Chemotherapy-Induced Nausea and Vomiting: Time for More Emphasis on Nausea?

Abstract

Despite advances in antiemetic therapy, chemotherapy-induced nausea and vomiting (CINV) remains the most feared and expected side effect of chemotherapy. Optimization of antiemetic therapy is important because CINV can lead to reduced quality of life, increased use of health care resources, and compromised treatment adherence. The evidence illustrates how antiemetic recommendations have evolved and raises ongoing issues and controversies in the management of CINV.

Despite advances in antiemetic therapy, chemotherapy-induced nausea and vomiting (CINV) remains the most feared and expected side effect of chemotherapy [1]. Optimization of antiemetic therapy is important because CINV can lead to reduced quality of life, increased use of health care resources, and compromised treatment adherence. The combination of anthracycline and cyclophosphamide is used extensively in breast cancer patients and traditionally has been classified as being moderately emetogenic [2]. However, in 2011, the American Society of Clinical Oncology reclassified the combination of an anthracycline with cyclophosphamide as highly emetogenic, given its propensity to induce CINV, particularly nausea, in breast cancer patients [3]. Since then, local [4], national [3, 5], and international [6] antiemetic guidelines have emphasized the use of aprepitant in combination with 5-hydroxytryptamine-3 (5-HT3) antagonists and corticosteroids in this population. Not surprisingly, aprepitant-containing regimens have become the standard recommendation at many institutions, and aprepitant use has soared. Has this optimism from both patients and physicians significantly improved patient care?

Despite the significant improvements in CINV reported in the aprepitant trials, control of CINV—and nausea in particular—remains suboptimal in patients receiving anthracycline and cyclophosphamide-based regimens [7, 8]. In the nontrial setting, ∼58%–71% of patients will have nausea and/or vomiting despite “optimal” antiemetic prescribing (C. Hernandez-Torres, personal communication) [9]. What is the cause of this discrepancy? It is likely a combination of factors including health care staff underestimating the incidence of CINV [10], suboptimal adherence to antiemetic guidelines [9], lack of an optimal antiemetic regimen [11], clinical trials overestimating the benefit of antiemetics [7], and, quite simply, current trial endpoints that do not fully reflect patient experience. We feel this is an opportune moment to review the evidence to illustrate how antiemetic recommendations have evolved and to raise ongoing issues and controversies in the management of CINV, especially in the breast cancer population.

As all oncologists are aware, there has been significant progress in CINV control. Historically, antiemetic regimens were developed using cisplatin (>50 mg/m²) as the benchmark because it causes CINV in almost all patients not receiving antiemetic prophylaxis. Based on the proposed pathophysiology of CINV, control of emesis has also been divided into acute (0–24 hours), delayed (>24–120 hours), and overall (0–120 hours) periods following chemotherapy [12], with antiemetic combinations being specifically chosen depending on their efficacy during these different time points. Following the results from antiemetic trials for patients receiving cisplatin-based chemotherapy (Table 1), most current antiemetic guidelines recommend a three-drug regimen consisting of aprepitant (days 1–3), 5-HT3 receptor antagonist (day 1), and dexamethasone (days 1–3 or 4) [3, 4, 6]; however, there are several areas to highlight regarding these trials. Although data for the aprepitant-based regimen from the cisplatin trials has shown significant improvement in control of chemotherapy-induced vomiting and modest improvement in controlling chemotherapy-induced nausea, closer examination of the data strongly suggests that that the same three-drug regimen with aprepitant does not necessarily confer the same benefits for patients receiving anthracycline and cyclophosphamide.

Table 1.

Summary of clinical trials published prior to the use of aprepitant in patients receiving anthracycline and cyclophosphamide

First, based on two multicenter trials in patients receiving cisplatin, use of an aprepitant-based regimen, as described above, improved CINV control during the acute period [13, 14] beyond that seen with regimens composed of 5-HT3 antagonist and dexamethasone only [15, 16]. Second, a pooled analysis of these two aprepitant trials in the cisplatin population showed an improvement of delayed vomiting and delayed nausea, albeit the latter effect being small (48% vs. 42%), in patients receiving cisplatin when added to dexamethasone given on days 2–4, regardless of CINV control in the acute period [17]. This subtlety is important because good control of CINV in the acute period in and of itself leads to better control of CINV in the delayed period [18, 19]. Third, ongoing controversy exists about the optimal regimen for CINV in the delayed period [20–22]. Trials in the late 1990s and early 2000s demonstrated that in the delayed period, for patients receiving cisplatin or anthracycline and cyclophosphamide, the addition of a 5-HT3 antagonist to dexamethasone was not superior to dexamethasone alone [19, 23–25]. In addition, the efficacy of 5-HT3 antagonists as a single agent for the control of delayed CINV is modest at best [20, 26, 27].

Following from the cisplatin trials, the same aprepitant-containing regimen was then evaluated in breast cancer patients receiving anthracyclines and cyclophosphamide. A large double-blind randomized controlled trial showed that an aprepitant-containing regimen provided superior overall complete response (i.e., no vomiting and no use of rescue antiemetics) compared with the control group (Table 2) [7]. Despite earlier data showing that the use of 5-HT3 antagonist in the delayed period is suboptimal, this pivotal trial used ondansetron instead of dexamethasone in the delayed period in the control group [7]. Moreover, although there was a trend showing that aprepitant provided better delayed complete response, the difference was not significant. In addition, contrary to the results from the pooled analysis of the cisplatin trials [17] aprepitant in the anthracycline trials did not confer an additional benefit for the two separate exploratory endpoints of no significant nausea (61% vs. 56%) and no nausea (33% vs. 33%). Subsequent studies in patients receiving anthracycline with cyclophosphamide regimens reaffirmed that aprepitant provided superior overall complete response, but they all used the same, likely suboptimal, control regimen [8, 28, 29]. In these studies, aprepitant also did not demonstrate a benefit for the control of nausea. Furthermore, a large systematic review published by the Agency for Healthcare Research and Quality Technology Assessment Program showed that although a regimen containing aprepitant, a 5-HT3 antagonist, and dexamethasone provided superior complete protection (no vomiting, no rescue, minimal nausea) and complete response in patients receiving cisplatin chemotherapy, there was no difference in complete protection with or without aprepitant in patients receiving moderately emetogenic chemotherapy (MEC), including patients receiving anthracycline and cyclophosphamide. However, there was significant advantage in complete response for patients receiving aprepitant in the MEC population, suggesting again that NK1 antagonists are not effective in controlling nausea in this population [30]. Another review focusing on chemotherapy-induced nausea drew the same conclusion [31]. Despite our hopes that NK1 antagonists will be the solution for CINV, the data do not strongly support the use of these costly drugs in the breast cancer population.

Table 2.

Summary of clinical trials testing aprepitant in breast cancer patients receiving anthracycline and cyclophosphamide

Clearly, there is an unmet need that current antiemetic guidelines must address. Despite these important trials, why is nausea control still such an issue? One reason that trials likely overestimate the benefit of antiemetic regimens is because of the large number of potentially confusing terms used to define CINV endpoints. This confusion arises from the need to measure control of both vomiting and nausea across the acute, delayed, and overall periods. Clinical trials, for example, have used a number of composite endpoints that represent various combinations of vomiting, nausea, both nausea and vomiting, or the use of rescue medications (Table 3). In an attempt to determine the optimal antiemetic regimen to prevent CINV in breast cancer patients receiving anthracycline and cyclophosphamide and to assess whether the randomized trial evidence supports current guideline recommendations of aprepitant use, we performed a systematic review that incorporated network meta-analysis [11]. In the 30 randomized controlled trials that met our criteria, 15 trial endpoints were used (Table 4). These included overall total control (6 trials), overall complete protection (6 trials), and overall complete response (15 trials). This made cross-trial comparisons challenging, and the results suggest that despite the rather firm recommendations of guideline groups, we still do not really know what the optimal antiemetic regimen is for our patients [32].

Table 3.

Definition of chemotherapy-induced nausea and vomiting outcomes used in randomized controlled trials

Table 4.

Summary of reported chemotherapy-induced nausea and vomiting outcomes across randomized controlled trials

Another confounding and more concerning issue is that the outcomes used in clinical trials might not reflect the patients’ actual experience of CINV. The use of composite endpoints is favored in clinical trials because a single dichotomous endpoint could be used in an attempt to summarize the patient’s experience of CINV. Few studies have used nausea as a primary endpoint [21]. Although there is value in using an endpoint such as total control (i.e., no vomiting, no rescue, and no nausea), nausea should also be measured and reported as a separate outcome because it is a subjective, patient-reported symptom, whereas vomiting is an objective and measurable phenomenon. Due to its subjective nature, the most reliable index is to measure “no nausea” rather than “no significant nausea,” as is used in endpoints such as complete protection. Several assessment tools exist to measure nausea as a single endpoint (e.g., visual analog scale [VAS]) or as part of a multiple symptom assessment tool (e.g., Edmonton Symptom Assessment System [ESAS], Functional Living Index-Emesis [FLIE], Functional Assessment of Chronic Illness Therapy [FACIT]). Because we were unable to find any literature that examined which CINV endpoints patients felt were the most important, we performed a survey of 168 patients with early stage breast cancer who had received anthracycline and cyclophosphamide-based chemotherapy to assess their experience of CINV and their perception of different CINV assessment tools (C. Hernandez-Torres, personal communication). Our survey was a 26-item questionnaire that asked patients to rate their nausea using five different scores based on the most common ones used in clinical trials. These included the VAS, the Likert scale, a scale that combined the Likert scale and descriptors of functional capacity, a modified version of the Common Terminology Criteria for Adverse Events 4.0, and a scale that combined the VAS and the Likert scale. The responses across the different scales were concordant most of the time. Patients consistently ranked nausea over vomiting as a “most feared side effect from chemotherapy.” In this survey, despite the use of guideline-recommended antiemetic regimens, 71% experienced nausea and 26% experienced vomiting. Of particular note regarding those composite endpoints that contained “use of rescue medications,” only 57% of those patients with nausea or vomiting took any rescue medication and did so only when the emesis was severe. Consequently, “use of rescue medication” as a surrogate measure for nausea control is inappropriate because it significantly underestimates nausea. Not surprisingly, patients strongly favored a CINV end point that included the absence of both nausea and vomiting.

In conclusion, there has been a significant improvement in CINV over past decades, and that is a testament to the randomized trials that have been performed; however, it is not the time to be complacent. Ultimately, all our patients deserve improved control of CINV. Where will progress come from? CINV remains important, and we need to continue to perform large randomized trials. We need to standardize CINV study endpoints and ensure that all existing studies allow open access to their data so that this process can begin. Standardization of data collection and reporting would allow objective cross-trial comparisons and meta-analyses to better inform treatment guidelines and health policy. We also have to accept that in this era of so-called personalized medicine, a cookie-cutter approach to prescribing antiemetics is not appropriate. We need better tools to stratify a patient’s CINV risk based on validated predictive tools [33, 34] to optimize care (NCT01913990). Clinical trials of more affordable, generic drugs, specifically olanzapine but also gabapentin, have shown some promising data and might offer better control of CINV without the need for more expensive, newer agents [35, 36]. Drugs that were recommended in the older guidelines [15] such as metoclopramide show some benefit compared with 5-HT3 antagonists in the delayed period when given with dexamethasone but have fallen to the wayside because of a low risk of extrapyramidal side effects [37]. The current evidence does not support the use of other agents such as cannabinoids and ginger for prevention or treatment of CINV [31]. Standardizing outcomes to what patients feel is important will improve their quality of life, leading to appropriate trial designs for future regimens and, hopefully, improving the cost-effectiveness of antiemetic treatment.

Author Contributions

Conception/Design: Terry L. Ng, Brian Hutton, Mark Clemons

Provision of study material or patients: Terry L. Ng, Brian Hutton, Mark Clemons

Collection and/or assembly of data: Terry L. Ng, Brian Hutton, Mark Clemons

Data analysis and interpretation: Terry L. Ng, Brian Hutton, Mark Clemons

Manuscript writing: Terry L. Ng, Brian Hutton, Mark Clemons

Final approval of manuscript: Terry L. Ng, Brian Hutton, Mark Clemons

Disclosures

Mark Clemons: Eisai Pharmaceuticals (RF). The other authors indicated no financial relationships.

(C/A) Consulting/advisory relationship; (RF) Research funding; (E) Employment; (ET) Expert testimony; (H) Honoraria received; (OI) Ownership interests; (IP) Intellectual property rights/inventor/patent holder; (SAB) Scientific advisory board