Clin (1993) 44:349-355

O Springer-Verlag 1993

Inhibition by paroxetine of desipramine metabolism in extensive

but not in poor metabolizers of sparteine
K. Brøsenll 2, J.G.Hansen2, K. K. Nielsenl,S. H.Sindrupl 2, and L.F.Graml2
Departmenl of Clinical Pharmacolow. Institute of Medical Biology. Odense University, Odense.
Dcnmark 2 Odense Drug Evaluation Center (ODEC.)F Odense, Denmark

Rece ived•- Julv 25.1992/Accepted in revised forrm December 22.1

Key words: Paroxetine, Dcsipramine; sparteine/dcbriso-

quinc, gcnctic polymorphism, drug-drug interaction
Summary. Nine extensive metabolizers (EMs) and eight
poor metabolizers (PMs) of sparteine took a single oral Paroxetine, a selective serotonin reuptake inhibitor
dose of III) mg of desipraminc HC.I before and while lak- (SSRI)
t l, is metabolized via both the CYP2D6, the source
ing paroxetine 20 mg per day.
Before paroxetine, the median of the total desipramine of lhe sparleine/debrisoquinc oxidation polymorphism,
clearance was 7 times higher in EMs than in PMs (102 and and low affinity enzymes which have not yet been idcnti•
fied (2-41.
15 1 CYP2D6 bccomcs saturated in the thcra-
respectively). This confirms that desipramine is
extensively metabolized via the sparteine/debrisoquine
oxidation polymorphism i. e. by CYP2D6. During paroxe-
tine, the median clearances were 22 1 •h -1 and 18

1 •h -I in
EMs and PMs respectively. The 5-fold decrease in clear-
ance in EMs when desipramine was co-administered with
paroxetine confirms that paroxetine is a potent inhibitor
of CYP2D6. The lack of effect on clearance in PMs shows
that paroxetine is a selective inhibitor of CYP2D6, which
is absent from the livers of PMs. Before
Lhe
median of desipramine clearance via 2-hydroxylation was
40-times higher in EMs than in PMs (56 and 1.4
re-
spcctivcly), but during paroxetine, it was only 2-times
higher (6 and 2.91
I respectively). The increase in this
clearance in PMs suggests that paroxetine is an inducer of
the altcrnative. unidentified P 450(s) which catalyze(s) the
formation of 2-0H-desipramine in this phenotype,
Before paroxetine, the median amounts of 2-0H-desi-
pramine glucuronide recovered in urine were 69% and
68% of the total recovery of 2-0H-desipramine in urine in
EMs and PMs respectively. During paroxetine, the corre-
sponding values were 77% and 84%. This increase in the
relative recovery of the glucuronide was statistically signi-
ficant in both phenotypes* suggesting that paroxetine is
a weak inducer of the glucuronidalion of 2-0H-desi-
pramine.
peulic dose range and the low-affinity enzymes become
more important for the elimination ofparoxetine when the
dose is increased or during repeated administration [3, 4].
Paroxetine, like some other SSRIs, is a polent inhibitor of
oxidations catalyzed by CYP2D6 in human liver micro-somes
[5—7]. [n accordance with this, we have shown that subjects
who are extensive metabolizers (EMs) of spar-teine either
become less extensive m etabo]izers or are con-verted to
poor metabolizers (PMs) while taking paroxetine 30
mgperday [3 5]. Ihei nhibition ofsparteine metabolism was
completely reversed a few days after withdrawal [3].
The CYP2D6 is a major enzyme catalyzing the meta bo-
lism of tricyclic antidepressants, certain neuroleptic drugs,
"-adrenoceptor blockers. and some antiarrhythmic drugs, as
well as the oxidative activation of codeine to morphine [8. 9].
Mutations in the CYP2D6 gene [10] cause the ab-sence of the
enzyme from the livers of PMs, [11], who make up about 7%
of white Caucasians [12]. In PMs the sub-strates of CYP2D6
are eliminated slowly by other low-af-finity P450s, by other
enzymcs. or by renal excretion.
spaneine/debrisoquine oxidation polymorphism has
a major impact on the elimination of desipramine [131. The
consequences for desipramine elimination kinetics are
genetic polymorphism [13—151, saturation kinetics in EMs
[15], and drug-drug interactions during the administration of
IN-dent inhibitors of CYP2D6. such as thioridazine [161.
perphenazine [17]. quinidine (181. and fluoxetine [19].
Paroxetine would also be expected to impair the
meta-bolism of desipramine. and we have therefore
studied the effects of daily paroxetine 20 mg on the
single oral dose kinetics of desipramine in 9 EMs and 8
PMs. We also re-port some new findings with regard to
thc metabolism and pharmacokinetics of desipramine.

Methods

Study population

We sclcctcd 9 EMs and 8 PMs of sparlcinc from a database of ap-

proximately 16(Å) healthy Danish subjects phenotyped with regard
to sparteine and mephenytoin oxidation. he metabolic ratio with
350

Table 1. The 0—48 h urinary recoveries of desiprammc and unconjugatu-l conjugaled metabolites (expressed as percentages of a single
oral dose of mg desipraminc HCI) before (Pcri(.xf I ) and during (Pcnod 2) daily paroxctine 20 mg. •lhe data arc reported as medians and
rangcs (in brackets) in nine extensive
(EMs) and eight pmr metatK>lizers (PMs) ofsparteine

Phenoty•pe Desipramine 2-OH-&sipramine 10-0} I-desipramine

Period I Period 1 Period 2 Total recoveryPeriod 2
Period 2 Period 1 Period 2 Period 1
1.0 44 3.1 45 20 *
All EMs 2.9* (13-26)

PMs 2.6
4.1
3.5
(2.9-5.6) (I. 5—4.9)
Mann-Whitney U-test: (2.8-4,2) 10 (10—14)
All EMs
PMs
om02 0.200 0.0025
Wilcoxon/Pra1t test: Period 1 vs Period 2:
* P < 001 ** P < 0.05 x not quantifiable
0.02 o.on

respect to sparteine (MR. see below) ranged from O. 1610 0.38 in four quantification was 10 nmol •l- and the coefficient of variation was
more extensive metabolizers. from 0.70 (0 0.91 in 5 less extensive All assays were run in duplicate.
metabolizers and from 22 to .5(Din 8 PMs (Table 1). Desipramineis hydroxylated to 2-0H- and 10-0H-desipramine,
Fifteen of the subjects had an SIR ratio (see below) of les than
which are excreted in the urine both unconjugated and as glucu-
0.1 and were therefore defined as EMs of mephenytoin* whereas two

sparteine PMs ronide conjugates. N-demethylation to didesipramine and side-

9 and 17) had SIR cbajn dealkylation to the
of metatX)1ite are minor path-
and 1.092 and ways. and
were defined as PMs of mephenytoin
The 30 400/0 of the dose is recovered as unidentified
assign, metabolites (fig. 1) [23). Thus, desipramine, 2-0H-desi-
ment was confirmed in all subjects by a rer:eat test using spa.rteine pramine,
and mephenytoin combined. and didesipramine in serum and
All the subjects were men (aged 20—24 y) and none took daily al- urine were assayed by reversed phase HPLC. A detailed account of
cohol or drugs. Clinical examination and clinical chemistry,maema-

tology screening confirmed that all were healthy.
The study was approved by the regional ethics u)mmittee and by
the Danisb National Board of Health. The volunteers consented to
participate on the basis of verbal and written information.
Study procedure
On day I in I of the study each subject
desipramine hydrcrhloride (Pertofrane, Ciba-Geigy A'S) at
(each tablet contained 25 mg of the hydrochloride or 82.5
the free base) and fasted for the next 3 h. of
samples for serum drug assay were drawn into dry töt-
tulrs hourly for the first 14 h and then at 24, 32.48* and 72 h- In the
four tablets of
h compounds werc monitored by IN-detection at 220 nm. All samples
the proædure has tren published scparately [241. In brief: sample
preparation for serum and urine. involved a three-step liquid-liquid
extraction procedure. The urinary concentrations of desipramine, 2-
OH-&sipramine. I()-OH-desipramine, and didesipramine were as-
sayed with and without hydrolysis. 'l'he glucuronide conjugates of
the hydroxylated metabolites were hydrolysed by incubating the
urine samples for 16 h at 3TCwith a mixture of Å-gtucuronidase and
arylsulphatase. •me hydrolysis process was followed by sample prep-
a-ration. in order to assess the total conczntration of unconjugated
plus conjugated 2-0H-desipramine and 10-0H-desipramine, 'Ihe
were asayed in duplicate.
Hydrolysis of the urine from paroxetine-treated subjects who
had not taken desipramine (3) showed two small extra peaks which

did not interfere with the chromatograms. Accurate determination

PMs additional blood samples were drawn at 120. 168, aud 240b.
was possible down to 20 nmol -l - in serum and 50 nmol -l - 1 in urine
Ihe tubes were centrifuged and the serum was kept at — 200C until
analysis. Urine was before hydrolysis. After hydrolysis of urine, the lower level of
for 0-12. 12-24. and 24—48 h; the vol- quantification was 1 p.mol•l-l, except for LO-OH-desipramine for
ume was recorded and 50 ml aliquots were kept at — 20 'C. which it was 4 pmol•l-i. At 50 nmol•l-' the within-day (n = 10)
During period 2 the subjects variations @efficients of variation) in plasma were respectively 6,
two paroxetine hydrochhride 9, 18, and 12% for 2-0H-desipramine, 10-0H-desipraminc, didcsi•
tablets (each containing 10 mg of base, corresponding to 1).4 pnol) pramine. and desipramine. In urine at 50 nmol •l -i tEfore hydro-
al 08.00 h daily for 20 days Each dose was given by a research
lysis the coefficients of variation were 5.4, 14, and 7% respectively.
nurse. On study days 9, 10, and II a blood sample was taken imme-
After hydrolysis of urine the coefficients of variation at 1 umol• I -L
diately before dosing, and the serum was kept frozen until assay
for paroxetine. On day II at were 4, 7, and 7 % for 2-0H-desipramine, didesipramme, and desi-
pramine rwpectively. For ICOH-desipramine at 4 umol•]
w h the subjects tCN)k T) mg of it was
desiprauüne HCI and fasted for the next 3 h. Blood sampling 3.3%.
took place as in period 1, except Ihal the samples at 120. 168. and Thus. the method permits detection of about 0.03 % of the dose
240 h were also drawn in the recovered as drug-related material
Urine was collected as in peri- hydrolysis and 0.6% after
{.xll. For 100H-desipramine the value. is about 2-2.5 % after
hydroly". depending on urine volume.

Analytical methods
Spartcinc, 2,3-dehydrospartcinc. and 5 "-dehydrospartcinc in urine
were assayed by gas chromatography (20] with a lower level of quan-
tification of 0.25 umol •l -l. Below this value, the coefficient of vari-
ation exceeded 20%. The chromatographic peak areas of R, and S-
mephcnyloin were assayed by gas chromatography [2 11.
The scrum concentrations of paroxcünc wcte dctermincd by
quantitative thin layer chromawgraphy (221. The lower level of
Pharmacokinetics
he area under the serum concentration time curve (AUCDMI) was
calculated by log-linear trapezoidal methods. and the area from the
last measurable concentration Chi* to infinity as Clast/A, where A,
was (Elermined by iu:rative non-linear least squares regression of
thc final monoexponential part of the concentration time curve
(251.

Unknown paroxetile): total clearance (CL), clearance via 2-hydroxylation

(CLQ0H), clearance via 10-hydroxylation
renal clearance
(CLR), and residud clearance (CLR.,). according to the fol-
lowing formulae:
Oasetamne
dose
(1)
AUCDMI
CH2
CHz

Dealkylation Hydroxylation Demethylatio

I mw19d.bMZYl 2-uq-d%ipra.mime 1001+desøam*19

CH2 CH2
CH2

Glucuronidation

mg. 1. The pattern of biotransformation ofdesipramine. The meta-

bolites 2-0H-desipramine and 10-0H-desipramine are excreted
both unconjugated and as glucuronide conjugates in the urine
•Ma-jor pathway; •Minor pathway

•10* and II in period 2 are rc-

Five diffcrcnt oral clcarances of desipramine were calculat-

ed both in period I (before paroxetine) and in 2 (duriug
 351
10 47 121 242 238 72 58

Table 2 Thc C and of desipramioe after single oral doses of 15 80 53 226 245 52 62
ICN) mg desipramine HCI before (Period 1) and during 17 113 71 334 24S 53 37
paroxetine 20 mg per day (Pcriod 2) 13 117 74 265 207 49 5-4
Sparqcinc Paroxetine Desipramizc 12 37 202 142 76 71
c 16 87 309 m 70
nmol• I -J (NS) 60 (NS)
Median 97 73 241 230 58

Period 1 Period2 Period 1 Period 2 Mann-Whitney CT-test:

EMs Fst EMs vs slow

4 29 o. (63 0.0556 0.413 0.2M

225 15 33
7 o. 16 18 45 180 11 57 o.uxg 1.00 0.114

6 024 18 109 285 12 26 Wilcoxon!Pratt test:

1 038 21 188 21 49 Petiod NS P > 0.05; * P < 0.01
Median 020 20 78 207 14 41
Slow EMs desipramine. Eq. 2 and eq 3 are based on the assumption that 2-0H-
3 0.70 176 143 342 12 76 desipramine and 10-0H-desipramine are only eliminated as uncon-
031 24 ] 43 184 16 so jugated and conjugated hydroxylated metabolites and that the total
2 47 145 265 17 51 0 48 h urinary recoveries reflect the rates of 2-hydroxylation and 10-
s (188 37 194 26 52 hydroxylation.
0.91 26 215 239 20 45 For 2-0H-desipramine.
Mcdiül 073 37 143 239 17 51 and a clearance via a ctearance via glucuronidation
Median 26 225* 16
renal excretion CL2-0H_ were calcu-
PMs laled as follows:
14 47 35 48
40 203 239 277 63 86
CL20H Aez-oe - OH — Ae•2 - uacmj.

(2) CL 2-1k ej-c (7)

Aen- OH,
AUCU& (3) CLT.M{ =
Ae.*2 -
(8)

CLF AeDMl 0-+48 (4)

where is the amount excreted

= CL-CLQOWCLIOOH-CLR as unconjugated metabolite and AUCåzßH is the AUC of 2-0H-

desipramine from 0 48 h.
The sparteine meqatk)lic ratio was calculatcd as:
(S)
The residual fraction of desipramine that was melaboli7*d 10 un- (9)
12 h rccovcry of sparteine
known metabolites, fre., was calculated as follows: 12 h recovery of 2.3 dehydt[xparteinc and 5.6 — dchydrøsparteine

(6) ne mephenytoin YR was calculated as the ratio between the chro-

and arc lhe total amou.ms of 2-0H-desi• matographic
pramine and 10-0H-desipreminc (aftcr hydrolysis) excreted during area ratios of S- and R-mephenytoin in the 12-h
()—48 h; is thc amoum of unchanged desiptaminc urine [21).
hydrolysis) excreted during 0-48 The arithmetic means of thc prc-dose steady-slate. serum con-
is (he 48 h ACC of ccntrations of paroxetine on
days pttcd as thc CX„,.
352

Statistics geiore puoxäine During paroxetine

The urinary recoveries ('IUb}e I) in 9 EMs and 8 PMs were com- Oe*remiM
pated using the Mann-Whitney Icsl. as were the plasrna kinetics
(Table 2) in 4 more extensive metabolizers versus 5 les extensive
metabolizers and in all EMs venus 8 PMs- The data in period 1 and

pcriLN32 were compared using the Wilcoxonmratt test for

paired observations. Spearrna-lß rank ajrrelation
was used (or 10

comparison of the MRs and of (he values 01 CL

(cq. 7)
and CI *201...
(eq 8) in nine EMs in period 1,
statistical tests
were camed out with the MEDSTAT program package version 21
EM
period 2. In period 1 thc median value was nearly three
times higher in thc more cxtensive metabo\izers than in
Results
thc less extensive metabolizers, and thc median value was
forty times highcr in all EMs compared with the PMs. In
The group average serum concentrations of desipramine pcriod 2 lhe median value was reduced by a factor of tcn
in 4 more extensive metabolizers 5 less extensive meta-
bolizers and 8 PMs are shown in Fig. 2.2-0H-desipramine
was only detcctcd in EMs before paroxetine, and 10-OH-
desipramine and didcsipraminc were undetectable in
serum during both periods.
In the urine. desipramine, 2-0H-desipramine, and 10-
OH-desipramine (but not didesipramine) were detect-
able in all subjects in both periods 1 and 2 (Table 1).
However, the concentrations of 10-0H-desipramine after
hydrolysis were below the level of quantification in EMs
in period 1. The amounts of the hydroxylated meta-
bolites in the urine increased several-fold after
hydrolysis with ß-glucuronidase/arylsulphatase, but
there were no statistically significant differences
between the desi-pramine concentrations measured
before and after hy-drolysis (P = 0.27, n = 102,
Wilcoxon's test). In EMs the total recovery [median
(range)] was 45% (38—54%) in period 1 and 20% (13—
26%) in period 2. In PMs, the total recovery was
(9—12%) in period I and 14%
(10—14%) in period 2. In EMs the dominating compo-
nent was 2-0H-desipramine in both periods 1 and 2
(Table 1). In PMs desipramine, 2-0H-desipramine, and
10-0H-desipramine were excreted in almost equal
amounts in both periods (Table 1).
The individual steady-state serum
concentrations of
paroxetine and the individual values of
total
and partial clearances, and tn of desipramine, as well
as the resulß of the statistical analysis are detailed in Ta-
bles2 and 3. Paroxetine serum concentrations in the
PMs were on average about twice those in the EMs
(Table 2).
In period I the total clearance of desiprarnine (eq. 1)
was 12-299
(all subjects). and in period2 8-321.
median clearance was nearly seven times higher io
the EMs than in the PMs in period 1, but in period 2. thc
difference was practically abolished because of a redoc-
lion in clearance in thc EMs and a minor but statistically
insignificant increase in the PMs.
Thc clearance of desipramine via 2-hydroxylation
(cq.2) was 0.8—132
- in period I and 1.8—9.91, h-
in
 10

10

o 200 050 too 150 200 250

Time {h)
'he average serum-concentrations of desipramine (v) and
fig.2 2-
OH-desipramine (V) in four more extensive met bolizers ('fast
five less extensive
EM'), metaboliz rs ('show EM'), and eight poor
HCI before
metabolizers (PM) after a single oral dose of IWmg desipramine
l) and during (period 2) paroxetine 20 mg per
day

in the EMs but increased by a factor of two in the PMs.

Ac-cordingty, the clearance of desipramine via 2-
hydroxyla-tion was only two times higher in the EMs
than in the PMs during paroxetine.
The residual clearance (eq.5) was 8—162
1 •h -l in peri-
Odl and
in period 2. The median was four
times higher in the EMs than in the PMs in period 1, but
the phenotype difference was abolished because of a
four-fold reduction in residual clearance in the EMs in
period 2 ("l'able 2).
median (range) of lhe clearances for glucuronida-
tion and for renal excretion of 2-0H-desipramine (eq. 7 and
8) were 79 (35—140) and 33 (25—43) I •h- respectively in
the EMs in period 1. and neither value correlated with the
MR (Spearman's (s: 0.329 and 0.563 respectively,

In period
1 the median amounts (ranges) of 2-0H-
desipramine glucuronide excreted were 69 (58-81 ) % and
68 (51—74) % ot the total recovcries of 2-0H-dcsipramine in
the urine in the EMs and PMs respectively. In period 2 there
was a statistically significant incrcase in thcsc values:
Tabk 3. l'he totat and partial clearances of desipramine after single oral doses of 100 mg desipramine BCI before (Period 1) and during
pa-roxctine 20 mg per day (Period 2)
Clearance (l •h l)
erotal 2-hydroxylation Rena) Residual
1 D-hydroxylation
Subject Period 1 Fast EMs Period Period
Period 2 Period 2 Period 2 Period 1 Period 2 Period 1 Period 2

Median
Slow EMs

Median
All EMs:
Median
1.4 (NS)

Median
Mann-Whitney U-test: 18 (.NS) 1.5 (NS) 12 (NS)

Fast EMs vs slow EMS

0.016
All EMs vs PMs 0190 0M)5 0.916 0.032
o.W01
0.277 ou)01 o (1)55 0.481 0963 0.888 omo 0.888
Wilcoxon/Prat( test:
Period 1 vs Period 2: NS:P > 0.05; * P < 0.01

77 (69-92) % and 84 respectively (Fig. 3). For
10-0H-desipramine glucuronide the corresponding
values were 58 (40—61) % of the total recovery of 10-01-1-
desipramine in the urine in the PMs in period l. The value
increased statistically significantly in period 2 to 71 (62—
Discussion
We have confirmed that paroxetine is partially and desi-
pramine extensively metabolized via the sparteine/debri-
soquine oxidation polymorphism (Fig. 2, Table 1, Tablc
2. and Table 3) [2-4, 13-15, 18, 281.
As expected from previous in vitro and in vivo studies (3,
5, 6, 7]. we have found that paroxetine is a potent inhibi-tor of
the metabolism of desipramine, as evidenced by the two-fold
increase in Cmax. the three-fold prolongation of tlü, the 5-fold
dccrease in total clearance, the ten-fold de-crease in
clearancc via 2-hydroxylalion, and thc 4-fold de-crease in
residual clearance in thc EMs (Table 2). This did not happen
in thc PMs (Table 2). The changes in the EMs
354
are similar to previously reported during quinidine 2(1)
mg per day, and like quinidine paroxetine appears to be
an inhibitor of both the first-pass and the systemic me-
tabolism of desipramine [271.
2-0H-desiprarnine is formed by low-affinity P 450(s) in
PMs, but the isozyme(s) have not yet been identified [291.
The doubling of 2-hydroxylation clearance in PMs in peri-
od 2 suggests that paroxetine is an inducer of the low-af-
finity P4SO(s). Our data show that 10-0H-desipramine is a
minor metabolite in vivo, and the data suggest that its
formation is independent of the sparteine/debrisoquine
oxidation
(Table 3). The results of our re-
cent study on the formation of 10-0H-imipramine by
human liver microsomes support this view [291. Thc
P450(s) which catalyzc the 10-hydroxylation have not
been identified, but the modest increase in the I()-hy-
droxylation clearance in PMs in period 2 (Table 3) sug-
gests that this isozyme(s) may also be induced by
paroxe-tine.
The increase in the fractions of 2-0H-desipramine and
10-0H-dcsipramine which was excreted as glucuronides
in period 2 suggests that paroxetine also stimulates con-
% of total
2EOH-desipramine
100 EM
80
60
40
20
Period 1 Period 2
Fig. 3. The 0-48 h urinary recoveries of 2-0H-desipra.mine glucu-
ronide in per ot the total recoveries of 2-OH-desiprarnine.. The data
are shown as median (bar), interquartile range (box), and com-plctc
range (whishers). Pcriod 1 vs. pcriod 2: P < 0.01 (Wilcoxon
jugation of the hydroxylated metabolites (Fig. 3). It is not
unusual for an inducing agent to stimulate several phase
I as well a.s phase II enzymes [301.
The residual fraction fre (cf. Table 2) was more than 50%
in most subjects, especially in period 2, and it has re-cently
been suggested that N-sulphation is an important pathway
in desipramine metabolism in man [31]. Our re-sults
contradict this assumption in several ways. First, hy-
drolysis with arylsulfatase did not increase the concentra-
tion of desipramine in the urine. Second* the residual
clearance was about four times higher in EMs than in PMs
in period 1. Third* the residual clearance was reduced by a
factor of four in the EMs in period 2. The last two obser-
vations independently suggest that tbe residual fraction
represents oxidative
formed mstly by
CYP2D6.
•rwo other SSRIst fluoxetine and sertraline, like pa-
roxetine, have apparent inhibitor constant (K) of less
than 1 gmobl 'for CYP2D6-mediateddrug oxidations in
human liver microsomes (6, 7, 32]. Thus, it is to be ex-
pected that those SSRIs will also be potent inhibitors of
CYP2D6-mediated drug oxidations in vivo. and
bas
already been shown for fluoxetine [19* 33]. Citalopram
and fluvoxamine are SSRIs which arc weak inhibitors of
sparteine oxidation and imipramine 2-hydroxylation in
human kivcr microsomes [K, about 4-70 gmol -I-I [6,
341]. and it has recently been shown that citalopram and
fluvoxamine are also weak in vivo inhibitors or desi-
pramine metabolism [.35, 36].
We conclude that clinically important drug-drug inter-
actions may occur if paroxetine is co-administered with
tricyclic antideprcssants certain neuroleptic drugs, or
certain antiarThythmic drugs
and dosage reductions
will bc necessary. However, our results suggest that the
interaction may bc used beneficially, since the large vari-
ability in desiptamine kinetics is greatly reduced during
paroxetinc administration. and it should be possible to
treat all patients with a combination of paroxetine and a
low fixed dose of desipraminc. Finally, our data suggest
that paroxetine is a weak inducer of some phase I and
phase II enzymes but the clinicat significance of this is
not clear.
Acknowldgements, This study was supported by the Danish Medi-
Research Council (grants nos. 12-92ff and 12082-1) and by
SmiLhKhne Bcccba.m Pharmaceuticals, Harlow, UK. Paroxetine and
desipramine were administered by MTS Susanne Fugmann Schmidt.
Mrs. Lone Hans. and Mrs. Anne Dorete Jensen, who also collected
blood and uine. Drugs in urine and plasma were assayed
by Mr. Anders Tiedje, Mrs. Karin Bøjeseo Nielsen, and Mrs. An-
nelise Casa. The manuscript was tmd by Mt. Henrik Hornebcrg.
ne valuable technical assistance of all those mentioned above is
very much appreciated. The study represents a contribution to
the aims0fCOSTB1.
Referenc«
I. Johnson AM (1.989) An overview of the animal pharmacology of
paroxetine. Acta Psychiatr Scand 80 [Suppl 3501: 14—20
2. Sindtup SH, Gram LR Brßen K, Eshøj O. Mogensen EF (19m)
The selective serotonin reuptake inhibitor paroxetine is effective in
the treatment of diabetic neuropathy symptoms. Pain 42: 135— 14
3. Sindmp SH, Brøsen K, Grun IF, Hallas Skjelbo E, Allen A,
Allen GD. Cooper SM, Mellows G, Tasker TC, Zussmann BD
(1992) The relationship between paroxetine and the sparteine
oxidation polymorphism. Clin Pharmacol Ther 51: 278—287
4 Sindrup SH, Brßen K, Gram LF (1992) Pharmacokinetics of the
selective seroronin reuptake inhibitor paroxetine. Non-linearity and
relation to the sparteine oxidation polymorphism. Clin Phar-
51:288-295
5. Brøsen K, Gram LE Kragh-Sørensen P (1991) Extremely slow
metabolism of amitriptyline but normal metabolism of imi-pramine
and desipramine in an extensive metabolizer of spar-teine,
&brisoquine, and mephenytoin. Ther Drug Monit 13; 177-182
6. Crewe HK. Letmard MS, Tucker
Woods FRT Haddock RE
( 1992) The effect of selective serotonin reuptake
on
cytcrhrome P4502D6 (CYP2D6) activity in human 'iver micro-
s.cxnö. Br J Clin Pharmacol 34: 262—265
7, Skjelbo E. Brøsen K (1"2) Inhibitors ofimipramine
metabolism by human Ever microsomes. Br J Clin Pharmacol
34: 265—261
& Brøsen K, Gram LF (1989) Clinical significance of the spar-
oxidation
Eur J Clin Pharma.
36:537-547
9. Eichelbaum M. Gross AS (19X)) Ihe genetic polymorphism of
debrisoquine/sparteine metabolism - clinical aspects, Pharmacol
Ther 46:377-394
10. Kagimoto M. Heim M, Kagimoto K, Zeugin
Meyer UA (19111)
Multiple mutations of the hu.tnan cytochrome P45011D6 gene
(CYP2D6) in poor metabolizcrs of debrisoquine. J Biol Chem
265: 17209-17214
I l. Zanger (JM. VillMis F, Hardwick JP. Meyer UA (1988) Absence
or hepatic cytochrome P450bufl causes genetically dcficicnt
dc-brisoquine oxidation in man. Biochemistry 27: 5447—5454
12. Alvin G, Bechtel P, Iselius L. Gunder1•Remy U (19%)) Hydroxy-lation
polymorphisms of debrisoquine and mephenytoin in Eu-ropcan
populations. Eur J Clin Pharmacol 39:533-537
13. Bertilsson L, Aaberg-WisIedI A (1%3) The debris(4.njn hy-
droxylation test predicts steady-state plasma levels or dtsi•
pramine. Br J Clin Pharmaco! 15'. 388-3X)
14. Brøseo K, Otton SV, Gram LF (1986) Imipramine demethyla-tion
and hydroxylation: impact of Lhe sparteine oxidation pheno-type.
Clin Pharmacol Ther 40; 543--549
15. Brøsen K, Klysner R, Gram LF. Olton SV, Bech P, Bertilson L
(1986) Steady-state
of imipramine and its meta-
bo[ites in relation to the sparteine/debrisoquine polyrnorphism.
Eur J Clin Pharmacol 30: 679—684
16. Ilirschowitz J, Bennett JA. Zemlan FP. Garver DL (1983)
Thioridazine effect on dcsipramine plasma lcvcls. J Clin Psycho-
pharmacol 3: 376--379
17. Gram LF (1975) Effects of perphenazine on imipramine metabo-
lism in man. Psychopharmacol Comm 1: 165—175
18. Steiner E, Dumont E, Spina
Dahlqvist R (1987) Inhibition
desipramine 2-hydroxylation by quinidinc and quinine. Qin
Pharmacol Ther 43: 577-581
19. Bell IR, Cole JO (1988) Huoxetine induces elevation of desi.
pramine level and exacerbation of geriatric non-psychotic de-
pression. J Clin Psychopharmacol 8: 447448
20. Vinks A, Inaba T, Otton SV. Kalow K (1982) Sparteine metabo-lism
in Canadian Caucasians. Clin Pharmacol Thet31: 23—29
21. SanzEJ. Villén T. Alm Cl Bertilsson L (198.9) S-mephenytoin hy-
droxylation phenotypes in a Swedish population determined after
coadministration with debrisoquin. Clin Pharmaco) Ther
45: 495-499
22. Gram LF. Bjerre M. KraghSørensen R Kvinesdal B. Molin J,
Pedersen OL, Reisby N (1983) Imipramine metabolites in bltl)d of
patients during therapy and after overdose Clin Pharmacol Ther
33:335--342
23. Gram LF (1974) Meeabolism of tricyclic antidepressants. Dan
Med Bull 21:218-231
241 Kramer Nielsen K. Brøsen K (1"3) High-performance liquid
chromatography of imipramine and six metabolites in human
plasma and urine. J Chromatogr 612: 87—94
25. Allen GD (1 BO) MODFIT, A pharmaa)kineLic wmpuler pro-
grain. Biopharm Drug Disp 11:477-498
26. Wulff HR. Schlichting P (1988) MEDSTAT version 2.1. The
Astra Group A'S, Albertslund
355
27. Brøseo K, Gram LF (1989) Quinidine inhibits the 2-hydroxyla-
tioa of imipramine and dcsipramine but not the demethylation of
imipramine. Eur J Clin Pharmacol 37: 155—10
JC. Woods FRT
2&
RE, Lennard MS, Tucker GT
( 1992) The role of
P4502D6 in the metabolism of pa•
roxeline by human liver microsomes. Br J Qin Pharmacol 33:
29. 521-523
Kt Zeugm T, Meyer UA (1991)
of P45011D6, tho
target of the sparteine/debrisoquin oxidation polymorphism, in
tl'E
of imipramine. Clin Pharmacol Ther 49; 0—617
30. Okey A (1990) Enzyme induction in the cyt(xhrome P450 sys-
tern. Pharmacol
45: 241—298
31. Romiti P Giuliani L, Pacifici GM (1992) Interindividual vari-
abiljly in the N-sulphation of desipramine in human liver and
platelets. Br J Clin Pharmacol 33: 17—23
32. Brøsen Skjelbo E (191) Fkuoxetine and norfluoxetine are po-
tent inhibitors or P4SOI D6 — (he source of the sparteine/debtiso-
quine oxidation polymorphism. Br J Clin Pharmacol 32: 136—137
33. Bergstrom RF. Peyton AL. Lemberger L (1"2) Ouantification
and mechanism of the fluoxetine and tricyclic antidepressant in-
teraction. Clin Pharmacol Ther 51:239-248
34. Olton SV (1%4) In vitro oxidation of sparteine by human liver.
PhD thesis. Department of Pharmacology, University of Toronto
35. Gram LF, Hansen MGJ. Sindrup SH. Brøsen K, Poulsen JH,
Aaes-Jørgensen T. Overø KF (193) Citalopram: interaction
studies with levomepromazine, imipramine and lithium. Ther
Drug Monit 15:18-24
36. Spina E, Gampo GM, Avenoso X Pollicino MA, Caputi AP
(1992) Interaction between fluvoxamine and imipramine/desi-
pramine in four patients, Ther Drug Monit 14: 194—196

Dr- K- Brøsen
Department of Clinical Pharmacology
Institute of Medical Biology
Odense University
Wmsløwparken 19
Oderse C
Denmark