Tariot PN, Solomon PR, Morris JC, et al. and The GalantamineUSA-10 Study Group. 2000. A 5-month, randomized, placebo-controlled trial of galantamine in AD. Neurology 54: 2269–2276.

ROGER BULLOCK

Kingshill Research Centre, Victoria HospitalSwindon, UK

LUC TRUYEN

Johnson & Johnson Pharmaceutical Research &Development, LLC, Titusville, NJ, USA

Published online in Wiley InterScience(www.interscience.wiley.com).

DOI: 10.1002/gps.1233

Response to letter by Bullock and Truyen

Dear Editor

We share the opinion of Drs Bullock and Truyenregarding the importance of head-to-head trials andthe critical evaluation of study designs and methods.As it is generally inappropriate to compare data fromdifferent clinical trials without considering differ-ences in study design and patient population, theJones et al. (2004) study was designed and carriedout to directly compare donepezil and galantaminein the same patient population, in a randomized fash-ion. Many factors cause variations between patientpopulations, even in studies of identical design, e.g.entry criteria, location of study sites (national/interna-tional), health and cognitive status of participants, andpatient and caregiver motivation and expectations.Improvements in cognition have been consistentlyobserved with donepezil treatment in clinical trials;the comment of Drs Bullock and Truyen regardingthe differences in magnitude of effect between thisstudy and the findings in pivotal studies emphasisesthe limitations of such comparisons, and the needfor randomised head-to-head trials.

In this 12-week study, significantly greater benefitswere seen for donepezil compared with galantamine atendpoint on the primary (Physician’s and Caregiver’sSatisfaction Questionnaires) and all secondary out-come measures (blinded assessments of cognitionusing the Alzheimer’s Disease Assessment-cognitivesubscale [ADAS-cog] and Mini-Mental State Exami-nation [MMSE], and a measure of functional ability,the Disability Assessment for Dementia [DAD] scale).In contrast, in the 52-week comparator studydescribed by Drs Bullock and Truyen (Wilcock et al.,

2003), both galantamine and donepezil provided ben-efits, with no significant differences between the twotreatments on the primary outcome measure (BristolActivities of Daily Living) throughout the study. Therewere also no significant differences between treat-ments in the total population in MMSE scores or 11-item ADAS-cog scores throughout the study, or in thenumber of patients remaining at or above baseline onthe 11-item ADAS-cog at Week 52.

Both short- and long-term clinical trials are impor-tant in evaluating the benefits of symptomatic treat-ments for Alzheimer’s disease (AD), although theyare designed to address different treatment questions.Short-term studies generally assist physicians andpatients in making decisions regarding initial efficacy,tolerability, and compliance, while longer term stu-dies assess the effectiveness and safety of mainte-nance therapy. Both are valid and desirable ifdesigned appropriately and reported accurately,although the objectives and outcomes differ. Theprimary objective of our study was to directly com-pare the ease of use and tolerability of donepeziland galantamine in the first three months of treatment.To meet this objective, the pre-specified outcomemeasures were the Physician’s and Caregiver’s Satis-faction Questionnaires, which capture many aspectsof treatment that are relevant and meaningful for doc-tors, patients, and caregivers. Included are itemsaddressing tolerability and the ease with which care-givers can follow treatment directions, in addition tosatisfaction with dosing frequency and titration sche-dule. In this study, all items favoured donepezil, notonly those assessing titration and dosing frequency.Furthermore, although Drs Bullock and Truyen are

letters to the editor 87

Copyright # 2004 John Wiley & Sons, Ltd. Int J Geriatr Psychiatry 2005; 20: 85–89.

correct that the questionnaires have not been formallyvalidated, their sensitivity and utility in detecting dif-ferences between treatments have been reported in aprevious head-to-head study of donepezil and rivas-tigmine, a study of which Dr Bullock was a co-author(Wilkinson et al., 2002). In addition, differences insatisfaction and ease of use of the two agents wereconsistent with the differences found in tolerability.It should be noted that patients in these two studieswere naı̈ve to cholinesterase inhibitor treatment;therefore, caregivers reported only on the treatmentpatients were receiving, and were not judging currenttreatment against previous experience with anothertreatment.

We agree that caregiver burden is an important con-sideration when assessing AD treatments. However,assessment of caregiver burden was not one of theobjectives of this study, and therefore no conclusionscan be drawn with regard to this outcome, despite theclaim that the study by Wilcock et al. (2003) demon-strated advantages for galantamine at 13 weeks on ameasure of caregiver burden. In that publication, theauthors note that the ‘number of caregivers was notwell recorded and many were not included in the ana-lysis’ (Wilcock et al., 2003; p. 784).

The description of the dose titration in our study asa ‘forced dose escalation’ is inaccurate and mislead-ing. Dosing of both donepezil and galantamine wascarried out according to the guidelines in the respec-tive approved product labelling, with 4 weeksbetween each dose escalation. Dosing was fully flex-ible; patients unable to tolerate a particular dose ofeither drug were able to have their dose reduced. Sub-sequently, the dose could then be increased again,with the intent to reach and maintain the highestrecommended dose. Patients were allowed to changedosing on unscheduled visits for full flexibility. Thesefactors ensured that the study reflected dosing ineveryday clinical practice as closely as possible.The differences in titration schedules between done-pezil and galantamine were not due to the studydesign, but are inherent characteristics of the twotreatments. Donepezil treatment begins at a clinicallyeffective dose whereas galantamine requires a four-week titration to reach a clinically effective dose.These differences may be important considerationsfor physicians when initiating treatment. Similarly,use of the recommended titrations and fully flexibledosing did not increase the likelihood of adverseeffects during titration; the tolerability differencesbetween the two treatments during the recommendedtitration period are again inherent. Twice as manypatients receiving galantamine had to reduce their

dose or temporarily discontinue medication due toAEs vs donepezil (21.4% vs 9.4%, respectively).

Data from a pivotal galantamine study did not showsignificant differences in efficacy between 16 and24mg/day (Tariot et al., 2000). As a result of usingthe recommended galantamine dose titration schedulein our study, patients were able to receive galantamine24mg/day for a maximum of four weeks only; how-ever, they were able to receive the clinically effectivedose of galantamine 16mg/day for up to eight weeks.While it is possible that a longer time on the maxi-mum dose may affect outcomes on functional mea-sures, it is unlikely that this would account for theobserved treatment differences in cognition at Week12. In addition, more time is required for donepezilto reach steady state compared with galantamine,owing to donepezil’s longer half-life. If this studyhad been continued for a longer time period, it is pos-sible that the outcomes would have remained infavour of donepezil—it is not appropriate to extrapo-late from a 12-week study that differences in treat-ment efficacy and tolerability will eventually bereduced.

The majority of decisions regarding treatment con-tinuation by patients and caregivers and, to a lesserextent, by physicians, are most likely to be madewithin the first three months of treatment. Short termtrials of this nature are therefore both relevant andimportant to physicians, patients, and their caregiverswhen selecting and initiating appropriate sympto-matic treatment for AD.

REFERENCES

Jones RW, Soininen H, Hager K, et al. 2004. A multinational, ran-domised, 12-week study comparing the effects of donepezil andgalantamine in patients with mild to moderate Alzheimer’sdisease. Int J Geriatr Psychiatry 19: 58–67.

Wilcock G, Howe I, Coles H, et al. 2003. A long-term comparisonof galantamine and donepezil in the treatment of Alzheimer’sdisease. Drugs Aging 20: 777–789.

Wilkinson DG, Passmore AP, Bullock R, et al. 2002. A multina-tional, randomized, 12-week comparative study of donepeziland rivastigmine in patients with mild to moderate Alzheimer’sdisease. Int J Clin Prac 56: 441–446.

Tariot PN, Solomon PR, Morris JC, et al. 2000. A 5-month, rando-mized, placebo-controlled trial of galantamine in AD. TheGalantamine USA-10 Study Group. Neurology 54: 2269–2276.

ROY JONESThe Research Institute for the Care of

the Elderly, St Martin’s Hospital, Bath, UK

HILKKA SOININENDepartment of Neurology

Kuopio University Hospital, Finland

Copyright # 2004 John Wiley & Sons, Ltd. Int J Geriatr Psychiatry 2005; 20: 85–89.

88 letters to the editor

KLAUS HAGER

Department of Medical Rehabilitation andGeriatrics Henriettenstiftung, Hannover, Germany

DAG AARSLAND

Section of Geriatric PsychiatryRogaland Central Hospital, Stavanger, Norway

PETER PASSMORE

Department of Geriatric MedicineQueens University, Belfast, UK

Published online in Wiley InterScience(www.interscience.wiley.com).

DOI: 10.1002/gps.1234

Copyright # 2004 John Wiley & Sons, Ltd. Int J Geriatr Psychiatry 2005; 20: 85–89.

letters to the editor 89