Potential Drug Interactions in Cancer Patients Receiving Supportive Care Exclusively

Article Outline

• Abstract
• Introduction
• Patients and Methods
  • Study Design and Setting
  • Drug Interaction Screening
  • Statistical Considerations
• Results
  • Study Cohort
  • Potential Drug Interactions
• Discussion
• References
• Copyright

Abstract 

Cancer patients at the end of life often take many medications and are at risk for drug interactions. The purpose of this study was to describe the epidemiology of potential drug interactions in cancer patients receiving supportive care exclusively. We retrospectively reviewed the charts of consecutive adult cancer outpatients attending palliative care clinics at the Princess Margaret Hospital, Toronto, Canada. Drugs were screened for interactions by the Drug Interaction Facts software, which classifies interactions by levels of severity (major, moderate, and minor) and scientific evidence (1–5, with 1=the strongest level of evidence). Among 372 eligible patients, 250 potential drug interactions were identified in 115 patients (31%, 95% confidence interval 26%–36%). The most common involved warfarin and phenytoin. Most interactions were classified as being of moderate severity (59%) and 42% of them were supported by Levels 1–3 of evidence. In multivariable analysis, increasing age (P<0.001), presence of comorbidity (P=0.001), cancer type (brain tumors, P<0.001), and increasing number of drugs (P<0.001) were associated with risk of drug interactions. Potential drug interactions are common in palliative care and mostly involve warfarin and anticonvulsants. Older patients, those with comorbid conditions, brain tumor patients, and those taking many medications are at greater risk of drug interactions.

Key Words: Drug interaction, drug therapy, palliative care, adverse drug event, cancer

Back to Article Outline

Introduction 

Drug combinations with potential to interact are common in medical practice, although their frequency in general medicine has been variable, depending on the patient population, study design, and the screening methods used to identify interactions. In general medical wards, the rate of potential drug interactions has been approximately 60%.¹, ², ³ Studies conducted in emergency departments found frequencies of potential drug interactions in the range of 16%–47%.⁴, ⁵, ⁶ Ambulatory patients with variable clinical conditions who were screened for potential drug interactions by their family physician were found to be exposed to a potential drug interaction in almost 70% of cases.⁷

Drug interactions are divided into three groups depending on the underlying mechanism of interaction: pharmacodynamic, pharmacokinetic, and pharmaceutical.⁸ A pharmacodynamic interaction results from combining two drugs with similar mechanisms of action, which may behave in a synergistic, additive, or antagonistic manner. A pharmacokinetic interaction occurs when a drug alters the absorption, distribution, metabolism (most often due to interaction with the cytochrome P450 (CYP) hepatic enzymes), and/or excretion of another medication.⁸ A pharmaceutical interaction occurs when mixing chemically incompatible drugs outside the body, as for example, incompatibility of phenobarbital with opioid analgesics when mixed in the same syringe, resulting in inactivation of one or both drugs.⁸ The majority of the literature on drug interactions published thus far describes pharmacokinetic interactions involving the CYP enzymes, in many cases with the use of drug probes.⁹

Cancer patients often receive numerous medications, including antineoplastic agents, drugs to treat comorbid conditions, and supportive care medications. These multiple treatments increase the risk of receiving drugs with potential to interact. Furthermore, even though there are no definitive studies about the pharmacokinetic parameters in patients with cancer, a drug's pharmacokinetics may theoretically be altered in these patients because of impaired drug absorption due to mucositis, variation in volume of distribution secondary to lower levels of serum-binding proteins and edema, and altered excretion in patients with renal and/or hepatic dysfunction. Although several studies have evaluated potential drug interactions in general medicine, only a few studies have examined the frequency of potential drug interactions in patients with cancer.¹⁰, ¹¹, ¹² An evaluation of causes of death in a Norwegian hospital identified 18% of more than 700 deaths to be associated with adverse drug reactions (including drug interactions) and 4% of all cancer-related deaths were likely to be associated with a serious interaction.¹⁰ A retrospective study of 100 hospitalized cancer patients found that 63% were exposed to at least one drug combination with the potential to interact.¹¹ A recent study of 405 ambulatory cancer patients receiving cancer-directed therapy found that one-third were at risk of drug interactions.¹²

Although the frequency of potential drug interactions has been examined in ambulatory cancer patients receiving cancer-directed therapy¹² and hospitalized cancer patients,¹¹ we are unaware of studies about the epidemiology of drug interactions in the palliative care setting. Patients receiving exclusively supportive care comprise a unique population with respect to symptomatology, emotional aspects, and medication profile.¹³, ¹⁴, ¹⁵ We hypothesized that cancer patients at the end of life may be at a greater risk of drug interactions, that is, are more likely to receive drug combinations with the potential to interact, as compared to our cross-sectional study with patients who were receiving anticancer treatment because of their high burden of disease, which in turn, requires more drugs to treat cancer-related symptoms, such as pain and cachexia.¹⁴, ¹⁶, ¹⁷, ¹⁸ The primary objective of this study was to evaluate the epidemiology of potential drug interactions in cancer patients receiving exclusive supportive care at the Princess Margaret Hospital (PMH).

Back to Article Outline

Patients and Methods 

Study Design and Setting 

This was a retrospective cross-sectional study conducted at PMH, the largest cancer facility in Canada, and was approved by the PMH Research Ethics Board. Approximately 1,000 new patients are seen by the PMH Palliative Care Service each year, 80% of whom are outpatients.¹⁹ Of these, 75% are scheduled visits to the palliative care clinics, which are available five full days a week, and 25% are urgent, same-day outpatient consultations seen by the urgent consultation team. Approximately 20% of patients are referred to the clinic for pain management, 20% for control of other symptoms and 60% for palliative care planning.¹⁵ During the initial medical visit to the palliative care clinic, which lasts approximately 90–120minutes, patients are initially seen by a palliative care Registered Nurse Case Manager, who briefly assesses the patient, collects their medication list, and screens them for distressful symptoms using the Edmonton Symptom Assessment System. The palliative care physician then performs a thorough medical and psychosocial assessment.

We reviewed the medical records of cancer patients attending the palliative care clinics at PMH. We included consecutive adult patients with solid or hematological malignancies, who were receiving supportive care exclusively and who had at least one consultation in the palliative care clinic. Patients still receiving systemic antineoplastic therapy, those seen as an urgent consultation, inpatients at the time of the first consultation and those whose clinical notes lacked information on medications were excluded. Patients who were being treated with antineoplastic therapy and those who were hospitalized were considered ineligible because they would not reflect the profile of the typical palliative care cancer outpatient; they were either too unwell (inpatients) or would represent a different cancer population, which has been evaluated for potential drug interactions in a previous study.¹² We also excluded patients seen on an urgent basis because a comprehensive medical assessment is not performed during such consultations, which are focused on the urgent matter(s) that precipitated the referral. In contrast, during the first nonurgent consultation with palliative care at PMH, patients are initially seen by an advanced practice nurse, who collects the names of their medications and screens them for distressful symptoms; the patients are then assessed by a palliative care physician. Following the consultation, physicians dictate a thorough clinical note following preset guidelines: reason(s) for referral, history of present illness, symptomatology, current medications, comorbidities, social and family history, allergies, physical examination, summary, and plan. A comorbid illness was defined as a clinical condition that required pharmacologic therapy.

We collected data from the first consultation only because not all patients are subsequently followed by the PMH Palliative Care Service, as they may be referred to services located in their local community. Two authors (SNC and AO) abstracted the following data from patients' medical records: age, sex, cancer type (breast, gastrointestinal, genitourinary, gynecologic, lung, brain, hematological, and other), presence and number of comorbid illnesses, and medications recommended by the attending palliative care physician. Multivitamins, alternative/experimental, and herbs were not considered.

Drug Interaction Screening 

The medications screened for drug interactions were those that were recommended by the palliative care service in the first consultation, that is, they consisted of drugs that patients were already using before the consultation and were continued on, in addition to those prescribed by the palliative care doctor. Drugs were screened for interaction by the Drug Interaction Facts software, version 4.0,²⁰ (available at the web site: www.factsandcomparisons.com) and manually by one of the authors (RPR) for drugs not recognized by the program, using pharmacology textbooks. The software screens for potential drug interactions, provides a description of pharmacological mechanisms underlying the interaction (pharmacokinetic, pharmacodynamic, or unknown when there was not sufficient evidence from the literature about the mechanism of interaction) and classifies them by level of severity and scientific evidence (Table 1). The Drug Interaction Facts software has both sensitivity and specificity of 97% in detecting previously known drug combinations with potential to interact²¹ and has been previously used in cancer patients.¹¹, ¹²

Table 1. Drug Interactions Facts^a Software Classification Scheme of Levels of Severity and Scientific Evidence of Drug Interactions

Level of Severity	Description
1	Major: an adverse effect can cause permanent damage or life risk
2	Moderate: an adverse effect can harm and treatment is required
3	Minor: small or no clinical effect, with no treatment required
Level of Scientific Evidence	Type of Scientific Data
1	Established: adverse effect confirmed by large clinical trials
2	Probable: adverse effect with high likelihood of occurrence but without definitive randomized clinical trials
3	Suspect: adverse effect likely to occur; data derived from case reports
4	Possible: adverse effect may occur but data are scarce
5	Unlikely: adverse effect may theoretically occur

Statistical Considerations 

Based on previous studies,¹¹, ¹² a sample of 350 patients was chosen as a compromise between reasonable size and feasibility. Approximately 800 new ambulatory patients are seen by the palliative care clinics each year¹⁹ and, based on the Palliative Care Service internal records, we estimated that approximately 60% would be eligible, that is, would not be receiving anticancer therapy. Therefore, an eight-month period would suffice to achieve a sample of 350 patients. We reviewed the medical charts from November 2005 to July 2006, when the target number was achieved.

We used summary statistics to describe patient characteristics, frequency, and types of potential drug interactions. The number of medications for each subject was calculated by summing all drugs. When a medication contained two or more pharmacological compounds (e.g., acetaminophen combined with codeine), each drug was considered an individual medication for analysis; when a patient was taking the same medication in different schedules (e.g., long-acting and short-acting insulin for glycemic control), the drug was counted only once.

Logistic regression was used to identify factors associated with drug combinations that have the potential to interact. The dependent variable, potential drug interaction, was defined as at least one instance of a potential drug interaction with level of evidence less than or equal to 4 (i.e., a drug interaction supported by at least a few case reports, excluding those of Level 5, which are theoretical). Independent variables tested in the model included: age (continuous variable), cancer type (breast, gastrointestinal, genitourinary, gynecologic, lung, brain, hematological, and other), presence of a comorbid illness (yes or no), and number of medications (continuous variable). Gender was not included as a risk factor for drug interactions because information that gender would contribute to the regression analysis is relevant only in certain cancer types, given that some cancer types only occur in men or women. For binary or nominal variables, the group at lower risk of the outcome was chosen as the referent. Variables with univariate P-values <0.1 were entered into the multivariable model. In the multivariable model, predictors were considered statistically significant if the P-value was <0.05. SAS, version 9.1 was used for all analyses.

Back to Article Outline

Results 

Study Cohort 

Between November 2005 and July 2006, 633 patients were seen in the palliative care clinics and 372 patients met our inclusion criteria. Reasons for exclusion included: 150 patients were receiving antineoplastic therapy, 25 had incomplete information about their medications, 52 were seen as urgent consults, 20 were inpatients, and 14 were excluded for other reasons. For 11 patients, it was noted in their chart that they remembered all but one of their medications and they were included in the study. The characteristics of the 372 eligible patients are described in Table 2. The median age was 66 years (range, 22–94), approximately half were male, and the most common cancer type was gastrointestinal. The most common pharmacological classes of drugs taken by patients were opioids (67% of patients), laxatives/stool softeners (54%), acetaminophen (40%), and corticosteroids (38%). The median number of medications per patient was six (range, 0–21). The most common comorbid illnesses were cardiovascular diseases (36% of patients), diabetes (14%), musculoskeletal disorders (11%), respiratory diseases and thromboembolism (10% each), hypercholesterolemia (9%), chronic liver diseases (7%), hypothyroidism and neurological diseases (6% each), and psychiatric disorders (5%).

Table 2. Patient Characteristics

Characteristic	n	%
Total number of patients	372	100
Age (years), median (range)	66 (range, 22–94)	—
Sex
Female	181	49
Male	191	51
Cancer type
Gastrointestinal	112	30
Lung	59	16
Gynecologic	47	13
Genitourinary	45	12
Brain	20	5
Hematological	18	5
Breast	17	4
Other	54	14
Patients with≥1 comorbid condition	247	66
Comorbid conditions, median (range)	1 (range, 0–7)	—
Pharmacological class of drugs taken by patientsa
Opioid	251	67
Laxative/stool softener	200	54
Acetaminophen	151	41
Corticosteroid	140	38
Antidyspeptic	136	36
Cardiovascular	128	34
Antiemetic	107	29
Benzodiazepine	87	23
Nonsteroidal anti-inflammatory drug	66	18
Antidepressant	59	16
Statin	49	13
Antidiabetic	46	12
Anticonvulsant	38	10
Antibiotic	34	9
Anticoagulant/heparin	25	7
Bisphosphonate	21	6
Drugs taken by patient, median no. (range)	6 (range, 0–21)	—

Potential Drug Interactions 

Among the 372 patients, 250 drug combinations with potential to interact were identified in 115 patients (31%; 95% confidence interval [CI] 26%–36%, Table 3). The majority of potential interactions were classified as of moderate severity (59%) and 41.5% of them were supported by Level ≤3 of scientific evidence (i.e., supported by at least several case reports). The most common pharmacological mechanism underlying the potential drug interactions was pharmacokinetic. The most frequent drug combinations at risk for interaction involved phenytoin (26 cases), warfarin (16 cases), and aspirin (15 cases). The most commonly found instances of drugs with potential to interact of either moderate or major severity are summarized in Table 4. We decided not to describe the drug interactions classified as of minor severity because their clinical consequences are likely small.

Table 3. Drugs with the Potential to Interact Among Ambulatory Palliative Care Cancer Patients

Potential Drug Interactions	n	%
Number	250	100
Number of patients with≥1 potential drug interaction	115	31 (95% CI 26%–36%)
Median (range) number of potential drug interactions per patient	0 (0–10)
Severity of potential drug interactions
Major	25	10
Moderate	147	59
Minor	78	31
Level of scientific evidence for identified potential drug interactionsa
1	23	9
2	66	26.5
3	15	6
4	80	32
5	66	26.5
Mechanism of identified potential drug interactions
Pharmacokinetic	109	44
Pharmacodynamic	70	28
Unknown	71	28

CI=confidence interval.

Table 4. Most Common Instances of Drug Combinations with Potential to Interact

Drugs with Potential to Interact	Number of Cases	Description	Severity	Level of Scientific Evidencea
Pharmacokinetic drug interactions
Phenytoin+corticosteroids22	10	Both drugs may increase or decrease each other's liver metabolism	Moderate	2
Warfarin+corticosteroids28	8	Increase or decrease in warfarin anticoagulation effects; mechanism unknown	Moderate	4
Phenytoin+statins30	5	Phenytoin induces statins' hepatic metabolism with consequent lower serum levels of statins	Moderate	4
Phenytoin+acetaminophen31	6	Phenytoin induces acetaminophen's hepatic metabolism with consequent increased hepatic toxicity	Moderate	2
Phenytoin+ranitidine32	5	Increased or decreased phenytoin plasma concentration, possibly because of inhibition or induction of its liver metabolism	Moderate	4
Benzodiazepine+omeprazole33	5	Inhibition of benzodiazepine's liver metabolism, with increased sedation	Moderate	4
ASA+corticosteroids34	5	Reduced serum salicylate levels because of increased liver metabolism and renal elimination by steroids	Moderate	2
Risperidone+SSRI29	3	Increased plasma concentration of risperidone, possibly due to inhibition of its liver metabolism; risk of serotonin syndrome	Major	1
Pharmacodynamic drug interactions
Warfarin+acetaminophen27	8	Increased vitamin K antagonism by acetaminophen, with risk of bleeding	Moderate	2
Prochlorperazine+ACE inhibitor35	7	Additive hypotensive effects	Moderate	4
NSAID+SSRI36	5	Increased risk of upper gastrointestinal bleeding; mechanism unknown, possible additive effects	Moderate	2
ASA×ACE inhibitor37	5	Hypotensive effects of ACE inhibitors may be reduced because of inhibition of prostaglandins synthesis by ASA	Moderate	2

ASA=acetylsalysilic acid, SSRI=selective serotonin reuptake inhibitor, ACE inhibitor=angiotensin converting enzyme inhibitor, NSAID=nonsteroidal anti-inflammatory drug.

Three hundred fifty-three patients were included in the logistic regression model to determine risk factors associated with drug interactions; eight patients who were not receiving any medication and 11 who were taking only one drug were excluded from analysis because they were not at risk of drug interactions. Results of the univariable and multivariable analyses are presented in Table 5. In unadjusted analysis, advanced age, increasing number of drugs, presence of comorbid illness and cancer type were significantly associated with potential drug interactions. In adjusted analyses, all variables remained significant: increasing age (odds ratio [OR]: 1.05 for each additional year of age, 95% CI 1.02–1.08; P<0.001), increasing number of medications (OR: 1.3 per each additional drug, 95% CI 1.2–1.4; P<0.001), presence of comorbid illness (OR: 6.5, 95% CI 2.1–19.8; P=0.001), and cancer type (OR for brain vs. gastrointestinal tumors: 28.5, 95% CI 7.2–113 and other types of cancer vs. gastrointestinal tumors: 3.1, 95% CI 1.2–7.9; P<0.001).

Table 5. Univariate and Multivariable Analyses for Factors Associated with Drug Interactions

Variable	Unadjusted OR (95% CI)	Unadjusted P-value	Adjusted OR (95% CI)	Adjusted P-value
Age	1.05 (1.03–1.07)	<0.001	1.05 (1.02–1.08)	<0.001
Number of medications	1.4 (1.3–1.5)	<0.001	1.3 (1.2–1.4)	<0.001
Presence of comorbid illness
No	Referent	<0.001	Referent	=0.001
Yes	16.7 (6–46.9)		6.5 (2.1–19.8)
Cancer type
Gastrointestinal	Referent	<0.001	Referent	<0.001
Brain	12 (4.1–34.9)		28.5 (7.2–113)
Breast	0.4 (0.05–3.3)		1.3 (0.1–12.1)
Genitourinary	2.6 (1.1–6.1)		2 (0.7–5.4)
Gynecological	1.8 (0.7–4.4)		2.2(0.8–6.2)
Hematological	1.3 (0.3–5)		1.2(0.3–5.9)
Lung	2.8 (1.3–6.2)		1.5 (0.6–3.7)
Other	3.4 (1.6–7.4)		3.1(1.2–7.9)

CI=confidence internal, OR=odds ratio.

Back to Article Outline

Discussion 

The present study demonstrates that one-third of cancer patients receiving exclusively supportive care were exposed to at least one drug combination with the potential to interact. These numbers raise concern because almost 70% of potential drug interactions were classified as of either major or moderate severity and almost 42% were supported by good scientific evidence (level of evidence ≤3). More than half of the potential drug interactions we identified in our sample were pharmacokinetic. The large majority of potential drug interactions related to warfarin and phenytoin, although the most commonly used drugs by patients were opioids, laxatives, acetaminophen, and corticosteroids. Older patients, those taking more medications, those with comorbid conditions and those with brain tumors were at greater risk of drug interactions.

Patients with brain tumors were almost 30 times more likely to be exposed to drug interactions than gastrointestinal cancer patients. Anticonvulsants, which are often used by these patients, not only interact with many different medications but also are routinely prescribed in conjunction with steroids to treat cancer-related cerebral edema and seizures. There is evidence that when phenytoin is combined with dexamethasone, its plasma level may increase or decrease.²²

Likely due to a high burden of comorbidities, older patients were also at increased risk of drug interactions. Patients with comorbid conditions usually take more medications or use drugs with potential for drug–drug interactions.¹² Our previous study showed that patients taking medications to treat comorbid conditions were at greater risk of drug interactions than were patients using supportive care drugs only.¹² Similar to other studies,², ⁴, ¹¹, ¹² an increasing number of medications was associated with a greater potential for interactions in our population.

The frequency of potential drug interactions in noncancer populations has varied from a low of 16% among emergency room patients to a high of 70% in a population of ambulatory patients being treated by their family physician.⁵, ⁷ To date, only a few studies have evaluated drug interactions in cancer patients,¹¹, ¹² and to our knowledge this is the first to report the epidemiology of potential drug interactions in patients receiving only supportive care. A retrospective study undertaken by one of the authors of this study identified drug combinations with the potential to interact in 63% of 100 consecutive cancer patients admitted to a general oncology ward for treatment complications, supportive care, or invasive procedures.¹¹ It is likely that the proportion of patients at risk of drug interactions was higher than in the present study because hospitalized patients are usually more acutely ill than ambulatory patients and are more likely to receive more medications. In the study of hospitalized patients, the median number of drugs per patient was eight, as compared to our study, where the median number was six. In addition, infections and thromboembolic events were among the most common causes of hospital admission in the previous study;¹¹ both of these illnesses require pharmacological treatment with drugs that may interact, such as oral anticoagulants and antibiotics. A cross-sectional study of ambulatory cancer patients receiving antineoplastic therapy found that approximately one-third of 405 patients were exposed to at least one potential drug interaction.¹¹

The similar frequency of potential interactions in this study and the large study was not expected, because we hypothesized that cancer patients at the end of life would be at a greater risk of drug interactions when compared to cancer patients receiving antineoplastic therapy. In both studies, it was found that almost one-third of patients were receiving drug combinations with the potential to interact. This is likely because almost 90% of potential interactions encountered in the large study¹² involved nonanticancer agents, many of which were the same as those identified in our study (e.g., warfarin and anticonvulsants). In both studies, the majority of interactions were pharmacokinetic, and in most instances the underlying mechanism was interactions with the hepatic CYP enzymes. This is because many drugs taken by cancer patients, such as proton-pump inhibitors, certain opioids, and benzodiazepines, are either substrates or inhibitors/inducers of CYP enzymes.⁹

Limitations of our study should be noted. First, the population is limited to a single institution and, therefore, the generalizability to other palliative care settings is unknown. However, the patients and their medication profile were representative of other previously described palliative care cancer populations.¹⁴, ¹⁶, ¹⁸, ²³, ²⁴ Second, the data were collected retrospectively, and although we were rigorous in our inclusion criteria and excluded 25 patients whose charts provided incomplete information about medications, it is possible that we missed information on medications. Thus, even though the 30% frequency of drug interactions is consistent with other studies,¹² it is possible that it has been underestimated. In addition, because of the accuracy of the screening method,²¹ approximately one-quarter of potential drug interactions were theoretical interactions, for which the clinical consequences are unknown. Lastly, one of the major limitations of studies examining drug interactions is the difficulty in measuring the number of potential interactions that resulted in clinical adverse events. These are difficult to ascertain retrospectively, and it would obviously not be ethical to evaluate the clinical outcomes of drug interactions in a prospective way without modifying or stopping the medications involved. We are aware that there is a growing use of herbal/dietary supplements by the general population, including cancer patients, and concerns have been raised about pharmacological interactions between drugs and herbal compounds.²⁵, ²⁶ However, the evaluation of herbal–drug interactions was beyond the scope of this study.

Anticonvulsants are well known to interact with other drugs. In this study, as well as in two previous studies about the risk of drug interactions in cancer patients,¹¹, ¹² the association of phenytoin and corticosteroids was among the most commonly identified drug combinations with the potential to interact. In this combination, dexamethasone affects the hepatic metabolism of phenytoin, which can lead to either an increase or decrease in phenytoin serum concentrations.²² Warfarin is metabolized by the CYP 2C6 hepatic enzyme, conferring a high risk of interactions with other drugs metabolized by this enzyme, such as corticosteroids; bleeding may result from such interactions.²⁷, ²⁸ Although there were only three patients who were prescribed risperidone in combination with selective serotonin reuptake inhibitors, this combination is life-threatening and supported by Level 1 evidence (i.e., supported by large trials).²⁹ An increased plasma concentration of risperidone, possibly due to inhibition of its hepatic metabolism by the antidepressant, may lead to serotonin syndrome.²⁹

The best way to prevent drug interactions remains unknown. Alert guidelines, such as flyers containing drug combinations with the potential to interact or electronic messages that alert physicians when entering a patient's order, and computerized programs can increase recognition of such interactions and provide an important tool for screening. We suggest that subjects at high risk, such as brain tumor patients, those with comorbidities, and those receiving warfarin and anticonvulsants, be routinely screened for drug interactions. This is particularly important in the palliative care setting, where patients are usually quite unwell and already have a high burden of symptoms. Palliative care physicians should be aware of possible clinical consequences of drug interactions and we suggest that health professionals who work with these patients check phenytoin serum levels when a patient develops symptoms of phenytoin toxicity (e.g., ataxia and confusion) and perform coagulation tests when a patient who takes long-term warfarin is prescribed a new medication (especially antibiotics, acetaminophen and steroids). In addition, routine reassessment of the need for continuing certain medications at the end of life, such as statins, aspirin, and angiotensin converting enzyme inhibitors, could decrease the risk of adverse drug reactions and drug interactions.

In summary, drug combinations with the potential to interact are common among ambulatory cancer patients receiving exclusively supportive care and mostly involve warfarin and anticonvulsants. Development of alert guidelines and computer-based screening could help physicians to recognize potentially dangerous drug interactions and avoid undesirable adverse events in this already frail population. Population-based studies are warranted to evaluate the epidemiology of clinical adverse events resulting from drug interactions in oncology.

Back to Article Outline

References 

1. Egger SS, Drewe J, Schlienger RG. Potential drug-drug interactions in the medication of medical patients at hospital discharge. Eur J Clin Pharmacol. 2003;58:773–778
   • View In Article
   • MEDLINE
2. Geppert U, Beindl W, Hawranek T, et al. [Drug interactions in clinical practice. A pilot project for quality assurance in prescribing], Hautarzt. 2003;54:53–57
   • View In Article
   • MEDLINE
3. Glintborg B, Andersen SE, Dalhoff K. Drug-drug interactions among recently hospitalised patients—frequent but mostly clinically insignificant. Eur J Clin Pharmacol. 2005;61:675–681
   • View In Article
   • MEDLINE
   • CrossRef
4. Karas S. The potential for drug interactions. Ann Emerg Med. 1981;10:627–630
   • View In Article
   • Abstract
   • Full-Text PDF (295 KB)
   • CrossRef
5. Herr RD, Caravati EM, Tyler LS, et al. Prospective evaluation of adverse drug interactions in the emergency department. Ann Emerg Med. 1992;21:1331–1336
   • View In Article
   • Abstract
   • Full-Text PDF (542 KB)
   • CrossRef
6. Goldberg RM, Mabee J, Chan L, et al. Drug-drug and drug-disease interactions in the ED: analysis of a high-risk population. Am J Emerg Med. 1996;14:447–450
   • View In Article
   • Abstract
   • Full-Text PDF (390 KB)
   • CrossRef
7. Davidson KW, Kahn A, Price RD. Reduction of adverse drug reactions by computerized drug interaction screening. J Fam Pract. 1987;25:371–375
   • View In Article
   • MEDLINE
8. Stiztel CCR. Modern pharmacology with clinical application, Philadelphia, PA: Lippincott-Williams, 1997.
   • View In Article
9. Haddad A, Davis M, Lagman R. The pharmacological importance of cytochrome CYP3A4 in the palliation of symptoms: review and recommendations for avoiding adverse drug interactions. Support Care Cancer. 2007;15:251–257
   • View In Article
   • MEDLINE
   • CrossRef
10. Buajordet I, Ebbesen J, Erikssen J, et al. Fatal adverse drug events: the paradox of drug treatment. J Intern Med. 2001;250:327–341
    • View In Article
    • MEDLINE
    • CrossRef
11. Riechelmann RP, Moreira F, Smaletz O, et al. Potential for drug interactions in hospitalized cancer patients. Cancer Chemother Pharmacol. 2005;56:286–290
    • View In Article
    • MEDLINE
    • CrossRef
12. Riechelmann RP, Tannock IF, Wang L, et al. Potential drug interactions and duplicate prescriptions among cancer patients. J Natl Cancer Inst. 2007;99:592–600
    • View In Article
    • CrossRef
13. Jenkins CA, Schulz M, Hanson J, et al. Demographic, symptom, and medication profiles of cancer patients seen by a palliative care consult team in a tertiary referral hospital. J Pain Symptom Manage. 2000;19:174–184
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (152 KB)
    • CrossRef
14. Nauck F, Ostgathe C, Klaschik E, et al. Drugs in palliative care: results from a representative survey in Germany. Palliat Med. 2004;18:100–107
    • View In Article
    • MEDLINE
    • CrossRef
15. Riechelmann RP, Krzyzanowska MK, O'Carroll A, et al. Symptom and medication profiles among cancer patients attending a palliative care clinic. Support Care Cancer. 2007;15(12):1407–1412
    • View In Article
    • CrossRef
16. Twycross RG, Bergl S, John S, et al. Monitoring drug use in palliative care. Palliat Med. 1994;8:137–143
    • View In Article
    • MEDLINE
    • CrossRef
17. Koh NY, Koo WH. Polypharmacy in palliative care: can it be reduced?. Singapore Med J. 2002;43:279–283
    • View In Article
    • MEDLINE
18. Mercadante S, Fulfaro F, Casuccio A. Pattern of drug use by advanced cancer patients followed at home. J Palliat Care. 2001;17:37–40
    • View In Article
    • MEDLINE
19. Zimmermann C, Seccareccia D, Clarke A, et al. Bringing palliative care to a Canadian cancer center: the palliative care program at Princess Margaret Hospital. Support Care Cancer. 2006;14:982–987
    • View In Article
    • MEDLINE
    • CrossRef
20. Facts and comparisons 4.0. St. Louis: Wolters Kluwer Health; 2006;Available at http://online.factsandcomparisons.comAccessed January 23, 2007
    • View In Article
21. Barrons R. Evaluation of personal digital assistant software for drug interactions. Am J Health Syst Pharm. 2004;61:380–385
    • View In Article
    • MEDLINE
22. Lackner TE. Interaction of dexamethasone with phenytoin. Pharmacotherapy. 1991;11:344–347
    • View In Article
    • MEDLINE
23. Drummond SH, Peterson GM, Galloway JG, et al. National survey of drug use in palliative care. Palliat Med. 1996;10:119–124
    • View In Article
    • MEDLINE
    • CrossRef
24. Curtis EB, Walsh TD. Prescribing practices of a palliative care service. J Pain Symptom Manage. 1993;8:312–316
    • View In Article
    • MEDLINE
    • CrossRef
25. Williamson EM. Interactions between herbal and conventional medicines. Expert Opin Drug Saf. 2005;4:355–378
    • View In Article
    • CrossRef
26. Yoon SL, Schaffer SD. Herbal, prescribed, and over-the-counter drug use in older women: prevalence of drug interactions. Geriatr Nurs. 2006;27:118–129
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (162 KB)
    • CrossRef
27. Mahe I, Bertrand N, Drouet L, et al. Paracetamol: a haemorrhagic risk factor in patients on warfarin. Br J Clin Pharmacol. 2005;59:371–374
    • View In Article
    • MEDLINE
    • CrossRef
28. Costedoat-Chalumeau N, Amoura Z, Aymard G, et al. Potentiation of vitamin K antagonists by high-dose intravenous methylprednisolone. Ann Intern Med. 2000;132:631–635
    • View In Article
    • MEDLINE
29. Boyer EW, Shannon M. The serotonin syndrome. N Engl J Med. 2005;352:1112–1120
    • View In Article
    • CrossRef
30. Murphy MJ, Dominiczak MH. Efficacy of statin therapy: possible effect of phenytoin. Postgrad Med J. 1999;75:359–360
    • View In Article
    • MEDLINE
31. Brackett CC, Bloch JD. Phenytoin as a possible cause of acetaminophen hepatotoxicity: case report and review of the literature. Pharmacotherapy. 2000;20:229–233
    • View In Article
    • MEDLINE
    • CrossRef
32. Tenni P, Lalich DL, Byrne MJ. Life threatening interaction between tamoxifen and warfarin. BMJ. 1989;298:93
    • View In Article
    • MEDLINE
    • CrossRef
33. Gerson LB, Triadafilopoulos G. Proton pump inhibitors and their drug interactions: an evidence-based approach. Eur J Gastroenterol Hepatol. 2001;13:611–616
    • View In Article
    • MEDLINE
    • CrossRef
34. Edelman J, Potter JM, Hackett LP. The effect of intra-articular steroids on plasma salicylate concentrations. Br J Clin Pharmacol. 1986;21:301–307
    • View In Article
    • MEDLINE
35. White WB. Hypotension with postural syncope secondary to the combination of chlorpromazine and captopril. Arch Intern Med. 1986;146:1833–1834
    • View In Article
    • MEDLINE
36. de Jong JC, van den Berg PB, Tobi H, et al. Combined use of SSRIs and NSAIDs increases the risk of gastrointestinal adverse effects. Br J Clin Pharmacol. 2003;55:591–595
    • View In Article
    • MEDLINE
    • CrossRef
37. Ahmed A. Interaction between aspirin and angiotensin-converting enzyme inhibitors: should they be used together in older adults with heart failure?. J Am Geriatr Soc. 2002;50:1293–1296
    • View In Article
    • MEDLINE
    • CrossRef