Types of studies

Randomized, double‐blind, parallel‐group, placebo (standard therapy with 5‐HT₃ receptor antagonist with or without corticosteroids and/or rescue therapy) comparator‐controlled trials.

Types of participants

Types of interventions

The treatment intervention is the use of a NK‐1 receptor antagonist (aprepitant, fosaprepitant, casopitant, ezlopitant, vestiputant, befetupitant, netupitant, rolapitant, maropitant, vofopitant, dapitant, nolpitantium along with their respective study‐drug identifier, or other NK‐1 receptor antagonists with only a study‐drug identifier; see Appendix 1) combined with the standard antiemetic treatment, as defined elsewhere (Gralla 1999), for acute and delayed nausea and vomiting.

The control intervention is placebo in addition to standard antiemetic therapy (5‐HT₃ receptor antagonist with or without corticosteroid). Studies including participants using rescue antiemetic therapy (e.g. metoclopramide, phenothiazines, butyrophenones and cannabinoids) in addition to treatment or control will be permitted. We will only include studies that randomize participants to the same HEC regimen in both groups.

Types of outcome measures

Electronic searches

We will search the following electronic databases for primary studies.

1. Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library)

2. MEDLINE (from 1948)

3. EMBASE (from 1980)

We will search the electronic databases using a strategy focusing on combining the terms for the participant group (i.e. CINV receiving HEC), the intervention (i.e. drugs classified as NK‐1 receptor antagonists) and the types of studies considered, i.e. randomized controlled trials (RCTs). We will use no language restrictions. We will translate the MEDLINE search strategy (Appendix 2) into the other databases using the appropriate controlled vocabulary as applicable, as determined by the Trials Search Co‐ordinator of the Cochrane Pain, Palliative and Supportive Care (PaPaS) Group. We will include full strategies for other databases in the Appendices of the full review.

Searching other resources

We will search the following other resources:

1. International Clinical Trials Registry Platform (ICTRP);

2. high‐yield journals and conference proceedings (by handsearching) which have not already been handsearched on behalf of The Cochrane Collaboration;

3. reference lists of all included studies and relevant reviews identified;

4. ISI Web of Science for papers which cite studies included in the review.

We will contact authors of included studies regarding any further published or unpublished work.

Selection of studies

One review author (HW) will screen titles and abstracts only from the list of citations identified through the search. We will retrieve citations that meet the inclusion criteria or are considered 'unsure' in full text. Both review authors will independently screen the full‐text articles in detail to determine the final list of included and excluded studies. We will resolve any discrepancies by discussion and consensus of the two review authors.

Data extraction and management

Once all studies have been identified, the two review authors (HW and AT) will examine those studies which have fulfilled the inclusion criteria in detail independently. The two review authors will use a standardized data extraction form independently and compare the extracted data for discrepancies. HW will then be responsible for inputting the data into Review Manager (RevMan) 5 (RevMan 2011) and both review authors will interpret the results.

Assessment of risk of bias in included studies

We will assess the following parameters:

• method and security of randomization;

• whether or not the individuals involved in the study (including healthcare provider, assessor and patient) were blinded to the treatment allocation;

• whether analysis by intention‐to‐treat (ITT) was performed;

• number of participants who completed the study (whether any participants were lost to follow‐up and why).

We will resolve any differences in interpretation of the data through consensus between the two review authors (HW and AT). If additional information is required, we will contact the original authors of the study and we will reassess the study once the missing information has been made available. The two review authors will collect the study characteristics and outcome measures of interest independently using a pre‐formed data extraction sheet. Regardless of compliance or completion of follow‐up, we will collect all data in order to allow for analysis by ITT.

Measures of treatment effect

For evaluation of the primary outcomes, we will record quality of life (QOL) scores (e.g. Functional Living Index ‐ Emesis (FLIE) or Functional Living Index ‐ Cancer (FLIC)), as well as the total number of participants with no emesis and no rescue therapy, no nausea, or greater than or equal to a 22 mm change in VAS for nausea per patient from days one to five of every chemotherapy cycle within each trial. We will calculate quality of life as a per‐patient average for each intervention group (total score of each group divided by the respective number of patients who reported in each group) and by the number of participants experiencing a minimally clinically important difference (MCID) in the QOL score used in the included studies (if the MCID is known for the QOL scales used). We will calculate the mean number of days free of each (emesis and nausea) per participant by dividing the total number days of each by the number of respective participants and present any comparisons between groups as risk ratios (RR) (with corresponding 95% confidence intervals (CI)). We will conduct all analyses using a fixed‐effect model first. If statistical heterogeneity is seen, we will use a random‐effects model to see if the effect remains statistically significant.

Unit of analysis issues

We will use data from all participants individually randomized to each intervention in the analyses. We will take care to identify situations in which data have been censored/excluded. We will contact the authors of the original studies for clarification if necessary (e.g. for participants that withdrew early, how were outcomes recorded or were they simply censored after the point of withdrawal?)

Dealing with missing data

If there are missing data, we will contact the authors of the studies for clarification. If standard deviation of the days free of emesis and nausea cannot be obtained, we will impute the value using standard deviation of the change data from other similar trials.

Assessment of heterogeneity

We will assess heterogeneity across the studies using the I² statistic (we will use a threshold of 30% to 60% to define important heterogeneity) and the Chi² test (with statistical significance being set at P < 0.10). If heterogeneity is detected for outcomes, we will use a random‐effects model to determine if the effects of the study interventions still are statistically significantly different from placebo. We will explore clinical and methodological sources of heterogeneity and characteristics for consideration will include: baseline risk factors for the outcomes of interest, age, gender, ethnicity, types of HEC agents, specific NK‐1 antagonists, differences in use of other antiemetic therapy in addition to study interventions, and duration of treatment distribution of patients across the studies.

Assessment of reporting biases

In the event that missing data are assumed to be due to a poor outcome or are imputed, we will perform sensitivity analyses to see if results are sensitive to the assumptions being made. We will review the potential impact of missing data in the Discussion section of the final report.

Data synthesis

We will use the Cochrane RevMan 5 software for all data analyses. We will base quantitative analyses of outcomes on the intention‐to‐treat (ITT) results. We will use relative risk ratios (RR) and the fixed‐effect model to combine outcomes across studies. We will calculate absolute risk reduction (ARR) (= risk difference (in percentage)) and numbers needed to treat to benefit (NNT) (= 1/risk difference) for all dichotomous outcomes. We will combine data for delayed nausea and emesis reduction using a weighted average method.

Subgroup analysis and investigation of heterogeneity

We will perform the following six subgroup analyses to see if there are differential effects on both acute and delayed emesis, nausea and on QOL or adverse events:

1. effect of age;

2. effect of gender;

3. effect of ethnicity;

4. effect of treatment duration (e.g. three days versus five days);

5. effect of multiple cycles of chemotherapy;

6. effect of different NK‐1 receptor antagonists.

Sensitivity analysis

We will carry out sensitivity analysis in order to test for the robustness of the results. We will undertake analysis of the following categories separately:

1. studies without proper randomization or concealment of allocation compared to those without these characteristics;

2. studies performed without ITT analysis compared to those with an ITT analysis;

3. unblinded versus blinded studies;

4. the effects NK‐1 receptor antagonists that have not been withdrawn from the market compared to the effect of all NK‐1 receptor antagonists;

5. the effects of NK‐1 receptor antagonists with the inclusion of trials where delayed nausea and vomiting standard deviations were not imputed versus trials where standard deviations were imputed or reported.

6. the effects of NK‐1 receptor antagonists in patients in trials receiving only HEC versus patients receiving either HEC or moderately emetogenic chemotherapy.