Determination of serum urea by mass fragmentography

199

Clinica Chimica Acta, 71 (1976) 199-205

@ Elsevier Scientific Publishing Company, Amsterdam -Printed

in The Netherlands

CCA 7903

DETERMINATION

INGEMAR &RKHEM

OF SERUM UREA BY MASS FRAGMENTOGRAPHY

*, ROLF BLOMSTRAND and &STA

GHMAN

Department of Clinical Chemistry at Huddinge Hospital, Karolinska Znstitutet, 141 86 Huddinge (Sweden)

(Received March 2nd, 1976)

Summary A mass fragmentographic method of high accuracy for determination of serum urea is described. A fixed amount of [“NJurea is added to a fixed amount of serum, then the urea is converted into 5,5diallyl barbituric acid by coupling with diallyl malonic acid diethyl ester. The barbiturate is then transferred from an alkaline water phase into an organic phase containing methyl iodine by ion-pair extraction using tetrabutyl ammonium as the positive counterion. The amount of urea is determined from the ratio between the recordings at m/e 236 and m/e 238 obtained after analysis with a combined gas chromatograph-mass spectrometer equipped with an MID-unit (multiple-ion detector). The two ions used correspond to the molecular peak in the mass spectrum of the methyl derivative of unlabeled and labeled 5,5diallyl barbituric acid, respectively. The relative standard deviation of the method was 3.6%. A comparison between the mass fragmentographic method and a routine method for determination of serum urea based on the urease-Berthelot reaction gave a high correlation (r = 0.99) and a regression coefficient of 0.95.

Introduction Determination of serum urea in clinical chemical routine is almost exclusively based upon the diacetyl reaction [l] or upon determination of ammonia after cleavage of urea with urease. In the latter procedure, the ammonia is determined by the Berthelot reaction [ 21, titration [ 31 or with an ammonia electrode [4]. The combined urease-titration method has been used as a reference for evaluation of other methods [ 51. The accuracy of all the different methods is not known with certainty, since no other more specific method has been available. As a part of a general plan to develop reference procedures based on * To whom correspondence

should be addressed.

200

mass fragmentography for validation of routine methods used in clinical chemistry we now describe a reference method for determination of serum urea. In this procedure, [‘5N,]urea is added to a fixed amount of serum and the mixture of labeled and unlabeled urea is converted into 5,5diallyl barbituric acid. The ratio between labeled and unlabeled molecules in the mixture is determined by means of mass fragmentography (selective-ion monitoring). In a very recent preliminary report [ 63 l “N,-labeled and unlabeled urea were determined by mass spectrometry. The ratio between labeled and unlabeled molecules was however determined in ammonia after the urease reaction. Any unspecificity of the urease would thus also affect the results obtained with this method. Materials and methods

Urea pa. with a purity claimed to be 99.5% was obtained from Merck, Darmstadt (GmbH). [‘SNz]Urea was obtained from the British Oxygen Company Limited (London, U.K.) and contained less than 1% unlabeled molecules as determined by mass spectrometry (cf. Fig. 1). The melting point of both the unlabeled and labeled urea was 134°C and was not affected by recrystallization from eth~ol/water. No imp~ities could be detected by thin-layer ~roma~graphy (silica gel G, chloroform/methanol/water, 7 : 5 : 1, v/v) in either of the two materials. Serum Serum was collected from kidney-transplanted patients in out care at 0800 a.m. None of the patients had taken drugs containing 5,5diallyl barbituric acid. The serum samples were stored for less than 6 hours at 4°C prior to analysis by the routine technique or lyophilization and treatment with diallyl malonic acid diethyl ester.

Serum urea was determined with a commercial kit from Boehringer Mannheim (GmbH) based on the urease-Berthelot reaction. The absorbance was measured with an LKB 2074 Calculating Absorptiometer. The relative standard deviation calculated from 30 duplicate determinations was 3.7%.

[‘“NJLJrea, 3.3 pmol, was added to 250 ~1 of serum and the mixture was lyophilized overnight. In preparation of samples for the standard curve, serum was substituted with water containing O-50 mmol of urea per 1. The residue was dissolved in 1 ml dry ethanol by sonication and centrifuged. The supernatant was evaporated at 80°C till complete dryness. Diallyl malonic acid diethyl ester, 20 gmol and 50 i.rf dry ethanol (distilled over sodium} containing 15 pmol sodium were added to the residue and allowed to react for 1 h at 70°C. PO4 buffer (pH lo), 0.5 ml, 0.27 mol/I containing tetrabutylammonium hydrogen sulfate, 0.1 mol/l, was then added. The mixture was shaken for 30 min with dichlorethane, 1 ml, containing 160 ,umol of methyl iodine and the two phases separated by centrifugation. The organic phase was collected and

201

dried with sodium sulfate. The solvent was reduced to give a final volume of about 50 ,ul.

under a stream of nitrogen

Mass fragmentography The organic phase, usually l-2 ~1, was analysed by gas chromatographymass spectrometry using an LKB 9000 equipped with a 3% SE-30 glass column (on Chromosorb W, 80-100 mesh, 2 mm X 2.5 m). The carrier gas was helium and the flow rate 30 ml/min. The column temperature was 150°C and the temperature of the flash heater and the ion source were about 250°C and 290°C respectively. The electron energy was set to 20 eV and the trap current to 60 ,uA. The electron-multiplier sensitivity was set to 180. The first channel of the MID-unit was focused on the ion at m/e 236 and the second at m/e 238 corresponding to the unlabeled and labeled molecular ion of the dimethyl derivative of 5,5diallyl barbituric acid, respectively. The amplification used for both channels was 30 x. The filter settings were 0.5 Hz for both channels and the measuring time was 20 ms. The MID recordings were made on UV-paper and the peak height was measured. Results The mass spectra of the dimethyl derivative of unlabeled and “N,-labeled 5,5diallyl barbituric acid are shown in Fig. 1. The fragmentation pattern was similar to that reported previously [ 71. As is evident from the mass spectra, the fragments at m/e 195 and 197, corresponding to the base peak in the unlabeled and labeled compound were less suited for mass fragmentography due to overlapping. Thus, there was a peak at m/e 195 of considerable intensity in the mass spectrum of the labeled compound. The fragments corresponding

0 H2C=CHCH H2C=CHCH

,c.H \t

N

3

2 2Q=O 0 0

N ‘CH

3

MW = 236

0 u H2C=CHCH H*C=CHCti

EN’ CH3

2 2c:,=O d’

‘k ‘CH

3

MW : 238

Fig. 1. Mass spectrum of the dimethyl derivative diallyl [1.3-“N21 barbituric acid (lower).

of unlabeled

5.5-diallyl

barbituric

acid (upper)

and 5.5

i m/e 238

A_ 1

2

Retention

3

L

mle 236

t

5

time (minutes)

--

mle 238

12345 Retention

ttme (mmutes)

Fig. 2. MID recording of the derivative of unlabeled 5,5-dial.W “Nzl barbituric acid (B). For experimental details, see Methods.

barbituric

acid (A) and 5.5-diallyl

[1.3-

to the molecular ions at m/e 236 and 238 were therefore chosen, in spite of their lower ~tensities. The MID recordings of the two ions at m/e 236 and 238 obtained after analysis of the unlabeled and labeled compound are shown in Fig. 2. The ratio between the peaks at m/e 236 and 238 obtained by MID recordings of different standard mixtures of unlabeled urea together with 3.3 pmol [ *SNz]urea was linear with the amount of unlabeled urea up to a concentration of 50 mmol/l (Fig. 3). An MID recording of a serum sample to which [“Nz]urea had been added is shown in Fig. 4. Analysis of serum samples to which no [‘5N2]urea had been added gave the same high ratio between the tracings at m/e 236 and m/e 238 as did unlabeled pure urea.

i

R

mm01 was

per

Retention

liter

Fig. 3. Standard curve for determination tails, see Methods and Results.

of serum

urea in the range O-50

Fig. 4. MID recording of the derivatives of a methylated added. For experimental details, see Methods.

extract

mmol/l.

i

m/e

236

i

m/e238

time (minutes)

For experimental

of serum to which [ * sN2]

de-

urea had been

203

/ E 3

30.

b 5

1. .:/

t a 8 ; =20 2 E

I. 10 / *. 7, L.,,..

mmol

Y=O95X-018

* *

10 urea

20 per

hter

30 serum

LO

50 MID

Fig. 5. Comparison between the MID method and the urease-Berthelot reaction method.

It should be mentioned that in the analysis of serum samples, the yield of barbiturate in the procedure, as reflected from the peak height of the recordings, varied considerably in some series of samples. The variations in the yield of barbiturate could however always be compensated for by varying the amount of sample injected and it was clearly shown that the amount had no influence on the ratio between the recordings at m/e 236 and m/e 238. After several subsequent analyses the column sometimes started to bleed out unidentified interfering compounds. This could easily be overcome by raising the column temperature to 250°C for a few minutes. The interfering compounds could be completely removed by introduction of an acid-extraction step prior to the transmethylation step. Under most conditions, however, the acid extraction step was considered to be too time-consuming to be justified. The accuracy of the method was tested by addition of unlabeled urea corresponding to 5.0 mmol/l and 10 mmol/l to the same serum sample containing 9.9 mmol/l. The results obtained were 15.1 mmol/l (error 1.5%) and 20.5 mmol/l (error 3.0%), respectively. The relative standard deviation of the method was 3.6% as calculated from duplicate analysis of 30 serum samples. These serum samples were also analyzed with the routine method. As is shown in Fig. 5, there was very good agreement between the two methods (r = 0.99) and the regression coefficient as calculated with two independent variables was 0.95. Discussion The reasons for conversion of urea into barbiturate prior to mass fragmentography in the present work were the following: (1) Urea in itself is very difficult to analyze by gas chromatography. A gas chromatographic procedure has been described with use of the trifluoroacetyl derivative of urea and a very short column of 0.3% ethylene glycol adipate[8]. In our hands this method’ gave broad tailing peaks and a very low reproducibility. Attempts to use other derivatives and other packing materials failed. (2) The specificity of a mass fragmentographic technique is dependent upon

204

the mass of the fragment chosen and in general the probability of interference increases with decreasing mass of the fragment. Since urea has a low molecular weight and since it is difficult to make derivatives, there is a high probability of interference when using urea alone. (3) A purification procedure is most probably necessary prior to the mass fragmentographic analysis. There is no easy way to separate urea from other water-soluble compounds with properties similar to urea which may interfere in the assay. The conversion of urea into barbiturate [ 93 and subsequent ionpair extraction ilO] gave an extract confining very little com~und of polarity similar to barbiturate. No interfering compounds at all could be detected in most of the MID recordings. The interfering compounds which sometimes were observed after several subsequent analyses could easily be removed by extraction with organic solvent from an acid-water phase prior to the transmethylation step. This was not considered necessary in the present work but can be recommended if the method is used on a larger scale. It should be pointed out that if the patients have a therapeutical concentration of 5,5diallyl barbituric acid in the blood, this may affect the results obtained with the MID method. The degree of disturbance should be low, but is not possible to predict in view of the varying yield in the reaction of urea with diallyl malonic acid diethyl ester. Interference of this kind can be avoided by using a derivative of mafonic acid diethyl ester which gives a barbiturate not used as a sedative. The sensitivity of the present method has not been tested in detail since the amount of serum urea was never critical. With modifications, however, the method may easily be used for determination of urea levels less than 10% of those used in the present study. The method can be used not only for determination of unlabeled urea when a high accuracy is needed, but also for determination of [“NJurea in metabolic experiments. In such experiments, the ratio between [“Nz]urea and unlabeled urea should be at least 1 : 20 or higher. It is evident from the present work that the urease-Berthelot method for dete~ination of serum urea has sufficient accuracy for most purposes. The agreement between the results obtained with this method and the mass fragmentographic method was excellent. As expected the mass fragmentographic method gave somewhat lower values, presumably due to determination of ammonia of origin other than the urease reaction.

This work was supported by the Bank of Sweden Tercentenary Fund. The skilful technical assistance of Mr. Olle Falk is gratefully acknowledged. References 1 2 3 4 5 6

Nat&son, S. (1961) Microtechniques of Clinical Chemistry, Vol. 3, 2nd edn., Thomas, Springfield Fawcett, J.K. and Scott, J.E. (1960) J. Clin. Pathol. 13, 156 Renfro, J.L. and Patel. Y. (1974) J. APPI. Phys. 37, 756 Conway, E.J. (1947) Microdiffusion and Volumetric Error, Crosby. Lockwood, London Keler-Bacoka, M., Kos, B. and Stojanovski-Bubati, A. (1972) Acta Med. Iugosl. 26,433 Schmidt, H.L., K&h. P., Traut, G. and Keller, H.E. Proc. 2nd Int. Conference on Stable Isotopes, October. 1975, in press

205 7 8 9 10

Skinner. R.F., Gallaher. E.G. and Predmore, D.B. (1973) Anal. Chem. 45. 573 Mee. J.M.L. (1974) J. Chromatogr. 93.302 Vogel, A.I. (1956) Practical Organic Chemistry, 3rd edn., Longmans. Green and Co, London Ehrsson, H. (1974) Anal. Chem. 46,922