Clarifying clinically significant abnormalities in endocrine function can represent a distinct challenge for numerous reasons. The challenges start with the numerous body systems that can be involved when an endocrine organ function is disrupted as well as the different ways these diverse body systems can react. They continue with the potential difficulties in establishing the clinical abnormalities that are truly due to the potential suspected endocrinopathy and conclude with all of the potential problems associated with measuring the activity of an endocrine gland and relating this to the clinical problem.

The clinical signs likely to develop as a result of impaired endocrine function are almost always expressed through altered function of a range of body systems but never directly the endocrine organ itself. Patients with low blood glucose because of uncontrolled overproduction of insulin develop signs referable to neuromuscular dysfunction as well as variably cerebrocortical dysfunction; however, the cause of the problem, a functional tumour of the B-cells of the islets of Langerhans, will not be in any way obvious. Patients with central diabetes insipidus develop clear indications of impaired renal function with no other abnormalities; those with diabetes mellitus develop changes again referable to impaired renal function but also malassimilation, while animals with impaired adrenocortical function express these deficiencies through abnormalities in the gastrointestinal, neuromuscular, cardiovascular and dermatological systems.

Perhaps the only real exception to this is those thyroid and testicular disorders where abnormal endocrine activity can be associated with a reasonably dramatic change in the shape of the gland. However, even then the question remains, is this abnormally shaped endocrine gland a reliable indicator of an organ with concurrent abnormal function? We are all aware that over 10% of older cats can have a palpably abnormal thyroid gland but not have clinically significant thyroid activity or that roughly 27% of dogs over the age of 12 had an adrenal "tumor" as an incidental finding at postmortem.

Consequently, the clinician is faced with the problem of logically approaching the question of whether or not the body system is primarily or secondarily involved; if appropriate, which one of the numerous explanations for secondary involvement of the affected body system is most likely; and, finally, how best to clarify if it is due to an endocrinopathy.

This brings us to the second area of difficulty in identifying clinically significant endocrine disorders. While some diseases of the endocrine system can be relatively easily recognised by specific markers of impaired hormonal activity, this is not the case for the majority of endocrine diseases.

Examples of relatively specific biological markers for altered endocrine function would include marked fasting hyperglucosemia and glucosuria suggesting significantly impaired insulin activity, ionised hypercalcemia with concurrent hypophosphatemia suggesting excessive PTH activity or marked unremitting hyposthenuria suggesting impaired vasopressin activity.

In these examples, it is worth considering two factors. First, you can instantly think of non-endocrine disorders that could produce similar changes in these biological "markers" and thus would need to be differentiated from the perhaps more likely endocrine causes. Second, these endocrinopathies are sending a rather obvious signal of their presence - the perturbation in a relatively specific "marker" or endpoint of the likely hormone involved. As such, we don't necessarily have to enter the realm of measuring serum or plasma concentrations of particular hormones we might be interested in.

This is a truly tricky business, as any interpretation of modified activity of a hormone will be based on an altered level of a particular hormone in serum. All hormones mediate their influence through receptors which are located on the surface of cells or within cells and all have effects on intracellular mechanisms, both cytosolic and intranuclear.

Thus the principal site of action for many hormones can be somewhat distant from and/or not particularly related to their concentration in serum or plasma. This "disconnect" between serum concentration and function is sometimes exacerbated by two further factors. First, the presence of "binding proteins" in serum resulting in variable amounts of inactive hormone being present and detectable in the circulation, even though it has no direct biological significance. Second, while a ligand-receptor interaction is going to be the direct activation pathway for a hormone's action, there are numerous factors that can influence how effective this first step in cell signalling is in bringing about a specific outcome.

As a result, interpreting alterations in a particular hormone's bioactivity on the basis of their concentration in the extracellular fluid, at a particular moment, is always going to be, at best, a very crude estimate of that hormone's function either currently or over the past days, weeks or months.

Consequently, there are numerous examples of where basal hormone concentrations are not going to provide sufficient discrimination, and some form of dynamic testing of the relevant endocrine system will be required. This opens a further area for debate regarding which dynamic test is more suitable for which specific endocrinopathy.

Finally, there is a third broad area where we need to be careful in how we estimate serum hormone concentrations. This is the very basic, but nonetheless massively important problem of the imprecision and inconsistency of the tools we generally use to measure serum hormone levels.

The vast majority of commercial assays used to measure hormone levels in serum or tissues are based on the concept of competitive binding. Effectively a limited number of binding sites specific for a hormone are set up in a closed system to which is added a very small amount of "labelled" hormone with sequentially increasing known concentrations of unlabelled hormone. The unbound hormone is removed and the amount of labelled hormone remaining (and hence bound to the specific binding sites) determined through whatever system is required to detect its "label". The means by which hormones are labelled is quite variable and ranges from chemicals which, when catalysed, change colour or become luminescent (chemiluminescence) to radioactive isotopes (radioimmunoassay).

Because the number of binding sites are limited, as the concentration of unlabelled hormone increases there will be increasing competition for these sites; thus with increasing concentrations of unlabelled hormone, less and less labelled hormone will remain bound. Consequently, if known concentrations of hormone displace predictable amounts of labelled hormone, the amount of labelled hormone displaced by a solution with an unknown concentration of the same hormone can be used to estimate the concentration of that hormone in the solution.

In general, two broad groups of hormone can be distinguished. One group are lipophilic and, with the exception of the iodothyronines, are cholesterol derivatives. The second group comprises various water-soluble hormones. Most of these are peptides which can be complex polypeptides (gonadotrophins), intermediate and smaller peptides (insulin, thyrotrophin-releasing hormone) and derivatives of single amino acids (catecholamines).

As a guideline, the lipophilic hormones are relatively stable and hence delays between collection and serum separation as well as storage conditions and time to analysis may not be particularly critical. However, the peptide hormones are less robust and often post-collection sample handling is critical in insuring a representative sample is available for analysis.

Generally speaking, the lipophilic hormones are well preserved across the species, while this is less often the case with the peptide hormones, especially the more complex molecules or molecules made up of mixed peptide chains (such as adrenocorticotrophic hormone, thyroid-stimulating hormone).

Most commercially available hormone assays are developed for use in people or rodents. Consequently these assays can often be adapted relatively easily for use in measuring lipophilic hormones and those peptide hormones which are well preserved across species. Equally, human or rodent assays designed to measure hormones that are not well preserved across species can be imprecise and less reliable. For example, this is particularly a problem with assays for hormones such as parathyroid hormone, thyroid-stimulating hormone, adrenocorticotrophic hormone and the gonadotrophins.

Regardless of the specificity of the binding site, imprecision can occur if the assay is being asked to detect small changes in hormone concentration or changes outside the assay's optimum range. This is particularly a problem when we are trying to measure total thyroid hormone concentrations, as generally commercial assays for total T4 are designed to operate optimal for concentrations of between 200 and 550 nmol/L, whereas we are usually interested in either values between 0 and 30 nmol/L or between 40 and 100 nmol/L.

Finally the whole system is based on how much labelled hormone is displaced from a set of specific binding sites. Any situation that alters the amount of labelled hormone bound can then be misinterpreted as unlabelled hormone. Depending upon the hormone in question and the assay system used to measure it, labelled hormone can be displaced by nonspecific compounds, and the assay will report a higher than actual level of the particular hormone in question. This phenomenon has certainly occurred on occasions, especially when certain assays designed to detect human hormones have been used to measure the same hormones in other species.