Terbina®ne: tolerability in general medical practice

D.P.O'SULLIVAN

Novartis Pharmaceuticals UK Ltd, Frimley Business Park, Frimley, Surrey, GU16 5SG, UK.

E-mail: des.osullivan@pharma.Novartis.com

Summary Four open, prospective, postmarketing studies with consistent protocols were undertaken, involvingover 25,000 patients and nearly 3600 physicians in four European countries (UK, Netherlands,

Germany and Austria). The purpose was to determine the tolerability of terbina®ne when prescribed

to a large, unselected patient population in general medical practice. Patients were recruited fromdermatology departments and from general and family practitioners. All received at least one dose of

terbina®ne. No speci®c exclusion criteria were applied. The only treatment instructions provided

were those contained in the manufacturer's product information. Patients were monitored foradverse events at baseline, during treatment, and at the end of treatment. Of the 25,884 patients

entering the study, 38´6% (9991) had concomitant disease, 42´8% (11,078) were taking other

medications, and 22´7% (5876) were older than 60 years. The predominant indication forprescribing terbina®ne was onychomycosis. Mean duration of treatment was 13´2 weeks. No adverse

events were reported in 89´5% (23,167) of patients. The remaining 10´5% (2717) reported one ormore adverse events, primarily gastrointestinal (4´9%) or dermatological (2´3%). A total of 115

serious adverse events were recorded, but of these only four were probably and eight possibly related

to terbina®ne. There were no reported drug±drug interactions, even in patients receiving con-comitant medications metabolized by cytochrome P-450 enzymes. There were also no clinically

signi®cant drug±disease interactions. In conclusion, terbina®ne is well tolerated in an unselected

general practice population. This con®rms its good tolerability previously established duringcontrolled clinical trials.

Terbina®ne is a member of the allylamine class ofsynthetic antifungal agents, and is thus both chemically

and mechanistically distinct from other available anti-

mycotics.1 In vitro, terbina®ne exhibits fungicidal activ-ity against dermatophytes, moulds, dimorphic fungi,

and the yeast Candida parapsilosis, based on its ability

to inhibit the biosynthesis of ergosterol at the level ofsqualene epoxidation (Fig. 1). Because of its high selec-

tivity for fungal squalene epoxidase1 and its lack of

af®nity for cytochrome P-450 enzymes, terbina®necan be expected to have a low potential for drug±drug

interactions.

To date, controlled clinical studies with terbina®nehave established a highly favourable tolerability pro®le,

with the great majority of adverse events (primarily

gastrointestinal and dermatological) being both mildand transient. However, the use of exclusion criteria

(concomitant medications, other diseases and age)

raises the issue whether these data can, by extrapola-tion, provide reliable evidence of good tolerability in the

general population. The purpose of this study was to

broaden the tolerability database of terbina®ne bydetermining the incidence of adverse events accompa-

nying its use in general medical practice.

Subjects and methods

Four open, prospective, postmarketing, observationalcohort studies of similar design were undertaken, invol-

ving a total of 25,884 patients and 3591 physicians

(dermatologists and general and family practitioners).The ®rst study commenced in the UK in 1991, and has

been reported separately;2 the other three studies com-

menced in 1993 in the Netherlands, Germany andAustria. All studies were completed by 1995, with the

UK and Germany providing some 10,000 patients each,

Austria and the Netherlands some 2500 patients each.The four study protocols were intentionally consistent,

allowing the data to be pooled and analysed collectively,

and making this the largest tolerability study everundertaken with an oral antifungal agent. The overall

results have been previously published.3

British Journal of Dermatology 1999; 141 (Suppl. 56): 21±25.

21q 1999 British Association of Dermatologists

Physicians were invited to participate by letter, and

were advised that any patient receiving a prescriptionfor terbina®ne was eligible for entry ± there were no

speci®c exclusion criteria. The only information provided

to physicians was the product information packagerelevant to their country. Patient management remained

the responsibility of the individual physician, with no

protocol stipulations as to the frequency of consultation,laboratory investigations and length of treatment.

Of the 25,884 patients enrolled, 54´5% were male,

45´4% female (0´1% unknown). Mean age was47´8 years (range 1±98), with 22´7% (n�5890) over

60 years of age. Patients were primarily caucasian

(93´8%), with smaller numbers of Asian (3´4%), black

(1´0%), oriental (0´8%) and other/unknown (1´0%).Reasons for prescribing terbina®ne are shown in Table 1.

Mean duration of treatment was 13.2 weeks, with a

range of 1 day to 3 years; 76% of patients receivedterbina®ne for more than 6 weeks and 59.3% for at

least 12 weeks. A total of 99% of patients received

terbina®ne at the recommended dose of 250 mg/day.Total exposure to terbina®ne amounted to 6350 patient

years.At enrolment each patient provided basic demo-

graphic data and a detailed medical history, particularly

concomitant medications and diseases, previous infec-tions and allergies. At the same time the reason for

prescribing terbina®ne was noted. This assessment

(excluding the demographic data) was repeated at theend of treatment; at 6 weeks if treatment were to be

continued after that timepoint; and at the time of any

reported adverse event.As would be expected in a large, unselected popula-

tion of this type, concomitant medications and diseases

22 D.P.O'SULLIVAN

q 1999 British Association of Dermatologists, British Journal of Dermatology, 141 (Suppl. 56): 21±25

α

Squalene

Squalene epoxidase Terbinafineblocks here

Squalene epoxide

Lanosterol

Ergosterol

14- -demethylase Triazolesblock here

Figure 1. Mechanism of action of terbina®ne, an allylamine, comparedto triazoles such as ¯uconazole and itraconazole. Terbina®ne inhibits

the non-cytochrome P-450 enzyme squalene epoxidase, leading to a

fungicidal accumulation of intracellular squalene. By contrast, tria-

zoles inhibit the cytochrome P-450 mediated enzyme 14-a-demethy-lase, preventing conversion of lanosterol to ergosterol. These

mechanisms suggest that terbina®ne is likely to have a lower level of

drug±drug interactions.

Table 1. Reasons for prescribing terbina®ne in patients enrolled in the

study. Some patients were diagnosed as having more than one

condition

% n

Clinical diagnosis

Onychomycosis 72´2 18,698Palmar or plantar mycosis 23´7 6135

Tinea corporis 13´7 3546

Interdigital mycosis 6´9 1786

Other indications 6´6 1702

Table 2. Main concomitant diseases (classi®ed by organ system or

condition) and medications declared by the 25 884 patients enrolled

in the study. Some patients declared more than one disease and/orwere receiving more than one medication

Patients (%)

Disease/disorderNone 61´4

General cardiovascular 11´0

Musculoskeletal 5´9Metabolic/nutritional 5´7

Skin and appendages 5´7

Whole body 5´2

Gastrointestinal 4´0Respiratory 4´0

Heart and valvular disorders 3´5

Diabetes mellitus 3´4

Extracardiac vascular 3´0Psychiatric 2´9

Central/peripheral nervous system 2´0

Medications

None 57´2

Cardiovascular 14´5Alimentary tract and metabolism 7´6

Musculoskeletal (incl. NSAIDs) 7´1

Dermatological 6´4

Central nervous system 6´4Oral contraceptives 5´6

Respiratory 5´2

Benzodiazepines 3´7Oral antidiabetic agents 2´4

were frequent, and strongly in¯uenced by age. In all,38´6% (n�9991) had a concomitant disease, and

42´8% (n�11,078) were taking other medications.

For those aged over 60 years (n�5890), the propor-tions were even higher at 65´2 and 65´8% of the

population, respectively. The main diseases and medica-

tions involved are shown in Table 2. These data areimportant because of the potential for drug±drug and

drug±disease interactions.

The de®nition of adverse events was intentionallybroad. Physicians were asked to report all adverse

events occurring during terbina®ne therapy ± including

intercurrent illness and elective surgery ± irrespective ofwhether they appeared to relate to terbina®ne. Informa-

tion was recorded under three headings: the adverse

event (and whether reported by the patient or observedby the physician); time of onset, duration and severity;

causal relationship and outcome. Serious adverse events

were reported immediately to the relevant authoritiesand to Novartis. Precise frequency estimates were cal-

culated, and permitted detection of rare adverse events

occurring at a rate of 4 per 100,000 population.

Results

A total of 89´5% (23,167/25,884) reported no adverse

event whatsoever. The remaining 10´5% (2717/

25,884) reported one or more adverse events, primarilygastrointestinal (4´9%) or dermatological (2´3%)

(Table 3).

Causality

Causality was determined by the physician. Data, again

divided by organ system, are shown in Table 3. Overall,

55´9% of adverse events were considered to be probablyor possibly related to terbina®ne, but with a wide

variation in causality across body systems.

Discontinuation

A total of 1370 patients (5´3%) discontinued terbina-

®ne therapy because of adverse events. As would beexpected, the frequency of discontinuation related to the

overall incidence of adverse events by organ system

(Table 3); thus 695 patients discontinued therapybecause of gastrointestinal events, 373 because of der-

matological events. However, 94´7% of patients

remained on terbina®ne therapy; the 89´5% whoexperienced no adverse events, plus the 5´2% who

suffered adverse events but continued with terbina®ne

therapy.

Gastrointestinal

These were primarily nausea (1´3%), diarrhoea (0´8%),abdominal pain (0´8%) and dyspepsia (0´6%); other

events accounted for 1´4%. Almost all were mild

and transient, and reversible on discontinuation oftreatment.

Dermatological

These were primarily rash (0´9%), pruritus (0´3%),

urticaria (0´2%), and eczema (0´2%); other events

accounted for 0´6%. The majority of rashes occurredwithin the ®rst 2 weeks, and were mild or moderate.

Other events were generally mild and transient.

TERBINAFINE TOLERABILITY 23

q 1999 British Association of Dermatologists, British Journal of Dermatology, 141 (Suppl. 56): 21±25

Causal relationship

Adverse event Probable/ Unlikely/ NotOrgan system (% of patients) possible not related assessed

Gastrointestinal 4´9 76´9 19´1 3´9

Dermatological 2´3 62´4 32´1 5´5

Whole body 1´5 33´8 59´6 6´6CNS/PNS 1´2 64´0 27´3 8´7

Respiratory 1´0 8´2 88´1 3´7

Musculoskeletal 0´8 24´0 72´1 4´9

Special senses 0´7 94´1 2´7 2´2Liver and biliary 0´2 72´4 19´0 8´6

Table 3. Frequency ($ 0´2%) and causality of

adverse events by organ system. Note thewide variation across organ system in a total

of 25 884 patients. CNS� central nervous

system; PNS� peripheral nervous system

Hepatobiliary

A total of 110 patients (0´4%) had a known history ofhepatobiliary disease at entry. Against this background,

0´2% of the study population (n�55) experienced

hepatobiliary adverse events. Of these, 45 were asymp-tomatic increases in hepatic enzymes which reverted to

normal on discontinuation of terbina®ne. In the opinion

of the physician, 38 of these were probably or possiblyrelated to terbina®ne. The other 10 cases were sympto-

matic; ®ve patients presented with elevated hepatic

enzymes and non-speci®c symptoms such as tiredness,diarrhoea and red palms (four cases probably/possibly

related to terbina®ne); while ®ve patients had chole-

static hepatic dysfunction (two cases probably/possiblyrelated to terbina®ne).

Taste disturbance

Taste disturbance or loss was seen in 0´7% of patients

(n�186). Taste disturbance (metallic, foul or alteredtaste) tended to occur within the ®rst week of therapy

(0´4%); taste loss occurred predominantly between

weeks 4 and 8 (mode 7 weeks, range 1±11) in 0´3%of cases. Most cases (94%) were considered to be

probably due to terbina®ne. Median time to recovery

on cessation of terbina®ne was 42 days. The proportionof females affected was signi®cantly greater, P<0´016.

Serious adverse events

In total, 115 serious adverse events were recorded(Table 4), of which only four were probably and eight

possibly related to terbina®ne, in the opinion of the

reporting physician. Thus, only 0´45% of all patientsreceiving terbina®ne suffered a serious adverse event,

and only 0´046% suffered a serious adverse event

probably or possibly attributable to terbina®ne. In fact,most events occurring in this category were consistent

with those expected to occur in any comparable popula-

tion over the same time span.There were 24 deaths, giving an overall mortality of

0´09%. Median age at death was 72 years (range 40±

94). None of the deaths was attributable to terbina®ne,and the causes were within the range expected for this

type of population.

Drug±drug interactions

In total, 42´8% of the study population were receivingone or more concomitant medications (Table 2) while

taking terbina®ne. The agents concerned ± oral contra-

ceptives, warfarin, digoxin, antidiabetics, antihista-mines and H2-receptor antagonists ± comprise many

of those known to react with other classes of antifungal

agent, such as triazoles.

Drug±disease interactions

Concomitant diseases were present in 38´6% of the

study population (Table 2), but the incidence of adverse

events remained at levels typically seen in clinical trials,where patients with concomitant diseases are

excluded.5 There is thus no evidence from this study

of clinically signi®cant drug±disease interactions.

Discussion

The adverse events reported in this prospective observa-

tional cohort study mirror those observed in controlledclinical trials. There is no evidence that the wide range

of age, concomitant diseases and medications had anyeffect on the incidence of adverse events. No serious or

life-threatening conditions were attributed to terbina-

®ne therapy. Serious events such as idiosyncratic hepa-tic dysfunction occur very rarely, and appear to be

reversible with prompt discontinuation of therapy. No

previously unrecognized risks were identi®ed.In vitro studies have shown that terbina®ne has

minimal potential to inhibit or induce the clearance of

24 D.P.O'SULLIVAN

q 1999 British Association of Dermatologists, British Journal of Dermatology, 141 (Suppl. 56): 21±25

Table 4. Serious adverse events by organ system, and those probably orpossibly related to terbina®ne therapy out of a total of 25 884 patients.

CNS� central nervous system; PNS� peripheral nervous system

Cases causally

Serious related to terbina®ne

adverse

events (n) Probably Possibly Not

Cardiovascular 33 0 0 33

Whole body 21 1 1 19

Neoplasia 14 0 0 14

Gastrointestinal 13 2 0 11Respiratory 7 1 0 6

Urinary 6 0 0 6

Musculoskeletal 6 0 0 6

Haematological 5 3 0 2Skin 2 0 2 0

CNS/PNS 2 0 0 2

Hepatobiliary 2 1 1 0

Psychiatric 2 0 0 2Hearing 1 0 0 1

Reproductive 1 0 0 1

Total 115 8 4 103

cyclosporin, tolbutamide or oral contraceptives.4 Theabsence of any reported drug±drug interactions

strongly supports the view that terbina®ne can be

safely prescribed in the community to patients receivingthese medications. The low potential for interaction

between terbina®ne and the cytochrome P-450

enzyme system (Fig. 1) also suggests that clinicallysigni®cant interactions between terbina®ne and agents

metabolised by these enzymes will be few.

In summary, terbina®ne has been shown to be welltolerated in an unselected general practice population,

con®rming the positive tolerability pro®le established

during controlled clinical trials.

Chairman's overview

The studies reported here, including the L.I.ON. Study,

have shown that treatment with terbina®ne is asso-ciated with a very low incidence of adverse events. In

addition, the open, prospective, postmarketing studies in

nearly 26,000 patients (reported by Dr O'Sullivan) haveestablished the excellent tolerability of terbina®ne in a

large, unselected population. It is particularly signi®-

cant that these postmarketing studies have found no

evidence of drug-drug interactions, even in patientsreceiving concomitant medications metabolised by cyto-

chrome P-450 enzymes, nor of any clinically signi®cant

drug-disease interactions.

Con¯ict of interest: Dr O'Sullivan is an employee of

Novartis Pharmaceuticals UK Ltd.

References

1 Ryder NS, Favre B. Antifungal activity and mechanism of action of

terbina®ne. Rev Contemp Pharmacother 1997; 8: 275±87.

2 O'Sullivan DP, Needham CA, Bangs A et al. Postmarketing surveil-lance of oral terbina®ne in the UK. Br J Clin Pharmacol 1996; 42:

559±65.

3 Hall M, Monka C, Krupp P et al. Safety of oral terbina®ne. Results of

a postmarketing surveillance in 25,884 patients. Arch Dermatol1997; 133: 1213±19.

4 Back DK, Stevenson P, Tija JF. Comparative effects of two antimy-

cotic agents, ketoconazole and terbina®ne, on the metabolism of

tolbutamide, ethinyloestradiol, cyclosporin and ethoxycoumarin byhuman liver microsomes in vitro. Br J Clin Pharmacol 1989; 28:

166±70.

5 Villars VV, Jones TC. Special features of the clinical use of oralterbina®ne in the treatment of fungal diseases. Br J Dermatol 1992;

126 (Suppl. 39): 61±9.

TERBINAFINE TOLERABILITY 25

q 1999 British Association of Dermatologists, British Journal of Dermatology, 141 (Suppl. 56): 21±25