Journal of Analytical Toxicology, Vol. 25, November/December 2001
Technical Note [
Analysisof the tidocaine Metabolite
2,6-Dimethylaniline in Bovine and Human Milk
Nell W. Puente and E David losephy*
Guelph-Waterloo Centre for Graduate Work in Chemistry (GWC2), Departmentof Chemistryand Biochemistry,
University of Guelph, Guelph, Ontario, CanadaN1G 2Wl
Abstract I
2,6-Dimethylaniline (2,6-xylidine; 2,6-DMA) is a nasal carcinogen
in rats. Humans may be exposed to this compound via several
routes: 2,6-DMA is found in cigarette smoke; it is a
pharmacologically inactive metabolite of some drugs (e.g., the
local anesthetic lidocaine) and pesticides(e.g., metalaxyl); and it is
an impurity in technical grade metalaxyl. The potential transfer of
2,6-DMA from mother to nursinginfant via milk is of toxicological
concern. Solid-phasemicroextraction with separation and
detection using gaschromatography-massspectrometry was
optimized and usedfor the analysisof 2,6-DMA in milk.
2,6-DMA-d9 was synthesizedand used for quantitation by the
isotope ratio method. At a concentration of 5 ppb 2,6-DMA, the
method detection limit was 0.20 ppb, and the relative standard
deviation was 3.6%. Samples of milk were obtained from bovines
administered lidocaine (2.9-3.9 mg/kg) during surgery. A breast
milk sample was also obtained from a human donor who received
36 mg lidocaine during dental work. 2,6-DMA was present at
levels rangingfrom 14.5 to 66.0 ppb in bovine milk and was
detected at 1.6 ppb in the human milk sample. Our results
demonstrate that 2,6-DMA, formed by the metabolism of
lidocaine, is transferable to bovine and human milk.
(4-6). 2,6-DMA metabolite excreted in the urine represents
about 1% of the administered dose of lidocaine in humans
(7); 4-hydroxy-2,6-dimethylaniline,the major urinary metabolite of lidocaine, is likely formed by further oxidation of 2,6DMA (8). Other potential routes of exposure to 2,6-DMA
include metabolism of the fungicide metalaxyl (9); metabolism
of the veterinary tranquilizer xylazine, which may remain as
a residue in meat (10); and exposure to cigarette smoke,
including side-stream smoke (11).
The transfer to human milk of lidocaine and its de-ethylated
metabolite monoethyl-glycinexylidide (MEGX) has been
demonstrated (12-14). Our laboratory has previously developed methods for the analysis of aromatic amines in biological
fluids (15,16), based on the technology of solid-phase microextraction (SPME) (17), with separation and detection using
gas chromatography-mass spectrometry (GC-MS). SPME has
also been used for the analysis of lidocaine in biological
matrices (18). We have now developed an SPME-GC-MS
method for analysis of 2,6-DMAin milk. The results presented
here demonstrate that 2,6-DMAproduced from the metabolism
of lidocaine is transferred to bovine and human milk.
~
Introduction
2,6-Dimethylaniline (2,6-DMA; chemical structures are
shown in Figure 1) has been identified as a nasal carcinogen in
rat long-term feeding studies (1). A recent study (2) indicates
that 2,6-DMAmay also act as a tumor promoter, enhancing the
effect of initiating carcinogens such as nitrosamines. Although
there is no direct evidence that 2,6-DMA is a human carcinogen, exposure to this aromatic amine, which may occur via
multiple routes, is of concern.
2,6-DMAis a metabolite of lidocaine. Lidocaine (2-(diethylamino)-N-(2,6-dimethylphenyl)acetamide) is commonly used
as a local anesthetic in medicine and dentistry; is indicated for
the treatment of cardiac arrythmia (3); and, administered intranasally, is effectivefor the treatment of migraine headaches
*Author to whom correspondenceshould be addressed.E-mail:[email protected].
.CH3
NH2 2,6-Dlmethylanlllne
CH~
CH3
/ f - ~ "H~
.CH3
/CH2CH3
CH2CH3
//--~ H ~
~
Monoethyl-
CH3 Udocalne
N
H
H
O[-13 glyclnexylldlde
~
/
I
\
~
CHs Xylazine
II
zCH--C--O--CH3
-N'g-CH2-O-CH3
CH3
Metalaxyl
FigureI. Structuresof 2,6-DMA,lidocaine,and relatedcompounds.
Reproduction (photocopying) of editorial content of this journal is prohibited without publisher's permission.
711
Journal of Analytical Toxicology, Vol. 25, November/December 2001
Experimental
Synthesis of 2,6-DMA-d9standard
2,6-DMA-d9was synthesized by nitration and reduction of mxylene-dl0 (CDN Isotopes, Inc., Pointe-Claire, QC, Canada) and
subsequent purification of the desired product, m-Xylene-dl0
(1.0 mL), water (0.3 mL), sulfuric acid (1.5 mL), and nitric
acid (0.90 mL) were mixed in a round bottom flask. The solution was refluxed at 59~ for 6 min. The product (nitro-xylene
isomers) was reduced with granulated tin (2.30 g) and hydrochloric acid (4.76 mL) and vigorously stirred for 20 rain at
58~ The product mixture was dissolved in methanol prior to
high-performance liquid chromatography (HPLC) purification.
Purification of the 2,6-DMA-d9standard was carried out on
a Hewlett-Packard (now Agilent Technologies, Wilmington,
DE) HP 1050 HPLC with Supelcosil PLC-SI column (25 cm x
4.6-ram i.d., 18-1Jm particle size, Supelco, Sigma-Aldrich
Canada), using conditions adapted from those of Schmeltz et
al. (19). The peak corresponding to 2,6-DMA-d9 (identity confirmed by GC-MS, see GC-MS section) was collected, quantitated by UVabsorbance, and used to prepare the internal standard dilution.
Milk samples
CA] was held at 60~ for I min and then ramped to 140~ (at
10~
no hold time at 140~ and finallyto 250~ (25~
l-rain hold). Total run time was 14.40 min. The injector temperature was set to 210~ and the detector was set to 300~ The
dwell time for SIM mode was 10 ms. Ion acceptance range was
--0.3 to +0.7 of nominal mass. Dailymidmass autotunes with perfluorotributylamine were carried out. Liquid injections (1 pL)
were made in the splitless mode. The head pressure was 10 psi.
For SPME injections, the purge was started at I min.
Figure 2 shows the mass spectrum of 2,6-DMA-dg. The
parent ion is m/z 130, and loss of a CD3 gives daughter ion
m/z 112. For 2,6-DMA, the corresponding ions occur at m/z
121 and 106, respectively.
Calibration curves, precision, and method detection limit
Calibration curves for 2,6-DMA in bovine milk were generated over the range 0.5-20 ppb. To test day-to-day variation,
three separate calibration curves were generated on different
days. The precision and method detection limit (MDL)were determined from 10 replicate extractions of 5 ppb 2,6-DMA + 5
ppb 2,6-DMA-dg.For quantitation of 2,6-DMAin milk, samples
were spiked with the internal standard (2,6-DMA-dg)to a final
concentration of 5 ppb.
Milk analysis
Bovine milk was obtained from animals treated at the Ontario
Veterinary College (Guelph, ON, Canada). Samples were kept
frozen at -70~ until analyzed. Samples were thawed and then
sonicated for 5 rain at 25~ An aliquot of milk (8 mL) was
pipetted into a Corning 15-mL conical centrifuge tube, spiked
with the 2,6-DMA-d9internal standard, and mixed thoroughly.
Samples were centrifuged (4000 rpm, 10 min, 4~ causing
the fat to separate. The fat plug was removed by use of a cotton
applicator, and the "skim" milk was transferred to a clean tube.
An aliquot of the skim milk (5 mL) was then pipetted into a 7mL Supelco clear vial with hole cap and PTFE/silicone septum,
prepared as described here, for SPME sampling.
Bovine milk. Milk samples (approximately 30 mL) were
taken from each of seven Holstein cows. Samples were
obtained before surgery and at a single time 2.5-6.0 h after
lidocaine injection.
Human milk. Control milk samples from 15 human donors
not exposed to lidocaine (or other known sources of 2,6-DMA)
were obtained with informed consent. A single human milk
sample was obtained from a woman who had received 36 mg
lidocaine by local injection during dental work; the sample was
collected 6 h after the ]idocaine injection.
SPME
Results and Discussion
The SPME sampling protocol was adapted from our earlier
method (15). The SPME fibers (CW/DVB coating) and the
SPME holder assembly were purchased from Supelco. One
milliliter of 12M KOH solution was pipetted into a 7-mL vial
containing 1.403 g NaCl (Fisher Scientific, Nepean, ON,
Canada), an octagonal stirbar (1/2in. x 1/8 in.), and 5 mL skim
milk. Vials were stirred (1500 rpm) at 60~ using a digital
hotplate/stirrer (Dataplate| PMC 730 series, Fisher Scientific).
Vials were equilibrated (3 rain) before the SPME fiber was inserted into the headspace. Headspace extraction was carried
out for 30 rain; the fiber was then retracted and injected into
the GC. Thermal desorption time was 1 min.
SPMEanalysis
Based on preliminary optimization studies, we selected the
SPME conditions (fiber type, temperature, and time) described
130
2
xlO
~
112
<
106
GC-MS
Analyseswere conducted using a Hewlett-Packard GC (5890
series) coupled to an HP 5971 mass selective detector in the
selected ion monitoring mode. The carrier gas was He at 20
cm/s. A silanized 0.75-ram insert was used. The column [(5%)diphenyl-(95%)-dimethylsiloxane copolymer, 30 m x 0.25-ram
i.d., 0.25-1Jmfilm thickness, DB-5MS,J&W Scientific, Folsom,
712
-i
....
20
i ....
30
i ....
40
i ....
50
i ....
60
i ....
70
i ....
80
m/z
Figure2. Massspectrumof 2,6-DMA-dg.
, ....
90
i ....
i ....
100 110
1-
120 130
Journal of Analytical Toxicology, Vol. 25, November/December 2001
in the Experimental section, as providing optimum extraction of 2,6-DMAin a reasonable period of time. A typical calibration curve (0.5-20 ppb) is shown in Figure 3. Calibration
curves were linear (r2 > 0.99). The day-to-day variation was
estimated as 1t.0%, based on calibration curves generated on
three days. Relative standard deviation (10 replicate extractions, 5 ppb) was 3.6%. For a 99% confidence interval, the
Student's t value is 2.821 (n = 10 with 9 degrees of freedom).
The MDL,calculated by multiplying this value by the standard
deviation for the 10 replicates (0.0725), was 0.20 ppb.
Lidocaine and certain metabolites, such as MEGX, have previously been detected in blood, urine, and human milk. In the
present work, we have shown that headspace SPME, using the
CW/DVBfiber, coupled with GC-MS provides a rapid, reliable,
and sensitive method for the analysis of the lidocaine metabolite 2,6-DMA in animal and human milk. In recent publications, Abdel-Rehim and colleagues (20,21) have also applied
SPME, in combination with GC or HPLC-tandem electrospray
mass spectrometry, to the analysis of lidocaine and DMA in
spiked samples of human plasma and urine.
2,6-DMA in bovine milk
Milk samples obtained before lidocaine injection were uniformly negative for 2,6-DMA and all samples obtained postinjection were positive, with 2,6-DMAconcentrations ranging
from 14.5-66.0 ppb. Typical blank (pre-injection) and positive
(post-injection) GC chromatograms are shown in Figure 4. The
concentrations of 2,6-DMA were calculated from the calibration curve. Table I reports, for each sample, the lidocaine dose
administered, the time of milk sampling, the measured concentration of 2,6-DMA, and the protein and lipid content.
Levels of 2,6-DMAin milk varied over an approximately fourfold range. There was no obvious relationship between milk
2,6-DMA level and dose, time of sampling, or milk protein or
fat levels, although the number of samples examined was
limited.
A
60
l
60
7.9
Fifteen milk samples from donors with no known exposure
to lidocaine or 2,6-DMA were analyzed and none contained
detectable levels of 2,6-DMA. The single milk sample obtained
from a donor who had received an injection of 36 mg of lidocaine contained 1.6 ppb 2,6-DMA. A typical blank chromatogram and the chromatogram for the positive sample are
shown in Figure 5. (Although it is not possible to state with
certainty that the 2,6-DMA is lidocaine-derived, this seems
very likely, because the compound was never found in unexposed control samples.)
~
8.0
8.1
8.2
8.3
RetentlonUm(mln)
B
5000] ion121
4000]
30001
20001
looo1
01
2,6-DMA in human milk
106
ion
.
.
.
.
5000 t lonl06
40001
30001
20001
looo 1
Ot
.
.
.
7.9
.
8.0
8.1
8.2
8.3
Retention time (mln)
Figure 4. Analysis of 2,6-DMA in bovine milk. Typical blank (pre-injection,
A) and positive (post-injection, B) GC chromatograms are shown. The vertical axis represents relative detector response, and the blank chromatograms are shown on an expanded scale.
Table I. 2,6-DMA in Bovine Milk Samples Obtained after
Lidocaine Injections*
"6
Dose
Post-injection
samplingtime
Case
(mg/kg)
(h)
(ppb)
(g/L)
(%)
261
259
254
260
255
258
256
3.9
3.5
3.4
3.2
3.0
3.0
2.9
3
4
2.5
6.5
3.5
2.5
3
14.5
66.0
47.9
16.8
36.8
24.0
39.7
40.5
50.4
28.9
46.4
36.6
33.3
31.0
10
11
13
6
4
7
8
o
5
10
15
20
2,6-DMA (ppb)
Figure 3. Calibration curve (0.5-20 ppb) for 2,6-DMA in bovine milk. The
parent ions of 2,6-DMA and 2,6-DMA-d9 are m/z 121 and m/z 130,
respectively.
2,6-DMA Protein
Fat
* Milk protein and fat concentrations were measured using the modified Lowry
assay(22) and modified Folch extraction (23), respectively.
713
Journa] of Analytical Toxicology, Vol. 25, November/December2001
The presence of 2,6-DMA as a lidocaine metabolite in milk
demonstrates that this xenobiotic can reach the mammary
tissue. The observation of 2,6-DMA in the milk of a nursing
mother treated with lidocaine suggests that the drug metabolite can reach the milk and be ingested by nursing infants. For
a compound to pass from the maternal circulation to the milk,
it must cross multiple membrane barriers. Compounds that
are relatively lipophilic, low molecular weight, un-ionized,
and not strongly protein-bound will diffuse into the breast
milk most readily. 2,6-DMA satisfies all of these criteria.
Conclusions
Treatment of nursing mothers with lidocaine or other compounds which may give rise to 2,6-DMA is likely to be a
potential route for exposure of infants. Although the levels of
2,6-DMA exposure are far below those which are associated
with its carcinogenicity in experimental animals, any exposure
of infants to the compound should probably be avoided, if possible. Monitoring human exposure requires the development of
reliable methods for analysis of 2,6-DMA in complex matrices.
A
2oo
. ' , i ~ w
1oo
200]
"~
.
.
.
.
.
.
.
7.8
.
.
.
7.9
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
8,2
R~ention time (mln)
8.0
8.1
.
.
.
.
.
.
.
.
.
.
.
.
.
.
8.3
B
1400 1
1200 1
i000 1 ion 121
800 1
600 ]
400 1
200
0t
. . . .
1400 1
1200 1 ion 106
1000 1
800 1
600 1
400 1
200
0~
7.0
7.9
8.0
0.1
8.2
s.3
Retention time (mln)
Figure 5. Analysis of 2,6-DMA in human milk. A typical blank GC
chromatogram (A) and the GC chromatogram for the positive sampleobtained from a donor who had receivedan injection of 36 mg lidocaine (B)
are shown.The vertical axis representsrelative detector response,and the
blank chromatograms are shown on an expandedscale.
714
Headspace SPME coupled with GC-MS is a rapid and sensitive
method for analysis of 2,6-DMA in milk.
Acknowledgments
We are grateful to Dr. Donald Trout, Ontario Veterinary College, who conducted the bovine anesthesia and collected the
bovine milk samples. Dr. Perry Martos, Lab. Services, and Dr.
Julie De Merchant, Chemistry and Biochemistry, University of
Guelph, generously provided advice and access to analytical instruments. We also thank Dr. Martos for critical reading of the
manuscript. Finally, we wish to thank Patricia Solbeck for
measuring the milk protein and lipid levels. Funding was provided by the Toxic Substances Research Initiative (Health
Canada) and NSERC Canada.
References
1. J.K. Haseman and J.R. Halley. An update of the National Toxicology Program database on nasal carcinogens. Mutat. Res. 380:
3-11 (1997).
2. T. Koujitani, K. Yasuhara, H. Kobayashi, A. Shimada, H. Onodera,
H. Takagi, M. Hirose, and K. Mitsumori. Tumor-promoting activity
of 2,6-dimethylaniline in a two-stage nasal carcinogenesis model
in N-bis(2-hydroxypropyl)nitrosamine-treated rats. Cancer Lett.
142:161-171 (1999).
3. J.A. Joglar and R.L. Page. Treatment of cardiac arrhythmias during
pregnancy: safety considerations. Drug Saf. 20:85-94 (1999).
4. M. Maizels, B. Scott, W. Cohen, and W. Chen. Intranasal lidocaine for treatment of migraine: a randomized, double-blind,
controlled trial. J. Am. Med. Assoc. 276:319-321 (1996).
5. T.M. Mills and J.A. Scoggin. lntranasal lidocaine for migraine
and cluster headaches. Ann. Pharmacother. 31:914-915 (1997).
6. R. Buckley, T. McCurry, and J. Gera. Intranasal lidocaine for migraine using a metered-dose spray. Headache 40:498 (2000).
7. S.D. Nelson, W.A. Garland, G.D. Breck, and W.F. Trager. Quantification of lidocaine and several metabolites utilizing chemicalionization mass spectrometry and stable isotope labeling.
J. Pharm. Sci. 66:1180-1190 (1977).
8. C.R. Short, M.L. Hardy, and S.A. Barker. The in vivo oxidative
metabolism of 2,4- and 2,6-dimethylaniline in the dog and rat.
Toxicology 57:45-58 (1989).
9. J.V. Headley, D.B. Maxwell, C. Swyngedouw, and J.R. Purdy.
Determination of combined residues of metalaxyl and 2,6dimethylaniline metabolites in urine by gas chromatography/mass
spectrometry. J. AOAC Int. 79:117-123 (1996).
10. Y. Govaert, P. Batjoens, K. Tsilikas, J.M. Degroodt, and S. Srebrnik.
Multi-residue analysis of tranquillizers in meat: confirmatory assays using mass spectrometry. Analyst 123:2507-2512 (1998).
11. F. Luceri, G. Pieraccini, G. Moneti, and P. Dolara. Primary aromatic amines from side-stream cigarette smoke are common
contaminants of indoor air. ToxicoL Ind. Health 9:405-413
(1993).
12. D. Ortega, X. Viviand, A.M. Lorec, M. Gamerre, C. Martin, and
B. Bruguerofle. Excretion of lidocaine and bupivacaine in breast
milk following epidural anesthesia for cesarean delivery. Acta
Anaesthesiol. Scand. 43:394-397 (1999).
13. T.H. Lebedevs,R.E.Wojnar-Horton, P.Yapp, M.J. Roberts, L.J.Dusci,
L.P. Hackett, and K. Ilett. Excretion of lignocaine and its metabolite monoethylglycinexylidide in breast milk following its use in
a dental procedure. A case report. J. Clin. Periodontol. 20:
606-608 (1993).
Journal of Analytical Toxicology,Vol. 25, November/December2001
14. M. Giuliani, G.B. Grossi, M. Pileri, C. Lajolo, and G. Casparrini.
Could local anesthesiawhile breast-feeding be harmful to infants?
J. Pediatr. GastroenteroL Nutr. 32:142-144 (2001).
15. L.S. DeBruin, P.D. Josephy, and J.B. Pawliszyn. Solid-phase microextraction of monocyclic aromatic amines from biological
fluids. Anal Chem. 70:1986-1992 (1998).
16. L.S. DeBruin, J.B. Paw[iszyn, and P.D. Josephy. Detection of
monocyclic aromatic amines, possible mammary carcinogens, in
human milk. Chem. Res. Toxicot. 12:78-82 (1999).
17. S. Ulrich. Solid-phase microextraction in biomedical analysis.
J. Chromatogr. A 902:167-194 (2000).
18. F. Sporkert and F. Pragst. Determination of lidocaine in hair of
drug fatal ities by headspace solid-phase microextraction. J. Anal.
Toxicol. 24:316-322 (2000).
19. I. Schmeltz, C.T.Dyer, E. Ratchik, and J. Matasker. Separation and
determination of xylidine isomers: comparison of gas and liquid
chromatographic methods. J. Chromatogr. 245:309-320 (1982).
20. M. AbdeI-Rehim, M. Bielenstein, and T. Arvidsson. Evaluation of
solid-phase microextraction in combination with gas
chromatography (SPME-GC) as a tool for quantitative bioanalysis. J. Microcot. Sep. 12:308-315 (2000).
21. M. AbdeI-Rehim, M. Bietenstein, Y. Askemark, N. Tyrefors, and
1",Arvidsson. High-performance liquid chromatography-tandem
electrospray mass spectrometry for the determination of Iidocaine and its metabolites in human plasma and urine. J. Chromatogr. B 741:175-188 (2000).
22. G.U Peterson. A simplification of the protein assay method of
Lowry et al. which is more generally applicable. Anal. Biochem.
83:346-356 (1977).
23. R.G. Jensen. The Lipids of Human Milk. CRC Press, Boca Raton,
EL, 1989.
Manuscript received March 2, 2001;
revision received June 11,2001.
715