Drug-Drug Interactions in the Hematology and Oncology Department: A Real-Life Assessment of Frequency and Severity

Drug-drug interactions (DDI) may lead to adverse drug events and constitute a risk for patients and a major challenge for treating physicians. Especially cancer patients receive a considerable number of concomitant drugs for the treatment of the underlying disease, the side effects of cancer therapy and due to preexisting comorbidities. As more and more older and severely compromised patients are treated today, polypharmacy and its related problems are of growing importance. However, data on the frequency and nature of DDI in cancer patients during hospitalization are rare.

To determine the prevalence and nature of potential DDI, we performed a detailed chart review of all hospitalized patients in our hematology and oncology department during two randomly appointed days. All medications administered on these very days were extracted from the electronically archived patients' medical charts. DDI were identified and rated for severity and level of evidence using two different electronic interaction databases, Micromedex 2.0 and Lexi-Interact. Only DDI of at least moderate severity (i.e. rated as moderate, major or contraindicated) were included in the analysis, and documented irrespective of their actual clinical consequences. At the time of analysis, no routinely performed DDI screening or electronic DDI alert system was in place in our department.

Medication profiles from 157 inpatients were analyzed. Patients' median age was 62 years (range: 25–94). Underlying diseases were leukemia (n=51), lymphoma (n=49), solid tumors (n=49) and non-malignant diseases (n=8). We evaluated a total of 1748 drug prescriptions with a median of 11 drugs per patient and day (range: 1–22). The median number of drugs was 9 in patients with solid tumors, 11 in patients with leukemia, and 12 in patients with lymphoma or non-malignant diseases. We found a total of 1153 drug interactions rated as of at least moderate severity by either Micromedex or Lexi-Interact, which corresponds to 734 DDI/100 patient days. The drug classes most frequently involved comprised antibiotics (23%), antifungals (21%), diuretics (20%), corticosteroids (17%), calcium channel blockers (11%), immunosuppressants (10%), benzodiazepines (9%) and opioid analgesics (8%). Antineoplastic drugs constituted 2.7% of all prescribed drugs, and were involved in 47 DDI (4%). The severity rating for all DDI was contraindicated, major and moderate in 2%, 24% and 74% of cases, respectively, according to the respective higher rating of both databases. The median number of DDI per patient was 5 (range: 0–31), and was highest for patients with leukemia with a median number of 7. 85% of patients experienced at least one DDI, and 26% of patients were exposed to ≥10 DDI on the evaluated day. The level of evidence of the 1153 DDI identified was excellent, good, fair and poor in 10%, 45%, 44% and 1% of cases, respectively.

To identify the most clinically relevant DDI, we also determined the frequency and nature of those DDI which were rated as major severity or contraindicated by both databases. 61 DDI fulfilled these criteria and occurred in 33/157 patients (21%). The level of evidence of those 61 DDI was excellent, good and fair in 18%, 28% and 54% of cases, respectively. The most frequently detected potential clinical consequences of the 61 DDI rated as major or contraindicated by both databases were QT-prolongation with a variety of drug combinations (n=36), extrapyramidal reactions of different combinations including metoclopramide (n=5), and cyclosporine toxicity in combination with voriconazole (n=4).

This retrospective DDI analysis revealed a high percentage of patients encountering DDI, including contraindicated drug combinations. Although some of these DDI are difficult to avoid or well manageable through close monitoring of the patients' clinical condition or laboratory results, there is a strong need for rising awareness of DDI to avoid patient harm. This is of particular relevance in the hematology and oncology setting, where numerous drugs are being applied, including cytotoxic agents. By identifying the most relevant DDI in this setting, these results allow targeted physician education and highlight the need to implement an electronic prescription system with DDI alerts in close cooperation with clinical pharmacists.

No relevant conflicts of interest to declare.