Background: Patients with cancer are particularly susceptible to drug interactions (DIs), but the extent of the problem has received limited attention. We aimed to evaluate the frequency of interactions with anticancer agents in a group of cancer patients. Methods: The study was performed in a Belgian teaching hospital. One hundred and twenty-two patients with solid malignancies were included. A comprehensive drug history was performed by a clinical pharmacist. Three renowned DI compendia were used to identify DIs. Results: Forty-one potential interactions involving an anticancer agent and considered to be clinically significant were identified among 25% of patients. The anticancer drugs mostly involved were cisplatin and methotrexate, and the most frequent co-medications involved were vitamin K antagonists, proton pump inhibitors and diuretics. In the majority of cases, the potential adverse consequence was increased toxicity of the anticancer agent and/or of the co-medication. Less than 10% of DIs were identified by the three compendia. Conclusions: Preventive measures should be taken to avoid increased toxicity or decreased efficacy of the drugs. Most of the time, this simply involves surveillance of biological or clinical parameters. Collaboration with a clinical pharmacist may be useful for the prescribing physician.

Cancer patients are particularly susceptible to drug interactions (DIs) for several reasons. Many of them are ≥65 years old, have several comorbidities and frequently take many medications [1]. The risk of interactions is further increased by (1) multiple prescribers for a single patient [2], and (2) the use, often without the awareness of prescribers, of over-the-counter (OTC) drugs, herbs or vitamins [3].

Anticancer agents usually have a narrow therapeutic index, and therefore, a slight increase or decrease in cytotoxic activity due to a DI can have major consequences [3]. An observational study using medical records and autopsy analysis showed that 4% of deaths among cancer patients were caused by chemotherapy itself, and serious drug-drug or drug-disease interactions were sometimes suspected [4].

Renowned databases [5, 6, 7] and literature reviews [3, 8, 9, 10, 11] display dozens of DIs involving anticancer drugs. For example, fluorouracil can increase the anticoagulant effect of coumarins [3, 9], cotrimoxazole or pantoprazole may increase the toxicity of methotrexate [8, 12], and even fatal cases of fluorouracil or methotrexate toxicity have been reported in patients who receive sorivudine or non-steroidal anti-inflammatory drugs (NSAIDs), respectively [3, 8]. Interactions can also occur with food, alcohol or herbs. For example, some components of grapefruit juice can increase the bioavailability of drugs through inhibition of the cytochrome P450 3A4 [13], St. John's wort may accelerate the metabolism of imatinib [14], and alcohol may increase the toxicity of methotrexate [15].

Therefore, the risk of DIs in cancer patients is important to consider and has been the object of several studies [16, 17, 18, 19, 20, 21, 22]. A recent systematic review found that approximately one third of ambulatory cancer patients had potential DIs, but only limited data were found on the clinical consequences of DIs [10]. Individual studies had important weaknesses. First, most studies were retrospective in nature. Second, none of them used more than one source of information to detect DIs, although there is good evidence that no single source has sufficient sensitivity to detect DIs [23, 24, 25, 26, 27]. Third, previous studies did not take into account interactions with complementary or alternative medications, herbs or food.

The objective of this study was to evaluate, in a group of cancer patients, the frequency of clinically relevant interactions between anticancer agents and any of the following: other prescribed drug, other anticancer agent, OTC drugs, herbs, vitamins, grapefruit juice, alcohol or tobacco. Secondary objectives were to identify the types of drugs mostly involved and to describe adverse consequences and management.

The study was conducted at CHU Mont-Godinne (Yvoir, Belgium), a 421-bed teaching hospital, from November 2008 to January 2009 (9 weeks). The study was approved by the local Ethics Committee. All patients with a solid tumor and receiving chemotherapy were eligible. Patients were included if, during the study period, they came to the day hospital to be administered an intravenous product (mostly their chemotherapy, and in a few cases disphosphonates or albumin) or if they were admitted to the oncology ward. Patients who did not take any drug in addition to their chemotherapy regimen or with only short-term treatment linked to chemotherapy (such as a setron or glucocorticoids) were excluded. Each patient provided oral informed consent. As the study was observational in nature, written consent was not required.

For each patient included, a clinical pharmacist performed a detailed drug history with the patient, using a piloted structured questionnaire (available upon request). The objective was to precisely identify all drugs taken by the patient in addition to the chemotherapy regimen (named ‘co-medications' in this paper), including the following: prescribed drugs, OTC drugs, plants and vitamins. In addition, the clinical pharmacist collected the following data from the electronic health record and through patient interview: age, type of cancer, chemotherapy regimen, tobacco, alcohol and coffee consumption, contact details of the general practitioner (GP) and the community pharmacist. When patients could not remember all their medications and if they agreed, the community pharmacist was called, and when necessary, the GP could also be contacted. Sometimes the patients themselves provided additional information when they were back home.

The clinical pharmacist then searched for DIs with each anticancer agent taken by each patient. All anticancer agents were taken into account, irrespective of the type of agent (e.g., monoclonal antibodies and protein kinase inhibitors were also included), route of administration (i.e. intravenous as well as oral agents were considered) and day(s) of administration. Three DI sources were used: Micromedex® Healthcare Series online, Epocrates© online (free version) and Stockley's© Drug Interactions [5, 6, 7]. These sources were selected because they have good performance rankings in terms of sensitivity, specificity, completeness or ease of use [23, 28, 29, 30, 31]. A selection of the most clinically relevant DIs was done. Micromedex classifies DIs into 5 categories of severity: contraindicated, major, moderate, minor, and unknown. For this study, only the first 3 were selected. In Epocrates, DIs are classified as ‘contraindicated', ‘avoid/use alternative', ‘monitor/modify treatment', ‘therapeutic advantage', and ‘caution advised'. Only the first 3 classes were selected. In Stockley's, there is no explicit ranking system in terms of clinical significance, severity or level of evidence. Using the narrative text describing each interaction, we excluded DIs described as ‘unlikely' or ‘of minor clinical significance'.

All interactions described in at least one of the aforementioned sources were recorded. Prevalence was calculated, and a descriptive analysis of the characteristics of DIs was done with regard to drugs involved, severity, adverse consequences, management, and source of information.

During the study period, 136 patients with a solid tumor and receiving chemotherapy were evaluated for inclusion. Fourteen patients were excluded because they had no co-medications or only a short-term treatment linked to chemotherapy. The characteristics of patients included (n = 122) are described in table 1 . The mean age was 63 years (range 29-85). Patients took an average of 7.5 co-medications (range 1-22).

Table 1

Characteristics of patients (n = 122)

Characteristics of patients (n = 122)
Characteristics of patients (n = 122)

A total of 41 potential interactions considered as clinically significant were identified. Thirty patients (24.6%) had at least 1 interaction involving their anticancer agent(s), 9 patients (7.4%) had 2 interactions and 1 patient (0.8%) had 3 interactions. Beside interactions between two anticancer agents, 23 patients (18.9%) had at least 1 interaction involving their usual therapy.

The characteristics of DIs are presented in table 2. Among the 18 interactions described in Micromedex, 10 (55.6%) were considered as of major severity and 11 (61.1%) got an excellent documentation. In contrast, only 2 (10%) of the 20 interactions found with Epocrates were set in the ‘contraindicated' or ‘avoid/use alternative' group. Only 7% of interactions were reported in the three sources, and 66% were reported in only one of the three sources. Potential adverse consequences were far more often increased toxicity than decreased efficacy. For interactions between an anticancer agent and a drug taken for a different purpose, the affected drug was the latter in two third of cases and the anticancer agent in one third of cases. With regard to management, most of the time surveillance was required and sufficient. Sixteen DIs (39.0%) were pharmacodynamic and 22 (53.7%) were pharmacokinetic in nature. Three (7.3%) interactions had an unknown mechanism.

Table 2

Characteristics of interactions (n = 41)

Characteristics of interactions (n = 41)
Characteristics of interactions (n = 41)

The anticancer drugs mostly involved in interactions were cisplatin (n = 12, including 8 with another anticancer agent), methotrexate (n = 6, including 1 with another anticancer agent), lapatinib (n = 4), and fluorouracil (n = 4). The most frequent co-medications interacting with chemotherapy were diuretics (n = 5), vitamin K antagonists (n = 4), proton pump inhibitors (n = 4), anticonvulsants (n = 2), and lipid-modifying agents (n = 2). The only OTC drug and the only vitamin involved in an interaction were an NSAID and folic acid. Proton pump inhibitors were not available as OTC drugs at the time of the study.

A detailed description of the 41 interactions is provided in table 3.

Table 3

Description of interactions

Description of interactions
Description of interactions

The present study shows that almost 25% of patients had at least 1 interaction of clinical significance involving their anticancer agent(s). Even after discounting interactions between two anticancer drugs, an interaction was still detected in 19% of patients.

Although direct comparison is difficult, the incidence seems to be higher than in previous studies. One study that focused on oxaliplatine and irinotecan found a clinically significant DI in 12% of patients [18]. Another study that focused on pharmacokinetic interactions affecting the activity of the anticancer agent found a 5% prevalence rate [20]. Other studies found similar [10, 16, 17] or higher [19, 21] rates of DIs, but the interactions considered were not limited to DIs with anticancer agents.

This high incidence might be a consequence of our attempt to maximize sensitivity. This was done in two ways. First, special attention was paid to obtain comprehensive drug history. Second, three different sources were used to identify DIs. Many studies have shown discrepancies in listing and rating systems between different DI sources [23, 24, 25, 26]. The three sources selected - even if they are renowned and widely used worldwide [23, 26, 31] - are no exception to the rule. Only 3 (7.3%) interactions were reported in the three sources and 27 (65.9%) were reported in only one of them. This highlights the need to use a combination of sources of information, even when focusing on clinically significant interactions.

One fourth of DIs involved two anticancer agents. These combinations usually belong to validated protocols, and adverse effects are usually carefully monitored by oncologists.

In contrast, more than half of the interactions involved a prescribed drug other than an anticancer agent. The interacting drugs were almost always prescribed by different prescribers (e.g., an oncologist and a GP). This increases the risk of not detecting the interaction, and therefore, the risk of an adverse drug event (ADE).

Similarly to previous studies, the use of complementary medicines or products was frequent [32, 33]. Prescribers are often unaware of their use for several reasons: for example, because patients fear disapproval [3, 8, 34, 35]. Nevertheless, the prevalence of interactions with OTC drugs, plants and vitamins was found to be low. This could be due to the low sensitivity of the sources used, at least for DIs with plants. In fact, no interaction involving a plant was detected with any of the three sources used, but a separate search in two specific sources of information [36, 37] identified 2 DIs with plants (garlic inducing the metabolism of etoposide, and black cohosh increasing the toxicity of paclitaxel). Stockley's Herbal Medicine Interactions textbook was not available at the time of the present study, but could be used in future studies [38].

Different recommendations can be made to decrease the risk of DIs in clinical practice.

First, oncologists and GPs should pay special attention when prescribing drugs more likely to interact. The need for co-medication has to be reassessed. If both drugs are necessary, most of the time, clinical or biological surveillance is required. Some doses might need adjustment (e.g., coumarins, fentanyl), and in some cases, the co-medication (e.g., NSAIDs) can be temporarily discontinued. If the latter is not possible, the drug can be replaced by another drug with a lower potential to interact (e.g., ranitidine instead of proton pump inhibitors) [5, 11, 39, 40]. Although these options seem relatively easy to implement in practice, oncologists are often reluctant to apply them [41]. They acknowledge little communication with pharmacists and have confidence in their ability to manage various degrees of toxicity.

Second, clinical pharmacists can contribute to better prevention and detection in different ways [12, 37]. They can compound educational tools for prescribers, answer specific questions asked by prescribers or even play a proactive role by checking each patient's therapy.

Third, patients themselves should be aware of the risk of DIs. They should keep an updated list of their medications and present this list to each doctor they see [42], they should not practice self-medication without talking to their doctor and ought to pay attention to the warning signs of ADEs [11].

Fourth, the classifications of the severity and scientific evidence of DIs in existing sources of information should be standardized to minimize confusion [24].

Finally, automated detection of DIs through computerized prescribing order entry and clinical decision support systems could be valuable to better detect DIs. However, this would require sharing prescription data across care settings. Unfortunately, this is currently not available in most cases.

The present study has several limitations. First, similarly to previous studies [16, 17, 18, 20, 41], we did not evaluate the clinical impact of DIs. A longitudinal cohort design and a larger sample would have been required. Nevertheless, we excluded DIs of minor clinical significance. Even though not all patients taking two interacting drugs will experience an ADE, harmful consequences must not be neglected [43]. In the present study, a patient taking warfarin had her international normalized ratio rising to 8.2, following the use of capecitabine and lapatinib. Second, the study was monocentric. However, the center has large activities in oncology, and the patients included were cared for by several oncologists. Third, the prevalence of DIs might have been underestimated, because few patients were taking a tyrosine kinase inhibitor or methotrexate, two drugs with a large potential to interact. A new study including patients with hematological cancers could provide valuable information.

In conclusion, this study confirms that most cancer patients take many drugs besides their chemotherapy and shows that a fourth of patients have at least 1 DI of clinical relevance. Unfortunately, little is known about the prevalence of ADEs due to these interactions. Despite this, it is essential that preventive measures are taken to avoid increased toxicity or decreased efficacy of the drugs. Most of the time, this simply involves surveillance of biological or clinical parameters. Collaboration with a clinical pharmacist might also be useful. An educational leaflet focusing on the most frequent and relevant DIs identified in the present study has been made available for prescribers.

We thank all patients and oncologists who participated in the study. We thank J.D. Hecq, L. D'Hondt and P. Gillet for their support and for facilitating the conduct of this study.

The authors declare that they have no conflict of interest.

1.
Hanigan MH, Dela Cruz BL, Thompson DM, Farmer KC, Medina PJ: Use of prescription and nonprescription medications and supplements by cancer patients during chemotherapy: questionnaire validation. J Oncol Pharm Pract 2008;14:123-130.
[PubMed]
2.
Green JL, Hawley JN, Rask KJ: Is the number of prescribing physicians an independent risk factor for adverse drug events in an elderly outpatient population? Am J Geriatr Pharmacother 2007;5:31-39.
[PubMed]
3.
Blower P, de Wit R, Goodin S, Aapro M: Drug-drug interactions in oncology: why are they important and can they be minimized? Crit Rev Oncol Hematol 2005;55:117-142.
[PubMed]
4.
Buajordet I, Ebbesen J, Erikssen J, Brors O, Hilberg T: Fatal adverse drug events: the paradox of drug treatment. J Intern Med 2001;250:327-341.
[PubMed]
5.
Stockley IH: Stockley's Drug Interactions. Chicago, Pharmaceutical Press, 2008.
6.
Thomson Reuters Micromedex® 1.0 (Healthcare series), 2009.
7.
Epocrates® online. 2009. https://online. epocrates.com.
8.
Beijnen JH, Schellens JH: Drug interactions in oncology. Lancet Oncol 2004;5:489-496.
[PubMed]
9.
Riechelmann RP, Saad ED: A systematic review on drug interactions in oncology. Cancer Invest 2006;24:704-712.
[PubMed]
10.
Riechelmann RP, Del Giglio A: Drug interactions in oncology: how common are they? Ann Oncol 2009;20:1907-1912.
[PubMed]
11.
Rédaction La Revue Prescrire: Eviter les effets indésirables par interactions médicamenteuses: comprendre et décider. Rev Prescrire 2006;26(suppl 278).
12.
Ranchon F, Vantard N, Gouraud A, Schwiertz V, Franchon E, Pham BN, Vial T, You B, Bouafia F, Salles G, Rioufol C: Suspicion of drug-drug interaction between high-dose methotrexate and proton pump inhibitors: a case report - should the practice be changed? Chemotherapy 2011;57:225-229.
[PubMed]
13.
Uno T, Yasui-Furukori N: Effect of grapefruit juice in relation to human pharmacokinetic study. Curr Clin Pharmacol 2006;1:157-161.
[PubMed]
14.
Smith P, Bullock JM, Booker BM, Haas CE, Berenson CS, Jusko WJ: The influence of St. John's wort on the pharmacokinetics and protein binding of imatinib mesylate. Pharmacotherapy 2004;24:1508-1514.
[PubMed]
15.
Neuman MG, Cameron RG, Haber JA, Katz GG, Malkiewicz IM, Shear NH: Inducers of cytochrome p450 2e1 enhance methotrexate-induced hepatotoxicity. Clin Biochem 1999;32:519-536.
[PubMed]
16.
Riechelmann RP, Tannock IF, Wang L, Saad ED, Taback NA, Krzyzanowska MK: Potential drug interactions and duplicate prescriptions among cancer patients. J Natl Cancer Inst 2007;99:592-600.
[PubMed]
17.
Riechelmann RP, Zimmermann C, Chin SN, Wang L, O'Carroll A, Zarinehbaf S, Krzyzanowska MK: Potential drug interactions in cancer patients receiving supportive care exclusively. J Pain Symptom Manage 2008;35:535-543.
[PubMed]
18.
Jansman FG, Idzinga FS, Smit WM, de Graaf JC, Coenen JL, Sleijfer DT, Brouwers JR: Classification and occurrence of clinically significant drug interactions with irinotecan and oxaliplatin in patients with metastatic colorectal cancer. Clin Ther 2005;27:327-335.
[PubMed]
19.
Riechelmann RP, Moreira F, Smaletz O, Saad ED: Potential for drug interactions in hospitalized cancer patients. Cancer Chemother Pharmacol 2005;56:286-290.
[PubMed]
20.
Lemachatti J, Leveque D, Beretz L, Bergerat JP: Potential pharmacokinetic interactions affecting antitumor drug disposition in cancer patients. Anticancer Res 2009;29:4741-4744.
[PubMed]
21.
van Leeuwen RW, Swart EL, Boom FA, Schuitenmaker MS, Hugtenburg JG: Potential drug interactions and duplicate prescriptions among ambulatory cancer patients: a prevalence study using an advanced screening method. BMC Cancer 2010;10:679.
[PubMed]
22.
Girre V, Arkoub H, Puts MT, Vantelon C, Blanchard F, Droz JP, Mignot L: Potential drug interactions in elderly cancer patients. Crit Rev Oncol Hematol 2011;78:220-226.
[PubMed]
23.
Vonbach P, Dubied A, Krahenbuhl S, Beer JH: Evaluation of frequently used drug interaction screening programs. Pharm World Sci 2008;30:367-374.
[PubMed]
24.
Wong CM, Ko Y, Chan A: Clinically significant drug-drug interactions between oral anticancer agents and nonanticancer agents: profiling and comparison of two drug compendia. Ann Pharmacother 2008;42:1737-1748.
[PubMed]
25.
Abarca J, Malone DC, Armstrong EP, Grizzle AJ, Hansten PD, Van Bergen RC, Lipton RB: Concordance of severity ratings provided in four drug interaction compendia. J Am Pharm Assoc 2004;44:136-141.
26.
Vitry AI: Comparative assessment of four drug interaction compendia. Br J Clin Pharmacol 2007;63:709-714.
[PubMed]
27.
Wang LM, Wong M, Lightwood JM, Cheng CM: Black box warning contraindicated comedications: concordance among three major drug interaction screening programs. Ann Pharmacother 2010;44:28-34.
[PubMed]
28.
Barrons R: Evaluation of personal digital assistant software for drug interactions. Am J Health Syst Pharm 2004;61:380-385.
[PubMed]
29.
Clauson KA, Polen HH, Marsh WA: Clinical decision support tools: performance of personal digital assistant versus online drug information databases. Pharmacotherapy 2007;27:1651-1658.
[PubMed]
30.
Clauson KA, Marsh WA, Polen HH, Seamon MJ, Ortiz BI: Clinical decision support tools: analysis of online drug information databases. BMC Med Inform Decis Mak 2007;7:7.
[PubMed]
31.
Perkins NA, Murphy JE, Malone DC, Armstrong EP: Performance of drug-drug interaction software for personal digital assistants. Ann Pharmacother 2006;40:850-855.
[PubMed]
32.
Sparreboom A, Cox MC, Acharya MR, Figg WD: Herbal remedies in the united states: potential adverse interactions with anticancer agents. J Clin Oncol 2004;22:2489-2503.
[PubMed]
33.
Werneke U, Earl J, Seydel C, Horn O, Crichton P, Fannon D: Potential health risks of complementary alternative medicines in cancer patients. Br J Cancer 2004;90:408-413.
[PubMed]
34.
McCune JS, Hatfield AJ, Blackburn AA, Leith PO, Livingston RB, Ellis GK: Potential of chemotherapy-herb interactions in adult cancer patients. Support Care Cancer 2004;12:454-462.
[PubMed]
35.
Yeung KS, Gubili J: Clinical guide to herb-drug interactions in oncology. J Soc Integr Oncol 2007;5:113-117.
[PubMed]
36.
Solimando DA: Drug Information Handbook for Oncology. Hudson, Lexi Comp, 2007.
37.
About herbs, botanicals and other products. Integrative Medicine Service, Memorial Sloan-Kettering Cancer Center, New York. http://www.mskcc.org/aboutherbs.
38.
Williamson E, Driver S, Baxter K: Stockley's Herbal Medicines Interactions, ed 2. London, Pharmaceutical Press, 2013.
39.
Mallet L, Spinewine A, Huang A: The challenge of managing drug interactions in elderly people. Lancet 2007;370:185-191.
[PubMed]
40.
Bink S, De Troy E: Interactions during oncological therapy: workshop on oncology. Leuven, ESCP Leuven, 2008.
41.
Sokol KC, Knudsen JF, Li MM: Polypharmacy in older oncology patients and the need for an interdisciplinary approach to side-effect management. J Clin Pharm Ther 2007;32:169-175.
[PubMed]
42.
Cancer patients at risk of drug interactions. CA Cancer J Clin 2007;57:258-259.
[PubMed]
43.
Riechelmann RP: Drug combinations with the potential to interact among cancer patients. Support Care Cancer 2007;15:1113-1114.
[PubMed]