The role of digitalis glycosides in heart failure has recently been re-evaluated [1-3]. In addition to their inotropic effect, they exert neuro-endocrine suppression, sympathetic inhibition and arterial vasodilation. Four large, controlled trials have shown digoxin to be of benefit in patients with heart failure and sinus rhythm . Digoxin is therefore an adjunct to diuretics and angiotensin-converting enzymes .
Digoxin and digitoxin are the two glycosides mainly used, the former due to its more rapid elimination in case of intoxication . Prospective toxicity studies with digoxin showed adverse effects similar to placebo , compared with 5.8% of digitoxin users [61. Digitoxin has a more stable absorption  and its elimination is not changed by reduced liver or kidney function [8-10].
The serum concentrations of digitoxin (but not digoxin) can be predicted from daily intake of the drug related to body weight . An intake of digitoxin of 0.05 mg/day is considered unlikely to give serum concentrations high enough to cause toxicity. This study shows that low doses of digitoxin in elderly subjects can result in increased serum concentrations and cause digitalis intoxication.
Materials and methods
The patients were admitted consecutively with an acute illness from homes or institutions to the geriatric unit at Aker Hospital over 2 months with a primary diagnosis of stroke (patients 3, 5 and 6), cardiac failure (patients 1, 2 and 4). The inclusion criterion was symptomatic digitalis intoxication  when on 0.05 mg/day digitoxin. Serum digitoxin concentrations were measured on admission. The patients with increased digitoxin concentrations were not given specific treatment to increase digitoxin elimination. Their serum digitoxin concentrations were usually measured weekly until they returned to normal. No drug known to interact with digitoxin was taken by any patient.
Serum sodium, potassium, creatinine, albumin and magnesium concentrations were analysed by a Hitachi 717 multianalyser (Boehringer-Mannheim, Mannheim, Germany). Serum digitoxin was analysed in the Clinical Chemistry Department of Ullevaal University Hospital, Oslo by TDx/TDxFlx Digitoxin assay (Abbott Laboratories, Abbott Park, IL, USA) which gives a suggested therapeutic range of 10-30 ng/ml (15-40 nmol/l) . The slope of the digitoxin elimination curve was calculated by linear regression by the data program Enzfit (Robin Leatherborrow, Cambridge, UK) and the half-life calculated from this curve.
Patients had a mean age of 86.5 [+ or -] 5.9 years and moderately increased serum creatinine concentrations (Table 1). Serum digitoxin concentrations were high or well above the usual therapeutic range (10-30 ng/ml) and all the subjects had reduced serum albumin concentration. The digitoxin/albumin ratio ranged from 0.80 to 1.93. Three patients had weights in the normal range, two had reduced weights. One patient had reduced serum magnesium and two had slightly reduced serum potassium concentrations. During the period of discontinuation of digitoxin, the patients recovered from general weakness or gastrointestinal symptoms. None of the patients had cardiac arrhythmia. The digitoxin concentrations decreased slowly, with a mean half-life of 25.2 [+ or -] 9 days (Figure 1 and Table 1).
(a) Reference value: 6.7 [+ or -] 1.2 .
The mean half-life of digitoxin in our patients was much longer than that previously reported (6.7 [+ or -] 1.7 days) . Such a prolonged half-life may lead to accumulation and increased serum concentrations even on a low dose of the agent. A high intake of digitoxin before hospitalization could possibly explain the increase in digitoxin concentrations, but this explanation is unlikely, due to the very slow elimination of digitoxin observed in our patients while in hospital.
The half-life of digitoxin in six free-living healthy elderly people with a mean age of 71 years was 8.3 days . However, these results can not be extrapolated to patients in their eight and ninth decade.
Digitoxin elimination is assumed not to be reduced by declining liver or kidney function [8-10]. When the conversion of digitoxin was studied in vitro in biopsies from human liver, the conversion of digitoxin to digitoxigenin-bis digitoxoside varied with a factor of 40 . This conversion could be reduced in elderly subjects and be responsible for the slow digitoxin elimination seen in our patients.
The serum concentration of albumin was reduced, either due to acute disease or due to malnutrition which is prevalent in this age group . The ratio of digitoxin to serum albumin is usually 0.37-0.60 x [10.sup.-6], based on a therapeutic level of digitoxin of 15-25 ng/ ml and serum albumin concentrations of 42-45 g/l, but in our patients was much increased at 0.80-1.93 x [10.sup.-6]. Digitoxin is mainly protein-bound with a free fraction of 1-3%, [15, 16]. The lack of predictable changes in the free fraction of digitoxin with changes in albumin concentrations already in the physiological range  makes it difficult to use assessment of serum digitoxin in a rational way in clinical practice.
Our patients had typical symptoms of digitoxin intoxication which disappeared after discontinuation of therapy. Reductions in serum potassium and magnesium concentrations both increase intracellular levels of digitoxin  and may in two of our patients have contributed to the symptoms. Digitoxin intoxication can cause non-specific symptoms. It is difficult to evaluate to what extent the general improvement was due to reduced digitoxin concentrations or to improvement in the disease for which they were admitted.
Digitoxin is partially converted to digoxin, which is eliminated through the kidneys. The reduced kidney function might have contributed to the symptoms of digitoxin intoxication, but not to the prolonged half-life time of digitoxin.
The accumulation of digitoxin in elderly people, makes it necessary to monitor serum digitoxin concentrations in serum and to be aware of the possibility of intoxication even when patients arc taking only 0.05 mg a day.
[1.] Kraus F, Rudolph C, Rudolph W. Effectivity of digitalis in patients with chronic heart failure and sinus rythm: an overview of randomised double-blind placebo-controlled studies. Herz 1993; 18: 95-117.
[2.] Smith TW. Digoxin in heart failure. N Engl J Med 1993; 329:51-3.
[3.] Packer M, Gheorghiade M, Young JB, Constantini PJ, Adams KE Cody RJ et al. for RADIANCE study. Withdrawal of digoxin from patients with chronic heart failure treated with angiotensin-converting-enzyme inhibitors. N Engl J Med 1993; 329: 1-7.
[4.] Dargie HJ, McMurray JJV. Diagnosis and management of heart failure. Br Med J 1994; 308: 321-8.
[5.] Hoffmann BF, Bigger JT Jr. Digitalis and allied cardiac glycosides. In: Goodman Gillman A, Rail TW, Nies AS, Taylor P eds. The Pharmacological Basis of Therapeutics, eighth edition. New York: Pergamon Press, 1990; 814-39.
[6.] Storstein O, Hansteen V, Hatle L, Hillestad L, Storstein L. Studies of digitalis XII. A prospective study of 649 patients on maintenance treatment with digitoxin. Am Heart J 1977; 93: 434-43.
[7.] Wagner JG. Appraisal of digoxin bioavailability and pharmacokinetics in relation to cardiac therapy: Am Heart J 1974; 88: 133-8.
[8.] Jeliffe RW. A mathematical analysis of digitalis kinetics in patients with normal and reduced renal function. Math Biosc 1967; 1: 305-25.
[9.] Kirch W, Ohnhaus EE, Dylcwicz P, Pabst J, Storstein L. Bioavailability and elimination of digitoxin in patients with hepatorenal insufficiency. Am Heart J 1986; 111: 325-9.
[10.] Nokhodian A, Santos SR, Kirch W Digitoxin and its metabolites in patients with liver cirrhosis. Eur J Drug Metab Pharmacokinet 1993; 18: 207-13.
[11.] Bussey HI, Hawkins DW, Gaspard JJ, Walsh RA. A comparative trial of digoxin and digitoxin in the treatment of congestive heart failure. Pharmacotherapy 1988; 8: 235-40.
[12.] Sawyer WT. The digitalis glycosides. In: Taylor WJ, Diers Caviness MH eds. A Textbook for the Clinical Application of Therapeutic Drug Monitoring. Irving, TX: Diagnostic Division, Abbott Laboratories, 1986; 83-96.
[13.] Donovan MA, Castleden CM, Pohl JEF, Kraft CA. The effect of age on digitoxin pharmacokinetics. Br J Clin Pharmacol 1981; 11: 401-2.
[14.] Eberhart DC, Gemzik B, Halvorsen MR, Parkinson A. Species differences in the toxicity and cytochrome p-450 III A-dependent metabolism of digitoxin. Mol Pharmacol 1991; 40: 859- 67.
[15.] Mowe M, Bohmer T, Kindt E. Reduced nutritional status in an elderly population ([is greater than] 70 y) is probable before disease and possibly contributed to the development of disease. Am J Clin Nutr 1994; 59:317-24.
[16.] Brors O, Fremstad D, Poulsson C. The affinity of human serum albumin for 3H-digitoxin is dependant on albumin concentration. Pharmacol Toxicol 1993; 72:310-3.
Received 14 May 1996
Department of Medicine, Aker University Hospital, 0514 Oslo, Norway
Address correspondence to T. Bohmer. Fax: (+47) 22 89 4008
COPYRIGHT 1998 Oxford University Press
COPYRIGHT 2000 Gale Group