In the human body, steroid hormones are involved in a variety of regulatory processes, which makes them also important diagnostic markers for a range of diseases. However, due to their high chemical similarity, they can represent a challenge for many assays – immunoassays in particular suffer from cross-reactivities. In comparison, LC-MS/MS-based assays provide high specificity in combination with the ability to determine several steroids in one run.

by Dr Marc Egelhofer

Steroids have a common distinct chemical structure – they consist of a cholesterol backbone with 3 hexane rings and a pentane ring. The hormones are synthesized in the adrenal cortex (corticosteroids) as well as in the reproductive organs (androgens, estrogens). Several doping agents are also artificial derivatives of the male sexual hormone testosterone, called anabolics, and are used abusively to increase muscle and bone synthesis.

With a distinguished role in regulatory processes of the human body, dysfunctional steroid release can be responsible for many diseases with sometimes extremely unspecific symptoms (see Table 1). One example is aldosteronism, where the adrenal glands produce excessive amounts of the steroid hormone aldosterone. This leads to lowered levels of potassium in the blood (hypokalemia) and an increased excretion of hydrogen ions (alkalosis). Patients suffer from muscle spasms, fatigue, headaches, high blood pressure, and muscle weakness. However, these symptoms can be attributed to many diseases, and only the clinical evaluation of aldosterone plasma levels can ensure a correct diagnosis. 

Challenging targets
The chemical similarity of the steroid structure can be a challenge, in particular in a clinical setting where requirements in specificity and selectivity need to be met. This problem becomes evident when looking at epidemiological studies of major diseases, where many different assay methods with a varying performance are used, resulting in an inability to compare data [1]. The discrepancies in assay performance also limit investigations where comparisons of absolute steroid concentration values are used, rather than relative levels.  For example, absolute steroid hormone concentrations are needed when analysing effects of hormonal threshold concentrations to obtain a certain disease outcome – or not.

Steroid profiling
A lot of the published literature and of our knowledge about the physiology of steroid hormones is based on radioimmunoassays (RIA). One of the reasons for discrepancies in values, however, is that immunoassays suffer from various interferences due to antibody cross-reactions with other steroid hormones. In contrast, mass spectrometry has been recognized as the best available method for the accurate analysis of steroids in biological samples [2]. It overcomes limitations of immunoassays, while also simplifying the sample preparation in comparison to GC-MS/MS analysis that requires lengthy derivatization processes to obtain the analytes in the gaseous phase for separation.

We have developed a CE-IVD assay for mass spectrometry (MassChrom Steroids) for the determination of 15 steroid hormones. The subsequent analysis takes place in multi reaction monitoring mode (MRM). In this mode, the first and second mass spectrometers are set to a fixed certain mass. MS1 selects only the molecular ion, and ions with a different mass are disregarded. The molecular ion then fragments in the collision cell and MS2 detects the characteristic fragment. The MRM mode makes it possible to determine several steroids in a single run, thereby reducing the time for analysis and increasing the effectiveness of the method. The 15 hormones that can be analysed with this method are divided into two panels for a clear separation of each of the analytes (see Figure 1).

The chromatographic setup, including the analytical column, is identical for all analytes, thereby eliminating the need to change columns or mobile phases between separate runs. Depending on requirements and throughput, sample preparation can be performed in 96 SPE well plates or SPE columns.  The assay has been tested on a range of systems, such as the AB Sciex Triple Quad 4500 or the Waters Xevo TQS instruments.

Salivary sampling 
Plasma sampling can represent a problem, in particular for parameters that need to be collected several times a day or under stress-free conditions. Saliva consists of 99.5% water, electrolytes, mucus, white blood cells, epithelial cells, glycoproteins and enzymes, though saliva is also a carrier of steroid hormones.  The speed at which they are transferred from blood into saliva is controlled by passage through the lipophilic layers of the capillaries and glandular epithelial cells. Consequently, the more lipophilic the molecules the faster is the transfer through these barriers. Salivary concentrations are therefore dependent on the lipophilic properties of the molecule ­­— lipid-soluble steroids such as cortisol have higher concentrations, whereas more hydrophilic substances such as dehydroepiandrosterone-sulfate (DHEA-S) have much lower concentrations relative to the free plasma levels [3].

One of the common medical indications of cortisol testing in saliva is the screening for Cushing’s syndrome, a pathological increase of cortisol [4]. This hypercortisolism can be due to an endogenous overproduction or based on the intake of exogenous glucocorticoids. Symptoms may include obesity, hypertension, hyperglycemia, muscle weakness and osteoporosis. However, these symptoms are also not specific – the majority of individuals with some or all of the symptoms will not suffer from Cushing’s syndrome, therefore, the analysis of cortisol plays a significant role in the identification of the disease.

Cortisol levels do vary significantly over the course of the day (see Figure 2), making it a requirement to measure several times a day. Salivary sampling represents a simple, non-invasive and, for the patient, stress-free sampling method [5]. After a short introduction, patients can collect their sample by themselves at home, which results in a simple process to obtain samples at different stages of the circadian cycle.

The non-invasive nature of the collection procedure also enables samples to be obtained from patients afraid of venipuncture without provoking an unwanted adrenal stress response, especially in children and phobic patients. A disturbing influence of stress-induced adrenal activity is less likely, making salivary sampling more reliable compared with serum, in particular in stress research and pediatric applications [3].

We have developed a CE-IVD method for the determination of cortisol and cortisone in saliva with a sample prep procedure that is performed by filtration and in just a few steps (see Table 2).

The use of stable isotopically labelled internal standards for both analytes ensures reproducible and reliable quantification of the parameters. The performance data are 96-105% for the recovery of spiked samples, an intraassay variation of CV = 2-5%, and interassay variation of CV = 2-7 %, and the lower limit of quantification is 0.27 µg/l (see Figure 3).

Conclusions
Immunoassays are widely used for the measurement of steroids, though it is accepted that these methods suffer from various interferences due to antibody cross-reactions with other steroid hormones. In contrast, LC-MS/MS has been recognized as the best available method for the accurate analysis of steroids in biological samples. LC-MS/MS overcomes many limitations of immunoassays, enhances diagnostic utility of the testing, and expands diagnostic capabilities in endocrinology. In addition to the superior quality of the measurements, LC-MS/MS can help in the standardization and harmonization of steroid testing among clinical laboratories. Commercial suppliers offer complete solutions from sample to result that allow the determination of steroids with LC-MS/MS as the gold standard and without the need to go through the development of an in-house method.

References
1. Stanczyk F. et al. Standardization of Steroid Hormone Assays: Why, How and When? Cancer Epidemiol. Biomarkers Prev. 2017; 16(9): 1713-1719.
2. Rosner W. et al. Position statement: Utility, limitations, and pitfalls in measuring testosterone: an Endocrine Society position statement. J Clin Endocrinol Metab 2017; 92(2): 405-13.
3. Gröschl M. Current Status of Salivary Hormone Analysis Clin. Chem. 2008; 1759 54(11): 1759-69.
4. Nieman L.K. et al. The diagnosis of Cushing’s syndrome: an Endocrine Society Clinical Practice Guideline. J Clin Endocrinol Metab. 2008; 93(5):1526-40.
5. De Palo EF et al. Human saliva cortisone and cortisol simultaneous analysis using reverse phase HPLC technique. Clin Chim Acta. 2009; 405(1-2): 60-5.

The author
Marc Egelhofer PhD*
Chromsystems Instruments & Chemicals GmbH, Am Haag 12, 82166 Gräfelfing, Germany

*Corresponding author,
egelhofer@chromsystems.de

Point-of-care glucose meters are used in a variety of settings to monitor glucose concentration in whole blood. Comparability between the results from these meters and results issued on plasma samples was examined by the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC), which in 2006 recommended that all glucose results should be reported as a plasma concentration. The group advised that a conversion factor of 1.11 be used to convert whole blood results to plasma equivalence. As neonatal hematocrit differs from that seen in adults, the IFCC recommendation is not appropriate in neonatal samples. It was decided to review this recommendation.

by Mary Stapleton and Ruth O’Kelly

Introduction
Neonates may be at risk of hypoglycemia in the first few hours and days after birth, the cause of which may be attributed to the stress of extra-uterine life [1]. However, it may also signal an underlying pathology, and prolonged episodes of hypoglycemia have been described as a cause of neurodevelopmental morbidity [2]. Identification of hypoglycemic episodes is, therefore, considered to be vital in the neonatal period, but the population in question often includes extremely premature and small infants. By regularly using point-of-care (POC) devices to measure glucose in this cohort of patients, it is hoped to obtain useful results while avoiding unnecessary blood loss.

In instances where glucose results obtained on POC devices do not fit the clinical picture, a fluoride-preserved sample may be sent for plasma analysis.

Discrepancies between POC whole blood and laboratory plasma results may be a cause of lack of confidence in bedside technology. There are several causes of such discrepancies, and while literature has suggested that hypoglycemia is missed by using POC devices, the role of glycolysis as a pre-analytical factor is starting to be recognized [3]. The second possible cause is that differing sample types are measured and unlikely to be comparable. In 2006, the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) published a recommendation that manufacturers of POC devices were to report glucose concentration as though it were a plasma sample rather than whole blood. A conversion factor of 1.11 was calculated to equate the results from the two sample types (whole blood × 1.11 = plasma) [4].

The aim of this study was to perform glucose measurements in neonatal and adult whole blood and plasma samples by a laboratory method and a POC method without a plasma correction factor. By comparing results, it was hoped to determine the appropriateness of the plasma conversion factor as recommended by the IFCC.

Methods
The HemoCue 201+ POC methodology that was used to analyse whole blood samples consists of an analyser and measuring cuvette containing dried reagents. The cuvette serves as a pipette, reaction chamber and measuring vessel. Analysis of plasma for glucose concentration was performed on an automated chemistry platform (Beckman Coulter AU640) using a hexokinase method in a laboratory accredited to ISO 15189 standards.

Samples for plasma glucose analysis were obtained in tubes containing fluoride as an antiglycolytic agent. When measuring glucose in the POC device, an aliquot of sample was taken from the sample in the blood tube before separation.

Statistical analysis, using Bland–Altman analysis to compare results by two different methods, was performed using Analyse-It software for Microsoft Excel (Analyse-It Software Ltd).

Study 1
Fluoride-stabilized plasma samples from 25 neonates (aged 3 days or less) received into the laboratory for routine glucose estimation were included in the study. An aliquot was taken from each sample before centrifugation and analysis, and glucose determination by POC was performed on a HemoCue 201+ analyser located in the laboratory.

Study 2
Fluoride plasma samples from pregnant women (n = 34) were also analysed for whole blood and plasma glucose in the same manner described in study 1.

Study 3
A portion of patients who were a part of the study had a sample sent for full blood count (FBC) analysis on the same day of the glucose request. Results were subdivided into greater and less than the median result for both hematocrit and mean corpuscular volume (MCV). These were then reviewed against the reported glucose concentrations.

Results
Studies 1 and 2
No significant difference was noted between neonatal samples analysed (Table 1, Fig. 1) (bias, 0.05mmol/L). However, a significant difference (P<0.0001) was noted between the two methods when samples had been obtained from adult patients (Table 2, Fig. 2) (bias, 0.6mmol/L).

Study 3
A standard calculation for determining the percentage of water in blood was reviewed (Equation 1). The data obtained from the FBC samples was used to propose plasma conversion factors for both adult and neonatal patients (Table 3). It was assumed that the median hematocrit in a healthy, non-pregnant adult is 0.43 L/L, with a resulting calculated conversion factor (CCF) of 1.11.

Discussion
This study investigated the reported difference between samples analysed for glucose using POC meters in a ward setting and those samples received for glucose analysis in a central laboratory. It may be seen that there is good correlation between POC and laboratory analyser methods in samples obtained from neonates.

This correlation was not seen in the set of adult samples analysed, and an average difference of up to 10% in results was reported from the two methods. By applying a plasma equivalence factor of 1.11 to the whole blood results from adults as recommended by the IFCC in 2006, the difference in results from adult patients could be explained.

The IFCC equivalence factor based on the hematocrit in neonates is 1.15, but this study confirms that the neonatal samples did not require this factor. POC glucose measurements in the HemoCue device include a cell lysis step and thus whole blood (intra-and extra-cellular) glucose is measured. However, neonatal blood is recognized as containing resistant cells and cells may not fully lyse causing the measured glucose to reflect extra-cellular glucose similar to plasma measurements.

In a previous study [5], Vadasdi and Jacobs compared heparinized samples from neonates that were analysed on the HemoCue immediately before centrifugation and assayed by the laboratory method. No significant difference was found between the mean values of the two methods over a hematocrit range of 0.185–0.72. Our study agrees with these findings.

Vadasdi and Jacobs suggested that the effect of hematocrit was decreased significantly by the hemolysis step in the cuvette. It is recognized that HemoCue POC meters are not affected by hematocrit [4, 5], which is why this meter is frequently used in a neonatal setting. Vadasdi and Jacobs also suggested that because the MCV (which describes the size of the red cells) is greater than seen in adults, there is less of a dilutional effect due to membrane proteins after lysis. Our study showed that the mean MCV in neonates was greater than seen in our adult (pregnant) subjects.

Conclusion
Laboratory measurements for glucose are usually performed on plasma samples while POC measurements are performed on whole blood. A difference in results may be expected as whole blood glucose is known to be approximately 11% lower than plasma glucose due to lower volume of water in the erythrocytes.

The difference between plasma and whole blood glucose in adults was similar to the recommended IFCC “plasma equivalent factor” of 1.11. The lack of difference between plasma and whole blood glucose in neonatal samples may be explained by the increased MCV or the presence of resistant red cells that may not undergo lysis in the POC device.

Many modern POC devices for measuring glucose now include the IFCC plasma conversion factor and such results should be carefully interpreted.

References
1. World Health Organization. Hypoglycaemia of the newborn. Review of the literature. WHO/CHD/97.1, 1997.
2. Lucas A, Morley R, Cole TJ. Adverse neurodevelopmental outcome of moderate neonatal hypoglycaemia. BMJ 1988; 297(6659): 1304–1308.
3. Stapleton M, Daly N, O’Kelly R, Turner MJ. Time and temperature affect glycolysis in blood samples regardless of fluoride- based preservatives: a potential underestimation of diabetes. Ann Clin Biochem 2017; 54: 671–676.
4. D’Orazio P, Burnett RW, Fogh-Anderson N, Jacobs E, Kuwa K, Külpmann WR, Larsson L, Lewenstam A, Maas AH, et al. Approved IFCC recommendation on reporting results for blood glucose: International Federation of Clinical Chemistry and Laboratory Medicine Scientific Division, Working Group on Selective Electrodes and Point of Care Testing (IFCC-SD-WG-SEPOCT). Clin Chem Lab Med 2006; 44: 1486–1490.
5. Vadasdi E, Jacobs E. HemoCue β-glucose photometer evaluated for use in a neonatal intensive care unit. Clin Chem 1993; 39(11): 2329–2332.

The authors
Mary Stapleton* FRCPath; Ruth O’Kelly FRCPath
Biochemistry Department, Coombe Women & Infants University Hospital, Dublin, Ireland

*Corresponding author
E-mail: mary.stapleton@nhs.net