Efficacy and Safety of Neurokinin-1 Receptor Antagonists for Prevention of Chemotherapy-Induced Nausea and Vomiting : Systematic Review and Meta-analysis of Randomized Controlled Trials

Chemotherapy-induced nausea and vomiting (CINV) has a significant adverse effect on health-related quality of life and even has negative impacts on the continuation of chemotherapy(Martin et al., 2003a; Martin et al., 2003b). Less toxic chemotherapy is important to retain good performance status and enable further treatment of cancer patients, and prevention of CINV remains the most important issue in supportive therapy of cancer patients. According to the frequency and power of emetic action, chemotherapy agents are categorized into a four-


Introduction
Chemotherapy-induced nausea and vomiting (CINV) has a significant adverse effect on health-related quality of life and even has negative impacts on the continuation of chemotherapy (Martin et al., 2003a;Martin et al., 2003b).Less toxic chemotherapy is important to retain good performance status and enable further treatment of cancer patients, and prevention of CINV remains the most important issue in supportive therapy of cancer patients.
According to the frequency and power of emetic action, chemotherapy agents are categorized into a four-et al., 2005).
Dexamethasone plays a major role in the prevention of acute and delayed CINV and is an integral component of almost all anti-emetic regimens (Grunberg, 2007).And improvement quickly followed with the addition of 5-hydroxytryptamine (5HT3) receptor antagonists (RAs) (Hesketh, 2008;Saito et al., 2009).5-HT3-RAs now form the cornerstone of the therapy for the control of acute emesis with MEC to HEC (Jordan et al., 2014).Navari et al firstly demonstrated that neurokinin-1 receptor antagonists (NK1-RAs) improve CINV when used in patients receiving cisplatin-based chemotherapy (Navari et al., 1999).NK1-RAs are thought to act centrally inhibiting emesis by blocking binding of substance P at the NK1 receptor in the brain stem emetic center (Tattersall et al., 1996).Recent studies and guidelines recommend that the addition of NK1-RAs to the 5-HT3-RAs plus corticosteroid combination as the most effective regimen for controlling both acute and delayed CINV (Aapro et al., 2015).Previous systematic review about the efficacy of NK1-RAs in the prevention of CINV was published in 2012 (dos Santos et al., 2012).dos Santos LV et al demonstrated that NK1-RAs increased CINV control in the acute, delayed, and overall phase, and NK1-RAs are effective for both HEC and MEC.However, further researches about NK1-RAs were still inconsistent.Roila F found that in cancer patients submitted to cisplatin-based chemotherapy, aprepitant plus dexamethasone was not superior to metoclopramide plus dexamethasone in preventing delayed emesis (complete response rate was 80.3% and 82.5%, respectively) (Roila et al., 2015).Meanwhile, Kitayama et al reported that palanosetron and 1-day dexamethasone is almost equivalent to the combination fosaprepitant, granisetron and dexamethasone for MEC (Kitayama et al., 2015).To be followed, there were studies evaluated the efficacy of NK1-RAs in pediatric patients (Kang et al., 2015) and adolescent patients (Gore et al., 2009).Therefore, the aim of this study is to provide an updated systematic review of the efficacy and safety of NK1-RAs in the prevention of CINV, and to evaluate the use of NK1-RAs in pediatric and adolescent patients.

Study searching
Sources such as MEDLINE, EMBASE, the Cochrane Library database, ISI Web of Science were searched (last search, April 30th, 2015).We searched for randomized controlled trials (RCTs) that compared the addition of NK1-RAs to standard antiemetic regimens for cancer patients receiving chemotherapy.We used a combination of the following terms: neurokinin-1 receptor, aprepitant, fosaprepitant, netupitant, casopitant, chemotherapy induced nausea and vomiting.Furthermore, we manually searched the reference sections of the selected studies and relevant reviews for additional publications.
We included human studies written in English, and we did not restrict publication date.When the same patient population was used in several researches, only the most recent, largest or complete study was included.

Inclusion criteria
Two of the authors (YDM, LQ) independently established the eligibility of the studies retrieved from the databases and bibliographies.Trials were included in this meta-analysis if they met the following criteria: (i) published randomized controlled clinical trials with a parallel design comparing NK1-RAs alone or in combination with other antiemetic therapy to antiemetic therapy without NK1-RAs (placebo, dexamethasone, or 5-HT3-RAs); (ii) sufficient data on adequate description of outcomes or toxicity by different treatment; (iii) The studies were prospective RCTs.Disagreement between the two authors were resolved by discussion.

Exclusion criteria
Trials were excluded if they met any of the following criteria: (i) case reports, reviews and conference reports; (ii) studies based on overlapping cohorts from the same institutions.

Data extraction and quality assessment
The name of the first author and year of publication of the article were used for the purpose of identification.The following data were also extracted: study population (country where the study was conducted, Number of patients, type of cancer, type of chemotherapy), methodological characteristics of the RCTs (method of randomization, drop-out description), drugs used in research group and control group, types of NK1-RAs, and most common adverse events.

Definitions of outcomes
In the present research, outcomes were defined as follows: (i) the primary outcome that we extracted was the proportion of patients who achieved CR during the overall period of assessment (acute phase: 0-24h after chemotherapy; delayed phase: 24 hours to several days after chemotherapy).CR was defined as the absence of vomiting and the absence of the need for rescue antiemetic therapy; (ii) the secondary outcomes were nausea, vomiting and need for rescue antiemetic therapy during the overall periods; (iii) safety and tolerability of the antiemetic regimens were also assessed.The most reported adverse events, such as constipation, neutropenia, hiccups, fatigue were also included in the meta-analysis as secondary outcomes.

Subgroup analyses
Predefined subgroup analyses were undertaken in clinically relevant subsets to evaluate the impact of these subgroups on the estimation of the effect size.The following comparisons were carried out: (i) different types of NK1-RAs (aprepitant vs casopitant vs fosaprepitant vs others); (ii) route of administration of NK1-RAs (oral vs intravenous vs both); (iii) age of included patients (adults vs children); (iv) different drugs used in the control arm (placebo vs others); (v) different types of malignancies included in the researches (solid tumor vs hematologic malignancy vs both); (vi) antiemetic regimens in combination with NK1-RAs in the research group (5-HT3-RAs vs 5-HT3-RAs plus dexamethasone).Sensitivity DOI:http://dx.doi.org/10.7314/APJCP.2016.17.4.1661Neurokinin-1 Receptor Antagonists for Chemotherapy-Induced Nausea and Vomiting: Systematic Review analyses based on methodological quality parameters were performed to test for possible variations in estimates of overall OR between subgroups.

Statistical analysis
The Mantel-Haenszel random-effects method was used to calculate odds ratios (ORs) and the corresponding 95% confidence intervals (CIs) (DerSimonian and Laird, 1986).We considered an OR more than 1 favored the NK1-RAs group in the primary endpoints.And an OR more than 1 favored the controlled group in the secondary endpoints and adverse events.
For the test of heterogeneity, we used Higgins I 2 , which measures the percentage of total variation across trials (Higgins and Thompson, 2002).I 2 ranges from 0 (no observed heterogeneity) to 100%.Heterogeneity was considered substantial if I 2 was equal to or more than 50%.When heterogeneity was detected, a possible explanation for it was intensively pursued.If a reasonable cause was found, a separate analysis was then performed.If the cause was not apparent and if heterogeneity was generated by divergent data, the data would not be pooled.
Publication bias was assessed by using Begg's funnel plot and Egger's test.If publication bias existed, the Begg's funnel plot was asymmetric or the P value was less than 0.05 by the Egger's test, the Trim and Fill method was subsequently used (Duval and Tweedie, 2000).
If a given study had more than one interventional arm, we chose to combine all intervention groups together to avoid multiple counting of the same individuals in the control arm (Altman and Bland, 1997).For example, if a trial had more than one research groups with different dose of NK1-RAs and on control arm without NK1-RAs (such as a research performed by Hesketh PJ et al (Hesketh et al., 2014)), the numbers of subjects in different research arms was added and then compared to control arm.
All the statistical tests used in our meta-analysis were performed with STATA version 11.0 software (Stata Corporation College Station, TX).A P-value <0.05 was considered statistically significant.

Literature search and study selection
Figure 1 shows a flow diagram depicting how we identified the relevant clinical trials.By searching four databases and by hand-searching relevant bibliographies, a total of 4722 articles were identified (last search: April 30th, 2015).After screening of title and abstract, 95 studies that potentially met the inclusion criteria were closely scrutinized.57 articles were further excluded for the following reasons: (i) 5 studies was not RCTs; (ii) 22 articles were removed because of duplication; (iii) 13 studies were about drug interaction; (iv)2 articles were economic analyses of aprepitant-containing regimens; (v) 15 studies were about the application of NK1-RAs in diseases other than CINV.

Complete Response
32 RCTs were included in the analysis of CR in the overall phase.In the NK1-RAs groups, 5354 out of 7561 patients (70.8%) had a CR, while only 2966 of 5259 patients (56.0%) achieved CR in the control groups (OR: 1.855, 95%CI: 1.668-2.062),(P<0.001)(Figure 2A).Although no significant heterogeneity was observed in most of the comparisons, we sought to minimize the potential heterogeneity and probe into detailed results in the sub-population by performing a subgroup analysis.Among patients receiving aprepitant, the CR rate was also significantly higher in the treatment group (64.2% vs 51.4%, OR: 1.817, 95%CI: 1.584-2.086),(P<0.001).However, we found that patients treated with fosaprepitant intravenously did not increase the CR rate of CINV (OR: 1.556, 95%CI: 0.662-3.660),(P =0.311).
Data from 33 RCTs and 13385 patients were included in the analysis for the evaluation of CR rate in the delayed phase.There was again a significantly greater frequency of CR among patients given with NK1-RAs (71.4% vs 58.2%, OR: 1.885, 95%CI: 1.671-2.126,P<0.001) (Figure 2C).And details about the results of subgroup analyses were shown in Table 2.

Adverse events
35 out of 38 included studies reported the safety of NK1-RAs.Toxicity was described according to the National Cancer Institute (NCI) Common Terminology Criteria for Adverse Events (CTCAE) in most studies.The adverse events reported in more than three researches were included in the final analysis of toxicity.The mostly reported adverse events were constipation, headache, and neutropenia.We demonstrated that the addition of NK1-RAs did not increase the incidence of patients experienced no less than one adverse event (68.9% vs 65.2%, OR: 1.061, 95%CI: 0.968-1.162,P=0.204).However, we found that constipation (13.9% vs 15.7%, OR: 0.836, 95%CI: 0.755-0.924,P<0.001) and insomnia (3.0% vs 5.0%, OR: 0.493, 95%CI: 0.308-0.789,P=0.003) were more common in the patients of control groups, whereas diarrhea (14.1% vs 11.0%, OR: 1.241, 95%CI: 1.063-1.447,P=0.006) and hiccups (17.8% vs 12.6%, OR: 1.265, 95%CI: 1.064-1.505,P=0.008) were more frequently to be detected in the patients receiving NK1-RAs.And the incidence rate of other adverse events, such as anemia, anorexia, asthenia, dizziness, etc, was not significantly different in two groups.Details about the analyses of toxicity were reported in Table 3.
Among the three researches included children or teenagers, Kang HJ, et al (Kang et al., 2015) and Gore L, et al (Gore et al., 2009) evaluated the incidence of any adverse events.The pooled analysis showed that the difference between NK1-RAs groups and control groups was not significant (79.9% vs 76.8%, OR: 1.193, 95%CI: No.= umber; OR=odds ratio; 95%CI=95% confidence interval; NK1-RAs=neurokinin-1 receptor antagonists.An OR less than 1 favored the NK1-RAs group, whereas an OR more than 1 favored the control group.

Bias analysis
Begg's funnel plot and Egger's test were performed to assess the publication bias of the included trials.The shapes of the funnel plot for the data of CR in acute, delayed and overall phase did not reveal any evidence of obvious asymmetry (Figure 3).Furthermore, Egger's test was used to provide statistical difference (acute CR: P=0.514, delayed CR: P=0.745, overall CR: P=0.593).

Discussion
More and more effective and better tolerated agents have been developed to prevent CINV.With the proper use of antiemetic drugs, CINV can be prevented in almost 70% to 80% of patients receiving chemotherapy.Till now, 5-HT3-RAs, NK1-RAs, and corticosteroids are considered to be the most effective therapeutic combination.In present study, we performed a systematic review and metaanalysis, and tried to figure out the efficacy and tolerability of the addition of NK1-RAs in the prevention of CINV.At the beginning, we demonstrated that patients receiving NK1-RAs had significantly higher CR rate compared to patients without NK1-RAs during the overall, acute, and delayed phase.These results provided evidence that the addition of NK1-RAs brought benefits for patients receiving chemotherapy.According to the positive results about NK1-RAs in the previous clinical researches, NK1-RAs have been added in several guidelines of the treatment of CINV (Herrstedt and Roila, 2009;Basch et al., 2011;Jordan et al., 2014).In the above guidelines, patients receiving HEC or MEC were recommended to be given with the combination of 5-HT3-RAs, NK1-RAs, and dexamethasone during the acute phase.And our research further confirmed the addition of NK1-RAs to other antiemetic regimens in the prevention of CINV.
As the first NK1-RAs approved by the FDA, aprepitant was the mostly often tested agent in our identified trials.And patients given aprepitant showed improvement in all of the interested outcomes, including CR rate, incidence of nausea and vomiting, and need of rescue therapy.However, in the subgroup analysis, we found that although patients treated with fosaprepitant intravenously had the trend to increase the CR rate of CINV, the difference between two groups was not significant.Fosaprepitant is an intravenous formulation of aprepitant that could convert to aprepitant in 30 minutes (Navari, 2007), and an intravenous dose of 115mg is area under the curve bioequivalent to aprepitant 125mg orally (Lasseter et al., 2007).Why the efficacy of fosaprepitant was not similar to aprepitant?One explanation to this question is that there were only two researches (Saito et al., 2013;Kitayama et al., 2015) and 375 patients included for the analysis of fosaprepitant, while Kitayama H, et al (Kitayama et al., 2015) got negative results.Secondly, studies used the combination of intravenous fosaprepitant and oral aprepitant were not included in the analysis of fosparepitant.Whether single intravenous fosaprepitant or combination of fosaprepitant and aprepitant could be an ideal choice for NK1-RAs?We need further clinical trials to solve this problem.
Currently, control of nausea is more difficult than control of vomiting (Grunberg et al., 2004).Previous results of NK1-RAs on the control of nausea were inconsistent.Saito H et al (Saito et al., 2013) did not find significant differences in terms of control of nausea in the overall, acute, and delayed phases.However, a combined analysis of the Poli-Bligelli et al trial (Poli-Bigelli et al., 2003) and Hesketh et al trial (Hesketh et al., 2003) showed a significant decrease of the incidence of nausea (Warr et al., 2005a).Our research got an agreement to this point, and demonstrated that the incidence of nausea and vomiting both significantly decreased after addition of NK1-RAs.
Another conclusion that could be drawn from our study was that NK1-RAs were well tolerated.The most frequently reported adverse events were constipation, headache, and neutropenia.It is quite interesting to find that patients receiving NK1-RAs had more risks of diarrhea and hiccups, while the risk of constipation and insomnia was decreased significantly.Diarrhea is a relatively common adverse effect from cytotoxic antineoplasic treatment and may be debilitating and potentially life threatening and dose limiting (Wadler et al., 1998).Men had a significantly higher incidence of hiccups post-chemotherapy, while women had significantly higher rates of vomiting and nausea (Liaw et al., 2001).We should take these results in consideration before we choose NK1-RAs for the patients receiving chemotherapy.Furthermore, the efficacy of safety of aprepitant have not been fully tested in other disease in which the substance P/NK-1 receptor system is involved (such as cancer, alcoholism, etc), clinical trials are now in progress (Munoz and Covenas, 2013).
A strong point of our study is that we pooled the results of NK1-RAs in children and teenagers.There are more and more researches focusing on the application of NK1-RAs in children.In a retrospective study, aprepitant was given to patients as young as 11 months old (Shillingburg and Biondo, 2014).There were three RCTs evaluated the efficacy and tolerability of NK1-RAs in children and adolescents in the identified studies.The pooled analysis demonstrated NK1-RAs were effective and safe in children and teens receiving chemotherapy, and different types of NK1-RAs might be another choice to prevent CINV in such patients.
Our study of CR rate in acute phase should be concerned with the problem of heterogeneity.There was significant heterogeneity among the 34 valuable studies used to assess the effect of NK1-RAs during acute phase (I 2 =68.2%,P<0.001).Some diversity in the designs of the different studies contributes to the heterogeneity.For example, there were different types of chemotherapy.The choices for therapy greatly influence the incidence of nausea and vomiting.To be continued, the present study also has the typical limitations of the meta-analytical methodology.Our findings and interpretations were limited by the quality and quantity of available evidence on the effects of NK1-RAs on the prevention of CINV, and only published data were used in this study.Publication DOI:http://dx.doi.org/10.7314/APJCP.2016.17.4.1661Neurokinin-1 Receptor Antagonists for Chemotherapy-Induced Nausea and Vomiting: Systematic Review bias is another concern in all forms of meta-analysis.
Still, there are several questions remaining to be solved about the application of NK1-RAs.Firstly, what types of antiemetic drugs should be used together with NK1-RAs?In our research, the antiemetic efficacy seemed to be better when NK1-RAs were used along with 5-HT3-RAs or 5-HT3-RAs plus dexamethasone.To be followed, what is the optimal dose of NK1-RAs?We did not compared different dosage of NK1-RAs in this study.Arpornwirat W et al (Arpornwirat et al., 2009) identified casopitant 150mg as the minimally effective dose.And Roila F et al (Roila et al., 2009) did not found significant difference between different doses of NK1-RAs.These questions have not yet been solved, the role of NK1-RAs for patients receiving antineoplasic therapy is still under active investigation.We look forward to more clinical trials to solve the above two points.
In conclusion, despite some limitations, our metaanalysis suggests that the addition of NK1-RAs could increase the CR rate for the prevention of CINV during acute, delayed, and overall phase.Meanwhile, NK1-RAs could decrease the incidence of nausea, vomiting, and need of rescue therapy.The use of NK1-RAs might be associated with an increased risk of diarrhea and hiccups, and a decreased risk of constipation and insomnia.For children and adolescents patients, NK1-RAs were still effective and well tolerated.

Figure 1 .
Figure 1.Flow chat of Identification and Selection of Studies

Table 1 . Main Characteristics of All Studies Included in the Meta-analysis (continued)
Dong-Mei Yuan et alAsian Pacific Journal of Cancer Prevention,Vol 17, 2016