The prevalence of potential alcohol–drug interactions in older adults

Abstract Objectives The aim of this study was to assess the possibility of clinically significant drug–alcohol interactions among home-dwelling older adults aged ≥ 65 years. Design This study was a cross-sectional assessment of a stratified random sample of 2100 elderly people (≥ 65 years) in Espoo, Finland. The response rate was 71.6% from the community-dwelling sample. The drugs were coded according to their Anatomical Therapeutic Chemical (ATC) classification index (ATC DDD 2012). Significant alcohol interactive (AI) drugs were examined according to the Swedish, Finnish, INteraction X-referencing (SFINX) interaction database, as well as concomitant use of central nervous system drugs, hypoglycaemics, and warfarin with alcohol. “At-risk alcohol users” were defined consuming > 7 drinks/week, or ≥ 5 drinks on a typical drinking day, or using ≥ 3 drinks several times/week, “moderate users” as consuming at least one drink/month, but less than 7 drinks/week, and “minimal/non-users” less than one drink/month. Results Of the total sample (n = 1395), 1142 respondents responded as using at least one drug. Of the drug users, 715 (62.6%) persons used alcohol. The mean number of medications was 4.2 (SD 2.5) among “at-risk users”, 4.0 (SD 2.6) among “moderate users”, and 5.4 (SD 3.4) among “minimal/non-users” (p < 0.001). The concomitant use of AI drugs was widespread. Among the “at-risk users”, “moderate users”, and “minimal/nonusers” 42.2%, 34.9%, and 52.7%, respectively, were on AI drugs (p < 0.001). One in 10 of “at-risk users” used warfarin, hypnotics/sedatives, or metformin. Conclusions Use of AI drugs is common among older adults, and this increases the potential risks related to the use of alcohol.


Introduction
Over the past few decades, the use of drugs has increased among older adults [1,2]. Approximately nine in 10 of older adults take medications [1, 3,4]. The number of drugs increases with advancing age [5]. Use of multiple medications increases the risk of adverse drug reactions (ADRs) and drug -drug interactions (DDIs) [6].
Although epidemiological studies have shown a decline in alcohol drinking along with age, the number of older adults drinking will increase when the age cohort born in the 1950s with their heavier drinking habits reach old age [7 -9]. Use of alcohol has increased particularly among men and women aged between 65 and 74, and abstinence confounders, moderate alcohol consumption was associated with a 24% increase in the risk of an adverse drug reaction [17]. Some of these interactions are due to age-related changes in the absorption, distribution, and metabolism of alcohol and medications. Others are due to disulfi ram-like reactions observed with some medications, exacerbation of therapeutic effects and adverse effect of medications when combined with alcohol, and interference of alcohol with effects of some medications. [15]. For example, warfarin has signifi cant drug, food, and alcohol interactions and concomitant use of alcohol may increase or decrease warfarin metabolism [18]. It has high inter-individual and intraindividual variations in international normalized ratios (INR). Alcohol enhances the sedative effects of antidepressants, barbiturates, benzodiazepines, and opioids, creating the potential for adverse events such as falls [13,19], or cognitive decline or central nervous system depression [15]. Alcohol consumption by diabetic patients on hypoglycaemic agents can also increase the risk of hypoglycaemia [20]. Alcohol can potentiate the effects of metformin on lactate metabolism, which may rarely result in lactic acidosis, particularly in acute alcohol intoxication [21].
There are only a few epidemiological studies investigating the concomitant use of alcohol and drugs presenting a potential risk for older adults [4,17,19,22 -24]. Of these studies, two examined emergency department or hospital admissions [17,19]. Thus, they are not representative of homedwelling older populations. Two focused only on the concomitant use of alcohol and psychotropic drugs [23,24]. In a large population based US study, 77% of older people were exposed to potential alcohol -drug interactions [22]. In a previous study from Eastern Finland, the use of AI drugs among moderate drinkers was common [4]. The aim of this study was to assess the possibility of clinically signifi cant drug -alcohol interactions among a random sample of home-dwelling older people aged 65 and older.

Materials and methods
The data were gathered in May -September 2007. A postal questionnaire was sent to a stratifi ed random sample of 2100 older individuals ( Ն 65 years) from the Espoo Population Register. A stratifi ed random sample of 350 was retrieved from each fi ve-year age cohort (65 -69 years, 70 -74 years, 75 -79 years, 80 -84 years, 85 -89 years, 90 years and over). The questionnaire was re-sent after three months to those who had not yet responded.
Alcohol consumption was charted with several questions developed from the clinical guidelines for alcohol use in older adults [12] and the AUDIT [26]. Quantity and frequency were ascertained by asking: (i) " How often do you have a drink containing alcohol (a drink in Finland contains 12 g of alcohol), including any beer, cider, wine, or liquor; spirits? " , (ii) " On a typical day when you drink, how many drinks do you have? (1 drink ϭ can or bottle (330 ml) of beer, 12 cl of wine, 4 cl of liquor; spirits (one shot glass), or 8 cl of sherry or madeira or aperitif) " , and (iii) " How often do you have three or more drinks on one occasion? " We counted the amount of alcohol consumed by the respondent by taking into account the frequency of use and the portions consumed on one occasion. We defi ned " at-risk users " as those consuming Ͼ 7 drinks/week, or Ն 5 drinks on a typical drinking day, or using Ն 3 drinks several times/week to be in line with the guidelines for persons Ն 65 years [12]. " Moderate users " were defi ned as consuming at least one drink/month, but less than 7 drinks/week, and " minimal/non-users " less than one drink/month. Falls related to alcohol use were inquired about as follows: " Have you fallen or injured yourself when you have used alcohol (never/yes, but not during the last year/yes, during the last year) " .
The concomitant use of drugs with potential • interactions with alcohol was common.
Among the " at-risk users " and " moderate • users " , 42.2% and 34.9% were on drugs potentially causing signifi cant interactions with alcohol. It is important that clinicians discuss medi-• cation and alcohol consumption with their patients and warn patients who are prescribed alcohol interactive drugs about alcohol -drug interactions.
We used the Statistical Package for the Social Sciences (SPSS) version 15.0 for Windows, and NCSS statistical analysis and graphics software to compute descriptive statistics. The " at-risk users " , " moderate users " , and " minimal or non-users " groups were compared with the chi-squared test in their categorical variables and the Kruskal -Wallis test was used for the non-normally distributed continuous variables.

Description of respondents
Of the random sample, 31 had a native language other than Finnish or Swedish (inquiry form languages), 16 were deceased, 92 lived in permanent institutional care, and the postal address of 14 individuals had changed and was therefore unknown. Thus, the number of potential respondents was 1947, of whom 1395 returned the questionnaire. The response rate for the community-dwelling older persons was 71. 6%.
Of the total sample (n ϭ 1395), 1142 persons reported using at least one drug and had data available concerning their alcohol consumption. Their mean age was 78.7 years and 64.5% were females. By defi nition, 90 could be defi ned as " at-risk users " , 625 as " moderate users " , and 427 as " minimal/ non-users " . The " at-risk users " were more often males and married; they were younger and had higher education than the " non/minimal users " . The mean number of medications in these groups was 4.2 among " at-risk alcohol users " , 4.0 among " moderate users " , and 5.4 among " minimal/non-users " (p Ͻ 0.001) ( Table I).
Participants had many diseases (Table II). The " minimal or non-users " usually had a higher number of diagnoses than the " at-risk " or " moderate " users. They also had poorer subjective health (see Table I).

Prevalence of potential alcohol -drug interactions
Of the drug users, 62.2% (715) persons also used alcohol. The use of drugs with potential alcohol interactions among participants is presented in Table  III. Among the " at-risk alcohol users " 42.2% were on AI drugs, whereas the respective fi gure among the " moderate users " was 34.9% and among the " non/minimal users " 52.7% (p Ͻ 0.001). The mean number of AI drugs in these groups was 0.67 (SD 0.98), 0.48 (SD 0.74), and 0.80 (SD 0.96), respectively. One in 10 of " at-risk users " and " moderate users " were on warfarin or hypnotics/sedatives (N05C). Of the groups, " at-risk users " used metformin more commonly than the other two groups (13.3% vs. 5.6% vs. 8.4%, p (adjusted) ϭ 0.0089) (Table III).
Among the " at-risk users " on AI drugs, 13.8% reported they had fallen or injured themselves when using alcohol, whereas the respective fi gure among drug-users was 4.1% (p Ͻ 0.001).

Summary
The concomitant use of drugs with potential interactions with alcohol was common. Of the drug users, 62.2% (715) persons also used alcohol. Among the " at-risk users " , " moderate users " , and " minimal/ nonusers " , 42.2%, 34.9%, and 52.7% were on drugs potentially causing signifi cant interactions with alcohol. Of the " at-risk users " , 11% were on warfarin, 11% on hypnotics/sedatives, and 13% on metformin, which is a serious concern for potential adverse events.

Strengths and weaknesses
The strength of this study lies in the large and representative sample of older home-dwelling persons in bigger cities in Finland. The high response rate also supports its validity. However, relying on selfreporting measures imposes a limitation on the study since self-reported alcohol consumption is likely to  be underestimated to some extent in drinking surveys [27 -29]. The cross-sectional nature of this study limits our ability to fully explore the associations identifi ed. Our data give a picture of the use of alcohol and potential AI drugs, but this does not allow us to evaluate whether alcohol and drugs are actually used concomitantly in everyday life and what kind of adverse events they expose these individuals to.
The problems and adverse events related to alcoholdrug interactions depend on both quality and quantity of medication and quantity of alcohol used as well as the regularity and simultaneity of their consumption in the daily rhythm of life [13,14,17].
The " at-risk users " and " moderate users " were younger, healthier, and more often males and more educated than the " non-users " . The cross-sectional nature of our study does not allow us to evaluate whether older people decrease their consumption of alcohol when they get older, have more illnesses and receive more medications, or whether these fi ndings refl ect changing trends in lifestyle of older and younger cohorts. Some prior studies have suggested that the cohorts born later actually use more alcohol than the older cohorts [7 -9].

Potential adverse effects related to AIs
Falls and injuries were more common among " atrisk users " using AI drugs concomitantly than among others. Both alcohol and drugs with central nervous system effects can increase the risk of falls [30,31]. Alcohol and sedatives can reduce awareness and balance, which in turn can increase the risk of injurious falls [31]. Alcohol consumption and the use of psychotropic drugs have become more prevalent among older adults [3,23]. In our study, almost half of the users of antipsychotics, anxiolytics, hypnotics, and antidepressants drank alcohol. In a previous Finnish study, almost 40% of the users of antidepressants, benzodiazepines, sleeping pills, and opiates drank alcohol [4].
The metabolism of warfarin is infl uenced by alcohol. The use of warfarin may expose individuals to higher risk of serious bleeding [13,18], especially when falling. Alcohol consumption can result in dangerously high or insuffi cient warfarin activity depending on the patient ' s drinking habits [13]. Occasional consumption of low to moderate amounts of alcohol may not have an effect on warfarin anticoagulation whereas the effect of short-term consumption of large amounts of alcohol is unknown [32]. The effect of chronic use of large amounts of alcohol is less clear: it may prolong warfarin halflife without having an effect on an international normalized ration [32]. In our study of the " at-risk alcohol users " , 11% were on warfarin. Warfarin requires skilful dose management and patient communication to achieve the best outcomes. It is important that patients are aware of warfarinalcohol interaction and their doctors of patients ' alcohol consumption.
Alcohol can potentiate the effects of metformin, which was the most commonly used oral hypoglycaemic in our study. Of the " at risk-users " 13.3% were on metformin, 6.7% used sulfonylureas, and 1.1% insulin. The use of metronidazole, tinidatzole, disulfi ram, griseofulvin, prazosin, and tacrolimus was negligible.

Clinical meaning of the study
The concomitant use of medications with potential interactions with alcohol was widespread. It is important that clinicians are screening the alcohol consumption (e.g. by using AUDIT) and discuss medication and alcohol consumption with their patients and warn patients who are prescribed alcohol interactive drugs about alcohol -drug interactions.

Funding and acknowledgements
This work was supported by the Yrj ö Jahnsson Foundation. The authors would like to thank statistician Hannu Kautiainen for his kind advice on the statistical analyses.

Ethical approval
The local ethics committee approved the study protocol.