Guidelines for the management of chemotherapy-induced emesisare necessary to help clinicians match the emetogenicity of antineoplasticagents with the abundance of antiemetic agents now available. Numerousguidelines for antiemetic therapy currently exist, but compliancewith them is inconsistent, in part because optimal antiemetic protectionis not yet possible, even with the best guidelines. For this reason,guidelines must be dynamic and evolve as knowledge increases.Revision of antiemetic guidelines should be prompted by changes ingeneral principles of treatment, not changes in specific details. Recentrecognition of the unique benefits of incorporating selective neurokinin-1 receptor antagonists into regimens for the prevention of nauseaand vomiting caused by highly emetogenic chemotherapy, particularlyin delayed emesis, justifies modification of existing antiemeticguidelines.
ABSTRACT: Guidelines for the management of chemotherapy-induced emesisare necessary to help clinicians match the emetogenicity of antineoplasticagents with the abundance of antiemetic agents now available. Numerousguidelines for antiemetic therapy currently exist, but compliancewith them is inconsistent, in part because optimal antiemetic protectionis not yet possible, even with the best guidelines. For this reason,guidelines must be dynamic and evolve as knowledge increases.Revision of antiemetic guidelines should be prompted by changes ingeneral principles of treatment, not changes in specific details. Recentrecognition of the unique benefits of incorporating selective neurokinin-1 receptor antagonists into regimens for the prevention of nauseaand vomiting caused by highly emetogenic chemotherapy, particularlyin delayed emesis, justifies modification of existing antiemeticguidelines.
The advent of the selective neurokinin-1 (NK-1) receptor antagonistsas a new therapeuticclass for the treatment of chemotherapy-induced emesis raises the questionof whether, when, and how theseagents should be incorporated intoguidelines for the management ofchemotherapy-induced emesis. Addressingthat question, however, firstrequires consideration of how treatmentguidelines are put together, themultiple functions they can serve, andsome general principles for revisingtreatment guidelines of any kind. Thisarticle explores these issues and concludeswith broad recommendationson the place of NK-1 receptor antagonistsin guidelines for the managementof chemotherapy-inducedemesis.Why Do We Need Guidelines?Guidelines are appropriate and necessaryfor the management of chemotherapy-induced emesis because of theproliferation of chemotherapeuticagents and of corresponding antiemeticagents. Chemotherapeutic drugsmay be classified into at least fourdifferent categories of emetogenicpotential, and at least four differentclasses of antiemetic agents are nowcommercially available (Table 1),yielding multiple potential therapeuticapproaches.Understandably, this has prompteda number of respected organizationsto organize data on the use ofthese agents into logical, useful schemasthat assign appropriate and consistentantiemetic regimens tochemotherapeutic agents of similaremetogenic potential. These organizationsinclude the Multinational Associationof Supportive Care inCancer, the American Society of ClinicalOncology, the American Societyof Health System Pharmacists, andthe National Comprehensive CancerNetwork.[1-4] In fact, so many guidelineson chemotherapy-induced emesishave emerged over the past decadethat a group of experts developed abrief, practical "consensus of consensus"document in 2002 to unify andclarify the available guidelines.Why Do Guidelines Vary?While all of these guidelines are toa certain extent evidence-based, theymay all differ in their conclusions,despite drawing on the same studiesand often sharing many of the sameauthors. These differences can stemfrom differences in the specific chargegiven to each guideline-writing committeeby its parent organization. Eachorganization may have a different goalin developing its guidelines, with correspondingadvantages and disadvantages.However, all of these differentguidelines may be legitimate whenconsidered within the context of thecharge to that particular committee.Examples of typical charges given toguideline-writing committees are outlinedbelow.
'Make the GuidelinesEvidence-Based'
The advantage of this charge is thehigh level of evidence; one can assumethat the recommendations insuch a guidelines document are supportedby level 1 data reflectingconsistent results from multiple randomizedtrials. A disadvantage of thischarge is that such guidelines are likelyto be incomplete. There will be norecommendations in areas where dataare poor or lacking, so that guidelinesof this type often will contain inconinconvenient"gaps."A second disadvantage is that compliancewith such guidelines will oftenbe variable, since evidence-based datausually concentrate heavily on efficacy,often at the expense of commonusage, convenience, cost, or sometimeseven toxicities. For example,most of the strictly evidence-basedguidelines recommend the use ofdexamethasone because of clear evidencethat dexamethasone works.However, these guidelines do not addressthe critical question of whetheror not the toxicities of dexamethasonemay be inconvenient for patientsor worrisome for physicians.'Make the GuidelinesComprehensive'
The advantage of this charge isthat there will be advice for all situations.The disadvantage is that givingadvice for all situations requires acceptinga variable level of evidence.Guidelines developed under such acharge begin with the statement thatthey are based on "the best availableevidence," which may be little morethan personal opinion. This puts theburden on the user of the guidelinesto keep track of which recommendationsare based on level 1 evidence,which are based on level 2 evidence,and which are based solely on expertopinion. This too may lead to complianceproblems, since recommendationsbased on expert opinion maynot always match common practice.'Make the Guidelines Acceptable'
Under this charge, guidelines maybe developed based on evidence supplementedwith expert opinion butthen sent out to clinician end-usersfor input. Guidelines may be modifiedto come closer to common practicebefore they are published.Guidelines developed in this way willbe likely to provide advice for all situationsand should have excellentcompliance since the end-user reviewis essentially a double-check to ensurecompliance before issuance.However, the level of evidence maybe even more variable than under theprevious charge since the feedbackprocess is not necessarily evidencedriven.Nevertheless, all of these methodsof guideline development are legitimateas long as the user understandswhere the guideline document is comingfrom.Suboptimal Compliance WithAntiemetic GuidelinesDespite differences among antiemeticguidelines, their basic recommendationscan be unified into asimple schema, as done by Koellerand colleagues in their 2002 "consensusof consensus" guidelines andas outlined in Table 2. Such a schemais a good starting point for the managementof chemotherapy-inducedemesis.Yet regardless of this significantdegree of consensus among numerousguidelines from respected organizations,compliance with antiemeticguidelines has been far from optimal.This was well illustrated in a studypresented by De Angelis and colleaguesat the 2003 annual meeting ofthe American Society of Clinical Oncology. They found that despitevigorous dissemination of antiemeticguidelines, substantial numbers ofoncologists at 92 oncology centers inItaly either undertreat patients receivinghighly or moderately emetogenicchemotherapy or overtreat patientsreceiving chemotherapy with low orminimal emetogenicity, particularly inthe setting of delayed emesis. Amongpatients receiving cisplatin-based chemotherapy,only 47% received an antiemeticregimen that was in keepingwith the "consensus of consensus"recommendations of Koeller et al fordelayed emesis, and 20% of cisplatintreatedpatients received no antiemeticat all for delayed emesis.Moreover, 30% of patients in thisItalian sample who were undergoingmoderately emetogenic chemotherapyreceived no antiemetic therapy fordelayed emesis, in spite of guidelinerecommendations, whereas 31% to46% of patients undergoing chemotherapywith low or minimal emetogenicitydid receive antiemetic therapyto prevent delayed emesis, despiteconsensus recommendations that notreatment should be given in thissetting.Why Are Guidelines NotFollowed More Closely?Findings like these raise the questionof why antiemetic guidelines arenot followed more consistently. Thereare several contributing factors.First, traditional educational interventionsare generally not effective inaltering physician habits. Mertensand colleagues recently showed thatphysician antiemetic prescribing behavioris unlikely to change withguideline dissemination and didacticsessions alone and that a convincingcorrelation with response among physicians'own patient populations isnecessary.Second, physicians may not followantiemetic guidelines because themagnitude of nausea and vomiting thatpatients experience is not fully appreciated,particularly in the setting ofdelayed nausea and vomiting. Thiswas demonstrated in a recent observationalstudy involving 14 oncologypractices in the United States and Europe. Twenty-four physicians andnurses at these practices were askedto estimate the incidence of nauseaand vomiting among their ownpatients following administration ofhighly or moderately emetogenicchemotherapy, assuming that theywere receiving standard antiemetictherapy for that practice. Then 298patients at these practices were prospectivelysurveyed to determine theactual incidence of nausea and vomitingduring the first 5 days followingchemotherapy.The results showed that the physiciansand nurses accurately estimatedthe incidence of patients' acute nauseaand vomiting but severely underestimatedthe incidence of delayed nauseaand vomiting. Delayed emesis rateswere underestimated by approximately25 to 30 percentage points withhighly emetogenic chemotherapy andapproximately 15 to 30 percentagepoints with moderately emetogenicchemotherapy.Perhaps a more fundamental reasonfor why antiemetic guidelines arenot followed more consistently is thateven the best guidelines do not resultin complete antiemetic protection forall patients, and a significant degreeof innovation and improvement is stillneeded. It should be remembered thatguidelines do not represent an immutablecode of behavior but rather initialtargets for clinicians and centersto aim for. They are a starting point,not the ending point. They should providethe basis for our actions but maynot provide rules for every situation.This viewpoint recognizes guidelinesfor what they fundamentally are: thedynamic results of a process in whichnew knowledge and new agents willlead to regular revisions in an attemptto improve the final outcome of antiemeticcontrol.
When Should GuidelinesBe Revised?The inherently dynamic nature ofguidelines calls for some principlesfor determining whether antiemeticguidelines should be revised at anygiven time (Table 3).When Revision Is Appropriate
Generally, guidelines should bemodified for changes in principles,not changes in details. It certainly isreasonable to revise guidelines whena greater understanding of the physiologyand mechanisms of emesis leadsto reclassification of the antineoplasticagents themselves. Revision alsois appropriate when a new strategyfor antiemetic control is developed. Italso is particularly important to reviseguidelines when a new family of antiemeticagents is introduced that willimpact current antiemetic strategiesand regimens.
When Revision Is Not Appropriate
It is likewise necessary to considersituations in which antiemetic guidelinesshould not be revised, since constantrevision will render any set ofguidelines immediately outdated andtherefore of minimal usefulness. Mi-nor modifications of dosing or administrationare a good example ofsituations in which guideline revisionis not warranted. While tables of suggesteddoses and routes of administrationfor antiemetic agents are useful,such tables are ancillary to the guidelinesthemselves. Modification of suchtables amounts to fine-tuning of antiemeticregimens without altering thegeneral principles that underlie theguidelines.
Most new antineoplastic agents canbe added to existing categories ofemetogenic potential without requiringthe creation of new categories orthe revision of guidelines. Similarly,even the introduction of a new antiemeticagent that is equivalent to previousmembers of the same drug classdoes not generally constitute reasonto update a set of guidelines. In fact,guidelines should, as much as possible,refer to drug classes rather thanto individual agents. If use of oneparticular antiemetic agent within aclass is suggested, the recommendationshould be based on objective evidenceof markedly increased efficacyor decreased toxicity.Finally, issues of "preference"should not dictate the revision of antiemeticguidelines. These include issuessuch as cost, convenience, or anynumber of other factors that cannotbe well supported by hard evidenceand that tend to change easily. Althougheconomic considerations maybe valid factors in choosing a particularagent, their inclusion in a set ofguidelines is not necessary or evenparticularly wise, since they are highlysubject to local patterns of care andreimbursement, the vagaries of drugpricing in a competitive therapeuticarea, and bundling or other health system-specific contracting issues. Similarly,convenience factors can varysubstantially from one clinical settingto another. Because cost, convenience,and other preference factors are subjectto so many potential variables,guidelines that give them too muchweight will inevitably fail to gain wideor enduring acceptance.
Assessing NK-1 Antagonists forGuideline InclusionThe NK-1 receptor antagonists representan interesting advance in antiemeticcare and, in light of theprinciples explored above, may justifymodification of existing antiemeticguidelines on several grounds.First, the NK-1 receptor antagonistsare a new class of antiemeticagents with a unique mechanism ofaction, as discussed in the previousarticle in this supplement.Second, the demonstration thatNK-1 receptor antagonists preferentiallyaffect the control of delayedemesis suggests a change in our understandingof the physiology of emesisand therefore the need to includethese agents in appropriate sectionsof guidelines. This was best illustratedin a study by Cocquyt and colleaguesthat compared monotherapywith an NK-1 receptor antagonist(L-758,298) to monotherapy with theserotonin receptor antagonist ondansetron(Zofran) for prevention ofcisplatin-induced emesis in 53 patients. The NK-1 receptor antagonistprevented delayed emesis ordelayed use of rescue medication insignificantly more patients than didondansetron (48% vs 17%; P = .005),although ondansetron demonstrated anabsolute but nonsignificant advantageover the NK-1 receptor antagonist inprevention of acute emesis or acuteuse of rescue medication. This suggeststhat NK-1 receptor antagonistsresult in a different pattern of efficacyfrom that of serotonin receptor antagonists.That, in turn, underscores a thirdjustification for guideline revision:NK-1 receptor antagonists necessitatenew treatment strategies since theiraction appears to complement that ofserotonin receptor antagonists. Simplesubstitution of one class for anotheris not appropriate, and rationaldouble- and triple-therapy antiemeticregimens must be considered. If oneclass is superior in the delayed settingand another class is superior in theacute setting, yet if both have someactivity in each setting, the logicalapproach is to combine them. Severallarge randomized clinical trials havedone so, combining ondansetron anddexamethasone with either the NK-1receptor antagonist aprepitant or placebo,with encouraging results for thetriplet combination.Two of these trials were in the settingof highly emetogenic cisplatinbasedchemotherapy,[10, 11] and bothfound statistically significant advantagesin complete response rates (noemesis and no use of rescue therapy)during all phases of therapy (acute,delayed, and overall) when ondansetronand dexamethasone werecombined with aprepitant rather thanwith placebo (P < .001). Figure 1 presentsdata from one of the studies,illustrating that the advantage fromthe addition of aprepitant was greaterin the delayed period than in the acuteperiod, a finding also observed in theother study.A third trial compared ondansetron/dexamethasone combined with eitheraprepitant or placebo in the setting ofmoderately emetogenic chemotherapy. It too found advantages incomplete response rates with theaprepitant-containing regimen in theacute and delayed phases of therapy,as well as overall, but the advantageswere less pronounced than in thestudy involving highly emetogenicchemotherapy.Since the long-acting serotoninantagonist palonosetron has been demonstratedto have some activity againstdelayed vomiting after moderatelyemetogenic chemotherapy, use ofa single dose of palonosetron (or multipledoses of a shorter-acting agent)as the serotonin antagonist componentof this three-drug regimen mightbe a reasonable strategy to furtherimprove antiemetic control in the delayedsetting.The Place of NK-1 Antagonistsin Antiemetic GuidelinesAs demonstrated by the above studies,high-level evidence supportinginclusion of NK-1 receptor antagonistsin the management of highlyemetogenic chemotherapy now exists,and evidence for inclusion of theseagents in the management of moderatelyemetogenic chemotherapy isbeginning to emerge. Although questionsof optimal dose and schedulewill continue to be addressed and maydepend substantially on the practicepreferences of individual physicians,it is clear that a new facet of antiemeticcontrol has been identified and thatnew treatment strategies must be devisedand incorporated into guidelinesto take advantage of this new knowledge.The likely implications that ourcurrent knowledge of NK-1 receptorantagonists will have on antiemetictreatment guidelines are depicted inTable 4, which is a suggested modificationof Table 2 from earlier in thisarticle.ConclusionsAntiemetic guidelines are helpfuland necessary for assisting cliniciansin matching the abundance of emetogenicantineoplastic agents with theabundance of antiemetic agents andtreatment strategies now at our disposal.At the same time, guidelinesare dynamic and must evolve asknowledge expands. Revision of antiemeticguidelines should be promptedby changes in principles, notchanges in treatment details. Our newunderstanding of the role of the NK-1pathway is a change in principle thathas justified modification of existingantiemetic guidelines.
Dr. Grunberg hasserved as a consultant for Merck,GlaxoSmithKline, and Helsinn.
Antiemetic Subcommittee of the MultinationalAssociation of Supportive Care inCancer (MASCC): Prevention of chemotherapy-and radiotherapy-induced emesis:Results of Perugia Consensus Conference. AnnOncol 9:811-819, 1998.
American Society of Clinical Oncology:Recommendations for the use of antiemetics:Evidence-based, clinical practice guidelines. JClin Oncol 17:2971-2994, 1999.
ASHP therapeutic guidelines on the pharmacologicmanagement of nausea and vomitingin adult and pediatric patients receivingchemotherapy or radiation therapy or undergoingsurgery. Am J Health Syst Pharm 56:729-764, 1999.
NCCN antiemesis practice guidelines.Oncology (Huntingt) 11:57-89, 1997.
Koeller JM, Aapro MS, Gralla RJ, et al:Antiemetic guidelines: Creating a more practicaltreatment approach. Support Care Cancer10:519-522, 2002.
De Angelis V, Roila F, Sabbatini R, et al,for the Italian Group for Antiemetic Research:Cancer chemotherapy (CT)-induced delayedemesis: Antiemetic prescriptions in clinical practice(abstract 2971). J Clin Oncol 22:739, 2003.
Mertens WC, Higby DJ, Brown D, et al:Improving the care of patients with regard tochemotherapy-induced nausea and emesis: Theeffect of feedback to clinicians on adherenceto antiemetic prescribing guidelines. J ClinOncol 21:1373-1378, 2003.
Grunberg SM, Deuson RR, Mavros P, etal: Incidence of chemotherapy-induced nauseaand emesis after modern antiemetics: Perceptionversus reality. Cancer 100:2261-2268,2004.
Cocquyt V, Van Belle S, Reinhardt RR, etal: Comparison of L-758,298, a prodrug for theselective neurokinin-1 antagonist L-754,030,with ondansetron for the prevention ofcisplatin-induced emesis. Eur J Cancer 37:835-842, 2001.
Poli-Bigelli S, Rodrigues-Pereira J,Carides AD, et al, Aprepitant Protocol 054Study Group: Addition of the neurokinin 1 receptorantagonist aprepitant to standard antiemetictherapy improves control of chemotherapy-induced nausea and vomiting: Resultsfrom a randomized, double-blind, placebo-controlledtrial in Latin America. Cancer 97:3090-3098, 2003.
Hesketh PJ, Grunberg SM Gralla RJ, etal, Aprepitant Protocol 052 Study Group: Theoral neurokinin-1 antagonist aprepitant for theprevention of chemotherapy-induced nauseaand vomiting: A multinational, randomized,double-blind, placebo-controlled trial in patientsreceiving high-dose cisplatin-TheAprepitant Protocol 052 Study Group. J ClinOncol 21:4112-4119, 2003.
Warr DG, Eisenberg P, Hesketh PJ, et al:Effect of aprepitant for the prevention of nauseaand vomiting after one cycle of moderatelyemetogenic chemotherapy: A randomizeddouble-blind trial in 866 patients (abstract8007). Proc Am Soc Clin Oncol 23:2004.
Eisenberg P, Figueroa-Vadillo J, ZamoraR, et al: Improved prevention of moderatelyemetogenic chemotherapy-induced nausea andvomiting with palonosetron, a pharmacologicallynovel 5-HT3 receptor antagonist: Resultsof a phase III, single-dose trial versusdolasetron. Cancer 98:2473-2482, 2003.