Skip to content

Drug interactions occur when a substance affects the activity of a drug when both are administered together. The interaction of substances in this way may have consequences for the health of patients, including increasing the risk that side effects will occur during treatment.

As part of the PhRMA Foundation’s Safe and Effective Prescribing Initiative, this educational module examines the different kinds of drug interactions, the principal mechanisms that cause them, and how to understand complex drug interactions. It also is designed to help prescribers interpret the clinical significance of various types of drug interactions and learn how to access drug interaction information in product labeling and tertiary references. Learn about about the Safe and Effective Prescribing Initiative and view the other modules.

How to Use the Module

  1. Complete the self-assessment. This will determine your current level of understanding about adverse drug interactions (ADRs).
  2. Watch the video presentation of the module.
  3. Review the case study and answer the questions.
  4. For additional learning, access the resources and references document, which includes further information to expand your knowledge about ADRs.
  5. Return to the self-assessment and test your knowledge again.

Question 1:
Perturbation of plasma protein binding is the primary mechanism causing most clinically significant drug interactions with warfarin.

True or False?

Question 2:
Co-administration of bosentan and glyburide results in ~40% lower glyburide exposure and ~30% lower bosentan exposure.

What is a possible mechanism for this interaction?

A. Inhibition of CYP3A or other CYP450 enzyme
B. Induction of CYP3A or other CYP450 enzyme
C. Inhibition of gut transporter BCRP
D. Reduced solubility of both agents in the gut
E. A pharmacodynamic interaction

Question 3:
Co-administration of glyburide and bosentan results in an increased risk of liver enzyme elevations.1  This is likely a result of the pharmacokinetic interaction described in Question 2.

True or False?

Question 4:
Consider a new drug being developed for the treatment of a common indication, such as diabetes mellitus or high blood pressure.  In vitro and preclinical studies indicate the compound is extensively metabolized, primarily by CYP3A and partially by CYP2C19.  The compound does not appear to be an inducer or inhibitor of any CYP450 enzymes, but it is a potential inhibitor of P-glycoprotein (P-gp).

What drug interaction studies should the company developing the compound consider?

A. A DDI study with digoxin
B. Two DDI studies – one with midazolam and one with itraconazole
C. Two DDI studies – one with itraconazole and one with fluconazole
D. All of the above
E. A and C

Case Study 1
A 39-year-old man was admitted to the hospital with new onset seizures and a 2-week history of blurred vision. After a work-up, diagnoses included advanced HIV infection, CMV retinitis, and probable toxoplasmic encephalitis. Sulfadiazine and pyrimethamine were started for treatment of toxoplasmic encephalitis. Phenytoin was initiated at a dosage of 300 mg daily for seizure. The patient also received valganciclovir for treatment of CMV retinitis. After experiencing another seizure while hospitalized, the phenytoin dose was increased to 300 mg twice daily on 19 September. Antiretroviral therapy with efavirenz (EFV), emtricitabine, and tenofovir was started on 22 September. A plan was made to monitor both phenytoin and EFV plasma concentrations to ensure adequate safety and therapeutic efficacy.

On 30 September, the first measurement of EFV plasma concentration was obtained, with a result of 0.34 µg/mL. Target EFV concentrations of 1 – 4 µg/mL or >2.2 µg/mL have been proposed in the literature, while concentrations <1.1 µg/mL have been associated with treatment failure. Because EFV has a long half-life (40-55 hours), it was recognized the plasma level could be low because the sample was obtained prior to steady-state.

1. Other than obtaining the level before achieving steady-state, what are other potential explanations for the low EFV exposure? Consider the patient’s concurrent medications.

There was concern that phenytoin was the cause of the lower-than-expected EFV concentrations. As persistently low EFV exposure could put the patient at risk for treatment failure and development of virologic resistance, it was decided to rapidly taper the phenytoin and commence therapy with levetiracetam —an anticonvulsant without CYP modulation potential. In the meantime, the EFV dosage was increased from 600 to 800 mg/day under the assumption that CYP induction could continue for up to 7-14 days after discontinuation of phenytoin. Prior to implementation of the EFV dosage increase, a 2nd test result returned an EFV concentration of 0.58 µg/mL.

The patient received his last dose of phenytoin on 22 October. A third measurement of EFV level was obtained on 9 November, after what was predicted to be sufficient time for reversal of phenytoin’s enzyme induction. The result was an EFV concentration of 2.5 µg/mL. At that time the EFV dosage was reduced back to the recommended dose of 600 mg/day. On 16 November, the patient had an HIV RNA of 189 copies/mL and a CD4 cell count of 100 cells/mm3, which indicated continued virologic response to antiretroviral therapy. The patient’s latest measured EFV, after 3 weeks at a dosage of 600 mg/day, was 2 µg/mL.

Phenytoin levels measured after EFV therapy initiation showed a gradual increase, despite a stable phenytoin dosage and no changes in medications.

2. Explain the likely reasons for this increase in phenytoin plasma levels.

Case Study 2
A 75-year-old man presented to the hospital with symptoms of muscle weakness, stiffness, and dark urine. A diagnosis of rhabdomyolysis and acute renal failure was made. The patient’s medical history included hypercholesteremia for which he took atorvastatin for approximately 2 years. Verapamil treatment had recently been initiated for hypertension. Upon admission, all medications were discontinued and the patient was treated with intermittent hemodialysis. Genotyping results showed that the patient had a single-nucleotide polymorphism (SNP) within genes encoding the organic anion-transporting polypeptide (OATP) 1B1. He was also found to be a poor metabolizer of CYP2C19.

1. Discuss the causes of the patient’s symptoms, including the interactions between atorvastatin and verapamil and the effect of the SNP.

Case Study 3
A 78-year-old man visited a hospital for cough and increased sputum, for which he was prescribed clarithromycin, a macrolide antibiotic. The following day he was admitted to another hospital for loss of consciousness. His medical history included hypertension, atrial fibrillation and chronic kidney disease. He reportedly took two calcium-channel blockers, nifedipine and diltiazem, as well as carvedilol (a beta-blocker), irbesartan (angiotensin receptor blocker), isosorbide dinitrate, and dypiridamole. At admission his blood pressure was 96/38 mm Hg, relative to a historical baseline value of 140/70 mm Hg. His pulse rate was 44 bpm. Physical examination showed no abnormal findings and his peripheries were warm. A 12-lead ECG revealed atrial fibrillation.

The differential diagnoses considered were septic shock, cardiogenic shock and hypovolaemic shock. However, the patient’s respiratory symptoms were mild, and he required no supplemental oxygen. Cardiogenic shock was ruled out by transthoracic echocardiogram. Even after his heart rate increased to 70 bpm after atropine administration, his blood pressure showed no improvement. He showed no signs of fluid loss to suggest hypovolemic shock. After other causes were ruled out, the hypotension was attributed to a drug interaction between a calcium-channel blocker and clarithromycin.

1. Discuss the complex drug interactions that contributed to hypotension in this patient.

Case Study 4
A 23 year-old female kidney transplant recipient (3 months post-transplant) presented with cutaneous nodules with fever and pulmonary nodules on CT. The patient was maintained on cyclosporine (300 mg/day), mycophenolate mofetil, and prednisone (5 mg/day) for immunosuppression. At this dose, the cyclosporine plasma trough (i.e. pre-dose) level was approximately 150 ng/mL, which is within the therapeutic target range. Cyclosporine levels are monitored to ensure an appropriate level of immunosuppression, while avoiding exposure-related nephrotoxicity and acute renal failure post-transplant

A work-up resulted in diagnoses of infection with Apergillus and cytomegalovirus (CMV). Treatment was initiated with IV ganciclovir for CMV and voriconazole for Apergillus. At the same time, the dose of cyclosporine was decreased to 50 mg/day.

1. Explain the drug interaction that required the reduction in dose of cyclosporine.

The cyclosporine trough level at the reduced dose was approximately 5 ng/mL, which is well below the therapeutic target. Respiratory symptoms and nodules began to improve over the next several weeks, and the immunosuppressive regimen and voriconazole were continued unchanged for 3 months. During this time, cyclosporine levels were maintained between 3 and 28 ng/mL. Despite the relatively low cyclosporine levels, there was no indication of graft dysfunction or rejection.

In addition to inhibiting CYP3A4, voriconazole is also an inhibitor of P-glycoprotein (P-gp), a transport protein that effluxes drug from inside cells such as T lymphocytes. Cyclosporine is a substrate for P-gp. It has previously been demonstrated that P-gp inhibitors increase cyclosporine concentrations inside T lymphocytes in a dose-dependent manner by decreasing efflux. PBMCs were collected from the patient for in vitro investigation of the effect of voriconazole on cyclosporine cellular uptake, as it was hypothesized increased intracellular concentrations of cyclosporine may explain the discrepancy between low cyclosporine serum levels and the well-maintained immunosuppressive state. The cellular accumulation of cyclosporine increased significantly in the presence of representative P-gp inhibitors in the patient’s PBMCs, demonstrating the feasibility of the in vitro experimental system. In this test system, voriconazole increased cyclosporine intracellular accumulation 3.2-fold, greater than the representative inhibitors.

The patient was genotyped for defective alleles in MDR1, the gene that encodes for P-gp expression.

2. Discuss why the caregivers genotyped the patient for MDR1. What might they have seen that would explain the observations in this patient?

Case Study 1

1. Other than obtaining the level before achieving steady-state, what are other potential explanations for the low EFV exposure? Consider the patient’s concurrent medications.

Phenytoin is a known CYP3A4 inducer.  It acts by increasing DNA transcription, which results in the synthesis of new CYP3A4.  EFV is extensively metabolized by CYP3A4 and CYP2B6.  Therefore, phenytoin may increase in EFV metabolism and causes a decrease in EFV plasma concentration.

2. Explain the likely reasons for the increase in phenytoin plasma levels

The increase in phenytoin plasma concentrations in this case was possibly the result of inhibition of CYP2C9 and CYP2C19 by EFV. In vitro studies have shown that EFV inhibits CYP2C9 and CYP2C19 enzymes, with Ki values in the range of observed EFV plasma concentrations (2.7–5.4 mg/mL). Interpretation of this interaction may be complicated by the concurrent use of sulfadiazine, which has been shown to decrease phenytoin clearance by 45%.

Case study adapted from:  Robertson SM, Penzak SR, Lane J, Pau AK, Mican JM. A potentially significant interaction between efavirenz and phenytoin: a case report and review of the literature. Clin Infect Dis. 2005;41:e15-18.

Case Study 2

1. Discuss the causes of the patient’s symptoms, including the interactions between atorvastatin and verapamil and the effect of the SNP

Atorvastatin is metabolized extensively by CYP3A4 in the liver, and verapamil is a moderate inhibitor of CYP3A4. Initiation of verapamil therapy led to inhibition of CYP3A4 metabolism, which is believed to have contributed to atorvastatin overexposure, thus increasing the risk of rhabdomyolysis.

Atorvastatin is also a substrate for OATP1B1 hepatic uptake transporter.  Therefore, the patient’s SNP in OATP1B1 may have contributed to the adverse event by decreasing the hepatic uptake of atorvastatin, thereby increasing the circulating levels of atorvastatin.

This case illustrates the clinical relevance and relationship between pharmacogenetics and drug-drug interactions in the development of statin-induced myopathy.

Case study adapted from:  Marusic S, Lisicic A, Horvatic I, Bacic-Vrca V, Bozina N. Atorvastatin-related rhabdomyolysis and acute renal failure in a genetically predisposed patient with potential drug-drug interaction.  Int J Clin Pharm.  2012;34:825-827.

Case Study 3

1. Discuss the complex drug interactions that contributed to hypotension in this patient

Calcium-channel blockers, including nifedipine and diltiazem, are metabolized extensively by CYP3A4. Clarithromycin is a strong inhibitor of CYP3A4, and diltiazem and its metabolites are also CYP3A4 inhibitors. Inhibition of CYP3A4 metabolism can cause excessive exposure of calcium-channel blockers, which can result in vasodilatory hypotension.

In addition, the combination of two calcium-channel blockers and a beta-blocker can lower cardiac output due to bradycardia and thus worsen hypotension. In this case, the hypotension presented one day after the patient began using a combination of calcium-channel blockers and clarithromycin. Typically inhibition of CYP3A by clarithromycin and erythromycin does not occur immediately, as these agents are mechanism-based inhibitors. In this case, however, the patient was taking two calcium-channel blockers and carvedilol, which may have contributed to the onset of symptoms.

Benzothiazepine calcium-channel blockers have negative inotropic and chronotropic effects and, in excess, can disturb the cardiac conduction system. β-Blockers prevent the compensating increase in heart rate. Consequently, these medications would effectively worsen hypotension. Since carvedilol is also metabolized by CYP3A4, its plasma concentrations may also have been increased by clarithromycin.  Finally, the patient’s age might be expected to contribute to the severity of the adverse outcome that resulted from this complex drug interaction.

Case study adapted from:  Takeuchi S, Kotani Y, Tsujimoto T.  Hypotension induced by the concomitant use of a calcium-channel blocker and clarithromycin. BMJ Case Rep. Jan 9, 2017

Case Study 4

1. Explain the drug interaction that required the reduction in dose of cyclosporine.

Cyclosporine is a substrate for CYP3A4, while voriconazole is an inhibitor. Therefore, an increase in exposure of cyclosporine is anticipated because of CYP3A4 inhibition by voriconazole.  In addition, the reduction in cyclosporine dose was intended to minimize immunosuppression in the face of a growing infection.

Furthermore, it has been demonstrated that voriconazole increases cyclosporine serum levels in kidney transplant recipients, including approximate 2.5-fold increases in trough concentrations.

2. Discuss why the caregivers genotyped the patient for MDR1.  What might they have seen that would explain the observations in this patient?

If the patient had an allele associated with decreased P-gp expression, this could explain the observation of adequate immunosuppression despite low circulating cyclosporine, as low P-gp expression would result in less efflux from cells and greater intracellular exposure of cyclosporine. However, in this case no alleles associated with decreased P-gp were identified, so the observation is hypothesized to be related to voriconazole-mediated P-gp inhibition. Another confounding factor which could explain the observed immunosuppression despite low circulating cyclosporine levels is the concomitant CMV infection, which enhances the immunosuppressive state.

Case study adapted from:  Park SJ, Song I-S, Kang SW, Joo H, Kim TH, Yoon YC, Kim E, Choi Y-L, Shin J-G, Son JH, Kim YH. Pharmacokinetic effect of voriconazole on cyclosporine in the treatment of aspergillosis after renal transplantation. Clin Nephrol.  2012;78:412-417.

Answer 1: False
While unbound (i.e. “free”) warfarin concentrations may be transiently increased with the addition of a concomitant highly protein bound drug, “free” drug concentrations return to baseline levels relatively quickly due to the elimination of additional unbound drug.

A review of the US Prescribing Informationindicates clinically significant drug interactions with warfarin may result via pharmacodynamic mechanisms (e.g. concomitant drugs that cause bleeding risk, such as antiplatelet drugs) or via pharmacokinetic interactions by induction or inhibition of CYP2C9, CYP1A2, and/or CYP3A4 isoenzymes.

Answer 2: B
This is a 2-way pharmacokinetic interaction, whereby both bosentan and glyburide concentrations are reduced, likely due to induction of CYP3A and CYP2C9 enzymes by each agent.1  Consequently, efficacy may be compromised for both agents during co-administration.

Answer 3: False
Decreased systemic exposure of drug due to a pharmacokinetic interaction typically results in loss of efficacy, not increased risk of toxicity.

The observed pharmacodynamic interaction (increased risk of liver enzyme elevations) cannot be readily explained by the pharmacokinetic interaction in this case.  Alternative hypotheses for the liver enzyme effects include potential inhibition of bile salt export pump (BSEP) by both bosentan and glyburide, which results in increased hepatocyte concentrations of toxic bile salts, or potential formation of a reactive metabolite.3

Answer 4: E (A and C)
As the compound is metabolized primarily by CYP3A and partially by CYP2C19, important DDI studies would include a study with itraconazole (strong CYP3A inhibitor) and fluconazole (strong CYP2C19 inhibitor). Fluconazole is also a moderate inhibitor of CYP3A, and thus would provide a good characterization of the “worst case” effect for CYP2C19 + CYP3A inhibition. Fluconazole is also a widely used antifungal agent, and therefore would provide useful information for a product intended to be used by a wide patient population.

In addition to characterizing the impact of perpetrator drugs on the compound, it is also necessary to characterize the effect of the compound as a potential inhibitor of P-gp. Digoxin is typically used characterize the effect of a potential perpetrator on P-gp.  Digoxin would also be a potential co-medication in a population of patients with diabetes or high blood pressure, and therefore would be an important drug to study with this compound.

A midazolam DDI study is not necessary, as the compound is not a potential inducer or inhibitor of CYP3A.


  1. TRACLEER® (bosentan) US Prescribing Information, 10/2012. Actelion Pharmaceuticals US, Inc. South San Francisco, CA, USA
  2. COUMADIN® (warfarin) US Prescribing Information, 10/2011. Bristol-Myers Squibb Company, Princeton, New Jersey, USA.
  3. van Giersbergen PLM, Treiber A, Clozel M, Bodin F, Dingemanse J. Clin Pharmacol Ther. 2002;71(4):253-62.

Presenter: Sarah Robertson, PharmD, Vertex Pharmaceuticals

Dr. Robertson is Senior Director in Clinical Pharmacology at Vertex Pharmaceuticals, a global biotechnology company. An expert in clinical drug interactions, she worked previously as a Clinical Pharmacology Research Fellow at the National Institutes of Health and a Clinical Pharmacologist at the U.S. Food and Drug Administration, where she was Team Leader for Antiviral Drug Products in the Office of Clinical Pharmacology at the Center for Drug Evaluation and Research (CDER). Dr. Robertson received a B.S. in chemistry from Marquette University, and a Doctorate of Pharmacy from the University of Wisconsin at Madison.