Potentially Inappropriate Medication Use in Mild Cognitive Impairment (MCI) Patients: Results from the Kerala Einstein Study

To the Editor

Potentially inappropriate medication (PIM) usage is a major safety concern, especially in cognitively impaired elderly. Despite the availability of guidelines for identifying PIMs,¹ a high proportion of older adults (8%–21%) continue to receive PIM.^2–4 Studies in developed countries suggest that patients with mild cognitive impairment (MCI) syndrome are at higher risk than the general population of receiving PIM.⁵ However, little is known about PIM usage among older adults, especially those with MCI, in developing countries. Elderly patients in developing countries have limited access to healthcare and PIMs may not only expendresources but the adverse side effects may not be readily recognized.

METHODS

We identified 27 patients with MCI (mean age 70.9, 40.7% women)attending neurology clinics at Baby Memorial Hospital, Kozhikode between October 2008 and October 2010 as part of the Kerala Einstein Study (KES).⁶ The goal of the KES is to identify risk factors for dementia in the southern Indian state of Kerala. We frequency matched cases by age and sex to 79 cognitively normal adults (mean age 70.4, 35.4% women)recruited from relatives of patients (70%) as well as from patients referred for non-cognitive complaints such as headaches, back pain, and peripheral neuropathies (30%). Inclusion criteria for KES were age 60 and older and willing to complete study procedures. Exclusion criteria included severe audiovisual loss or presence of active medicalor psychiatric illness that in the judgment of the study clinician would interfere with completion of study procedures. Informed consent was obtained from all participants and study evaluations followed protocols approved by the local institutional review board.

A neurologist, experienced in dementia evaluation, obtained medical history and recorded medication use. The clinician reviewed medication bottles, medical records, and verified information with informants. A battery of cognitive tests in the local language (Malayalam) and validated in the Kerala elderly population was administered by another clinician with behavioral neurology expertise. The cognitive battery assessed general cognition: Addenbrooke Cognitive Examination (ACE), Memory: Rey Auditory Verbal Learning Test, Attention/Executive function: Digit span test, and Mood: 15-item Geriatric Depression scale. The ACE incorporates the Mini-Mental State Examination (MMSE).

MCI was diagnosed in non-demented individuals with cognitive complaints and objective impairment (1.5 SD or below age and sex appropriate means) on tests of memory, attention executive function, or language at consensus diagnostic conferences attended by two or more study clinicians.

An independent clinician who did not do study evaluations or diagnosis reviewed medications blinded to cognitive status to identify PIM usage as per the updated 2003 Beer’s criteria¹ with the addition of benzodiazepines.⁵

Descriptive statistics were used to test group differences by MCI status. Logistic regression analysis adjusted for cognitive status (MMSE), Charlson Comorbidity Index and depressive symptoms was used to further examine association of MCI with PIM.

RESULTS

Seven participants were prescribed PIM (5.6%). Patients on PIM had higher Charlson Comorbidity Indexscores compared to those not on PIM(1.43 vs. 0.70, p=0.034). There were no significant group differences in age, education, depressive symptoms, cognitive scores, and prescription medication count.

Table 1 shows baseline and cognitive test characteristics by MCI status. The 27 MCI patients (20 amnestic and 7 non-amnestic) had higher PIM usage than the 79 normal controls (18.5% vs. 2.5%, p 0.011). Among MCI cases, PIMs included two medications with anti-cholinergic properties, two benzodiazepines and one cardiac glycoside. In the controls, PIMs consisted of one tricyclic antidepressant and one non-steroidal anti-inflammatory drug. MCI patients were at higher risk of being on a PIM (adjusted odds ratio 8.75; 95% CI 1.58 to 48.23).

Table 1. Baseline Characteristics and Cognitive Test Performance by MCI Status.

Values expressed as mean ± SD unless otherwise specified.

Variables	MCI (n = 27)	Normal elderly (n = 79)	P-value
Age	70.9 ± 5.9	70.4 ± 5.7	0.686
Sex, female %	40.7	35.4	0.649
Education, years	5.7 ± 3.8	8.7 ± 3.4	<0.001
Total prescription medication count	2.0 ± 0.8	1.9 ± 0.4	0.765
ACE	61.5 ± 12.0	80.7 ± 10.7	<0.001
Mini Mental Status Exam	22.9 ± 3.7	28.2 ± 2.3	<0.001
RAVLT 20 minute recall	1.5 ± 1.7	4.5 ± 2.9	<0.001
Digit span forward	5.1 ± 1.1	6.1 ± 1.1	<0.001
Digit span backward	2.6 ± 1.5	4.5 ± 4.5	0.037
Geriatric Depression Scale	6.6 ± 3.4	4.0 ± 2.9	<0.001

ACE: Addenbrooke Cognitive Examination

RAVLT: Rey Auditory Verbal Learning Test

DISCUSSION

Higher PIM use (18.5%) was seen in MCI patients in a clinic-based population from Kerala, India. Patients with MCI had an eight-fold increased risk of being on PIM compared to normal adults. A US based study of 689 MCI patients reported PIM usage in 20.8%, with anticholinergics and benzodiazepines being the most common.⁵ U.S. based studieshave reported that higher prescription medication countincreases risk of PIM usage,^2,4,5 but this was not the case in our study. The lower PIM usage (5.6%) in our overall sample may be a result of cultural differences in medical or social practices as well as resource availability but is still of concern in MCI patients. The high prevalence of medications with anti-cholinergic effects is concerning as they may negatively affect cognition.⁷ Furthermore, concomitant use of cholinesterase inhibitors and anticholinergics diminishes effectiveness of dementia medications.⁸ Although limited by sample size, the results serve as a reminder that extra care be taken when managing medications for MCI patients in developing countries.