Evaluation of the Urinary Free Cortisol (UFC) Test to Measure Cortisol

The Urinary Free Cortisol (UFC) test is used to measure the total amount of cortisol excreted into urine in a 24 hour period. The Queen Elizabeth Hospital Birmingham currently measures UFC using a Waters tandem mass spectrometer. This project will investigate moving service of the UFC test from the Waters tandem mass spectrometer to a newer AbSciex 6500 mass spectrometer. To transfer the test a series of investigations need to be carried out to ensure that the analyser performs adequately. It is hoped that transferring the UFC test to the AbSciex 6500 tandem mass spectrometer will improve the ability to detect UFC of lower values. Additionally, the drugs prednisolone and hydrocortisone are known to interfere with UFC measurements. The effect these drugs have on the measurement of UFC will be investigated.

Urinary Free Cortisol (UFC) assay is used to measure the levels of the hormone cortisol in a 24-hour urine sample of a patient. The Queen Elizabeth Hospital Birmingham currently measures UFC using a in house solvent extraction method on the Waters Alliance 2690 HPLC module coupled to a Waters Quattro premier XE API mass spectrometer. The project’s aim is to evaluate the UFC assay using the ABSciex 6500 LC-MS/MS system (Sciex) to determine if the performance is acceptable to transfer analysis of UFC to the Sciex analyser. This will be achieved by performing method validation to investigate the linearity, bias, recovery, precision of the assay. Ion suppression and sigma metrics will also be evaluated. The reasoning behind this is due to the age of the Waters analyser, the analyser is no longer reliable and prone to break down. An advantage the Sciex analyser has over the Waters is, due to having three quadrupoles it is more sensitive and specific, resulting in lower detection and quantitation. Both prednisone and dexamethasone can interfere with the results of a UFC test by suppressing the levels of cortisol in a patient's blood, creating a false positive result for Addison's Disease (Kushnir, 2003). The secondary aim of the project is to investigate interference from these drugs.

 Urinary Free Cortisol (UFC) test is used to measure the levels of the hormone cortisol in a 24-hour urine sample of a patient. UFC tests are ordered when a patient reports symptoms that indicate either excessively high or abnormally low levels of cortisol in their blood, or acute adrenal failure (Young et al. 2017).

 There are numerous serious chronic illnesses that can cause abnormal cortisol levels within the patient's blood, or adrenal failure. One of these illnesses is Cushing's Syndrome (CS), which results when a patient has overly high levels of cortisol in their blood for long periods of time. In some cases, the underlying cause is steroidal medications that are being taken by the patient, and in other cases, it is simply that the adrenal glands located on top of the patient's kidneys are malfunctioning and producing too much cortisol. Additionally, Addison disease may cause insufficient production of cortisol.

 According to Moloney, Mercado, Ludlam et al. (2016) the UFC test is the most common screening method for the diagnosis of CS because UFC levels are not affected by circadian rhythm. The levels of UFC recorded in a patient is related to the levels of free cortisol in the plasma over the given time period. Issues associated with obtaining UFC levels are mostly pre analytical. These issues include impaired renal function which lowers the UFC levels (Rosmalen et al, 2014). Additionally, UFC levels do not reflect plasma levels in patients who suffer from alcohol abuse, polycystic ovary syndrome and those on Carbamazepine (Petersenn et al, 2014).

Prednisolone and its metabolites may interfere with the measurement of cortisol by LC-MS/MS (Kushnir, 2003). Both prednisone and dexamethasone are corticosteroid drugs that are used to suppress the body's natural immune response, and are commonly used to treat asthma, bronchitis, colitis, skin conditions such as psoriasis. They are also given to organ transplant recipients to help ensure that their body does not reject the transplanted organ. While prednisone and dexamethasone help numerous patients, they can interfere with the results of a UFC test, since they also suppress the levels of cortisol in a patient's blood, creating a false positive result for Addison's Disease (Kushnir, 2003). Therefore, interference from these drugs will be investigated in the evaluation of the UFC method.

  At the Queen Elizabeth Hospital Birmingham, UFC analysis is performed using an in house method on the Waters Alliance 2690 HPLC module coupled to a Waters Quattro premier XE API mass spectrometer (Waters) using an electrospray ionization (ESI) probe. UFC is extracted from 24-hour urine samples using Dichloromethane. The extract is analysed and Cortisol levels quantified using liquid chromatography-tandem mass spectrometry (LC-MS/MS). The Waters LC-MS/MS are atmospheric pressure ionization (API) tandem mass spectrometers that use liquid chromatography to separate compounds of interest and a mass spectrometer to detect the compounds eluting from the column.

The LC module houses a sample probe assembly and reverse phase HPLC (high performance liquid chromatography) column with a non-polar stationary phase. The column’s function is to separate molecules based on their polarity. Various concentrations of solvents are washed through the column and molecules are eluted when the polarity matches that of the solvent. The Waters mass spectrometer uses the multiple reaction monitoring (MRM) function to measure cortisol. Tandem mass spectrometry means there are two quadrupoles and is referred to as MS/MS. The first quadrupole has a small alternating electrical field which corresponds to the trajectory of native cortisol. Only ionized compounds with the required mass divided by charge number (m/zs) will pass through the first quadrupole. The second quadrupole focuses on a specific fragment of cortisol. The ions strike a photomultiplier tube which amplifies the signal to give a chromatographic readout of time versus signal intensity. The MRM function allows more than one ion to be monitored over a time period.

The UFC assay will be transferred to the Shimadzu LC-20ADXR solvent delivery module attached to AB Sciex 6500 LC-MS/MS system (Sciex). The reasoning behind this is due to the age of the Waters analyser, the analyser is no longer reliable and prone to break down. An advantage the Sciex analyser has over the Waters is, due to having three quadrupoles it is more sensitive and specific, resulting in lower detection and quantitation limits (Sciex.com, 2019).

The project’s aim is to evaluate the UFC assay using the Sciex 6500. This will be achieved by performing the necessary method validation steps discussed in the method section. Method validation assesses a measurement’s procedure to determine if the performance is acceptable for purpose. Validation should be undertaken for non-standard methods and laboratory developed methods such as the one being evaluated for this project.

No cross-discipline work is involved in this project.

Null hypothesis:

The performance of the Urine Free Cortisol assay is unacceptable when performed on the ABSciex 6500 LC-MS/MS system. Comparisons to the Waters Quattro premier XE API mass spectrometer show the assay is less sensitive and precise when performed on the ABSciex 6500 LC-MS/MS system. Additionally, prednisolone and hydrocortisone do not interfere with the Urine Free Cortisol assay when performed on the ABSciex 6500 LC-MS/MS system.

The project’s primary aim is to evaluate the UFC assay using the ABSciex 6500 LC-MS/MS system (Sciex) to determine if the performance is acceptable to transfer analysis of UFC to the Sciex analyser.

The Secondary aim of the project is to investigate interferences from exogenous steroids such as dexamethasone and prednisolone.

The verification experiments described below are based on recommendations taken from The Association for Clinical Biochemistry and Laboratory Medicine (ACB). (Acb.org.uk, 2019).

Imprecision

Precision evaluates the analyser’s ability to reproduce results. Quality control (QC) materials will be used to assess precision. The controls used will be Randox Assayed Urine Quality Control (QC) level 2 and 3. An addition level 1 control will be produced by diluting the QC2 material times 2 in 0.1 mol/L hydrochloric acid. Each QC will be analysed over 10 days, with 5 replicates per day, giving a total of 50 data points per sample. SPSS statistical software using analysis of variance (ANOVA) will be used to calculate within run, between run and total imprecision. The acceptance criteria will be a %CV of less than 15% (rcpaqap.com, 2019).

Sample size for precision can be calculated using within subject (Sw) deviation. Where n represents the number of subjects and m the number of observations. Using the following formula (Www-users.york.ac.uk, 2019):

m=1+1.962%CV2 X 2 X n

%CV is 15 and number of subjects is 3 (QC’s) we can calculate that a minimum of 29 repeat measurements per sample should be used.

Bias

 Bias is the means by which the accuracy of the experiment is measured. The bias will be assessed in two ways:

1. Comparison to patients’ results produced from the Waters LC/MS/MS analyser.
2. Comparison to assigned targets from EQA data.

1. For the comparison to the Waters analyser paired patient samples will be analysed on both analysers in duplicate. For comparison purposes the mean of the duplicates will be used. The results will be processed using passing-Bablok regression for linearity and Bland-Altman for bias. To obtain a suitable confidence interval (CI) using 95% limits of agreement, a minimum of 50 patient samples will be analysed. This will give a maximum CI of ±0.48 standard deviation. For analysis using patients samples the UFC values will span the full measuring range.

2. During the given time scale of the project, it is estimated that a minimum of 10 EQA samples will be available to calculate the bias from. Bland-Altman and Passing-Bablok tools will be used to assess the data. The EQA scheme’s allowable bias (B-score target) will be used as the assay’s acceptance criteria.

Sigma Metrics

Sigma metrics provides a way to assess a method or instruments performance on a universal scale. The sigma metric score shows the number of standard deviations that can fit within the defined tolerance.

The sigma metric score will be calculated using the following formula (Westgard, Bayat and Westgard, 2018):

Sigma Metric Score= Total Allowable Error (%) – Bias (%)Total Imprecision (CV%)

The higher the Sigma Metric Score the better. The acceptance criteria will be set as a Sigma Score of ≥3.0 as this is the typical value used in industry to demonstrate excellence (Westgard, Bayat and Westgard, 2018).

Assay Limits

The limit of detection is the lowest value UFC can be detected, this is defined as a signal to noise ratio of >10. The lowest concentration UFC can be measured with acceptable precision shows the lower limit of quantification. Acceptable precision is usually defined as a CV of <20%. Assay limit calculations will be based on the CLSI guideline EP17 (Pierson-Perry, Vaks and Durham, 2012). The following will be calculated:

Limit of the Blank (LoB): A blank will be obtained and measured 20 times and the mean and SD will be calculated.

Limit of Detection (LoD): A low concentration sample will be used and measured 20 times and the mean and SDs calculated.

Limit of Quantitation (LoQ): measure each of the low-level samples 5 times and calculate the CVs.

Linearity

Linearity is assessed to show the ability of the assay to return UFC values that are directly proportional to the concentration. To assess this a patient sample with a high concentration of UFC will be chosen. A second patient sample with a low concentration of UFC will also be used. The table below shows a 5-sample dilution series that will be created using the high and low patient samples.

Each sample in the dilution step will be analysed in triplicate.

Expected and measured UFC results will be compared against each other on a graph. For a full linearity validation experiment, statistical analysis is not needed. The data should be visually assessed to determine the linear portion of the assay to ensure the graph does not tail off at either the low or high end.

Stability

For the assay to be practical in a routine laboratory it must be stable, ideally any delays in analysis due to hardware issues should not affect the results produced. The stability of the prepared UFC extracts will be examined over a period of 3 days by measuring them each day in duplicate. The results will be compared using a t-test, with the resulting p-values indicating whether there is a statistical difference from the baseline results.

Recovery

Recovery experiments are used to give an estimate of proportional systematic error.

Three pairs of test samples will be prepared. The paired samples will then be spiked with a set volume of either spike material or a blank. All the samples are then analysed in duplicate and the recovery calculated using the following equation:

Recovery (%)= ConcentrationSpike- ConcentrationDilutionAmount of analyte added x 100

Ideal recovery is 100%, and any deviation from this is the proportional systematic error:

Proportional Error=100%-Observed Recovery (%)

In this instance, an acceptable recovery rate would be between 80 and 120% on three different pools of patients,

Analyte Specificity (Interference)

 The project will investigate interference from the corticosteroid medications, prednisone and dexamethasone. To assess interference a series containing samples of varying concentrations of prednisone and dexamethasone will be produced, along with a suitable blank matrix containing none of the interfering substance. The chromatograms generated will be examined for elution of peaks close to the UFC retention time. Additionally, the data will be statistically compared for each concentration of interfering substance to the blank-spiked (control) sample using t-tests, with the resulting p-values indicating whether there is a statistical difference from the blank (control) results.

Ion Suppression

Ion suppression will be assessed by direct infusion of cortisol into the LC-MS/MS source at a constant rate, whilst also running a blank sample that contains the internal standard through the HPLC system. Each sample will be analysed and the analyte ion count obtained. The matrix effect and process efficiency are then calculated using the following equations (Furey et al., 2013):

Matrix effect %= Sample 2-Sample 3Sample 3 x 100

Process efficiency %= Sample 1Sample 2 x 100

This will determine whether the assay has any ion suppression issues.

References

• Acb.org.uk. (2019). Measurement Verification Terms. [online] Available at: http://www.acb.org.uk/whatwedo/science/best_practice/measurement_verification.aspx [Accessed 15 Dec. 2019].
• Furey, A., Moriarty, M., Bane, V., Kinsella, B. and Lehane, M. (2013). Ion suppression; A critical review on causes, evaluation, prevention and applications. Talanta, 115, pp.104-122.
• Kushnir, M. (2003). Liquid Chromatography-Tandem Mass Spectrometry Analysis of Urinary Free Cortisol. Clinical Chemistry, 49(6), pp.965-967.
• Moloney, K.J., Mercado, J.U., Ludlam, W.H. and Broyles, F.E., 2016. Diagnosis of Cushing's disease in a patient with consistently normal urinary free cortisol levels: a case report. Clinical case reports, 4(12), p.1181.
• Petersenn, S., Newell‐Price, J., Findling, J.W., Gu, F., Maldonado, M., Sen, K., Salgado, L.R., Colao, A. and Biller, B.M.K., 2014. High variability in baseline urinary free cortisol values in patients with Cushing's disease.Clinical endocrinology, 80(2), pp.261-269.
• Pierson-Perry, J., Vaks, J. and Durham, A. (2012). Evaluation of detection capability for clinical laboratory measurement procedures; approved guideline - second edition. Wayne, Pa., U.S.A: CLSI.
• Sciex.com. (2019). Triple Quad Advantages. [online] Available at: https://sciex.com/x51131 [Accessed 15 Dec. 2019].
• rcpaqap.com. (2019). RCPAQAP Allowable limits of performance. [online] Available at: http://www.rcpaqap.com.au/docs/ 2014/ chempath/ALP.pdf [Accessed 15 Dec. 2019].
• Rosmalen, J.G., Kema, I.P., Wüst, S., van der Ley, C., Visser, S.T., Snieder, H. and Bakker, S.J., 2014. 24h urinary free cortisol in large-scale epidemiological studies: Short-term and
• Westgard, S., Bayat, H. and Westgard, J. (2018). Analytical Sigma metrics: A review of Six Sigma implementation tools for medical laboratories. Biochemia Medica, 28(2).
• long-term stability and sources of variability. Psychoneuroendocrinology, 47, pp.10-16.
• Www-users.york.ac.uk. (2019). How can I decide the sample size for a repeatability study?. [online] Available at: http://www-users.york.ac.uk/~mb55/meas/sizerep.htm [Accessed 15 Dec. 2019].
• Young, J., Hatipogulu, B., Molitch, M.E., Bertagna, X., Barbier, N., Sauter, N., Biller, B.M. and Pivonello, R., 2017, May. Osilodrostat maintains normalized urinary free cortisol levels in a majority of patients with Cushing. In 19th European Congress of Endocrinology (Vol. 49). BioScientifica.