View / Download pdf version of this article

The Biochemistry Subgroup

As part of the wider review of the New Zealand Laboratory schedule, a biochemistry subgroup was formed to identify tests where special expertise was considered appropriate for interpretation of results, and tests where guidelines or restrictions on requesting were thought to be necessary. The group was also asked to identify tests which were outdated or of no clinical value and for which funding should be withdrawn, as well as to identify underutilised tests which should be encouraged first-line.

The key drivers for this process were:

  • The desire for a national schedule that was relevant to the current evidence base and best practice
  • The desire to develop more consistency of testing across DHBs
  • A lack of clarity regarding appropriate and cost-effective testing, as there was no guidance on limiting testing
  • The intention for the schedule to interface with an e-labs initiative, and electronic test requesting

Ultimately there was concern not only because of the increasing volume of laboratory testing in general, but also because the requesting of certain “vogue” tests had increased dramatically in a way not justified by current overall evidence. Particular attention was suggested for those tests which “create issues in terms of volume and requesting appropriateness”. The background rationale was to allow appropriate, evidence-based spending on pathology testing by DHBs facing increasingly constrained laboratory budgets. The intention of the review was not to place blanket restrictions on tests, but rather to provide guidance on appropriate test requesting.

The guidelines produced are not mandatory but were developed as a resource for individual DHBs to use. They are not intended to replace well-established local protocols or clinical pathways, but rather to support them where judged appropriate by local clinicians and policy setters.

Composition and process of the biochemistry subgroup

The biochemistry subgroup was composed of six Chemical Pathologists representing different DHBs, from both public/academic and private (community) backgrounds, along with a convener from DHB Shared Services. Individual members were each allocated a range of tests to evaluate and present recommendations for wider discussion among the group. Specialists from related clinical disciplines were consulted when appropriate. In all cases where guidelines or restrictions were put in place the strength of evidence base, and the opinions of local experts were considered, and there was ultimately unanimous agreement among the group. Third party stakeholders also had the opportunity to provide feedback on an initial draft set of guidelines, and suggestions were incorporated into the final document.

In biochemistry there were a significant number of “esoteric” tests identified, which were considered to be Tier 2 tests, i.e. requiring special expertise in interpretation. Many of these tests are rarely requested and, while detailed criteria or guidelines for requesting them have not yet been recommended, requestors are encouraged to contact the laboratory or a specialist in the relevant clinical discipline to discuss appropriate requesting and interpretation.

It is intended that the Laboratory Schedule and Test Guidelines will be updated and modified as new evidence comes to light, new tests are added and others become outdated. As electronic ordering becomes standard practice there will be opportunity to guide testing based on clinical presentation and minimise inappropriate testing frequency, e.g. requesting HbA1c more often than every three months without special circumstances.

Biochemistry tests (referred to as chemical pathology in the schedule) were divided into four groups:

  • Tests where it is appropriate to recommend ordering restrictions and/or criteria for funding based on clinical circumstances and/or expertise of referrer
  • Tests which are outdated and which should be funded only in very limited circumstances
  • Tests where public funding was not considered justified based on current evidence
  • Tests which were considered underutilised, but for which requesting guidelines were appropriate to optimise clinical utility

Tests with restrictions

The following tests are examples of those that have recommended guidelines or criteria for their use and should be requested only in specific clinical situations.

Androgen tests

Restricted tests include androstenedione (ASD), dehydroepiandrosterone sulphate (DHEAS), sex hormone binding globulin (SHBG) and free testosterone:

  • In the assessment of hirsutism measurement of ASD and DHEAS is not justified unless testosterone is also elevated (except when requested by specialist Endocrinologists, or pre-authorised by a Chemical Pathologist)
  • Measurement of sex hormone binding globulin (SHBG) and calculated free testosterone is not justified unless the initial total testosterone result is in a range where SHBG/free testosterone is likely to provide additional clinical value
  • Measurement of dihydrotestosterone is only justified in isolated rare clinical scenarios of defective androgen action or response, e.g. partial or complete androgen insensitivity

DHEAS and ASD are androgens sometimes measured in addition to testosterone in the assessment of women with hirsutism and possible polycystic ovary syndrome (PCOS). Free testosterone, derived from measurement of total testosterone and SHBG, is also sometimes advocated as providing a better measure of tissue androgen exposure.

The added value of measuring these hormones is very limited in the large majority of patients being evaluated for possible PCOS. The main reason for initially performing such tests is to exclude other secondary causes, particularly virilising ovarian or adrenal tumours. However, these conditions occur very rarely and patients will virtually always have an unusual clinical presentation with relatively severe and rapidly progressive hirsutism, and/or evidence of virilisation. Even for these patients, it is extremely uncommon for there to be isolated elevation of DHEAS or ASD without testosterone elevation (which is usually marked).

N.B. Testosterone levels are not always raised in females with PCOS. Measurement of testosterone levels (total testosterone), while often carried out, is not required for diagnosing PCOS. The diagnosis is based on a constellation of findings related to clinical and/or biochemical evidence of androgen excess, menstrual irregularity and ovarian dysmorphology (usually multiple peripheral ovarian cysts).

Exclusion of other secondary causes such as Cushing’s syndrome and congenital adrenal hyperplasia (mostly late onset 21 hydroxylase deficiency) involves measurement of other specific tests (urine free cortisol and/or overnight dexamethasone suppression, and 17OH-progesterone).

Measurement of DHEA or ASD has also been advocated in patients taking these as supplements. However, the biochemistry subgroup consider supplementation with DHEA or ASD (“andro”) to be of unproven clinical value (and unclear long-term clinical risk), except in certain situations, such as in patients with premature ovarian failure, hypopituitarism and possibly some other limited settings, such as some female patients with SLE.1 Even in these patients, measurement of DHEAS and ASD is of unclear and unproven value in monitoring their treatment.

Sex hormone binding globulin (SHBG), which is used to calculate free testosterone, is also of limited value in most patients. Evaluation at LabPlus shows that all female patients with a testosterone > 5 nmol/L will also have a raised free testosterone, and those with total testosterone < 1.3 nmol/L have a free testosterone within reference limits. There is little additional clinical value therefore in measuring SHBG/free testosterone for samples with total testosterone outside these limits. Even for patients with total testosterone within this range, only those with unusually high (e.g. taking oral contraceptives, hyperthyroidism) or low (e.g. obese, insulin resistant) SHBG levels are likely to have a reclassification of testosterone to within or above reference limits based on their free testosterone result. For similar reasons, in males, free testosterone adds value only if the total testosterone is between 7 – 15 nmol/L.

Dihydrotestosterone measurement is extremely expensive and adds little to the clinical management of patients with hirsutism (even those taking 5-alpha-reductase blockers, such as finasteride). This test is of established clinical utility only in patients being evaluated for very rare defects in androgen action or response (e.g. partial or complete androgen insensitivity) in specialist settings.

For further information see: “Reproductive hormones: the right test, at the right time, for the right patient”, Best Tests (Feb, 2013).

Tests of adrenal function

24h urine free cortisol (UFC) has well-established value in the initial evaluation of patients with possible Cushing’s syndrome.2 A 24 hour urinary excretion result over four times the upper reference value makes Cushing’s highly likely. Lesser degrees of elevation can reflect a broad range of other factors, such as stress, illness, insomnia, depression, anorexia and alcoholism, as well as Cushing’s.

The clinical utility of 24h cortisol excretion for the evaluation of possible primary or secondary hypoadrenalism is, however, very limited and the group did not consider this to be an appropriate clinical indication for this test. There are other established means with much better clinical utility to make this diagnosis, such as synacthen testing and, for primary adrenal disease, plasma adrenocorticotropic hormone (ACTH).

While there is a loose correlation between 24h urine cortisol production and cortisol output, excretion can be affected by a range of factors and can vary significantly from day-to-day, even in healthy patients exposed to temporary physical or psychological stress. Patients with primary adrenal insufficiency may also have daily excretion well within reference limits, but output is stimulated by increased ACTH stimulation (in a similar way to patients with mild hypothyroidism with free T4 maintained within reference limits by increased TSH).

Many requests for UFC are made in the belief that functional adrenal insufficiency (“adrenal fatigue”) is a cause for chronic fatigue syndrome. There is no substantive evidence for “adrenal fatigue” as a real clinical entity. The use of hydrocortisone treatment in chronic fatigue syndrome is not supported by randomised controlled trial evidence,3, 4 and both United Kingdom and Australasian guidelines specifically state that hydrocortisone should not be used in chronic fatigue syndrome.5, 6

Cortisol binding globulin (CBG) measurement is considered to have no clinical utiltity other than in rare situations where calculation of free cortisol adds clinical value to the patient’s management, almost always in specialist settings. This would typically be where a total cortisol result (usually on stimulation testing) seemed inconsistent with the patient’s clinical presentation. CBG is therefore considered a specialist test (Tier 2).

Salivary cortisol measurement is appropriate for the evaluation of patients with possible Cushing’s syndrome.2 Since saliva reflects the level of free cortisol in the tissues (salivary glands), it provides an indirect measurement of tissue cortisol exposure. Normal, unstressed patients show a marked fall in salivary cortisol in the late evening, whereas in patients with Cushing’s syndrome cortisol levels, and salivary cortisol, remain elevated.7 However, as with 24 hour urine free cortisol tests, other non-Cushing’s causes of elevation can occur, such as patients with significant physical or psychological stress. A late night (10 – 11 pm) saliva sample can be collected by patients before bed and sent to the laboratory the following day.

Measuring salivary cortisol samples or profiles at other times of the day as a means of assessing tissue cortisol exposure, and thereby diagnosing cortisol excess or deficiency (organic or functional, “adrenal fatigue”) is considered unproven and lacks sufficiently robust evidence at this time to justify public funding.

Tests of thyroid function

No restrictions or guidelines around thyroid stimulating hormone (TSH), Free T4 (FT4)and thyroid antibody testing have been included in the recommendations (these are all Tier 1 tests), but formal schedule guidelines on tests of thyroid function are planned.

It is important to note that:

  • FT4 is not considered an appropriate initial request for the routine assessment of thyroid status unless an unusual cause, such as pituitary disease (secondary hypo- or hyperthyroidism) is suspected. When this is not specified, reflex addition of FT4 occurs in most laboratories when TSH is abnormal.
  • The FT4/FT3 ratio may be influenced by a range of factors including drug treatment, illness and fasting status. While it may also be influenced by some trace elements such as iodine and selenium it was not considered a sufficiently reliable marker for this purpose.
  • Thyroid peroxidase (anti-TPO) is considered the appropriate first-line antibody test for autoimmune thyroid disease. Anti-thyroglobulin may add some value when anti-TPO is raised but can cause confusion when raised in isolation. Anti-thyroglobulin testing is important, however, in the management of patients with thyroid cancer. Repeated monitoring of anti-TPO titre has been advocated in the monitoring of iodine status, but there is little substantive evidence base for its value in this context.


Free T3 (FT3), and its precursor FT4, levels are patient-specific with an individual “set point” much narrower than the population range. This is mostly due to individual variation in tissue sensitivity to thyroid hormone, but also other factors, such as the enzymatic conversion of T4 to T3 by tissue deiodinases (mainly type 1 in the liver). This is influenced by factors such as recent calorie intake, mineral status (such as iodine and selenium), growth hormone levels and thyroid status itself.

While all routine thyroid tests (TSH, FT4, FT3) can be affected temporarily by factors such as illness and drugs, FT3 is particularly affected by illness and also by reduction in calorie intake, with both of these causing a rapid decrease in plasma level.

FT3 requests are justified in the following circumstances:

  • If TSH is low and FT4 is normal (to exclude T3 toxicosis): FT3 is routinely added by most laboratories in this situation, even if not requested
  • When hyperthyroidism (including secondary hyperthyroidism) is suspected or monitored based on clinical details
  • If there is known or suspected pituitary/hypothalamic disease: FT3 is not considered appropriate, however, for routine monitoring of primary hypothyroidism
  • In patients with thyroid cancer, where FT3 measurement is occasionally helpful to monitor the degree of replacement (which in advanced cases can be above physiological requirements)

In early hyperthyroidism or primary hypothyroidism (thyroid failure, most often Hashimoto’s disease) the serum level of TSH falls, or rises, early and is a sensitive biomarker of tissue exposure. It is therefore the single most useful initial test when either primary hyper- or hypothyroidism is suspected. Serum levels of FT4 and FT3 may rise and fall compared with the patient’s individual set point, but typically initially remain within population limits.

In primary hyperthyroidism FT3 may rise above population limits before FT4 (so-called “T3-toxicosis”), and it is useful to perform a FT3 assay when TSH is low (typically suppressed to unmeasurable levels in true hyperthyroidism) but FT4 is within reference limits.

In secondary hyper- or hypothyroidism (pituitary/hypothalamic disease) TSH measurement alone is unreliable, and it is very important to measure FT4 in such patients, both for initial screening/evaluation and in monitoring. FT3 measurement can also be useful, especially if there is an abnormality of growth hormone production (growth hormone insufficiency can reduce the conversion of FT4 to FT3).8

While theoretically the plasma level of FT3 can be of value in assessing patients with hypothyroidism, there are many factors that confound interpretation, such as the individual patient set-point (which is unknown), recent illness or calorie and iodine intake. In patients with primary hypothyroidism and in iodine deficiency FT3 levels are generally preserved within population limits until relatively late (unlike in hyperthyroidism), making it an insensitive marker.

In patients taking T3 replacement, either alone or in combination with T4 (e.g. whole thyroid extract), FT3 levels rise and fall significantly depending on time of last dose and are not considered sufficiently reliable for monitoring. As with patients taking conventional replacement treatment, TSH is considered the primary analyte by which to adjust dose.

Tests of pituitary function

Insulin-like growth factor 1 (IGF-1) is an accepted test for the initial investigation of growth hormone excess (acromegaly, gigantism), and in monitoring the treatment of such patients. Since identification of acromegaly is important and the test has well-established clinical utility (even though the diagnosis is rare), writing “possible or known acromegaly” on the request form is sufficient for the test to be funded.

IGF-1 may also be requested, when recommended by a Chemical Pathologist or Endocrinologist, as an initial investigation of the possibility of growth hormone deficiency. However, interpretation is much more likely to be confounded by other factors, such as nutritional status, oestrogen and thyroid hormone status. A low result is more likely to be clinically significant when prior suspicion is high, e.g. patients with other anatomical or biochemical evidence for pituitary disease. Formal diagnosis of growth hormone deficiency (i.e. to qualify for publically funded treatment) requires further testing in a specialist setting.

Measurement of IGF-1 in patients on certain weight loss diets, e.g. the intermittent fasting (“5+2”) diet, is not considered sufficient reason to justify public funding.

Growth hormone measurement can be helpful in the evaluation of patients with pituitary disease, particularly when acromegaly is suspected or in children or adults when there is suspicion of hypopituitarism. The test is funded if one of these indications is specified on the request form, or when ordered by an Endocrinologist.

A major problem limiting interpretation, however, is that growth hormone is secreted in a pulsatile fashion, so unless a result is clearly high or low, a single isolated result can be impossible to interpret. Stimulation or suppression tests, or serial measurements throughout the day, provide additional information; this should only be carried out under specialist management or recommendation.

Assessment of pancreatic disease and obesity

Plasma insulin levels are a key measurement when establishing a diagnosis of insulinoma as a cause of recurrent hypoglycaemia; since insulin has a plasma half-life of minutes and insulin secretion is shut off by hypoglycaemia in normal patients, plasma insulin levels should be suppressed. As evaluation of possible insulinoma is complex, prior discussion with an Endocrinologist or Chemical Pathologist is recommended before requesting this test.

When considering possible insulinoma it is critical to:

  • Measure venous plasma glucose concurrently, so that the plasma insulin level can be properly interpreted. If the plasma glucose is > 3 mmol/L, then there is no stimulus to shut off pancreatic insulin release and plasma insulin level will be unhelpful
  • Document any hypoglycaemic symptoms at the time, particularly those associated with poor glucose supply to the brain (neuroglycopaenic symptoms), such as confusion, “absence” and disorientation
  • Document fasting status or time since last meal

Patients who have had bariatric surgery can develop excessive inappropriate pancreatic insulin secretion. For these patients, measuring insulin and glucose together at the time they describe symptoms is considered reasonable for any referrer, as long as the clinical information details that the patient had previous bariatric surgery.

While controversial, the biochemistry subgroup felt that evidence to justify funding of plasma insulin to identify insulin resistance and the metabolic syndrome was not sufficiently robust to justify public funding, except in specialist settings and then preferably when used as part of a calculation incorporating concurrent glucose level. For example, calculation of the HOMA index of insulin resistance may be useful in assessing the probability of non-alcoholic steatohepatitis (NASH) and the need for liver biopsy to assess fibrosis.9, 10

Insulin levels are not useful in patients with diabetes, as they can range from very high to unmeasurably low. They should also not be used to decide whether a patient has type 1 or type 2 diabetes; other tests such as diabetes-related antibodies (anti-GAD, anti-IA2) and plasma C-peptide have greater utility.

C-peptide is stored in secretory granules with insulin and co-released in equimolar amounts. Measuring plasma C-peptide is useful in the context of evaluating possible excess endogenous insulin secretion (e.g. insulinoma) and distinguishing this from exogenous insulin administration or another cause. Fasting status or relationship to meals should be well defined and plasma glucose should be measured concurrently. Ideally the sample should be taken during a spontaneous hypoglycaemic attack or a controlled fast, with careful correlation with symptoms. C-peptide is filtered by the glomeruli and caution should be exercised in patients with reduced GFR as this may lead to elevated values independent of any changes in pancreatic status. C-peptide may also be helpful in classifying some patients, when there is uncertainty as to whether they have type 1 or type 2 diabetes.11 The utility of C-peptide for assessing insulin resistance is limited and it is not recommended for this purpose.

Nutritional markers: Essential fatty acids, vitamins, iodine and trace elements

Essential fatty acids (EFAs) are divided into two main classes: omega-3 and omega-6. The shortest chain omega-3 essential fatty acid is linolenic acid, and the shortest omega-6 is linoleic acid.

The most well known longer chain EFAs are:

  • Omega-6 – arachidonic acid (C20:4n6), a precursor to prostaglandins and leukotrienes
  • Omega-3 – eicosapenatenoic acid (C22:5n3 – EPA) and docosahexaenoic acid (C24:6n3 – DHA) (‘fish oils’)

There is considerable literature on the biology and benefits of n3 and n6 EFAs, and increased intake of omega-3 rich foods has been reported to have beneficial cardiovascular and anti-thrombotic effects, as well as a wide range of other less well substantiated benefits. There are also some isolated reports that higher plasma levels of some EFAs in plasma and/or red cells are associated with better long-term outcomes, but randomised trial evidence using plasma levels as a marker is currently limited.

EFA testing is technically difficult and very expensive. This test is not appropriate for patients who are considering or taking EFA supplements. Based on current evidence, knowing the detailed composition of EFAs in plasma and red cells was not considered sufficient to justify publically funding such requests at this time. Targets to guide treatment are not clearly established, correlation with tissue levels is imperfect, and there is potential for confusion due to the range of other biological and dietary influences. Achieving an appropriate balance of EFAs is important in some limited clinical settings, such as patients with severe liver disease or short bowel syndrome on intensive nutritional support. An EFA test would be appropriate in this setting.

Vitamins B1 (thiamine), B2 (riboflavin), and B6 (pyridoxine). Plasma levels of these vitamins are sometimes requested as part of an overall nutritional or wellness screen. However, clinically significant deficiency is rare in New Zealand, except in the context of significant malnutrition or malabsorption, and/or liver disease (e.g. alcoholism). All of these vitamins are water soluble with very limited storage in tissues such as fat, hence plasma levels will be very influenced by recent short-term intake.

The assays are all expensive and there are significant pre-analytical factors of collection, processing and storage to consider which, if not addressed correctly, will invalidate the result. Even if the patient is suspected to have a deficiency, testing is often unhelpful as the turnaround is slow. The clinical response to vitamin supplementation is more helpful in confirming the diagnosis, and is the only way to prove that symptoms leading to the suspected diagnosis were related to deficiency of that particular vitamin.

Patients who have had bariatric surgery are predisposed to vitamin and trace element deficiency, in some cases leading to short and long-term neurological complications, including Wernicke’s encephalopathy, polyneuropathy and visual defects. Post-operative monitoring of nutritional status is considered appropriate in this situation and requests for vitamin B1 and B6 are approved.12 Measurement of vitamin B6 (pyridoxine) is justified in a specialist setting, when investigating a patient with raised homocysteine levels.

Vitamin D has a central role in bone and calcium metabolism and vitamin D tests were developed for investigation of abnormalities of calcium metabolism as well as metabolic bone disorders, such as rickets and osteomalacia. In recent years an association has been reported between low vitamin D levels and a very wide range of disorders (cancers, cardiovascular disease, diabetes, autoimmune disorders and infectious diseases). However, a causal link has yet to be demonstrated for any of these conditions.13–15

Despite this, the number of requests for vitamin D tests has increased dramatically, with many patients who get reasonable sun exposure and who are otherwise at relatively low risk, wishing to know their vitamin D level.

A comprehensive literature review for the Ontario Ministry of Health concluded that there is little evidence that it is useful to test vitamin D concentrations in patients without symptoms of metabolic bone disease.16

It is not necessary to routinely measure vitamin D in patients with low bone density. It is reasonable to routinely provide vitamin D supplements (1.25 mg or 50,000 IU cholecalciferol per month), without testing vitamin D, to frail housebound or institutionalised elderly people, or those in the community who avoid sunlight for cultural or medical reasons.

Requests for a vitamin D test should clearly indicate a high risk of vitamin D/calcium abnormalities for investigation, e.g:

  • Rickets or osteomalacia, known osteoporosis, abnormalities of calcium/phosphate metabolism, raised ALP with likely bone cause
  • Cystic fibrosis, special diets (e.g. PKU), renal transplant, anticonvulsant use
  • Children (16 years and under) and refugees
  • Prior to treatment with interferon for hepatitis C

For further information see: “Vitamin D supplementation: navigating the debate”. BPJ 36 (Jun, 2011).

Vitamin K is a fat-soluble vitamin important in the post-translational modification (gamma-carboxylation) of a number of proteins, importantly some clotting factors (II, VII, IX and X), and also certain bone proteins. Measuring vitamin K levels directly is rarely helpful except in limited specialist settings.

People at risk of vitamin K deficiency include those with fat malabsorbtion (e.g. chronic pancreatitis, cystic fibrosis, parenteral nutrition) and some neonates. However, a vitamin K test is not indicated as part of the general investigation of nutritional status and possible malabsorption.

The appropriate investigation of patients with clotting disorders due to possible vitamin K deficiency is the direct assessment of clotting status (raised prothrombin time and, if more severe, raised activated partial thromboplastin time). Echis ratio (a further test of clotting) may also sometimes be helpful. Plasma levels of individual clotting factors can also be measured if required.

Coenzyme Q10 (CoQ10, vitamin Q, ubiquinone) is important in mitochondrial oxidative metabolism and energy production, as well has having natural antioxidant effects.The most clearly established reason for measurement is the investigation of rare inborn metabolic defects, in which there may be primary or secondary CoQ10 deficiency.

Plasma CoQ10 measurement has been suggested to be useful in statin-induced myopathy, heart failure and neurological disorders such as Parkinson’s disease. There is biological rationale for an intracellular deficiency of CoQ10 as a factor in these conditions. However, the correlation between plasma and intracellular (e.g. muscle biopsy) levels of CoQ10 is limited. Since CoQ10 is also mostly carried in the lipid fraction, statin treatment will inherently lower CoQ10 levels independent of those in tissues. Therefore this test is not recommended for this purpose.

Some evidence suggests that low CoQ10 predicts worsened mortality in heart failure and achieving a higher level may be associated with a better outcome in patients taking supplements. However, other trials have suggested no benefit and the value of measuring CoQ10 in these conditions at this time awaits further evidence.17, 18

For these reasons the group recommended CoQ10 measurement should be restricted to Cardiologists, Neurologists and Paediatricians managing patients with the above disorders.

Although it has been advocated, the use of CoQ10 measurement and treatment in chronic fatigue syndrome has weak evidence-base.

Urine iodine levels reflect recent iodine intake and vary widely from day to day depending on recent food intake; even a patient with relatively low body stores can have normal excretion if analysed within two to three days of an iodine-rich meal (foods rich in iodine include most seafood and seaweed, eggs/poultry, milk and sometimes soy products). Routine urine iodine testing has no established role in general practice, and there is no evidence that it leads to any beneficial outcomes in patients who are appropriately monitored for hypothyroidism and appropriately supplemented in pregnancy. Routine inclusion of iodine in a vitamin supplement (but not iodine testing) has been recommended in women who are pregnant by the Royal Australasian College of Obstetricians and Gynaecologists.19

The median urine iodide level in a population can be used as an index of population iodine status, however, urine iodine excretion (both spot urine iodine creatinine ratio and 24h excretion) has very low predictive value for iodine deficiency in an individual patient. WHO guidelines for population medians do not apply to individual subjects and will grossly over-diagnose iodine deficiency if misapplied in this way.20 At least ten urine iodine collections are needed to provide a reasonable estimate of iodine status.21 The earliest functional evidence of iodine deficiency is a rise in TSH, which can be treated with iodine supplementation.21

Currently the only clearly established use of measuring urine iodine in individual patients is in the assessment of patients undergoing radioiodine treatment, where high urine iodine suggests poor thyroid radioiodine uptake and reduced treatment efficacy. It is also sometimes helpful in the evaluation of patients with hyperthyroidism.

Zinc, copper, and selenium, mercury, chromium and cobalt. Unless there is a high pre-test probability of deficiency (i.e. a pre-disposing condition, such as gastrointestinal disease), or toxicity (e.g. workplace exposure) it is rarely necessary to measure plasma copper, zinc, selenium or blood mercury in patients in general practice. Deficiencies of zinc or selenium do not occur in people who consume a reasonable diet and have normal gastrointestinal function.

Measurement of these trace elements may be useful in the management of patients predisposed to deficiency by malnutrition and/or gastrointestinal disorders and especially in patients taking parenteral nutrition.

Measurement of plasma and urine copper levels are also useful in the diagnosis and management of Wilson’s Disease (clinical details should state “? Wilson’s Disease” or “raised LFTs”) and in rare genetic disorders of copper metabolism (e.g. Menke’s syndrome).

These tests are also helpful in cases of zinc, copper and selenium poisoning, and cases of suspected poisoning are an indication for referral. Measurement of whole blood and urine mercury are of value in monitoring workplace exposure and when mercury poisoning is suspected.

Measurement of serum cobalt and chromium is indicated in patients with concern over possible overexposure. The most common situation is patients with a metal-on-metal joint prosthesis where there is concern over possible deterioration of the joint surfaces, and who may present with symptoms such as pain, swelling, limping or trouble walking, or noise coming from the joint. If cobalt and chromium levels are abnormally elevated, it is recommended to repeat the tests after three months. If levels from the second test remain abnormally elevated, discussion with the Orthopaedic Surgeon is recommended.

For further information see: “Testing serum cobalt and chromium in people with metal-on-metal hip replacements”. Best Tests (Dec, 2012).

High levels of cobalt and chromium can also occur in people working with ceramics or metals, excessive supplement intake or renal impairment. Urine testing is more appropriate than serum for assessing chronic occupational exposure.

Evidence was not considered sufficiently robust to justify the public funding of measurement of plasma zinc or the zinc/copper ratio in patients with depression, autism, other mental health disorders or chronic fatigue syndrome. Results of these tests are often misleading because low plasma zinc and raised copper levels are non-specific changes commonly seen in inflammatory states and chronic disease.

The presence of amalgam dental fillings or symptoms of fatigue, depression, cognitive decline etc. are not sufficient indications for measurement of blood or urine mercury levels. The major determinant of blood mercury is dietary fish intake, and amalgam fillings do not cause a clinically significant increase in blood mercury levels.22

Tumour markers

These include:

  • Acid phosphatase
  • CEA
  • CA125, C15-3, CA19-9, CA72-4

Apart from acid phosphatase, no formal restrictions have been placed on these tests at this time (Tier 1), however, guideline recommendations for requesting them have been developed.

The guidelines recognise the value of these tests for monitoring known malignancies of specific types in specific clinical settings. They can also be useful for diagnosis in patients with a high probability of cancer at presentation, e.g. CA125 in patients presenting with a suspicious ovarian mass, and can provide prognostic information.

Virtually none of the typical tumour markers are completely specific for malignancy, or for a particular type of malignancy. For example, while often thought of as useful in ovarian cancer, CA125 can also sometimes be raised in other malignancies such as pancreas, lung, breast, endometrium and non-Hodgkins lymphoma. It can also be raised in a wide range of benign disorders such as acute and chronic liver diseases, acute and chronic pancreatitis, rheumatoid arthritis, ulcerative colitis, endometriosis, menstruation, non-malignant ascites and pleural effusions and SLE. Similarly, while a very high CEA is strongly suspicious for malignancy, it can be raised in a wide range of cancers (e.g. gastrointestinal, lung, thyroid, breast), and also in benign diseases such as hepatitis.

The role of most soluble tumour markers in screening is still under evaluation but they are not currently recommended for this purpose in the general population based on insufficient large trial evidence for benefit.

As an example of the recommendations, the indications for measurement of CA125 are:

  • Patients with symptoms or signs associated with high suspicion of ovarian cancer: persistent continuous or worsening unexplained abdominal or urinary symptoms, pelvic mass
  • Case detection in patients at high risk of familial ovarian cancer
  • At diagnosis of ovarian cancer to provide prognostic information
  • After treatment to monitor response and detect relapse

However, measurement of CA125 is not indicated for:

  • Investigation of non-specific symptoms, when probability of malignancy is low
  • Screening of asymptomatic low risk population (in a low risk patient a mildly raised result is much more likely to be a false positive rather than a true positive)
  • Investigation of other suspected malignancies

Lipid and cardiovascular disease related tests

Apolipoproteins B (ApoB) and A1 (ApoA1). These tests measure the protein component of lipid particles, LDL (ApoB) and HDL (ApoA1) respectively. Since there is only one ApoB or ApoA1 molecule per particle, they give an estimate of particle concentration rather than total cholesterol concentration in those particles.

At present there are no restrictions on requesting these tests as the demand for them is very low, and there is little evidence that they are being inappropriately ordered.

A number of epidemiological studies (but not all) suggest that these tests, and their ratio, may be marginally more predictive than lipid measurements themselves. They may identify some patients with genetic dyslipidaemias, and possibly help identify residual risk in patients on aggressive statin treatment.

These tests are significantly more expensive than lipid tests and while there measurement is improving, they are less well standardised internationally. Their advantage of being able to be measured in the non-fasting state is of limited practical value as non-fasting lipid tests themselves are usually reliably interpreted in most patients.

For further information see: “Fasting may be unnecessary for lipid testing”, Best Tests Nov, 2013.

Lipoprotein (a) is a weak independent risk factor for premature coronary artery disease and thrombosis in the general population. Lp(a) levels are mainly genetically determined, change little over time, and are poorly responsive to diet or to lipid-lowering treatment. There is very limited evidence to support whether Lp(a) reduction reduces the incidence of cardiovascular events.

Based on current evidence, the group considered that measuring Lp(a) is not indicated as part of routine cardiovascular risk assessment in primary care.23 If the clinical approach is otherwise clear based on other risk factors, then measuring Lp(a) has little additional value. The group recommended that requests for Lp(a) be funded (once only per patient) when requested by Cardiologists, as part of a specialist lipid/metabolic clinic, or with prior Chemical Pathologist approval.

Measurement should be limited to certain uncommon situations, particularly:

  • Patients in whom assessment using traditional Framingham risk markers may be unreliable, e.g. an unexpectedly early personal history of CVD, or significant family history in the absence of clear Framingham risk factors
  • Where measurement may influence the decision of whether or not to start the patient on pharmacological treatment based on other risk factors

For further information, see: “Assessing cardiovascular risk: what the experts think”. BPJ 33 (Dec, 2010).

Lipoprotein electrophoresis was historically used to classify patients with likely familial dyslipidaemias (Frederickson classification), with interpretation being based on the staining pattern and intensity of different lipid fractions. However, this classification is now rarely used, electrophoresis is expensive and there are other clinical and laboratory means of recognising primary lipid disorders (e.g. apolipoprotein measurements, genetic tests). The group considered that lipoprotein electrophoresis should only be funded in specific clinical circumstances when requested by Cardiologists, Endocrinologists/metabolic specialists or Internal Medicine specialists.

The major remaining application of electrophoresis is when considering the rare diagnosis of type III dysbetalipoproteinaemia (broad beta or remnant removal disease). Such patients have palmar xanthomas and increased concentrations of apoB-containing remnant particles (VLDL remnants, IDL).

High sensitivity CRP. Inflammation is now considered to play an important role in atherosclerosis. In well, asymptomatic patients the baseline level of CRP (referred to as high sensitivity CRP or hs-CRP) is thought to reflect the underlying level of inflammation and to have a graded association with CVD risk. There is epidemiological evidence linking levels of CRP with levels of cardiovascular risk, however, recent data has suggested that the risk is not as strong as originally stated. Genetic studies also fail to support a clear causal link of hs-CRP with cardiovascular disease. The group recommended that hs-CRP is funded when requested or pre-authorised by a Cardiologist, specialist lipid, metabolic or cardiovascular disease clinic or a Chemical Pathologist.

It is thought that hs-CRP is able to refine CVD risk in people rated at intermediate risk with traditional risk factors, and thereby re-categorise them above or below a treatment threshold. However, no current guideline (including local guidelines) recommends using hs-CRP as part of routine risk assessment. The American Heart Association suggests that this use be at the physician’s discretion, especially in the context of deciding whether or not to prescribe a statin.

Recent data has suggested that using the value for hs-CRP in the Reynolds modification of the Framingham equation does not sufficiently alter risk in most patients at intermediate risk to be cost-effective.24 The current risk calculator used in New Zealand also does not allow data for hs-CRP to be used.

There is also debate about the validity of the main intervention trial (Jupiter trial) that has been quoted to support the use of stratification by hs-CRP to guide treatment with statins. Further analyses of this and other large randomised trials shows the relative benefit from statin treatment is similar regardless of initial CRP level, i.e. the test does not identify a unique group that is likely to benefit.25, 26

Homocysteine is a sulphur-containing amino acid interconverted with methionine in a very important cycle of intermediary metabolism (methylation cycle), in which folate and vitamin B12 are required co-factors. Deficiency of folate and vitamin B12 may be associated with raised homocysteine, but measurement of these vitamins directly is generally considered adequate to assess the patient’s nutritional status.

Population evidence shows raised plasma homocysteine levels to be associated with long-term cardiovascular risk, however, intervention trials using B vitamin supplementation (folate, B12, B6) to lower homocysteine have been disappointing, suggesting such supplementation may be associated with worse outcomes.27 It is therefore most likely that mild/borderline homocysteine elevation is not itself causative of vascular disease, but rather may be a marker of other more complex predisposing nutritional factors. Regardless, since modifying homocysteine has been proven to be of little benefit its measurement as a cardiovascular risk marker was not considered sufficient to justify public funding.

Measuring plasma homocysteine is indicated when a monogenic disorder of methionine and homocysteine metabolism is suspected, e.g. patients with early or atypical thrombosis (including presentations such as retinal vein thrombosis), and when homocystinuria is otherwise suspected on clinical grounds.

Homocysteine elevation has also been suggested to be a marker of long-term risk of neurodegenerative diseases, such as Alzheimer’s disease. A recent systematic review suggested there may be a weak association between raised homocysteine and dementia risk, but the evidence was of very low quality.28 As with vascular disease, there was no proof of causal relationship, and no proof that lowering homocysteine mitigates this risk. Raised homocysteine is also associated with other factors which are themselves known to increase long-term dementia risk, such as diabetes, renal impairment, and advancing age.

Outdated tests

The following tests have been replaced in favour of other tests with greater clinical utility in most situations.

Prostatic acid phosphatase. For the diagnosis and monitoring of prostate cancer this test has been almost entirely superceded by PSA, which has much higher sensitivity for early disease, better correlation with tumour burden and treatment response and is more sensitive in identifying residual disease. Acid phosphatase is also more affected by prostatic hyperplasia (BPH) and digital rectal exam (DRE) than PSA. International guidelines have therefore not recommended its use, as in the large majority of patients it has no proven clinical benefit in addition to PSA.29, 30

Prostatic acid phosphatase is raised in certain uncommon disorders such as Gaucher’s disease, however, other markers are preferred. It has also been used historically as a marker of bone resorption, but has been replaced by other markers with better biological and analytical performance.

The group recommended measurement of acid phosphatase when referred or pre-authorised by an Urologist, Internal Medicine Specialist, Paediatrician or Haematologist (or when pre-approved by a Chemical Pathologist).

Creatine kinase MB (CKMB). This isoenzyme of CK is present in highest concentration in heart muscle, but is also widely present at lower concentrations in skeletal muscle. It was widely used historically in the diagnosis of myocardial infarction. However, troponin (T or I) testing is far more sensitive and specific and has a much wider diagnostic window, with detection of myocardial injury generally before CKMB is increased and for up to 10 – 14 days. Recent guidelines, both internationally and from the New Zealand Cardiac Society, recommend troponin as the marker of choice in the investigation of patients presenting with possible acute coronary syndrome.31–33

CKMB testing has been suggested to be useful in the evaluation of possible reinfarction, but with modern troponin assays a change in troponin is usually reliable. In some patients where there may be an analytical issue with a particular troponin assay, an alternative (either Troponin T or I, or a different manufacturer’s assay) will usually solve the problem, avoiding the need for CKMB testing.

Faecal fat. Although used historically for identifying and monitoring patients with steatorrhoea, this is a poor screen as typically over 90% of pancreatic function must be lost before it becomes elevated. It is also a very unpleasant test for both the patient and laboratory. Most laboratories no longer offer faecal fat testing.

Measuring fat content in a small faeces sample can be performed by measuring a “steatocrit”, or by visualising fat droplets using a fat stain (this detects the large majority of patients with moderate/severe fat absorption). Other tests such as faecal elastase are both more sensitive and less onerous for evaluating pancreatic enzyme insufficiency. The only remaining use of faecal fat estimates (as steatocrit) is in specialist settings, e.g. as a means of quantitating the degree of fat malabsorption in patients on close monitoring of replacement regimens.34

Fructosamine. For a wide range of reasons, both biological and analytical, fructosamine is an inferior test compared with HbA1c for monitoring patients with diabetes. It has a much shorter window of monitoring glucose levels, has greater biological variation, and is affected by albumin turnover (especially significant proteinuria) and hydration status. International evidence for the long-term prognostic value of HbA1c is far greater and treatment targets are much better established.

Fructosamine should only be measured when a reliable HbA1c result cannot be obtained, e.g. in situations of altered haemoglobin turnover (e.g. ongoing active blood loss or venesection) and with certain uncommon haemoglobin variants. If a HbA1c analytical interference is identified then other HbA1c methods without interference can usually be found, which is the preferred approach (if in doubt the laboratory should be contacted to discuss).

In the rare situations where fructosamine testing is indicated, there is little value in measuring it more often than monthly.

Tests with insufficient evidence

These tests lack sufficient evidence to justify funding their analysis under any circumstances.

Red cell magnesium (RBC Mg). Plasma magnesium is considered to be adequate for assessment of magnesium status and there is insufficient evidence to justify the additional expense of RBC Mg measurement for any clinical purpose. Evidence linking red cell magnesium to chronic fatigue syndrome was felt to be unconvincing.35, 36

Salivary progesterone measurement has been advocated as a means of monitoring transdermal progesterone treatment in peri- and post-menopausal women. Serum progesterone levels in such women are very low, reflecting perhaps the poor systemic absorption of progesterone creams through the skin. The evidence base to justify public funding of the salivary progesterone test was considered insufficient by the group.37

Salivary testosterone levels add little clinical utility to a serum testosterone measurement. Levels in saliva are very low and in current assays the precision at these levels also hampers interpretation.

Underutilised, but expensive tests

The following tests have increasing evidence for their clinical utility when requested within appropriate clinical guidelines, but are relatively expensive.

In some cases tests were recognised as being very good tests in specific clinical circumstances and, even though expensive, were probably underutilised. However, there were also situations where their clinical utility was limited and when the temptation to request them should be avoided.

BNP and NTProBNP is an example of such a test.

It is recommended that BNP or NT-ProBNP is requested in the following situations:

  • Exclusion of heart failure as a cause of unexplained breathlessness and other non-specific symptoms
  • Management of anti-heart failure treatment (secondary role only, usually for difficult to treat patients). There were no formal restrictions recommended for non-cardiologists, but it is recommended that repeat testing occur no sooner than two weeks between tests and, additionally, no more than four tests per year, per patient (more frequent need than this suggests excessive use or need for specialist involvement)

These tests have high negative predictive value for the exclusion of undiagnosed heart failure in patients presenting with non-specific symptoms and not already taking anti-heart failure treatment. Conversely, a clearly high result supports the diagnosis of heart failure and also carries adverse prognosis, independent of other variables (although in most acute cases this is clinically obvious through other means). However, mild-moderate elevation does not exclude the possibility of some other cause of breathlessness besides, or in addition to, heart failure. These tests also do not completely avoid the need for echocardiography, which provides other important information on cardiac structure and function, such as cardiac valve anatomy and (regional) myocardial contractility and relaxation.

The value of BNP and NTProBNP is much less well established for guiding ongoing anti-heart failure treatment. While a rise or fall can sometimes help guide treatment, proof of outcome benefit is much more limited and at present these tests have a secondary role only. NHF/NZGG guidelines do not specifically restrict use in this setting but have not encouraged it and NICE guidelines (UK) recommend their use be restricted to challenging patients under specialist management.

It takes at least two weeks for a new equilibrium level to be established and repeat measurement within this time frame is not recommended. Patients with heart failure who are difficult to manage should be referred for specialist review.

The Laboratory Schedule Test List and Laboratory Test Guidelines are available from:


Thank you to Dr Cam Kyle, Chemical Pathologist, Auckland for contributing this article and Dr Michael Crooke, Dr James Davidson, Dr Chris Florkowski and Dr Geoff Smith for expert review.


  1. Arlt W, Callies F, van Vlijmen J, et al. Dehydroepiandrosterone replacement in women with adrenal insufficiency. N Engl J Med. 1999;341: 1013-20.
  2. Nieman L, Biller B, Findling J, et al. The diagnosis of Cushing’s syndrome: An Endocrine Society Clinical Practice Guideline. J Clin Endocrinol Metab. 2008; 93: 1526-40.
  3. Van Den Eede F, Moorkens G, Van Houdenhove B, et al. Hypothalamic-pituitary-adrenal axis function in chronic fatigue syndrome. Neuropsychobiol 2007;55(2):112-20.
  4. Blockmans D, Persoons P, Van Houdenhove B, et al. Combination therapy with hydrocortisone and fludrocortisone does not improve symptoms in chronic fatigue syndrome: a randomized, placebo-controlled, double-blind, crossover study. Am J Med 2003;114(9):736-41.
  5. National Institute for Health and Care Excellence (NICE). Chronic fatigue syndrome/myalgic encephalomyelitis (or encephalopathy): diagnosis and management of CFS/ME in adults and children. NICE clinical guideline 53. NICE, 2007. Available from: (Accessed Feb, 2014).
  6. Chronic fatigue syndrome. Clinical practice guidelines-2002. Med J Aust 2002;176 Suppl:S23-56.
  7. Viardot A, Huber P, Puder J, et al. Reproducibility of nighttime salivary cortisol and its use in the diagnosis of hypercortisolism compared with urinary free cortisol and overnight dexamethasone suppression test. J Clin Endocrinol Metab. 2005;90: 5730-6.
  8. Alcantara M, Salvatori R, Alacantra P, et al. Thyroid morphology and function in adults with untreated isolated growth hormone deficiency. J Clin Endocrinol Metab. 2006; 91(3):860-4.
  9. Jung K, Cho S, Kim H, et al. Nonalcoholic steatohepatitis associated with metabolic syndrome: relationship to insulin resistance and liver histology. J Clin Gastroenterol. 2014;[Epub ahead of print].
  10. Bugianesi E, Marchesini G, Gentilcore E, et al. Fibrosis in genotype 3 chronic hepatitis C and non-alcoholic fatty liver disease: Role of insulin resistance and hepatic steatosis. Hepatology 2006;44(6): 1648-55.
  11. Jones A, Hattersley A. The clinical utility of C-peptide measurement in the care of patients with diabetes. Diabet Med. 2013;30:803-17.
  12. Ba F, Siddiqi Z. Neurologic complications of bariatric surgery. Rev Neurol Dis 2010;7(4):119-24.
  13. Grey A, Bolland M. Vitamin D: a place in the sun? Arch Intern Med. 2010;170:1099-100.
  14. Ross A, Manson J, Abrams S, et al. The 2011 report on dietary reference intakes for calcium and vitamin D from the Institute of Medicine: what clinicians need to know. J Clin Endocrinol Metab. 2011;96:53-8.
  15. Manson J, Mayne S, Clinton S. Vitamin D and prevention of cancer-ready for prime time? N Engl J Med. 2011;364:1385-7.
  16. Clinical utility of vitamin D testing: an evidence-based analysis. Ontario Health Technology Assessment Series 2010;10(2). Available from: (Accessed Feb, 2014).
  17. Molyneaux SL, Florkowski CM, Richards M, et al. Coenzyme Q10; an adjunctive therapy for congestive heart failure? N Z Med J. 2009;122(1305).
  18. Stocker R, Macdonald P. The benefit of coenzyme Q10 supplements in the management of chronic heart failure: a long tale of promise in the continued absence of clear evidence. Am J Clin Nutr 2013;97:233-4.
  19. The Royal Australian and New Zealand College of Obstetricians and Gynaecologists (RANZCOG). Vitamin and Mineral supplementation in pregnancy. RANZCOG bulletin C-Obs 25. RANZCOG, 2012. Available from: (Accessed Feb, 2014).
  20. World Health Organisation, UNICEF, ICCIDD. Assessment of iodine deficiency disorders and monitoring their elimination: a guide for programme managers; 3rd edition. WHO, Geneva: 2007.
  21. Konig F, Anderson M, Hotz K, et al. Ten repeat collections for urinary iodine from spot samples or 24-hour samples are needed to reliably estimate individual iodine status in women. J Nutr 2011;141:2049-54.
  22. Eyeson J, House I, Yang YH, Warnakulasuriya KA. Relationship between mercury levels in blood and urine and complaints of chronic mercury toxicity from amalgam restorations. Br Dent J 2010; 208(4):E7.
  23. Nordestgaard BG, Chapman M, Ray K, et al. Lipoprotein (a) as a cardiovascular risk factor: current status. Eur Heart J 2010;31(23):2844-53.
  24. Lee K, Cipriano L, Owens DK, et al. Cost-effectiveness of using high sensitivity C-Reactive protein to identify intermediate and low cardiovascular risk individuals for statin therapy. Circulation 2010;122: 1478-87.
  25. Lloyd-Jones DM, Liu K, Tian L, Greenland P. Narrative review: assessment of C-reactive protein in risk prediction for cardiovascular disease. Ann Intern Med 2006;145(1):35-42.
  26. Emerging Risk Factors Collaboration, Kaptoge S, Di Angelantonio E, et al. C-reactive protein concentration and risk of coronary heart disease, stroke, and mortality: an individual participant meta-analysis. Lancet 2010;375(9709):132-40.
  27. Marti-Carvajal AJ, Sola I, Lathyris D, Salanti G. Homocysteine lowering interventions for preventing cardiovascular events. Cochrane Database Syst Rev 2009;(4):CD006612.
  28. Ford AH and Almeida OP. Effect of homocysteine lowering treatment on cognitive function: a systematic review and meta-analysis of randomized controlled trials. J Alzheimers Dis 2012; 29(1):133-49.
  29. Sturgeon CM, Duffy MJ, Stenman UH, et al. National Academy of Clinical Biochemistry laboratory medicine practice guidelines for use of tumor markers in testicular, prostate, colorectal, breast, and ovarian cancers. Clin Chem 2008;54(12):e11-79.
  30. Bunting PS. Is there still a role for prostatic acid phosphatase? CSCC Position Statement. Canadian Society of Clinical Chemists. Clin Biochem 1999;32(8):591-4.
  31. Thygesen, Alpert JS, Jaffe AS, et al. Third universal definition of myocardial infarction. Euro Heart J 2012;33(20):2551-67.
  32. Non ST-Elevation Acute Coronary Syndrome Guidelines Group and the New Zealand Branch of the Cardiac Society of Australia and New Zealand. New Zealand 2012 guidelines for the management of non-ST elevation acute coronary syndromes. N Z Med J 2012;125(1357):122-47.
  33. Reichlin T, Hochholzer W, Bassetti S, et al. Early diagnosis of myocardial infarction with sensitive cardiac troponin assays. N Engl J Med 2009; 361: 858-67.
  34. Hill, PG. Faecal fat: time to give it up. Ann Clin Biochem 2001; 38:164-7.
  35. Hinds G, Bell NP, McMaster D, McCluskey DR. Normal red cell magnesium concentrations and magnesium loading tests in patients with chronic fatigue syndrome. Ann Clin Biochem 1994;31( Pt 5):459-61.
  36. Reid SF, Chalder T, Cleare A, et al. Chronic fatigue syndrome. Clin Evid (Online) 2008: 1101.
  37. Lewis JG, McGill H, Patton VM, Elder PA. Caution on the use of saliva measurements to monitor absorption of progesterone from transdermal creams in postmenopausal women. Maturitas 2002; 41(1):1-6.