Adverse Drug Reactions in Older People with Dementia

To The Editor: We read with interest the paper by Giron et al.1 regarding appropriateness of drug use in an older nondemented and demented population. The authors demonstrated that there was substantial exposure to presumptive inappropriateness of drug use in nondemented and demented older subjects with the equivalent number of prescribed drugs. Because noncompliance resulting from low medication management capacity in demented subjects might lead to the higher incidence of adverse drug reactions (ADRs), it is important to study the actual rates of ADRs in older people with dementia, particularly in association with the number of drugs. To do so, we conducted a database analysis of the patients (n = 517) who were admitted to the acute geriatrics ward of the University of Tokyo Hospital from 1995 through 1998. The presence or absence of ADRs was recorded in 95% (n = 490; mean age ± standard deviation = 69 ± 14) of the patients. ADRs were found in 62 subjects (12.7%); 24% were due to antibiotics, 20% to antihypertensives, 18% to nonsteroidal anti-inflammatory drugs, and 9% to psychotropics. The incidence of ADRs increased after 75 years of age and correlated with the number of prescribed drugs; ADRs were found in less than 10% of the patients who were taking five drugs or fewer, and in more than 15% of the patients who were taking more than five drugs. The relationship between medication management capacity and ADRs was analyzed. Medication management capacity was assessed by Lawton and Brody's2 instrumental activities of daily living (IADLs) as independent, assistance, or dependent. As expected, impairment of medication management capacity was associated with cognitive dysfunction, as measured by the Hasegawa Dementia Scale—Revised3 and with the disturbance of activities of daily living, as measured by Barthel Index (Table 1). Although most of the subjects in the dependent group were severely demented, there was a nonsignificant lower incidence of ADRs in this group than in other groups and subjects in this group were taking a smaller number of drugs than those in the assistance group. Polypharmacy may lead to the high incidence of ADRs in older people, as was shown in our total subjects. Accumulation of diseases contributes, at least in part, to polypharmacy. In our study, the assistance group had a larger number of diseases and prescribed drugs than the independent group (Table 1). Why then was the number of drugs smaller in the dependent group than in the assistance group even though the number of diseases was comparable? In our department, it has been recommended that the number of drugs be limited to three in patients with severe cognitive dysfunction. Comprehensive geriatric assessment, including cognitive function and IADLs, might have been helpful in detecting those at high risk for noncompliance and ADRs. In fact, since comprehensive geriatric assessment was introduced to our geriatrics ward, the incidence of ADRs has decreased from 18% in 1990 to 12.7% in 1995–1998. Taken together, the incidence of ADRs was not higher in older people with dementia and low medication management capacity. Detection of the high-risk subjects and restriction of drug use in these subjects may prevent ADRs. Our results will add further information to the findings by Giron et al.1 about medication use in older people.