Drugs Aging DOI 10.1007/s40266-014-0180-6

ORIGINAL RESEARCH ARTICLE

Concurrent Use of Drugs and Supplements in a CommunityDwelling Population Aged 50 Years or More: Potential Benefits and Risks Jure Peklar • Martin Charles Henman • Mitja Kos • Kathryn Richardson • Rose Anne Kenny

Ó Springer International Publishing Switzerland 2014

Abstract Background The use of vitamin and mineral (VMs) and non-vitamin/non-mineral supplements (non-VMs) in the general population and the older population in developed countries has increased. When combined with drugs, their use can be associated with benefit and potential risks. Objective The aims of this study were to determine the extent and associated factors of the combined use of drugs and VM/non-VM supplements, and to examine the potential major drug–supplement interactions Methods Cross-sectional analysis of first-wave data of TILDA, The Irish Longitudinal Study on Ageing, nation-

J. Peklar
M. Kos Faculty of Pharmacy, University of Ljubljana, Ljubljana, Slovenia J. Peklar (&)
M. C. Henman School of Pharmacy and Pharmaceutical Sciences, Trinity College Dublin, Dublin, Ireland e-mail: [email protected] K. Richardson
R. A. Kenny The Irish Longitudinal Study on Ageing (TILDA), Trinity College Dublin, Dublin, Ireland R. A. Kenny Department of Medical Gerontology, Trinity College Dublin, Dublin, Ireland R. A. Kenny Trinity College Institute of Neuroscience, St James’s Hospital, Dublin, Ireland

ally representative a cohort including 8,081 communitydwelling persons aged C50 years. Prevalences including 95 % confidence intervals (CI) were weighted to the population. Group differences in drug and supplement use were assessed using Pearson’s Chi-square test, and associations between concurrent drug–supplement use and covariates were assessed using logistic regression. Potential interactions between drugs and supplements were assessed using relevant sources. Results Every seventh respondent (14.0 %; 95 % CI 13.1–15.0) reported regular concurrent use of drugs and supplements; 7.9 % (95 % CI 7.3–8.6) took only VMs, 3.9 % (95 % CI 3.4–4.4) took only non-VMs, and 2.2 % (95 % CI 1.8–2.6) took at least one of each concurrently with drugs. Concurrent use was more prevalent in women and in the oldest (C75 years) group. Chronic disease, female sex, third-level education and private medical insurance were associated with an increased likelihood of use of both supplement types, whereas those classed as employed were much less likely to use any supplements. Supplements were combined with drugs in all of the commonly prescribed therapeutic groups, ranging from just under 60 % with drugs for bone diseases to 15.7 % with drugs for diabetes. Potential major drug–supplement interactions were detected in 4.5 % (95 % CI 3.4–5.8) of concurrent drug–supplement users, and were more prevalent in older respondents. Conclusions Concurrent use of drugs and supplements among those aged over 50 years in the Irish population is substantial and increases with age. There is considerable variation in usage, and the outcome of this approach is evidence of unmet need and therefore unrealised benefits among some subgroups, and of exposure to avoidable and potential serious drug interactions among others.

J. Peklar et al.

Key Points Concurrent use of drugs and supplements among those aged over 50 years in the Irish population is substantial and increases with age and number of chronic conditions. Due to considerable variation in drug–supplement use, potential benefits of both valuable treatment intensification and prevention are not fulfilled. Vulnerable groups for potential harmful combined use are persons aged C75 years, women, those with C3 chronic conditions, and/or polypharmacy.

1 Introduction In past 20 years, the use of herbal and dietary substances in the general population and the older population in developed countries has increased [1–4]. For example, in a national UK survey including a general population aged 19–64 years, Denison et al. [4] reported that in 1986/87, 9 % of men and 17 % of women used at least one food supplement in the past 3 months, while the proportion in 2008/09 increased to 27 % and 41 % of men and women, respectively. Published population prevalence data on concurrent use varies markedly from 15 % to 74 % [3, 5, 6] because of differences in the populations studied and the methodology used during assessment, the recall period, and the definition of the food supplement used. In a recent Irish study of a national sample of people aged 50 years or more, 14 % of those who reported regular (‘like every day or week’) drug use, were also regularly taking at least one supplement [7]. Supplements may contain vitamins, minerals and other dietary components, as well as herbs and other substances. Some have an evidence-based role in the prevention or treatment of clinical conditions and may be prescribed, while others are widely used as ‘alternative’ or ‘complementary’ therapies to prescribed drugs and may be purchased without prescription. The combinations may be intentionally used to meet common treatment needs or to meet separate needs, but both may incur risk as well as benefit. Because many supplements are promoted as ‘natural’ and have long been available, they may be considered as safe and effective by the public. Despite their widespread use, the potential risks associated with combining them with other medications are poorly understood by consumers and only a minority of supplement users disclose their use to their health care practitioners [8]. Furthermore, only few practitioners actively assess their patients’ usage of supplements [5] as they may be challenged when

counselling them about concurrent use since knowledge of the effectiveness and safety of many of these preparations is limited and sometimes conflicting [9]. The aims of this paper were to determine the extent of the combined use of drugs and supplements and to identify factors associated with concurrent use, and to examine the potential major drug–supplement interactions among the population aged 50 years or more in the Republic of Ireland.

2 Methods 2.1 Study Design and Population The Irish Longitudinal Study on Ageing (TILDA) is a prospective cohort study led by Trinity College Dublin to provide systematically collected data that describes the social (social network, homecare), economic (income, employment, life standard) and health status (physical and mental health, drug and supplement use, need of health services and its utilisation) of older Irish adults, and try to identify the factors that influence healthy ageing. More details about the study cohort profile were published elsewhere [10]. Data on drug and supplement use were collected during the in-home interviews performed by trained interviewers using Computer-Aided Personal Interview software (Quancept SPSS). Inclusion criteria for participation in this study were age 50 years or older, and being resident in the Republic of Ireland. Institutionalized people were excluded from recruitment. Participants were selected by means of a three-stage RANSAM sampling procedure process [10] using the sampling frame of Irish Geodirectory, an up-todate listing and mapping of all residential addresses in the Republic of Ireland. First, all addresses in the country were clustered using proportionate stratification by socioeconomic status, age structure and geographical location; the second step was selection of a systematic random sample of addresses of fixed size within each cluster; and in the last step, a random list of residents who were C50 years of age was made. Of the 10,128 households randomly chosen using the RANSAM sampling procedure, 6,282 were eligible households, where at least one resident aged C50 years agreed to participate (response rate 62.0 %). The sample was designed to give each household in the country an equal probability of selection and, since all members aged C50 years in each household were eligible, each person aged C50 years also had an equal probability of selection. A total of 8,175 people aged 50 years and over participated in in-home interviews. In 94 cases, data on drug and supplement use were not collected (participants did not want to give, or did not know, the details about the drug/ supplement use during the interview), which corresponds to a participation rate of 98.9 %.

Concurrent Use of Drugs and Supplements

This study was approved by the Research Ethics Committee of Trinity College Dublin, and subjects were required to sign an informed consent document prior to participation. All experimental procedures adhered to the tenets of the Declaration of Helsinki. 2.2 Data Collection For this study, data was used from the first cross-sectional study that took place from October 2009 to February 2011. The in-home inventory of drugs and supplements was conducted by asking the question, ‘‘Now I would like to record all medications that you take on a regular basis, like every day or every week. This will include prescription and non-prescription medications, over-the-counter medicines, vitamins, and herbal and alternative medicines’’. No information about dose, frequency, quantity or prescription status was obtained. The definition of a food/dietary supplement according to Directive 2002/46/EC of the European Parliament and of the Council, 2002 [11] was used to describe the supplements in this study: ‘‘‘food supplements’ means foodstuffs the purpose of which is to supplement the normal diet and which are concentrated sources of nutrients or other substances with a nutritional or physiological effect, alone or in combination, marketed in dose form, namely forms such as capsules, pastilles, tablets, pills and other similar forms, sachets of powder, ampoules of liquids, drop dispensing bottles, and other similar forms of liquids and powders designed to be taken in measured small unit quantities’’. For the purpose of comparing concurrent drug and supplement use more in detail, they were divided into ‘vitamin’ (containing any vitamin and/or mineral as a main ingredient, e.g. multivitamin products, single vitamin/products, iron) [VMs], and ‘non-vitamin’ (containing herbal [e.g. echinacea] and other ingredients [e.g. glucosamine, omega-3-tryglicerides, garlic]) [non-VMs]. Medication data was coded using the Anatomical Therapeutic Chemical (ATC) classification system. (http:// www.whocc.no/atc). Products with ATC groups C10AX06 (omega-3-triglycerides), M01AX05 (glucosamine), A12AA (calcium), A11D, A11E and B03B (vitamins B), B03A (iron), A11A and A11B (multivitamins with minerals), A11GA01 (vitamin C) and A11CC (vitamin D and analogues) were classed as food supplements.

threatening and/or require medical intervention to minimize or prevent serious adverse effects’’ [12] were noted. Where discrepancies were found between these two sources a third, Stockley’s Drug Interactions [14], was used to decide on the clinical significance. The number of available references for each identified drug–supplement interaction with its potential interaction effect is listed in Table 5. 2.4 Statistical Analysis Concurrent use of drugs and supplements was compared across sex and age groups (50–64 years, 65–74 years and 75 years or more), and descriptive statistics were used to assess the differences in the prevalence of drug and supplement use. Prevalence of use is presented as a percentage, including the 95 % confidence interval (CI). Prevalence data were weighted adjusting for the non-respondents and the survey design to provide the national estimates. Group differences were assessed by Pearson’s Chi-square test and associations between prescription medication use and food supplement use was assessed using logistic regression. Adjusted and unadjusted odds ratios were presented with 95 % CI; the adjusted model considers variables previously found to be significantly associated with the use of drugs and supplements and were entered into the final model simultaneously: age, sex, education, location of living, self-perceived health status, income, smoking status, difficulties with activities of daily living or instrumental activities of daily living, body mass index, frequency of exercise. We also included factors that we thought could have an impact on combining drugs and food supplements: polypharmacy, marital status and employment status. Given the problem of missing income data, we included a ‘dummy’ variable for those with missing income data to avoid limiting the number of respondents available for the logistic regression analysis [6, 15]. Two-sided significance tests were used throughout and a p-value\0.05 was considered statistically significant. As there is a greater risk associated with potential interactions in those who take drugs with non-vitamin/nonmineral supplements, we assessed associated factors separately for those combining drugs with VMs (dependent variable 1) and non-VMs (dependent variable 2), respectively. All analyses were performed using SPSS 18 (Released 2009. PASW Statistics for Windows, Version 18.0; SPSS Inc., Chicago, IL, USA).

2.3 Identification of Potential Interactions Between Drugs and Supplements Reported 3 Results Potential interactions between drugs and supplements, and assessment of the clinical significance of the interactions, were assessed using the Thomson Micromedex [12] and Lexi-Comp [13] databases. Only major drug–supplement interactions, described as, ‘‘The interaction may be life-

3.1 Cohort Baseline Characteristics Of the 8,081 participants with medication information in our cohort, 4,375 (51.9 %; 95 CI 51.0–52.8) were female

J. Peklar et al.

and the mean ± standard deviation (SD) age was 63.8 ± 9.7 years, with the oldest participant being 99 years. Taking at least one drug was reported by 5,572 respondents (69.2 %; 95 % CI 68.0–70.4), and taking at least one supplement of any kind was reported by 1,404 respondents (16.6 %; 95 % CI 15.5–17.7). The mean number of drugs reported (±SD) was 2.35 ± 2.55, and mean (±SD) supplement of any type was 0.25 ± 0.64. Overall, every seventh respondent (14.0 %; 95 % CI 13.1–15.0) reported taking drugs and supplements. However, almost three-quarters of those took only one supplement with drugs, while the range of combined use varied from 1 supplement with 19 drugs to 10 supplements with 2 drugs. Multiple supplement use increased with age in both sexes, but most notably in men; from 0.10 ± 0.45 for men and 0.31 ± 0.76 for women aged 50–64 years, to 0.16 ± 0.51 and 0.37 ± 0.72 in those aged 65–74 years, and to 0.22 ± 0.51 and 0.38 ± 0.69, respectively, in the 75 years and older group. The majority of participants (657 [7.9 %]; 95 % CI 7.3–8.6) took only VMs with drugs, while 329 (3.9 %; 95 % CI 3.4–4.4) took only non-VMs with drugs and 189 (2.2 %; 95 % CI 1.8–2.6) took at least one of each of VMs and non-VMs (Table 1). 3.2 Associated Factors of Concurrent Drug– Supplement Use Fifteen factors were tested in the logistic regression (Table 2) and 12 were shown to have some association to either or both types of supplements, with 11 associated with VMs and 6 with non-VMs. After adjustment for all 15 factors, age, marital status and level of physical activity did not show any association with concurrent use of any type of supplements (collectively) and are not presented in Table 2. Furthermore, even when tested for concurrent use of any type of supplement (dependent variable) with drugs, age, marital status and level of physical activity again did not show any association when adjusted for same factors. Chronic disease, female sex, third-level education and private medical insurance were associated with an increased likelihood of use of both supplement types, whereas those classed as employed were much less likely to use any supplements. Most notably, chronic disease was strongly associated with increased supplement use and showed a clear gradient for both types of supplements as the number of conditions reported increased. For those reporting fair or poor health, there was an increased likelihood (p \ 0.05) of drug and VM use, and those less able to carry out activities of daily living were more likely to take a non-VM (p \ 0.01). Although polypharmacy was associated with VM use, further analysis showed that the prevalence of combined use increased from just over 18 %

in both those taking B4 and those taking 5–9 drugs, to over 36 % in those taking C10 drugs. The associations in the population between increased combined drug and supplement use in women versus men, with increasing age and comorbidity are the aggregation of contrasting patterns of use when the subgroups are examined (Table 3). Only those reporting three or more conditions show a progressive increase in use across the age groups, with a consistently greater prevalence among women. Notably, with the exception of osteoporosis, the disparity in use between the sexes was much less for other conditions such as arthritis, and for cardiac conditions and diabetes, combined use was more prevalent among men than women in all of the age ranges. 3.3 Drug–Supplement Combinations Supplements were combined with drugs in all of the commonly prescribed therapeutic groups, ranging from just under 60 % of those taking drugs for bone diseases to 15.7 % of those taking drugs for diabetes (Table 4). VMs were most frequently combined with drugs for bone disease (56.4 %; 95 % CI 51.2–61.4) and least frequently taken with drugs for diabetes (10.8 %; 95 % CI 8.2–14.1), while non-VMs were most frequently used with non-steroidal anti-inflammatory drugs (NSAIDs) [15.5 %; 95 % CI 12.5–19.2] and least frequently combined with systemic antibiotic preparations (4.5 %; 95 % CI 1.8–10.5). Usage of supplements with clear therapeutic indications showed considerable variation, such as the twelvefold variation in the use of calcium preparations, which was in marked contrast to other supplements; for example, omega-3-triglycerides only varied threefold across the therapeutic groups examined. 3.4 Potential Major Drug–Supplement Interactions In total, at least one potential interaction of any kind was detected in 335 (4.1 %; 95 % CI 3.6–4.5) respondents. No combination was found to be absolutely contraindicated. The numbers of supplements used ranged from 1 to 9 and of drugs from 1 to 16; four cases involved only one or two drugs with one or two supplements. Nevertheless, multiple supplement use was found in almost half (49.1 %) of this group compared with only 7.8 % of the whole drug–supplement-using cohort. Over twice as many took VMs alone (27) compared with non-VMs (12), while 14 took both supplement types. Almost equal numbers of interactions involved VMs (27) and non-VMs (29) despite the greater prevalence of drug–VMs combined use (Table 5). Ten different potential drug–food supplement major adverse interactions were identified in the cohort and non-vitamin supplement

Concurrent Use of Drugs and Supplements Table 1 Baseline characteristics of the study cohort (absolute count and weighted proportions adjusted to non-respondents and survey design to provide national estimates) Category

Subgroup

All Age (years)

Sex Location of household

Education

Marital status

50–64

Yearly household income [€] (missing = 147)

Self-rated health (missing = 5) Number of chronic conditions

Taking drugs with nonvitamin/non-mineral supplements (n = 518)

The rest (n = 6,900)

n

n

% (95 % CI)

n

n

8,081

846

10.1 (9.4–10.9)

518

6.1 (5.5–6.8)

6,900

85.9 (84.9–86.8)

4,635

355

7.1 (6.4–8.0)

248

5.0 (4.2–5.8)

4,127

89.7 (88.7–90.7)

% (95 % CI)

% (95 % CI)

2,137

259

11.6 (10.1–13.1)

165

7.5 (6.3–8.8)

1,766

83.2 (81.2–84.9)

C75

1,309

232

18.0 (15.8–20.5)

105

8.0 (6.5–9.8)

1,007

76.9 (74.2–79.3)

Male

3,706

192

5.0 (4.3–5.8)

165

4.2 (3.5–4.9)

3,386

91.8 (90.7–92.7)

Female

4,375

654

14.8 (13.6–16.1)

353

7.9 (7.0–8.9)

3,514

80.5 (79.0–81.8)

Dublin

1,915

237

12.3 (10.7–14.0)

132

7.0 (5.8–8.4)

1,600

83.6 (81.5–85.4)

Other city

2,290

233

9.8 (8.6–11.3)

150

6.1 (5.0–7.4)

1,956

86.1 (84.4–87.6)

Rural

3,864

376

9.3 (8.3–10.5)

236

5.7 (4.8–6.8)

3,332

86.8 (85.4–88.1)

None/primary Secondary

2,456 3,236

262 329

10.9 (9.7–12.3) 9.6 (8.6–10.7)

148 194

6.1 (5.2–7.2) 5.8 (4.9–6.7)

2,088 2,785

84.7 (83.1–86.2) 86.8 (85.5–88.0)

CThird-level

2,386

255

9.7 (8.6–11.0)

176

6.9 (5.8–8.1)

2,024

86.1 (84.6–87.5)

Married

5,592

508

8.8 (8.0–9.7)

358

6.0 (5.3–6.8)

4,839

87.1 (86.0–88.1)

Never married

780

83

9.3 (7.3–11.8)

34

3.8 (2.7–5.5)

677

88.3 (85.7–90.5)

Separated/divorced

548

54

8.7 (6.6–11.4)

34

5.7 (4.0–7.9)

475

87.9 (85.0–90.4)

1,161

201

17.1 (14.8–19.6)

92

8.1 (6.6–10.0)

909

78.4 (75.6–80.9)

Employed

2,914

155

4.7 (4.0–5.6)

122

3.7 (3.0–4.5)

1,452

82.1 (80.0–84.0)

Retired

2,995

412

13.5 (12.1–15.0)

233

7.4 (6.4–8.6)

2,686

84.0 (82.5–85.4)

Other

2,172

279

12.7 (11.2–14.3)

163

7.5 (6.3–8.8)

2,762

89.9 (88.7–91.1)

866

69

7.8 (6.0–10.0)

44

4.8 (3.4–6.7)

2,679

92.8 (91.7–93.7

10,001–20,000

1,778

223

12.5 (10.9–14.2)

133

7.1 (5.9–8.6)

2,431

81.6 (80.0–83.2)

20,001–40,000

2,668

288

10.1 (8.9–11.4)

178

6.4 (5.9–7.5)

1,790

82.6 (80.7–84.3)

40,001–70,000

1,465

118

7.4 (6.1–9.0)

81

5.0 (3.9–6.3)

762

88.3 (85.7–90.5)

626

48

7.2 (5.3–9.7)

35

4.9 (3.4–7.0)

1,473

83.1 (81.1–84.9)

Not covered Private

831 3,268

35 288

4.0 (2.8–5.8) 8.0 (7.0–9.1)

26 219

2.9 (2.0–4.3) 6.2 (5.3–7.3)

2,265 1,294

85.8 (84.2–87.3) 89.3 (87.4–90.9)

Medical carda

3,978

523

12.9 (11.8–14.1)

273

6.7 (5.8–7.7)

560

90.2 (87.3–92.4)

0–10,000

[70,000 Health insurance

Taking drugs with vitamin/ mineral supplements (n = 846)

65–74

Widowed Employment

All

Excellent/very good 4,429

364

7.7 (6.9–8.7)

266

5.8 (5.0–6.9)

779

94.1 (82.2–95.6)

Good

2,400

275

11.1 (9.8–12.5)

152

5.9 (5.0–6.9)

2,847

88.2 (86.8–89.4)

Fair/poor

1,239

204

15.9 (13.9–18.1)

98

7.4 (5.9–9.1)

3,270

82.5 (81.1–83.9)

None

1,822

46

2.4 (1.8–3.3)

29

1.5 (1.0–2.1)

3,891

88.4 (87.3–89.5)

1

2,257

160

6.8 (5.8–8.0)

108

4.7 (3.8–5.7)

2,030

85.1 (83.5–86.6)

2

1,873

220

11.3 (9.8–13.0)

135

6.7 (5.6–8.0)

970

79.2 (76.8–81.4)

C3

2,129

420

19.1 (17.3–21.0)

246

11.0 (9.6–12.5)

1,756

96.6 (95.7–97.4)

CI confidence interval a

Medical card is most inclusive public health insurance available in Ireland

interactions made up eight of the ten types; of these, five involved interactions with anticoagulant/antithrombotic drugs (e.g. warfarin, aspirin, clopidogrel), and in three of these the risk derived from either gastric or generalised bleeding with NSAIDs. Not surprisingly, the single most prevalent potential interaction involved calcium supplements (15 cases) and digoxin and the risk of arrhythmia and cardiovascular collapse.

The 56 instances of potential major interactions concerned 53 respondents, which indicates that 0.7 % (95 % CI 0.5–0.9) of the national population aged 50 years and over could be at risk. Among them, the proportion of the oldest (C75 years) was highest (1.6 %; 95 % CI 1.0–2.5). However, among those aged 50 years and over who reported taking drugs and supplements concurrently (n = 1,181), the prevalence of potential major drug–

J. Peklar et al. Table 2 Effect of sociodemographic lifestyle and health-related factors on concurrent use of drugs and two different types of supplements— vitamin/mineral and non-vitamin/non-mineral Variables

Drugs with vitamin/mineral supplements

Drugs with non-vitamin/non-mineral supplements

Adjusted for all listed variables

Adjusted for all listed variables

OR

OR

95 % CI Lower

Upper

2.57

3.76

1.77***

95 % CI Lower

Upper

Sex Male

1a

Female

3.11***

1 1.43

2.20

Location of household Dublin city/county

1

Another town/city

0.81*

0.66

0.99

1.01

1 0.78

1.30

Rural area

0.85

0.70

1.03

0.98

0.78

1.24

0.88 1.02

1.41 1.74

Highest education achieved Primary/none

1

Secondary Third-level/higher

1.22* 1.28*

1 1.00 1.02

1.48 1.60

1.12 1.33*

Composition of household Living alone

1

Living with spouse

0.89

0.62

1.29

1 1.14

0.72

1.82

Living with othersb

0.68*

0.50

0.92

0.79

0.53

1.18

Employment status Retired

1

Employedc

0.58***

0.46

0.75

0.71*

1 0.54

0.94

Otherd

0.84

0.69

1.02

0.98

0.76

1.25

Yearly household income (€) 0–10,000

1

10,001–20,000

1.51**

1.12

2.04

1.43

1 1.00

2.05

20,001–40,000

1.73***

1.28

2.33

1.25

0.87

1.78

40,001–70,000

1.83***

1.28

2.62

1.11

0.73

1.69

[70,000 ‘No data’ (n = 674)

1.81** 2.13***

1.16 1.49

2.82 3.03

1.13 1.10

0.68 0.70

1.89 1.72

Number of reported chronic conditions None

1

1

2.44***

1.74

3.44

2.90***

1 1.91

4.41

2

3.58***

2.56

5.01

4.27***

2.82

6.46

C3

4.61***

3.29

6.46

6.96***

4.58

10.56

Disabilility Not disabled

1

IADL disability only

1.11

0.78

1.57

0.89

1 0.55

1.45

Any ADL disability

1.22

0.96

1.55

1.47**

1.10

1.96

Self-rated health Excellent/very good

1

Good

1.16

0.97

1.39

0.84

1 0.68

1.04

Fair/poor

1.35*

1.07

1.69

0.84

0.63

1.12

2.29

1 1.08

0.85

1.37

Polypharmacy (C5 drugs) No Yes Smoking

1 1.90***

1.57

Concurrent Use of Drugs and Supplements Table 2 continued Variables

Drugs with vitamin/mineral supplements

Drugs with non-vitamin/non-mineral supplements

Adjusted for all listed variables

Adjusted for all listed variables

OR

OR

95 % CI Lower

Upper

95 % CI Lower

Upper

Non-smoker

1

1

Currently smoker

0.65***

0.51

0.82

0.85

0.63

1.14

Ex-smoker

0.93

0.79

1.10

1.15

0.94

1.41

Medical cost cover (health insurance) Not covered

1

Private medical insurance

1.50*

1.03

2.20

1.65*

1 1.07

2.54

Medical card

1.43

0.97

2.09

1.13

0.73

1.75

OR odds ratio, CI confidence interval, IADL instrumental activities of daily living, ADL activities of daily living p-Values * \0.05; ** \0.01, *** \0.001 a

Indicates the reference group

b

Other relatives or unrelated people

c

Employed including self-employed Unemployed, permanently sick or disabled, looking after home of family

d

supplement interactions was much higher at 4.5 % (95 % CI 3.4–5.8). The data in this study were adjusted for non-respondents and survey design, and showed that significantly more major potential interactions were found in women than men: 74.5 % (95 % CI 61.1–84.4) vs. 25.5 % (95 % CI 15.6–38.9); in those who were retired: 52.3 % (95 % CI 38.5–65.8); and those who reported three or more chronic conditions: 58.8 % (95 % CI 44.4–71.8). Significantly more major potential interactions were also found in people with five or more drugs (polypharmacy): 70.6 % (95 % CI 57.9–80.8). Participants receiving health care from the State comprised 83.0 % (95 % CI 70.6–90.9) of the group, well above the proportion of those with private health insurance and those without any. Univariate logistic regression of those identified with interacting combinations of drug(s) and supplement(s) showed that the associated factors at p \ 0.05 were female sex, being C75 years of age, widowed, not employed, having C3 chronic conditions, disabled, having fair or poor self-rated health, taking C5 drugs (polypharmacy) and not being highly physically active. Further examination of the completed questionnaires and assessments of the individuals in this group showed that many of them reported a diagnosis of a cardiovascular disease (64.2 %; 95 % CI 51.3–77.1) and arthritis (39.6 %; 95 % CI 26.4–52.8), which, since the consequences of 44 of the potentially serious interactions were either cardiac arrhythmias or an increased risk of bleeding, would suggest a need for careful re-appraisal of their clinical condition. Additionally, 6 of 10 (63.6 %; 95 % CI 50.7–76.6)

reported regularly drinking alcohol, and 8 of the 53 (15.1 %; 95 % CI 5.5–24.7) were current smokers, which, given the inverse relationship between smoking and supplement use in this population [8], was notable.

4 Discussion This first study of a European population has shown that drug–supplement use is frequent and that drug–VM combinations are more prevalent than combinations with nonVMs, and that this pattern occurs across all ages, many common conditions and with all the major therapeutic drug groups. Although much of the combined use is associated with a few common self-reported conditions and seems appropriate, there appears to be unmet need in some groups, and in the most vulnerable of the elderly this combination use predisposes them to the risk of potentially serious drug–supplement interactions. Concurrent use of drugs with VMs was one-third higher than drug use with non-VMs (10.1 % vs. 6.1 %) and that there were three times as many women taking drug–VM combinations than men. Although there was increasing use with age, this was more marked with VMs than with nonVMs as usage more than doubled in the very old. Similar results have been reported in other countries for comparable cohorts regarding sex [3, 15, 16] and also age [3]. Women were also more likely to take non-VMs, and our study shows that the covariates associated with increased concurrent use were being female, with higher education and the presence of chronic diseases, which is similar to a

140 35.9 % (31.2–41.0)

40

35.6 % (27.2–45.0)

56.5 % (51.5–61.4)

55.2 % (45.8–64.3)

7.5 % (5.3–10.6) 227

9.1 % (5.0–16.2)

59

a

High blood pressure, atrial fibrillation, transient ischaemic attack

COPD chronic obstructive pulmonary disease

Data are expressed as n, % (95 % CI)

C3

1–2

0

32

14.8 % (11.7–18.6)

9.9 % (5.6–16.9)

10

36.7 % (31.7–41.9) 56

7.9 % (4.0–15.1) 11

Eye conditions

No. of chronic conditions

140

18.0 % (14.2–22.4)

14.6 % (8.9–22.8)

8

70

16

4.6 % (2.9–7.1)

7.3 % (3.6–14.1)

37.3 % (32.8–42.0) 19

32.2 % (24.1–41.6)

8

149

42.2 % (37.1–47.4)

52.1 % (43.0–61.0)

37

164

58

51.2 % (41.0–61.2)

58

41.1 % (31.8–51.2)

43

7.7 % (3.9–14.6)

8

25.3 % (17.7–34.7)

9.2 % (4.9–16.6) 28

10

16.7 % (10.4–25.6)

18

11.6 % (6.8–18.9)

14

44.2 % (34.6–54.2)

47

68.6 % (58.3–77.3)

75

Male (n = 109)

Male (n = 109)

Female (n = 399)

65–74

50–64

Age (years) and sex

Osteoporosis

COPD/asthma

Diabetes

Arthritis

Cardiaca

Self-reported condition

56.1 % (50.2–61.9)

147

42.3 % (36.6–48.3)

110

1.5 % (0.6–3.8)

5

28.3 % (22.8–34.5)

42.3 % (35.9–48.9) 77

113

15.5 % (11.5–20.6)

43

8.5 % (5.5–12.9)

21

55.1 % (48.7–61.4)

142

49.5 % (43.6–55.5)

127

Female (n = 262)

53.1 % (43.3–62.7)

54

41.2 % (31.8–51.3)

42

5.7 % (2.5–12.3)

6

45.9 % (35.8–56.4)

6.0 % (2.7–12.8) 48

6

17.4 % (10.9–26.7)

16

16.7 % (10.7–25.0)

17

46.4 % (37.2–55.9)

46

78.7 % (69.8–85.5)

79

Male (n = 102)

C75

67.6 % (60.7–73.9)

135

29.5 % (23.6–36.2)

60

2.8 (1.1–6.9)

5

51.5 % (44.4–58.6)

43.7 % (36.5–51.1) 103

89

13.9 % (9.6–19.8)

26

7.9 % (4.8–12.9)

16

55.7 % (48.0–63.2)

112

67.4 % (60.2–73.8)

133

Female (n = 200)

574

541

66

323

366

189

95

533

636

(n = 1,181)

Total

Table 3 Weighted prevalence estimates of the most commonly reported medical conditions and comorbidity by age and sex in those who reported both drug and supplement use

J. Peklar et al.

Lipid modifying (C10)

Antiepileptics (N03A)

Anxiolytics (N05B)

Hormone replacement therapy (G03CA03, G03FA01, G03AA05, G03AA05, G03CA57, G03FB05)

Antibiotics for systemic use (J01)

Thyroid therapy (H03)

Antidepressants (N06A)

NSAID and aspirin as painkiller (M01, N02BA01)

Hypnotics (N05C)

Drugs for acid-related disorders (A02)

Opioid analgesics (N02A)

Antiarrhythmics (C01B, C07AA07)

Other analgesics (N02B)

122 4.6

8.9

7.8–10.2

236

3.8

9.5

6.2–14.2

10

5.6

9.1–20.2

13.7

25

12

7.1

24

7.8–23.1

1.2 5

13.8

1

5.9–18.1

10.5 12

5.3

14

32

10.9

8.6–13.8

73

22 3.2

9.9–15.5

78

7.4

12.4

8.4–14.0

3.4 44

10.9

13

13.5–21.9

17.3 59

4.1

65

56

10.5–14.5

12.4

3.3

165

9

12.6–22.0

16.8

5.8

44

4

18.1

9.6–31.3

9

18 5.3

15.0–25.2

51

19.6

2.5

25.6

18.5–34.2

5

33

3.8–5.7

2.0–7.0

2.9–10.5

2.9–16.0

0.2–8.1

3.7–7.6

2.1–4.9

5.3–10.2

1.9–6.1

3.1–5.5

1.7–6.3

2.1–15.1

3.2–8.6

1.0–6.3

3.4–7.8

5.2

51.3

46.4–56.2

25

230

Bone diseases (M05B, G03XC01)

Antineoplastics (L01, L02)

Omega-3triglycerides (C10AX06)

Calcium with or without D vitamin (A12A)

Therapeutic group (ATC)

3.7

95

3.2

7

3.0

5

0.8

2.8 1

3

2.9

18

4.1

24

6.4

3.2 37

11

5.5

67

6.9

16

6.3

5

6.9

16

1.2

2

5.8

25

3.0–4.6

1.4–7.1

1.2–7.2

0.1–5.8

0.9–8.1

1.8–4.7

2.7–6.2

4.6–8.8

1.7–5.9

4.3–7.0

4.2–11.0

2.5–14.8

4.2–11.1

0.3–4.7

3.9–8.6

Glucosamine (M01AX05)

3.0

74

3.8

9

1.0

2

2.4

6.0 2

6

3.8

22

2.7

15

4.4

3.2 27

13

4.2

51

6.0

17

4.5

2

5.0

12

27.4

35

4.4

22

2.4–3.7

2.0–7.1

0.2–4.0

0.6–9.7

2.7–12.8

2.5–5.8

1.6–4.5

3.0–6.4

1.8–5.6

3.1–5.5

3.8–9.5

1.1–16.6

2.8–8.6

20.4–35.8

2.8–6.8

Vitamins B (A11, A11E, B03B)

1.7

37

1.8

3

1.7

3

2.3

1.0 2

1

2.9

16

1.4

7

2.2

3.2 15

10

2.1

25

1.1

3

0.0

0

2.2

7

2.1

3

3.3

14

1.2–2.5

0.6–5.6

0.5–5.7

0.5–9.5

0.1–6.6

1.7–4.9

0.7–3.0

1.2–4.1

1.5–6.4

1.3–3.1

0.3–3.3

0.0

1.0–4.8

0.6–6.4

1.8–6.0

Multivitamins with minerals (A11A, A11B)

2.0

45

2.2

5

0.6

1

0.0

4.2 0

3

2.2

12

2.5

12

1.7

2.8 9

8

4.3

50

2.2

6

2.2

1

5.1

12

1.0

1

3.1

11

1.5–2.7

0.9–5.3

0.1–4.3

0.0

1.4–12.1

1.2–3.9

1.4–4.6

0.9–3.3

1.3–6.0

3.3–5.6

1.0–4.9

0.5–9.0

2.9–8.8

0.1–6.5

1.7–5.7

Iron (B03A)

0.3

8

0.2

1

0.6

1

0.0

0.0 0

0

0.6

4

1.0

6

0.8

0.6 5

2

0.6

7

0.2

1

0.0

0

1.3

4

0.9

1

1.7

7

0.2–0.6

0.0–1.7

0.1–4.4

0

0

0.2–1.8

0.5–2.3

0.3–2.0

0.1–2.3

0.3–1.2

0.0–1.4

0

0.5–3.6

0.1–6.1

0.8–3.5

Evening primrose oil

0.4

10

0.0

0

0.5

1

3.2

0.0 2

0

1.2

7

0.2

1

0.8

0.3 5

1

0.3

4

0.3

1

2.9

2

0.6

2

0.5

1

0.5

3

0.2–0.7

0

0.1–3.6

0.8–11.8

0

0.6–2.6

0.0–1.3

0.3–2.0

0.0–2.2

0.1–0.8

0.0–2.2

0.7–10.9

0.1–2.5

0.1–3.3

0.2–1.7

Vitamin C (A11GA01)

21.2

536

21.2

50

21.3

38

22.8

24.1 19

24

24.2

145

25.3

145

27.4

28.4 155

98

28.4

353

31.2

81

32.2

19

34.7

94

49.0

64

58.9

263

19.4–23.0

16.1–27.3

15.6–28.3

14.8–33.4

16.2–24.3

20.6–28.3

21.8–29.0

23.6–31.4

23.6–33.8

25.7–31.3

25.7–37.4

21.3–45.4

29.2–40.7

39.9–58.0

53.8–63.9

Any supplement

14

14

13

12

11

10

9

8

6

6

5

4

3

2

1

Rank

Table 4 Concurrent use of most prevalent supplements with drugs from one of the most prevalent therapeutic groups in the whole cohort (n = 8,081) [weighted values adjusted to nonrespondents and survey design to provide national estimates]

Concurrent Use of Drugs and Supplements

20.5

19.2–21.9

19 12.4–19.6 83

15.7 0

0

0.0 0.2–1.6

3

0.5 2.2–5.8

16

3.6 0.3–2.6

3

0.8

Data are expressed as n, % (95 % CI)

Any therapeutic group

ATC Anatomical Therapeutic Chemical, NSAID non-steroidal anti-inflammatory drug

0.9–4.0 11

1.9 1.6–4.9

14 18

2.9 2.8–6.5

28 Diabetes (A10)

7.0–9.0

4.3

3.8 8.0

1.8–4.8

100

3.1–4.6 129 264

Antihypertensives (C02, C03, C07, C08, C09)

2.8

18 17.7–20.8 604 18 12 63 49 79

2.5–3.8

2.6

2.0–3.2

1.6

1.2–2.3

2.2

1.8–2.9

0.4

0.2–0.7

0.5

0.3–0.8

19.2

17.4–24.4 2.7 10.0–15.8 12.6

1.7–4.5

2.0

1.1–3.6

2.1

1.1–3.7

1.8

1.0–3.4

1.2

0.6–2.6

1.0

0.4–2.2

1.2

0.6–2.4

20.7

121 8 6 7 12 11 12 16 73

Drugs for obstructive respiratory disease (R03)

3.1

17

16 18.9–22.8 425

20.8 0.2–0.7

7

0.3 0.1–0.6

5

0.2 1.8–3.3

43

2.4 1.2–2.5

31

1.7 2.7–4.5

71

3.5 2.6–4.2

69

3.3 3.2–5.2

93

4.1

182

8.0

Antithrombotics (B01A)

6.8–9.3

Calcium with or without D vitamin (A12A) Therapeutic group (ATC)

Table 4 continued

Omega-3triglycerides (C10AX06)

Glucosamine (M01AX05)

Vitamins B (A11, A11E, B03B)

Multivitamins with minerals (A11A, A11B)

Iron (B03A)

Evening primrose oil

Vitamin C (A11GA01)

Any supplement

Rank

J. Peklar et al.

US national cohort [15]. For both types of drug–supplement combinations, the correlation and gradient between the extent of supplement use and the number of chronic conditions is clear-cut but, notably, the association was stronger and the gradient greater with non-VMs, fewer of which have precisely defined therapeutic indications, compared with VM combinations. Interestingly, for those reporting three or more conditions, combined usage increased in both sexes and in all age groups, whereas for those reporting one to two conditions, combined use decreased with age. The major transition point was 65 years of age; in those with three or more conditions, usage increased by over 15 % and, by contrast, in those with one to two conditions, usage decreased by over 10 % (Table 3). Moreover, taking five or more drugs (polypharmacy) was only significantly associated with VM but not non-VM use, which may reflect the use of frequent ‘fixed’ combination therapy which includes VMs as a part of a treatment, as the prevalence of combined use in self-reported conditions and with drugs for specific indications confirms (Tables 4, 5). Interestingly however, usage doubled in those taking more than ten drugs, suggesting that for several related reasons drug– supplement use increases in a non-linear fashion in the oldest age group and may reflect the use of all available means to treat multimorbidity. Thus, the highest frequency of combined use is associated with a constellation of factors which are more prevalent in the very old: low income, which is the criterion for receipt of free health care from the State, together with poor self-reported health status, multimorbidity and polypharmacy. Despite the extensive prescribing of calcium with or without vitamin D for women among those receiving medical care from the State, there was lower concurrent use of VMs and drugs in the poorest subpopulation in our study, even after controlling for all socioeconomic factors and for health insurance. In contrast, greater use of drug– supplement combinations was associated with third-level education. Furthermore, smoking, which is more prevalent in low-income groups, [17] was also associated with low combined use in this study and with low overall use in our previous analysis [7]; this association has also been reported elsewhere [18, 19]. Low use among patients in public health schemes may arise from inadequate care or from non-persistence with drugs or supplements, but as it may be combined with harmful behaviours this clearly indicates that these groups need to be targeted. Use of combinations was lowest overall in those who were employed, suggesting that those with more disposable income are less likely to use supplements of any type with drugs and this may indicate that this group is also not accessing all available means to control their conditions or to maintain their health. Like other countries, Ireland is

Concurrent Use of Drugs and Supplements Table 5 Potential major drug–supplement interactions stratified by age (weighted values adjusted to non-respondents and survey design to provide national estimates) 50–64 years n = 4,635

65–74years

C75 years

n = 2,137

n = 1,309

Total n = 8,081

Supplement

Drug

Potential interaction effect

References available

Calcium (with or without D vitamin)

Digoxin

Increased risk of arrhythmia and cardiovascular collapse

M,L,S

2

3

10

15

Iron

Levothyroxine

M,L,S

1

5

6

12

Omega-3-triglycerides

Warfarin

Increased risk of hypothyroidism Increased risk of bleeding

M,L,S

4

1

2

7

Evening primrose oil including combinations

Antithrombotics

Increased risk of bleeding

M,L

3

2

0

5

NSAID agents

Increased risk of stomach bleed

L

5

0

0

5

Garlic

Antithrombotics

Increased risk of bleeding

M,L,S

1

3

1

5

NSAID agents

Increased risk of stomach bleed

L

4

0

0

4

Dong quai

Antithrombotics

Increased risk of bleeding

M,L,S

1

0

0

1

Ginseng

Antithrombotics

Increased risk of bleeding

L

0

1

0

1

Ginger

NSAID agents

Increased risk of stomach bleed

L

1

0

0

1

22 (0.4, 0.2–0.6)

15 (0.7, 0.4–1.2)

19 (1.6, 1.0–2.5)

56 (0.7, 0.5–0.9)

Any interaction (%, 95 % CI)

M Thomson Micromedex, L Lexi-Comp, S Stockley’s Drug Interactions, NSAID non steroidal anti-inflammatory drugs, CI confidence interval

adopting a policy of Health and Wellbeing [20] and these findings show that its implementation will have to be tailored to the particular needs of different groups within the cohort of older people if it is to improve their health status. Combined drug–supplement use can provide therapeutic benefits and the most prevalent therapeutic drug group– supplement combination in our study, calcium with or without vitamin D supplements taken with drugs for osteoporosis, is recommended in evidence-based clinical guidelines [21]. However, only every second person (51.3 %; 95 % CI 46.4–56.2) who reported use of drugs indicated for osteoporosis, and only just over 42 % of women over 65 years of age who reported a diagnosis of osteoporosis (Table 4), were taking calcium supplements (with or without vitamin D). There was very little use of calcium with or without vitamin D preparations in both men and women not receiving drugs for bone disease, suggesting negligible use without prescription of an appropriate drug. This was confirmed in a separate analysis (not presented in this study) of the bone health assessment data of the cohort which showed that around half of those identified with scores consistent with osteopenia or osteoporosis were not taking a calcium with or without a vitamin D supplement and that fewer than 1 person in 25 of those without the criteria were taking this type of supplement [unpublished observations]. However, calcium and vitamin D supplementation is needed by many elderly in Ireland because a low intake of these nutrients among this group has been detected despite the finding that people over 65

years and women of all ages took more food supplements than other groups of adults in a recent national survey that measured both dietary intake and food supplement use over a 4-day period [22]. Taken together, these findings show that that there is a consistent pattern of unmet need that is not being met either by health service practices or addressed by current public health measures. Men reporting a diagnosis of diabetes were more likely than women to be taking a supplement, but fewer than one in ten men under 75 years of age were concurrent users, and use by respondents of both sexes taking drugs for diabetes (insulin or oral therapy) was the lowest prevalence for both supplement types at all ages, as in other countries. [6, 15] This patient group and their health care providers have been more intensively and comprehensively addressed about appropriate medical, dietary and lifestyle interventions; consequently, they may have a very cautious approach to the use of supplements. Overall, supplement use among men was also low, except for those with cardiac and arthritic conditions, which may reflect men’s reluctance to take action about their health as well as ineffective health care for this group. However, a conservative approach to supplement use with drugs may be a general tendency in those aged over 50 years as it was apparent in circumstances where there have been reports of clinical benefit that have been extensively discussed in the general media; for example, B vitamin use has been linked to prevention of cognitive decline and depression but its use was low in participants taking drugs for depression and

J. Peklar et al.

there was no use of evening primrose oil among participants taking hormone replacement therapy. The concomitant use of drugs and supplements for unrelated indications is not common but it may still result in unintended risk because it is unplanned. The risk in this study of potentially serious drug–supplement interactions was found to be more frequent in those C75 years of age and in women, which correlates with the high prevalence of both types of supplements in these groups. An increased prevalence with greater age was reported in a US national study, although more potential interactions were detected in men [3]. Using the national census data [23] within the included age groups to estimate the cohort-derived proportions as absolute numbers of people living in Ireland at the time, potentially suffering from the identified interactions, around 2,000 men and 6,200 women aged 50 years and over are at risk of a major drug–supplement interaction, most of whom are in the oldest age group (n = 3,550). The Irish population is predicted to grow substantially [24] and the proportions of those aged over 74 years and over 80 years will increase more than any others [24]. Concomitantly, multimorbidity and use of ten or more drugs would be expected to rise proportionately. Projection to the population level (adjusted for nonrespondents and survey design) suggests that the major potential interactions that are likely would occur among those taking antiarrhythmics, as 18.1 % (95 % CI 9.6–31.3) of them were also taking calcium supplements; those taking thyroid hormone replacement therapy, since 2.2 % (95 % CI 1.2–3.9) of them were also taking iron supplements; and in 0.5 % (95 % CI 0.2–1.4) and 0.8 % (95 % CI 0.3–2.0) of NSAID users with garlic or evening primrose oil, respectively. It was striking that drug–non-VM combinations were disproportionately associated with potentially serious interactions and also that both polypharmacy and polysupplement use was common in this group. Another interesting finding was the wide range of drug–food supplement combinations that gave rise to risk, and that often these added to existing drug–drug interactions, which illustrates the great diversity of sources of potential harm that health care practitioners need to be aware of when providing care, as others have argued [9]. Potential drug interactions cannot be assessed in isolation and so a more detailed examination of the characteristics of those identified as at risk of serious interactions in this study showed that there were other factors that would need to be taken into account to assess thoroughly the degree of risk incurred. Furthermore, besides the risk from the drug–food supplement combinations, some of the same respondents were also taking combinations of drugs that, whilst therapeutically appropriate, also require careful monitoring; in

one case not only were omega-3-triglycerides and glucosamine combined with warfarin, both of which may increase the risk of bleeding, but so also were aspirin and rosuvastatin, both of which increase the risk of bleeding with warfarin. Although polypharmacy is likely to predispose to drug interactions, this particular case illustrates that the type of drugs and the type of supplements are crucial— the respondent reported only four drugs and two supplements being taken, yet a complex set of multiple potentially serious interactions were generated. Not all potential drug–food supplement interactions will result in harm in practice, but without the detection and evaluation of many cases it will not be possible to distinguish the signal from the noise and understand the relevant factors associated with substantial risk. Irish national pharmacovigilance data (Irish Medical Board, personal communication) provided by the Irish Medical Board for the years 2007–2011 showed that of 13,755 spontaneous adverse reaction reports (ARRs) regarding (human) medications collected, only 35 related to herbal products and 25 to vitamin and mineral products. However, many products containing supplements are not subject to medicines regulation and there is therefore no requirement for pharmacovigilance reporting or for risk management plans. Given that only 0.4 % of ARRs contained information on supplement use in Ireland, which is less than in some other countries [25, 26], passive pharmacovigilance may not be sufficient to identify signals. To improve that, more should be done in the promotion of adverse drug reaction reporting, or even active surveillance may be necessary, as has been suggested in Canada [27], and where regulators of foods and medicines are separate organisations, collaboration would be necessary. 4.1 Strengths and Limitations The strengths of the present study are the nationally representative cohort with validated response quality [28] and the high response rate from which the sample of regular users of drugs and supplements was derived. Furthermore, drug data was collected by in-home interview, which is more reliable than the self-report recall method [29]. The limitations of the study are the absence of dose information, whether supplements were either prescribed or purchased, and information on the reasons for taking them, all of which constrained the interpretation of the results. In our analysis of drug–supplement use, only potential interactions were identified, as detailed clinical data, in particular data on clinical outcomes, were not available. Although the seriousness of the interactions was assessed using three standard clinical references, this estimate depends upon the robustness of the methods used by these sources to compile and interpret the interactions. The

Concurrent Use of Drugs and Supplements

overall potential risks may well be different from those identified in these sources, as a systematic review of the extensive literature suggests [9]. However, these reference sources are readily available to health care professionals and, if used, would trigger a further assessment of the patient’s case and therefore utilisation of additional health care resources irrespective of the provenance of the recommendation. Almost half the cases identified in this study involved multiple drugs and multiple supplements, and in these circumstances only rigorous clinical assessment and monitoring could verify and determine the seriousness of the interactions. 5 Conclusions Concurrent use of drugs and supplements among those aged over 50 years in the Irish population is substantial and has increased. There is considerable variation in usage from one subgroup to another as, with the exception of those drugs and supplements indicated for the treatment of osteoporosis, most of the patterns of use appear to be the result of individual case-by-case choices. The outcomes of this approach are evidence of an unmet need and therefore unrealised benefits among some subgroups, and of exposure to avoidable and potentially serious drug interactions among others. Both of these outcomes will increase in the future as the proportion of those aged over 50 years with multimorbidity grows, and polypharmacy and supplements use grows. Drug–supplement use could provide opportunities for both valuable treatment intensification and prevention, but without reliable information, adequate public awareness, professional guidance and practice support the extent of inadequate benefits and unanticipated risks will increase. This study suggests that national medicines policies must address each of the stakeholders to ensure responsible supplement use and to tailor pharmacovigilance initiatives to ensure that benefits are not foregone nor harm incurred. Acknowledgments The authors wish to thank the Irish Medicines Board for the pharmacovigilance data on food supplement use and to acknowledge the contribution of the TILDA participants and research staff. This work was supported by Irish Life; the Department for Health and Children; and by The Atlantic Philanthropies. The sponsors had no role in the design and conduct of the study; collection, management, analysis and interpretation of the data; or preparation, review and approval of the manuscript. The authors have no potential conflicts of interest that are directly relevant to the content of this study. The presented work has not been submitted elsewhere. Preliminary results were presented at a poster session at the British Geriatric Society Spring meeting in Belfast, April 2013. J. Peklar and K. Richardson had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: All authors. Analysis and interpretation of data: J. Peklar, M.C. Henman, M. Kos. Drafting of the manuscript: J. Peklar and M.C. Henman. Statistical analysis: J. Peklar, K. Richardson. Study supervision: M.C. Henman, M. Kos, R.A. Kenny.

References 1. Skeie G, Braaten T, Hjarta˚ker A, et al. Use of dietary supplements in the European Prospective Investigation into Cancer and Nutrition calibration study. Eur J Clin Nutr. 2009;63(Suppl 4):S226–38. 2. Kelly JP, Kaufman DW, Kelley K, et al. Recent trends in use of herbal and other natural products. Arch Intern Med. 2005;165(3):281–6. 3. Qato DM, Alexander GC, Conti RM, et al. Use of prescription and over-the-counter medications and dietary supplements among older adults in the United States. JAMA. 2008;300:2867–78. 4. Denison HJ, Jameson KA, Syddall HE, et al. Patterns of dietary supplement use among older men and women in the UK; findings from the Hertfordshire Cohort Study. J Nutr Health Aging. 2012;16(4):307–11. 5. Peng CC, Glassman PA, Trilli LE, et al. Incidence and severity of potential drug–dietary supplement interactions in primary care patients: an exploratory study of 2 outpatient practices. Arch Intern Med. 2004;164(6):630–6. 6. Nahin RL, Pecha M, Welmerink DB, et al. Concomitant use of prescription drugs and dietary supplements in ambulatory elderly people. J Am Geriatr Soc. 2009;57(7):1197–205. 7. Peklar J, Henman MC, Richardson K, et al. Food supplement use in community dwelling population aged 50 and over in the Republic of Ireland. Complement Ther Med. 2013;21(4):333–41. 8. Izzo AA, Ernst E. Interactions between herbal medicines and prescribed drugs. Drugs. 2009;69(13):1777–98. 9. Tsai HH, Lin HW, Pickard AS, et al. Evaluation of documented drug interactions and contraindications associated with herbs and dietary supplements: a systematic literature review. Int J Clin Pract. 2012;66:1056–78. 10. Kearney PM, Cronin H, O’Regan C, et al. Cohort profile: the Irish Longitudinal Study on Ageing. Int J Epidemiol. 2011;40(4):877–84. 11. Directive 2002/46/EC of the European Parliament and of the Council of 10 June 2002 on the approximation of the laws of the Member States relating to food. Official Journal L 183, 12/07/ 2002 P.0051–0057. http://eur-lex.europa.eu/LexUriServ/ LexUriServ.do?uri=CELEX:32002L0046:EN:HTML. Accessed 26 Jan 2013. 12. MicromedexÒ Healthcare Series. Thomson Healthcare. http:// www.micromedexsolutions.com/micromedex2/librarian/ND_T/ evidencexpert/ND_PR/evidencexpert/CS/5D1DCB/ND_App Product/evidencexpert/DUPLICATIONSHIELDSYNC/8AA6D1/ ND_PG/evidencexpert/ND_B/evidencexpert/ND_P/evidencexpert/ PFActionId/evidencexpert.FindDrugInteractions. Accessed 19 Feb 2013. 13. Lexi-Comp OnlineTM, Lexi-InteractTM. Hudson: Lexi-Comp, Inc.; 2011. Accessed 19 Feb 2013. http://online.lexi.com 14. Baxter K, editor. Stockley’s drug interactions [online]. London: Pharmaceutical Press. http://www.medicinescomplete.com. Accessed 19 Feb 2013. 15. Gardiner P, Graham RE, Legedza ATR, et al. Factors associated with dietary supplement use among prescription medication users. Arch Intern Med. 2006;166:1968–74.

J. Peklar et al. 16. Morgan TK, Williamson M, Pirotta M, et al. A national census of medicines use: a 24-hour snapshot of Australians aged 50 years and older. Med J Aust. 2012;196(1):50–3. 17. Barrett A, Savva G, Timonen V, et al. Fifty plus in Ireland 2011. First results from the Irish Longitudinal Study on Ageing, 2011 TILDA: Dublin; 2011. 18. Harrison RA, Holt D, Pattison DJ, et al. Are those in need taking dietary supplements? A survey of 21 923 adults. Br J Nutr. 2004;91(4):617–23. 19. Touvier M, Niravong M, Volatier JL, et al. Dietary patterns associated with vitamin/mineral supplement use and smoking among women of the E3N-EPIC cohort. Eur J Clin Nutr. 2009;63(1):39–47. 20. Irish Government, Department of Health. Healthy Ireland—a framework for improved health and wellbeing 2013–2025. 2013. http://www.dohc.ie/publications/Healthy_Ireland_Framework. html. Accessed 27 Nov 2013. 21. National Institute for Health and Clinical Excellence. Alendronate, etidronate, risedronate, raloxifene, strontium ranelate and teriparatide for the secondary prevention of osteoporotic fragility fractures in postmenopausal women. http://www.nice. org.uk/nicemedia/live/11748/42447/42447.pdf. Accessed 7 Nov 2012. 22. Irish Universities Nutrition Alliance. National Adult Nutrition Survey. http://www.iuna.net/?p=106. Accessed 27 Nov 2013. 23. Central Statistics Office. Statistical yearbook of Ireland 2011 edition. http://www.cso.ie/en/releasesandpublications/statistical

24.

25.

26.

27.

28.

29.

yearbookofireland/statisticalyearbookofireland2011edition/. Accessed 26 Sep 2012. Central Statistics Office. Population and labour force projection 2011—2041. 2012. http://www.cso.ie/en/releasesand publications/population/populationandlabourforceprojections 2011-2041/. Accessed 27 Nov 2013. Jacobsson I, Jo¨nsson AK, Gerde´n B, et al. Spontaneously reported adverse reactions in association with complementary and alternative medicine substances in Sweden. Pharmacoepidemiol Drug Saf. 2009;18:1039–47. Patel DN, Low WL, Tan LL, et al. Adverse events associated with the use of complementary medicine and health supplements: an analysis of reports in the Singapore Pharmacovigilance database from 1998 to 2009. Clin Toxicol. 2012;50(6):481–9. Vohra S, Cvijovic K, Boon H, et al. Study of Natural Health Product Adverse Reactions (SONAR): active surveillance of adverse events following concurrent natural health product and prescription drug use in community pharmacies. PLoS One. 2012;7(9):e45196. Richardson K, Kenny RA, Peklar J, et al. Agreement between patient interview data on prescription medication use and pharmacy records in those aged older than 50 years varied by therapeutic group and reporting of indicated health conditions. J Clin Epidemiol. 2013;66(11):1308–16. Qato DM, Schumm LP, Johnson M, et al. Medication data collection and coding in a home-based survey of older adults. J Gerontol B Psychol Sci Soc Sci. 2009;64B(S1):i86–93.