Real-world data on herb-drug interactions in oncology: A scoping review of pharmacoepidemiological studies

Abstract

Background

The concurrent use of conventional drugs and herbal medicines is becoming popular among patients with cancer. However, the potential risk of herb-drug interactions (HDI) remains under-addressed in the literature. Previous reviews have mainly focused on the prevalence of interactions, with less attention paid to the methods used by pharmacoepidemiological studies on evaluating HDI. This scoping review aims to summarize the existing pharmacoepidemiological studies that evaluate HDI using real-world data and to identify gaps to be addressed in future research.

Methods

A comprehensive search was performed in nine English- and Chinese-language databases from their inception to May 2021. Gray literature and manual searches were conducted to identify additional studies. The recommended components of the pharmacoepidemiological studies and key findings related to HDI were summarized. The proportion (%) of patients with cancer at risk of HDI was estimated by combining data from eligible studies.

Results

Twenty-eight studies were included in the review. More than half of these studies were cross-sectional studies (n = 18, 64.3%), followed by retrospective cohort studies (n = 5, 17.9%) and prospective cohort studies (n = 2, 7.1%). The three cancer drugs most commonly studied for their interaction potential with herbs were tamoxifen (n = 11, 39.3%), cyclophosphamide (n = 6, 21.4%), and paclitaxel (n = 6, 21.4%). Most cross-sectional studies identified potential HDI using tertiary databases and primary literature searches. Conversely, prospective and retrospective studies mainly investigated actual clinical outcomes, such as adverse events and secondary cancer occurrences. Most interaction outcomes identified using real-world data did not lead to negative clinical consequences. Collectively, 45.4% of herbal medicine users of the included studies were found to be at risk of HDI. We infer from this review that the common limitations of these studies were limited sample size, lack of data on herbal medicine use and details of HDI, and lack of evidence of HDI. Based on the study limitations, several recommendations to enrich the data sources and optimize the study designs were proposed.

Conclusions

There is a high demand for pharmacoepidemiological research on HDI, considering the increasing popularity of herbal medicine among patients with cancer. It is anticipated that emerging real-world data in this field can guide the development of safe and effective approaches to integrative oncology.

Introduction

The proportion of patients with cancer who seek benefits from complementary medicines has increased significantly, from approximately 25% prior to 1990 to an average of 51% from 2009 to 2018 worldwide (Horneber et al., 2012). Herbal medicines are the most commonly used type of complementary medicine (Bautista et al., 2011; Molassiotis et al., 2005). Herbal medicines refer to “herbs, herbal materials, herbal preparations and finished herbal products that contain, as active ingredients, parts of plants, other plant materials or combinations thereof” (WHO, 2019). In some countries, herbal medicines may also contain natural ingredients derived from animals or minerals. A recent systematic review of 155 studies showed that the pooled prevalence of herbal medicine use is 22% among patients with cancer, with the highest prevalence reported in Africa and Asia (Asiimwe et al., 2021).

While patients with cancer take herbs to improve their general health or reduce the side effects of cancer treatment, many use them concomitantly with conventional drugs (Alsanad et al., 2014). However, this usage can cause herb-drug interactions (HDI) that increase or decrease the pharmacological or toxicological effects of either component (Fugh-Berman, 2000). These interactions can lead to both beneficial and harmful health outcomes. For example, the combination of Yunzhi or Lingzhi with cytotoxic drugs may lead to improved survival and the alleviation of chemotherapy-related side effects (Lam et al., 2020). However, for patients with cancer, concerns about negative interactions may be significant, considering the narrow therapeutic windows of oncologic drugs. One example is the use of the cytochrome inducer, St. John's wort, which may lower the plasma concentrations of anticancer drugs, such as docetaxel and irinotecan (Meijerman et al., 2006). A previous review of five studies found that 13.9% of patients with cancer are at risk of herb/food supplement-drug interactions (Alsanad et al., 2014).

Due to methodological and cost considerations, randomized controlled trials (RCTs) often do not have sufficient sample size and follow-up duration in investigating herb-drug interactions. Moreover, the usefulness of RCTs may be particularly limited as they are carried out using selected populations under carefully controlled conditions. Therefore, the results are often less generalizable to a broader population (Garrison Jr et al., 2007). Consequently, pharmacoepidemiological studies are often the principal method used to study drug interactions (Hennessy et al., 2016). Pharmacoepidemiology is “the study of the use of and the effects of drugs in large numbers of people” (Strom, 2019a). These studies aim to collect real-world data, which is commonly defined as data used for decision-making that are not collected in conventional randomized controlled trials (Garrison Jr et al., 2007).

Although regulatory authorities have issued guidelines for methods to evaluate the risk of drug-drug interactions, less attention has been given to HDI (EMA, 2012; FDA, 2020). There is less data on HDI compared with conventional drug-drug interactions (DDIs), and often the data on a specific herb cannot be extrapolated to other herbs (EMA, 2012). A previous review of HDI focused on summarizing the proportion of patients with cancer at risk of HDIs, but it did not provide a critical appraisal of the methods used to study the interactions (Alsanad et al., 2014). The goal of this review was to provide an overview of how existing pharmacoepidemiological research on HDI has been conducted (including study designs, data sources, and interaction analyses); and to identify gaps to be addressed in future research initiatives.