Chemical pathology case conference - common endocrine tests

Morris H L Tai 戴學良, Chloe M Mak 麥苗, Yuet-ping Yuen 袁月冰, Ching-wan Lam 林青雲, Rossa W K Chiu 趙慧君, Michael H M Chan 陳浩明, Albert Y W Chan 陳恩和, Anthony C C Shek 石志忠, Tony W L Mak 麥永禮, Wing-tat Poon 潘永達, Sidney Tam 譚志輝

HK Pract 2007;29:83-89

Summary

The endocrine system is essentially an information system of the body. The interpretation of the endocrine tests is difficult because of the complex interaction between different hormones and organs. Analysis of hormones is technically demanding because the hormone levels are usually low, and the tests are also prone to interferences. Useful results could be obtained only if the appropriate tests are chosen, the patient is properly prepared and the analysis is correctly performed. Unawareness of clinical or technical pitfalls may lead to failure to diagnose an important endocrine disease or unnecessary treatment of a condition which does not exist. In this article, the use and the pitfall of selected endocrine tests are discussed.

摘要

內分泌系統實質上是身體的資訊系統。由於不同激素和器官之間複雜的相互作用，解讀內分泌檢查結果會比較困難。 對激素進行分析在技術上要求較高，因為激素水平通常較低，檢查又易受干擾。只有檢查方法選擇得當， 病人完全做好準備，而且正確的進行了分析，才能得出有用的結果。不瞭解臨床或技術上的難點， 就可能會導致某個重要的內分泌疾病的漏診，或對一個本不存在的疾病進行不必要的治療。本文對一些內分泌檢查的應用和難點進行了討論。

Introduction

Our body's endocrine system is a complex system involved in the transmission of information from the endocrine glands to the effector organs. An example of a complex endocrine system includes releasing hormones from the hypothalamus, trophic hormones from the pituitary and hormones from peripheral glands. Levels of individual hormone are affected by negative feedback and many other factors like circadian rhythm, stress, and diet. The hormone bioactivity is also affected by protein binding as only free hormone and the portion bound to protein loosely (e.g. albumin-bound testosterone) are active in vivo. Interpretation of endocrine investigation results is challenging because of the complicated interaction between the hormones and the effector organs. Meaningful conclusion could only be reached when the appropriate tests are chosen, the patient is properly prepared, and the sample is taken at the correct timing.

What makes it more challenging is the technical requirement in hormone measurement. The quantity of hormones is low when compared with other analytes like proteins and electrolytes, and the bulk of circulating hormone is protein-bound. Immunoassay is the main method involved, although use of mass spectrometry in hormone analysis has gained popularity and the outcome is promising. Immunoassay is a headache in endocrine investigation because hormone may be present in several isoforms and metabolites, which could have altered cross-reactivities in various analysers. The immunoreactivity is not necessarily correlated with the in-vivo bioactivity of the hormones. Furthermore, a legion of interferents have the potential to invalidate the immunoassay results, examples include commonly encountered proteins like fibrinogen, and unexpected substances like human against mouse antibody, which is developed in subjects with history of contacts with animals or animal products. It is therefore common to observe discrepant results if the same specimen is analyzed by different immunoassays.¹ To improve method comparability, standardization of hormone assays is in active progress.²

In this article, the use and pitfalls of common endocrine tests will be discussed. Complicated dynamic function tests that are seldom performed in out-patient setting are not included. Thyroid function tests, which represented a major workload in the laboratory, will be discussed in details in another upcoming article.

Is a normal level really normal?

Case 1:

A 53-year old male complained of gross haematuria for three months. The presence of renal stone was confirmed by radiological investigation. Blood tests revealed hypercalcaemia and parathyroid hormone (PTH) was subsequently requested.

Serum

Question: Was the PTH level normal?

Answer: PTH level is negatively controlled by ionized calcium level. The relationship of serum calcium and PTH is steeply sigmoidal³ and a depressed PTH is thus expected even in case of mild hypercalcaemia. The PTH concentration in this context is considered to be inappropriately normal in the presence of hypercalcemia. This patient was suffering from primary hyperparathyroidism and a parathyroid adenoma was subsequently localized in the right lower pole by sestamibi scan.

Case 2:

A 36-year old female, with long-standing vitiligo, complained of decreased appetite and increased body weight in the recent year. Only free T4 (FT4) and antimicrosomal antibody (AMC) were requested.

Serum

Question: Was the patient hypothyroid?

Answer: The reference interval of a laboratory test reflects the inter-individual variation. Reference intervals may occasionally provide misleading information if inter-individual variation is significantly greater than intra-individual variation. In this situation, the hormone level may still fall within reference interval even it is grossly deviated from the normal set-point. The typical one-year inter- and intra-individual coefficients of variation of FT4 are 17.1 % and 9.2 % respectively.⁴ Therefore a mild or moderate thyroid derangement may present with a normal FT4. Thyroid stimulating hormone (TSH) is the most sensitive marker for screening primary thyroid disorders as a minor deviation of serum FT4 from its set-point will cause a significant change in TSH level. In this patient, TSH was 8 mIU/L (0.35 - 5.5) indicating mild hypothyroidism. The high AMC concentration reflected a higher risk of conversion of overt hypothyroidism and thus closer follow-up is needed.

Is spot hormone level useful?

Case 3:

A 45-year old woman presented with weight gain, coarsening facial features and hypertension.

Question: Was acromegaly excluded?

Answer: The secretion of GH, like other pituitary hormone, is pulsatile. In acromegaly, the episodic secretion is preserved despite an increase in number, duration and amplitude of pulses. Very high random GH level may be diagnostic of acromegaly, but about 10% of patients with acromegaly have random GH level falling within the reference interval.⁵ An oral glucose tolerance test is viewed as gold standard and a non-suppressible GH is diagnostic of acromegaly. Random insulin-like growth factor-1 (IGF-1) is considered as an accurate reflection of integrated GH production, and has been recommended as the initial test for suspected acromegaly.⁶

Case 4:

A 55-year old male had secondary adrenal insufficiency after pituitary surgery for acromegaly. He was put on hydrocortisone supplement. He had an assessment of adrenal function and hydrocortisone was replaced by dexamethasone 24 hours before blood taking.

Question: Could the steroid replacement be tailed down?

Answer: The secretion of ACTH and cortisol are also episodic and the circadian rhythmicity is well-known. Dynamic function tests like short synacthen stimulation test (SST) and insulin tolerance test (ITT) are considered more reliable for investigation of adrenal insufficiency than a spot level. However, the dynamic function tests are time consuming, troublesome and potentially dangerous. It has been shown that morning cortisol 550 nmol/L⁷ (or > 450 nmol/L⁸) are highly correlated with the SST findings and has been recommended as a screening test. Morning cortisol less than 100 nmol/L is highly suggestive of cortisol insufficiency whereas level greater than 550 nmol/L reflects adequate adrenal function. Levels between 100 and 550 nmol/L should be confirmed by SST or ITT. In this case the result is equivocal and dynamic function tests should be performed.

Importance of patient's preparation

Case 5:

A 25-year old male, with history of epilepsy and was put on carbamazepine, was found to be hypertensive. A lowish potassium was observed and secondary hypertension was suspected.

Serum

Question: Was the patient suffering from Cushing's syndrome?

Answer: Cushing's syndrome cannot be excluded nor diagnosed in this gentleman. The apparently non-suppressible cortisol concentration could be caused by the minimized dexamethasone effect. Dexamethsone is metabolized by microsomal enzyme in liver. Carbamazepine is an enzyme inducer that enhances the metabolism of dexamethasone and thus reduces its pharmacological effect. Therefore, patients on long-term carbamazepine treatment or any enzyme inducers may have a false positive result in overnight dexamethasome suppression test because less dexamethasone is present to exert the negative feedback on the hypothalamus-pituitary-adrenal axis. In this case, 24 hours urinary free cortisol should be used for screening.

What is being measured?

Case 6:

A 36-year old lady with history of hydatidiform mole, treated with chemotherapy, was recently transferred to Hospital B from Hospital A for management after the completion of chemotherapy.

Serum

hCG (pre-chemotherapy level performed by Hospital A)
2426 IU/L   (non pregnant < 1 IU/L)
hCG (three months after the completion of chemotherapy performed by Hospital B)
530 IU/L   (non pregnant < 1 IU/L)

Question: Was there a real regression?

Answer: There are several hCG-related molecules like intact-hCG, subunits, nicked forms, and metabolites in blood. Different assays are designed for detection of a single hCG form or multiple forms.⁹ Trophoblastic disease secretes high proportions of nicked hCG and beta-hCG subunit,¹⁰ which could not readily be detected by all assays. Although the two assays may both be valid and accurate, they are not able to provide comparable result. Interhospital or interlaboratory comparison is therefore not recommended. The chemical pathologist should be consulted for the specificity of the assay if in doubts.

Case 7:

A 34-year old woman complained of oligomenorrhoea.

Serum prolactin (PRL)

1422 IU/L   (female non pregnant < 496 IU/L)
640 IU/L   (after bromocriptine treatment)

Question: Was this a definite case of prolactinoma?

Answer: In most individuals with increased prolactin, 60-90% of prolactin is monomeric form, 10-30% is "big" PRL, and 0-10% is "big-big" PRL.¹¹ It has been shown that latter two types of prolactin may aggregate with immunoglobulin G and the renal clearance is thus prolonged. The "big-big" PRL is also known as macroprolactin and is biologically inactive in vivo. The macroprolactin has been found to be the dominant form of immunoreactive PRL in some subjects without any evident cause and features of hyperprolactinaemia.¹² Clinically the presence of macroprolactin is difficult to confirm since response to dopamine agonist may be noted as well, and radiological abnormality is probably present as pituitary incidentaloma is not uncommon. Currently, macroprolactinaemia can be detected using screening method of polyethylene glycol (PEG) pretreatment and it has become a routine practice of some laboratories to re-analyze all hyperprolactinaemic samples with PEG pretreatment. A more cost-effective approach is to request PEG pretreatment for the first sample if no specific acromegalic features (e.g. galactorrhoea) has been noted, and obvious causes like drug-induced hyperprolactinaemia has been excluded.

Case 8:

A 24-year old female complained of secondary amenorrhoea and persistent vomiting.

Serum hCG /L

656 IU   (non pregnant < 1 IU/L)
Urine pregnancy test   negative

Question: Was the patient pregnant?

Answer: Urine pregnancy test usually has a detection limit of 25 IU/L and comparable serum level could achieve a positive result if concentrated urine (i.e. morning urine) is used. The discordant result in this case could be due to the failure of the urine pregnancy test or serum hCG assay. Urine pregnancy test is considered one of the most reliable laboratory test¹³ but it is prudent to test the specimen with another lot or brand of test kit to confirm the true negativity. Phantom hCG immunoreactivity is a notorious interferent of serum hCG assay and results in a false positive result. It can be caused by trypsin-like molecules, cholera toxin, transforming growth factor- and certain bacteria.⁹ The unawareness of this condition may lead to unnecessary operation and chemotherapy if malignancy is suspected. Phantom hCG can be confirmed by distorted stoichiometric relationship upon dilution, abnormal recovery with PEG precipitation and discordant result on another analyser. Analytical interferences can be a serious pitfall and clinical suspicion should be raised when there is a significant discrepancy between clinical and biochemical results.

Conclusion

In summary, correct interpretation of endocrine investigation relies on proper selection of test and preparation of patient. A normal result does not necessarily imply absence of disease and interference must be suspected if the result is not compatible with the clinical presentation.

Key messages

1. Appreciation of normal physiology is essential in correct interpretation of endocrine tests. Like other biochemical investigations, simple comparison of the test results with reference interval may lead to incorrect diagnosis. The result must be interpreted in the clinical context and a normal hormone level may be indicative of an underlying disease.
2. Dynamic function tests are still the gold standard of most endocrine disorders. Spot sample may occasionally provide important information if appropriate test is chosen and the result is interpreted properly.
3. The effect of patient's medication on endocrine tests is usually overlooked and this may result in incorrect interpretation.
4. Endocrine tests are prone to interferences. The interferents may be exogenous or endogenous. The clinicians must be aware of their presence and to consult the laboratory if the results are not compatible with the clinical condition.