WHAT is Diabetes Mellitus?
Diabetes mellitus is a metabolic disorderis characterized by hyperglycemia that results from defective insulin production, secretion, or utilization. The classification of diabetes includes:
In type 1 DM, the pancreas produces little or no endogenous insulin.
In type 2 DM, disease results from a defect in insulin manufacture and release from the beta cells and from insulin resistance in the peripheral tissues. It has a strong genetic component and is commonly associated with obesity. Onset is usually in adulthood; however, cases are increasingly occurring in teenagers and older children.
Glycosuria
Urine testing is a valuable pointer to diabetes mellitus (DM) but is insufficient to establish the diagnosis. Although modern glucose oxidase test strips are free from interference by other reducing substances, glycosuria does not always indicate DM; the converse also pertains. Thus, glycosuria has low sensitivity and specificity.
Causes of glycosuria
2 DM.
2 Impaired glucose tolerance (IGT).
2 Lowered renal threshold for glucose (esp. pregnancy, children).
2 Casual (random)—usually the first line investigation.
2 Fasting plasma glucose (FPG)—an alternative to a casual reading.
2 75g oral glucose challenge—if necessary.
Neither the confirmation nor exclusion of DM should rest on measurement of longer term indicators of glycaemia such as glycated haemoglobin or fructosamine. Although of high specificity, these tests are not sufficiently standardised nor do they have sufficient sensitivity. False negative results are particularly likely with less marked degrees of hyperglycaemia, especially in subjects with IGT or IFG.
Blood glucose
A blood glucose measurement is the essential investigation in the diagnosis of DM. A glucose-specific assay is required. An appropriate sample must be collected, usually venous plasma (in a tube containing fluoride
oxalate as an inhibitor of glycolysis) and the sample tested without undue delay in an accredited laboratory.
Reagent test strips
Although convenient and readily available, reagent test strips for monitoring of capillary glucose (even when used in conjunction with a calibrated reflectance meter) are unsuitable for diagnosing DM; a confirmatory laboratory measurement must therefore always be performed. In the absence of typical symptoms the diagnosis should be confirmed by a repeat measurement on a separate day; this may be either a
casual or FPG sample.
Confirmation of the diagnosis is especially important in asymptomatic individuals.
An oral glucose tolerance test (OGTT) is rarely required to confirm the diagnosis and should not be regarded as a first line investigation. The OGTT is time consuming, requires trained staff and is less reproducible than FPG.
The diagnostic FPG is lower than the previous National Diabetes Data Group (1979) and WHO (1980, 1985) criteria which specified a diagnostic fasting plasma glucose >7.8mmol/L.
The 1997 criteria introduced the new intermediate category of impaired fasting glucose (IFG) defined as:
2 FPG 6.1–6.9 mmol/L.
False +ve diagnoses may arise if the subject has prepared inadequately. This possibility is more likely following the reduction in the diagnostic threshold for diabetes based on FPG in the 1997 revised criteria.
Impaired glucose tolerance
The diagnosis of IGT can only be made using a 75g oral glucose tolerance test; a random blood glucose measurement will often point to the diagnosis when other results are non-diagnostic.
This category denotes a stage intermediate between normal glucose levels and DM (OHCM, section 9). By definition, plasma glucose levels are not raised to DM levels so typical osmotic symptoms are absent. Although subjects with IGT are not at direct risk of developing chronic microvascular tissue complications, the incidence of macrovascular complications (i.e. CHD, CVD, PAD) is increased. Presentation with one of these conditions should therefore alert the clinician to the possibility of undiagnosed IGT (or type 2 DM). Note that a proportion of individuals who are diagnosed by an OGTT may revert to normal on re-testing.
Impaired fasting glucose If an OGTT is performed, the 120min glucose concentration must be <7.8mmol/L. This category is also usually asymptomatic. To date, crosssectional studies suggest that IGT and IFG may not be synonymous in terms of pathophysiology and long-term implications.
Oral glucose tolerance test
The OGTT (see table below) continues to be regarded as the most robust means for establishing the diagnosis of diabetes in equivocal cases. The WHO suggests that only when an OGTT cannot be performed should the diagnosis rely on FPG. OGTTs should be carried out under controlled conditions after an overnight fast.
The interpretation of the 75g glucose tolerance test. These results apply to venous plasma. Marked carbohydrate depletion can impair glucose tolerance; the subject should have received adequate nutrition in the days preceding the test.
Effect of intercurrent illness on glycaemia
Patients under the physical stress associated with surgery, trauma, acute MI, acute pulmonary oedema or stroke may have transient 4 of plasma glucose—often settles rapidly without antidiabetic therapy. However, the hormonal stress response in such clinical situations is liable to unmask pre-existing DM or to precipitate DM in predisposed individuals. Blood glucose should be carefully monitored and urine tested for ketones.
Sustained hyperglycaemia, particularly with ketonuria, demands vigorous treatment with insulin in an acutely ill patient.
Oral glucose tolerance test
Anhydrous glucose is dissolved in 250ml water; flavouring with sugar-free lemon and chilling increase palatibility and may reduce nausea. The subjecrt sits quietly throughout the test.
Blood glucose is sampled before (time 0) and at 120min after ingestion of the drink, which should be completed within 5min.
Urinalysis may also be performed every 30min although is only of interest if a significant alteration in renal threshold for glucose is suspected.
Acute myocardial infarction (OHCM section 5)
Hyperglycaemia at presentation is associated with a higher mortality—even in subjects with previously undiagnosed DM; tight metabolic control using an intravenous insulin-dextrose infusion (followed by subcutaneous insulin) significantly reduced mortality in a recent multicentre Swedish study.
Stroke (OHCM section 10)
Hyperglycaemia on admission may be associated with a poorer outcome; however, there is no clinical trial evidence to date that intensive control of hyperglycaemia improves prognosis.
Re-testing is usually indicated following resolution of the acute illness—an OGTT at a 4–6-week interval is recommended if glucose levels are equivocal.
Potential diagnostic difficulties of DM
Type 1 DM
This is diagnosed principally on clinical and biochemical features (OHCM section 12). The presence of serum islet cell antibodies (ICA, in ~30–60% of patients) at diagnosis may confirm the diagnosis.
The proportion of patients testing positive for ICA 5 with increasing duration of type 1 DM. If there is doubt, treat with insulin if indicated on clinical and biochemical criteria; the need for insulin can be considered
at a later date. However, discontinuation of insulin can be disastrous in patients with type 1 DM. The decision to stop insulin should be made only by an experienced diabetologist. iA –ve test for ICA
does not necessarily exclude type 1 DM.
Other humoral markers of autoimmunity, e.g. anti-GAD65 antibodies, anti-insulin antibodies, are generally only available in research laboratories.
Stiff man syndrome: rare condition presenting as a progressive spastic paraparesis with polyglandular endocrine involvement (p246). Anti- GAD65 antibodies are present; approximately 30% of patients develop insulin-requiring DM.
MODY: a small percentage of young patients with relatively minor hyperglycaemia and no ketonuria, will prove to have relatively uncommon inherited forms of DM, e.g. MODY. Such patients often receive insulin therapy from diagnosis, the assumption being that they have type 1 DM. Prerequisites for the diagnosis include:
– A family history with an autosomal dominant inheritance.
– Diagnosis under the age of 25 years.
In some subtypes of MODY (glucokinase mutations; MODY 2), good glycaemic control may be maintained life-long without insulin or even oral antidiabetic agents. The exception is pregnancy; insulin may be
required temporarily in order to ensure optimal control—oral antidiabetic agents should be avoided. The diagnosis of MODY may be confirmed by molecular genetic testing although presently this is not widely available. Appropriate counselling is required. Seek expert advice through your local hospital diabetes unit.
2 Early-onset type 2 diabetes—In recent years there has been a dramatic increase in the incidence of type 2 diabetes in younger patients (children and adolescents) from non-white ethnic minorities. This may present diagnostic difficulties but some pointers suggest the diagnosis:
– Obesity is usually a prominent feature.
– Serum autoantibody tests for type 1 DM are negative.
– A skin marker of insulin resistance (acanthosis nigricans) may be present.
If in doubt, it is usually safer to treat younger patients with insulin; this is especially true if ketosis is present.
Monitoring diabetic control
Self-testing by diabetic patients
Self-testing of urine and/or capillary blood glucose testing can readily be performed by the majority of patients. Measurements of longer term gly- caemic control are laboratory-based or require specialised techniques generally suitable only for use in a hospital clinic.
Urine testing
Glycosuria: Semiquantitative testing for glucose using reagentimpregnated test strips by patients is of limited value. Urinalysis provides retrospective information over a limited period of time. Other limitations:
The renal threshold for the reabsorption of glucose in the PCT is ~10mmol/L on average but varies between individuals. Subjects with a low threshold will tend to show glycosuria more readily, even with normal glucose tolerance (‘renal glycosuria’). Children are particularly liable to test positive for glucose. The renal threshold is effectively lowered in pregnancy. Conversely, a high threshold, common among the elderly, may give a misleadingly reassuring impression of satisfactory control. Fluid intake and urine concentration may affect glycosuria.
Renal impairment may elevate the threshold for glucose reabsorption. Delayed bladder emptying, e.g. due to diabetic autonomic neuropathy (OHCM section 9), will reduce the accuracy of the measurements
through dilution.
Hypoglycaemia cannot be detected by urinalysis.
Ketonuria: Semiquantitative test strips for acetocetate (e.g. Ketostix®) are available for patients with type 1 DM. Useful when intercurrent illness leads to disturbance of metabolic control. The presence of ketonuria on dipstick testing in association with hyperglycaemia indicates marked insulin deficiency (absolute, or more commonly, relative). Increased doses of insulin are indicated in such circumstances to avert more severe metabolic decompensation (DKA, see below and OHCM section 21). Occasionally, patients with type 2 DM develop ketosis during severe intercurrent illness, e.g. major sepsis. Neither Ketostix® nor Acetest® tablets detect 3-hydroxybutyrate (although acetone is detected by Acetest®). Occasional underestimation of the degree of ketonaemia using these tests is a well-recognized, albeit uncommon caveat of alcoholic ketoacidosis (OHCM section 9).
Self-testing of blood glucose
Self-testing of capillary blood glucose obtained by fingerprick has become an established method for monitoring glycaemic control. Frequent testing is a prerequisite for safe intensive insulin therapy such as that employed in the DCCT. Enzyme-impregnated dry strip methods are available which can be used in conjunction with meter devices to improve accuracy. Most are based on the glucose oxidase reaction:
Glucose + O2————7Gluconic acid + H2O2
Glucose oxidase
The hydrogen peroxide generated by the reaction reacts with a reduced dye in the test strip producing an oxidised colour proportional to the amount of H2O2 formed. This reflects the amount of glucose that was oxi- dised. In most strips, blood cells are excluded by a layer within the strip.
Thus, the glucose concentration in capillary plasma is measured.
Adequate training and a system of quality control are important; even when trained health professionals use such systems in clinics or hospitals misleading results are possible, particularly in the lower range of blood
glucose results. Where there is doubt, an appropriate sample (in a tube containing the glycolysis inhibitor fluoride oxalate) should be collected immediately for analysis by the clinical chemistry laboratory. However,
acute treatment of hypoglycaemia, where indicated, should not be delayed.
Laboratory assessment of glycaemic control Glycated haemoglobin HbA1c (comprises 60–80% total glycated haemoglobin, HbA1) is formed by the slow, irreversible, post-translational non-enzymatic glycation of the N-terminal valine residue of the chain of haemoglobin. In retina and renal glomerulus this process is implicated in the pathogenesis of the longterm complications of diabetes (OHCM section 9). The proportion of HbA1c : total haemoglobin (normal non-diabetic reference range approximately 4–6%) provides a useful index of average glycaemia over the preceding 6–8 weeks. The result is disproportionately affected by the blood glucose levels during the final month before the test (~50% of value).
Average HbA1c levels collected over a longer period (i.e. years) provide an estimate of the risk of microvascular complications. Sustained high concentrations identify patients in whom efforts should be made to improve long-term glycaemic control.
In patients with type 1 DM, a landmark randomised, controlled clinical trial (the DCCT) confirmed a causal link between degree of metabolic control and the development and progression of microvascular complications of diabetes (especially retinopathy) and neuropathy. Consensus panels in the USA and Europe have suggested targets for HbA1c of approximately 7–8% for most patients (if circumstances and frequency of hypoglycaemia allow). By contrast, tight glycaemic control may be contraindicated
by advanced complications, e.g. clinical nephropathy with renal impairment. It is recommended that HbA1c should be measured every 6 months in younger patients with type 1 DM. Pregnancy requires
monthly monitoring of HbA1c concentrations (although other measures may be preferable in pregnancy—see below: fructosamine). Blood can be collected by venesection ahead of the clinic visit (in primary care, by the hospital phlebotomy service or even by a nurse in the community if necessary). Alternatives include rapid assays for use in the clinic, or self-collection in advance of a fingerprick sample (in a capillary tube or on filter
paper) which is mailed to the laboratory.
Limitations of HBA1c measurements
Although glycated haemoglobin levels are a reliable indicator of recent average glycaemic control they do not provide information about the daily pattern of blood glucose levels; this supplementary information required
for logical adjustment of insulin doses is derived from home blood glucose monitoring (see below). More recent changes in glycaemia (i.e. within the preceding 4 weeks or so) will influence HbA1c level more than glucose levels 12 or more weeks ago.
Spurious HbA1c levels may arise in states of
- Blood loss/haemolysis/reduced red cell survival (low HbA1c).
- Haemoglobinopathy (OHCM section 16).
- 4 levels of HbS.
- 4 levels of HbF (high HbA1c).
Uraemia due to advanced diabetic nephropathy is associated with anaemia and 5 RBC survival thereby falsely lowering HbA1c levels. Fructosamine: refers to protein-ketoamine products resulting from the
glycation of plasma proteins. The fructosamine assay measures glycated plasma proteins (mainly albumin) reflecting average glycaemia over the preceding 2–3 weeks. This is a shorter period than that assessed using
glycated haemoglobin measurements and may be particularly useful when rapid changes in control need to be assessed, e.g. during pregnancy. Levels can be misleading in hypoalbuminaemic states, e.g. nephrotic syndrome (OHCM section 8). Some fructosamine assays are subject to interference by hyperuricaemia or hyperlipidaemia.
Measurements of fructosamine are less expensive than glycated haemoglobin assays; this may be an important consideration for some laboratory services. The methodology is suitable for automation and
rapid results can be obtained for use within a clinic attendance obviating the requirement for a prior blood test.
Diabetic emergencies: diabetic ketoacidosis, hyperosmolar non-ketotic syndrome & lactic acidosis Diabetic ketoacidosis (DKA) should be considered in any unconscious or hyperventilating patient. The hyperosmolar non-ketotic (HONK) syndrome is characterised by marked hyperglycaemia and dehydration in
the absence of significant ketosis or acidosis. Lactic acidosis (LA) associated with metformin is uncommon. A rapid clinical examination and bedside blood and urine tests should allow the diagnosis to be made
(OHCM, section 12). Treatment (IV rehydration, insulin, electrolyte replacement) of these metabolic emergencies should be commenced without delay.
Confirm diagnosis by bedside measurement of 2 Capillary blood glucose (glucose-oxidase reagent test strip)>
- Urinary dipstick for glucose and ketones (e.g. Ketostix®). Note: nitroprusside tests detect acetoacetate, but not 3-hydroxybutyrate. This may be relevant in some circumstances, e.g. alcoholic ketoacidosis. Venous plasma may also be tested for ketones.
- Urine for nitrites and leucocytes (UTI).
- Venous blood for urgent laboratory measurement of Plasma glucose (fluoride-oxalate; itrue ‘euglycaemic’ DKA is rare).
- U&E (arterial K+ can be measured by some gas analysers). Plasma Na+ may be depressed as a consequence of hyperglycaemia or marked hyperlipidaemia.
- Plasma creatinine (imay be falsely elevated in some assays by DKA).
- Plasma lactate (if indicated—can also be measured by some gas analysers).
- Indicated if acidosis without heavy ketonuria is present. LA is a complication of tissue hypoxia (type A) and is a rare complication of metformin treatment in patients with type 2 DM (type B).
- Plasma osmolality in HONK—either by freezing point depression or calculated: 2 × [plasma Na+] + [plasma K+] + [plasma glucose] + [plasma urea].
- FBC (non-specific leucocytosis is common in DKA).
- Blood cultures (signs of infection, e.g. fever, may be absent in DKA).
- ABGs (corrected for hypothermia) for: – Arterial pH, bicarbonate, PCO2 and PO2 (if shock or hypotension).
- Repeat laboratory measurement of blood glucose, electrolytes, urea at 2, 4 and 6h and as indicated thereafter. Electrolyte disturbances, renal impairment or oliguria should prompt more frequent (1–2 hourly) measurements of plasma K+. Capillary blood glucose is monitored hourly at the bedside.
Avoidance of hypokalaemia and hypoglycaemia are most important during therapy.
Other investigations, as indicated
- CXR.
- Microbial culture of urine, sputum, etc.
- ECG (acute MI may precipitate metabolic decompensation; note that serum transaminases and CK may be non-specifically elevated in DKA).
- Sickle cell test.
- Venous plasma PO4 3– (if there is respiratory depression).
Performance of investigations should not delay initiation of treatment and transfer to a high-dependency or intensive care unit.
A severe metabolic acidosis in the absence of hyperglycaemia (or other obvious cause of acidosis such as renal failure) raises the possibility of Lactic acidosis, Endocrinology & metabolism, Alcoholic ketoacidosis—this occurs in alcoholics following a binge.
Alterations in hepatic redox state may result in a misleading negative or ‘trace’ Ketostix® reaction. A similar caveat may occasionally be encountered when significant LA coexists with DKA. Venous plasma glucose may be normal or 4.
Anion gap (p432) >15mmol/L. Normally, the anion gap (<10mmol/L) results from plasma proteins, SO4
2–, PO4 3– and lactate ions. When the anion gap is increased, measurement of plasma ketones, lactate, etc. usually confirms the aetiology.
Investigation of hyperlipidaemia
Primary dyslipidaemias are relatively common and contribute to an individual’s risk of developing atheroma (e.g. CHD, CVD). Prominent examples include familial combined hyperlipidaemia (FCHL, ~2–3% of UK
population) and heterozygous familial hypercholesterolaemia (FH, UK incidence 1 in 500). Major hypertriglyceridaemia also predisposes to pancreatitis.
The key features of familial FH, FCHL and diabetic dyslipidaemia are considered later.
Investigations
Although many subtle alterations in plasma lipids have been described, therapeutic decisions rest on measurement of some or all of the following in serum or plasma (plasma being preferred since it can be cooled rapidly):
- Total cholesterol (may be measured in non-fasting state in first instance since levels are not greatly influenced by meals).
- Triglycerides (after 12h fast).
- Low-density lipoprotein (LDL)-cholesterol (calculated using the Friedewald formula when triglycerides are <4.5mmol/L):
HDL-cholesterol (regarded as the ‘cardioprotective’ subfraction—HDL particles are synthesised in the gut and liver and thought to be involved in ‘reverse transport’ of cholesterol from peripheral tissues to the liver where it can be excreted as bile salts.
Notes on sampling in relation to lipoprotein metabolism
- Triglycerides (triacylglycerols) are measured after a ~12h overnight fast in order to clear diet-derived chylomicrons.
- Alcohol should be avoided the evening prior to measurement of triglycerides (can exacerbate hypertriglyceridaemia).
- A weight-maintaining diet is recommended for 2–3 weeks before testing.
- Lipid measurements should be deferred for 2–3 weeks after minor illness and 2–3 months after major illness, surgery or trauma since cholesterol may be temporarily 5 and triglycerides 4. Following acute myocardial infarction it is generally accepted that plasma cholesterol is reliable if measured within 24h of the onset of symptoms.
- The effects of certain drugs on lipids should be considered.
- Glycaemic control should be optimised wherever possible before measuring plasma lipids in patients with diabetes.
Important additional considerations are
- Day-to-day variability—generally, decisions to treat hyperlipidaemia should be based on more than one measurement over a period of 1–2 weeks.
- Exclusion of secondary hyperlipidaemia—many common conditions, drugs and dietary factors can influence plasma lipids.
- Family members should also have their plasma lipids measured if a familial hyperlipidaemia is suspected in a proband.
Both cholesterol and triglycerides may be affected to some degree by these factors, but one often predominates. Pre-existing primary hyperlipidaemias may be exacerbated.
Clinical features
E.g. xanthelasma, tendon xanthomas, etc. should always be sought. A detailed family history, drug history and medical history (for diabetes and other cardiovascular risk factors such as hypertension) should always be obtained. Certain endocrine disorders, impaired hepatic or renal function can influence circulating lipid composition and cardiovascular risk. A classification of the major familial dyslipidaemias. Specialist advice should be sought in the management of major or resistant hyperlipidaemias.
Test protocols
Insulin tolerance test (insulin stress test)
Indication: suspected ACTH or GH deficiency.
Contraindications: patients with epilepsy, coronary heart disease (check
ECG).
Children: use no more than 0.1U/kg. Considerable care should be exercised; the test should only be performed in a centre with expertise.
Alternatives: short synacthen test for hypocortisolism; stimulation tests for growth hormone deficiency
Preparation: patient fasting overnight. Bed required (though day case procedure). Patient must be accompanied home and may not drive. OMIT morning hydrocortisone or other steroid hormone replacement if patient is taking this and previous day’s growth hormone. Physician must be present throughout the test. Requires written consent.
Procedure: early morning outpatient test in fasting patient. Indwelling venous cannula and constant medical supervision required throughout.
Cannula is kept patent by running saline infusion with three-way tap for sampling. Discard initial 2–3mL when each sample is taken. Label all samples clearly with time and patient details. Near-patient testing
glucometer required.
1. Take baseline blood for glucose, cortisol and GH. Check IV access working well. Review test with the patient and explain symptoms he/she is likely to experience.
2. Draw up 25mL of 50% dextrose for immediate administration IF REQUIRED.
3. Give soluble (regular) insulin as an intravenous bolus in a dose of 0.15U/kg after an overnight fast. Consider 0.1U/kg (lower dose) if suspected profound hypocortisolism. iThis appears a very small dose,
e.g. typically around 10 units. CHECK DOSE CALCULATION CAREFULLY.
Usually an insulin syringe is used to draw it up and then transfer it to a 2mL syringe containing saline.
4. Take blood at 15min intervals (0, 15, 30, 45, 60min) for glucose, cortisol and GH.
5. Observe for symptoms and signs of hypoglycaemia. First sign is usually profuse sweating. Patient may then be aware of symptoms such as palpitations, hunger, paraesthesiae. This typically occurs 30–45min into
the test. Check near-patient glucose to confirm <3.5mmol/L. Continue to talk to and reassure patient. If patient becomes very drowsy or unrousable then given 25mL of 50% glucose. This does not invalidate
the test as the hypoglycaemic stimulus has already occurred. Continue blood sampling at standard times.
6. If patient has not experienced hypoglycaemia by 45min and nearpatient glucose is >4mmol/L, give a further intravenous bolus of 0.15U/kg or 0.3U/kg if patient known to be very insulin resistant (e.g.
acromegalic). Repeat sampling at 15min intervals for 60min after this second bolus.
7. At end of procedure (usually 60min), give IV 25mL dextrose if patient still has symptoms of hypoglycaemia. 8. Give patient a meal including complex carbohydrate (e.g. sandwiches or lunch) and observe for a minimum of 1h further before accompanied discharge.
Unwanted effects: severe hypoglycaemia with depressed level of consciousness or convulsion requires immediate termination of test with 25mL of 50% dextrose IV. Repeat if necessary and follow with 5 or
10% dextrose infusion. Continue to collect samples for hormone and glucose measurements.
Interpretation: test is only interpretable if adequate hypoglycaemia is achieved (<2.2mmol/L). Normal maximal cortisol response >550nmol/L. Normal GH response >20mU/L. Impaired responses (if
hypoglycaemic stimulus adequate) denote corticotrophin (assuming adrenal glands are normal) or GH deficiency or both. Peak GH response <10mU/L is sufficient to consider GH replacement; peak GH response <5mU/L is severe growth hormone deficiency.
Combined anterior pituitary function testing
Indication: assessment for anterior pituitary hypofunction.
Contraindications: previous reaction to stimulatory hormones.
Alternatives: insulin tolerance testing for GH and adrenal axis; metyrapone test for adrenal axis.
Preparation: test usually performed in morning for basal sampling.
Procedure: IV cannula inserted. Basal blood samples taken for cortisol, oestradiol (3) or testosterone (9), free T4 and IGF-1. Hypothalalmic hormones are given sequentially intravenously each as a bolus over
around 20s: LHRH 100μg, TRH 200μg and ACTH 250μg. Additionally GHRH (1μg/kg body weight) may be given. (Reduce doses in children.)
Samples are drawn at 0, 20, 30, 60 and 120min for LH, FSH, TSH cortisol and prolactin. If GHRH is given, samples are drawn at the same time points for GH.
Interpretation: normal values as follows:
TRH: Suspect secondary hypothyroidism if peak response (at 20min) <20mU/L (Note: low levels also seen in hyperthyroidism—ensure free T4 or total T4 not raised).
ACTH: Peak cortisol response >550nmol/L at 30 or 60min.
LHRH: Peak LH/FSH response 2–5 × basal value.
LH: Peak at 20min, FSH later.
GHRH: Normal GH peak response >15mU/L.
Water deprivation test
Indication: diagnosis of diabetes insipidus (DI) and to distinguish cranial and nephrogenic diabetes insipidus.
Contraindications: none if carefully supervised. For correct interpretation, thyroid and adrenal deficiency should be replaced first.
Interpretation in the presence of diabetes mellitus and uraemia can be difficult.
Alternatives: morning urine osmolality of >600mOsmol excludes significant degrees of DI. No other definitive test for diabetes insipidus.
Patient preparation: usually an outpatient procedure. Correct thyroid and adrenal insufficiency in advance. Renal function and blood glucose should have been checked in advance. Steroid and thyroid hormone replacement should be taken as normal on the day of the test. If the patient is on DDAVP, omit the dose on the evening before the test (or if not possible, halve this dose). Free fluids, but not too excess, up to 0730h on the day of the test. No alcohol on the night before the test or in the morning of the test. Light breakfast but no tea, coffee or smoking on the morning of the test. Empty bladder before attending for the test.
If urine volume is <3L/day (‘mild cases’), ask patient to have no fluids or food from 1800h on the evening before the test (‘prolonged water deprivation test’).
Requirements for test: accurate weighing scales. Supervision for the whole test (up to 8h). DDAVP for injection (2μg). Immediate access to serum electrolyte, plasma and urinary osmolality assays. Access to a
plasma AVP (ADH) assay desirable.
Procedure: 0730h
1. Weigh patient and calculate 97% of body weight.
2. Mark this target on the chart.
3. No food or fluid for next 8h.
4. Insert cannula for repeated blood sampling and flush. 0800h
5. Obtain plasma for Na+ and osmolality and urine for osmolality.
6. Then collect urine hourly for volume and osmolality and plasma every 2h for Na+ and osmolarity.
7. Weigh patient before and after passing water if unobserved.
8. If patient loses 3% body weight, order urgent plasma osmolality and Na+.
9. If plasma osmolality >300mOsmol (Na+ >140mmol/L) stop test, allow patient to drink and give DDAVP
10.If plasma osmolality <300mOsmol, patient may have been fluid overloaded before test and water deprivation can continue.
11.Stop test at 8h (4pm) and take final recordings of urine and plasma.
12.Save an aliquot of plasma for vaspressin levels in case of difficulties in test interpretation.
13.Ideally urine osmolalities will have reached a plateau (<30mOsmol rise between samples).
14.Now give 2μg DDAVP IM (or 20μg intranasally) and collect urine samples only for a further 2h. Allow free fluids at this stage.
Interpretation: normal response: plasma osmolality remains in the range 280–295mmol, urine osmolality rises to >2 × plasma (>600mOsmol). If urine volumes during water deprivation do not reduce and yet the
plasma does not become more concentrated (rising osmolality) and weight does not fall, suspect surreptitious drinking during test.
Diagnostic trial of DDAVP
Indication: distinction of partial diabetes insipidus from primary polydipsia.
Contraindications: cardiac failure. Current diuretic use (test uninterpretable). Note that this test may precipitate severe hyponatraemia in primary polydipsia and should be preformed in an inpatient unit with clinical and biochemical regular review.
Preparation: admission to assessment unit. First line tests for polydipsia/polyuria should have been performed.
Procedure:
1. 24h urine volume, morning urine osmolality, weight, fluid intake (as far as possible), serum osmolality, Na+, urea and creatinine should all be performed daily and the results reviewed the same day.
2. Subjects should have access to fluid ad libitum but should be reminded that they should only drink if they are thirsty.
3. After an initial 24h period of observation, desmopressin (DDAVP) is administered at a dose of 2mg bd SC for 3 days.
Stop test if serum Na+ falls to <130mmol/L.
Interpretation: reduction in urine volume to <2L/day, 4 in urine osmolality to >600mOsmol/L without fall in serum Na+ to <140mmol/L suggests central diabetes insipidus. Reduction in urine volume with no increase in urine osmolality >600mOsmol/L and without a fall in serum Na+ suggests partial nephrogenic diabetes
insipidus. Limited reduction in urine volume, with some increase in urine osmolarity but a fall in serum Na+ suggests primary polydipsia.
Low dose dexamethasone suppression test
Indication: to distinguish hypercortisolism from normality. The dexamethasone suppressed CRH test is believed to have less false positives in cases of alcholic or depressive pseudo-Cushing’s syndrome.
Patient preparation: patients should not be on oral steroids or drugs that increase steroid metabolism.
Overnight dexamethasone suppression test: 1mg dexamethasone is taken PO at midnight. Serum sample for cortisol is taken the following morning between 0800 and 0900h.
Interpretation: serum cortisol should suppress to <140nmol/L (usually <50nmol/L). Values 140–175nmol/L are equivocal and suggest a 2-day test should be performed. 10–15% false +ve rate.
2-day low dose dexamethasone suppression test (preferred): dexamethasone 0.5mg is giv
en PO every 6h for 8 doses (2 days) starting in the early morning. Ideally tablets are taken strictly at 6-hourly intervals (0600, 1200, 1800, 0000h) which may necessitate an inpatient stay. A 24h collection for urine free cortisol is taken on the second day of the test and serum cortisol is measured at 0600h on the 3rd day, 6h after the last dose. IV administration of dexamethasone can be used if there are concerns over absorption or compliance.
Interpretation: serum cortisol 6h after the last dose should be <140nmol/L, usually <50nmol/L. Urinary free cortisol on the second day should be <70nmol/L, normally <30nmol/L. The 2-day test strictly
performed has less false +ves than the overnight test.
Dexamethasone suppressed CRH test: dexamethasone 0.5mg is given PO every 6h for 8 doses (2 days) but starting at midnight and ending at 0600h. Tablets are taken strictly at 6-hourly intervals (0000, 0600,
1200, 1800h) which may necessitate an inpatient stay. Last dose is taken at 0600h and an injection of CRH (100μg IV or 1μg/kg) is given at 0800h. A blood sample for cortisol is taken at 0815h (i.e. 15min later).
Interpretation: serum cortisol level should be <38nmol/L (normal).
High dose dexamethasone suppression test
Indication: to distinguish between patients with Cushing’s disease (ACTH-secreting pituitary tumour) and ectopic ACTH production in patients with established hypercortisolism.
Patient preparation: as low dose test except that the test can be performed immediately following the 2-day low dose test.
Procedure:
1. 2 × 24h urine free cortisol collections are made to calculate the mean basal 24h urine free cortisol.
2. Baseline serum cortisol measurement is also taken before the first dexamethasone dose, ideally at 0600h. If the low dose test is performed first, the baseline values (urine and blood) must be taken prior to the
low dose test (i.e. any doses of dexamethasone).
3. Dexamethasone 2mg is given PO every 6h for 8 doses (2 days) starting in the early morning. Ideally tablets are taken strictly at 6-hourly intervals (0600, 1200, 1800, 0000h) which may necessitate an inpatient stay.
4. A 24h urine collection for urinary free cortisol (final) is taken on day 2 and a blood sample is taken for (final) cortisol 6h after the last dexamethasone dose (0600h on day 3). Creatinine excretion should be
measured and compared between urine samples to confirm true 24h collections.
Interpretation: % suppression of basal cortisol is calculated as: (basal cortisol–final cortisol)/basal cortisol × 100.
The same calculation is made for basal and day 2 urine free cortisol. 50% suppression is suggestive of pituitary-dependent disease. 90% suppression increases the likelihood (strict criteria). Thymic carcinoids
and phaechromocytomas releasing ACTH are source of false positives.
Short synacthen test
Indication: suspected adrenal insufficiency. Will not detect recent-onset
secondary adrenal insufficiency.
Contraindication: asthma/allergy to ACTH—risk of allergic reaction (can be performed with careful medication supervision of patient).
Preparation: patient must not take hydrocortisone on the morning of the test as this will be detected in the cortisol assay. The test can be performed on low dose dexamethasone but the morning dose should
be omitted until after the test. May have some value in patients on higher dose steroid therapy to indicate the degree of suppression of adrenocortical function.
Procedure: 250μg of synthetic ACTH (synacthen) given IM or IV. Blood taken at times 0, 30 and 60min for serum cortisol. A value at any time >550nmol/L makes the diagnosis very unlikely.
Low dose test: the test can be performed with a very low dose of ACTH (e.g. 1μg). This may detect more subtle degrees of hypoadrenalism but the clinical significance of these findings remains uncertain.
Long (depot) ACTH test
Indication: distinguishing 1° and 2° adrenal failure.
Patient preparation: a short synacthen test should be performed prior to the test to diagnose adrenal failure. If patient is on steroid replacement, change to dexamethasone 0.5mg/day.
Procedure: blood is taken at 0900h for basal cortisol. 1mg of depot synthetic ACTH (synacthen) is then given IM on 2 consecutive days and blood collected 5h after each dose (1400h). A final cortisol sample is
taken at 0900h on the 3rd day.
Interpretation: serum cortisol should rise to >1000nmol/L on the last day and, if adrenal failure previously indicated by a short synacthen test, such a rise indicates secondary adrenal failure (pituitary/hypothalamic
cause inc. suppressive drugs).
