Prevalence and significance of potential drug-drug interactions among cancer patients receiving chemotherapy

Background Cancer patients often receive multiple drugs to maximize their therapeutic benefit, treat co-morbidities and counter the adverse effects of chemotherapy. Concomitant administration of multiple drugs increases the risk of drug interactions leading to compromised therapeutic efficacy or safety of therapy. The purpose of this study was to identify the prevalence, levels and predictors of potential drug-drug interactions (pDDIs) among cancer patients. Methods Six hundred and 78 patients receiving chemotherapy from two tertiary care hospitals were included in this cross-sectional study. Patient medication profiles were screened for pDDIs using the Micromedex® database. Logistic regression analysis was performed to identify the predictors of pDDIs. Results The overall prevalence of pDDIs was 78%, majority of patients had 1–2 pDDIs (39.2%). A total of 1843 pDDIs were detected. Major-pDDIs were most frequent (67.3%) whereas, a significant association of pDDIs was found between > 7 all prescribed drugs (p < 0.001) and ≥ 3 anti-cancer drugs (p < 0.001). Potential adverse outcomes of these interactions include reduced therapeutic effectiveness, QT interval prolongation, tendon rupture, bone marrow suppression and neurotoxicity. Conclusions Major finding of this study is the high prevalence of pDDIs signifying the need of strict patient monitoring for pDDIs among cancer patients. Patients at higher risk to pDDIs include those prescribed with > 7 any types of drugs or ≥ 3 anticancer drugs. Moreover, list of most frequently identified major and moderate interactions will aid health care professional in timely identification and prevention of pDDIs.


Background
The global cancer burden is on the rise due to increased prevalence of risk factors such as smoking, environmental pollution, obesity, unhealthy diet, physical inactivity, infections (hepatitis, helicobacter pylori and human papilloma virus) and use of oral contraceptives [1][2][3][4]. The Global burden of cancer (GLOBOCAN) 2012, reported an estimated 14.1 million new cases of cancer, which is anticipated to rise by 70% over the next 20 years [2,3].
Cancer is equally prevalent around the world [3,5], however differences in the pattern of cancer and its subsequent therapy exist from region to region [5]. The overall survival of cancer is high in developed countries due to timely and easy access to standard health care facilities [1,2,6,7].
The use of cytotoxic agents is inevitable for the cure of cancer despite the availability of a number of other alternatives such as surgery and radiation [8,9]. The main aim of chemotherapy is to cure cancer, extend life years and improve the overall quality of life [7,8]. Cytotoxic agents are often administered in combination containing two or more cytotoxic drugs as multiple drug regimens along with other medicines to achieve maximum therapeutic benefit and counter the adverse effects of chemotherapy or to treat other co-illnesses [10,11]. However, interpatient variability is frequently observed especially with oral cytotoxic agents due to drug interactions as a result of polypharmacy [12,13].
Drug-drug interactions (DDIs) are drug combinations that may result in therapeutic failure or potentially serious adverse events than from solitary administration [14][15][16][17][18]. An estimated 2% of hospital admissions are due to adverse effects caused by DDIs [19,20] and approximately 4% of the cancer patients die because of adverse effects caused by drug interactions [21] however, DDIs are often predictable and preventable [22,23]. Advance age, prolonged hospital stay and increased number of prescribed medicines are strong predictors of DDIs [24][25][26][27]. The risk of DDIs and its associated adverse events are higher in cancer patients as they frequently receive multiple drugs concomitantly [17,25]. Moreover, there has been a recent increase in the availability and use of anticancer agents because of favorable therapeutic outcome and cost effectiveness [12,28].
DDIs among cancer patients have been well studied in developed countries [20,24,[26][27][28][29]. However, such studies have been poorly addressed in developing countries like Pakistan. There are a few studies which have their own scope and limitations such as involving only QT prolonging DDIs [30], small sample size and different drug interactions screening tool [31], and primarily focused on medication safety [32]. Further, we cannot generalize the findings of developed countries to the developing countries because of variations in drug prescribing pattern, utilization of anticancer drugs, pattern of cancers, drug interactions screening before prescribing, and non-availability of standard healthcare services.
DDIs involving anticancer drugs are a major concern in oncology practice due to their potential to cause severe adverse effects [1,28,33,34]. Moreover, knowledge about the most common interacting drugs used in cancer patients and identification of predictors of pDDIs are essential to reduce avoidable drug-related problems and increase the efficacy and compliance of chemotherapy [14,17]. Additionally, this study will be helpful for the promotion of rational drug use and for the prevention and management of drug interactions leading to improved therapeutic outcome and patients' quality of life [34,35]. Therefore, the study aimed to identify the prevalence, levels, predictors and potential adverse outcome of pDDIs among cancer patients receiving chemotherapy.

Study settings and design
A cross-sectional study was performed in two tertiary care hospitals of Peshawar, Khyber Pakhtunkhwa,

Selection criteria
Patients of any age and either gender diagnosed with any type of cancer and treated with anticancer agents (either intravenous and/ or oral) were included in this study. Whereas, patient profiles were excluded if the required information were lacking.

Data collection and screening for pDDIs
Data regarding patient's demographic, symptoms, laboratory results and prescribed medications were collected after acquiring written permission from the administration of respective hospitals.
Micromedex Drug-Reax® (Truven Health Analytics, Greenwood Village, Colorado, USA) was used for the screening of patients' medication profile for pDDIs [36]. We select this software because it has got highest sensitivity and specificity score [37,38]. Further, it has got sensitivity score of 70% in identifying drug interactions involving oral anticancer drugs [39]. According to the description of this database, all detected interactions were categorized on the basis of severity-levels and documentation-levels [36]. All available lab values were reviewed to identify abnormal results.

Statistical analysis
In statistical analysis, quantitative data were presented as frequencies and percentages. Logistic regression analysis was applied in order to identify association of pDDIs presence with patients' gender, age, prescribed medications, hospitalization status, cancer type, presence of metastasis, treatment type and treatment intent. For each predictor odds ratio (OR) a 95% confidence intervals (CIs) was determined by performing univariate logistic regression analysis. For variables with significant univariate p-values multivariate analysis was performed. In this study p-value of 0.05 or less was considered significant. SPSS version-23 was used for statistical analysis.

Levels of pDDIs
Overall, 1843 pDDIs were detected, of which, 1240 (67.3%) were of major and 507 (27.5%) were of moderate severity while contraindicated pDDIs were least frequent accounting for 25 (1.4%) pDDIs. The documentary evidence of majority of pDDIs were fair (66.4%) and good (23.8%) (Fig. 2).    -Percentages do not add up to 78% because many patients were exposed to multiple pDDIs of different severities

Wide spread interacting drug combinations
Most frequently detected pDDIs are enlisted in Table 5 along with their severity, documentation levels and potential adverse outcomes. Reduced therapeutic effectiveness, QT interval prolongation, drug toxicity such as tendon rupture, bone marrow suppression, seizures, serotonin syndrome, neurotoxicity and cardiomyopathy were the potential adverse outcomes of these interactions. Potential drug-drug interactions involving anticancer agents are enlisted in additional Table 1.

Discussion
This study presents the frequency, severity and predictors for pDDIs and list of most frequent pDDIs among cancer patients undergoing chemotherapy. An overall 78% prevalence of pDDIs is higher in comparison with other studies conducted in oncology setting. A study from Iran reported 62.8% prevalence of pDDIs in patients with hematological malignancy [40]. Another study from Netherland reported a prevalence rate of 46% among patients using oral anticancer drugs [41]. Whereas, a study from the United States of America reported 40% prevalence rate of pDDIs [29]. Similarly, the prevalence rate of present study is higher in comparison with studies from other specialties such as internal medicine (52.8%) [25], psychiatry (64.8%) [42], pediatrics (25.8%) [43] and pulmonology (45%) [44]. Such widespread variability in prevalence may be attributed to differences in study designs, inclusion and exclusion criteria, study population & their characteristics, study settings, presence or absence of clinical pharmacy services, prescribing pattern, drugs involving, and high sensitivity of drug interactions screening databases/sources.  The high prevalence of pDDIs identified in our study demands thoughtfulness regarding the issue of pDDIs in cancer patients receiving chemotherapy. Levels of pDDIs are imperative for healthcare professionals to evaluate their potential clinical significance and rationalize the patients' treatment. All interactions are not equally harmful, therefore, classification of the identified interactions into different levels helps in proper management of these interactions. The more frequent occurrence of major pDDIs is an important finding of this study, necessitating the need of strict monitoring of patients as these interactions carry higher potential for causing life threatening adverse reactions. A study reported similar results in hematological malignancies [40] however, in majority of studies moderate pDDIs are more frequent [41,45,46].
The use of polypharmacy is prevalent among cancer patients. The significant association of > 7 prescribed drugs and ≥ 3 anti-cancer drugs with pDDIs in present study are coherent with other studies both in oncology setting and other specialties [17,22,27,47]. The presence of polypharmacy in cancer patients demands the screening of prescribed medications for timely prediction and prevention or minimization of any unwanted negative consequences as polypharmacy is inevitable among cancer patients. The univariate analysis estimated a significant association of pDDIs with hospitalized patients, combination chemotherapy and patients having solid malignancy however, they were insignificant in multivariate analysis. Moreover, like other studies age, gender, metastasis and treatment intent had insignificant association with pDDIs [17,34,45] however, few studies have reported a significant association of pDDIs with gender, age and type of cancer [28,34,48,49].
List of most frequent pDDIs particularly those of contraindicated, major and moderate severity are of utmost importance for health care providers. It can aid in the selective screening of pDDIs by overburdened health care professionals. Such information is needed for health care professionals to estimate the risk in specific patients and guide their therapeutic decision making [8]. Patients at risk of these interactions may be given special attention and their therapy may be closely monitored for any potential adverse effect.
There are a few potential limitations of this study. Although, this work explored pharmacoepidemiology of pDDIs in cancer patients, the exact extent of patient suffering due to these interactions were not studied. The study was conducted only in two hospitals, which is the second point which may limit the generalizability of this study. Moreover, only one drug interactions screening database (Micromedex Drug-Reax®) was used for the identification of pDDIs, however, other sources are also available which may not necessarily give the same results. Further, we have only included hospitalized cancer patients receiving intravenous and/ or oral anticancer agents. In hospital settings, mostly intravenous anticancer therapy is provided. We don't include cancer patients treated in homebased care settings which could provide a different DDIs pattern due to frequent use of oral anticancer agents.

Conclusions
This study points out a high prevalence of pDDIs among cancer patients treated with anti-cancer agents. Majority of interactions were of major and moderate severity. Patients with polypharmacy i.e. > 7 all prescribed drugs or ≥ 3 anticancer drugs had a significantly increased risk of pDDIs. Whereas, list of most frequently identified major and moderate interactions will aid in timely identification, prevention and management of pDDIs and their adverse outcome in cancer patients. Moreover, strict patient monitoring is recommended especially in patients with > 7 all prescribed drugs or ≥ 3 anticancer drugs for timely prevention and/or management of negative clinical outcomes associated with these interactions particularly those involving cytotoxic drugs.