European Journal of Internal Medicine 25 (2014) 843–846

Contents lists available at ScienceDirect

European Journal of Internal Medicine

j ourna l homepage: www.e lsev ie r .com/ locate /e j im

Original Article

Understanding adverse drug reactions in older adults through drug–drug interactions

A. Marengoni a,⁎, L. Pasina b, C. Concoreggi c, G. Martini d, F. Brognoli d, A. Nobili b, G. Onder e, D. Bettoni f

a Geriatric Unit, Department of Clinical and Experimental Sciences, University of Brescia, Italyb Department of Neuroscience, IRCCS Istituto di Ricerche Farmacologiche Mario Negri, Milan, Italyc Intensive Brief Observation Unit, Emergency Room, Spedali Civili, Brescia, Italyd Haemostasis Centre Laboratory, Spedali Civili, Brescia, Italye Centro Medicina dell'Invecchiamento, Policlinico A. Gemelli, Università Cattolica del Sacro Cuore, Rome, Italyf Pharmacovigilance, Spedali Civili Pharmacy, Brescia, Italy

⁎ Corresponding author at: Department of ClinicaUniversity of Brescia, Geriatric Unit, Spedali Civili, PiBrescia, Italy. Tel.: +39 030 2528554/478; fax: +39 030

E-mail address: alessandra.marengoni@med.unibs.it (

http://dx.doi.org/10.1016/j.ejim.2014.10.0010953-6205/© 2014 European Federation of Internal Medi

a b s t r a c t

a r t i c l e i n f o

Article history:

Received 28 August 2014Received in revised form 29 September 2014Accepted 1 October 2014Available online 11 October 2014

Keywords:Adverse drug reactionDrug–drug interactionElderly

Aims: The aims of this study are to evaluate prevalence and characteristics of adverse drug reactions (ADRs) andto evaluate the potential contribution of specific medications, therapeutic categories and drug–drug interactions(DDIs) in older adults.Methods: All ADR reporting forms of persons aged 65+ years collected by the pharmacovigilance of one of themain hospitals in Italy during 2013 were evaluated. DDIs were analysed by a computerized prescription system(INTERCheck) and based on the interactions' databasemanaged by the Istituto di Ricerche FarmacologicheMarioNegri. DDIs were classified according to their clinical relevance as contraindicated, major, and moderate.Results: Amongst all the ADR reporting forms (n = 1014) collected during 2013, 343 affected older adults. Themost frequent ADRs were: haemorrhages (n = 122, 35.5%), allergic reactions (n = 56, 16.3%), and elevated

International Normalized Ratio (INR N 6, n = 54, 15.7%). The specific medications that contributed to ADRswere warfarin (42.5%), acenocumarol (9%), and allopurinol (8.5%); while the therapeutic categories werehaematological agents (67%) and proton pump inhibitors (13%). A total of 912 DDIs were found; onethird of them were contraindicated or major and 31.5% of them potentially contributed to ADRs; of these,the most frequent were: warfarin and heparin (contraindicated, n = 5); warfarin and a statin (major,n = 38); warfarin and a proton pump inhibitor (moderate, n = 40). At least one DDI contributed to 66haemorrhages out of 122 (54%) and to 41 elevated INR out of 54 (76%).Conclusion: DDIs significantly contribute to the onset of ADRs in older adults and intervention programmes,e.g., the employment of a computerized system,may reduce the burden of iatrogenic illnesses in the elderly.

© 2014 European Federation of Internal Medicine. Published by Elsevier B.V. All rights reserved.

1. Introduction

Due to the persistent exclusion of older persons from clinicaltrials, data on benefit–risk balance of several drugs in this sectionof the population are unknown. Research literature showed thatadverse drug reactions (ADRs) are frequent and severe, and areresponsible for considerable morbidity and mortality in older persons,leading to multiple admissions to the emergency room and acute hospi-talizations [1,2]. This means that ADRs have a significant and detrimentalimpact on the healthcare system's economy.

l and Experimental Sciences,azzale Spedali Civili 1, 251232528476.A. Marengoni).

cine. Published by Elsevier B.V. All rig

In a recent editorial, the EuropeanMedicinesAgencydefined a strategyto overcome several key issues in the pharmacological treatment of olderpersons. The Agency suggested that ‘spontaneous reports of adverseevents can be used to identify patterns of drug-disease and drug–druginteractions (DDIs) that were not apparent before authorization’ [3],with DDI being the ability of a drug to modify the action or effect ofanother drug administered successively or simultaneously. Theseinteractions have rarely been considered to cause ADRs in geriatricpersons compared to single pharmacologic agents, although they increasewith increasing age and the number of drugs consumed. In particular,amongst hospitalized older adults, the overlap of acute and chronicdiseases may increase the susceptibility to ADRs and their severity.

This study aimed to evaluate the occurrence and characteristics ofADRs in persons aged 65+ years attending one of the inpatient unitsand/or to the anticoagulant outpatient unit of one of the main hospitalsin Italy during 2013. The contribution of specificmedications, therapeutic

hts reserved.

844 A. Marengoni et al. / European Journal of Internal Medicine 25 (2014) 843–846

categories, and all possible contraindicated, major, and moderate DDIswere assessed.

2. Methods

In Italy, spontaneous ADRs are collected through the NationalNetwork of Pharmacovigilance, an extensive linkage that includes theRegional Centers of Pharmacovigilance, the Local Health Authorities,about one hundred hospitals, research institutes, and pharmaceuti-cal companies, as well as the Italian Agency of Medicine. The systemalso operates in connection with the European network forpharmacovigilance, EudraVigilance, which collects all data providedby the EU countries at national level in a single database.

Data analysed in this study were all spontaneous ADR reportingforms collected by the pharmacovigilance of the Spedali Civili, Brescia.The Spedali Civili is one of the main hospitals in Italy: in 2013, 72,132persons were admitted to the hospital and 1868 were examined in theanticoagulant outpatient unit. Of these, 27,700 and 1444 were aged65+ years, respectively. Usually, all the physicians are required tocomplete specific reporting forms and to submit them to the hospitalpharmacy if an ADR is suspected.

ADRs were here defined as a noxious and unintended response to amedicinal product that occurs at doses normally used in humans forthe prophylaxis, diagnosis, or therapy of disease or for the restoration,correction, or modification of physiological function as well as harmcaused by medication errors (in prescription, administration, or moni-toring of drug therapy) or patient non-adherence (adherence definedas ‘the extent to which patients take the medications as prescribed bytheir health care providers’) [4]. Severity of ADRs was classified asi)mild-moderate, ii) severewith hospitalization, iii) severewith clinicalcondition not requiring hospitalization, iv) life-threatening, andv) leading to permanent invalidity and to death.

Potential DDIs were analysed by a computerized prescriptionsupport system, i.e., INTERCheck, whichwas based on an Italian interac-tions' database managed by the Istituto di Ricerche FarmacologicheMario Negri, which has been already validated and described in detail[5,6].We classified all drug interactions in terms of clinical relevance,considering potential health outcomes, type, quality, and relevance ofsupporting clinical and pharmacological documentation. Coherentlywith other computerized support systems [7], each potential DDI wasclassified by clinical relevance as follows: contraindicated (drugcombination should be avoided or may potentially lead to seriousclinical consequences), major (the combination can be handled e.g. bydose adjustments or temporal separation, closemonitoring is required),moderate (clinical outcome of the interaction is uncertain and/or mayvary) and minor (drug combination probably has no clinical relevanceor has not been completely assessed). Spontaneous ADRs reportingformswere collected by the hospital pharmacy and stored in a databaseincluding age, sex,medications possibly implicated in the ADR, concom-itant medications, and narrative description of the adverse reaction aswell as its severity and outcome.

Number and prevalenceper 100 of ADRswere calculated for differentage groups; specific medications and therapeutic categories potentiallycontributing to ADRs were evaluated. The whole drug regimen of allthe subjects was analysed using INTERCheck for the calculation of DDIspotentially contributing or not to ADRs. In this study, only contraindi-cated, major, and moderate interactions were analysed.

The Ethical Committee of the Spedali Civili approved the study.

3. Results

Amongst all the spontaneous ADR reporting forms (n = 1014)collected by the hospital pharmacovigilance from January 1st toDecember 31st 2013, 343 (34%) affected people aged 65 years orolder (Fig. 1). The mean age of the patients was 76.4 years. The averagenumber of drugs per patient was 8.5. ADRs were equally distributed in

men (n = 165) and women (n = 178), whereas a lower prevalencewas found in the 60 very old (85+ years) persons compared to theyounger age group (17.5 versus 82.5%). Prevalence of ADRs in thetotal number of patients admitted during 2013 was 1.3%; prevalenceof ADRs in patients aged 65+ years was 1.2%. Spontaneous ADRreporting forms were mostly sent by the emergency room (n = 152)and the anticoagulant drug outpatient unit (n = 120), followed bythe clinical inpatient wards (20 cases by haematology, 10 cases bydermatology, 7 cases by oncology, 7 cases by neurology, 26 cases byother units). Amongst all the ADRs, 194 (56.5%) were mild-moderate,123 (35.9%)were severe requiringhospitalization, 16 (4.6%)were severebut not requiring hospitalization, 5 (1.5%)were life-threatening, 4 (1.2%)led to death, and 1 (0.3%) led to permanent invalidity. Themost frequentADRs were haemorrhages (n = 122, 35.5%), allergic reactions (n = 56,16.3%), elevated International Normalized Ratio (INR N 6, n = 54,15.7%), and neuropsychiatric/neurological effects (n = 26, 7.5%)(Table 1). Table 1 also shows the most frequent medications contribut-ing to ADRs, which were warfarin, acenocumarol, and allopurinol. Alsoshown are the most frequent therapeutic categories, which werehaematologic agents and proton pump inhibitors (Table 1).

Fig. 1 shows the number of the DDIs found in the pharmacologicalregimenof the patients that could have contributed to the correspondingADRs (Fig. 1). Amongst the DDIs potentially contributing to ADRs (n =288, 31.5%), the most frequent contraindicated DDIs were warfarin andheparin (5 cases). The most frequent major DDIs were warfarin and astatin (38 cases), warfarin and allopurinol (27 cases), and warfarin andamiodarone (22 cases) (Table 2). The most frequent moderate DDI waswarfarin and a proton pump inhibitor (41 cases).

Forty-eight out of 123 (39%) ADRs that were classified as ‘severerequiring hospitalization’ were potentially caused by DDIs, but noneled to death or were life-threatening. The analysis of the associationbetween DDIs and the most frequent ADRs showed that at least oneDDI contributed to 66 haemorrhages out of 122 (54%) and to 41 elevatedINR out of 54 (76%).

4. Discussion

Data for these analyses were gathered from 1-year ADR reportingforms of one of the main hospitals in Italy. Specific medications(i.e., warfarin, acenocumarol and allopurinol), as well as therapeuticcategories (i.e., haematological agents and proton pump inhibitors)mostly contributed to ADRs. Furthermore, a considerable number ofDDIs were found in the therapeutic regimens, potentially related orunrelated to the specific ADR and more importantly, about one thirdof them were contraindicated or major. Finally, at least one DDIexplained almost 40% of the ADRs classified as ‘severe requiring hospi-talization’ as well as the majority of the most frequent ADRs, i.e. morethan half of haemorrhages and two-third of elevated INRs.

We evaluated specific medications potentially contributing to ADRs,and our results highlighted the presence of drugs that typically causeADRs in older persons, such as warfarin [8], but also medications thatare less known, such as allopurinol and proton pump inhibitors. Thesedrugs are often inappropriately prescribed and considered safe in theelderly. For instance, proton pump inhibitors are frequently used fornon-evidence-based indications or in case of the use of a high numberof concomitant drugs [9]. Allopurinol is often inappropriately usedin asymptomatic hyperuricaemia [10], following some epidemiolog-ical and experimental studies that suggested a linkage betweenhyperuricaemia and hypertension [11,12], although there is poor evi-dence indicating a role of allopurinol in preventing cardiovasculardiseases [13]. A rational use of these drugs is warranted as not onlycan they have serious short- and long-term consequences, such ashypersensitivity syndrome or Steven–Johnson syndrome (allopurinol)[14], and intestinal and pulmonary infections (proton pump inhibitors)[15], but they are also frequently involved in DDIs.

All ADR repor�ng forms in 2013n=1014

ADR repor�ng forms in children (0-18 yrs)

n=266

ADR repor�ng forms in adults (19-64 yrs)

n=405ADR repor�ng forms in older persons (65+ yrs)

n=343

Total number of DDIsn=912

Number of DDIs poten�ally contribu�ng to ADRs

n=288

Number of DDIs poten�ally not contribu�ng to ADRs

n=624

Contraindicatedn=8

Severen=133

Moderaten=147

Contraindicatedn=43

Severen=124

Moderaten=457

Fig. 1. Number (N) of ADRs in the whole sample according to age and number of DDIs potentially contributing or not to ADRs in older persons.

Table 1Number of ADRs according to specific medications and therapeutic categories. Numberand proportion per 100 (on a total number of ADRs = 343).

Number Proportion (%)

Most common ADRsAll haemorrhages 122 35.5

Epistaxis 31 9.0Skin or wound haemorrhage 29 8.4Intracranial haemorrhage 25 7.2Gastrointestinal haemorrhage 17 4.9Genitourinary haemorrhage 14 4.0Others 6 1.7

Allergic reactions 56 16.3INRaN6 54 15.7Neurologic or neuropsychiatric effects 26 7.5Dizziness and syncope 16 4.6Gastrointestinal effects 15 4.3Hypoglycaemia or acidosis 8 2.3Osteonecrosis 5 1.4Most commonly implicated medications

Warfarin 146 42.5Acenocumarol 31 9.0Allopurinol 29 8.5Acetylsalicylic acid 26 7.6Amiodarone 23 6.7Simvastatin 19 5.5Amoxicillin/clavulanic acid 18 5.2Pantoprazole 16 4.6Omeprazole 15 4.4Atorvastatin 15 4.4

Most commonly implicated therapeutic categoriesHaematologic agents 230 67.0Proton pump inhibitors 44 12.8Statins 41 11.9NSAIDs 24 6.9Antibiotics 22 6.4

a INR denotes International Normalized Ratio.

845A. Marengoni et al. / European Journal of Internal Medicine 25 (2014) 843–846

Epidemiological studies of DDIs and the related risk of adverse clinicaloutcomes in elderly patients are rare and the prevalence of DDIs respon-sible for ADRs is not well-defined [16–19]. The prevalence of potentialDDIs in geriatric patients may vary according to the design of the study,the clinical setting, and the severity of DDIs. According to a large prospec-tive observational study, 1% of all hospital admissions are due toDDIs [19],but adverse events can be serious in the elderly. A literature review indi-cated that in the elderly, DDIswere responsible for 4.8% of the admissions[16]. The high prevalence of DDIs in elderly patients was suggested byanother review that showed a prevalence of DDIs between 15 and 45%in hospital, higher in patients with heart diseases and elderly people[20].

Our findings suggest that DDIs are an important risk factor for ADRs,which explain around 40% of the ADRs classified as severe requiringhospitalization, as well as the majority of the most frequent ADRs (i.e.haemorrhages and elevated INR). In line with previous studies in theelderly [21], the medications most frequently involved in DDIs were

Table 2Most frequent contraindicated and major DDIs potentially contributing to ADRs andpotential adverse effect.

N Potential adverse reaction

Contraindicated DDIs (all) 8Warfarin and heparin 5 Increased bleeding riskDabigatran and dronedarone 1 Increased dabigatran effectAcenocumarol and acetylsalicylic acid 1 Increased bleeding riskItraconazole and vincristine 1 Increased vincristine toxicity

Major DDIs (all) 133Warfarin and allopurinol 27 Increased bleeding riskWarfarin and amiodarone 22 Increased bleeding riskWarfarin and simvastatin 17 Increased anticoagulant effectWarfarin and acetylsalicylic acid 16 Increased bleeding riskWarfarin and atorvastatin 14 Increased anticoagulant effectWarfarin and rosuvastatin 6 Increased bleeding riskWarfarin and amoxicillin 4 Increased bleeding risk

846 A. Marengoni et al. / European Journal of Internal Medicine 25 (2014) 843–846

anticoagulants, warfarin above all. Indeed, DDIs between anticoagulantsand another medication explained the majority of ADRs linked to theuse of these drugs.

4.1. Limitations and strengths

The reported prevalence of ADRs in older persons varies in theliterature. Previous studies have found a higher effect than in our study[22]. However, it is worth noting that in our sample the prevalence ofADRs is significantly lower in the oldest old than in the younger elderly.In the oldest old, the presentation of ADRs may be unusual and they canbe difficult to recognize, especially by physicians who are not specializedin geriatrics. For example, the onset of geriatric syndromes, such as falls,delirium, and incontinence, may be wrongly attributed to diseasesrather than to ADRs. Secondly, despite the fact that all drug prescriptionsof patients were collected, we did not have any information on co-morbidities. Furthermore, we did not have information on the time intarget therapeutic range of warfarin and so we could not adjust therisk of haemorrhage for this variable.

5. Conclusions

Older adults are the main users of drugs. Therefore, the number ofADRs andDDIs is the highest in the population. Recentfindings, includingours, might have relevant implications to plan appropriate preventionprogrammes aimed at reducing the rate of iatrogenic illness in the oldpopulation. Given valuable information that can be obtained by alreadyexisting instruments, e.g., computerized prescription systems, our worksuggests to use these systems more in clinical settings and integratethem in a global approach capable of targeting all potential factorsinvolved in the onset of ADRs. In addition to a computer-based prescrib-ing system, a global approach in complex elderlymay includemedicationreview so that potentially inappropriate medications can be avoided anda comprehensive geriatric assessment can be achieved.

Learning points

• Specific medications (i.e., warfarin, acenocumarol and allopurinol), aswell as specific therapeutic categories (i.e., haematological agents andproton pump inhibitors) mostly contribute to ADRs in older adults.

• A large number of DDIs were found in the therapeutic regimens of thepatients and about one third of them were contraindicated or major.

• At least oneDDI explained almost 40% of the ADRs classified as ‘severerequiring hospitalization’ as well as the majority of the most frequentADRs, i.e. more than half of haemorrhages and two-third of elevatedINR.

Conflict of interest

The authors do not have any conflict of interest to disclose.

Acknowledgements

The authors gratefully thank Dr. Simona Ghibelli and Leonardo Peronifor their valuable contribution in analyzing drug–drug interactions.

References

[1] Atkin PA, Veitch PC, Veitch EM, Ogle SJ. The epidemiology of serious adverse drug reac-tions among the elderly. Drugs Aging 1999;14(2):141–52.

[2] Onder G, van der Cammen TJ, Petrovic M, Somers A, Rajkumar C. Strategies to reducethe risk of iatrogenic illness in complex older adults. Age Ageing 2013;42(3):284–91.

[3] Cerreta F, Eichler HG, Rasi G. Drug policy for an aging population—the EuropeanMedicines Agency's geriatric medicines strategy. N Engl J Med 2012;367(21):1972–4.

[4] Osterberg L, Blaschke T. Adherence to medication. N Engl J Med 2005;353(5):487–97.

[5] Garattini S, Nobili A. Interazioni tra farmaci: una valutazione della loro rilevanzaclinica. Pavia: Selecta Medica Eds.; 2001.

[6] Ghibelli S, Marengoni A, Djade CD, Nobili A, Tettamanti M, Franchi C. Prevention ofinappropriate prescribing in hospitalized older patients using a computerizedprescription support system (INTERcheck(®)). Drugs Aging 2013;30(10):821–8.

[7] Vonbach P, Dubied A, Krahenbuhl S, Beer JH, et al. Prevalence of drug–drug interac-tions at hospital entry and during hospital stay of patients in internal medicine. Eur JIntern Med 2008;19:413–20.

[8] Budnitz DS, Lovegrove MC, Shehab N, Richards CL. Emergency hospitalizations foradverse drug events in older Americans. N Engl J Med 2011;365:2002–12.

[9] Pasina L, Nobili A, Tettamanti M, Salerno F, Corrao S, Marengoni A, et al. REPOSI investi-gators. Prevalence and appropriateness of drug prescriptions for peptic ulcer andgastro-esophageal reflux disease in a cohort of hospitalized elderly. Eur J Intern Med2011;22(2):205–10.

[10] Mikuls TR, Farrar JT, Bilker WB, Fernandes S, Saag KG. Suboptimal physician adher-ence to quality indicators for the management of gout and asymptomatichyperuricaemia: results from the UK General Practice Research Database (GPRD).Rheumatology 2005;44(8):1038–42.

[11] Luk AJ, Levin GP, Moore EE, et al. Allopurinol and mortality in hyperuricaemic patients.Rheumatology 2009;48(7):804–6.

[12] Thanassoulis G, Brophy JM, Richard H, et al. Gout, allopurinol use, and heart failureoutcomes. Arch Intern Med 2010;170(15):1358–64.

[13] Gois PH, Souza ER. Pharmacotherapy for hyperuricemia in hypertensive patients.Cochrane Database Syst Rev Jan 31 2013;1:CD008652.

[14] Calogiuri G, Nettis E, Di Leo E, Foti C, Ferrannini A, Butani L. Allopurinol hypersensi-tivity reactions: desensitization strategies and new therapeutic alternative molecules.Inflamm Allergy Drug Targets 2013;12(1):19–28.

[15] Tleyjeh IM, Bin Abdulhak AA, Riaz M, Alasmari FA, Garbati MA, AlGhamdi M, et al.Association between proton pump inhibitor therapy and clostridium difficile infec-tion: a contemporary systematic review and meta-analysis. PLoS One 2012;7(12):e50836.

[16] Becker ML, Kallewaard M, Caspers PW, Visser LE, Leufkens HG, Stricker BH.Hospitalisations and emergency department visits due to drug–drug interactions:a literature review. Pharmacoepidemiol Drug Saf 2007;16:641–51.

[17] Hines LE, Murphy JE. Potentially harmful drug–drug interactions in the elderly: areview. Am J Geriatr Pharmacother 2011;9(6):364–77.

[18] Nobili A, Pasina L, Tettamanti M, Lucca U, Riva E, Marzona I, et al. Potentially severedrug interactions in elderly outpatients: results of an observational study of anadministrative prescription database. J Clin Pharm Ther 2009;34(4):377–86.

[19] PirmohamedM, James S, Meakin S, Green C, Scott AK,Walley TJ, et al. Adverse drug re-actions as cause of admission to hospital: prospective analysis of 18 820 patients. BMJ2004;329(7456):15–9.

[20] Espinosa-Bosch M, Santos-Ramos B, Gil-Navarro MV, Santos-Rubio MD, Marín-Gil R,Villacorta-Linaza P. Prevalence of drug interactions in hospital healthcare. Int J ClinPharmacol 2012;34(6):807–17.

[21] Obreli-Neto PR, Nobili A, de Oliveira Baldoni A, Guidoni CM, de Lyra Júnior DP, PilgerD, et al. Adverse drug reactions caused by drug–drug interactions in elderly outpa-tients: a prospective cohort study. Eur J Clin Pharmacol 2012;68(12):1667–76.

[22] Petrovic M, van der Cammen T, Onder G. Adverse drug reactions in older people:detection and prevention. Drugs Aging 2012;29(6):453–62.