Drug-drug interactions in patients with malaria: a multicenter retrospective study

Background: Hospitalized patients with malaria often present with comorbidities or associated complications for which a variety of drugs are prescribed. Multiple drug therapy often leads to drug-drug interactions (DDIs). Therefore, we investigated the prevalence, levels, risk factors, clinical relevance, and monitoring parameters/management guidelines of potential DDIs (pDDIs) among inpatients with malaria. Methods: A retrospective cohort study was carried out at multiple hospital settings. A total of 398 patients’ profiles were evaluated for pDDIs using the Micromedex Drug-Reax ®. Odds ratios were calculated to identify the strength of association between presence of DDIs and potential risk factors via logistic regression analysis. Further, the clinical relevance of frequent pDDIs was investigated. Results: Of 398 patients, pDDIs were observed in 37.2% patients, while major-pDDIs in 19.3% patients. Total 325 interaction were found, of which 45.5% were of major- and 34.5% moderate-severity. Patients with the most common pDDIs were found with signs/symptoms and abnormalities in laboratory findings representing nephrotoxicity, hepatotoxicity, QT interval prolongation, and reduced therapeutic efficacy . The adverse events were more common in patients prescribed with the higher doses of interacting drugs. Multivariate regression analysis showed statistically significant association of pDDIs with 5-6 prescribed medicines (p=0.01), >6 prescribed medicines (p<0.001), >5 days of hospital stay (p=0.03), and diabetes mellitus (p=0.04). Conclusions: PDDIs are commonly observed in patients with malaria. Healthcare professional’s knowledge about the most common pDDIs could help in preventing pDDIs and their associated negative effects. Pertinent clinical parameters, such as laboratory findings and signs/symptoms need to be checked, particularly in patients with

polypharmacy, longer hospital stay, and diabetes mellitus.

Background
Malaria is one of the infectious diseases that cause burden on the healthcare system. According to WHO, malaria accounts for 216 million cases in 91 countries in the year 2016. This was an increase of five million cases over 2015 [1]. Moreover, in 2016, an approximately 85% of vivax malaria cases were identified in five countries including Pakistan [2]. Worldwide, malaria remains one of the causes of death due to infectious diseases [3].
Hospitalization in malaria occurs due to disease severity, managing the associated symptoms or comorbid illnesses [4]. Anti-malarial drugs, anti-pyretic, and analgesics are usually prescribed to treat hospitalized malaria patients [5]. Besides these medicines, a variety of other medicines are also prescribed so as to manage the comorbid illnesses and associated symptoms [4][5][6]. Concomitant use of several drugs increased the chance of drug-drug interactions (DDIs)-affecting drug's pharmacokinetic parameters and pharmacodynamics profile [7,8]. DDIs may lead to a variety of negative clinical outcomes such as hospitalization, reduced or abolished therapeutic efficacy, prolongation of hospital stay, toxicity, and adverse effects [7][8][9]. An approximately, 20-30% of adverse effects have been reported as due to DDIs, of which 1-2% are life-threatening and 70% need clinical intervention [10]. Hence, particular consideration of DDIs and their timely management is crucial for the rational use of medicines in patients with malaria.
Potential DDIs (pDDIs) issue has been addressed generally in hospitalized patients [7] as well as in specific diseases such as liver cirrhosis [11], hypertension [12], diabetes mellitus (DM) [13], bone marrow transplant [14], cancer [15], stroke [16], pneumonia [17], urinary tract infections [18], and hepatitis C [19]. Despite, being the most prevalent causes of hospitalization [20], DDIs particularly among inpatients with malaria remains unaddressed. Moreover, in developing countries, literature has been least reported as well as irrational use of medicines is a common issue. Consequently, specific consideration is required to conduct studies evaluating pDDIs and their clinical relevance among hospitalized patients with malaria. Afterward, such studies will improve patients' safety and help healthcare professionals to manage pDDIs and reduce their associated negative clinical consequences.
This study aimed to evaluate the prescriptions of inpatients with malaria for pDDIs prevalence, and their levels. Investigate the risk factors contributing towards pDDIs prevalence, and clinical relevance of pDDIs. Secondary aim was to identify monitoring parameters and develop management guidelines for the most frequent pDDIs. Both male or female patients.

Study settings and design
All medications, that were prescribed during hospitalization of the patient were included in analysis.
Patients' profiles lacking relevant data required for the study were excluded.

Data source
A total of 398 patients were included for the study based on above criteria. The following date were collected from the patients' profiles such as hospital admissions, patients' demographics, diagnoses, comorbidities/complications, medications therapy, sign/symptoms, and laboratory test reports.

Medications profiles screening for pDDIs
Medicines prescribed to patients were evaluated for pDDIs using Micromedex Drug-Reax® [21]. This software classify drug interactions on the basis of severity-(contraindicated, major, moderate, and minor) and documentation-levels (excellent, good, and fair) [21]: Overall-prevalence of pDDIs as well as prevalence of pDDIs based on severity-levels were reported.
Widespread (most common) and clinically important pDDIs were enlisted along with their potential adverse consequences.

Statistical analysis
Data were presented in the form of frequencies and percentages alone or with median and interquartile range (IQR), where appropriate. A statistical method of logistic regression analysis was used to calculate odds ratios (OR) for various risk factors of pDDIs such as patients' gender, age, number of prescribed medicines, hospital stay, and comorbidities.
Dependent variable in the model was exposure to pDDIs. While, patients' characteristics (gender, age, number of prescribed medicines, hospital stay, and comorbidities) were taken as independent variables in the model. Odds ratios and 95% confidence intervals (CIs) were calculated for each independent variable. Univariate logistic regression analysis was run initially. Then, multivariate analyses were performed for variables with p-values of ≤ 0.15. A p-value of ≤ 0.05 was considered as statistically significant. SPSS-v23 was used for statistical analyses of the data.

General characteristics of study patients
Patients' demographics are presented in Table 1. Of 398 patients, males were more prevalent (51.8%). Most of the patients were aged 21-40 years (44.2%). Majority were prescribed with > 6 drugs (54.8%). Most frequent hospital stay was ≥ 4 days (64.6%). The median (IQR) age, prescribed drugs and hospital stay was 30 years (22-50), 7 drugs (5-9), and 4 days (3-6), respectively. Hypertension (n = 52), DM (45), urinary tract infections (34), hepatitis (23), and ischemic heart diseases (IHD) (15) were the most prevalent comorbidities of the studied patients. Moreover, exposure to pDDIs stratified against the patient's characteristics are also shown in Table 1. In males, pDDIs were more prevalent as compared to females. Similarly, pDDIs were commonly reported in patients aged > 40 years, prescribed with > 6 medicines, and hospitalization of > 5 days. Moreover, pDDIs were mostly reported in patients with DM, hypertension, urinary tract infections, and hepatitis as comorbidities. Prevalence of potential drug-drug interactions 8 Out of total 398 patients, 148 (37.2%) met at least one pDDI. Based on severity-wise prevalence, 19.3% patients were identified with at least one major-pDDI while, 15.8% with at least one moderate-pDDI. However, a smaller number of patients were found with contraindicated-and minor-pDDIs ( Figure 1).
Levels of potential drug-drug interactions Figure 2 illustrates categorization of pDDIs based on severity-and documentation-levels.
Total number of interactions were 325, among which 45.5% were of major-and 34.5% moderate-severity. Based on documentation-levels, 49.5% were of fair and 44.9% good scientific-evidence.
Risk factors of potential drug-drug interactions Table 2 shows logistic regression analysis based on exposure to pDDIs. In the univariate logistic regression analysis, association for pDDIs was statistically significant with 5-6 prescribed medicines (p = 0.005), > 6 prescribed medicines (p < 0.001), hospital stay of 4-5 days (p = 0.003), and > 5 days hospitalization (p < 0.001). Moreover, concerning comorbidities, association of pDDIs with DM (p = 0.001) and IHD (p = 0.07) was statistically significant.  Clinical relevance of potential drug-drug interactions Table 3 presents daily prescribed dosage of the ten most frequent interacting drug pairs.
In this study, the term high and low doses were used relatively. It was observed that the drugs were prescribed in varying doses and administration frequencies. Interacting drugs were prescribed more frequently in low doses, whereas, higher doses of the drugs were prescribed less frequently. Most frequent pDDIs along with their potential adverse consequences and levels are presented in Additional Table 1, while Additional Table 2 and Table 3 enlists most prevalent antimicrobial agents (AMAs) and drugs besides AMAs, respectively.   (2) Isoniazid -Acetaminophen (9) High + High (5)  -BUN, blood urea nitrogen; ALT, alanine aminotransferase; ALP, alkaline phosphatase; LFTs, liver function tests; FBS, fasting blood sugar.
a Frequencies are given in parenthesis and calculated among patients with respective interaction.

Discussion
DDIs remains one of the therapeutic challenges among inpatients [7]. Studies addressing pDDIs issues among hospitalized patients with malaria are lacking. The prevalence of pDDIs reported in the current research is higher (37.2%) in comparison to that among patients with acquired immune deficiency (33.5%) [23], liver cirrhosis (21.5%) [11], and hypertension (21.1%) [12]. Contrary, it is lower (37.2%) as compared to that among patients with hypertension (48%) [24], DM (52.2%) [13], and bone marrow transplant (60%) [14]. Furthermore in current study, prevalence of major-pDDIs is higher (19.3%) as compared to that reported among patients with cancer (16%) [15]. Whereas, it is lower in comparison to that reported among patients with liver cirrhosis (21.4%) [11], hepatitis C (30-44%) [19], and stroke (61%) [16]. Similarly, the prevalence of contraindicated-pDDIs in patients with malaria is also lower (14.3%) in comparison to the prevalence reported among patients with hepatitis C (16.7%) [25]. This contradiction may be due to variable study population, drug prescribing patterns, study design, considering pDDIs types, and drug interaction screening software. Considering the findings of this study, malaria patients are more at risk to pDDIs. Published literature has proposed some evidence based approaches to minimize, prevent or manage DDIs in hospital settings such as screening medication profiles for pDDIs by using computerized screening programs [26], engaging clinical pharmacists in assesing patients' medication profiles for pDDIs [27][28][29], procedure for structured assessment of pDDIs [30], and checking pertinent laboratory findings for clinical relevance of interactions [7,31].
Healthcare professionals can manage adverse outcomes related to interactions, by taking into considerations the levels of interactions. In our study, pDDIs of major and moderate types were commonly observed, while concerning documentation levels, pDDIs of fair and good types were more prevalent. These findings are inconsistent with the findings from other studies [11,20,32]. This situation is alarming as our results warrant about the exposure of malaria patients towards negative consequences of pDDIs. Therefore, identifying the type of interaction, by healthcare professional is crucial for managing pDDIs, minimizing the related risk, and designing prophylactic measures for prevention.
Hospitalized patients with malaria receive a variety of medications for the management of underlying disease, related complications, and/or comorbid illnesses [4][5][6]. Our findings support that provision of multiple therapy has been positively associated with pDDIs prevalence [15,[32][33][34]. Moreover, the statistically significant association of pDDIs with prolong hospitalization reported by our study is in accordance with the published reports [20,35]. Furthermore, we observed a significant association of pDDIs with DM as comorbidity of malaria. The reason is that, in patients with DM, such drugs are prescribed, having higher risk of DDIs [36]. In this regard, hospitalized malaria patients having any of the above-mentioned risk factors are at higher risk to pDDIs. Healthcare professionals should have knowledge regarding the factors contributing towards pDDIs prevalence. This will help in reducing the risk of pDDIs-patients more at risk to pDDIs should be individualized to improve drug therapy and reduce the adverse outcomes of pDDIs.
All types of pDDIs are not clinically significant. Hence, developing the list of clinically

Consent for publication
Not applicable

Availability of data and material
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.    -The total recorded pDDIs 325 were classified based on severity-and documentation-levels.

Supplementary Files
This is a list of supplementary files associated with the primary manuscript. Click to download.