Potentially Harmful Drug–Drug Interactions and Their Associated Factors Among Hospitalized Cardiac Patients: A Cross-Sectional Study

Abstract

Background

Cardiovascular diseases are responsible for a significant proportion of mortalities worldwide. Elderly patients are the most affected by cardiovascular diseases, and because of factors such as polypharmacy, multimorbidity, and age-related changes in drug availability and metabolism, they are highly susceptible to the occurrence of drug–drug interactions. Drug–drug interactions are among the many drug-related problems leading to negative outcomes among inpatients and outpatients. Thus, it is important to investigate the prevalence, involved drugs, and factors related to potential drug–drug interactions (pDDIs) to properly optimize pharmacotherapy regimens for these patients.

Objective

We aimed to determine the prevalence of pDDIs, drugs most frequently implicated, and significant predictors associated with these interactions among hospitalized patients in the Cardiology Unit at Sultan Qaboos University Hospital in Muscat, Oman.

Methods

This retrospective cross-sectional study included 215 patients. Micromedex Drug-Reax^® was used to identify pDDIs. Data extracted from patients’ medical records were collected and analyzed. Univariable and multivariable linear regression was applied to determine the predictors associated with the observed pDDIs.

Results

A total of 2057 pDDIs were identified, with a median of nine (5–12) pDDIs per patient. Patients with at least one pDDI accounted for 97.2% of all the included patients. The majority of pDDIs were of major severity (52.6%), fair level of documentation (45.5%), and pharmacodynamic basis (55.9%). Potential drug–drug interactions between atorvastatin and clopidogrel were the most frequently observed (9%). Of all the detected pDDIs, around 79.6% of them included at least one antiplatelet drug. Having diabetes mellitus as a comorbidity (B = 2.564, p < 0.001) and the number of drugs taken during the hospitalization period (B = 0.562, p < 0.001) were factors positively associated with the frequency of pDDIs.

Conclusions

Potential drug–drug interactions were highly prevalent among hospitalized cardiac patients at Sultan Qaboos University Hospital, Muscat, Oman. Patients having diabetes as a comorbidity and with a high number of administered drugs were at a higher risk of an increased number of pDDIs.

Key Points

Of 215 patients in the study, 97.2% (n = 209) of them had at least one potentially harmful drug–drug interaction during their stay in the cardiology ward.

Anticoagulants and antiplatelet drugs were responsible for the vast majority of the observed potential drug–drug interactions among hospitalized cardiology patients.

The number of drugs taken and having diabetes mellitus were associated with an increased number of potential drug–drug interactions.

Introduction

Mortalities caused by cardiovascular diseases (CVDs) account for the largest proportion of disease-related mortalities worldwide. The number of prevalent cases of patients with CVDs nearly doubled from 271 million in 1990 to reach 523 million in 2019, and the number of CVD deaths steadily increased from 12.1 million in 1990, reaching 18.6 million in 2019 [1]. In the Middle East, CVD is by far the leading cause of death in all of these countries, making up 45% of all deaths. In Oman, the estimated age-standardized annual mortality rate from CVDs was reported to be around 400 per 100,000 persons [2]. In addition to the significant mortality rate, a huge economic burden is ascribable to CVDs. In the USA, the average annual direct and indirect cost of CVDs was estimated at $363.4 billion in the period from 2016 to 2017 [3]. In Europe, these diseases were estimated to cost the European Union €169 billion annually [4].

Elderly patients, who are the most affected by CVDs, tend to have a high prevalence of multimorbidity and polypharmacy, in addition to having age-related alterations in terms of drug availability and metabolism [5–8]. The previous four factors make elderly patients with CVDs highly susceptible to the occurrence of drug–drug interactions (DDIs) [7, 9–17].

A DDI is defined as a clinically meaningful alteration in the exposure and/or response to a drug (object drug) that has occurred as a result of the co-administration of another drug (precipitant drug). A DDI differs from a potential DDI (pDDI), which is defined as the co-prescription of two drugs known to interact, and therefore a DDI could occur in the exposed patient [18]. The mechanisms of DDIs can be broadly classified into either a pharmacokinetic or pharmacodynamic basis. Drug–drug interactions of a pharmacokinetic basis involve the alteration of the processes by which drugs are absorbed, distributed, metabolized, and excreted. Such interactions may cause a change in the drug concentration at the site of action with subsequent toxicity or diminished efficacy. Pharmacodynamic interactions involve the alteration of the pharmacological effect of one drug owing to the presence of another drug at its site of action. They could happen because of the competition for specific receptor sites but mostly they are indirect and involve interference with physiological systems. These interactions could have an additive, synergistic, or antagonistic effect [19].

Drug–drug interactions, among other drug-related problems, are associated with negative outcomes among both outpatients and inpatients [20, 21]. Drug interactions are an important cause of adverse drug reactions. Around 3–26% of all adverse drug reactions leading to hospital admission are caused by DDIs. Additionally, a DDI is associated with an increased length of hospital stay and other additional healthcare costs [22]. It was estimated that 17% of all preventable adverse drug events among hospitalized patients are due to the presence of DDIs and that around 1% of patients will experience an adverse drug reaction caused by DDIs during their hospitalization [23].

Overall, we noticed a high prevalence of pDDIs among hospitalized cardiology patients in different countries [7, 9–17, 24]. Information about pDDIs, drugs most frequently implicated, and factors associated with them could be used to aid healthcare workers to optimize patients’ pharmacotherapy by reducing the number of these interactions. No such studies have been conducted in Oman. Therefore, our main aim was to evaluate pDDIs and risk factors associated with them among hospitalized patients with CVDs at Sultan Qaboos University Hospital in Muscat, Oman.

Methods

This was a retrospective cross-sectional study for patients hospitalized in the Unit of Cardiology between January and December of 2021 at Sultan Qaboos University Hospital, Muscat, Oman. The main data source was electronic patient records “TrakCare,” the hospital information system. A standard data collection form was used to collect data about demographic characteristics (age and sex), causes of admission, comorbidities, medications taken, and length of hospital stay.

Study Population

The included patients were aged at least 18 years, received more than one drug during the stay, and hospitalized for at least 24 hours at the Unit of Cardiology. Patients who had incomplete data such as sex, age, reason for hospitalization, length of hospital stay, received drugs, and comorbidities were excluded from the study.

Study Measures

Patients’ electronic medical charts were screened using Micromedex Drug-Reax^® to identify pDDIs. This software was chosen as it is the highest in terms of completeness and consistency scores among other software used for pDDI detection [25]. It classifies pDDIs according to severity and documentation levels. The severity classification is as follows:

• Contraindicated: concurrent use of the interacting pair is contraindicated.

• Major: the interacting pair may result in permanent damage/death; medical intervention is needed to prevent or minimize the adverse outcome.

• Moderate: the combination may worsen the patient’s condition and/or require an alteration in therapy.

• Minor: there are limited clinical effects of the interaction. These may include an increase in the severity or frequency of adverse effects, and major alteration of therapy is not required.

Whereas classification according to documentation levels (scientific evidence) is as follows:

• Excellent: controlled studies have demonstrated the existence of an interaction.

• Good: well-controlled studies are lacking, but documentation strongly suggests the existence of an interaction.

• Fair: existing documentation is less, but physicians suspect the presence of interaction on the basis of pharmacological considerations, or evidence is good for interactions involving pharmacologically similar drugs [26].

The overall prevalence of pDDIs, defined as the presence of at least one pDDI in a patient’s regimen, and the prevalence based on severity and documentation levels were provided. Identified interactions were classified according to their mechanisms as well. We considered all drugs given to patients during their hospital stay, whether they were urgently given drugs (STAT drugs), normally given drugs, own drugs, and drugs given during coronary angiography, percutaneous coronary intervention (PCI), and coronary artery bypass graft surgery. Only interactions between drugs given on the same day were deemed pDDIs in the study. Clinical notes were also screened to make sure the drugs prescribed were actually administered to patients, and to include any drug given but missing from the prescription charts. For a better estimation of the prevalence and total number of pDDIs and as we were interested in the pDDIs with a potential harm to patients, combinations used according to local guidelines were excluded from the analysis. Interactions between aspirin and clopidogrel were not considered potentially harmful if the combination was given to patients admitted because of a previous diagnosis of acute coronary syndrome (with doses not exceeding 100 mg for aspirin and 75 mg for clopidogrel), or when it was given to patients after undergoing PCI or coronary artery bypass graft surgery (with doses not exceeding 100 mg for aspirin and 150 mg for clopidogrel given for 1–2 weeks). Heparin and nitroglycerin, and heparin and eptifibatide interactions were not considered potentially harmful if given during PCI.

Statistical Analysis

Descriptive statistics were used to describe the data. For categorical variables, frequencies and percentages were reported. For continuous variables, if they were normally distributed, mean and standard deviation (SD) were used to present the data. For non-normally distributed continuous variables, median and interquartile range (IQR) were used to summarize the data. The Kolmogorov–Smirnov test was used to assess the normality of continuous variables distribution.

Univariable and multivariable linear regression analysis was performed to determine the effect of predictor variables on the total number of pDDIs. Only variables found to have a linear relationship with the dependent variable and had a statistically significant association in the univariable linear regression analysis (at p-values of 0.05 or less) were included in the multivariable linear regression analysis. The multivariable regression analysis was performed using the stepwise method. Related assumptions were tested for each proposed regression model. The included predictors were number of drugs taken, length of hospital stay, age, number of comorbidities, sex, reason for hospitalization (myocardial infarction vs other causes), and having a certain comorbidity (diabetes mellitus, hypertension, dyslipidemia, and chronic kidney disease). The previous four comorbidities were chosen in particular because they were the most prevalent among the patients in the cohort. The other variables were chosen as predictors because they were used in previous similar studies conducted in several countries [7, 9–14, 16, 17].

Data were analyzed using SPSS 23.0 (IBM SPSS Inc. Chicago, IL, USA) and Stata software version 13.1 (STATA Corporation, College Station, TX, USA). P-values of 0.05 or less were considered statistically significant for all statistical tests.

The following formula was used for the sample size calculation: $n=Z^{2}P\left(1-P\right)/d^2$ [27], where n is the required sample size, Z is the Z-statistic for a level of confidence of 95% (Z = 1.96), P is the expected prevalence (P = 87.74%), which was derived from a previous study on a similar topic [17], and d is the absolute error or precision (d = 0.05). By including these values in the above-mentioned formula, the calculated sample size was 166. Eventually, a total of 215 patients were included in the study.

Results

The mean age of the patients was 60.1 ± 13.825 years. Male individuals represented 66.5% of the sample. The median length of hospital stay was 3 (2–4) days, while the median number of drugs taken during this period was 13 (11–17). Non-ST-elevation myocardial infarction was the main cause of hospitalization in the cardiology unit (28.37%), followed by ST-elevation myocardial infarction (26.52%). Unstable angina was the third most common cause (11.16%). Patients with stable angina and heart failure accounted for 9.77% and 8.37% of all the admitted patients, respectively. Finally, around 15.81% were hospitalized because of other causes. Patients with comorbidities made up 79.5% of the sample. The most common comorbidities were hypertension (57.7%), diabetes (46%), and dyslipidemia (30.2%). Of the 215 included patients, 93.5% of them had at least one moderate pDDI followed by a major pDDI occurring in 92.1% of patients. The characteristics of the included 215 patients are presented in Table 1.

Table 1.

Patients’ characteristics

Patients’ characteristics	N (%)/mean ± SD/median (IQR)
Sex
Male	143 (66.5)
Female	72 (33.5)
Age (years)	60.1 ± 13.825
Length of hospital stay (days)	3 (2–4)
Number of drugs taken	13 (11–17)
Reason for hospitalization
NSTEMI	61 (28.37)
STEMI	57 (26.52)
Unstable angina	24 (11.16)
Stable angina	21 (9.77)
Heart failure	18 (8.37)
Acute coronary thrombosis	4 (1.86)
Aortic valve stenosis	4 (1.86)
Bradycardia	4 (1.86)
Heart block	3 (1.4)
Tachycardia	3 (1.4)
Atrial fibrillation	2 (0.93)
Hypertension + pulmonary edema	2 (0.93)
Spontaneous coronary artery dissection	2 (0.93)
Left ventricular inflow obstruction	1 (0.47)
Atrial fibrillation and flutter	1 (0.47)
Atrial flutter	1 (0.47)
Heart failure + NSTEMI	1 (0.47)
Hypertension	1 (0.47)
Ischemic cardiomyopathy	1 (0.47)
Mitral valve insufficiency	1 (0.47)
Mixed mitral valve disease + pulmonary hypertension + left ventricular dysfunction	1 (0.47)
Pericardial effusion	1 (0.47)
Pulmonary hypertension	1 (0.47)
Comorbidity
Yes	171 (79.5)
No	44 (20.5)
Number of comorbidities	2 (3–1)
Types of comorbidities
Hypertension	124 (57.7)
Diabetes mellitus	99 (46)
Dyslipidemia	65 (30.2)
Chronic kidney disease	25 (11.6)
Gastrointestinal disease	16 (7.4)
Previous stroke	12 (5.6)
Stable angina	11 (5.1)
Hypothyroidism	11 (5.1)
Cardiac arrhythmias	10 (4.7)
Benign prostatic hyperplasia	9 (4.2)
Heart failure	8 (3.7)
Asthma	8 (3.7)
Valvular heart disease	7 (3.3)
Myocardial infarction	6 (2.8)
Obstructive sleep apnea	5 (2.3)
Psychiatric disorders	5 (2.3)
Cardiomyopathy	4 (1.9)
Malignancies	4 (1.9)
Gout	3 (1.4)
Anemia	3 (1.4)
Osteoporosis	3 (1.4)
Epilepsy	2 (0.9)
Psoriasis	2 (0.9)
Cervical spondylosis	2 (0.9)
Peripheral neuropathy	1 (0.5)
Abdominal aortic aneurysm	1 (0.5)
Cervical dystonia	1 (0.5)
Benign paroxysmal positional vertigo	1 (0.5)
Hyperthyroidism	1 (0.5)
Peroneal neuropathy	1 (0.5)
Hypopituitarism	1 (0.5)
Fibromuscular dysplasia	1 (0.5)
Limb lymphedema	1 (0.5)
Post-renal transplantation	1 (0.5)
Severity of pDDIs per patients
Contraindicated	3/215 (1.4)
Major	198/215 (92.1)
Moderate	201/215 (93.5)
Minor	38/215 (17.7)

IQR interquartile range, NSTEMI non-ST elevated myocardial infarction, pDDIs potential drug–drug interactions, SD standard deviation, STEMI ST-elevated myocardial infarction

Frequency and Classifications of Detected pDDIs

The total number of potential pDDIs observed was 2380, but after excluding the potentially beneficial interactions the number was 2057, with a median of nine (5–12) pDDIs per patient. Only six patients (2.8%) did not have any pDDI in their pharmacotherapy regimen, while 97.2% of the included patients had at least one DDI. Major interactions represented the majority of the pDDIs (52.6%), followed by moderate (45.1%) and minor (2.1%) pDDIs. Contraindicated pDDIs were present in the regimens of three patients only. In regard to the documentation level, a fair level was seen in 45.5% of all the noted pDDIs, whereas levels of good and excellent documentation were seen in 35.8% and 18.7% of all pDDIs, respectively. Interactions of a pharmacodynamic basis were the most common (55.9%). Data on the classifications of detected pDDIs are presented in Table 2.

Table 2.

Classifications of pDDIs

Classification	N (%)/median (IQR)
pDDIs	2057/9 (5–12)
Severity
Contraindicated	3 (0.1)
Major	1082 (52.6)
Moderate	928 (45.1)
Minor	44 (2.1)
Documentation
Excellent	385 (18.7)
Good	737 (35.8)
Fair	935 (45.5)
Mechanism
Pharmacokinetic	702 (34.1)
Pharmacodynamic	1150 (55.9)
Unknown	206 (10)

IQR interquartile range, pDDIs potential drug–drug interactions

The most common interacting pair observed was atorvastatin and clopidogrel (9%). The vast majority of pDDIs involved at least an antiplatelet drug (79.6%). The only contraindicated interactions detected were between verapamil and colchicine. As for the excluded interactions, of 217 pDDIs detected between aspirin and clopidogrel, 191 were excluded because of meeting previously defined criteria. The combination was used 163 times in patients with acute coronary syndrome, and 28 times in patients after undergoing PCI, with proper doses documented in both cases. Three cases were detected in which a heparin and nitroglycerin combination was inappropriately used, whereas the combination was used correctly in 119 cases (during PCI). In addition, we detected 12 pDDIs between heparin and eptifibatide in which they were used according to local guidelines (during PCI). Drugs involved in the 15 most frequent pDDIs and their classifications are given in Table 3. Table 4 shows the most common drug classes involved in the detected pDDIs.

Table 3.

Top 15 potential drug–drug interaction combinations

Combination	Severity	Documentation	Mechanism	N (%)
Atorvastatin + clopidogrel	Moderate	Excellent	PK	168 (9)
Clopidogrel + esomeprazole	Major	Excellent	PK	151 (7.3)
Aspirin + bisoprolol	Moderate	Good	PD	113 (5.5)
Clopidogrel + heparin	Major	Fair	PD	99 (4.8)
Aspirin + nitroglycerin	Moderate	Good	PK	93 (4.5)
Aspirin + heparin	Major	Fair	PD	85 (4.1)
Aspirin + insulin	Moderate	Fair	Unknown	82 (4)
Aspirin + lisinopril	Moderate	Good	PD	72 (3.5)
Aspirin + furosemide	Major	Good	PD	72 (3.5)
Bisoprolol + insulin	Moderate	Good	PD	65 (3.2)
Clopidogrel + verapamil	Major	Fair	PK	51 (2.5)
Aspirin + fondaparinux	Major	Fair	PD	45 (2.2)
Clopidogrel + fondaparinux	Major	Fair	PD	44 (2.1)
Furosemide + insulin	Moderate	Fair	PD	41 (2)
Insulin + lisinopril	Moderate	Fair	Unknown	35 (1.7)

PD pharmacodynamic, PK pharmacokinetic

Table 4.

Most common classes of drugs involved in the detected potential drug–drug interactions

Drug class	N (%)a
Antiplatelets	1637 (79.6)
Anticoagulants	409 (20)
Antidiabetic agents	360 (17.5)
Beta-blockers	296 (14.4)
Diuretics	248 (12)
Statins	225 (11)
Calcium-channel blockers	204 (9.9)
Angiotensin-converting-enzyme inhibitors and angiotensin II receptor antagonists	180 (8.8)
Opioids	172 (8.4)
Proton pump inhibitors	166 (8.1)
Nitrates	97 (4.7)
Benzodiazepines	68 (3.3)
Digitalis glycosides	24 (1.2)

^aThe total number and proportion of participations of a drug belonging to a particular class in one of the 2057 detected potential drug–drug interactions

Factors Associated with the Total Number of Detected pDDIs

A linear relationship existed between the number of drugs taken and the total number of pDDIs. This was not the case for age, length of hospital stay, and number of comorbidities. The univariable analysis showed that the total number of pDDIs was significantly associated with the number of drugs taken (B = 0.768, p < 0.001), hospitalization because of a myocardial infarction (B = 2.132, p = 0.017), and having diabetes as a comorbidity (B = 4.949, p < 0.001). Thus, the three previous variables were used as predictors in the first multivariable regression model. However, the model was not accepted because the constant was not statistically significant (p = 0.118). A new model was proposed, containing the same previous variables without the constant. The reason for hospitalization was excluded from the new model (p = 0.071). The model met all the assumptions except for the presence of homoscedasticity. To fix the heteroscedasticity issue, the Huber–White standard errors (robust standard errors) were used [28]. By following this approach, all the parameters produced were the same as the previous model except for standard errors, t values, confidence intervals, and coefficient p-values. Coefficients p-values were still statistically significant.

The multivariable analysis indicated that two factors, the number of drugs taken (B = 0.562, p < 0.001) and having diabetes as a comorbidity (B = 2.564, p < 0.001) were associated with an increased number of pDDIs (R² = 0.798, F = 202.05, p < 0.001). The final model results are shown in Table 5.

Table 5.

Univariable and multivariable linear regression analysis for the total number of potential drug–drug interactions

Variables	Univariable				Multivariable
	B	95% CI	P-value	β	B	95% CI	P-value	β
Number of drugs taken	0.768	0.621 to 0.916	< 0.001	0.576	0.562	0.468 to 0.655	< 0.001	0.421
Sex	1.424	− 0.431 to 3.279	0.132	0.103	–	–	–	–
Reason for hospitalization	2.132	0.387 to 3.877	0.017	0.163	1.313	− 0.128 to 2.754	0.074	0.1
Diabetes mellitus	4.949	3.315 to 6.583	< 0.001	0.379	2.564	1.160 to 4.032	< 0.001	0.199
Hypertension	0.611	− 1.168 to 2.390	0.499	0.046	–	–	–	–
Dyslipidemia	− 0.070	− 1.986 to 1.846	0.942	− 0.005	–	–	–	–
Chronic kidney disease	− 0.149	− 2.895 to 2.596	0.915	− 0.007	–	–	–	–

CI confidence interval

The total number of observed DDIs was the dependent variable in the model. The following variables were included in the model as dichotomous predictor variables: sex (1 = male, 0 = female), reason for hospitalization (1 = myocardial infarction, 0 = other causes), hypertension (1 = having hypertension, 0 = not having hypertension), diabetes (1 = having diabetes, 0 = not having diabetes), dyslipidemia (1 = having dyslipidemia, 0 = not having dyslipidemia), and chronic kidney disease (1 = having chronic kidney disease, 0 = not having chronic kidney disease). The number of drugs taken was the only continuous predictor variable included in the model. P ≤ 0.05 was considered statistically significant.

Discussion

This study evaluated the prevalence of pDDIs and the factors that contributed to an increased number of interactions among patients hospitalized in the cardiology wards at a university hospital in Oman. To the best of our knowledge, this is the first study of its type conducted in Oman.

The study revealed a high prevalence of pDDIs (97.2%), which is higher than what was reported in similar studies conducted in other countries, and slightly lower than the reporting of one study [7]. The median number of pDDIs per patient was higher than what was reported in studies conducted elsewhere [7, 9, 11–17]. The majority of the detected pDDIs were of major severity (52.6%). Minor severity pDDIs represented most of the detected interactions (51.66%) in only one previous study done in Morocco [15]. In studies conducted elsewhere, the majority of the observed pDDIs were of moderate severity [7, 9, 11, 12, 14, 16, 17, 24]. Potential DDIs of a pharmacodynamic nature accounted for the majority of interactions, at 55.9%, which was in line with other studies conducted in India [9, 13], UAE [17], and Serbia [14]. However, studies conducted in Nepal and Morocco revealed a higher proportion of pDDIs of a pharmacokinetic nature [11, 15].

The atorvastatin and clopidogrel combination was the most common combination implicated in observed pDDIs in our study. In studies conducted in India [13], Morocco [15], Pakistan [12], and China [24] the most common combination was aspirin and clopidogrel. However, the previous studies did not mention any exclusion of combinations in the case of proper use according to guidelines, and in the case of our study, the most common combination resulting in a pDDI would have been also aspirin and clopidogrel had we not excluded the cases in which this combination was beneficial for patients and was used according to local guidelines. This combination was the second most common in a study done in UAE [17], whereas aspirin and bisoprolol was the most common combination in the same study. Heparin and aspirin, another highly observed interaction, was the most common combination in another Indian study [9]. Another study conducted in Pakistan reported that aspirin and enoxaparin was the most observed combination in the pDDI list [7]. Antiplatelets and anticoagulants were the drug classes most frequently involved in the observed pDDIs in our study. This was the case in studies conducted in India [9, 13], Pakistan [7, 12], UAE [17], and China [24]. The discrepancies in the previous results were probably because of the usage of different screening software and different preferred therapeutic regimens.

Our study found that some factors were associated with the total number of pDDIs. Those factors were the number of drugs taken during the hospitalization period (B = 0.562, p < 0.001) and having diabetes as a comorbidity (B = 2.564, p < 0.001). Polypharmacy is consistently associated with pDDIs [23]. In our study, antidiabetic agents (mainly insulin) were among the most frequently involved drugs in the detected pDDIs. Statin therapy has been recommended by most clinical guidelines in patients with diabetes [29], and statins (mainly atorvastatin) were commonly implicated in the detected pDDIs in our study. The previous arguments could explain why diabetes was associated with an increase in the pDDIs among the study cohort.

A study conducted in Pakistan revealed that hospitalization because of a myocardial infarction is a significant predictor of pDDIs [7], and in our study, the same factor was not significantly associated with the increased number of pDDIs after adjusting for other variables. A Serbian study found that patients with arrhythmia and heart failure as the main diagnosis were also associated with a higher prevalence of pDDIs [14]. When it comes to the influence of comorbidities on the pDDI prevalence, the results were conflicting. Similar to our findings, the number of comorbidities did not affect the prevalence of pDDIs as reported in two previous studies [7, 16]. Three other studies found that the presence of comorbidities was a significant predictor for the prevalence of pDDIs [11, 14, 17]. Having diabetes as a comorbidity in particular was found to increase the total number of pDDIs for patients in our study. Authors of the Serbian study revealed a positive association between having an infectious or respiratory diseases and the presence of pDDIs [14]. Age did not have an effect on the presence of pDDIs as reported in three previous studies [7, 16, 17], which was consistent with our findings. The same factor did have an effect per three other studies [12, 13, 24]. There was no association between pDDIs and sex in our study. This was supported by similar findings in studies conducted in Egypt [16], and Pakistan [7, 12]. In regard to the length of hospital study, there were inconsistencies in the results of previous pertinent research. Some studies showed a significant association between this factor and the prevalence of pDDIs [9, 11, 12, 14], while others did not [7, 16, 17, 24], which was in line with our results. Methodological differences could explain these discrepancies regarding the impact of comorbidities, length of hospital stay, and age. Among studies demonstrating the presence of comorbidities as a significant predictor [11, 14, 17], only one study had adjusted for other variables [17]. The same scenario also applies to the previous study carried out in Pakistan when it comes to the effect of length of hospital stay [12]. Among the four studies that controlled for other variables while examining the effect of age [7, 12, 16, 17, 24], only two had demonstrated a positive association [12, 24]. The number of drugs taken during hospitalization was a risk factor for the prevalence of pDDIs in our study and in previous related research [7, 9, 11–14, 16, 17, 24].

Dealing with pDDIs could happen while a certain drug is added to a patient’s regimen. Internationally, decision support alerts designed to warn prescribers of pDDIs when a new drug is ordered are incorporated in electronic prescribing systems in increasingly more hospitals. However, there is limited evidence of the effectiveness of such electronic alerts and they may also be counterproductive as they could result in a high volume of alerts, leading to user frustration and constant interruptions to workflow. This is because many of these alerts may be of little or no clinical relevance, and many of them lack specific instructions or alternative therapy suggestions. In order to improve the design and content of pDDI alerts, every hospital should design its own pDDI alert system as a ‘one-fit-for-all’ solution might not be applicable because of the diversity that exists in different clinical and organizational settings. Most importantly, this requires determining the proportion of pDDIs that result in actual harm to patients (clinically relevant DDIs) [30, 31].

Alternatively, pDDIs could be dealt with after the administration of the interacted drugs. Some management strategies were suggested to deal with the top 15 pDDIs (Table 6). Of the 15 most commonly observed combinations involved in pDDIs, 12 included an anticoagulant, an antiplatelet drug, or both. Monitoring for the signs and symptoms of bleeding, coagulation parameters, and platelet reactivity could be enough to deal with these types of pDDIs. Insulin was a commonly observed drug implicated in many pDDIs, including four of the most frequently detected interactions. Monitoring for the signs and symptoms of hypoglycemia and altering the insulin dose would be viable strategies for the management of these interactions. Verapamil and colchicine interactions were the only contraindicated interactions observed in the study. Verapamil could increase the serum concentrations of colchicine and increase the risk of its toxicity. This combination should be avoided in general, but could be continued in patients with normal hepatic and renal functions. In these cases, the dose for gout flare should not exceed 1.2 mg (two tablets for one dose) with the repeat dose to be given no earlier than 3 days, and for gout prophylaxis, the dose should not exceed 0.6 mg daily [26].

Table 6.

Potential adverse outcomes for the 15 most frequent potential drug–drug interactions and their management strategies [19, 26, 33–35]

Combination	Potential adverse outcome	Management strategies
Clopidogrel + atorvastatin	Decreased formation of clopidogrel active metabolite and high on-treatment platelet reactivity	Monitoring of platelet reactivity is necessary. If a patient develops high on-treatment platelet reactivity during treatment, atorvastatin should be discontinued and substituted with a statin that is not metabolized by CYP3A4 (i.e., pravastatin or rosuvastatin)
Clopidogrel + esomeprazole	Reduction in clopidogrel active metabolite serum concentrations and reduced anti-platelet activity	It is recommended to use an alternative acid-lowering drug with less CYP2C19 inhibitory effect such as rabeprazole,, and possibly also pantoprazole or lansoprazole, a histaminergic (H2) blocker (except cimetidine), or an antacid
Aspirin + bisoprolol	Increased blood pressure	Patients’ blood pressure and hemodynamic parameters should be monitored
Clopidogrel + heparin	Increased risk of bleeding	The concurrent use of heparin with clopidogrel is indicated in specific conditions such as acute coronary syndrome. However, unless specifically indicated, this combination should probably be avoided. If they are used together, careful clinical and laboratory monitoring is recommended
Aspirin + nitroglycerin	Increased nitroglycerin concentrations and additive platelet function depression	This combination could be beneficial to patients with acute myocardial infarction. Other patients should be monitored for hypotension, headache, syncope, and for signs of bleeding. If any of these effects occurs, then it is recommended to consider discontinuing aspirin or tailor the dose of nitroglycerin
Aspirin + heparin	Increased risk of bleeding	This combination could be beneficial in some situations such as prophylaxis of ischemic complications of unstable angina. In other patients, it is better to avoid this combination. If they are used together, it is suggested to monitor coagulation parameters (PPT) and signs of bleeding
Aspirin + insulin	Increased risk of hypoglycemia	Monitoring the patient’s blood glucose and clinical signs of hypoglycemia is suggested. Adjusting the dose of insulin could be necessary
Aspirin + lisinopril	Decreased antihypertensive effect of lisinopril	Monitoring patients’ blood pressure, hemodynamic parameters, and renal function is recommended. If an adverse effect is evident, then reducing the dose of aspirin to less than 100 mg per day or using another anti-platelet drug is advised. Substituting lisinopril for an angiotensin receptor blocker is an additional option
Aspirin + furosemide	Decreased furosemide effectiveness and increased risk of nephrotoxicity	Monitoring for signs of renal toxicity and salicylate toxicity is recommended. Diuretic effectiveness should be assured including its effects on blood pressure. High doses of aspirin should be avoided
Bisoprolol + insulin	Altered effect of insulin on blood glucose lowering and decreased or obscured signs and symptoms of hypoglycemia	Increasing the frequency of blood glucose monitoring or a dose adjustment of insulin may be required. Close monitoring of signs and symptoms of hypoglycemia is also warranted
Clopidogrel + verapamil	Decreased anti-platelet effect of clopidogrel	Monitoring patients for loss of clopidogrel efficacy is advised (platelet reactivity test). The addition of cilostazol is an available option
Aspirin + fondaparinux	Increased risk of bleeding	Monitoring patients for signs and symptoms of bleeding is necessary
Clopidogrel + fondaparinux	Increased risk of bleeding	Monitoring patients for signs and symptoms of bleeding is necessary
Furosemide + insulin	Decreased efficacy of insulin due to furosemide inducing a rise in blood glucose levels	Monitoring of blood glucose levels is advised
Insulin + Lisinopril	Increased risk of hypoglycemia	Monitoring for clinical signs and symptoms of hypoglycemia is suggested

CYP cytochrome P450, PPT partial thromboplastin time

There are multiple limitations in our study. First, this was a retrospective study, thus certain data regarding some drugs prescribed and administered could be missing. Generalizability of the results could be limited as this was a unicentric study. There could be an overestimation of the prevalence of pDDIs in patients’ regimens as screening software used for their detection deal with drugs as classes. Another cause of this possible overestimation is the strict inclusion criteria applied and the exclusion of a considerable number of medical charts. Furthermore, not all of the detected pDDIs could be of clinical significance, even the major severity class. Despite the fact that interactions between drugs and supplements were investigated, interactions between drugs and dietary products taken during the hospital stay (such as green tea and coffee) were not accounted for. Such products could potentially interact with many drugs given to hospitalized cardiology patients [32]. Finally, this study did not provide data about the adverse events plausibly caused by the observed pDDIs. To address some of the previous issues, prospective multicenter studies should be conducted to improve the results’ generalizability, and to account for drugs taken and dietary products possibly missing from patients’ medical records. Studies to determine which of these pDDIs resulted in actual harm to patients are also warranted.

Conclusions

Our study revealed a very high prevalence of pDDIs among hospitalized patients in the cardiology wards in Oman. The number of drugs taken during a hospital stay and having diabetes as a comorbidity were factors associated with an increased number of pDDIs. Antiplatelets and anticoagulants were the culprit classes of drugs in the majority of the detected pDDIs. These results will enable healthcare workers to reduce the prevalence of these interactions, which will help optimize pharmacotherapy for cardiology patients in inpatient settings.

Declarations

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Conflicts of Interests/Competing Interests

Abdulrahman Kalash, Aly Abdelrahman, Ibrahim Al-Zakwani, and Yousuf Al Suleimani have no conflicts of interest that are directly relevant to the content of this article.

Ethics Approval

Ethics approval was obtained from the Medical Research Ethics Committee at Sultan Qaboos University before conducting the study (ref. no. SQU-EC/ 091/2022. MREC #2739). The study was also performed in accordance with the ethical standards as laid down in the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards.

Consent to Participate

As personal identification information was masked prior to the analysis, informed consent was not sought.

Consent for Publication

Not applicable.

Availability of Data and Material

The authors confirm that the data supporting the findings of this study are available upon request.

Code Availability

Not applicable.

Authors’ Contributions

Conceptualization: all; data curation: IA-Z, AA, YAS; formal analysis: AK, IA-Z; investigation: AK; methodology: all; project administration: all; supervision: YAS, AA; validation: IA-Z, AA, YAS; writing, original draft: AK; writing, review and editing: YAS, AA, IA-Z. All authors read and approved the final version.