CH3

H

OC

CH3O

H

O

CH3

H

H

O

OH

Trenbolone acetate α−Trenbolone (bovine metabolite)

Production and stability testing of incurred reference materialsfor the anabolic steroid trenbolone in bovine urine†

Saskia Sterk, Frederike van Tricht, Bich Le, Leendert van Ginkel and Rainer Stephany

National Institute of Public Health and the Environment (RIVM), European Union CommunityReference Laboratory, P.O. Box 1, NL-3720 BA Bilthoven, The Netherlands

Received 29th June 1998, Accepted 23rd October 1998

The production of stable reference materials with incurred residues of veterinary drugs is necessary for thevalidation of methods of analysis, including the determination of critical performance characteristics. A referencematerial for trenbolone in bovine urine was produced and the long-term stability was tested. From a pilot 16 weekstability study on seven batches containing different additives it was concluded that the use of preservatives doesnot improve the stability of the residue. A final batch of reference material of 800 vials each containing 5 ml ofurine with a target concentration of 5 mg l21 was prepared. The homogeneity and long-term stability of thematerial were tested. The material was found to be homogeneous. Based on the results of a 52 week stability studyit was concluded that the material is stable, using the current analytical methodology. For the development ofreference materials, highly accurate and precise analytical methods are necessary. However, the current analyticalmethodology is not suitable for full evaluation and certification. Currently, a new LC-MS method is beingdeveloped. After validation of this method, the stability and homogeneity study will be repeated.

Introduction

As one of the four European Union Community ReferenceLaboratories for residues (CRLs) it is one of our tasks to provideNational Reference Laboratories (NRLs), in co-operation withthe European Commission DG XII SMT program, with incurredmaterials to test in-house methods of analysis.1–4 In particular,for trenbolone, a veterinary growth promoter that in routineanalysis is difficult to determine, a reference material is neededto evaluate e.g., hydrolysis conditions and method performancecharacteristics.

Until now it was not possible to develop a stable referencematerial for a-trenbolone, the major metabolite, in bovine urine.The production and stability testing of such a reference materialare described.

Experimental

Collection of the sample material

At the Institute for Animal Science and Health (ID-DLO) atLelystad, The Netherlands, a calf (Zwart-bont breed), weighing140 kg, was injected intramuscularly, in the neck muscle, with140 mg of trenbolone acetate (Fig. 1) dissolved in 40 ml of 10%

Intralipid. The total dose was distributed over four injections of10 ml each.

Trenbolone acetate was fully characterised and checked forpurity before injection using GC-MS, HPLC and FTIRtechniques.

Urine was collected for 5 d, pooled, and stored at 220 °Cimmediately after collection.

Materials and method

To a 2 ml test portion of urine the internal standard b-trenbolone-d2 was added (RIVM, code H146903), the amountdepending on the mass concentration expected. The sample washydrolysed at pH 5.2 for 2 h at 37 °C with 0.050 ml of Helixpomatia juice. The deconjugated trenbolone was extracted withtert-butyl methyl ether (2 3 5 ml). The extract was dried andredissolved in 0.100 ml of methanol and 0.5 ml of water. Thissolution was purified further by solid-phase extraction (SPE) onC18 columns. The eluate was collected, dried and dissolved inHPLC mobile phase [0.120 ml of acetonitrile–water (65 + 35,v/v)]. Pharmacia HPLC equipment was used. The HPLCcolumn was a 5 mm LiChrospher RP-18 end-capped column; theflow rate was 0.8 ml min21, the detection wavelength 350 nmand the temperature 40 °C. The trenbolone metabolite a-trenbolone and b-trenbolone-d2 were collected in a singlefraction. This fraction was dried under a stream of nitrogen andderivatised using 0.050 ml of N,O-bis(trimethylsilyl)trifluoro-acetamide (BSTFA) + 1% trimethylchlorosilane (TMCS) for 1h at 60 °C. A 0.003 ml aliquot was injected (splitless mode) intoa Hewlett-Packard 5890 gas chromatograph equipped with a HP5989A mass spectrometer (type Engine). A CP-SIL 5 column(60 m 3 0.2 mm id) was used. Mass fragments monitored werem/z 342 and 344 of the mono-trimethylsilyl (TMS) derivativesof a-trenbolone and b-trenbolone-d2.

Depending on the concentration, a-trenbolone was quantifiedusing HPLC-UV and/or GC-MS. The concentration of a-trenbolone in the collected urine was determined using HPLC-UV and was found to be 243 ± 3 mg l21, n = 5, RSD =1.1%.

† Presented at the Third International Symposium on Hormone andVeterinary Drug Residue Analysis, Bruges, Belgium, June 2–5, 1998.

Fig. 1 Structures of trenbolone acetate and a-trenbolone.

Analyst, 1998, 123, 2563–2566 2563

Publ

ishe

d on

01

Janu

ary

1998

. Dow

nloa

ded

on 2

7/10

/201

4 23

:10:

24.

View Article Online / Journal Homepage / Table of Contents for this issue

Validation of the method

The method was validated on spiked samples of urine.Validation focused on determining (inter-day) repeatabilitysince this parameter is essential for determining the homo-

geneity and stability of the materials. The criteria used werethose applicable to methods to be used for official residuecontrol programmes.4

Table 1 Factors that may increase stability of lyophilised referencematerials

Factor

Decreasing the residual water content, below 5%Removing oxygen, storage under nitrogenInhibition of bacterial/enzymatic activity by addition of preservatives

such as surfactantsRemoval of catalysts, such as heavy metals by addition of chelating

agents (e.g., EDTA)Prevention of photolysisAddition of antioxidants such as ascorbic acid or sodium sulfite

Fig. 2 Results of pilot stability study. All data normalised to 220 °C values.

Table 2 Results of the short pilot stability study for trenbolone in bovineurine

PreservativeT = +4 °C,16 weeks

T = +21 °C,2 weeks

T = +37 °C,1 week

None Stable Not stable Not stable0.01% thiomersal Stable Not stable Not stable0.1% Na2EDTA Stable Not stable Not stable0.4% ascorbic acid Stable Not stable Not stable0.2% Na2SO3 Not stable Not stable Not stable0.1% Na2EDTA +

0.4% ascorbic acid Stable Not stable Not stable0.1% Na2EDTA +

0.2% Na2SO3 Not stable Not stable Not stable

2564 Analyst, 1998, 123, 2563–2566

Publ

ishe

d on

01

Janu

ary

1998

. Dow

nloa

ded

on 2

7/10

/201

4 23

:10:

24.

View Article Online

The limit of determination was below 2.5 mg l21 (GC-MS),the precision was 1.1–5.9%, the inter-day repeatability RSDwas 9%, the analytical recovery was 81% and the analysis waslinear over the range 2.5–25 mg l21 with a correlation coefficient> 0.9966. Validation was based on the analysis of approx-imately 50 samples.

Lyophilised samples from the final batch were used for amixed homogeneity test. On five days, five vials werereconstituted and homogenised. From this mixed sample, fivesample portions were taken and analysed. The mean concentra-tion of the final reference material was 4.3 ± 0.6 mg l21 (n =24), measured over 5 d with an inter-day repeatability RSD of14.1%. For the precision, the RSD ranged from 5.9 to 16.4%.

Pilot stability study

The goal of this pilot study was to determine the conditionsnecessary for the production of a stable reference material forincurred trenbolone in urine.

In this study we used the following criterion to define theterm ‘stable’: ‘no changes (meaning significant changes of themass concentration during storage) in the starting concentration

at room temperature over 2 weeks, and no changes at 37 °C over1 week’. In addition to visual inspection, first-order kinetictransformation (criterion slope of zero indication that the rate-constant for decomposition is zero) was used for evaluation ofthe data.

In Table 1 a number of factors influencing the stability aregiven.

The collected calf urine was diluted with blank calf urine toa target concentration of approximately 100 mg l21. Sevenbatches were prepared containing different preservatives.Portions of 5 ml were lyophilised in amber-coloured glass vialsand sealed under a stream of nitrogen. The samples were stored(n = 2 for each temperature/time combination) at 220 °C,4 °C, room temperature and 37 °C for t = 1, 2, 4, 8 and 16weeks.

The samples of t = 1 week were analysed using HPLC-UV.Some chromatograms showed interferences at the retentiontime of the internal standard. From week 1, 37 °C, onwards, allsamples were analysed by GC-MS.

From the results it was concluded that at +4 °C a number ofbatches are stable for 16 weeks (Table 2). In Fig. 2 the resultsare presented normalised to the value of the 220 °C point.

For preparation of the final batch of reference material it wasdecided that no preservatives were to be added. None of thepreservatives could prevent instability at high temperatures(room temperature and higher) whereas some preservativescaused interferences in HPLC-UV analysis. Consequently,storage of the material at +4 °C or lower became necessary.

Final reference material

Preparation

A final batch of reference material, 800 vials each containing 5ml of incurred urine and 800 vials each containing 5 ml of thecorresponding blank urine, was lyophilised.

The urine collected was diluted with blank bovine urine to thecalculated target concentration of approximately 5.8 mg l21.

The materials were lyophilised in amber-coloured glass vialsand sealed under nitrogen.5

For quality control, 30 units of blank and incurred materialwere weighed. The RSD in the weight of the dry residue wasdetermined to be less than 1.1%. The content of residual waterwas determined using Karl-Fisher titration and was found to be4.8 and 5.1%, respectively, for the incurred and blankmaterial.

The batches were stored at 270 °C except for the vialsundergoing the prolonged stability study.

Fig. 3 Homogeneity test on incurred reference material [n = 20, analysedon two different days (1–10 and 11–20].

Fig. 4 Results of 52 week stability study of incurred reference material. All data normalised to 270 °C values.

Analyst, 1998, 123, 2563–2566 2565

Publ

ishe

d on

01

Janu

ary

1998

. Dow

nloa

ded

on 2

7/10

/201

4 23

:10:

24.

View Article Online

Homogeneity study of incurred reference material

Twenty vials, randomly chosen, were analysed individually (tenper day) by GC-MS. The mean value ranged from 4.9 to 6.4mg l21, the standard deviation from 1.9 to 2.7 mg l21 (Fig. 3).

Prolonged stability study

Samples were stored at different temperatures (270, 220, +4,+21 and +37 °C) for a period of 2 years (1, 2, 4, 13, 25, 52 and104 weeks). Results are presented up to 52 weeks and arenormalised to the corresponding 270 °C values (Fig. 4).

From Fig. 4 it can be concluded that all values stay above thetarget value minus twice the RSD. From these data it can beconcluded that the material is stable. The results indicate that thematerial is also stable at temperatures above +4 °C, a findingthat is not in agreement with the conclusion obtained during thepilot stability study. The main difference between the prepara-tion of the final batch and the pilot batch is the lyophilisationprocess, which was optimised during the preparation of the finalbatch, resulting in a significant decreased residual watercontent.

Even though from both the homogeneity and stability study itcannot be concluded that the material is not suitable, reserva-tions with respect to the suitability of the material as a potentialcertified reference material remain.

Conclusion

For the development of reference materials (RMs), highlyaccurate, precise and robust analytical methods are necessary.

During this study, much information was gathered on thedevelopment of a homogeneous and stable RM for trenbolone inbovine urine. However, the current analytical methodology isnot suitable for full evaluation and certification. Variability dueto inhomogeneity and instability up to 20% cannot be excluded.Currently, a new LC-MS method is being developed. Aftervalidation of this method, the stability and homogeneity studywill be repeated.

Acknowledgement

The authors thank Ir. P. L. M. Berende (ID-DLO, Lelystad, TheNetherlands) for his help in collecting the materials. Thematerials were prepared, partially, within a contract with theEuropean Commission DG XII SMT/MAT (former BCR)program, and performed within the RIVM Multi Annual Planunder project number 388710.

References

1 R. J. Heitzman, Fresenius’ J. Anal. Chem., 1993, 345, 207.2 R. Walker, ISO REMCO, 1993, 282.3 EC BCR Document 48/93 of 15 December 1994.4 Council Directive 96/23/EC of 29 April 1996 on measures to monitor

certain substances and residues thereof in live animals and animalproducts and repealing directives 85/358/EEG and 86/469/EEG anddecisions 89/187/EEG and 91/664/EEG,10.

5 S. Sterk, F. van Tricht, A. van Soeren-kieft, H. Herbold, R. Stephanyand L. van Ginkel, Fresenius’ J. Anal. Chem., 1998, 360, 454.

Paper 8/04944K

2566 Analyst, 1998, 123, 2563–2566

Publ

ishe

d on

01

Janu

ary

1998

. Dow

nloa

ded

on 2

7/10

/201

4 23

:10:

24.

View Article Online