New Developments in the Management of Chemotherapy-Induced Emesis: Do They Impact on Existing Guidelines?

The advent of the selective neurokinin- 1 (NK-1) receptor antagonists as a new therapeutic class for the treatment of chemotherapy- induced emesis raises the question of whether, when, and how these agents should be incorporated into guidelines for the management of chemotherapy-induced emesis. Addressing that question, however, first requires consideration of how treatment guidelines are put together, the multiple functions they can serve, and some general principles for revising treatment guidelines of any kind. This article explores these issues and concludes with broad recommendations on the place of NK-1 receptor antagonists in guidelines for the management of chemotherapy-induced emesis. Why Do We Need Guidelines? Guidelines are appropriate and necessary for the management of chemotherapy- induced emesis because of the proliferation of chemotherapeutic agents and of corresponding antiemetic agents. Chemotherapeutic drugs may be classified into at least four different categories of emetogenicpotential, and at least four different classes of antiemetic agents are now commercially available (Table 1), yielding multiple potential therapeutic approaches. Understandably, this has prompted a number of respected organizations to organize data on the use of these agents into logical, useful schemas that assign appropriate and consistent antiemetic regimens to chemotherapeutic agents of similar emetogenic potential. These organizations include the Multinational Association of Supportive Care in Cancer, the American Society of Clinical Oncology, the American Society of Health System Pharmacists, and the National Comprehensive CancerNetwork.[1-4] In fact, so many guidelines on chemotherapy-induced emesis have emerged over the past decade that a group of experts developed a brief, practical "consensus of consensus" document in 2002 to unify and clarify the available guidelines.[5] Why Do Guidelines Vary? While all of these guidelines are to a certain extent evidence-based, they may all differ in their conclusions, despite drawing on the same studies and often sharing many of the same authors. These differences can stem from differences in the specific charge given to each guideline-writing committee by its parent organization. Eachorganization may have a different goal in developing its guidelines, with corresponding advantages and disadvantages. However, all of these different guidelines may be legitimate when considered within the context of the charge to that particular committee. Examples of typical charges given to guideline-writing committees are outlined below. 'Make the Guidelines Evidence-Based'
The advantage of this charge is the high level of evidence; one can assume that the recommendations in such a guidelines document are supported by level 1 data reflecting consistent results from multiple randomized trials. A disadvantage of this charge is that such guidelines are likely to be incomplete. There will be no recommendations in areas where data are poor or lacking, so that guidelines of this type often will contain inconinconvenient "gaps." A second disadvantage is that compliance with such guidelines will often be variable, since evidence-based data usually concentrate heavily on efficacy, often at the expense of common usage, convenience, cost, or sometimes even toxicities. For example, most of the strictly evidence-based guidelines recommend the use of dexamethasone(Drug information on dexamethasone) because of clear evidence that dexamethasone works. However, these guidelines do not address the critical question of whether or not the toxicities of dexamethasone may be inconvenient for patients or worrisome for physicians. 'Make the Guidelines Comprehensive'
The advantage of this charge is that there will be advice for all situations. The disadvantage is that giving advice for all situations requires accepting a variable level of evidence. Guidelines developed under such a charge begin with the statement that they are based on "the best available evidence," which may be little more than personal opinion. This puts the burden on the user of the guidelines to keep track of which recommendations are based on level 1 evidence, which are based on level 2 evidence, and which are based solely on expert opinion. This too may lead to compliance problems, since recommendations based on expert opinion may not always match common practice. 'Make the Guidelines Acceptable'
Under this charge, guidelines may be developed based on evidence supplemented with expert opinion but then sent out to clinician end-users for input. Guidelines may be modified to come closer to common practice before they are published. Guidelines developed in this way will be likely to provide advice for all situations and should have excellent compliance since the end-user review is essentially a double-check to ensure compliance before issuance. However, the level of evidence may be even more variable than under the previous charge since the feedback process is not necessarily evidencedriven. Nevertheless, all of these methods of guideline development are legitimate as long as the user understands where the guideline document is coming from. Suboptimal Compliance With Antiemetic Guidelines Despite differences among antiemetic guidelines, their basic recommendations can be unified into a simple schema, as done by Koeller and colleagues in their 2002 "consensus of consensus" guidelines[5] and as outlined in Table 2. Such a schema is a good starting point for the management of chemotherapy-induced emesis. Yet regardless of this significant degree of consensus among numerous guidelines from respected organizations, compliance with antiemetic guidelines has been far from optimal. This was well illustrated in a study presented by De Angelis and colleagues at the 2003 annual meeting of the American Society of Clinical Oncology.[ 6] They found that despite vigorous dissemination of antiemetic guidelines, substantial numbers of oncologists at 92 oncology centers in Italy either undertreat patients receiving highly or moderately emetogenic chemotherapy or overtreat patients receiving chemotherapy with low or minimal emetogenicity, particularly in the setting of delayed emesis. Among patients receiving cisplatin(Drug information on cisplatin)-based chemotherapy, only 47% received an antiemetic regimen that was in keeping with the "consensus of consensus" recommendations of Koeller et al for delayed emesis, and 20% of cisplatintreated patients received no antiemetic at all for delayed emesis. Moreover, 30% of patients in this Italian sample who were undergoing moderately emetogenic chemotherapy received no antiemetic therapy for delayed emesis, in spite of guideline recommendations, whereas 31% to 46% of patients undergoing chemotherapy with low or minimal emetogenicity did receive antiemetic therapy to prevent delayed emesis, despite consensus recommendations that no treatment should be given in this setting. Why Are Guidelines Not Followed More Closely? Findings like these raise the question of why antiemetic guidelines are not followed more consistently. There are several contributing factors. First, traditional educational interventions are generally not effective in altering physician habits. Mertens and colleagues recently showed that physician antiemetic prescribing behavior is unlikely to change with guideline dissemination and didactic sessions alone and that a convincing correlation with response among physicians' own patient populations is necessary.[7] Second, physicians may not follow antiemetic guidelines because the magnitude of nausea and vomiting that patients experience is not fully appreciated, particularly in the setting of delayed nausea and vomiting. This was demonstrated in a recent observational study involving 14 oncology practices in the United States and Europe.[ 8] Twenty-four physicians and nurses at these practices were asked to estimate the incidence of nausea and vomiting among their own patients following administration of highly or moderately emetogenic chemotherapy, assuming that they were receiving standard antiemetic therapy for that practice. Then 298 patients at these practices were prospectively surveyed to determine the actual incidence of nausea and vomiting during the first 5 days following chemotherapy. The results showed that the physicians and nurses accurately estimated the incidence of patients' acute nausea and vomiting but severely underestimated the incidence of delayed nausea and vomiting. Delayed emesis rates were underestimated by approximately 25 to 30 percentage points with highly emetogenic chemotherapy and approximately 15 to 30 percentage points with moderately emetogenic chemotherapy. Perhaps a more fundamental reason for why antiemetic guidelines arenot followed more consistently is that even the best guidelines do not result in complete antiemetic protection for all patients, and a significant degree of innovation and improvement is still needed. It should be remembered that guidelines do not represent an immutable code of behavior but rather initial targets for clinicians and centers to aim for. They are a starting point, not the ending point. They should provide the basis for our actions but may not provide rules for every situation. This viewpoint recognizes guidelines for what they fundamentally are: the dynamic results of a process in which new knowledge and new agents will lead to regular revisions in an attempt to improve the final outcome of antiemetic control. When Should Guidelines Be Revised? The inherently dynamic nature of guidelines calls for some principles for determining whether antiemetic guidelines should be revised at any given time (Table 3). When Revision Is Appropriate
Generally, guidelines should be modified for changes in principles, not changes in details. It certainly is reasonable to revise guidelines when a greater understanding of the physiology and mechanisms of emesis leads to reclassification of the antineoplastic agents themselves. Revision also is appropriate when a new strategy for antiemetic control is developed. Italso is particularly important to revise guidelines when a new family of antiemetic agents is introduced that will impact current antiemetic strategies and regimens. When Revision Is Not Appropriate
It is likewise necessary to consider situations in which antiemetic guidelines should not be revised, since constant revision will render any set of guidelines immediately outdated and therefore of minimal usefulness. Mi-nor modifications of dosing or administration are a good example of situations in which guideline revision is not warranted. While tables of suggested doses and routes of administration for antiemetic agents are useful, such tables are ancillary to the guidelines themselves. Modification of such tables amounts to fine-tuning of antiemetic regimens without altering the general principles that underlie the guidelines. Most new antineoplastic agents can be added to existing categories of emetogenic potential without requiring the creation of new categories or the revision of guidelines. Similarly, even the introduction of a new antiemetic agent that is equivalent to previousmembers of the same drug class does not generally constitute reason to update a set of guidelines. In fact, guidelines should, as much as possible, refer to drug classes rather than to individual agents. If use of one particular antiemetic agent within a class is suggested, the recommendation should be based on objective evidence of markedly increased efficacy or decreased toxicity. Finally, issues of "preference" should not dictate the revision of antiemetic guidelines. These include issues such as cost, convenience, or any number of other factors that cannot be well supported by hard evidence and that tend to change easily. Although economic considerations maybe valid factors in choosing a particular agent, their inclusion in a set of guidelines is not necessary or even particularly wise, since they are highly subject to local patterns of care and reimbursement, the vagaries of drug pricing in a competitive therapeutic area, and bundling or other health system- specific contracting issues. Similarly, convenience factors can vary substantially from one clinical setting to another. Because cost, convenience, and other preference factors are subject to so many potential variables, guidelines that give them too much weight will inevitably fail to gain wide or enduring acceptance. Assessing NK-1 Antagonists for Guideline Inclusion The NK-1 receptor antagonists represent an interesting advance in antiemetic care and, in light of the principles explored above, may justify modification of existing antiemetic guidelines on several grounds. First, the NK-1 receptor antagonists are a new class of antiemetic agents with a unique mechanism of action, as discussed in the previous article in this supplement. Second, the demonstration that NK-1 receptor antagonists preferentially affect the control of delayed emesis suggests a change in our understanding of the physiology of emesis and therefore the need to include these agents in appropriate sections of guidelines. This was best illustrated in a study by Cocquyt and colleagues that compared monotherapy with an NK-1 receptor antagonist (L-758,298) to monotherapy with the serotonin receptor antagonist ondansetron(Drug information on ondansetron) (Zofran) for prevention of cisplatin-induced emesis in 53 patients.[ 9] The NK-1 receptor antagonist prevented delayed emesis or delayed use of rescue medication in significantly more patients than did ondansetron (48% vs 17%; P = .005), although ondansetron demonstrated an absolute but nonsignificant advantage over the NK-1 receptor antagonist in prevention of acute emesis or acute use of rescue medication. This suggests that NK-1 receptor antagonists result in a different pattern of efficacyfrom that of serotonin receptor antagonists. That, in turn, underscores a third justification for guideline revision: NK-1 receptor antagonists necessitate new treatment strategies since their action appears to complement that of serotonin receptor antagonists. Simple substitution of one class for another is not appropriate, and rational double- and triple-therapy antiemetic regimens must be considered. If one class is superior in the delayed setting and another class is superior in the acute setting, yet if both have some activity in each setting, the logical approach is to combine them. Several large randomized clinical trials have done so, combining ondansetron and dexamethasone with either the NK-1 receptor antagonist aprepitant or placebo, with encouraging results for the triplet combination. Two of these trials were in the setting of highly emetogenic cisplatinbased chemotherapy,[10, 11] and both found statistically significant advantages in complete response rates (no emesis and no use of rescue therapy) during all phases of therapy (acute, delayed, and overall) when ondansetron and dexamethasone were combined with aprepitant rather than with placebo (P < .001). Figure 1 presents data from one of the studies,[11] illustrating that the advantage from the addition of aprepitant was greater in the delayed period than in the acute period, a finding also observed in the other study.[10] A third trial compared ondansetron/ dexamethasone combined with either aprepitant or placebo in the setting of moderately emetogenic chemotherapy.[ 12] It too found advantages in complete response rates with the aprepitant-containing regimen in the acute and delayed phases of therapy, as well as overall, but the advantages were less pronounced than in the study involving highly emetogenic chemotherapy. Since the long-acting serotonin antagonist palonosetron has been demonstrated to have some activity against delayed vomiting after moderately emetogenic chemotherapy,[13] use of a single dose of palonosetron (or multiple doses of a shorter-acting agent)as the serotonin antagonist component of this three-drug regimen might be a reasonable strategy to further improve antiemetic control in the delayed setting. The Place of NK-1 Antagonists in Antiemetic Guidelines As demonstrated by the above studies, high-level evidence supporting inclusion of NK-1 receptor antagonists in the management of highly emetogenic chemotherapy now exists, and evidence for inclusion of these agents in the management of moderately emetogenic chemotherapy is beginning to emerge. Although questions of optimal dose and schedule will continue to be addressed and may depend substantially on the practice preferences of individual physicians, it is clear that a new facet of antiemetic control has been identified and that new treatment strategies must be devised and incorporated into guidelines to take advantage of this new knowledge. The likely implications that our current knowledge of NK-1 receptor antagonists will have on antiemetic treatment guidelines are depicted in Table 4, which is a suggested modification of Table 2 from earlier in this article. Conclusions Antiemetic guidelines are helpful and necessary for assisting clinicians in matching the abundance of emetogenic antineoplastic agents with the abundance of antiemetic agents and treatment strategies now at our disposal. At the same time, guidelines are dynamic and must evolve as knowledge expands. Revision of antiemetic guidelines should be prompted by changes in principles, not changes in treatment details. Our new understanding of the role of the NK-1 pathway is a change in principle that has justified modification of existing antiemetic guidelines.