Keywords: Warfarin, Ageing.
Introduction
Cross-sectional studies are consistent in reporting that warfarin requirements to achieve the same degree of anticoagulation are lower in elderly than in younger patients [1-3]. Although it has been inferred from these observations that warfarin requirements fall with ageing, such differences could result from birth cohort effects alone. Only the longitudinal method of study, in which serial measurements are made on the same subjects, can identify true age-related changes. The study described here was designed as a longitudinal study to test the hypothesis, suggested by the results of cross-sectional studies, that there is an age-related fall in warfarin requirements. We set out to investigate the rate of this change in a population of anticoagulated patients.
Methods
Patients who had been attending the anticoagulant clinics at the three Newcastle Hospitals, for a minimum period of 6 years, were invited to take part. The hospitals perform a full range of secondary care, as well as cardiothoracic surgery. One hundred and four community-dwelling patients (47 men) took part. They had been stabilized on warfarin for a median period of 10 years, range 6-24 years. Indications for warfarin were cardiac valve replacement (37), cardiac valve disease and atrial fibrillation (21), atrial fibrillation and cerebral embolism (16), recurrent deep-vein thrombosis or pulmonary embolism (28), and ischaemic cardiomyopathy (2). All patients were clinically stable and none was suffering from acute medical illness such as uncontrolled cardiac failure. Patients taking, or who had taken, drugs known to increase or inhibit warfarin sensitivity were excluded, as were those with abnormal blood tests of renal or hepatic function. Eighty-six patients were taking a range of drugs chronically, including loop diuretics (39), thiazide diuretics (15), amiloride or potassium (8) digoxin (61), beta-blockers (20), calcium channel antagonists (9), nitrates (5) angiotensin converting enzyme inhibitors (11), aspirin (13), thyroxine (4), allopurinol (5), iron (1), quinine (1), cinnarizine (1) and paracetamol (3).
All patients had retained, and provided for data collection, a complete record of their warfarin dosage over time, and degree of anticoagulation, measured before 1984 as the Thrombotest [4] and after 1984 as the International Normalised Ratio [5]. Each patient's age at the start of therapy was recorded, as was the average weekly dose during this and each subsequent year of therapy.
Statistical Methods
The data were evaluated by regression analysis using the MINITAB statistical package. Correlation coefficients (r) and the regression equations are reported, along with the p value. Ninety-five per cent prediction intervals for individuals with specified values of an x variable, which is an indication of the variability in the prediction of an individual's y variable using the regression equation in question were calculated.
Results
Median age of the patients at start of therapy was 59 years (range 31-74 years). Median dosage at commencement of therapy was 31 mg/week (range 9-70 mg/week). There was a significant negative correlation between dosage and age at the start of therapy (p = 0.002, r = -0.30) (Figure 1). The regression equation for this relationship was: dosage at start of therapy (mg/week) = 58.0 - 0.43 age at start of therapy. There was a significant fall in warfarin requirements over time, dosage difference being significantly correlated with age difference (p [is less than] 0.01, r = 0.25) (Figure 2). The regression equation for the change in warfarin requirements over time was: dosage difference = - 0.29 + 0.47 age difference. Use of this equation indicates a 21% fall in warfarin requirements in this population over a 1 5-year period. Estimations of actual fall in warfarin requirements over time, based upon this equation are shown in the Table.
[Figures 1 and 2 ILLUSTRATION OMITTED]
* Warfarin dosage requirements at start of therapy were significantly negatively correlated with age.
* Patients demonstrated a significant fall in warfarin requirements over time.
* The relationship between warfarin requirements and age established in this longitudinal study is similar to that established by cross-sectional studies and suggests that no major birth cohort effect is influencing requirements in our warfarin-treated population.
References
[1]. Wynne H, Cope L, Kelly P, Whittingham T, Edwards C, Kamali F. The influence of age, liver size and enantiomer concentrations on warfarin requirements. Br .T Clin Pharmacol 1995;40:203-7.
[2]. Gurwitz JH, Avron J, Ross-Degnan D, Choodnovsky I, Ansell J. Ageing and the anticoagulant response to warfarin. Ann Intern Med 1992;116:901-4.
[3]. Redwood M, Taylor C, Bain BJ, Matthews JH. The association of age with dosage requirement for warfarin. Age Ageing 1991;20:217-20.
[4]. Janson T, Jacobsen K, Brynildsrud T. Ten years of experience with standardization of thrombotest, according to the WHO model for ISI determination. Nicomed Pharma AS Publication, Diagnostics, Oslo, Norway, 1994.
[5]. Poller L. Laboratory control of anticoagulants. BM J 1987,294:118.
[6]. Calloway NO, Foley GF, Lagerbloom P. Uncertainties in geriatric data. J Am Geriatr Soc 1965;13:20-8.
[7]. Rowe JW, Andies R, Tobin JD, Norris AH, Shock NW. The effect of age on creatinine clearance in men: a cross-sectional and longitudinal study. J Gerontol 1976;31:15563.
[8]. Mungall DR, Luden TM, Marshall J, et al. Population pharmacokinetics of racemic warfarin in adult patients. J Pharmacokinet Biopharm 1985;13:213-27.
[9]. Shepherd AMM, Hewick DS, Moreland TA, Stevenson IH. Age as a determinant of sensitivity to warfarin. Br. J Clin Pharmacol 1977,4:315-20.
[10]. Chan E, McLachlan AJ, Pegg M, MacKay AD, Cole RB, Rowland M. Disposition of warfarin enantiomers and metabolites in patients during multiple dosing with racwarfarin. Br J Clin Pharmacol 1994;31:563-9.
Authors' addresses
H. A. Wynne(*), A. Long Department of Medicine,
F. Kamali, C. Edwards Department of Pharmacological Sciences,
P. Kelly Department of Medical Statistics
University of Newcastle upon Tyne
(*) Address correspondence to: Dr H. A. Wynne, Royal Victoria Infirmary, Queen Victoria Road, Newcastle upon Tyne NEI 4LP
Received in revised form 30 April 1996
COPYRIGHT 1996 Oxford University Press
COPYRIGHT 2004 Gale Group