Determination of 27 Glucocorticoids in Urine by Ultra-High-Performance Liquid Chromatography-Tandem Mass Spectrometry Using UniSpray^TM Source

Round 1

Reviewer 1 Report

The manuscript proposes an UHPLC method coupled to tandem MS/MS for simultaneous determination of 27 glucocorticosteroids (GCs). An extensive validation has been carried out. The method is useful for routine analysis of GCs in different situations. However, before acceptation, minor revision is recommended.

Specific comments:

Section 3.1. More detailed information than Table 2 about method optimization is necessary. For example, how injections were carry out (individually, mixtures of GCs, etc), conditions (if different gradients were assayed), etc.

Table 3. Please indicate slope ± s and intercept ± s, at least for a selection of compounds. R² values provide limited information, especially without graphical support.

Figure 4. Impossible to read the information given by this Figure.

Section 3.2.3. The information given about the robustness study is very limited. A table including the results with the tested variables should be provided.

The manuscript includes very large Tables. Some of them could appear as supporting material.

Author Response

We appreciate for review, our response is attached.  

Author Response File: Author Response.docx

Reviewer 2 Report

The introduction should be rewritten, and more valuable sources of information should be used. For example, if the biological activity of a GCs is described, the relevant reference should be used, not the analytical paper (l.n. 33 [2]).

[3, 5, 12] references are not available in English.

Reference [14] is not relevant to the aims of the present work.

A significant part of the introduction is devoted to WADA restrictions and information on GCs therapy. However, in the reviewer’s opinion, more attention should be paid to the analytical methods available for the quantitative analysis of GCs in biological samples. Indeed, the summary of available methods for identifying and analyzing GCs in different matrices and method validation parameters needs to be included.

The abbreviations GC and GCs should be unified in the text.

l.n. 31- 33: Statements such as “Like narcotics, they can be addictive if taken at increasing doses to achieve greater analgesic effects….” must be provided with reference. In reference [2], which the author referred to, there is not a single word about addiction.

l.n. 30: Reference [1] – reviewer believes that an alternative source of literature would be helpful.

l.n. 37: “They also affect immune response” -> “They also affect the immune response”.

l.n. 37-38: “…water electrolyte homeostasis” -> “water-electrolyte homeostasis”.

l.n. 38: “…blood pressure regulation and bone metabolism” -> “blood pressure regulation, and bone metabolism”.

l.n. 150: What is “…oraz…”?

l.n. 158: “…The basic reagents...” -> “the common reagents”.

l.n. 180: “… the material was shaken and frozen (until the water layer froze completely)…” How did you determine that the water layer is completely frozen? How long did it take? Conditions need to be described.

l.n. 186: “…using high-performance UPLC (Waters, 187 Milford, MA, USA) liquid chromatography…” -> “…using ultra-high-performance liquid chromatography system UPLC (Waters, 187 Milford, MA, USA)…”.

Table 2: Why are the formulae and exact mass missing for internal STDs?

l.n. 236: “3.2.1. Linearity, LOD, LoQ, RRT” -> “3.2.1. Linearity, LOD, LOQ, RRT”

l.n. 238: “The method’s linearity was determined by analysing four independent measurements for each analyte”. Please explain what means four independent measurements for each analyte. Do you mean four repeatable injections of one sample? It is not clear.

l.n. 292: Legend for table 6 has a different font.

Figure 4: the quality should be improved; it is impossible to see reconstructed ion chromatograms and retention times. Red error lines should be deleted. Moreover, the compounds should be written in English.

Terminology Substances and Compounds should be unified in tables.

The abbreviation RSD% is unnecessarily introduced several times.

l.n. 309: “Figure 4. Sample chromatogram of blank urine (1A) and analysed glucocorticoids (1B)”. What are 1A and 1B? What do you mean by “chromatogram of…analyzed glucocorticoids…”?

l.n. 322: please use English and “…22.5 ng/mL i 30 ng/mL…”

l.n. 322: “… with eight repeated measurements for each substance…”. Did you use eight measurements for each sample? It is unclear.

l.n. 325: “…The basis for assessing accuracy was the value of standard error.” Unclear. Please add the reference to the validation guidelines used for the analytical method with requirements and limits.

Figure 3: What are biomarkers? Why are they marked in figure 3? The information about biomarkers needs to be included if necessary.

The description of how ISTDs were chosen and used needs to be included. For example, were responses for ISTDs used for calculations during method validation? It needs to be described.

 

The reviewer understood that the validation was made according to the reference [16]. However, only some required parameters for the new quantitative method were validated, and the explanation or comment needs to be included. Moreover, mismatched terminology is used for validation parameters compared to the reference [16]. Therefore, it needs to be corrected. For example, “intra-series precision”, “intra-test accuracy”, “short-term precision”, “intra-day accuracy”, and “intra-day precision”. In addition, the reference [16] shows that precision should be expressed as a CV%.

Author Response

We appreciate for review, our response is attached. 

Author Response File: Author Response.docx

Reviewer 3 Report

The topic of the manuscript is highly interesting, but the scientific quality of the work is questionable. The organisation of the article is confusing (e.g. materials and methods in results and discussion) and several important experimental parameters are missing, so that replication of the method is not possible. Additionally, some mistakes happened in the method validation. Moreover, the authors did not comment on the applicability of their method. Are the achieved LOD and LOQ values sufficiently low for doping analysis? How is the different water content in urine samples considered? To make this paper acceptable, lots of explanations and re-writing of the article is required.

 

Introduction:

 

·         Lines 80 ff: Authors should write more precisely about the thresholds. Which concentrations of glucocorticoids (GCs) in urine are to be expected after allowed, which after prohibited use? This is of great importance because it defines which limits of detection (LOD) and quantification (LOQ) the analytical method needs to have. This information may also be provided in Table 4, together with the analytical method’s LOD and LOQ.

Materials and methods:

 

·         Line 167: When were the urine samples taken? How much of urine samples was collected? What is the difference between analyte-free and blank urine or is that the same?

·         Line 174: Which internal standards were used? Which compounds, which concentrations, where purchased?

·         Line 175: Describe the “prepared mixtures of standards”. Which compounds, which concentrations?

·         Line 176: What does “TM” in TM ß-glucuronidase mean?

·         Line 178: How was that 20% mixture of potassium carbonate and potassium bicarbonate prepared? 10 g of each plus 80 mL of water? Currently, the instructions do not allow replication.

·         Line 180: How were samples “extracted in portions”?

·         Line 183: Did you verify if all analytes dissolve into ACN/water (50/50, v/v)?

·         Line 185 ff: Which flow rate was used?

·         Line 198: Which mass spectrometer was used?

·         Line 208ff: How were the precursor – fragment ion pairs selected? How were the optimum parameters of the collision energy determined?

·         Table 2: Does it make sense to give two digits after the comma for the protonated precursor ion and for the product ions? Usually, only 1 digit after the comma is given if MRM is used as scan type. In addition, if 4 or 5 product ions are given, it would be very useful for unequivocal compound identification to provide the intensity ratios of all fragment ions compared to the quantifier fragment ion. Normally, the qualifier / quantifier ratio for the one qualifier ion is given. Here, the qualifier / quantifier ratios should be given for all qualifiers.

·         All information about how method validation was performed should be provided in materials and methods, not in the results.

 

Results:

 

The following comments are to be understood as if the information provided in the results was given in materials and methods (where it should be):

·         How were the calibration curves in urine prepared? Were analytes spiked before work-up or after work-up?

·         The linearity was tested from 0.75 ng/mL to 30 ng/mL. This range is very limited and linearity should be given in this range!

·         How can a 12-point calibration curve be established from 7 non-zero concentration levels? This looks like a 7-point calibration curve.

·         Line 243: add the unit of the specific weights.

·         Line 249: How were LOQ values determined? In that respect: the LOQ = limit of quantification cannot be lower than the lowest calibration standard used! The lowest calibration standard was 0.75 ng/mL. All values presented in Table 4 are below 0.75 ng/mL. Therefore, they do not make sense, as quantification must not be performed below the lowest calibration level. Another comment to Table 4: are these values the values in urine or in the measurement solution?

·         Table 3: This table is not necessary, as calibration curves in mass spectrometry are not stable. The correlation coefficients are irreproducible values and do not have to be presented in a table. However, I wonder why the correlation coefficients are that bad? Usually, at least 0.99 should be achieved for a calibration curve, and usually values are above 0.997.

·         What is the additional benefit of Figure 1 compared to Table 5? It would be possible to save space if all compounds of Figure 1 were included in Table 5.

·         Table 5: What is the use of one column showing the retention time of just one internal standard for all compounds?

·         Line 286: How was the extraction recovery determined? Is really the extraction recovery presented or are also matrix effects included? What are the apparent recoveries for the individual compounds?

·         Figure 4: The axes of the figures are not readable. Are these chromatograms with relative intensity versus retention time? It would be better for the reader to provide high quality chromatograms in the supplementary material than to provide low quality inserts in the main manuscript. In addition, it would be interesting to see the absolute intensity on the y-axis instead of the relative intensity.

·         Lines 319 ff: How was the accuracy determined? What does the term intra-day accuracy mean? There is an intra-day precision, but no intra-day accuracy.

·         How do authors correct for different water content in urine samples? There should be some kind of normalization, e.g. using the creatinine content in urine samples.

 

Discussion:

 

·         The first paragraph of the discussion is rather a description of results than a discussion.

·         There should be a discussion on how the obtained LOD and LOQ values compare with GC concentrations in real urine samples of people who consumed GCs. Is the method sensitive enough for doping analysis? Are there any routine methods used or is the presented method the first for determining GCs in urine in the context of doping?

Author Response

We appreciate your review, our review is attached. 

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

As was already stated in the previous reviewer’s report, the validation part needs improvement. The authors referred to [15, 16] guidelines. However, the mismatched parameters are described. For example, the precision should be expressed as CV%. The LOQ should be LLOQ, according to [16]. Moreover, according to [15], the identification criteria differed from the present work. Therefore, what the author determines according to which guidelines must be clear.

l.n. 167: “LOD was determined by calculating the signal-to-noise ratio of the analyte in four repetitions, which had to exceed 3. The results of LOD, LOQ, and linearity are summarized in Table 3” – How was LOQ determined? It must be described.

l.n. 213, Fig. 5: “Calibration curve for urine 9 for 6α-fluprednisolone”. Change to “Calibration curve for 6α-fluprednisolone in urine 9”. Same for Fig. 3-4.

Fig. 6, 7 – impossible to read. It must be improved.

Fig. 6 – what are ES positive substances? Does the author mean ions? Not clear; it must be explained.

Fig. 6 – column graph should have error bars. Where can be found results for retention time and RRT?

l.n. 175: “…acceptance criterion for RT and RRT < 2%.” According to TD2023IDCR? There is a limit of 0.5%: “If the CRC is the stable isotope-labeled Analyte, the RRT of the Analyte in the Sample shall not differ by more than ± 0.5% from that of the same Analyte in a Reference Specimen analyzed in the same analytical batch”

l.n. 194: the same number for decimal places should be given for RE (%mean and SD). Also true for %RSD and precision.

l.n. 198: “correlation coefficient of at least r ≥ 0.95”. This parameter should be then given, not R2.

l.n. 159: Carry-over and carryover should be unified. Check the validation guidelines.

l.n. 194: The LOQ cannot be lower than the lowest concentration level in the calibration curve (l.n. 165).

l.n. 249: What is the QC sample? It must be explained.

Reference [15] is unavailable. It must be refreshed.

References [1, 3, 5, 8-9] are unavailable in English. Must be found alternatives. The information is general so that it can be easily found in English.

l.n. 287: “Kim N.S et al. [17] obtained much higher LOD and LOQ values, however, they identified targeted compounds in cosmetics.” Reference [14] is not relevant to the aims of the present work. LOD or LOQ cannot be compared to a cosmetic matrix. Results must be compared to the urine matrix or the biological fluids.

 

l.n. 284: “It seems that these values were also affected by the use of the UniSpray source, which improved the method’s sensitivity.” There would be no proof for this statement unless the author compared the same method using an alternative ion source. This statement is not valid, then. The “new source UniSpray” is mentioned in the title, so it needs to be discussed more in the present work.

Author Response

We appreciate your revision. Please find enclosed our answers to your comments. 

Author Response File: Author Response.docx

Reviewer 3 Report

see attached file

Comments for author File: Comments.pdf

Author Response

We appreciate your revision. Please find enclosed our answers to your comments. 

Author Response File: Author Response.docx

Round 3

Reviewer 2 Report

l.n. 20 - "...and LOQ values (Table 4), ranging ..." - please delete "Table 4" from Abstract body.

l.n. 92 - "... that offers greater efficiency compared to ion sources" – please change to "... that offers greater ionization efficiency compared to other available ionization techniques"

l.n. 190 - "...LOQ was determined at 10 X SD/a, where ‘a’ is a slope." - In the Table the LOQ is 0,75 for each analyte. It seems not clear how could the calculation with SD/a provide the same number for each analyte. Analytes must have different slopes of calibration curves, do they?

l.n. 331 - "...method is (fit-for-purposes?). " - Why is it written with a question mark?

l.n. 335 - " It seems that these values were also affected by the use UniSpray source, which improved the method’s sensitivity..." - Improved sensitivity compare to which method? It must be explained, or an appropriate reference should be added.

l.n. 340 - "Others [20] studies were conducted on horses and dogs..." - please change to "Another study [20] was ..."

l.n. 345 - "Obtained relative standard deviation at levels below and/or equal (≤) 15 %, regardless

of the set concentration" – it seems that sentence is not completed.

l.n. 250 - "The precision assessment was specified based on standard deviation values while accuracy was based on standard error …" - in Table 4 is CV%. Please add to the text body.

The list of references is not unified. Please unify the style and add the missing information.

[15] please add the title of U.S. Patent

[18] reference is not valid, please refresh it.

 

References [1, 3, 5, 8, 9] - general information. References must be changed to English sources for readers.

Author Response

We appreciate your revision. Please find enclosed our answers to your comments. 

Author Response File: Author Response.docx

Reviewer 3 Report

The authors made the requested changes, so that the manuscript can be accepted.

Just one comment on the determination of the LOQ: Authors state that LOD and LOQ were determined in the same way, just considering different signal to noise ratios. If they do so, the LOQ values are below the lowest calibration standard. As quantification below the linear range is not allowed, authors have to give the concentration of the lowest calibration standard as the LOQ, and this is what they did. However, in this case, the LOQ is acutally the concentration of the lowest calibration standard for every compound, and not 10 X SD/a. I suggest to write in the manuscript that the LOQ = 10 X SD/a criterion was checked for every compound, but that the LOQ was actually determined by the lowest calibration standard used.

Author Response

We appreciate your revision. Please find enclosed our answers to your comments. 

Author Response File: Author Response.docx