Scolaris Content Display Scolaris Content Display

Cochrane Database of Systematic Reviews Protocol - Intervention

Nivolumab for adult individuals with Hodgkin lymphoma (an exemplar rapid review using RobotReviewer)

Authors' declarations of interest

Version published: 17 February 2017 Version history

https://doi.org/10.1002/14651858.CD012556

This is not the most recent version

view the current version 12 July 2018
Collapse all Expand all

Abstract

This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:

To assess the benefits and harms of nivolumab in adult individuals with Hodgkin lymphoma (HL).

Background

Description of the condition

Hodgkin lymphoma (HL) is a cancer of the lymphatic system, and involves the lymph nodes, spleen and other organs such as the liver, lung, bone or bone marrow, depending on the tumour stage (Lister 1989). HL typically shows a bimodal age distribution with a first peak around the age of 30 years and a second peak after the age of 60 years. It accounts for 10% to 15% of all lymphoma in industrialised countries, with an incidence of 2 to 3 per 100,000 inhabitants. It can therefore be regarded as a relatively rare disease, but is nevertheless one of the most common malignancies in young adults. (Thomas 2002).

The disease usually develops in lymph nodes in the upper part of the body, mostly the latero‐cervical lymph nodes, and results in painless swelling of the lymphatic tissue involved. Normally HL appears within these parts of the body, with peripheral extranodal involvement being rare. As a sign of large tumour size or spreading, 25% of individuals present with B‐symptoms such as fever, drenching night sweats and a loss of more than 10% body weight (Connors 2009; Pileri 2002).

The World Health Organiztion (WHO) Classification of Tumours of Haematopoietic and Lymphoid Tissues distinguishes between two types of HL: classic HL (cHL), which represents about 95% of all HL, and lymphocyte predominant HL, which represents about 5% of all HL (Mathas 2016). Both types differ in morphology, phenotype and molecular features, and therefore in clinical behaviour and presentation (Re 2005).

The Ann Arbor Classification is used for staging and distinguishes between four different tumour stages (Rosenberg 1971). Stages I to III indicate the degree of lymph node and localised extranodal organ involvement or both, and stage four includes disseminated organ involvement, which can be found in 20% of cases. Factors associated with a poor prognosis include a large mediastinal mass, three or more involved lymph node areas, a high erythrocyte sedimentation rate, extranodal lesions, B‐symptoms (weight loss of greater than 10%, fever, drenching night sweats) and advanced age, but the factors considered significant vary slightly between different study groups (German Hodgkin Study Group (GHSG), European Organisation for Research and Treatment of Cancer (EORTC) and the National Cancer Institute of Canada (NCIC)). HL is classified into early favourable, early unfavourable and advanced stage (Engert 2007). In Europe, the early favourable‐stage group usually comprises Ann Arbor stages I and II without risk factors. The early unfavourable‐stage group includes individuals with Ann Arbor stages I or II and one or more risk factors. Most individuals with stages IIB, III or IV disease are included in the advanced‐stage risk group (Engert 2003).

With cure rates of up to 90%, HL is one of the most curable cancers worldwide (Engert 2010; Engert 2012; von Tresckow 2012). A combination of adriamycin, bleomycin, vinblastine and dacarbazine (ABVD) is widely accepted as the gold‐standard chemotherapy regimen in people with HL (Canellos 1992; Engert 2010). Individuals with limited‐stage disease usually receive a combination of chemotherapy and involved‐field radiation (IF‐RT) (Engert 2010; von Tresckow 2012), whereas those with advanced‐stage disease usually receive an intensified regimen, such as BEACOPP (bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine and prednisone) (Borchmann 2011; Engert 2012; Skoetz 2013) or ABVD. A large randomised trial showed that two cycles of ABVD followed by 20 Gy of IF‐RT is sufficient for the treatment of early favourable HL (Engert 2010). Two cycles of escalated BEACOPP (BEACOPPesc) followed by two cycles of ABVD can improve progression‐free survival (PFS) in comparison to four cycles of ABVD in individuals with early unfavourable HL (von Tresckow 2012).

Approximately 10% of people with HL will be refractory to initial treatment or will relapse; this is more common in those people with advanced stage or bulky disease. Standard of care for these individuals is high‐dose chemotherapy and autologous stem cell transplantation (ASCT), but only 55% of patients treated with high‐dose chemotherapy and ASCT have been shown to be free from treatment failure at three years (Rancea 2013). For patients progressing after ASCT, brentuximab vedotin can improve PFS and is the preferred treatment (Younes 2012). However, most patients eventually become refractory to brentuximab vedotin, with limited treatment options.

Description of the intervention

The European Commission has approved nivolumab for the treatment of patients with relapsed/refractory classical HL after autologous stem cell transplant and treatment with brentuximab vedotin. The approval was based on an objective response rate (ORR) of 66% in a combined analysis of 95 patients with relapsed or refractory classic HL who received nivolumab either in the phase II CheckMate‐205 trial or the phase I CheckMate‐039 trial (Ansell 2015; Younes 2016). A recent data report for CheckMate‐205 stated that the median duration of response was prolonged to 13.1 months. The 12‐month PFS was 54.6% and 12‐month overall survival (OS) was 94.9% (Timmermann 2016).

The most common drug‐related adverse events (AEs) included fatigue, infusion‐related reaction, arthralgia and rash. The most common drug‐related grade 3 or 4 AEs were neutropenia and increased lipase concentrations. The most common serious adverse events (SAEs) include fever, pneumonia, tumour progression, arrhythmia, infusion reaction and meningitis (≤ 4% each) (Timmermann 2016; Younes 2016).

Recent data show that nivolumab is now also used in combination with other drugs to treat patients with relapsed or refractory HL (Ansell 2016).

How the intervention might work

Checkpoint inhibitors that target the interaction of the programmed death (PD)‐1 immune checkpoint receptor, and its ligands PD‐L1 and PD‐L2, have shown remarkable activity in a wide range of malignancies. Development started in solid tumours and is most advanced in malignant melanoma and lung cancer (Brahmer 2015; Hamid 2013). In classical HL, malignant Hodgkin Reed‐Sternberg (HRS) cells are dispersed within an extensive inflammatory/immune cell infiltrate (Küppers 2009; Mathas 2016). HRS cells frequently overexpress PD‐L1 and PD‐L2 due to alterations in chromosome 9p24.1 and HL tumours may thus be genetically susceptible to blockade of the PD‐1 pathway (Green 2012; Roemer 2016). Nivolumab is an anti‐(PD)‐1 monoclonal antibody and currently approved by the US Food and Drug Administration (FDA) for the treatment of melanoma, non‐small cell lung cancer, renal cell carcinoma (Matsuki 2016), and, since 2016, for classical HL after treatment with ASCT and brentuximab vedotin.

Why it is important to do this review

To our knowledge, no systematic review on the effectiveness of nivolumab in individuals with HL has been performed to date. As nivolumab is now approved by the European Commission and the FDA based on non‐randomised data, we will critically appraise all published trials and conduct a rapid review on nivolumab. If we identify controlled clinical trials, we will meta‐analyse these data, which will lead to a more precise and reliable evaluation of the benefits and harms of nivolumab. In this way we aim to overcome the limitations of individual studies, such as small sample sizes and a lack of statistical power.

For this review we will use the software RobotReviewer (Marshall 2016; RobotReviewer 2015) to extract study data and assess risk of bias. As this software has not been not validated yet, one review author will extract manually all these data and a second review author will compare the results from the software tool and the first review author. Any discrepancies between the software results and the manually extracted data will be resolved by discussion between both review authors.

Objectives

To assess the benefits and harms of nivolumab in adult individuals with Hodgkin lymphoma (HL).

Methods

Criteria for considering studies for this review

Types of studies

We will include randomised controlled trials (RCTs). In case we do not identify any RCTs, we will include quasi‐RCTs (e.g. assignment to treatment by alternation or by date of birth) and cross‐over trials. In case we do not identify any published RCT, quasi‐RCT or cross‐over trial, we will include published reports of prospectively planned studies.

We will include both full‐text and abstract publications if sufficient information is available on study design, characteristics of participants, interventions and outcomes.

Types of participants

We will include studies that evaluate adult individuals with a confirmed diagnosis of HL, with no gender or ethnicity restrictions. We will consider individuals with all stages, first‐line and relapsed and refractory people and all subtypes of HL. In trials that include mixed populations of individuals with haematological malignancies we will use only data from participants with HL. We will exclude trials in which fewer than 80% of participants had HL, unless the trial authors provide the subgroup data for these individuals in the publication or after we contact the trial authors.

Types of interventions

The main experimental intervention is nivolumab treatment (with or without other drugs). In case we identify RCTs that meet the inclusion criteria of the review, the comparison of interest will be nivolumab (with or without other drugs) versus control treatment. We will conduct separate analyses for trials that evaluate nivolumab and nivolumab combined with other drugs.

Types of outcome measures

Primary outcomes

  • Overall survival (OS)

  • Quality of life (QoL), if measured using reliable and valid instruments

Secondary outcomes

  • Progression‐free survival (PFS)

    • The time interval from random treatment assignment onto the study to first confirmed progression, relapse or death from any cause, or to the last follow‐up

  • Response rate

    • Measured as overall response, complete response and partial response

  • Treatment‐related mortality (TRM)

  • Overall rate of grade 3 and grade 4 adverse events (AEs), including potential relationship between intervention and adverse reaction

  • Overall rate of serious adverse events (SAEs)

Search methods for identification of studies

We will adapt search strategies from the Cochrane Handbook for Systematic Reviews of Interventions (Lefebvre 2011). We will search for studies in all languages in order to limit language bias.

Electronic searches

We will search the following databases and sources. We will start the search in 2000 as PD‐L1 blockade for tumour control, the underlying mechanism of nivolumab, has been first mentioned in 2002 (Iwai 2002)

  • Databases of medical literature

    • Cochrane Central Register of Controlled Trials (CENTRAL; latest issue) (Appendix 1)

    • MEDLINE (Ovid) (2000 to present) (Appendix 2)

    • Embase (2000 to present) (Appendix 3)

    • International Pharmaceutical Abstracts (Appendix 4)

  • Conference proceedings of the annual meetings of the following societies for abstracts (2000 to present, if not included in CENTRAL)

    • American Society of Hematology

    • American Society of Clinical Oncology

    • European Hematology Association

    • International Symposium on Hodgkin Lymphoma

  • Databases of ongoing trials

  • Databases and websites of relevant institutions, such as pharmaceutical organisations, agencies and societies.

Searching other resources

  • Handsearching

    • We will check the reference lists of all identified trials, relevant review articles and current treatment guidelines for further literature.

  • Personal contacts

    • We will contact experts in the field, drug manufacturers and regulatory agencies in order to retrieve information on unpublished trials.

Data collection and analysis

Selection of studies

Two review authors will independently screen the results of the search strategies for eligibility for this review by reading the abstracts using Covidence software (Covidence 2016). We will code the abstracts as either 'retrieve' or 'do not retrieve'. In the case of disagreement or if it unclear whether we should retrieve the abstract or not, we will obtain the full‐text publication for further discussion. Two review authors will assess the full‐text articles of selected studies. If the two review authors are unable to reach a consensus, we will consult a third review author to reach final decision (Higgins 2011b).

We will document the study selection process in a flow chart, as recommended in the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) statement (Moher 2009), and will show the total numbers of retrieved references and the numbers of included and excluded studies. We will list all articles we exclude after full‐text assessment and their reasons for exclusion in a 'Characteristics of excluded studies' table.

Data extraction and management

One review authors will extracted data on the characteristics of included studies and a second review author will compare with the results from the software RobotReviewer (Marshall 2016; RobotReviewer 2015). The two review authors will resolve any discrepancies between the results from the first review author and the software by discussion.

Two review authors will extract data using a standardised data extraction form developed in Covidence (Covidence 2016). If the authors are unable to reach a consensus, we will consult a third review author. If required, we will contact the authors of specific studies for supplementary information (Higgins 2011a).

We will extract the following information.

  • General information: author, title, source, publication date, country, language, duplicate publications

  • Quality assessment: (as specified in the 'Assessment of risk of bias in included studies' section)

  • Study characteristics: trial design, aims, setting and dates, source of participants, inclusion/exclusion criteria, comparability of groups, subgroup analysis, statistical methods, power calculations, treatment cross‐overs, compliance with assigned treatment, length of follow‐up

  • Participant characteristics: age, gender, ethnicity, number of participants recruited/allocated/evaluated, participants lost to follow‐up, additional diagnoses, stage of disease, previous treatment (type of (multi‐agent) chemotherapy (intensity of regimen, number of cycles), field and dose of radiotherapy, autologous stem cell transplantation (ASCT), brentuximab vedotin dosage and duration)

  • Interventions: nivolumab dosage, duration of treatment, duration of follow‐up, for RCTs: comparator (type, dosage)

  • Outcomes: overall survival (OS), QoL, PFS, response rate, TRM, AEs (including assessment of causality, how it was determined, relation between intervention and adverse drug reaction, method of AEs ascertainment (passive or active methods), method of measurement, how severity or seriousness was measured)

Assessment of risk of bias in included studies

Randomised controlled trials

One review author will assess the risk of bias for each RCT. Therafter, we will use the RobotReviewer software to assess risk of bias (RobotReviewer 2015) and a second review author will compare these results with the results from the first review author. Both review authors will resolve any discrepancies between the results from the first review author and the software by discussion. We will use the following criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011a).

  • Sequence generation

  • Allocation concealment

  • Blinding (participants, personnel, outcome assessors)

  • Incomplete outcome data

  • Selective outcome reporting

  • Other sources of bias

For every criterion we will make a judgement using one of three categories.

  • 'Low risk': if the criterion was adequately fulfilled in the study (i.e. the study was at a low risk of bias for the given criterion)

  • 'High risk': if the criterion was not fulfilled in the study (i.e. the study was at high risk of bias for the given criterion)

  • 'Unclear risk': if the study report did not provide sufficient information to allow for a judgement of 'low risk' or 'high risk', or if the risk of bias was unknown for one of the criteria listed above

Non‐randomised prospectively planned trials

As reported in the 'Types of studies' section, we will include non‐randomised studies only if we do not identify any RCTs.

Two review authors will independently assess eligible studies for methodological quality and risk of bias (using the Risk Of Bias in Non‐randomised Studies ‐ of Interventions (ROBIN‐I) tool) (Sterne 2016). The quality assessment strongly depends upon information on the design, conduct and analysis of the trial. The two review authors will resolve any disagreements regarding the quality assessments by consulting a third review author until they reach a consensus.

We will assess the following domains of bias.

  • Bias due to confounding

  • Bias in selection of participants into the study

  • Bias in classification of interventions

  • Bias due to deviations from intended interventions

  • Bias due to missing data

  • Bias in measurement of outcomes

  • Bias in selection of the reported result

For every criterion we will make a judgement using one of five response options.

  • Yes

  • Probably yes

  • Probably no

  • No

  • No information

Measures of treatment effect

We will estimate the dichotomous outcomes of individual studies as rates by extracting the number of events and the total number of participants (overall and complete response rate, TRM, AEs). For RCTs, we will extract dichotomous outcomes from both study arms and will report them as risk ratios (RRs) with 95% confidence intervals (CIs) (Deeks 2011).

We will estimate survival data (OS, PFS) using Kaplan–Meier methods. From RCTs we will extract and report hazard ratios (HRs). If HRs are unavailable, we will estimate the HR by using the available data as described by Parmar 1998 and Tierney 2007.

We will measure continuous outcomes (e.g. QoL) as mean difference (MD) values. For RCTs we will extract and report the mean or mean change from baseline, standard deviation and total number of participants in both the experimental and control arms. If the same scale is used to measure effect, we will perform analyses using the MD with 95% CIs. If the included studies used different scales to measure effect, we will use standardised mean difference values with 95% CIs.

Unit of analysis issues

Studies with multiple treatment groups

As recommended in Chapter 16.5.4 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011b), for studies with multiple treatment groups we will combine arms as long as they can be regarded as subtypes of the same intervention.

When arms can not be pooled this way we will compare each arm with the common comparator separately. For pair wise meta‐analysis, we will split the ‘shared’ group into two or more groups with smaller sample size, and include two or more (reasonably independent) comparisons. For this purpose, for dichotomous outcomes, both the number of events and the total number of patients will be divided up, and for continuous outcomes, the total number of participants will be divided up with unchanged means and standard deviations.

Dealing with missing data

Chapter 16 of the Cochrane Handbook for Systematic Reviews of Interventions suggests a number of potential sources for missing data are suggested (Higgins 2011b), which we will need to taken into account: at study level, at outcome level and at summary data level. In the first instance it is of the utmost importance to differentiate between data 'missing at random' and 'not missing at random'.

If data are missing, we will request these data from the original investigators. If, after this, data are still missing, we will have to make explicit assumptions of any methods the included studies used: for example, we will assume that the data were missing at random or we will assume that missing values had a particular value, such as a poor outcome.

Assessment of heterogeneity

We will assess heterogeneity of treatment effects between trials using a Chi2 test with a significance level at P value < 0.1. We will use the I2 statistic to quantify possible heterogeneity (I2 statistic value > 30% to signify moderate heterogeneity, I2 statistic > 75% to signify considerable heterogeneity) (Deeks 2011). If heterogeneity is above 80%, and we identify a cause for the heterogeneity, we will explore potential causes through sensitivity and subgroup analyses. If we cannot find a reason for heterogeneity, we will not perform a meta‐analysis, but will comment on results from all studies and presented these in tables.

Assessment of reporting biases

In meta‐analyses involving at least 10 trials, we intend to explore potential publication bias by generating a funnel plot and statistically testing this by conducting a linear regression test (Sterne 2011). We will consider a P value of < 0.1 as significant for this test.

Data synthesis

If the clinical and methodological characteristics of individual studies are sufficiently homogeneous, we will pool the data in a meta‐analysis. We will perform analyses according to the recommendations of theCochrane Handbook for Systematic Reviews of Interventions (Deeks 2011). We will use the Review Manager 5 (RevMan 5) software for analyses (Review Manager 2014). One review author will enter the data into the software, and a second review author will check the data for accuracy. As we expect there will be some heterogeneity in trial design, we will use a random‐effects model.

We will not conduct meta‐analyses by including both RCTs and non‐RCTs. We will not meta‐analyse data from uncontrolled trials, as there might be no additional benefit in meta‐analysing data without a control group. In the case of uncontrolled trials, we will report results of each included trial.

In case meta‐analysis is feasible for non‐randomised but controlled trials, we will only analyse outcomes with adjusted effect estimates if these are adjusted for as recommended in the Cochrane Handbook for Systematic Reviews of Interventions (Reeves 2011).

If data do not allow quantitative assessment, we will present outcome data individually per study.

'Summary of findings' table

We will use the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach to assess the quality of the evidence. We will use GRADEpro Guideline Development Tool (GDT) software to create a 'Summary of findings' table (GRADEpro GDT 2014), as suggested in the Cochrane Handbook for Systematic Reviews of Interventions (Schünemann 2011). In addition, we will provide an interactive 'Summary of findings' table for a better user‐experience and for improved dissemination of the findings of this Cochrane Review (Schünemann 2016). We will avoid use of lengthy text.

We prioritise outcomes according to their relevance to patients.

  • OS

  • QoL

  • PFS

  • Response rates

  • TRM

  • AEs

Subgroup analysis and investigation of heterogeneity

We will perform subgroup analyses of the following characteristics.

  • Age

  • Stage of disease (first‐line treatment versus relapsed and refractory disease, early versus advanced stage)

  • Type of previous therapy (ASCT, brentuximab vedotin)

  • Duration of follow‐up.

We will use the tests for interaction to test for differences between subgroup results.

Sensitivity analysis

We will perform only one sensitivity analysis for the following.

  • Quality components

  • Preliminary results versus mature results