Impact of findings on practice

  • There is a high prevalence of potential DDIs in prescriptions dispensed in community pharmacies in Greece.

  • Physicians such as psychiatrists, cardiologists and neurologists need to be more aware in order to avoid clinically meaningful DDIs than other medical specialties.

  • The implementation of an electronic surveillance system, which would provide pharmacists with the appropriate information to assess the degree of risk to the patient and to prevent potentially harmful DDIs, seems to be an effective action.

Introduction

Drug–drug interactions (DDIs) result in a modification of the action of one or more concurrently administered medications. DDIs are classified as pharmacodynamic or pharmacokinetic, and may result in increased or decreased efficacy, in treatment failure as well as in increased toxicity of medications [1, 2]. The problem is made even more complex by the concomitant use of over-the-counter medications and herbal remedies (e.g., St. John’s wort), certain types of food, ethanol and smoking. Although many DDIs exist, only a small part of these is clinically relevant [3–5]. Multiple drug treatment has been associated with the occurrence of DDIs, adverse drug reactions (ADRs), medication errors, and increased risk of hospitalization [6, 7] with several studies suggest that DDIs may be the cause of up to 3% of all hospital admissions [8–12]. Usually when multiple drug treatment is unavoidable, the potential benefits of drug combinations are weighted against the risk of the occurrence of a clinically significant DDI, taking into account the availability of alternatives.

Pharmacists play an important role in protecting the patients from the dangers posed by potential DDIs, especially with respect to drugs with a narrow therapeutic index [13]. Manual review of medications in a prescription can be performed by pharmacists, but the efficiency in the detection of DDIs is approximately 70% of DDIs in a two drug prescription, and the proportion decreases substantially as the number of medications increases [14]. By using computerized DDI screening programs we can significantly improve the identification of potentially harmful DDIs, beyond what can be achieved with manual review alone [15].

In Greece there is a lack of an effective screening system for detecting DDIs both at the level of prescribing by physicians and also at dispensing from community pharmacies. Almost all prescriptions are handwritten. Usually patient’s medication history can be accessed through manual observation of patient’s prescriptions book and all DDIs have to be identified manually. Although issues of improving primary health care and pharmaceutical care are of wide concern, due to the lack of DDIs monitoring systems, inappropriate prescription of drugs with potential DDIs causing serious risks to patients’ health has not been studied extensively [16–19]. Monitoring of potential DDIs may improve the quality of patient care [1].

Aim of the study

The objectives of this study were to evaluate the nature, type and frequency of potential DDIs in prescriptions dispensed from three community pharmacies in Thessaloniki, Northern Greece, during a 3-month period (November 2007–January 2008). Secondary objectives were the investigation of the relationship between specialty of prescribing physicians and the frequency of DDIs and to determine drug classes which were involved in potential DDIs.

Methods

The major region of Thessaloniki was stratified into three geographical areas, according to different population groups, including urban and rural regions, which are represented the city’s population. Following randomization, three community pharmacies (one from each geographical area) were selected in order to obtain a representative sample of the population of community pharmacies of Thessaloniki. All 1,553 prescriptions dispensed in these three community pharmacies through a 3-month period (November 2007–January 2008) were collected. Each participating pharmacist, following oral consent by the patient, removed all personal data from each prescription, including patients’ name, address and insurance information. The information collected included age, gender, date of the prescription, diagnosis, specialty of the prescribing physician, name of the medications in each prescription, dosage and quantity of medications dispensed.

Potential DDIs were detected using the Drug Interactions Checker within www.drugs.com database [20]. Furthermore, all medications were classified according to The Anatomical Therapeutic Chemical Classification System with Defined Daily Doses (ATC/DDD) [21]. The detected DDIs were classified as major, moderate and minor, depending to their severity of clinical significance and cross-over checked manually for the presence of enough published scientific evidence for the identified interacting agents [1, 22]. Major interactions are either well documented and have the potential of being harmful to the patient, or have a low incidence of occurrence (and perhaps limited documentation) and have the potential of serious adverse outcome. Moderate interactions are of moderate clinical significance, are less likely than major interactions to cause harm to the patient, or are less well documented. Minor interactions are of minor clinical significance. These interactions, regardless of the degree of their documentation, are least significant because of limited risk to the patient [23]. Due to interpatient variability, an interaction labeled as major may produce no harmful effects in some patients, while a moderate interaction may cause significant adverse effects. Only major and moderate DDIs were further analyzed statistically using Microsoft Access® and Microsoft Excel®.

Results

A total of 3,853 medications were prescribed to 1,553 prescriptions dispensed in three community pharmacies in Thessaloniki, Greece, during a 3-month period (November 2007–January 2008). The average number of medications per patient was found to be 2.5 (range 1–9). Of the 1,553 prescriptions 416 (26.8%) had only one medication and the remaining 1,137 prescriptions (73.2%) had two or more medications.

Prescriptions with two or more medications were checked for potential DDIs both electronically and manually [20, 22]. DDIs identified were classified as major and moderate, according to their level of clinical significance. In 213 prescriptions (13.7% of all prescriptions) at least one major or moderate DDI was identified and the total number of interactions was found to be 287 (18.5% of all prescriptions). Of these 287 DDIs, 30 were major (1.9% of all prescriptions corresponding to 10.5% of all interactions) and the remaining 257 DDIs (16.6% of all prescriptions corresponding to 89.5% of all interactions) were moderate (Table 1).

Table 1 DDIs identified in prescriptions dispensed in three community pharmacies
Full size table

Of the 30 major DDIs, 20 cases involved pharmacodynamic interaction and 10 cases involved pharmacokinetic interaction. Major DDIs were identified in prescriptions of patients with chronic diseases. The most frequent diagnosis was found to be coronary heart disease. The higher ratio (%) of major DDIs in relation to the number of prescriptions written from each medical specialty was observed by psychiatrists (6.6%), followed by cardiologists (5.0%), neurologists (4.7%), general practitioners (2.0%), primary care physicians (1.4%) and orthopaedics (1.3%; Fig. 1). The most frequent major DDIs identified in the prescriptions are given in Table 2. The ratio of DDIs/number of prescriptions increased with the number of medications prescribed per patient (Fig. 2).

Fig. 1
figure 1

Number of major and moderate DDIs per medical specialty. The number on the top of bars represent ratio (%) of DDIs/number of prescriptions of each medical specialty (n)

Full size image
Table 2 Frequency of most common major DDIs recognized in prescriptions dispensed in three community pharmacies
Full size table
Fig. 2
figure 2

Relationship between ratio of DDIs/number of prescriptions and number of medications prescribed per patient

Full size image

Medications that were involved in major DDIs belonged to several ATC system groups: 58.3% of them were agents acting on the cardiovascular system (ATC: C) 23.3% were agents acting on the nervous system (ATC: N) and 18.3% of the drugs were agents acting on other systems.

Discussion

The present study revealed that the overall incidence of potential DDIs in prescriptions dispensed in community pharmacies in Northern Greece was 18.5%. The identified DDIs were classified according to their severity [23], as major (the interaction may be life-threatening and/or require medical interventions to minimize or prevent serious adverse effects) and moderate (the interaction may result in an exacerbation of the patient’s condition and/or require an alteration in therapy). The prevalence of the major DDIs was 1.9% of all prescriptions dispensed (10.5% of all DDIs), whereas the prevalence of the moderate DDIs was 16.6% of all prescriptions dispensed (89.5% of all DDIs). Similar findings have been reported in several studies from other countries [24, 25]. Our findings clearly show that the ratio of potential DDIs/number of prescriptions is directly proportional to the number of medications prescribed per patient, in accordance with the results of previous studies [25].

Most potential DDIs were recognised in prescriptions written by cardiologists, primary care physicians and general practitioners. These physicians usually prescribe a greater number of medications per prescription than others. In the present study the most common major DDI was between ACE inhibitors (e.g., ramipril) or angiotensin II receptor antagonists (e.g., irbesartan) with potassium-sparing diuretics (e.g., amiloride) [26]. Some specific drugs were strongly associated with clinically relevant DDIs, such as amiodarone which interacted with a wide range of prescribed medications, including potassium-wasting diuretics (e.g., furosemide and hydrochlorothiazide) [27], digoxin, HMG-CoA reductase inhibitors (e.g., simvastatin) and coumarin anticoagulants (e.g., acenocoumarol) [28, 29].

The results give an indicative picture of the problem of DDIs in prescriptions dispensed in community pharmacies in Greece and demonstrate the need for the use of an improved reliable screening system for DDIs. Currently, all Greek pharmacies are equipped with computer software for dispensing medication, and therefore adding a drug-interaction utility to that software seems feasible in the near future. A larger study with more community or hospital pharmacies may give more reliable results. The use of electronic prescriptions, bar codes to assist in the identification of patients and their medications, the implementation of an accurate system providing novel scientific evidence, as well as careful selection of medication, are some of the suggestions strongly recommended nowadays to physicians [30–36]. Moreover, pharmacists have a key role in identifying and preventing potential DDIs in prescriptions dispensed to patients [37]. In order to improve surveillance of DDIs, several computer-based programs which recognize and warn about potential DDIs have been developed and evaluated [2, 38, 39]. However, several studies have demonstrated the inconsistency and overall limited reliability of DDI screening software to warn dispensing pharmacists of potential serious DDIs [40]. Often there is conflicting information among the various drug interaction compendia because classification of the severity of potential DDIs is not always straightforward. Therefore, the implementation of computer software capable of detecting DDIs at community or hospital pharmacies has not obviated the need for pharmacists to have a solid understanding of DDIs.

Limitations of this study were related to a certain degree of underreporting of potential DDIs because DDIs were detected only within one handwritten prescription and not reflecting a history, as well as without taking into account over-the-counter medications and herbal preparations (e.g., St. John’s wort) which may contribute to DDIs.

Conclusion

The main outcome of this descriptive study is that prescriptions dispensed in community pharmacies in Greece revealed a high frequency of DDIs in prescribed medications. In particular, patients on multiple drug therapy for cardiovascular diseases and patients receiving antidepressant medications should be closely monitored for adverse outcomes from potential DDIs. Identifying and preventing potentially harmful DDIs is a critical component of a pharmacist’s mission and the role of pharmacist should be shifted from drug-oriented to patient-oriented. Physicians especially psychiatrists, cardiologists, neurologists, primary care physicians and general practitioners, and pharmacists must remain vigilant in their monitoring of potential DDIs and make appropriate dosage or therapy adjustments. The implementation of a computer-based electronic system, which would provide pharmacists with the appropriate information to assess the degree of risk to the patient and to prevent an adverse outcome as a result of exposure to the interacting drugs, seems to be an appropriate action.