anterior pituitary function using thecombined insulin stress test

D. BHATNAGAR

INTRODUCTIONThe insulin stress test came into use as a test of anteriorpituitary function in the late 1960s after it had been estab-lished that insulin-induced hypoglycaemia stimulated therelease of cortisol and growth hormone." Since then it hascome to be used along with TRH (thyrotrophin releasinghormone) and LHRH (luteinizing hormone releasinghormone) tests to assess anterior pituitary function.?

As a general rule endocrine deficient states arediagnosed by tests which stimulate the production of thedeficient hormone, thereby testing its reserve. Excessautonomous hormone production is documented by testswhere an attempt is made to suppress hormone secretion.

In states of deficiency, especially when it is borderline,basal hormone concentrations may be within normallimits. However, these might have been achieved throughsupra physiological stimulation at a point of control abovethat of the failing gland. Basal hormone concentrationsare also influenced markedly by diurnal variation andfactors such as stress, thereby decreasing the diagnosticvalue of single estimations. Dynamic tests offer severaladvantages over basal hormone estimations as they testthe function of a pathway for control. They also helpdifferentiate between primary and secondary causes offailure, particularly when used in conjunction with othertests.

INDICATIONSThe combined insulin stress test is essentially a measureof the integrity of the hypothalamic-pituitary axis (Fig. 1).Insulin hypoglycaemia tests the ACTH (adrenocortico-trophic hormone) reserve by stimulating the release of thehormone from the pituitary, while the TRH and LHRHtests (done at the same time or sequentially) assessthyrotrophin and gonadotrophin secretion respectively(Table I).

Insulin hypoglycaemia can also be helpful in establish-ing the diagnosis of Cushing's syndrome, especially inpatients with depression or chronic alcoholism, whereresults from the dexamethasone suppression test maybe equivocal. A delayed or exaggerated response toinsulin hypoglycaemia may also indicate hypothalamicdysfunction.

Manchester Royal Infirmary, Oxford Road, Manchester M13 9WL, UKD. BHATNAGAR Department of Medicine

© The National Medical Journal of India 1992

Hypoglycaemia

Limbic system

11)

LHRH

TRH

Anteriorpituitary

Posteriorpituitary

AjGrowth LH FSH TSH

\. hormone I~------------~~-------------Cortisol Prolactin

Hormones sampled during the combinedinsulin stress test

FIG. 1. The hypothalamic-pituitary axis (adapted withpermission, From Color Atlas of Physiology. EdsDespopoulos A and Silbernagl S. 1984. Georg ThiemeVerlag, Thieme-Stratton Inc, Stuttgart.)

130

TABLE I. Common indications

1. Pituitary tumours prior to surgery2. Post-pituitary surgery3. Post-pituitary radiation4. Assessment of pituitary reserve or hypothalamic damage

(suprasellar extension) in pituitary tumours causinga. Cushing's diseaseb. Acromegalyc. Prolactinomad. Craniopharyngiomae. Parasellar tumours

5. Hypothalamic tumours or infiltration

CONTRAINDICA TIONSThe test is contraindicated in patients with coronary heartdisease or epilepsy. It should be avoided in the elderly andin patients with cerebrovascular disease. In acutely illpatients and those at risk from hypoglycaemia, the deci-sion to do the test should be based on the likely benefit tothe patient from the diagnostic information obtained.Particuiar care should be exercised in patients withadrenocortical insufficiency who are extremely sensitiveto insulin hypoglycaemia. The combined insulin stress testis safe if the relative and absolute contraindicationsareobserved, the dose of insulin is correctly calculated andmeasured, and the patient is observed closely lest hedevelops prolonged hypoglycaemia.

METHODFasting patients are given intravenous insulin along withTRH and LHRH. Blood samples are taken, at 30 minuteintervals over a period of 2 hours and tested for glucose,cortisol, prolactin, growth hormone, thyroxine, thyroidstimulating hormone (TSH) and follicle stimulatinghormone (FSH) together with 17 ~-oestradiol in femalesand testosterone in males. The test can be performed onout-patients most of whom can go home the same dayfollowing an afternoon meal.

Patients should be given clear instructions to fast(unflavoured water is allowed) and asked not to take anymedication for 12 hours prior to the test. Drugs such ashydrocortisone cross-react in the assay for serum cortisoland will make interpretation of the results impossible,while ~-blockers may mask the symptoms of hypo-glycaemia. The following should be noted: dose and timeof consumption of any thyroxine tablets or oral contracep-tives and the use of depot testosterone injections andhormone implants or skin patches. The nature of the testshould be described and the transient nature of thesymptoms of hypoglycaemia explained so that the patientis not alarmed by them.

Items needed for the test should be laid out on a nursingtrolley (Table II) and checked before proceeding. Patientsshould be weighed and the general physical examinationincluding the pulse and blood pressure recorded. There-after, 10 ml of normal saline and 5 ml of sodium heparin100 units/ml (for flushing the indwelling cannula) aredrawn up in separate syringes and labelled for identifica-tion with an indelible marker pen.

THE NATIONAL MEDICAL JOURNAL OF INDIA VOL. 5, NO.3

TABLE II. Items needed on the nursing trolley

Drugs50% dextrose 15mlx Z Actrapid insulinnormal saline IOmlx2 TRH 200"gheparin sodium l00ulml LHRH lOO'"gHydrocortisone 200 mg for intravenous injectionOral glucose solution 500 mlOther Items50 ml syringe xt5 ml syringe x 41 ml syringe x 1 (not insulin syringe)Venflon cannulae 18G Needles size 18 GAdhesive tape GauzePaper tissue or cotton wool Sharps binThree way tap or rubber cap for intermittent injectionFluoride oxalate bottles x 8Blood specimen bottles without preservative x 10

10 ml syringe x 152 ml syringe x 15

A size 18 G (coded green) intravenous cannula orbutterfly needle (the former is preferable) is inserted intoa large vein in the antecubital fossa and anchored withtape. A 3-way tap or a rubber cap for intermittent injec-tion is fitted to the cannula and 2 to 5 ml of normal salineis injected to check proper positioning. Good venousaccess is required in case of severe hypoglycaemia. Aftereach sampling, the cannula should be flushed with normalsaline and enough heparin instilled to fill the dead space.The following medication is then drawn up in to separatesyringes and labelled clearly: (a) 100 IJ.gLHRH (b) 200 IJ.gLHRH (c) 100 mg hydrocortisone and (d) 20 ml 50%glucose.

The dose of insulin to be given is then calculated. Thisvaries from 0.1 to 0.3 units/kg body weight (usually it is0.15 units/kg). Patients known to have hypopitutarism orsuspected to have a low cortisol reserve should be given alower dose of insulin. Patients with diabetes mellitus,obesity, acromegaly and Cushing's syndrome will need ahigher dose-O.3 units/kg body weight. The insulin isdrawn up in a conventional 1 ml syringe rather than aninsulin syringe, as the latter is not designed to fit on todevices for intravenous injections.

Sample bottles for glucose and hormone estimationsshould be labelled on the top and side with a marker penfor times ranging from 0 to 120 minutes (at 30 minuteintervals) and placed in order in a test tube rack. Thewhole process of labelling and drawing up of drugs usuallytakes about 10 minutes.

Before sampling a 2 ml syringe is used to withdrawabout 0.5 ml of blood and heparin, which is then dis-carded. A basal sample is then taken. The dose of alldrugs, especially insulin, should be checked beforeintravenous injection. After injection the cannula shouldbe flushed with normal saline to ensure the required dosesare injected and do not remain in the cannula. Somepatients notice flushing and a peculiar taste in the mouthafter being given TRH and LHRH. Patients who have hadpituitary surgery may experience a transient headache.

Hypoglycaemic symptoms normally appear between 20and 40 minutes, but can occur earlier in a patient given toomuch insulin. Frequent monitoring of the pulse will reveal

BHATNAGAR : COMBINED INSULIN STRESS TESt

tachycardia which occurs before the onset of symptoms.In some patients pallor and sweating are not marked so itis important to keep track of the patient's sensorium toassess the degree of hypoglycaemia. With an ideal dosesymptoms should start to improve within a few minutes.There is no advantage in allowing hypoglycaemia to con-tinue beyond 2 minutes because reversing it has no effecton subsequent sampling and results. A glass of water anda fan switched on make the patient who is recovering fromhypoglycaemia more comfortable. Most patients tendto fall off to sleep, but constant awareness should bemaintained in case the hypoglycaemia recurs. The lowestlevel of blood sugar reached together with the symptomsexperienced should be recorded in the case notes. It isimportant to ensure that the patient eats a good mealbefore going home.

If the hypo glycaemia is prolonged, it should be treatedwith oral and intravenous glucose. It is not uncommon tosee hypoglycaemia recur 5 to 10 minutes later even afteran intravenous injection of dextrose. An additionaldrink of glucose should therefore always be given.Hypoglycaemia must not be allowed to proceed to neuro-glycopaenia, which often appears suddenly. Some patientsmay have tonic and clonic fits and in some transientischaemic attacks may occur. These are rare and respondto intravenous glucose. Intravenous diazepam can also beused if necessary.

If at the end of 40 minutes hypoglycaemia has notoccurred two-thirds of the original dose of insulin shouldbe given again. The hypo glycaemia that follows is usuallymore severe than that seen with a single dose andadequate precautions should be taken so that it does notdo lasting harm. A good history and examination shouldalso reveal a cause for the apparent resistance to insulin(Tables III and IV).

INTERPRETATION OF RESULTSLaboratory considerationsIt is usual for most laboratories to establish referenceranges on 50 to 100 'healthy' individuals for basal hormonelevels. It is, however, not routine to determine thehormonal response to insulin hypoglycaemia. Throughthe years the cortisol and growth hormone responses toinsulin hypoglycaemia have been characterised andaccepted as conventional. Nonetheless, it is worth enquir-ing of the laboratory as to the methods used, particularlyfor cortisol. Newer cortisol assays are more precise andless prone to interference compared to the older fluori-metric methods. The same caveat applies to assays for

TABLE III. Precautions

1. Check patient is fasting2. Check patient has not taken hydrocortisone3. Document all medication being taken4. Have ready access to intravenous and oral glucose5.. Recheck dose of insulin calculated and drawn up in the syringe6. Remember to flush the insulin, TRH and LHRH with normal saline7. Observe the patient throughout the test .8. Ensure patient eats a good meal before going home

131

TABLE IV. Common causes for failure of adequate hypoglycaemiaor unusual results

1. Concurrent medication2. Obesity3. Insulin resistance4. Inadequate insulin due to:

a. wrong calculationb. failure to tlush the cannula after giving insulinc. loose connection between syringe or 3-way tap and venous cannula

TSH, lutenizing hormone (LH) and FSH, all of whichhave a common alpha subunit. If highly sensitive TSHassays are being used, these will alter, to a certain extent,reference values for basal TSH levels although not for theresponse to TRH.

Diagnostic considerationsAdequacy of hypoglycaemia: The plasma glucose value

will generally indicate if adequate biochemical hypo-glycaemia has been achieved «45 mg/dl), although thisis best judged clinically.

Hormonal responses: Each hormone result should beassessed for basal levels, time and extent of peak values,and time of return to basal values.

ProlactinIn a combined insulin stress test the rise in prolactinconcentrations is a marker of the adequacy of the stressresponse. An absent response together with a diminishedresponse of other hormones suggests hypopituitarism. Anelevated basal prolactin value (>500 mUlL in malesand >400 mUlL in females) followed by an exaggeratedrise is suggestive of a prolactinoma or pressure on thehypothalamic-hypophyseal tract from the suprasellarextension of a pituitary tumour.

CortisolA low basal value (<200 nmol/L) with an inadequate rise(a normal rise is about 500 nmol/L at its peak and anincrement of 200 nmol/L over the basal value) is sugges-tive of failure of the hypothalamic-anterior pituitary-renal axis. Sometimes an elevated basal cortisol level isfollowed by a progressive decline. This can result if thepatient is stressed initially or has Cushing's syndrome.In some patients with long standing hypopituitarisminsulin stress may produce cortisol increments that areinterpreted to be normal, although the patient may still beat risk from a hypoadrenal crisis.

Growth hormoneGrowth hormone deficiency is suggested by an absence ofa rise in hormone levels. There is no consensus on thelevel of rise in growth hormone, but in normal subjectsone post-insulin sample should have a concentration ofmore than 10 mUlL. About 10% of normal individualsshow no growth hormone response to an insulin stresstest. Other causes of a blunted growth hormone responseinclude diabetes mellitus, obesity, Cushing's syndrome orsteroid therapy. Post-insulin responses may be blunted or

132

exaggerated in acromegaly, but the diagnosis is best madeclinically and on the basis of the suppression of growthhormone during a glucose tolerance test.

Thyroid stimulating hormoneBasal TSH and thyroxine levels are not reliable indicatorsof secondary hypothyroidism. On stimulation with TRHthe normal TSH response would be an approximate four-fold rise at 20 minutes, followed by near basal levels at60 minutes. In pituitary hypothyroidism basal levelsmay be normal, but the response is impaired at 20 and60 minutes. A blunted response can also been seen inCushing's syndrome, acromegaly, depression or systemicillness. A delayed response, or a prolonged rise in TSH at60 minutes suggests hypothalamic disease, although thisis not characteristic.

Lutenizing hormone and follicle stimulating hormoneElevated basal LH and FSH levels with low concentra-tions of 17 ~~oestradiol or testosterone indicate primaryfailure. Low basal concentrations of LH and FSHgenerally indicate hypothalamic or pituitary failure. Inmales, a low testosterone level with an elevated basal LHconcentration, in the presence of a normal FSH level,is suggestive of Leydig cell failure. If FSH levels areelevated this suggests failure of spermatogenesis.

After administration of LHRH the normal responsevaries widely in both males and females. FSH concentra-tions tend not to rise as much as LH levels. In femalesthe response may vary with the phase of the cycle. Largerises in LH and FSH at 20 and 60 minutes with low basalconcentrations would be suggestive of an LHRH deficiency.An exaggerated response with a high basal level of LHand FSH indicate primary gonadal failure. A poorresponse to LHRH indicates pituitary or hypothalamicdysfunction. However, a normal LHRH test does notexclude hypothalamic disease.

OTHER TESTSA number of other tests are also available to assess ACTHand growth hormone secretion. A comparison of thesetests with the insulin stress test however is not straight-forward. The predictive value of any test is dependent onthe prevalence of the disease being investigated so that fora disease which has a low prevalence, the sensitivity andthe specificity will be poor. The prevalence of disorders ofthe anterior pituitary is small and there have been fewdirect comparisons between the insulin stress test andother tests. Nevertheless, the insulin stress test has cometo be accepted as the 'gold standard' for assessing cortisoland growth hormone reserve.

The glucagon-propranolol test is perhaos the only testcurrently available for the combined assessment ofACTH and growth hormone reserve.! The hormoneresponses to glucagon have not been definitively charac-terised. It is now recommended that glucagon be givenintramuscularly (rather than subcutaneously) as thisproduces a higher and more consistent cortisol response.The cortisol response is generally lower and delayed afterglucagon compared to that after insulin. There is some

THE NATIONAL MEDICAL JOURNAL OF INDIA VOL. 5, NO.3

evidence to suggest that the cortisol response in healthysubjects and patients with established pituitary diseasemay differ." It remains a useful alternative in patientswhere the insulin stress test is contraindicated.

It has recently been suggested that continuous intra-venous insulin infusion may be a more reliable andcontrolled way of producing hypoglycaemia, and that itproduces hormone responses equivalent to those seen inthe conventional stress test. 5 This variation of the test isnot commonly used.

With the availability of pituitary hormone releasingfactors and hormones attempts have been made to infusethese in combination.s? However, their cost andrestricted availability have meant that they are unlikelyto be in common use for some time. Moreover, theresponses to corticotrophin releasing hormone andgrowth hormone releasing hormone have not proved tobe as useful in diagnosis as expected. At present their useis perhaps best restricted to measurement of samples fromvenous sinuses draining the pituitary.

ACTH reserve may also be tested by metyrapone. Thiscompetitive inhibitor of 11 ~,-hydroxylase blocks theconversion of Ll-deoxycortisol to cortisol thus stimulatingthe release of ACTH. The test requires the collection of24 hour urine specimens although an overnight version ofthe test can be performed if a direct assay for l l-deoxy-cortisol is available. It is possibly a more dangerous testbecause it can precipitate an adrenal crisis. Besides it onlyselectively tests a part of the hypothalamic-pituitary-adrenal axis. The use of ACTH to assess pituitaryinsufficiency is seemingly illogical, but proponents of itsuse in this manner base this on empirical observations.The release of ACTH can also be potentiated by argininevasopressin when it is used with other releasing factors.fTransient flushing and nausea seem to be the main side-effects but again, like the insulin stress test, it does not testthe cerebra-hypothalamic-pituitary axis. It may be usefulin assessing reserve in patients with existing pituitarydisease.

Growth hormone response can also be tested byexercise or ingesting L-dopa or infusing arginine. Themechanisms by which responses are elicited in each caseare not clear. Although they are likely to involve releaseof somatomedin and growth hormone releasing hormonea final diagnosis is better made by the insulin stress test.

CONCLUSIONThe insulin stress test (carried out with or without theTRH or LHRH test) can be of diagnostic value in anteriorpituitary dysfunction when the results of basal hormoneestimations are equivocal. It is also used to assess 'reserve'in patients with pituitary tumours and followinghypophysectomy. The test is not without hazards andshould be carried out only when the results are likely to beof clinical use. It is contraindicated in patients with coro-nary heart disease, profound adrenal insufficiency,untreated hypothyroidism and epilepsy. It is best carriedout, with appropriate precautions, in a standardizedmanner to avoid problems in the interpretation of results.Alternative dynamic tests of anterior pituitary function

BHATNAGAR : COMBINED INSULIN STRESS TEST

that avoid hypoglycaemiado exist, but the insulin stresstest remains the 'gold standard', particularly whenassessingACfH and growthhormone reserve. Even withthe availabilityof pituitary hormone releasing hormonesit is unlikely that the insulin stress test will become aredundant investigation. It is cheaper than other testoptions and, when carried out correctly, providespracticalinformationofdiagnosticand therapeutic value.

ACKNOWLEDGEMENTSI would like to thank my former colleaguesDr M. Labiband Dr G. P. Weavind for helpful discussions. I wouldalso like to thank Professor J. G. Ratcliffe and ProfessorD. C. Anderson at Hope Hospital, Salford, UK foradvice.

REFERENCESGreenwood FC, London J, Stamp TCB. The plasma sugar, free fattyacid, cortisol and growth hormone response to insulin. J Clin Invest1966;45:429-49.

133

2 Fish HR, Chernow B, O'Brian JT. Endocrine and neurophysiologicresponses of the pituitary to insulin-induced hypoglycaemia: Areview. Metabolism 1986;35:763-80.

3 Vanderschueren-Lodeweyckx M, Wolter R, Malvaux P, EggermontE, Eekels R. The glucagon stimulation test: Effect on plasma-growthhormone and on immunoreactive insulin, cortisol and glucose inchildren. J Pediatr 1974;85:182-7.

4 Littley MD, Gibson S, White A, Shalet SM. Comparison of theACTH and cortisol responses to provocative testing with glucagonand insulin hypoglycaemia in normal subjects. Clin Endocrino/1989;31:527-33.

5 Newman GH, MacDonald lA, Allison SP. Testing the anteriorpituitary: Hypoglycaemia produced by continuous intravenousinsulin infusion. Br Med J 1983;287:571-4.

6 Sheldon WR, DeBold CR, Evans WS, et al. Rapid sequentialintravenous administration of four hypothalamic releasing hormonesas a combined anterior pituitary function test in normal subjects.J Clin Endocrinol Metab 1985;60:623--30.

7 Hall R, Fink P, Hetzel WD. Combined pituitary stimulation test withreleasing hormones. Acta Endocrino/1985;108 (supp1267):20.

8 Sandler LM, Burrin JM, Joplin GF, Bloom SR. Combined use ofvasopressin and synthetic hypothalamic releasing factors as a newtest of anterior pituitary function. Br M ed J 1986;292:511-14.

ObituariesMany doctors in India practise medicine in difficult areas under trying cir-cumstances and resist the attractions of better prospects in western countriesand in the Middle East. They die without their contributions to our countrybeing acknowledged. .

The National Medical Journal of India wishes to recognize the efforts ofthese doctors in a new section 'Obituaries'. We invite short accounts of thelife and work of a recently deceased colleague by a friend, student or relative.The account in about 500 to 1000 words should describe his education andtraining and highlight the achievements as well as the disappointments. Aphotograph should accompany this article.

-Editor