Cardiovascular drug adverse reactions in hospitalized patients in cardiac care unit

The aim of the present survey was to evaluate different aspects of cardiovascular drug adverse reactions in a cardiac care unit (CCU). All patients admitted to CCU during a 16 months period were recruited into this study. Detection of adverse drug reactions (ADRs) was based on daily chart review and face to face interview with patients. Causality assessment was performed by the World Health Organization (WHO) probability criteria. Seriousness of ADRs was determined by WHO definition. Schumock and Thornton scale was applied to assess preventability of ADRs. Statistical analysis was performed. Among 740 cardiovascular patients admitted to CCU, 70 ADRs were recorded from 44 patients. Headache (15.71%) was the most frequent ADR. The highest ADR rates were attributed to digoxin (44.29%) and atenolol (12.86%). Fifty five (78.57%) of ADRs were serious. The rate of preventable ADRs was 62.86%. Regarding outcome, one (1.43%) ADR led to death. Multivariate logistic regression showed that length of CCU stay (OR = 1.09, 95%CI = 1.02-1.17) and non-ischemic heart diseases (OR = 3.26, 95% CI = 1.57-6.78) were risk factors for ADR occurrence. Cardiovascular drugs could develop fatal adverse reactions in CCU patients. Primary admission diagnosis and duration of CCU stay were risk factors for ADR development.


INTRODUCTION
Adverse drug reactions (ADRs) are an important cause of morbidity and mortality (Ramirez et al., 2009).Current studies reveal that ADRs occur in 6.5% to more than 20% of hospitalized patients.These result in prolonged hospitalization, increased cost and complicated treatment (Classen, 2003).According to results of a study on 8,208,960 medicare patients admitted to the United States (US) hospitals in 1998, 141,398 patients (1.73% experienced at least one ADR.The drug classes most frequently associated with ADRs in that study include cardiotonic glycosides, adrenal corticosteroids, antineoantineoplastic agents, anticoagulants and analgesics.*Corresponding author.E-mail: khgholami@sina.tums.ac.ir.Tel:/Fax: +98 (21) 66954709.
ADRs caused $516,034,829 increase in total charges, $37,611,868 in drug charges and $9,456,698 in laboratory charges (Bond and Raehl, 2006).Cardiovascular drugs were the most common class of drugs (36.3%) involved in ADR-related admissions to medicine and cardiology wards of a university hospital in Sweden (Mjörndal et al., 2002).Similarly, in a cohort study conducted in the Netherlands, antithrombotics cardiovascular drugs, antineoplastics immunosuppressives and central nervous system (CNS) drugs accounted for the majority of ADR-related hospitalizations (van der Hooft et al., 2008).Adverse reactions induced by cardiovascular drugs during hospital stay were the subject of several studies.However, to the best of our knowledge, different aspects of cardiovascular drug adverse reactions in coronary care settings have been considered only in very few studies.The aim of the present study was to evaluate adverse reactions of cardiovascular drug in a cardiac care unit (CCU) at a teaching hospital in Iran.

METHODS
From early March 2006 to early July 2007, all patients admitted to CCU of Namazi hospital, a multispeciality healthcare university setting in Shiraz, were included in this study.The hospital's Ethics Committee approved the study.Cardiovascular medication group is defined as drug classes including diuretics, peripheral and central adrenergic inhibitors, direct vasodilators, calcium channel blockers, angiotensin-converting enzyme inhibitors (ACEIs), angiotensin II receptor blockers, positive inotropic agents, antiarrhythmics, anticoagulants, fibrinolytics, and antiplatelets.
At the beginning of this cross-sectional study, yellow cards used in Iran for reporting ADRs were introduced to healthcare team in CCU including doctors and nurses.They were asked to record ADRs due to cardiovascular agents during CCU stay or ADRs of cardiovascular medications which led to CCU admission.Daily chart review and face to face interview with patients were done for ADR detection.The WHO definition of ADR was used in this study: "Any noxious or unintended response to a drug, which occurs at doses normally used in human for the prophylaxis, diagnosis or treatment of disease or for the modification of physiological function" (Edwards and Aronson, 2000).Data on ADR yellow cards including patient information (demographic characteristics, other simultaneous diseases, and possible allergy or past ADR history), information about the suspected medicine (name, dose, frequency, route of administration, and indication), concomitantly drugs used and information about the observed ADRs (onset, duration, outcome, clinical manifestation, and laboratory findings), and possible actions for management of detected ADRs were registered.To confirm the collected data, yellow cards were reviewed by the clinical pharmacist.ADR yellow cards were then sent to Iranian Pharmacovigilance Centre.Reported ADRs were categorized according to WHO adverse reaction terminology (WHO-ART) organ-system classification (Meyboom et al., 1997a).In order to determine the causality relationship between reported ADRs and suspected medications, WHO probability scale was used (Meyboom et al., 1997b).According to WHO criteria, ADRs which result in death, life-threatening situation, persistent or significant disability/incapacity, hospital admission or prolonged existing hospital stay were categorized as serious (Safety monitoring of medicinal products, 2000).Preventability of ADRs was assessed based on Schumock and Thornton questionnaire (Schumock and Thornton, 1992).

Statistical analysis
Categorical variables were expressed as percentage.Continuous variables were reported as mean ± standard deviation (SD).The study subjects were divided into 2 groups: those with ADRs and those without.The association between ADR occurrence and demographic and clinical characteristics including sex, age, length of CCU stay, primary admission diagnosis, concomitant cardiovascular diseases, and total number of drugs taken during CCU stay was examined by both univariate and multivariate logistic regression analysis to calculate odds ratios (OR) and their 95% confidence intervals (CI).P value less than 0.05 was considered to be statistically significant.Statistical analysis was performed by the SPSS 11.5 program (SPSS Inc., Chicago, IL, USA).

RESULTS
During the study period (16 months), 740 patients were admitted to the CCU.Demographic characteristics and primary admission diagnosis of patients are shown in Table 1.Of the entire subjects, 54.32% were males.More   Table 4 provides causality assessment results of detected ADRs.Most ADRs (62.86%) were recognized to be possible.Regarding the outcome of ADRs, 67was as a result of gastrointestinal (GI) bleeding attributed to streptokinase.The reaction occurred in a 67-year old male patient who received 1,500,000 units streptokinase intravenously for ST-elevated myocardial infarction.GI bleeding developed 1 day after streptokinase initiation.Despite GI washing through nasogastric tube, he died 2 days later.Among 70 ADRs, 44 (62.86%) was identified as preventable.Details of ADRs considered to be preventable are listed in Table 5.Among seven preventability criteria of Schumock and Thornton's questionnaire, drug-drug interactions and lack of regular therapeutic drug monitoring accounted for most of preventable ADRs.
The management of ADRs involved discontinuation of causative medications in combination with additional therapeutic measurements (68.57%), withdrawal or reduction the dosage of drugs causing ADRs without any further treatment (18.57%), and additional therapeutic measurements without cessation of the offending medications (10%).Two (2.86%) ADRs need no management.According to results of univariate logistic regression analysis, age (p = 0.022), gender (p = 0.017), length of CCU stay (p = 0.001), primary admission diagnosis (p < 0.001), and concomitant cardiovascular diseases (p < 0.001) were identified as possible risk factors of cardiovascular drug adverse reactions.However, multivariate logistic regression analysis revealed that only length of CCU stay (OR = 1.09, 95% CI = 1.02-1.17,p = 0.008) and non-ischemic heart diseases as primary admission diagnosis (OR = 3.26, 95% CI = 1.57-6.78,p = 0.002) were independent risk factors for experiencing an ADR (Table 6).

DISCUSSION
The frequency of ADRs in the present study was 44/740 (5.95%).This rate is much lower than the previous study performed in outpatients with cardiovascular diseases.In that study, 105 out of 518 (20.3%) outpatients developed at least 1 cardiovascular drug adverse reactions (Gholami et al., 2008).In a study conducted by Mohebbi and co-workers on 677 patients admitted to CCU over an 8-months period, the rate of ADRs was 24.2% (Mohebbi et al., 2010).In another study in Danish department of cardiology, the rate of ADRs attributed to cardiovascular agents was 15.3% (Davidsen et al., 1988).
In the current study, 2.84% of CCU admissions were ascribed to ADRs.The rate of ADR-related admissions reported in other studies on Iranian population was quite different.According to a study conducted on 370 Iranian patients in general wards, ADRs has been identified as the cause of 8% of total hospitalizations (Gholami and Shalviri, 1999).In another survey at infectious disease department of a hospital in Iran, 2.2% of all admissions were due to anti-infective ADRs (Gholami et al., 2005).In two further studies in Iranian population, the rate of ADRs causing hospitalization in infectious disease and internal medicine wards were 5.4% (Javadi et al., 2007).and 1.75% (Pourseyed et al., 2009), respectively.In line with result of our present study, of 2,559 admissions to an in tensive cardiac care unit (ICCU) in Israel, 64 (2.5%) were due to major cardiac iatrogenic events.Major cardiac iatrogenic events were defined as life-threatening or serious problems caused by diagnostic procedure or therapy (Hammerman and Kapeliovich, 2000).Digoxin, atenolol, and streptokinase were the most offending cardiovascular drugs in the current study.In the survey in outpatients, calcium channel blockers especially diltiazem had the highest rate of ADRs (Gholami et al., 2008).The highest number of ADRs in Mohebbi et al. (2010) study was caused by nitroglycerin.In Danish trial, almost 60% of all ADR-related admissions as well as 80% of all detected ADRs were attributed to thiazide diuretics, beta blockers, and calcium channel antagonists (Davidsen et al., 1988).Among 64 admissions to ICCU which were due to major cardiac iatrogenic events, 58 (91%) patients suffered from arrhythmias caused by beta blockers, calcium channel blockers, amiodarone or combination of these drugs (Hammerman and Kapeliovich, 2000).In contrast to our results, digitoxicity was only proven in 1 patient as the cause of ICCU admissions.The authors explained their latter finding by a decrease in the number of patients treated with digitalis as well as regular monitoring of digoxin blood levels in the study population (Hammerman and Kapeliovich, 2000).According to results of Zaidenstein and co-workers study in an internal medicine ward in Israel, beta blockers and warfarin together were responsible for 40% of detected ADRs (Zaidenstein et al., 2002).
The highest rate of ADRs reported in this study were headache (15.71%), dizziness (14.29%), and nausea and vomiting (11.43%).Central nervous (37.14%) and gastrointestinal (21.43%) systems were considered to be the most frequent affected organ-systems.These results are in exact agreement with the previous survey (Gholami et al., 2008).Similar results were observed in Mohebbi et al. (2010) study.Lazarou et al. (1998) reported a 6.7% overall incidence of serious ADRs and mortality of 0.32%.The incidence of serious ADRs in our survey was 4.05% which is comparable to Zaidenstein et al. (2002) finding (4%).In contrast, incidence of serious ADRs in Iranian outpatients with cardiovascular disease (0.58%) was much lower than those reported in the present study (Gholami et al., 2008).Mohebbi et al. (2010) reported a higher incidence of serious ADRs in their study (5.47%).Mean ± SD age of patients with serious ADRs was more than the opposite group (69.53 ± 9 vs. 57.43± 15.18 years).However, given the limited number of nonserious ADRs, comparison between age as well as other demographic charactristics and number of cardiovascular drug given in patients with and without serious ADRs was not statistically feasible.
Similar to seriousness, the rate of preventable ADRs in this study (62.86%) is much higher than the value we observed previously (1.9%) (Gholami et al., 2008).
According to a systematic review conducted on eight studies published between 1991 and 2006, the median overall incidence of in-hospital adverse events was 9.2% and almost 50% of  (Ferner, 2003).In Zaidenstein and co-workers study, patients with adverse drug events induced by cardiovascular agents stayed about 2 days more in hospital compared to those without adverse drug events (p = 0.018) (Zaidenstein et al., 2002).Duration of hospitalization was significantly longer in patients who experienced an ADR in an infectious diseases ward in Iran (14.6 vs. 10.2 days, p < 0.001) (Kourorian et al., 2009).In the current study, patients with non-ischemic heart diseases as primary admission diagnosis and those who had a longer CCU stay were at risk of experiencing an ADR.
The difference between results of our survey and other studies about cardiovascular drug adverse reactions could be attributed to ethnic differences, different possible risk factors for ADR development (e.g.demographic characteristics, type and number of coadministered drugs, duration of therapy, length of stay in hospital, and comorbidities), and variable sources of detecting and reporting ADRs.In this regard for example, a pharmacist looked for ADRs in Mohebbi et al. (2010) study whereas in the current survey, doctors and nurses detected and reported ADRs.

Conclusion
Our data demonstrate that cardiovascular drugs could develop serious and even fatal adverse reactions in CCU patients.Primary admission diagnosis and duration of CCU stay were risk factors for ADR development.Since most of detected ADRs in this study were preventable, strategies such as ADR surveillance centre development in hospitals and healthcare team training could substantially reduce ADR occurrence in medical settings.

Table 1 .
Demographic characteristics and primary admission diagnosis of total patients admitted to the CCU (n = 740).

Table 2 .
Clinical manifestations, frequency, sex ratio, and cardiovascular drugs suspected of causing adverse reactions (n = 70).
clopidogrel, captopril, furosemide, and aspirin (1.43%) the least ADRs.All reported ADRs attributed to digoxin and atenolol resulted in admission to the CCU.Organsystems affected by cardiovascular drug adverse reactions are demonstrated in Table3.The 3 most common involved organ-systems were central and

Table 6 .
Demographic and clinical characteristics of patients with and without adverse drug reactions (n = 740).

Total number of drugs taken during CCU stay
ADRs: adverse drug reactions; OR: odds ratio; CI: confidence interval; CCU: cardiac care unit.† Other than primary admission diagnosis.