QT prolongation and torsades de pointes induced by an antifungal agent, D0870, in conscious dogs

Toxicology 132 (1999) 187–199

QT prolongation and torsades de pointes induced by anantifungal agent, D0870, in conscious dogs

Toshiyuki Matsunaga a, Takuma Harada a,*, Tsuyoshi Harada a, Masanori Inokuma a,Takeshi Mitsui a, Paul A. Duffy b, Andrew Aldridge b, Yasunori Shibutani a

a Toxicology Di6ision, Fuji Central Research Laboratory, Mochida Pharmaceutical, 342 Gensuke, Fujieda,Shizuoka 426-8640, Japan

b Safety of Medicines Department, Zeneca Pharmaceuticals, Mereside, Alderley Park, Macclesfield, SK10 4TG, UK

Received 6 August 1998; accepted 11 December 1998

Abstract

D0870 ((R)-2-(2,4-difluoro-phenyl)-1-[3-[(E)-4-(2,2,3,3-tetrafluoropropoxy)-styryl]-1H-1,2,4-triazol-1-yl]-3-(1H-1,2,-4-triazole-1-yl) propan-2-ol), a novel bis-triazole antifungal agent, induced sudden deaths at a high dose of 5mg/kg/day in a 6-month toxicity study in dogs. In the present study, we intended to elucidate the cause of the suddendeath in dogs. When used in a single dose, D0870 induced prolongation of QTc intervals in proportion to its plasmaconcentration, and the threshold plasma concentration of the drug causing 10% QTc prolongation was estimated tobe 3.8 mg/ml. Then, we conducted a study to induce sudden death in dogs using loading (50 mg/kg) and maintenance(5 mg/kg/day) doses with long-term ambulatory electrocardiographic monitoring. Marked QTc prolongation (52–96%), ventricular premature contractions, and T-wave alternans were observed in all 10 animals treated with the drug,and seven out 10 animals died of ventricular fibrillation (VF) associated with torsades de pointes (TdP) when the dogswere treated with D0870 for 14 or 16 days. The TdPs were elicited in both tachycardia and bradycardia, and someof them in the former proceeded to VF. Consequently, we clarified that D0870-induced sudden death is primarilyattributable to the development of TdP preceding VF and may be enhanced by sympathetic nervous tone producedby emotional or physical stress. © 1999 Elsevier Science Ireland Ltd. All rights reserved.

Keywords: D0870; Antifungal agent; QT prolongation; Sudden death; TdP; Dogs

1. Introduction

D0870, which is chemically (R)-2-(2,4-difluoro-phenyl)-1-[3-[(E)-4-(2,2,3,3-tetrafluoropropoxy)-styryl]-1H-1,2,4-triazol-1-yl]-3-(1H-1,2,4-triazole-

* Corresponding author. Tel: +81-54-6367035; fax: +81-54-6362447; e-mail: [email protected].

0300-483X/99/$ - see front matter © 1999 Elsevier Science Ireland Ltd. All rights reserved.

PII: S 0300 -483X(99 )00002 -5

T. Matsunaga et al. / Toxicology 132 (1999) 187–199188

1-yl) propan-2-ol (Fig. 1), is a novel antifungalagent that is active orally and parenterally againsta wide range of systemic mycosis, even in theimmunocompromised models (Yamada et al.,1993). D0870 caused sudden deaths in the high-dose group (5 mg/kg/day) of a 6-month dog study(personal communication). Because the dogs werefound dead or had died suddenly without anyremarkable symptoms and showed no histopatho-logic changes in the heart, the cause of suddendeath remains unknown. In contrast, we havealready obtained apparent evidence for D0870-in-duced QT prolongation in dogs and monkeys.These findings suggest that the cause of the sud-den death induced by D0870 may be drug-in-duced long QT interval syndromes (LQTS). TheLQTS in humans presents characteristic electro-cardiograms (ECGs) showing a prominent QTprolongation, ventricular premature contractions(VPCs), T-wave alternans, and repeated torsadesde pointes (TdP) at the occurrence of seizureswith syncope (Schwartz et al., 1995). AlthoughQT prolongation and TdP are produced by manydrugs (Jackman et al., 1988; Symanski and Gettes,1993; Napolitano et al., 1994; Haverkamp et al.,1995; Schwartz et al., 1995) and many experimen-tal models of TdP have been developed (Weis-senburger et al., 1993), there are few reports onTdP induced by only one drug itself in consciousdogs. In the present study, we intended to eluci-date the cause of the sudden death in dogs treatedintravenously with D0870 in detail, based on theresults of electrocardiographic analysis. To solvethis issue quickly, it is essential to establish exper-imental conditions for a conscious dog modelwith a high incidence of sudden death induced byD0870. First, we evaluated the relation betweenQT prolongation and plasma concentration ofD0870 in dogs by single intravenous dosing.Then, we conducted the study to induce, shortlyand with high frequency, sudden deaths in dogsby using loading (50 mg/kg) and maintenance (5mg/kg/day) doses with long-term ambulatory elec-trocardiographic monitoring in consideration ofthe long elimination half-life (ca. 3 days) of thedrug in dogs (personal communication). Wedemonstrated that TdP preceding ventricularfibrillation (VF) is primarily responsible for the

development of the sudden death induced byD0870, and have thus established a conscious dogmodel of TdP.

2. Methods

2.1. Test materials

An injectable preparation containing 50 mg ofD0870 per 20 ml of a sterile lipid emulsion sup-plied from Zeneca Pharmaceuticals (Cheshire,UK) was used for this study. A 10% Intralipid™emulsion purchased from Kabi Pharmacia(Dubendorf, Sweden) was used as the controlvehicle. Since the chemical and biological stabilityof the preparation diluted with vehicle had notbeen validated, the preparation without dilutionwas used in this study. Therefore, the dosingvolume was altered for each dose level.

2.2. Animals

Five male and two female beagle dogs weighing10.0–12.5 kg (White Eagle Laboratories,Doylestown, PA) were used for a single intra-venous dose study; and five of each sex, weighing8.5–13.0 kg (BMR Laboratory, Gifu, Japan), fora loading and maintenance dose study. The ani-mals were housed individually in stainless-steelcages in a climate-controlled room (temperature,2392°C; relative humidity, 55915%; 12-h light/12-h dark cycle). They were given 300 g/day of acertified solid diet (Lab Rations c4360; PurinaJapan, Tokyo, Japan) and tap water ad libitumduring the experimental period. The guidelines ofMochida Pharmaceutical were followed for thecare and use of the animals in this study.

Fig. 1. Chemical structure of D0870.

T. Matsunaga et al. / Toxicology 132 (1999) 187–199 189

Fig. 2. Changes in individual plasma concentrations of D0870 and QTc (log 600×QT/log RR) intervals in dogs after a singleintravenous dosing at 1, 2.5, 10, and 20 mg/kg. M, male; F, female.

2.3. Study design

2.3.1. Single dose studyIn order to clarify the relation between plasma

concentrations of D0870 and QT prolongation,we conducted a single dose study at dose levels ofcontrol, 1, 2.5, 10, and 20 mg/kg. The dosing

volume for each dose was 8, 0.4, 1, 4, and 8ml/kg, respectively, and the rate of infusion wasset at approximately 5 ml/min. The number ofanimals for each dose was 1, 1, 1, 2, and 2,respectively. The control animal was given 10%Intralipid™ alone. The dogs were medicated withD0870 or vehicle Intralipid™ via the cephalic vein


by a sterile plastic syringe or an infusion set witha 25 G butterfly needle.

Heparinized blood samples were taken from thecephalic vein at 0, 0.5, 1, 2, 3, and 6 h, and ondays 1, 2, 3, 7, 10, and 14, after treatment with 10or 20 mg/kg in consideration of the long elimina-tion half-life of this drug in dogs, whereas sampleswere obtained at 0, 0.5, 1, 2, 3, and 6 h afterdosing with 1 or 2.5 mg/kg. Plasma concentra-tions of D0870 were measured by a high-perfor-mance liquid chromatographic (HPLC) method.

In the animals given 10 or 20 mg/kg, ECGs(standard limb leads) were recorded with an elec-trocardiograph (503 FB-D; Fukuda ME, Tokyo,Japan) before dosing at 0, 1, 3, and 6 h, and atdays 1, 2, 3, 7, and 14, after dosing; and in theanimals given 1 or 2.5 mg/kg, readings wererecorded before dosing, and at 0, 1, 3, and 6 hafter dosing. QT and RR intervals were measuredwith an ECG processor (SP-2; Softron, Tokyo,Japan), and the QTc interval normalized for aheart rate of 100 beats per minute (bpm) wascalculated by the logarithmic formula QTc=

log 600×QT/log RR (RR in ms) since it is thebest formula for correcting the QT interval forheart rate in dogs (Matsunaga et al., 1997).

2.3.2. Loading and maintenance dose studyIn the 6-month study with D0870 in dogs, four

sudden deaths out of 14 animals were observed ata daily dose of 5 mg/kg after about 1 month oftreatment (personal communication). The plasmaconcentrations of the compound in the suddendeaths were approximately 29–45 mg/ml. There-fore, we considered that the plasma concentra-tions of D0870 should be maintained between 30and 50 mg/ml to reproduce the sudden deaths. Toshorten the time for obtaining those plasma con-centrations of D0870, we determined 50 mg/kg tobe the loading dose and 5 mg/kg/day as themaintenance dose based on the results of a pre-liminary dog study, in which the plasma concen-trations at 24 h after 50 mg/kg intravenous dosingwere 20–30 mg/ml.

Five male and five female dogs were treatedwith D0870 for 14 days via the cephalic vein byuse of a sterile plastic syringe or an infusion setwith a 25 G butterfly needle. The dosing volumewas 20 ml/kg for the loading dose (50 mg/kg), and2 ml/kg for the maintenance dose (5 mg/kg/day).The rate of infusion was set at approximately 2ml/min. Of the surviving animals during the dos-ing period for 14 days, one male (number 005)and one female (number 504) were dosed continu-ously for a total of 16 or 35 days, respectively.

Heparinized blood samples were collected fromthe cephalic vein at 24 h after each dosing for 14or 16 days, and on days 21, 28, and 35 based onthe result of the single dose study showing thatthe plasma concentration of D0870 was main-tained a plateau from 3 to 24 h after the intra-venous dosing. The blood samplings were alsoperformed on the dead animals. Plasma concen-trations of D0870 were measured with a HPLC.

Ambulatory electrocardiographic recordingwith a Holter recorder (SM-50; Fukuda Denshi,Tokyo, Japan) was conducted in all animals usedin the loading and maintenance dose study beforethe dosing and throughout the dosing period. TheECG of one male (number 005) was recordedcontinuously during a 7-day recovery period after

Fig. 3. Relationship between plasma concentrations of D0870and prolongation rate of QTc intervals in dogs treated withD0870 by a single intravenous infusion or a loading andmaintenance dosing regimen.�, data obtained from the singledose study; �, the loading and maintenance dose study. Thebroken line indicates the 10% prolongation level.


Table 1Mortality and the number of episodes of ventricular tachyarrhythmias in dogs treated intravenously with 50 mg/kg loading and 5mg/kg/day maintenance doses of D0870

Status DayaSex Rate dependentbAnimal number Pause dependentc

nsVT sVT nsVT

Died 4 0Male 1001 0Died 4 1002 1 0

003 Survived 14 1 0 0Died 2 10 1004 2Died W7d 0005 1 0

501Female Died 11 4 Unknowne 0Died 4502 1 1 0Survived 14 0503 0 0

504 Survived 35 0 0 0Died 6 0 1 0505

a The day when the animals died or the treatment period for survivors.b Ventricular tachycardia (VT) elicited in tachycardia. nsVT, non-sustained VT that spontaneously reverted to a sinus rhythm;

sVT, sustained VT preceding ventricular fibrillation.c VT elicited in bradycardia.d This animal was found dead on day 7 of the withdrawal period after a 16-day treatment.e Electrocardiogram just before death could not be recorded since this animal died during the change of the electrodes.

the 16-day treatment. Self-adhesive electrodes(Skintact; Leonhard Lang, Innsbruck, Austria)were used for two animals; however, one animal(number 501) died during the change of the elec-trodes, and the ECG just before the death couldnot be recorded. Therefore, we changed the self-adhesive electrodes to unipolar stimulation heart-wires (TME 60-Z; Dr. Osypka Medizintechnik,Grenzach-Wyhlen, Germany), which have beenused as chronic electrodes, in the remaining eightanimals. Two sets of Holter recorders were pre-pared for each animal, and exchanged every 24 hto obtain the electrocardiographic recordings ascontinuously as possible. The self-adhesive elec-trodes for the MX lead were attached to the skinof the manubrium sterni region of the secondthoracic vertebra (M, − ) and the xiphoid process(X, + ), whereas those for the supplemental leadwere placed on either side of the chest. Theground electrode was placed on the lateral part ofthe abdomen. The unipolar stimulation heart-wires were embedded subcutaneously in the samepositions as the adhesive electrodes under sodiumthiopental (25 mg/kg, i.v.) anesthesia. The adhe-sive electrodes were changed every 2 or 3 days,

and the heart-wires were applied chronically dur-ing the experimental period.

The records were digitized and analyzed withthe ECG processor. QT and QTc intervals weremeasured before the commencement of dosing, at24 h after each dosing, and just before death. Theincidence of VPCs and T-wave alternans for 24 hwas determined from the full disclosures of ambu-latory ECGs.

A full necropsy was performed as soon as pos-sible when the animals were found dead, and anyabnormalities were recorded.

2.4. Statistical analysis

The relations between QTc intervals andplasma concentrations of D0870 in the single dosestudy, and the loading and maintenance dosestudy were assessed by the least-squares regressionmethod. The analyses were carried out with SASsoftware (SAS Institute, NC). A statistically sig-nificant linearity of the regression lines wasdefined at PB0.05. The other parameters werenot subjected to the statistical analyses but evalu-ated individually.


Fig. 4. Changes in individual plasma concentrations of D0870 in male and female dogs treated intravenously with D0870 as loadingdose (50 mg/kg) and maintenance doses (5 mg/kg/day) for 14, 16, or 35 days. Open symbols, the survivors; closed symbols, the dead.Six dead animals had been treated for at most 14 days, and one dead animal for 16 days. Two surviving animals had been treatedfor 14 days, and one surviving animal for 35 days.

3. Results

3.1. Single dose study

Plasma concentrations of D0870 in the animalsgiven 10 or 20 mg/kg decreased steeply for 2 or 3h after dosing, maintained a plateau until 24 h,and then decreased slowly (Fig. 2). The plasmaconcentrations of D0870 in the animals given 1 or

2.5 mg/kg were 0.71 and 3.45 mg/ml, respectively,just after dosing, and were 0.37 and 0.79 mg/ml at6 h.

No QTc prolongation was found in the animalsgiven 1 or 2.5 mg/kg. The QTc intervals just afterthe dosing were prolonged by 23 and 28%, com-pared with their pretreatment values, in the ani-mals given 10 mg/kg, and 59 and 57% in thosethat received 20 mg/kg. The QTc intervals in the


Fig. 5. Changes in individual QTc intervals in male and female dogs treated intravenously with D0870 as loading dose (50 mg/kg)and maintenance doses (5 mg/kg/day) for 14, 16, or 35 days. Open symbols, the survivors; closed symbols, the dead. Six deadanimals had been treated for at most 14 days, and one dead animal for 16 days. Two surviving animals had been treated for 14 days,and one surviving animal for 35 days.

animals given 10 mg/kg returned to their pretreat-ment levels within 3 or 6 h, whereas the animalsgiven 20 mg/kg showed a slight prolongation ofQTc interval even at 7 days after dosing (Fig. 2).The plasma concentrations of the compound inthe animals given 10 mg/kg were 4.51 and 4.57mg/ml at 3 h after dosing, and those in the animalsthat had received 20 mg/kg were 5.55 and 3.16

mg/ml at 7 days after dosing. The prolongationrate of QTc intervals (Y%) was proportionate tothe plasma concentrations of D0870 (X mg/ml),and the correlation formula between the twoparameters was Y=2.61X, giving a correlationcoefficient, r, of 0.911 (Fig. 3). The thresholdplasma concentration of D0870 causing 10% QTcprolongation, a major diagnostic criterion for


Fig. 6. Changes in individual incidence of VPCs in male and female dogs treated intravenously with D0870 as loading dose (50mg/kg) and maintenance doses (5 mg/kg/day) for 14, 16, or 35 days. Open symbols, the survivors; closed symbols, the dead. Six deadanimals had been treated for at most 14 days, and one dead animal for 16 days. Two surviving animals had been treated for 14 days,and one surviving animal for 35 days.

LQTS (Schwartz, 1985), was estimated to be 3.8mg/ml (Fig. 3).

3.2. Loading and maintenance dose study

3.2.1. MortalityA total of seven animals died during the treat-

ment or withdrawal period: one male (number004) on day 2, two males (numbers 001 and 002)and one female (number 502) on day 4, onefemale (number 505) on day 6, one female (num-ber 501) on day 11 of the treatment, and onemale (number 005) on day 7 of the withdrawalperiod after a 16-day treatment (Table 1). The


remaining animals, one male (number 003) and twofemales (numbers 503 and 504) survived during theexperimental period for 14 or 35 days.

As syndrome just before death, three animals(numbers 001, 501, and 502) showed avoidance orfear-like behavior toward venipuncture and re-straint for drug dosing and blood sampling, andsuddenly convulsed and died during the dosing orthe change of the adhesive electrodes. One male(number 002) convulsed suddenly in the cage dur-ing cleaning of the animal room, and died soonthereafter. The symptoms just before the death ofthe other three animals (numbers 004, 005, and505) could not be observed, because they werefound dead.

3.2.2. Plasma concentrations of D0870Plasma concentrations of D0870 after the load-

ing dose (50 mg/kg) were 20.9–28.4 mg/ml, andthen continued to increase as a result of themaintenance doses (5 mg/kg/day). D0870 levels inthe plasma obtained from the dead animals were19.7–36.4 mg/ml, whereas those in the survivors at24 h after the final dose were 33.9–52.6 mg/ml. Theplasma concentration–time curves of the dead andsurviving animals were comparable (Fig. 4).

3.2.3. Ambulatory electrocardiographic findingsA marked prolongation of QT and QTc intervals

was noted in all of the 10 animals used for thestudy. The QTc intervals started to lengthen duringthe administration of the loading dose, and main-tained approximately the same values from thesecond or third maintenance dose until the end ofthe experimental period (Fig. 5). The QTc intervalsat 24 h after the loading dose produced 48–84%prolongation compared with the pretreatment val-ues, and those after the final dose or before deathexhibited 52–96% prolongation. The QTc intervalsof the dead and surviving animals were comparable(Fig. 5). A linear relation between plasma concen-trations of D0870 and prolongation rate of QTcintervals was also seen in the loading and mainte-nance dose study; however, the slope of the regres-sion line and the correlation coefficient value wereapparently smaller compared with those of thesingle dose study (Fig. 3).

According to the detailed analysis of the full

disclosures, alternans of T-waves in polarity or inamplitude, morphological changes in T-waves, andVPCs were observed in all of the animals during thedosing and withdrawal periods. The animals witha high incidence of VPCs (numbers 004 and 502)died within a short treatment period (Fig. 6);however, there was no relation between the inci-dence of T-wave alternans and the mortality (Fig.7). The morphology of T-waves was altered fre-quently; the negative component of the T-wavestended to increase in association with the occur-rence of bradycardia, tachycardia, and T-wavealternans; and VPCs were elicited from the termi-nal portion of the enhanced negative T-waves(Figs. 8 and 9). The T-wave alternans was observedin tachycardia, especially when the heart rateincreased suddenly from the normal or slow rate.The VPCs were observed mainly during tachycar-dia before and after dosing, and were also seen inbradycardia early in the morning. T-wave alternansand VPCs observed in tachycardia often proceededto ventricular bigeminy or trigeminy. On the otherhand, VPCs elicited in bradycardia were observedsolitarily. Ventricular tachycardias (VTs) were ob-served in all males and in three females (numbers501, 502, and 505), and these animals, except forone male (number 003), all died.

ECGs just before death could be recorded in sixanimals (numbers 001, 002, 004, 005, 502, and 505).All of the VTs observed before death showed atwisting morphology of QRS complexes aroundthe imaginary baseline, and proceeded to VF. Sinceone animal (number 501) died during the change ofthe adhesive electrodes on day 11, the ECG justbefore death could not be recorded; however,T-wave alternans and ventricular trigeminy wereobservable at a few minutes before the death. Thesearrhythmias were also observed in this animalwhen VTs were elicited on days 4 and 5.

The VTs observed in this study were elicitedby R on T type VPC or by T-wave alternans(Figs. 8 and 9), reached in total 25 times, andwere classified into two types (Table 1). One wasrate-dependent VT elicited in tachycardia or aftera sudden increase in heart rate: some VTs ofthis kind did not stop but proceeded to VF (sixepisodes in six animals; numbers 001, 002, 004, 005,502, and 505), whereas the remaining ones


Fig. 7. Changes in individual incidence of T-wave alternans in male and female dogs treated intravenously with D0870 as loadingdose (50 mg/kg) and maintenance doses (5 mg/kg/day) for 14, 16, or 35 days. Open symbols, the survivors; closed symbols, the dead.Six dead animals had been treated for at most 14 days, and one dead animal for 16 days. Two surviving animals had been treatedfor 14 days, and one surviving animal for 35 days.

spontaneously reverted to a sinus rhythm within afew seconds (17 episodes in five animals; numbers002, 003, 004, 501, and 502). Some of the non-sus-tained VTs were polymorphic, whereas the twist-ing morphology was unclear because of theirshort duration. The other was pause-dependentVT elicited in severe bradycardia, which also

showed a twisting morphology of QRS com-plexes, and spontaneously reverted to a sinusrhythm within 10 s (two episodes in one animal;number 004). The pause-dependent VTs were elic-ited after a long pause, and showed a typicallong–short sequence and a large TU complex-likewave form at the initiation (Fig. 8d,e).


Fig. 8. ECG patterns (MX lead) obtained from a male dog (number 004) treated intravenously with D0870 as loading dose (50mg/kg) and maintenance doses (5 mg/kg/day). (a) Normal ECG pattern with respiratory sinus arrhythmia (pre-dose). (b) T-wavealternans in tachycardia (on day 2). (c) Rate-dependent ventricular tachycardia elicited after a sudden increase in heart rate fromthe normal heart rate (on day 2). (d,e) Pause-dependent torsades de pointes (TdP) elicited in severe bradycardia (on day 2). (f) VPCsand rate-dependent TdP preceding VF observed just before death (on day 2).

3.2.4. Necropsy findingsNo abnormalities were detected in the dead

animals except in two females, one showing anenlargement of the adrenal cortex and the other, ahemorrhage in the thymus.

4. Discussion

To elucidate the cause of the sudden death indogs treated intravenously with D0870, an anti-fungal agent, we first conducted a single dosestudy in dogs to evaluate the relation betweenplasma concentrations of D0870 and QT prolon-gation. We found a linear relation between QTcprolongation rate and plasma concentration ofD0870, and estimated the threshold causing 10%QTc prolongation to be 3.8 mg/ml. This concen-tration is sufficiently higher than the 50% antibi-otic concentration (0.0015 mg/ml) of D0870against Candida albicans strains (Yamada et al.,1993). Then, based on this result and the resultsobtained from the 6-month toxicity study per-

formed previously, we designed a loading andmaintenance dose study in dogs to reproduce,with high frequency, the sudden death induced byD0870. The study revealed that D0870 inducedmarked QT prolongation, VPCs, and T-wave al-ternans as changes in ECGs in all 10 animalsused. These characteristic electrocardiographic al-terations coincided with those in the LQTS(Schwartz et al., 1995). In addition, the suddendeaths were observed in seven of 10 animals, i.e.at an incidence of 70%. Some of the dead animalsshowed avoidance or fear-like behavior towardthe experimental handling and suddenly con-vulsed and died. The ECGs before the episodes ofsudden death were successfully recorded in sixanimals, and all of them indicated VTs with twist-ing morphology of QRS complexes. The VTs wereelicited in both tachycardia and bradycardia, andsome of them in the former degenerated into VF.

The remaining three animals (survivors) showedthe same extent of QT prolongation as the deadanimals. In addition, they were similar to the deadanimals in alterations of ECGs. These findings


Fig. 9. ECG patterns (MX lead) obtained from a female dog (number 505) treated intravenously with D0870 as loading dose (50mg/kg) and maintenance doses (5 mg/kg/day). (a) Normal ECG pattern with respiratory sinus arrhythmia (pre-dose). (b) Gradualincrease in negative component of T-waves accompanied by increasing heart rate during respiratory sinus arrhythmia (on day 5). (c)T-wave alternans and ventricular bigeminy in tachycardia (on day 6). (d) Bradycardia, sudden increase in heart rate, T-wavealternans, and rate-dependent TdP preceding VF observed just before death (on day 6).

suggest that QT prolongation plus certain factorsare necessary to cause the sudden death in dogstreated with D0870. However, the sudden deathsoccurred in animals with high or low incidence ofVPCs. This finding seems to indicate that thedevelopment of VPCs is not an indispensablecondition, but augments the probability of theoccurrence of sudden death.

All of the ECGs recorded just before deathshowed VTs with twisting morphology of QRScomplexes around the imaginary baseline. Thetwisting pattern, however, was also seen in thenon-sustained VTs that spontaneously reverted toa sinus rhythm. We regard most of the VTsobserved in this study as TdPs, since the twistingmorphology and QT prolongation are majorcharacteristics of TdP (Haverkamp et al., 1995). Ithas been reported that alternation of the T-waveis commonly induced during emotional or physi-cal stress in the congenital LQTS, and may pre-cede the occurrence of TdP (Pernot et al., 1972;Schwartz and Malliani, 1975). We also observedT-wave alternans during tachycardia or after asudden increase in heart rate, and TdP elicited byT-wave alternans in this study. These characteris-tics of ECGs and the symptoms observed before

death suggest that D0870-induced LQTS resem-bles the congenital one. Jackman et al. (1988)proposed that LQTS could be classified intoadrenergic-dependent type, pause-dependent type,and intermediate type with respect to the cause ofTdP. In the adrenergic-dependent LQTS, TdP isoften associated with an increase in cate-cholamines by stress, fright, etc., and is sup-pressed by b-adrenergic blocking agents, whereasTdP in the pause-dependent LQTS usually occursduring bradycardia. Although drug-inducedLQTS is generally classified into the pause-depen-dent type (Jackman et al., 1988), we propose thatthe D0870-induced LQTS should be classified intothe intermediate LQTS with characteristics ofboth above-mentioned types, since the TdPs wereelicited in both bradycardia and tachycardia inthis study. Most of the TdPs were elicited intachycardia, and some TdPs provoked in tachy-cardia did not disappear but proceeded to VF. Onthe other hand, the TdPs elicited in bradycardiaspontaneously reverted to a sinus rhythm. VPCswere also observed mainly in tachycardia andpartially in bradycardia; and those in the formeroften proceeded to ventricular bigeminy ortrigeminy, whereas VPCs in the latter were ob-


served solitarily. The differences in incidence andconsequence of these arrhythmias may be depen-dent on the extent of autonomic nervous tone,since it has been reported that the sympatheticnervous system promotes cardiac arrhythmias, es-pecially malignant ventricular arrhythmias, suchas VT and VF, whereas the parasympathetic ner-vous system plays a protective role (Schwartz andVanoli, 1981). Accordingly, we propose that thesudden death due to TdPs in D0870-treated dogsmay be provoked by QT prolongation plus animbalance between the sympathetic nervous toneand the parasympathetic nervous tone, namely inthe case of predominance of the former. Themechanism of initiation of TdP induced by D0870remains a topic for further study.

In conclusion, we clarified that D0870-inducedsudden death in dogs is primarily attributable toTdP associated with QT prolongation and thatthe sympathetic nervous tone collaboratively leadsto the occurrence of the sudden death. Thus, wepropose the D0870-treated conscious dog as anew experimental model of TdP. This model maybe suitable for studies on the mechanisms ofinitiation of TdP and on the treatment of it.

References

Haverkamp, W., Shenasa, M., Borggrrefe, M., Breithardt,G., 1995. Torsades de pointes. In: Zipes, D.P., Jalife, J.(Eds.), Cardiac Electrocardiology: From Cell to Bedside,2nd edn. WP Saunders, Philadelphia, pp. 885–899.

Jackman, W.M., Friday, K.J., Jerome, L.A., Aliot, E.M.,Clark, M., Lazzara, R., 1988. The long QT syndrome: acritical review, new clinical observations and a unifyinghypothesis. Prog. Cardiovasc. Dis. 31, 115–172.

Matsunaga, T., Mitsui, T., Harada, T., Inokuma, M., Mu-rano, H., Shibutani, Y., 1997. QT corrected for heartrate and relation between QT and RR intervals in beagledogs. J. Pharmacol. Toxicol. Meth. 38, 201–209.

Napolitano, C., Priori, S.G., Schwartz, P.J., 1994. Torsadesde pointes: mechanism and management. Drugs 1, 51–65.

Pernot, C., Henry, M., Aigle, J.C., 1972. Syndrome car-dioauditif de Jarvell et torsades de pointes. Arch. Mal.Coeur Vaiss. 65, 261–274.

Schwartz, P.J., 1985. Idiopathic long QT syndrome: progressand question. Am. Heart J. 2, 399–411.

Schwartz, P.J., Malliani, A., 1975. Electrical alternation ofthe T-wave: Clinical and experimental evidence of its re-lationship with the sympathetic nervous system and withthe long QT syndrome. Am. Heart J. 89, 45–50.

Schwartz, P.J., Vanoli, E., 1981. Cardiac arrhythmias elicitedby interaction between acute myocardial ischemia andsympathetic hyperactivity: A new experimental model forthe study of antiarrhythmic drugs. J. Cardiovasc. Phar-macol. 3, 1251–1259.

Schwartz, P.J., Locati, E.H., Napolitano, C., Priori, S.G.,1995. The long QT syndrome. In: Zipes, D.P., Jalife, J.(Eds.), Cardiac Electrocardiology: From Cell to Bedside,2nd edn. WP Saunders, Philadelphia, pp. 788–811.

Symanski, J.D., Gettes, L.S., 1993. Drug effects on theECG: a review of their clinical importance. Drugs 46,219–248.

Weissenburger, J., Davy, J.-M., Chezalviel, F., 1993. Experi-mental models of torsades de pointes. Fundam. Clin.Pharmacol. 7, 29–38.

Yamada, H., Tsuda, T., Watanabe, T., Ohashi, M., Mu-rakami, K., Mochizuki, H., 1993. In vitro and in vivoantifungal activities of D0870, a new triazole agent. An-timicrob. Agents Chemother. 37, 2412–2417.

.

Documents

QT prolongation and torsades de pointes induced by an antifungal agent, D0870, in conscious dogs