Adverse Drug Reactions (ADR) and Emergencies
The prevalence of suspected ADR in four emergency departments in Germany
; ; ; ; ; ; ; ; ;
Background: Adverse drug reactions (ADR) are a common reason for emergency room visits and for hospitalization. An ADR is said to have occurred when the patient’s symptoms and signs are considered to be possibly, probably, or definitely related to the intake of a drug.
Methods: In four large hospital emergency departments, one in each of four German cities ( Ulm, Fürth, Bonn, and Stuttgart), the percentage of suspected ADR cases among all patients presenting to the emergency room was determined during a 30-day period of observation. ADRs were ascertained by screening the digital records of all patients seen in the emergency room; causality was assessed as specified by the WHO-UMC (Uppsala Monitoring Center).
Results: ADR were sought in a total of 10 174 emergency department visits. 665 cases of suspected ADR were found, yielding a prevalence of 6.5%. The prevalence of ADR among patients with documented drug intake was 11.6%. Among the patients with documented suspected ADRs, 89% were hospitalized (in contrast to the 43.7% hospitalization rate in the entire group of 10 174 emergency department visits). A possible causal relationship between the patient’s symptoms and signs and the intake of a drug was found in 74–84% of cases. Patients with ADR were found to be taking a median of 7 different drugs simultaneously.
Conclusion: Adverse drug reactions are a relevant cause of emergency department visits, accounting for 6.5% of the total visits in this study, and often lead to hospital admission. The ADRED (Adverse Drug Reactions in Emergency Departments) study, which is now being conducted, is intended to shed further light on their causes, patient risk factors, and potential avoidability.
Adverse drug reactions (ADRs) are among the most common causes of iatrogenic injury in healthcare and have been identified as an essential research topic in the German Federal Ministry of Health‘s Action Plan for Improving Medication Safety in Germany (AMTS Action Plan 2016–2019) (1). ADRs can cause complications and inpatient hospital treatment, especially among vulnerable patients, such as the elderly (2–6). Furthermore, ADRs lead to increased costs in the healthcare system (4, 7–9). About 5–10% of hospital admissions can be attributed to ADRs (3, 4, 10, 11). The reported frequency, however, strongly depends on the study design. It is therefore important to understand the clinical setting in which the ADRs occurred, the method used to monitor for ADRs, the definition used for ADRs, and the patient population in the study (11–13). To date, only few systematic investigations on ADR-related hospital admissions have been carried out in Germany. An analysis within the network of regional pharmacovigilance centers showed that 3.25% of inpatient admissions were due to ADRs (7). In addition, there are a number of estimates, model analyses, and projections that report ADR rates between 2% and 20% in data from hospitals, emergency departments, or internist wards in Germany (8, 14–17).
An ADR is defined in the guideline of the European Medicines Agency (EMA) as a response to a medicinal product which is noxious and unintended. In this context, a drug is assumed to be the cause for an ADR when a causal relationship is considered to be at least possible (18). An ADR resulting from a medication error, that is, an unintended failure in the drug treatment process that can harm the patient, is considered to be preventable (19). While this implies that some ADRs can be avoided (4, 10, 11), there is still little known about how to accomplish this. Medication errors such as improper dosage or incorrect use of drugs due to contraindications (5, 20) may be a cause of preventable ADRs, but the proportion of ADRs due to these errors is not clear and until now has not been studied to a great extent. Therefore, a prospective case cohort of patients with severe ADRs leading to presentation in one of four large hospital emergency departments in Germany was collected prior to implementation of the ADRED study (Adverse Drug Reactions in Emergency Departments; DRKS-ID: DRKS00008979). This study focuses on collecting suspected ADR cases and analyzing them for causes and risk factors. In particular, the study aims to elucidate the proportion of potentially preventable ADRs and medication errors.
Before implementation of the ADRED study, a prospective collection of suspected ADR cases collection of suspected was conducted over a 30-day observation period from patients presenting at the emergency department of participating hospitals and then analyzed regarding frequency. The aim was to determine the proportion of suspected ADR cases in the total number of emergency treatments in the participating centers and to gain initial knowledge about the patient collective.
The study used a multicenter case collection from four emergency departments of maximum care or specialist care hospitals (namely, the Interdisciplinary Emergency Department of the University Hospital of Bonn, the Internal Medicine Emergency Department of the Ulm University Medical Center, the Central Emergency Department of the Hospital Fürth, and the Interdisciplinary Emergency Department of the Robert Bosch Hospital Stuttgart).
Study design and data collection
Using an observation period of 30 workdays, all cases of patients presenting to one of the respective emergency departments were assessed to determine whether, based on the clinical symptoms, the emergency visit was due to an ADR. This assessment was carried out by systematic screening of the patient files by the study staff, as well as by direct reporting by the hospital medical team to the study staff of suspected cases during the observation period. An assessment was also made as to whether, and to what extent, the symptoms had a possible temporal and causal relationship with a drug therapy. The study staff received advanced training with respect to both the definition of ADR and the causality assessment. They were sensitized to clinical issues, such as:
- Does the patient take medication?
- Has anything changed in the medication?
- Has the patient discontinued using a drug or recently started using an additional drug?
- Is the reason for the current emergency department visit an unexpected event?
- Are underlying diseases responsible for the acute event in the patient?
- Could a drug side effect be the reason for the current problem (for example, a fall due to dizziness)?
Additionally, the ADRED study includes frequent telephone conference calls about cases, to increase sensitivity to suspected ADRs. The term “suspected ADR” means that while a causal relationship in the individual case has not been proven, a relationship is initially assumed to exist between the medicine(s) taken and the symptoms or diagnoses. Adult patients (≥ 18 years old) with suspected ADRs were informed of the study both orally and in writing and were included in the study if they gave consent (Figure 1).
A causality assessment of the ADR-associated treatment cases was standardized according to the WHO-UMC (Uppsala Monitoring Center) system (21). Suspected ADR cases include patient treatment cases for which a likelihood of being related to a drug therapy was scored as “possible”, “probable”, or “certain” after causality assessment. Note that all recorded ADR cases are formally considered to be “suspected” ADRs, as causality is assessed retrospectively by a physician, even in cases for which causality is considered certain. According to the §62 German Medicinal Products Act (Arzneimittelgesetz, AMG), however, cases of suspected ADR are precisely those that must be collected for the pharmacovigilance and spontaneous reporting system. Cases deemed “unlikely”, “unclassified”, or “unclassifiable” were not included in the data analysis.
Cases were pseudonymized and prospectively documented in an electronic documentation form in a study database (OpenClinica, version 3.9). Questions about the cause and preventability of ADRs were also systematically collected and documented according to Schumock and Thornton (22). An overview of the collected data is shown in the eTable. For emergency department patients whose treatment was not associated with an ADR, no additional demographic or clinical data were collected in this study. After successful evaluation, a positive vote from the ethics committee was granted for conducting the study.
Data analysis and statistics
Demographic and clinical data of the study population, such as sex, age, number of active substances taken, causality assessment, type of treatment, and discharge status, were analyzed descriptively. The relationship between age and number of drugs was assessed using Spearman‘s rank correlation coefficients. In terms of sex distribution, causality assessment, type of treatment, and discharge status, comparisons were made between the two groups of adults—those 18–64 years old and the elderly (65 years and older)—with either chi-squared test or Fisher´s exact test.
If multiple drugs were found to be causally related in a suspected ADR case, only the highest probability score was included for the causality analysis. To test the hypothesis that more cases were considered “possible” in the older cohort than in the younger cohort, the ratings “certain” and “probable” were combined into one criterion. Preventability assessment of the ADR cases was not evaluated, as this is the primary endpoint of the ongoing ADRED trial, and higher case numbers are needed for such an analysis. Data were analyzed with the statistical software IBM® SPSS® Statistics Version 21.
In total, 10 174 treatment cases were analyzed for suspected ADRs during the 30-day observation period in the four emergency departments. Of this, 665 cases of suspected ADR were detected (Figure 1), corresponding to 6.5% of all treatment cases during the observation period. The proportion increased to 12% when only treatment cases were analyzed for which a drug therapy was recorded in the emergency admission protocol (n = 5719). A medication plan was recorded in the emergency departments in Bonn, Fürth, and Stuttgart for about every second patient, and was asked for each patient in the internal medicine emergency department in Ulm (Table). Of the patients included with suspected ADR, only 11% received outpatient care (versus 51% of total treatment cases).
A total of 352 suspected ADR cases were included in the ADRED study involving 350 patients (Figure 1); two patients with suspected ADR were treated twice in an emergency department during the observation period. The median age was 74.5 years (with a range from 18 to 97 years of age). 264 patients (75%) were 65 years old or older, and 55% of the patients were women (Figure 2).
Patients with suspected ADR used an average of seven different active substances at the same time (with a range of 1 to 18 active substances). The number of drugs taken increased with increasing age (Figure 3).
In the causality assessment, 287 cases were categorized as “possible”, 57 as “probable”, and 8 as “certain” (Figure 4). At the time of discharge, 265 patients (75%) had recovered without harm or were discharged in an improved condition (Figure 5).
The most common symptoms for ADR were gastrointestinal complaints and symptoms affecting the nervous system, each documented for approximately 32% of the cases, respectively (Figure 6). This included, among other things, abdominal pain, nausea, vomiting, hematemesis, melena, or vertigo and syncope. Frequently, symptoms were nonspecific and could not be further differentiated; around 30% of patients reported general symptoms, such as worsened general condition, pain, fever, and fatigue.
For the 352 ADR cases, 687 drugs were considered to have a probable causal relationship with the clinical picture. Of the ADRs classified as probable, antithrombotic and antihypertensive agents were the most commonly suspected cause of ADR (Figure 7).
A comparison between the age groups <65 years and ≥ 65 years showed that the cohort of the elderly had a statistically significantly more frequent admission rate (P <0.001) (Figure 8a). Both the health status at discharge and sex were similarly distributed in both age cohorts. The analysis of the causality assesment revealed a trend that in the elderly cohort the ADR cases were more frequently categorized as “possible”, and less frequently as “probable”, than the under-65 cohort (P = 0.056) (Figure 8b).
With a total of more than 10 000 cases of hospital emergency department visits collected in four large German hospitals over a period of 30 days, this a priori study is to our knowledge one of the largest systematic analysis of this topic in Germany. Additionally, following the 2012 update of the EMA Pharmacovigilance Guidelines and the redefinition of ADRs, this is the first study of its type that also includes medication errors and the use of drugs outside their intended use.
Our analysis showed that a suspected ADR was present in 6.5% of the more than 10 000 cases of emergency department visits. This proportion of suspected ADR cases doubles if only treatment cases with documentation of drug therapy are considered; however, both results are in line with the estimates of other German studies on the incidence of severe ADRs in clinical practice (8, 14–17).
Emergency departments of hospitals are often primary care providers of outpatient ADRs, as many of the numerous emergency patients are not hospitalized but rather treated on an outpatient basis (23, 24). This makes it difficult to detect an ADR in the emergency department or in an emergency situation, or to analyze and document its causes. Indeed, ongoing drug therapy was only documented for about half of all treatment cases. In particular, the rate of ADRs that result in a traumatic event could be underrepresented in the current dataset, as patients in an acute situation (for instance, trauma from syncope) may not have been asked about medication. A considerable amount of time and effort is needed for first questioning about whether drugs were taken and then filling out the corresponding documentation, which is not always possible in emergency departments. Additionally, not all emergency department patients are able to remember all the medications they take (25). In order to detect a suspected ADR as a reason for the emergency department visit by patients receiving drug therapy, both doctors and patients must be trained for awareness about ADR. The number of ADR cases in emergency patients who are registered as receiving drug therapy is significantly higher than in the overall study group. Notably, 12% of emergency patients who were documented to take medication had a suspected ADR.
In all four emergency departments, there was a tendency for patients with documented suspected ADRs to be hospitalized more frequently: only 11% of the enrolled patients with suspected ADRs were treated in the emergency room exclusively on an outpatient basis, as compared to 51% of the total treatment cases. This rate could also be due to outpatients leaving the hospital too quickly to be included in the study; this is especially conceivable in cases of treatment at night. However, it is also possible that many cases with ADR were so severe that patient hospitalization was necessary.
ADRs are difficult to recognize. As they often cause nonspecific symptoms, only the causal link to the medication can lead to the identification of an suspected ADR case. Indeed, general symptoms were among the most commonly reported symptoms in suspected ADR cases. However, it is much easier to suspect an ADR if the symptoms are obvious. For instance, antithrombotic drugs were assumed in many cases to have led to an ADR, and antihypertensive drugs that can cause dizziness and syncope were also often suspected (Figure 7). Still, it is always possible that the underlying diseases themselves were responsible for the symptoms. Therefore, only a few relationships could be classified as certain. An ADR can also aggravate the symptoms of an underlying disease, so that many possible relationships were taken into consideration.
Our data show that the more drugs are taken, the higher the chance of a suspected ADR; this is especially true for the elderly. Our data further highlight the impact of polypharmacy on the onset of ADRs, as patients with suspected ADR took seven different active substances (26–28) on average. By taking multiple medicines (29, 30), elderly patients in particular are more frequently affected by complications related to drug therapy (31). Furthermore, because of their state of health, elderly people are often more vulnerable to the development of an ADR (32). This could explain why elderly patients with ADRs are more frequently hospitalized than younger patients with ADRs. In addition, the ADR symptoms that led to the emergency department often cannot be associated with one precise drug and often include several nonspecific symptoms or complaints (33). However, falls or delirium can often be attributed to drug therapy (34, 35). It is also more difficult to exclude other factors in elderly patients, who often are multimedicated and have several co-morbidities. Thus, the trend towards a weaker assessment of the causal relationship (“possible” rather than “probable”) in the elderly can be explained by the fact that emergency department visits are often due to numerous, nonspecific symptoms, which cannot be traced back to a specific drug. For the total study population, the causality of the individual ADR cases was predominantly rated as “possible”; a relationship was established as “certain” in only in 2% of cases (Figure 4). The generally low proportion of ADRs identified as certain reflects the difficulty of classification. However, in order to first better understand and then address ADRs, possible and probable causal relationships should also be considered in line with the EMA Guideline on Good Pharmacovigilance Practice (18).
At more than 6%, ADRs are among the most common causes of health problems severe enough to lead to an unplanned hospital visit. Within the ADRED study, these cases will be prospectively recorded from four emergency departments in Germany, as a basis for a more precise analysis of the causes with respect to preventability and medication errors, as well as for analyzing patient risk factors and specifics of pharmacogenetics. In the future, this will provide an important contribution to a better understanding of drug therapy safety in care and clinical practices. Raising physicians’ and patients’ awareness to drug-related problems in routine care is important and should not be undervalued. For patients with nonspecific symptoms, and especially for the elderly, an ADR can only be suspected once a possible relationship with medication is brought to the attention of the caregivers. Increasing awareness of drug-related complaints, supported by the introduction of a mandatory medication management plan, will make an important contribution to the future improvement of drug therapy safety.
The study is supported by a financial project funding within the framework of the AMTS focus of the German Federal Ministry of Health (BMG), grant number ZMVI5–2514ATA004.
Conflict of interest statement
The authors declare that no conflict of interest exists.
Manuscript submitted on 7 July 2017, revised version accepted on
23 January 2018.
Translated from the original German by Dr. Veronica A. Raker.
Prof. Dr. med. Julia C. Stingl
Federal Institute for Drugs and Medical Devices
53175 Bonn, Germany
Dr. med. A. Marlen Schurig, Miriam Böhme, Dr. med. Katja S. Just, Dr. rer. nat. Catharina Scholl,
Prof. Dr. med. Julia C. Stingl
Central Emergency Department, Hospital Fürth: Prof. Dr. med. Harald Dormann, Bettina Plank-Kiegele
Internal Medicine Emergency Department, Ulm University Medical Center: Prof. Dr. med. Thomas Seufferlein
Interdisciplinary Emergency Department (INZ), University Hospital of Bonn: PD Dr. med. Ingo Gräff
Dr. Margarete Fischer Bosch Institute of Clinical Pharmacology (IKP), Stuttgart; Institute for Clinical Pharmacology and Institute for Pharmacy and Biochemistry, University Hospital Tübingen; German Cancer Consortium (DKTK) and German Cancer Research Center (DKFZ) Heidelberg: Prof. Dr. med. Matthias Schwab
Center for Translational Medicine, Medical Faculty, University Bonn:
Prof. Dr. med. Julia C. Stingl
|1.||Bundesministerium für Gesundheit, Aktionsplan 2016–2019 zur Verbesserung der Arzneimitteltherapiesicherheit in Deutschland. https://www.bundesgesundheitsministerium.de/fileadmin/Dateien/3_Downloads/A/Arzneimittelversorgung/Aktionsplan_AMTS_2016–2019.pdf (last accessed on 27. June 2017).|
|2.||Budnitz DS, Pollock DA, Weidenbach KN, Mendelsohn AB, Schroeder TJ, Annest JL: National surveillance of emergency department visits for outpatient adverse drug events. JAMA 2006; 296: 1858–66 CrossRef MEDLINE|
|3.||van der Hooft CS, Dieleman JP, Siemes C, et al.: Adverse drug reaction-related hospitalisations: a population-based cohort study. Pharmacoepidemiol Drug Saf 2008; 17: 365–71 CrossRef MEDLINE|
|4.||Pirmohamed M, James S, Meakin S, et al.: Adverse drug reactions as cause of admission to hospital: prospective analysis of 18 820 patients. BMJ 2004; 329: 15–9 CrossRef MEDLINE MEDLINE PubMed Central|
|5.||Chen YC, Huang HH, Fan JS, et al.: Comparing characteristics of adverse drug events between older and younger adults presenting to a Taiwan emergency department. Medicine (Baltimore) 2015; 94: e547 CrossRef MEDLINE PubMed Central|
|6.||Kim J, Parish AL: Polypharmacy and medication management in older adults. Nurs Clin North Am 2017; 52: 457–68 CrossRef MEDLINE|
|7.||Rottenkolber D, Schmiedl S, Rottenkolber M, et al.: Adverse drug reactions in Germany: direct costs of internal medicine hospitalizations. Pharmacoepidemiol Drug Saf 2011; 20: 626–34 CrossRef MEDLINE|
|8.||Schneeweiss S, Hasford J, Gottler M, Hoffmann A, Riethling AK, Avorn J: Admissions caused by adverse drug events to internal medicine and emergency departments in hospitals: a longitudinal population-based study. Eur J Clin Pharmacol 2002; 58: 285–91 CrossRef MEDLINE|
|9.||Schrappe M: [Patient safety: a topic of the future, the future of the topic]. Bundesgesundheitsblatt Gesundheitsforschung Gesundheitsschutz 2015; 58: 4–9 CrossRef MEDLINE|
|10.||Leendertse AJ, Egberts AC, Stoker LJ, van den Bemt PM: Frequency of and risk factors for preventable medication-related hospital admissions in the Netherlands. Arch Intern Med 2008; 168: 1890–6 MEDLINE|
|11.||Stang AS, Wingert AS, Hartling L, Plint AC: Adverse events related to emergency department care: a systematic review. PLoS One 2013; 8: e74214 CrossRef MEDLINE PubMed Central|
|12.||Muehlberger N, Schneeweiss S, Hasford J: Adverse drug reaction monitoring—cost and benefit considerations. Part I: frequency of adverse drug reactions causing hospital admissions. Pharmacoepidemiol Drug Saf 1997; 6, Suppl 3: 71–7 CrossRef|
|13.||Bouvy JC, de Bruin ML, Koopmanschap MA: Epidemiology of adverse drug reactions in Europe: a review of recent observational studies. Drug Saf 2015; 38: 437–53 CrossRef MEDLINE PubMed Central|
|14.||Meier F, Maas R, Sonst A, et al.: Adverse drug events in patients admitted to an emergency department: an analysis of direct costs. Pharmacoepidemiol Drug Saf 2015; 24: 176–86 CrossRef MEDLINE|
|15.||Schwake L, Wollenschlager I, Stremmel W, Encke J: Adverse drug reactions and deliberate self-poisoning as cause of admission to the intensive care unit: a 1-year prospective observational cohort study. Intensive Care Med 2009; 35: 266–74 CrossRef MEDLINE|
|16.||Dormann H, Criegee-Rieck M, Neubert A, et al.: Lack of awareness of community-acquired adverse drug reactions upon hospital admission: dimensions and consequences of a dilemma. Drug Saf 2003; 26: 353–62 CrossRef MEDLINE|
|17.||Dormann H, Sonst A, Muller F, et al.: Adverse drug events in older patients admitted as an emergency: the role of potentially inappropriate medication in elderly people (PRISCUS). Dtsch Arztebl Int 2013; 110: 213–9 VOLLTEXT|
|18.||EMA: Guideline on good pharmacovigilance practices (GVP), annex I—definitions. www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2013/05/WC500143294.pdf (last accessed on 6 December 2017).|
|19.||EMA: Good practice guide on recording, coding, reporting and assessment of medication errors. www.ema.europa.eu/docs/en_GB/document_library/Regulatory_and_procedural_guideline/2015/11/WC500196979.pdf (last accessed on 6 December 2017).|
|20.||Huckels-Baumgart S, Manser T: Identifying medication error chains from critical incident reports: a new analytic approach. J Clin Pharmacol 2014; 54: 1188–97 CrossRef MEDLINE|
|21.||The use of the WHO-UMC system for standardised case causality assessment: www.who-umc.org/media/2768/standardised-case-causality-assessment.pdf (last accessed on 12 June 2017).|
|22.||Schumock GT, Thornton JP: Focusing on the preventability of adverse drug reactions. Hosp Pharm 1992; 27: 538 CrossRef|
|23.||Plank-Kiegele B, Burkle T, Muller F, et al.: Data requirements for the correct identification of medication errors and adverse drug events in patients presenting at an emergency department. Methods Inf Med 2017; 56: 276–82 CrossRef MEDLINE|
|24.||Schneeweiss S, Gottler M, Hasford J, et al.: First results from an intensified monitoring system to estimate drug related hospital admissions. Br J Clin Pharmacol 2001; 52: 196–200 CrossRef MEDLINE PubMed Central|
|25.||Davis CP: Emergency department visits: we are not prepared. Am J Emerg Med 2012; 30: 1364–70 CrossRef MEDLINE|
|26.||Krüger-Brand H: Arzneimitteltherapie: Medikationsplan für mehr Sicherheit. Dtsch Arztebl 2015; 112: A-1410.|
|27.||Zopf Y, Rabe C, Neubert A, Hahn EG, Dormann H: Risk factors associated with adverse drug reactions following hospital admission: a prospective analysis of 907 patients in two German university hospitals. Drug Saf 2008; 31: 789–98 CrossRef|
|28.||Köberlein-Neu J, Mennemann H, Hamacher S, et al.: Interprofessional medication management in patients with multiple morbidities: a cluster-randomized trial (the WestGem Study). Dtsch Arztebl Int 2016; 113: 741–8 VOLLTEXT|
|29.||Hajjar ER, Cafiero AC, Hanlon JT: Polypharmacy in elderly patients. Am J Geriatr Pharmacother 2007; 5: 345–51 CrossRef MEDLINE|
|30.||Moßhammer D, Haumann H, Mörike K, Joos S: Polypharmacy—an upward trend with unpredictable effects. Dtsch Arztebl Int 2016; 113: 627–33 VOLLTEXT|
|31.||Nguyen JK, Fouts MM, Kotabe SE, Lo E: Polypharmacy as a risk factor for adverse drug reactions in geriatric nursing home residents. Am J Geriatr Pharmacother 2006; 4: 36–41 CrossRef MEDLINE|
|32.||Hubbard RE, O‘Mahony MS, Woodhouse KW: Medication prescribing in frail older people. Eur J Clin Pharmacol 2013; 69: 319–26 CrossRef MEDLINE|
|33.||Nickel CH, Ruedinger JM, Messmer AS, et al.: Drug-related emergency department visits by elderly patients presenting with non-specific complaints. Scand J Trauma Resusc Emerg Med 2013; 21: 15 CrossRef MEDLINE PubMed Central|
|34.||Huang AR, Mallet L, Rochefort CM, Eguale T, Buckeridge DL, Tamblyn R: Medication-related falls in the elderly: causative factors and preventive strategies. Drugs Aging 2012; 29: 359–76 CrossRef MEDLINE|
|35.||Young J, Inouye SK: Delirium in older people. BMJ 2007; 334: 842–6 CrossRef MEDLINE PubMed Central|