Aprepitant in pediatric patients using moderate and highly emetogenic protocols: a systematic review and meta‐analyses of randomized controlled trials

Abstract

Aims

To review the efficacy and safety of aprepitant in combination with ondansetron and dexamethasone (triple therapy) in children and adolescents on moderate to highly emetogenic chemotherapy.

Methods

Medline, Embase, Scielo, Lilacs, Cochrane and congress abstracts published until September 2016 were used as data sources. Two reviewers independently selected manuscripts and extracted data. A third reviewer solved discrepancies in study selection and data extraction. The primary outcome was overall complete response (no vomiting from 0 to 120 h). Secondary outcomes were: response in acute phase, delayed phase and reported toxicities. Each study was considered a unit of analysis. Summarized relative risks were recalculated based on reported data. All meta‐analyses used a random‐effects model and heterogeneity was reported using the I² method.

Results

From 1004 studies, we screened 288 titles and abstracts and included three trials for data extraction. The population comprised 451 patients. Most patients were males, ranging from 6 months to 19 years of age, and weighing from 6 to 134 kg. Bone cancer was the most incident (≥50%) neoplasm, followed by rhabdomyosarcoma and Hodgkin's lymphoma. Triple therapy was associated with a reduced risk of developing chemotherapy‐induced vomiting (CIV) (RR = 0.48; 95% CI 0.34–0.67). There were no differences in incidence of febrile neutropenia between groups (RR = 1.02; 95% CI 0.66–1.58).

Conclusions

Triple therapy decreased CIV risk, without increasing the occurrence of febrile neutropenia. However, this review could not address which subpopulations would most benefit from using this strategy. Future studies should focus on assessing risk factors for nausea and vomiting, as many patients did not achieve a complete antiemetic response.

Tables of Links

TARGETS
GPCRs 2
NK1 receptor
Nuclear hormone receptors 3
Glucocorticoid receptor

LIGANDS
Ondansetron http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=3726
Substance P
Dexamethasone http://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=388
Aprepitant

These Tables list key protein targets and ligands in this article that are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY 1, and are permanently archived in the Concise Guide to PHARMACOLOGY 2015/16 2, 3 .

Introduction

Chemotherapy‐induced nausea and vomiting (CINV) is one of the most impactful events during cancer treatment 4, and accounts for significant losses in overall quality of life and reduced adherence to treatment 4, 5, 6, 7, 8, 9, 10.

CINV was once known as an activation of the vomiting centre 11, 12, 13. Currently, two of the most accepted mechanisms of CINV are: an efferent emetic reflex generated after neurohormonal vagal pathway activation by 5‐hydroxytryptamine (5‐HT₃) and substance p; and a direct antineoplastic stimuli on the postrema area, located in the fourth ventricle's floor 11, 14, 15.

Based on this physiological rationale, the association of three drugs that antagonize the aforementioned biological mediators is considered a strategy to prevent CINV:

• 5‐HT₃ antagonists are known to significantly reduce acute CINV 11, 16, 17, 18, 19, 20;

• Aprepitant, a neurokinin 1 receptor (NK) antagonist, is associated with improved rates of complete response (absence of CINV and need for medication rescues) 11, 16, 17, 18, 19, 20;

• Corticosteroids are the oldest agents used to reduce acute and delayed nausea and vomiting. Although its mechanism of action on CINV prophylaxis is still undetermined, the association with 5‐HT₃ antagonists has demonstrated important CINV risk reduction 11, 17, 18, 19, 20, 21.

The American Society of Clinical Oncology (ASCO) 17, the Pediatric Oncology Group of Ontario (POGO) 18, the National Comprehensive Cancer Network (NCCN) 19 and the European Society of Medical Oncology (ESMO)/Multinational Association of Supportive Care in Cancer (MASCC) 20 endorse the use of triple therapy‐based antiemetics in patients undergoing highly emetogenic chemotherapy.

Despite MASCC/ESMO guidelines 20, the recommendations for using aprepitant in children and adolescents from other scientific societies are still based on the cumulative experiences of adults 22, 23, 24, small observational studies 25, 26, 27, 28 and the first randomized trial 29. In 2015, other randomized controlled trials 30, 31 were published, which in turn, provided us with the opportunity to summarize all reported information into one single review. To the best of our knowledge, there are no systematic reviews with meta‐analyses on aprepitant within the oncology paediatrics specialty.

Besides the need to explore the efficacy of aprepitant in combination with dexamethasone and ondansetron backbone in paediatric patients, safety features regarding aprepitant – a moderate inducer and inhibitor of cytochrome 3A4 – were left open by a previous multicenter trial 29. In that case, Gore and collaborators found an increased number of infections and febrile neutropenia in the triple therapy‐based group, which promoted several discussions throughout the scientific community 32, 33, 34.

The aim of this review is to systematically assess the efficacy and safety of aprepitant in combination with ondansetron and dexamethasone in children and adolescents undergoing moderate to highly emetogenic therapy.

Methods

Protocol registration and rationale of review

The study protocol was registered at PROSPERO, a database of ongoing and finalized systematic reviews from York University, and is available online to the general public 35.

In this review, we specifically sought studies that could fulfil the following research and daily practice problem: ‘what is the efficacy and safety of aprepitant in combination with ondansetron plus dexamethasone in children and adolescents undergoing moderate to highly emetogenic chemotherapy?’.

Data sources and searches

We reviewed five formal international database: MEDLINE, Embase, Cochrane Central Register of Controlled Trials, Lilacs and Scielo. Other relevant database were also reviewed: clinical trials information sources (NIH/clinicaltrials.gov) and abstracts published in Oncology Specialty Congress were searched: (a) Journal of Clinical Oncology (American Society of Clinical Oncology), (b) Annals of Oncology (European Society for Medical Oncology), (c) Supportive Care in Cancer (the journal of the Multinational Association of Supportive Care in Cancer), (d) International Society for Pediatric Oncology and (e) American Society of Pediatric Hematology/Oncology. Searches included published manuscripts until 30 September 2016. There were no date and language restrictions. A summary of our eligibility criteria included:

• Randomized controlled trials design;

• Children and adolescents with cancer undergoing moderate to highly emetogenic chemotherapy, as per POGO 18 and MASCC/ESMO 20 guidelines and the definitions used by the authors in the trial;

• Use of aprepitant in combination with dexamethasone and ondansetron to prevent chemotherapy‐induced vomiting (CIV);

• Comparison with a group that received an ‘active placebo’, defined as double therapy (ondansetron plus dexamethasone). In the present case, not using active placebos is not ethically acceptable.

The complete strategy can be assessed in Table S1). All studies found were downloaded at each reviewer's shared repository (myendnoteweb.org).

Study selection

Two independent reviewers (LMO and FDR) assessed the title and abstract of each study taking into account our predefined eligibility criteria. All abstracts that fulfilled the protocol's inclusion criteria were selected for complete text reading (second screening). After comparing the selected manuscripts, in cases where the two reviewers did not agree, a third reviewer (MAPF) assessed the manuscripts to decide if they should be included for data extraction.

Data extraction

Two independent reviewers (LMO and FDR) extracted in duplicate from all included studies the following information: study design, baseline population, dose of antiemetics, neoplasm, types of antineoplastic protocols, incidence of nausea and vomiting and adverse events. The data was extracted into a standardized data collection sheet. After comparing the data, a third reviewer (MAPF) was invited to resolve any discrepancies.

Outcomes, data synthesis and analysis

The main outcome of this review was overall complete response, defined as absence of CIV from 0 to 120 h. Secondary outcomes included: (a) adverse events reported in the studies (according to each study description); (b) response in acute phase (from 0 to 24 h); (c) delayed response (from 25 to 120 h).

We performed a qualitative assessment of the included manuscripts by describing them in a summary table (Table 1). Moreover, all reports were assessed regarding the risk of bias, as suggested by Cochrane's handbook 36. Thus, we classified each of the six types of bias (selection, performance, detection, attrition, reporting and other types of bias) into three categories: unclear information, low or high risk of bias (Table S2).

Table 1.

Study characteristics

Reference	Population described as intervention (left) versus control (right)						Antiemetic regimen
	Age rangea (years)	Weighta (kg)	Sex (female)	Ethnic background	Neoplasms	Chemotherapy protocol or drugs
Bakhshi et al. 30	5–18	15–65	25 vs. 31%	Black 13 vs. 22% White 44 vs. 44% Hispanic 25 vs. 17% Other 18 vs. 17%	HL 25 vs. 25% Bone 39 vs. 25% Ewing 25 vs. 38% Rhabd. 12 vs. 9%	ABVD 25 vs. 25% VAC 39 vs. 25% VAdC 25 vs. 38% CisDo 41 vs. 25%	All patients received onda (0.15 mg kg−1; max. 16 mg; IV bolus) + dexa (0.15 mg kg−1) as 15‐min IV in 100 ml normal saline 30 min prior to chemotherapy. All patients received oral onda (0.3 mg kg−1 q 8 hourly) and oral dexa (0.15 mg kg−1 q 8 hourly) starting from the day of chemotherapy till 48 h after completion of chemotherapy. Intervention group: Aprepitant capsules 1 h prior to chemotherapy on days 1–3 as follows: 15–40 kg received 80 mg on days 1–3 and 41–65 kg received 125 mg on day 1 followed by 80 mg on days 2–3. Rescue agents: additional onda (max. 32 mg day−1) dexa, metoclopramide.
Gore et al. 29	11–19b	30–105	25 vs. 33%	Not reported	Bone 53 vs. 83% Others not described	Not described	Control group: D1 dexa 16 mg + onda (0.15 mg kg−1 for 3 doses); D2 dexa 8 mg + onda same doses; D3–D4 dexa 8 mg. Intervention group: D1 dexa 8 mg + onda (0.15 mg kg−1 for 3 doses) + aprepitant 125 mg (1 h prior chemotherapy); D2 dexa 4 mg + onda (0.15 mg kg−1 for 3 doses) + aprepitant 80 mg; D3 dexa 4 mg + aprepitant 80 mg; D4 dexa 4 mg. Rescue medication permitted for nausea or vomiting but not described.
Kang et al. 31	0.5–17.8	6–134	45 vs. 47%	Asian 7 vs. 11% White 78 vs. 73% Other 14 vs. 16%	Ewing 11 vs. 11% Bone 11 vs. 11% Neurobl. 9 vs. 7% ALL 9 vs. 5% Rhabd. 8 vs. 9% Medull 6 vs. 8% Nephro 5 vs. 5%	Vinc 43 vs. 49% VP‐16 38 vs. 36% Ifo 30 vs. 32% Doxo 30 vs. 29% Carbo 26 vs. 18% Cis 23 vs. 26% Cy 20 vs. 22% Mtx 15 vs. 15% D‐Act 11 vs. 10%	Control group: D1–D3 of onda 0.18 mg kg−1 in average + dexa 0.2–0.44 mg kg−1 (30% only of patients received dexa in both groups). Intervention group: In addition to onda + dexa from D1 to D3, if 12–17 years old, D1 125 mg aprepitant 1 h prior chemotherapy; D2–D3 80 mg. If 6 months to 12 years old, 3 mg kg−1 on D1 and 2 mg kg−1 on D2 to D3. Rescue medications (5‐HT3 antagonists, phenothiazines, butyrophenones, benzamides, corticosteroids, benzodiazepines, and domperidone) were permitted to alleviate established nausea or vomiting (i.e. not prophylactically).

Notes: Ethnic background, neoplasms and chemotherapies used were intervention (left) vs control group (right).

^aAge and weight expressed as one group.

^bPatients were 15 years old on average.

HL (Hodgkin's Lymphoma); Bone (osteosarcomas other than Ewing's sarcomas); Neurobl. (neuroblastoma); ALL (acute lymphocytic leukaemia); Rhabd. (rabdomiosarcoma); Medull. (medulloblastoma); Nephro (nephroblastoma); onda (ondansetron); dexa (dexamethasone); IV (intravenous); ABVD (doxorubicin, bleomycin, vinblastine and dacarbazine); VAC (vincristine, actinomycin‐D and cyclophosphamide); VAdC (vincristine, doxorubicin and cyclophosphamide); CisDo (cisplatin and doxorubicin); Vinc (vincristine); VP‐16 (etoposide); Ifo (ifosfamide); Doxo (doxorubicin); Carbo (carboplatin); Cis (cisplatin); Cy (cyclophosphamide); Mtx (methotrexate); D‐Act (dactinomycin). Days of chemotherapy: ABVD, VAC, VAdC and CisDo are administered for 1 day.

Outcomes were analysed at the study level. Meta‐analyses were conducted to determine the efficacy and safety of aprepitant in a triple antiemetic therapy regimen. For all estimates, relative risks were recalculated based on the reported number of events and nonevents (Table S4). We used the random‐effects Mantel–Haenszel method to estimate the summarized effect, and relative risks were described with their respective 95% confidence intervals. Heterogeneity was assessed according to the I² method, where values above 30% were considered heterogeneous.

In case of heterogeneity, sensitivity analyses were conducted according to systematic review protocol, by removing one study at a time: (a) the one with the lowest sample size, (b) higher risk of bias, and (c) studies with chemotherapy naïve patients.

All analyses were performed with Revman v.5.3.5 (free Cochrane's Group software) 37.

Results

Of the 1004 studies found, 716 did not have titles compatible with our inclusion criteria. We screened 288 abstracts, from which only four trials were selected for complete text reading. Three studies were included for data extraction and one was excluded because it was a crossover trial with mixed data on adults and adolescents (Figure 1) 38.

Figure 1.

Flowchart describing manuscript selectionASCO (American Society of Clinical Oncologists), MASCC (Multinational Association of Supportive Care in Cancer), ASH (American Society of Hematology), ESMO (European Society for Medical Oncology), ASPHO (American Society of Pediatric Hematology/Oncology); SIOP (The International Society of Paediatric Oncology), RCT(Randomized Controlled Trial), TIAB (title and abstract); * NCT01661335 and NCT02116530

This systematic review comprised 451 patients. Most of the included population were male subjects, ranging from 6 months to 19 years of age, and weighing 6 to 134 kg. Of the three included studies, two were multicenter trials and two were published in 2015. Only Bakhshi et al. and Kang et al. reported their population's ethnic background. Most of patients were white, followed by Hispanics, Blacks and Asians (Table 1).

Efficacy

Overall complete response attributed to triple therapy (aprepitant, ondansetron and dexamethasone) was associated with a 52% relative risk reduction of CIV. Meta‐analysis also showed that heterogeneity was low (Figure 2).

Figure 2.

Overall complete response5HT3 (5‐hydroxytryptamine 3; ondansetron), CTC (corticosteroids; dexamethasone) and NK1 (neurokinin 1, aprepitant)

By analysing secondary outcomes, such as absence of CIV in the first 24 h, two studies suggested that triple therapy was protective. However, by combining the three studies, there were no differences between groups (RR = 0.55, 95% CI 0.30–1.03) (Figure S1, Meta‐analysis A). Notably, heterogeneity was the highest (73%) between all meta‐analyses performed in this review.

Finally, regarding delayed action, meta‐analysis suggested that only Kang et al. contributed positive results of delayed CIV prevention. This result was further confirmed after performing a sensitivity analysis, by removing the Kang et al. study and keeping the other two trials (Figure S1, Meta‐analysis B). Heterogeneity for this analysis was also considered high (53%).

Adverse events

We analysed all adverse events from studies but only febrile neutropenia could be meta‐analysed, due to different drug reactions reported in randomized controlled trials (Figure 3 and Table S2). There were no differences between the groups exposed to triple vs. dual therapy (RR = 1.02, 95% CI 0.66–1.58).

Figure 3.

Risk of febrile neutropenia5HT3 (5‐hydroxytryptamine 3; ondansetron), CTC (corticosteroids; dexamethasone) and NK1 (neurokinin 1, aprepitant)

Overall, the adverse events were heterogeneously reported, as well as the definition of the events observed (Table S2). For example, febrile neutropenia was well defined in two trials, while other observations, such as infection rate, were not completely described. Despite the lack of definitions, it should be noted that most safety outcomes were equally observed between the two comparison groups (with and without aprepitant), as follows:

• Gore et al. reported a difference in the incidence of infection (22% in aprepitant group vs. 9%) and hiccups (five cases in aprepitant group and none in the control group);

• Bakhshi et al. included diarrhoea, constipation, anorexia and fever, but they were equally observed between the groups;

• Kang et al. was the only study to classify the severity of adverse events (grade 1 to 4), especially from the haematologic perspective (leucopoenia, thrombocytopenia and anaemia), but there were no differences between groups.

Primary cancer

Types of cancers were fully described in two studies, while only one reported the antineoplastic protocols. When reported, bone cancers, including Ewing's sarcoma, were presented in 50% of all included patients, followed by rhabdomyosarcomas, Hodgkin's lymphoma and other solid tumours commonly found in early childhood.

Chemotherapy protocols and antiemetic regimens

When protocols were reported, the most common antineoplastics used were: VAC (vincristine, dactinomycin and cyclophosphamide), VAdC (vincristine, adriamycin and cyclophosphamide), cisplatin plus doxorubicin and ABVD (doxorubicin, bleomycin, vinblastine and dacarbazine) (Table 1). Only the Bakhshi et al. study included chemotherapy‐näive patients, while other trials recruited chemotherapy treated and untreated subjects. In summary, the Bakhshi et al. study included solely patients on highly emetogenic chemotherapies, such as dacarbazine, cisplatin, cyclophosphamide plus anthracycline and dactinomycin; whereas Kang et al. had a mixed population on moderate to highly emetogenic agents.

Regarding the types of antiemetic regimens, all studies used aprepitant, ondansetron and dexamethasone from day 1 to day 3 (1 h prior to chemotherapy infusion). However, there were clear differences in dosing regimens of aprepitant and dexamethasone among the clinical trials:

• Bakhshi et al. provided full dexamethasone doses (0.15 mg kg⁻¹ q8h) to both intervention and control groups. Aprepitant dosing was based on weight categories. In this study, there were dosing differences on day 1, given that patients weighing less than 40 kg received 80 mg as first dose and patients >40 kg received adult‐like dosing (125 mg on day 1).

• Kang et al. provided 50% dexamethasone dose reduction in patients receiving aprepitant, due to drug interactions on CYP3A4 level. All adolescents received adult‐like aprepitant doses (125 mg on day 1 and 80 mg on days 2 and 3). Patients younger than 12 years old received 3 mg kg⁻¹ on day 1 and 2 mg kg⁻¹ on subsequent days,

• Gore et al. also provided a 50% reduction in corticosteroid dosing and all patients received aprepitant based on adults' usual regimen.

Risk of bias

Out of the six domains assessed, one study had all domains classified as unclear or high risk of bias, while the other two identified fewer problems (more information is provided in Table S3). The most remarkable biases were: (a) Gore et al. did not explain how they performed the random allocation, and why some patients were excluded or included as open label protocol; (b) Bakhshi et al. showed CIV reduction in the acute phase, but the power of the studys was likely to limit the ‘unequivocal’ nature of the trial's conclusion; and (c) Kang et al. did not provide corticosteroids to all patients.

Sensitivity analysis

As per protocol, we performed four sensitivity analyses by removing the Bakhshi et al. or Gore et al. studies from the acute CIV and delayed CIV meta‐analyses (Figure S2).

By analysing acute CIV prevention, by removing one study at a time, when Gore et al. was taken out, results remained without statistical significance. However, when we removed Bakhshi et al.'s report, the final estimate suggested a 23% relative risk reduction and low heterogeneity (Figure S2, Meta‐analyses A and B).

Regarding delayed CIV prevention, by removing one study at a time, results did not change and the final estimate remained without statistical significance (Figure S2, Meta‐analyses C and D).

Discussion

This systematic review assessed the efficacy and safety of aprepitant in combination with ondansetron and dexamethasone (triple therapy) in children and adolescents on moderate to highly emetogenic chemotherapy. The meta‐analysis of three randomized controlled trials showed a 52% reduction in the relative risk of CIV occurrence.

No additional risk was seen with febrile neutropenia between the groups. Other safety outcomes did not show significant differences in the incidence of adverse events (Table S2). Problems regarding the lack of definitions on safety parameters were seen in all studies. To illustrate, infection was not well documented and there was no information on cultures or other infection‐related data to guide the reader. Most of the reported adverse events could be attributed to the chemotherapy or disease.

When analysing those patients who failed to achieve overall complete antiemetic response (secondary outcomes), meta‐analyses showed that triple therapy was not effective to prevent acute CIV and delayed CIV.

I² values also revealed important results. While low heterogeneity was found in overall response, it was undetected for secondary outcomes (I² > 50%).

Exploring the heterogeneity of the studies on secondary outcomes: delayed CIV

Delayed CIV heterogeneity was 73%, which could be due to several differences in the studies. Firstly, by removing the Gore et al. study, the one with highest risk of bias and lowest sample size, heterogeneity remained high (Figure S2, Meta‐analysis C). However, by removing Bakhshi et al., the study with chemotherapy‐naïve patients, heterogeneity reduced to 30% (Figure S2, Meta‐analysis D).

In both situations, the new summarized relative risks did not evidence a statistically significant reduction in delayed CIV. However, it seemed that patients' characteristics might determine heterogeneity values: studies that included a mixed population, with chemotherapy‐exposed and chemotherapy‐naïve patients (Kang et al. and Gore et al.), might have incorporated a relevant source of confounding.

The Kang et al. study demonstrated a reduction in the incidence of delayed CIV. This positive result might be attributed to the lack of dexamethasone use among the included patients: only 30% were on corticosteroids. It is well known that corticosteroids can prevent acute and delayed CINV, while 5‐HT₃ antagonists are likely to reduce the incidence of acute CINV 16, 21. The use of corticosteroids is a clear recommendation from the main CINV treatment guidelines 17, 18, 19, 20. Results from this trial should be carefully interpreted, as the number of chemotherapy cycles and the protocols were also poorly described.

Exploring the heterogeneity of the studies on secondary outcomes: acute CIV

Meta‐analysis on acute CIV did not reveal a statistically significant reduction in vomiting, and presented high heterogeneity values.

The Gore et al. study had the potential to disturb the final meta‐analysis results on acute CIV because their small sample size might have influenced the size of the 95% confidence interval (Figure S1, Meta‐analysis A). On the other hand, by removing Bakhshi et al., the relative risk changed from a non‐significant to a statistically significant estimate (RR = 0.77, 95% CI 0.64–0.93) (Figure S2, Meta‐analysis A). This was attributed to the higher contribution from the Kang et al. study (weight > 90%).

It is worthwhile to remember that Bakhshi et al. had the lowest relative risk in acute CIV in comparison to the other studies. Such antiemetic protection during the acute phase could be attributed to: (a) the highest dosing strategy of antiemetics (Table 1); and (b) the chemotherapy‐naïve baseline population.

Aprepitant use in the ‘real world’ versus our meta‐analyses

Previously, four observational studies (three retrospective and one prospective cohort) 25, 26, 27, 28 depicted the use of triple therapy without the strict controls from randomized trials. Together, they represented 89 patients, and their results are similar to the present overall CIV response results.

Bauters et al. included 20 previously untreated children and adolescents ranging from 8 to 16 years old, with lymphoma and bone sarcoma 27. They offered a triple therapy‐based antiemetic regimen, with 50% oral dexamethasone reduction (3 mg m⁻²) and 125 mg (day 1) plus 80 mg (days 2 and 3). They found that 50% achieved overall complete response. Six out of ten patients did not experience acute CINV, and seven did not experience delayed CINV. These results are closely related to our meta‐analysis, by fitting our overall complete response confidence interval.

Another study published in 2014 also reported similar results 26. Bodge et al. included the youngest population, in comparison to other observational cohorts 25, 26, 27, 28. Patients were 9 years old on average (range 1–17) and received the lowest aprepitant dosing regimen in the literature for patients weighing less than 10 kg (day 1 = 40 mg and days 2 and 3 = 20 mg). They had a lower proportion of patients achieving overall complete response (7 out of 18, or 39%) in comparison to other observational studies. The results of Bodge et al. were similar to our triple therapy arm, where 80 out of 230 patients (35%) did not experience acute or delayed CIV.

The biggest observational study 25 included 32 patients. In summary, 8 were chemotherapy‐näive, 23 were on highly emetogenic chemotherapy and 9 were on moderate emetogenic chemotherapy. Researchers found that most of the patients (n = 19) did not experience clinically significant nausea and vomiting symptoms after using triple therapy. In this study, subjects were 10 years old on average (ranging from 1 to 18 years of age). The highest dosing strategy of aprepitant was given to patients. In total, 24 children received adult doses, as follows: 125 mg (day 1) and 80 mg (days 2 and 3), plus ondansetron 0.15 mg kg⁻¹ dose⁻¹, plus dexamethasone 7 mg m⁻².

Based on the last two studies, it was observed that the lower the age and antiemetic dosing strategy, the higher is the proportion of patients not achieving complete response. These observations have not yet been addressed in children and adolescent trials, so it could be considered in future studies. Notably, aprepitant has lower area‐under‐the‐curve values in comparison to adults. So, the variation in dosing strategy might be an important source of heterogeneity 29, 30.

One population‐based pharmacokinetic (PK) study assessed the drug interaction between aprepitant and dexamethasone. The results of this study suggest that dexamethasone dose should be adjusted when used together with aprepitant. Despite this strategy being adopted in the Kang et al. study, the clinical benefits (in terms of CIV outcomes) of this dose reduction are still to be determined in children and adolescents 40.

Future directions

There are several differences between the studies on aprepitant and paediatric patients. As a result, it is difficult to identify who the main candidates are to benefit from using a triple antiemetic therapy, and what dosing strategy should be provided.

Until recently, all studies 25, 27, 29, 30, 31 that included bone sarcomas had the greatest overall complete response, ranging from 48% to 66% 30, 31. This observation is reinforced by the following:

• Gore et al. randomized more patients with bone cancers to the control group (83%), while the aprepitant group had almost 20% fewer patients with such primary cancer. This selection bias might have improved their results on overall complete response 29.

• Bone cancers have the worst emetogenic chemotherapy profile, when considering all types of existing protocols. Most include doxorubicin plus cyclophosphamide, high doses of cisplatin, dactinomycin and other highly emetogenic drugs.

• Some studies included acute lymphoblastic leukaemia patients, where it is well known that some chemotherapy courses do not include highly emetogenic drugs. If having a previous CIV is a predictor of other vomiting events, future studies should consider analysing these patients separately, as they might experience less nausea and vomiting per chemotherapy cycle than bone sarcoma patients 31.

Following studies in adults, paediatric research should focus on better selection criteria to reduce heterogeneity. Historically, standardizing antineoplastic therapy as ‘high or low emetogenic treatments’ allowed clinicians to identify drugs or protocols with higher occurrence of CINV 20. However, considering drugs as the only source of risk factor for CINV fails to consider 40, 41 other important clinical issues that are associated with higher occurrence of nausea and vomiting, such as: tumour location (including staging and cerebral metastasis), gastrointestinal obstructions, gastroparesis, opioid‐induced nausea and vomiting, and other features reported elsewhere 11. Neither the current published literature, nor the ongoing trials in the paediatric population explored such variables and their influence on antiemetic response. Despite the lack of controlling risk factors, new molecules with few pharmacokinetic advantages are being tested in adults (e.g. netupitant), and might be tested in the paediatric population in future studies 42, 43.

Limitations

This review is not free of limitations. Three randomized controlled trials with aprepitant as triple therapy were insufficient to draw definite conclusions on the efficacy of aprepitant. Moreover, all included studies were underpowered 40 and had consistent variation among included patients, types of chemotherapy (when reported) and antiemetic strategy. Selecting patients based on their primary cancer, likewise performed in studies in adults 23, might be a differential to identify populations that could most benefit from using aprepitant, and even to explore further ways to prevent CIV. Finally, despite the fact that other 5‐HT₃ would be relevant for clinical practice (e.g. palonosentron), they were not included in the present meta‐analyses because all randomized clinical trials used ondansetron.

Conclusions

According to the meta‐analyses, aprepitant as triple therapy (ondansetron, dexamethasone and aprepitant) reduced the overall risk of experiencing CIV in children and adolescents undergoing moderate to highly emetogenic chemotherapy. Moreover, it was not associated with an augmented risk of developing febrile neutropenia. Despite focusing on the emetogenicity of chemotherapies or adding other newer drugs with few pharmacokinetic advantages, interesting results might rely on determining the characteristics of patients who are at higher risk of developing CINV, which was poorly addressed in all paediatrics studies that used aprepitant in association with a 5‐HT₃ receptor antagonist and corticosteroids.

Competing Interests

There are no competing interests to declare. This work was supported by the Brazilian Ministry of Education, by giving monthly scholarships to LMO as a paediatric pharmacy (PGY4) resident (Residência Integrada Multiprofissonal em Saúde, Hospital de Clínicas de Porto Alegre).

Supporting information

Figure S1 Acute (A) or delayed (B) CIV response 5HT3 (5‐hydroxytryptamine 3; ondansetron), CTC (corticosteroids; dexamethasone), NK1 (neurokinin 1), A (complete response on acute CIV) and B (complete response on delayed CIV)

Figure S2 Sensitivity analysis of heterogeneous meta‐analysis 5HT3 (5‐hydroxytryptamine 3; ondansetron), CTC (corticosteroids; dexamethasone), NK1 (neurokinin 1), A (sensitivity analysis by removing 1 study, complete response on acute CIV), B (sensitivity analysis by removing 1 study, complete response on acute CIV), C (sensitivity analysis by removing 1 study, complete response on delayed CIV) and D (sensitivity analysis by removing 1 study, complete response on delayed CIV)

Table S1 Search strategy

Table S2 Safety outcomes based on adverse events reported in the trials

Table S3 Risk of bias assessment

Table S4 Efficacy data included in meta‐analyses

Supporting info item

Supporting info item

Supporting info item

Supporting info item

Supporting info item

Supporting info item

Okumura, L. M. , D'Athayde Rodrigues, F. , Ferreira, M. A. P. , and Moreira, L. B. (2017) Aprepitant in pediatric patients using moderate and highly emetogenic protocols: a systematic review and meta‐analyses of randomized controlled trials. Br J Clin Pharmacol, 83: 1108–1117. doi: 10.1111/bcp.13193.