Retinoic acid postconsolidation therapy for high‐risk neuroblastoma patients treated with autologous haematopoietic stem cell transplantation

Abstract

Background

Neuroblastoma is a rare malignant disease and mainly affects infants and very young children. The tumours mainly develop in the adrenal medullary tissue, with an abdominal mass as the most common presentation. About 50% of patients have metastatic disease at diagnosis. The high‐risk group is characterised by metastasis and other features that increase the risk of an adverse outcome. High‐risk patients have a five‐year event‐free survival of less than 50%. Retinoic acid has been shown to inhibit growth of human neuroblastoma cells and has been considered as a potential candidate for improving the outcome of patients with high‐risk neuroblastoma. This review is an update of a previously published Cochrane Review.

Objectives

To evaluate the efficacy and safety of additional retinoic acid as part of a postconsolidation therapy after high‐dose chemotherapy (HDCT) followed by autologous haematopoietic stem cell transplantation (HSCT), compared to placebo retinoic acid or to no additional retinoic acid in people with high‐risk neuroblastoma (as defined by the International Neuroblastoma Risk Group (INRG) classification system).

Search methods

We searched the Cochrane Central Register of Controlled Trials (CENTRAL) in the Cochrane Library (2016, Issue 11), MEDLINE in PubMed (1946 to 24 November 2016), and Embase in Ovid (1947 to 24 November 2016). Further searches included trial registries (on 22 December 2016), conference proceedings (on 23 March 2017) and reference lists of recent reviews and relevant studies. We did not apply limits by publication year or languages.

Selection criteria

Randomised controlled trials (RCTs) evaluating additional retinoic acid after HDCT followed by HSCT for people with high‐risk neuroblastoma compared to placebo retinoic acid or to no additional retinoic acid. Primary outcomes were overall survival and treatment‐related mortality. Secondary outcomes were progression‐free survival, event‐free survival, early toxicity, late toxicity, and health‐related quality of life.

Data collection and analysis

We used standard methodological procedures expected by Cochrane.

Main results

The update search did not identify any additional studies. We identified one RCT that included people with high‐risk neuroblastoma who received HDCT followed by autologous HSCT (N = 98) after a first random allocation and who received retinoic acid (13‐cis‐retinoic acid; N = 50) or no further therapy (N = 48) after a second random allocation. These 98 participants had no progressive disease after HDCT followed by autologous HSCT. There was no clear evidence of difference between the treatment groups either in overall survival (hazard ratio (HR) 0.87, 95% confidence interval (CI) 0.46 to 1.63; one trial; P = 0.66) or in event‐free survival (HR 0.86, 95% CI 0.50 to 1.49; one trial; P = 0.59). We calculated the HR values using the complete follow‐up period of the trial. The study also reported overall survival estimates at a fixed point in time. At the time point of five years, the survival estimate was reported to be 59% for the retinoic acid group and 41% for the no‐further‐therapy group (P value not reported). We did not identify results for treatment‐related mortality, progression‐free survival, early or late toxicity, or health‐related quality of life. We could not rule out the possible presence of selection bias, performance bias, attrition bias, and other bias. We judged the evidence to be of low quality for overall survival and event‐free survival, downgraded because of study limitations and imprecision.

Authors' conclusions

We identified one RCT that evaluated additional retinoic acid as part of a postconsolidation therapy after HDCT followed by autologous HSCT versus no further therapy in people with high‐risk neuroblastoma. There was no clear evidence of a difference in overall survival and event‐free survival between the treatment alternatives. This could be the result of low power. Information on other outcomes was not available. This trial was performed in the 1990s, since when many changes in treatment and risk classification have occurred. Based on the currently available evidence, we are therefore uncertain about the effects of retinoic acid in people with high‐risk neuroblastoma. More research is needed for a definitive conclusion.

Plain language summary

Retinoic acid after intensive chemotherapy and bone marrow transplantation in people with high‐risk neuroblastoma

Review question

We reviewed the evidence about the effect of adding retinoic acid as part of a postconsolidation therapy after intensive chemotherapy (high‐dose chemotherapy) followed by autologous (from the same person) bone marrow transplantation (haematopoietic stem cell transplantation) in people with high‐risk neuroblastoma. A consolidation therapy tries to destroy possible remnant cancer cells after a preceding therapy has achieved the elimination of detectable tumour. A postconsolidation therapy is applied after that consolidation therapy. The addition of retinoic acid was compared to the same pretreatment but placebo (inactive) retinoic acid or no addition of retinoic acid for two primary and five secondary outcomes. The primary outcomes were overall survival (participants who did not die) and treatment‐related mortality (participants who died due to complications of the intervention). Secondary outcomes were progression‐free survival (the condition did not worsen), event‐free survival (staying free of any of a particular group of events), early and late toxicity (harmful effects), and health‐related quality of life.

Background

Neuroblastoma is a rare cancerous disease and mainly affects infants and very young children. The high‐risk group is prone to spread of the disease and other characteristics that increase the risk of an adverse event. Retinoic acid stops uncontrolled cell growth in laboratory cell cultures and might reduce the return of the tumour after high‐dose chemotherapy followed by autologous haematopoietic stem cell transplantation in people with high‐risk neuroblastoma.

Study characteristics

The evidence is current to 24 November 2016. We included a single randomised trial with 50 people allocated to the addition of retinoic acid after high‐dose chemotherapy followed by autologous haematopoietic stem cell transplantation, and 48 people allocated to the same treatment but without the addition of retinoic acid.

Key results

The update search did not identify any new studies. Overall survival and event‐free survival were no different between the two treatment alternatives. Other outcomes, including those concerning adverse events, were not adequately reported. Additional retinoic acid as part of a postconsolidation therapy after high‐dose chemotherapy followed by autologous haematopoietic stem cell transplantation may not improve survival in people with high‐risk neuroblastoma, and we lack information on its safety. More research is needed before we can draw solid conclusions.

Quality of the evidence

The evidence is based on a single study. The quality of the evidence in this single included study is low.

Summary of findings

Summary of findings for the main comparison. Retinoic acid postconsolidation therapy compared to no further treatment for high‐risk neuroblastoma patients treated with autologous haematopoietic stem cell transplantation.

Retinoic acid postconsolidation therapy compared to no further treatment for high‐risk neuroblastoma patients treated with autologous HSCT
Patient or population: high‐risk neuroblastoma patients treated with autologous HSCT Settings: paediatric oncology departments Intervention: retinoic acid postconsolidation therapy Comparison: no further treatment
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	No further treatment	Retinoic acid post‐consolidation therapy
Overall survival (reported as mortality)	583 per 10001	533 per 1000 (331 to 760)	HR 0.87 (0.46 to 1.63)	98 (1 study)	⊕⊕⊝⊝ low2,3	The length of follow‐up was not mentioned for the 98 participants eligible for this review
Treatment‐related mortality ‐ not reported	See comment	See comment	Not estimable	‐	See comment	No adequate information on this outcome was provided
Progression‐free survival ‐ not reported	See comment	See comment	Not estimable	‐	See comment	No information on this outcome was provided
Event‐free survival (reported as relapse, disease progression, death from any cause, or second neoplasm)	604 per 10001	549 per 1000 (371 to 749)	HR 0.86 (0.50 to 1.49)	98 (1 study)	⊕⊕⊝⊝ low2,3	The length of follow‐up was not mentioned for the 98 participants eligible for this review
Early toxicity ‐ not reported	See comment	See comment	Not estimable	‐	See comment	No adequate information on this outcome was provided
Late toxicity including secondary malignancies ‐ not reported	See comment	See comment	Not estimable	‐	See comment	No adequate information on this outcome was provided
Health‐related quality of life ‐ not reported	See comment	See comment	Not estimable	‐	See comment	No information on this outcome was provided
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; HR: Hazard ratio; HSCT: haematopoietic stem cell transplantation
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.

¹The assumed risk is based on the number of events in the control group at the final time point of the survival curve presented in the included study.
 ²The presence of selection bias, performance bias, attrition bias and other bias was unclear, so we downgraded one level for study limitations.
 ³Since this is a small study and the 95% CIs include values favouring both the intervention and the control treatment we downgraded one level for imprecision.

Background

Description of the condition

Neuroblastoma is a rare malignant disease and mainly affects infants and very young children (GARD 2017). Tumours develop in the sympathetic nervous system such as adrenal medullary tissue or paraspinal ganglia, and may be localised or metastatic at diagnosis (Cole 2012). The median age at diagnosis is 17 months and the incidence rate of neuroblastoma is age‐dependent, with an incidence rate of 64 per million children in the first year of life reducing to 29 per million children in the second year of life (Goodman 2012). The incidence rate in adults is less than one per million a year, but adults have a considerably worse prognosis (Esiashvili 2007). Cohn 2009 proposed the International Neuroblastoma Risk Group (INRG) classification system (see Table 2). Of 8800 people with neuroblastoma, 36.1% had high‐risk neuroblastoma. Berthold 2005 estimated that 50% of people with neuroblastoma have metastatic disease at diagnosis. Patients in the high‐risk group had a five‐year event‐free survival rate (defined as the time from diagnosis until the time of first occurrence of relapse, progression, secondary malignancy, or death, or until the time of last contact if none of these occurred) of less than 50%. Matthay 2012 addressed new approaches with targeted therapy that may improve the outcome in people with high‐risk neuroblastoma.

1. The INRG consensus pretreatment classification scheme.

INRG stage	Age (months)	Histologic category	Grade of tumour differentiation	MYCN	11q aberration	Ploidy	Pretreatment risk group
							Code	Interpretation
L1/L2	–	Ganglioneuroma maturing; ganglioneuroblastoma intermixed	‐	–	–	–	A	Very low
L1	–	Any, except ganglioneuroma or ganglioneuroblastoma	–	Not amplified	–	–	B	Very low
				Amplified	–	–	K	High
L2	< 18	Any, except ganglioneuroma or ganglioneuroblastoma	–	Not amplified	No	–	D	Low
					Yes	–	G	Intermediate
	≥ 18	Ganglioneuroblastoma nodular; neuroblastoma	Differentiating	Not amplified	No	–	E	Low
					Yes	–	H	Intermediate
			Poorly differentiated or undifferentiated	Not amplified	–	–	H	Intermediate
			–	Amplified	–	–	N	High
M	< 18	–	–	Not amplified	–	Hyperdiploid	F	Low
	< 12	–	–	Not amplified	–	Diploid	I	Intermediate
	12 to < 18	–	–	Not amplified	–	Diploid	J	Intermediate
	< 18	–	–	Amplified	–	–	O	High
	≥ 18	–	–	–	–	–	P	High
MS	< 18	–	–	Not amplified	No	–	C	Very low
					Yes	–	Q	High
				Amplified	–	–	R	High

Reference: Cohn 2009.

The INRG consensus classification schema includes the criteria INRG stage, age, histologic category, grade of tumour differentiation, MYCN status, presence/absence of 11q aberrations, and tumour cell ploidy. Sixteen statistically or clinically different pretreatment groups of patients (lettered A through R), or both, were identified using these criteria. The categories were designated as very low (A, B, C), low (D, E, F), intermediate (G, H, I, J), or high (K, N, O, P, Q, R) pretreatment risk subsets.

An abdominal mass is the most common presentation of neuroblastoma. In general, neuroblastoma occurs at a single location, usually the medulla of the adrenal gland or along the paravertebral sympathetic chain. Organ‐specific symptoms may be caused by the local presence of metastases, such as eye problems associated with retrobulbar tumours, pancytopenia associated with bone marrow infiltration, abdominal distension and respiratory problems associated with liver enlargement, paralysis, and Horner syndrome associated with ganglion involvement (Berthold 2005; NCI PDQ 2017). Furthermore, there are general signs and symptoms like tiredness, weakness or pain. Some neuroblastomas regress spontaneously without therapy, while others progress with a fatal outcome despite therapy. One study of infants younger than 12 months showed that nearly half of the study population within three years of follow‐up had a spontaneous regression at diagnosis (Hero 2008). A tumour mass may be confirmed by ultrasound, X‐rays, computed tomography, or magnetic resonance imaging. Guidelines for using imaging methods have been developed in response to the increased importance of image‐defined factors in staging and risk assessment (Brisse 2011).

The International Neuroblastoma Staging System provides the current definitions for diagnosis, the stages 1, 2A, 2B, 3, 4, and 4S shown in Table 3, and treatment response shown in Table 4 (Brodeur 1993). The INRG classification system provides the current definitions for the very low, low, intermediate, and high‐risk groups shown in Table 2 (Cohn 2009).

2. The International Neuroblastoma Staging System.

Stage	Definition
1	Localised tumour with complete gross excision, with or without microscopic residual disease; representative ipsilateral lymph nodes negative for tumour microscopically (nodes attached to and removed with the primary tumour may be positive)
2A	Localised tumour with incomplete gross excision; representative ipsilateral nonadherent lymph nodes negative for tumour microscopically
2B	Localised tumour with or without complete gross excision, with ipsilateral nonadherent lymph nodes positive for tumour. Enlarged contralateral lymph nodes must be negative microscopically
3	Unresectable unilateral tumour infiltrating across the midlinea, with or without regional lymph node involvement; or localised unilateral tumour with contralateral regional lymph node involvement; or midline tumour with bilateral extension by infiltration (unresectable) or by lymph node involvement
4	Any primary tumour with dissemination to distant lymph nodes, bone, bone marrow, liver, skin and/or other organs (except as defined for stage 4S)
4S	Localised primary tumour (as defined for stage 1, 2A or 2B), with dissemination limited to skin, liver, and/or bone marrowb (limited to infants < 1 year of age)

Reference: Brodeur 1993.

Note: Multifocal primary tumours leg, bilateral adrenal primary tumours should be staged according to the greatest extent of disease, as defined above, and followed by a subscript letter M e.g. 3_M.
 ^aThe midline is defined as the vertebral column. Tumours originating on one side and crossing the midline must infiltrate to or beyond the opposite side of the vertebral column.
 ^bMarrow involvement in stage 4S should be minimal, i.e. < 10% of total nucleated cells identified as malignant on bone marrow biopsy or on marrow aspirate. More extensive marrow involvement would be considered to be stage 4. The (Meta‐iodobenzylguanidine) MIBG scan (if performed) should be negative in the marrow.

3. Response to treatment.

Response	Primary tumour	Metastatic sites
Complete response	No tumour	No tumour; catecholamines normal
Very good partial response	Decreased by 90% to 99%	No tumour; catecholamines normal; residual 99Tc bone changes allowed
Partial response	Decreased by more than 50%	All measurable sites decreased by > 50%. Bones and bone marrow: number of positive bone sites decreased by > 50%; no more than 1 positive bone marrow site allowed
Minimal response	No new lesions; > 50% reduction of any measurable lesion (primary or metastases) with < 50% reduction in any other; < 25% increase in any existing lesion
No response	No new lesions; < 50% reduction but < 25% increase in any existing lesion
Progressive disease	Any new lesion; increase of any measurable lesion by > 25%; previous negative marrow positive for tumour

Reference: Brodeur 1993.

Description of the intervention

Retinoic acid is a derivative of vitamin A (retinol) that includes 13‐cis retinoic acid, also known as isotretinoin, among others. Retinoic acid regulates the growth and development of epithelial cells, and inhibits growth of human neuroblastoma cells (Sidell 1982). It reduces morphological signs characteristic of several malignant human neuroblastoma cell lines (Sidell 1983). In a phase I clinical trial, 13‐cis retinoic acid was used in children with neuroblastoma after autologous haematopoietic stem cell transplantation (HSCT) without signs of myelosuppression (Villablanca 1995). In a phase III clinical trial, 13‐cis retinoic acid was added in people with high‐risk neuroblastoma after receiving high‐dose chemotherapy (HDCT) followed by autologous HSCT as well as after receiving standard‐dose chemotherapy (SDCT) (Matthay 1999). The test intervention of this review is the addition of retinoic acid as part of a therapy that comes after the consolidation therapy, which constitutes HDCT followed by autologous HSCT. Yalçin 2015 included three randomised controlled trials (RCTs) in a Cochrane Review including the study Matthay 1999. The objective of the review was to compare the efficacy of HDCT followed by autologous HSCT with standard‐dose chemotherapy in children with high‐risk neuroblastoma.

How the intervention might work

Retinoic acid induces the differentiation of human neuroblastoma cell lines and stops uncontrolled cell growth in vitro (Reynolds 2003). Thus, retinoic acid might reduce the relapse rate, which occurs frequently after intensive chemotherapy with or without autologous HSCT in people with high‐risk neuroblastoma.

Why it is important to do this review

People with high‐risk neuroblastoma might have improved survival if retinoic acid as part of a postconsolidation therapy were added after HDCT followed by autologous HSCT. However, it is possible that a considerable number of patients may not respond to the addition of retinoic acid. It is therefore important to evaluate the current evidence base for the efficacy and the possible adverse events associated with this treatment. This review is an update of an earlier published Cochrane Review (Peinemann 2015).

Objectives

To evaluate the efficacy and adverse events of additional retinoic acid as part of a postconsolidation therapy after high‐dose chemotherapy (HDCT) followed by autologous haematopoietc stem cell transplantation (HSCT), compared to the same pretreatment with placebo retinoic acid or to no additional retinoic acid in people with high‐risk neuroblastoma (as defined by the International Neuroblastoma Risk Group (INRG) classification system).

Methods

Criteria for considering studies for this review

Types of studies

Randomised controlled trials (RCTs).

Types of participants

People with high‐risk neuroblastoma according to the INRG or the Children's Oncology Group classification of risk groups shown in Table 2 and Table 5.

4. Children's Oncology Group assignment to low, intermediate, and high‐risk group.

INSS stage	Age	MYCN	INPC classification	DNA index	Risk group
1	0 to 21 y	Any	Any	Any	Low
2A/2B	< 365 d	Any	Any	Any	Low
	≥ 365 d to 21 y	Nonamplified	Any	‐	Low
	≥ 365 d to 21 y	Amplified	Favorable	‐	Low
	≥ 365 d to 21 y	Amplified	Unfavorable	‐	High
3	< 365 d	Nonamplified	Any	Any	Intermediate
	< 365 d	Amplified	Any	Any	High
	≥ 365 d to 21 y	Nonamplified	Favorable	‐	Intermediate
	≥ 365 d to 21 y	Nonamplified	Unfavorable	‐	High
	≥ 365 d to 21 y	Amplified	Any	‐	High
4	< 548 d	Nonamplified	Any	Any	Intermediate
	< 365 d	Amplified	Any	Any	High
	≥ 548 d to 21 y	Any	Any	‐	High
4S	< 365 d	Nonamplified	Favorable	> 1	Low
	< 365 d	Nonamplified	Any	= 1	Intermediate
	< 365 d	Nonamplified	Unfavorable	Any	Intermediate
	< 365 d	Amplified	Any	Any	High

Reference: NCI PDQ 2017 
 DNA index: favorable > 1 (hyperdiploid) or < 1 (hypodiploid); unfavorable = 1 (diploid).
 Abbreviations. d: days of age; y: years of age.

Types of interventions

• Intervention

  • Addition of retinoic acid as part of a postconsolidation therapy after HDCT followed by autologous HSCT

• Comparator

  • Placebo retinoic acid or no addition of retinoic acid to postconsolidation therapy as described above

Types of outcome measures

We did not use the outcomes listed here as criteria for including studies, but these are the outcomes of interest within studies we identified for inclusion.

Primary outcomes

• Overall survival: the event is death by any cause from the start of retinoic acid treatment

• Treatment‐related mortality: incidence of deaths that were classified as treatment‐related, or the participants died of treatment complications

Secondary outcomes

• Progression‐free survival: time staying free of disease progression from start of retinoic acid treatment; participants may still have the disease but their disease is stable or showing a partial response to treatment; the events are death from all causes or any progression of the disease

• Event‐free survival: time staying free of any of a particular group of defined events from the start of retinoic acid treatment; participants may still have the disease; the events are death from all causes, any sign of the disease in participants who had a complete response to treatment, any relapse or progression of the disease, or events that were defined by the individual study protocol

• Early toxicity: adverse events within 90 days of the therapy; incidence of all reported adverse events, severe events (grades 3 and 4 of toxicity), and incidence of toxicity‐related discontinuation of treatment. Examples of possibly‐reported classifications: the Cancer Therapy Evaluation Program (CTEP) Common Terminology Criteria for Adverse Events (CTEP 2010); WHO Toxicity Grading Scale for Determining the Severity of Adverse Events (ICSSC 2003)

• Late toxicity: adverse events including secondary malignancy occurring 90 days or more beyond therapy

• Health‐related quality of life measured by validated questionnaires

Search methods for identification of studies

We used search methods as suggested in the Cochrane Handbook for Systematic Reviews of Interventions and by Cochrane Childhood Cancer (Higgins 2011; Kremer 2008). We applied no language restrictions.

Electronic searches

We searched in the following medical literature databases. We tailored the terms and syntax used for the search in MEDLINE to the requirements of the other two databases. Cochrane Childhood Cancer ran the searches in the three mentioned electronic databases; the review authors ran all other searches.

• Cochrane Central Register of Controlled Trials (CENTRAL) in the Cochrane Library (2016, Issue 11) using the search strategy shown in Appendix 1.

• MEDLINE in PubMed (1946 to 24 November 2016) using the search strategy shown in Appendix 2.

• Embase in Ovid (1947 to 24 November 2016) using the search strategy shown in Appendix 3.

We searched for ongoing trials by scanning the following online registries on 22 December 2016:

• ClinicalTrials.gov (clinicaltrials.gov/). Search: diagnosis 'neuroblastoma'; intervention 'retinoic acid'. Limits: study type 'interventional studies'; phase '2' or '3'.

• World Health Organization International Clinical Trials Registry Platform (ICTRP) (apps.who.int/trialsearch/). Search: diagnosis 'neuroblastoma'; intervention 'retinoic acid'. Limits: none.

We searched abstracts of the following annual meeting proceedings on 23 March 2017:

• Annual Meetings of the American Society of Clinical Oncology (ASCO) (meetinglibrary.asco.org/abstracts): 2012, 2013, 2014, 2015, and 2016 (ASCO Meetings 2012 to 2016). Search online by browsing abstracts by meeting and by category (Pediatric Oncology, Pediatric Solid Tumors): 'retinoic'. Search online all abstracts: 'retinoi'.

• Annual Meetings of the American Society of Hematology (ASH) (www.bloodjournal.org/page/ash‐annual‐meeting‐abstracts): ASH meetings in 2012, 2013, 2014, 2015, and 2016. Search online all abstracts: 'neuroblastoma'.

• Congresses of the International Society of Paediatric Oncology (SIOP) (Pediatric Blood & Cancer at onlinelibrary.wiley.com/): 2012, 2013, 2014, 2015, and 2016. Search in the downloaded PDF: 'retinoi'.

Searching other resources

We searched for information about trials not registered in electronic databases in reference lists of relevant articles and review articles.

Data collection and analysis

Selection of studies

In preparing this updated Cochrane Review, we endorsed the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) statement, adhered to its principles and conformed to its checklist (Moher 2009). We downloaded all titles and abstracts retrieved by electronic searching and removed duplicates. Two review authors independently examined the remaining references. We obtained the full texts of potentially relevant references. Two review authors independently examined the eligibility of retrieved papers, resolving any disagreements by discussion and consultating with a third review author if necessary. We documented reasons for exclusion. If we identified multiple reports of one study we used the most recently published results. We checked the multiple reports for possible duplicate data, addressed the issue, and did not include duplicate data in our analyses. We include a study selection flow chart in the review.

Data extraction and management

For each included study, two review authors independently abstracted study characteristics and outcomes, including information on study design, participant characteristics (such as inclusion criteria, age, stage, co‐morbidity, previous treatment, number enrolled in each arm), interventions (such as type of retinoic acid, dose applied, duration of therapy, control treatment), risk of bias, follow‐up duration, outcome measures, and deviations from protocol, onto a data extraction form specifically designed for this review. We resolved any differences in opinion between the review authors by discussion or by appeal to a third review author.

Where possible, all data extracted were those relevant to an intention‐to‐treat (ITT) analysis, in which all participants were analysed in the groups to which they were assigned. If this was not possible, we reported it. We noted the time points at which outcomes were collected and reported.

Assessment of risk of bias in included studies

Two review authors independently appraised the risks of bias in the included studies, resolving differences between review authors by discussion or by appeal to a third review author. We used the items listed in the Cochrane Childhood Cancer module (Kremer 2008), which is based on the Cochrane tool for assessing risk of bias (Higgins 2011):

1. Random sequence generation (selection bias)

2. Allocation concealment (selection bias)

3. Blinding of participants (performance bias)

4. Blinding of personnel (performance bias)

5. Blinding of outcome assessment (detection bias) for each outcome separately

6. Incomplete outcome data such as missing data for each outcome separately (attrition bias)

7. Selective reporting such as not reporting prespecified outcomes (reporting bias)

8. Other sources of bias such as bias related to the specific study design (other bias)

We applied the Cochrane tool for assessing risk of bias (Higgins 2011). In general, a low risk of bias means that plausible bias is unlikely to seriously alter the results (for example, participants and investigators enrolling participants could not foresee assignment). A high risk of bias means that plausible bias seriously weakens confidence in the results (for example, participants or investigators enrolling participants could possibly foresee assignments). Unclear risk of bias means that plausible bias raises some doubt about the results (for example, the method of concealment is not described or not described in sufficient detail to allow a definite judgement). In addition to the 'Risk of bias' tables, we included 'Methodological Quality' summaries. We took the results of the 'Risk of bias' assessment into account when interpreting the review results.

Measures of treatment effect

For analyses of time‐to‐event data, the primary effect measure was the HR. If the HR was not directly given in the publication, we estimated HRs according to methods proposed by Parmar 1998 and Tierney 2007. For all analyses we reported the 95% CIs.

We would have calculated the risk ratio for dichotomous outcomes. In the case of rare events, we planned to use the Peto odds ratio instead. We planned to analyse continuous data and present them using the mean difference (MD) if all results were measured on the same scale (e.g. length of hospital stay). If this was not the case (e.g. pain or quality of life), we planned to use the standardised mean difference (SMD). However, we did not identify either dichotomous or continuous data.

Dealing with missing data

We conformed to the Cochrane guidance for dealing with missing data (Higgins 2011). If data were missing or only imputed data were reported, we planned to contact trial authors to request data on the outcomes among participants who were assessed. When relevant data for study selection, data extraction, and risk of bias assessment were missing, we contacted study authors to retrieve the missing data.

Assessment of heterogeneity

We planned to assess heterogeneity between studies by visual inspection of forest plots, by estimation of the percentage of heterogeneity between trials which cannot be ascribed to sampling variation (I² statistic) (Higgins 2003) and, if possible, by subgroup analyses. If there was evidence of substantial heterogeneity, we planned to investigate and report the possible reasons for this. We considered an I² statistic value greater than 50% to indicate substantial heterogeneity. However, as only one trial met our inclusion criteria, this was not applicable.

Assessment of reporting biases

In addition to the evaluation of reporting bias as described in the Assessment of risk of bias in included studies section, we planned to assess reporting bias (such as publication bias, time lag bias, multiple publication bias, location bias, citation bias, language bias) by constructing a funnel plot when there were enough included studies (that is, at least 10 studies included in a meta‐analysis) because otherwise the power of the tests is too low to distinguish chance from real asymmetry (Higgins 2011). Since we only included one trial in the review, this was not applicable.

Data synthesis

One review author analysed the data using Review Manager 2014, and another review author checked them. If sufficient clinically‐similar studies were available, we planned to pool their results, but since we included only one study this was not applicable. We did not identify trials with multiple groups that 'shared' a comparison group, so it was not necessary to divide the 'shared' comparison group into the number of treatment groups and comparisons between each treatment group and treat the split comparison group as independent comparisons. We used a random‐effects model with inverse variance weighting for all analyses (DerSimonian 1986).

We used GRADEpro 2015 to create the 'Summary of findings' table as recommended in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We presented overall survival, treatment‐related mortality, progression‐free survival, event‐free survival, early toxicity, late toxicity, and health‐related quality of life, provided that data were presented for both treatment arms. For each outcome two review authors independently assessed the quality of the evidence by using the five GRADE considerations, that is, study limitations, inconsistency, indirectness, imprecision and publication bias as described in the Cochrane Handbook for Systematic Reviews of Interventions (Schünemann 2011).

Subgroup analysis and investigation of heterogeneity

We planned a subgroup analysis by age (younger than 18 months versus older than 18 months) since the International Neuroblastoma Risk Group classification has established 18 months as the optimum cut‐off for age. We also planned a subgroup analysis on MYCN gene amplification (with versus without MYCN amplification). However, the included study did not provide enough data to enable us to do this.

Sensitivity analysis

We planned to conduct sensitivity analyses of studies at low risk of bias versus studies at high or uncertain risk of bias, but since we include only one trial this was not applicable.

Results

Description of studies

Results of the search

See Figure 1 for the study flow diagram of the search for the previous and the current version of this Cochrane Review.

1.

1Study flow diagram of current review version.

Abbreviations. CT.gov: ClinicalTrials.gov; ICTRP: International Clinical Trials Registry Platform

For the original version of the review we retrieved 591 records after searching the databases CENTRAL, MEDLINE, and Embase (on 1 October 2014) and removing duplicates. We screened the title or abstract, or both, and excluded 492 records. We screened the full texts of 99 publications that reported potentially relevant data and excluded 91 publications. We included eight publications associated with one RCT (Matthay 1999). The study by Matthay et al., published in 1999, reported the main information and results and the publication by Matthay et al. 2009 updated these data. Four further publications were also associated with the study. Two conference proceedings (Reynolds 1998; Reynolds 2002; see Characteristics of studies awaiting classification table) were associated with the study Matthay 1999. However, as it is known that information provided in conference proceedings often differs substantially from information provided in subsequent full‐text publications (Yoon 2012), we did not include these results in our analyses. We retrieved 31 records searching the study registries ClinicalTrials.gov and the International Clinical Trials Registry Platform Search Portal, duplicates removed (27 December 2013), but found no additional relevant studies. We found no additional studies by screening the reference lists of relevant articles and reviews.

For this update we retrieved 67 records after searching the databases CENTRAL, MEDLINE, and Embase (on 24 November 2016) and removing duplicates. We screened the title or abstract, or both, and excluded 62 records. We screened the full texts of five publications that reported potentially relevant data and excluded all five of them. We retrieved nine records by searching the study registries ClinicalTrials.gov and the International Clinical Trials Registry Platform Search Portal, duplicates removed (22 December 2016), but found no additional relevant studies. We found no additional studies by screening the reference lists of relevant articles and reviews or by screening the conference abstracts. We knew about an erratum associated with the Matthay 1999 study (Erratum 2014), which we incorporated in this update. It was not necessary to contact authors to for missing information.

Included studies

We have described the characteristics of the included RCT in the Characteristics of included studies table.

Design

We include a randomised, prospective, parallel, controlled clinical trial. Matthay 1999 enrolled participants from 1991 to 1996.

Sample sizes

Matthay 1999 randomised 379 participants with high‐risk neuroblastoma to either HDCT followed by autologous HSCT (N = 189) or to continuous standard‐dose chemotherapy without transplantation (N = 190) in a first randomisation (Figure 2). After the first randomisation, 98 of the transplanted participants progressed to a second randomisation. Of the 98 transplanted participants, 50 were randomised to additional retinoic acid and 48 to no further therapy (Figure 2).

2.

Flow of patients in the Matthay 1999 study (as prepared by review author FP).
 Abbreviations. BMT: bone marrow transplantation; ContCT: continuation chemotherapy; CT: chemotherapy; HDCT: high‐dose chemotherapy; RA: retinoic acid.

Setting

Matthay 1999 was a multicentre study conducted in the United States of America.

Participants

Matthay 1999 included participants with high‐risk neuroblastoma stages 1, 2, 3, and 4, although most had stage 4. Patients who had progressive disease before week eight of the protocol were deemed ineligible for the trial. The characteristics of 98 participants with autologous HSCT randomised to retinoic acid or to no further therapy were not separately reported. The duration of their follow‐up was not mentioned.

Interventions

Participants in the retinoic arm received six cycles of retinoic acid. One cycle consisted of 13‐Cis retinoic acid at a dose of 160 mg/mg/m² a day for 14 consecutive days in a 28‐day cycle (14 days on, 14 days off), resulting in a cumulative dose after 28 days of 2240 mg/m². The control group received no further treatment. Regarding the consolidation therapy, high‐dose chemotherapeutic substances and doses were specified, such as carboplatin, etoposide, melphalan, and total‐body irradiation.

Primary outcomes

Matthay 1999 reported overall survival. Treatment‐related mortality was not reported separately for the treatment groups of interest.

Secondary outcomes

Matthay 1999 reported event‐free survival. It did not report early and late toxicity separately for the treatment groups of interest. Progression‐free survival and health‐related quality of life were not reported.

Excluded studies

We excluded 96 records of potentially relevant full‐text articles, 91 records in the previous version and five records (Chen 2015; Kushner 2015; Kushner 2016; Mora 2015; Mostoufi‐Moab 2016) in the current version (Figure 1) based on:

• not the diagnosis of interest: n = 7 (seven in the previous version and none in the current version);

• not the intervention of interest: n = 21 (20 in the previous version and one in the current version); this includes Kohler 2000 in which 49 of 175 participants (28%) did not receive high‐dose chemotherapy and bone marrow transplantation and no separate data for eligible participants were available;

• not the comparator of interest: n = 47 (46 in the previous version and one in the current version);

• not the outcome of interest or not reported separately: n = 5 (five in the previous version and none in the current version);

• not the publication type of interest: n = 16 (13 in the previous version and three in the current version).

We have described the excluded studies in more detail in the Characteristics of excluded studies table.

Risk of bias in included studies

The 'Risk of bias' tables in the Characteristics of included studies section provide details of each item of the 'Risk of bias' tool for RCTs. Figure 3 and Figure 4 provide an overview.

3.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

4.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Allocation

Matthay 1999 clearly described the random sequence generation, which we judged to be at low risk of bias. The concealment of the allocation was not specified, which prompted us to judge it as an unclear risk of selection bias.

Blinding

Matthay 1999 did not report blinding of investigators and participants and we judged it to be at unclear risk of performance bias. We judged a low risk of detection bias, as the study committee and investigators were unaware of participants' treatment assignments, and the study was monitored by an independent committee according to a group sequential monitoring plan.

Incomplete outcome data

The study did not report whether in all eligible participants all outcomes (i.e. overall and event‐free survival) were assessed, so there was an unclear risk of attrition bias.

Selective reporting

Matthay 1999 reported the outcomes in accord with their protocol. Concerning multiple (duplicate) publication bias, we found eight publications that reported results from the Matthay 1999 study. It appears probable that data for the same participants were included in several publications. However, we did not include duplicate data in the review. Overall, we judged there to be a low risk of reporting bias.

Other potential sources of bias

Analysis of data in Matthay 1999 was conducted using the data of the randomised participants, which should be consistent with the ITT principle. However, it was unclear if all 98 participants received the treatment to which they were randomised. Also, not all participants treated with transplantation in the first randomisation and without progressive disease afterwards were included in the second randomisation. It is unclear how many eligible participants did not undergo the second randomisation. The consequences of two randomisations in one study for the risk of bias are unclear. We judged there to be an unclear risk of other potential sources of bias.

Effects of interventions

See: Table 1

Primary outcomes  

Overall survival

We extrapolated the appropriate data on overall survival from the Kaplan Meier survival curve displayed in Figure 4B of the 2009 update article of Matthay 1999. We then calculated a HR of 0.87 (95% CI 0.46 to 1.63; low quality of evidence; Analysis 1.1; Figure 5). The difference between the groups was not statistically significant (P = 0.66). It was not reported if data for all 98 eligible participants were available, so it is unclear if this is an ITT analysis.

1.1. Analysis.

Comparison 1 Retinoic acid versus no further therapy, Outcome 1 Overall survival.

5.

Forest plot of comparison: 1 Retinoic acid versus no further therapy, outcome: 1.1 Overall survival.
 Abbreviations. CI: confidence interval; IV: inverse variance; SE: standard error.

We calculated the HR using the complete follow‐up period of the trial. The study also reported five‐year overall survival rates: 59% (standard error 8%) for the retinoic acid group and 41% (standard error 8%) for the no‐further‐therapy group (P value not reported). The values comply with the erratum, which corrected the value of the second standard error.

Treatment‐related mortality

We did not identify any results for treatment‐related mortality reported separately for the retinoic acid group and the control group.

Secondary outcomes  

Progression‐free survival

We did not identify any results for progression‐free survival.

Event‐free survival

We extrapolated the appropriate data on event‐free survival from the Kaplan Meier survival curves displayed in Figure 4A of the 2009 update article of Matthay 1999. Then, using the complete follow‐up period of the trial, we calculated a HR of 0.86 (95% CI 0.50 to 1.49; low quality of evidence; Analysis 1.2; Figure 6). The difference between the groups was not statistically significant (P = 0.59). It was not reported if data for all 98 eligible participants were available, so it is unclear if this is an ITT analysis.

1.2. Analysis.

Comparison 1 Retinoic acid versus no further therapy, Outcome 2 Event‐free survival.

6.

Forest plot of comparison: 1 Retinoic acid versus no further therapy, outcome: 1.2 Event‐free survival.
 CI: confidence interval; IV: inverse variance; SE: standard error.

Early toxicity

Matthay 1999 did not separately report severe adverse events for the eligible 98 participants.

Late toxicity including secondary malignancy

We did not identify any results for late toxicity including secondary malignancy reported separately for the retinoic acid group and the control group.

Health‐related quality of life

We did not identify any results for health‐related quality of life.

Discussion

Summary of main results

This Cochrane Review evaluates the current state of evidence on the efficacy of retinoic acid versus control treatment in people diagnosed with high‐risk neuroblastoma, pretreated with HDCT followed by autologous HSCT. This is an update of a previously published Cochrane Review. We did not identify any new studies.

We include one RCT (Matthay 1999) of 98 participants with high‐risk neuroblastoma who received autologous HSCT after a first randomisation and then 13‐cis‐retinoic acid (n = 50) or no further treatment (n = 48) after a second randomisation. This group had no progressive disease after HDCT followed by autologous HSCT. The objectives of study Matthay 1999 were to assess "whether myeloablative therapy in conjunction with transplantation of autologous bone marrow improved event‐free survival compared with chemotherapy alone, and whether subsequent treatment with 13‐cis‐retinoic acid (isotretinoin) further improves event‐free survival". The comparisons were between groups that differed in more than only the intervention of interest. Our review, however, was only interested in the efficacy of retinoic acid versus control treatment in people who were pretreated with transplantation. As a result, only a subset of the participants from the Matthay 1999 were eligible for for inclusion in the review.

There was no statistically significant difference between the treatment groups in either overall survival (HR 0.87, 95% CI 0.46 to 1.63; one trial; P = 0.66, low quality of evidence) or event‐free survival (HR 0.86, 95% CI 0.50 to 1.49; one trial; P = 0.59, low quality of evidence). We calculated the HRs using the complete follow‐up period of the trial. The study also reported survival estimates at a fixed point in time. At the time point of five years, overall survival was estimated to be 59% (SE 8%) for the retinoic acid group and 41% (SE 8%) for the no‐further‐therapy group (P value not reported). No data were available for the other outcomes of interest (i.e. treatment‐related mortality, progression‐free survival, early and late toxicity, and health‐related quality of life) (see Table 1).

Overall completeness and applicability of evidence

The inclusion of only one study in this review limits the inferences we can draw from the extracted data. 'No evidence of effect', as identified in this review, is not the same as 'evidence of no effect'. The reason that we found no statistically significant difference between study groups could be the fact that the number of participants included was too small to detect a difference between the treatment groups (i.e. low power). Also, the included study was not designed to identify a superior treatment combination.

Furthermore, the study was conducted from 1991 to 1996 (Matthay 1999). The applicability of these data to current clinical practice is limited, as medical knowledge and terms of health care have progressed and changed significantly since then. The results may therefore not be applicable to patients who are treated today. Also, the applicability may depend on the dosing regimen. According to Veal 2013, the medication dose may have a marked influence on retinoic acid plasma concentrations in children with high‐risk neuroblastoma. Matthay 2013 suggested that the medication dose may have the potential to cause different outcomes between studies. Prior treatment and the response to that treatment may also play a role. Since we only found only one eligible study we could not assess these issues.

Another point is that the included RCT used an age of one year as the cut‐off point for pretreatment risk stratification. Recently the age cut‐off for high‐risk disease was changed from one year to 18 months (Cohn 2009). As a result, it is possible that people with what is now classified as intermediate‐risk disease were included in the high‐risk groups. Consequently, the relevance of the results of these studies to current practice can be questioned. Survival rates may be overestimated due to the inclusion of participants with intermediate‐risk disease.

Only data on overall and event‐free survival were available for the patient population we were interested in (high‐risk neuroblastoma patients pretreated with HDCT followed by autologous HSCT). The included study did not provide data on, for example, toxicity and quality of life in these participants. As a result we cannot draw any conclusions regarding those outcomes, but they are of course important for clinical practice.

Cheung 2014 addressed the problem of end point selection and questioned the importance of HDCT followed by autologous HSCT: "A rethinking of the consolidation strategy for HR‐NB [explanation: high‐risk neuroblastoma] would be consistent with the general consensus among pediatric oncologists that ASCT [explanation: HDCT followed by autologous HSCT] in all other extracranial pediatric solid tumours is no longer recommended based on unsatisfactory risk‐benefit and cost‐benefit ratios assembled worldwide [the following references was used: Ratko 2012]."

Since information on the toxicity of an intervention is very important we decided to perform a post hoc search in the previous version for studies describing toxicity possibly associated with retinoic acid treatment, to get an impression of the importance of this outcome. We did not update the post hoc search for studies on toxicity data. We screened the 91 studies excluded in the search for the previous version for results on adverse events after retinoic acid application after HDCT followed by autologous HSCT in people with high‐risk neuroblastoma. We included all types of clinical studies and accepted meeting abstracts. We excluded narrative publication types such as nonsystematic reviews, letters and editorials. We did not report on those who received retinoic acid but did not receive transplantation. Thirty‐three out of these 91 potentially relevant publications were relevant (see Table 6); more than 1000 participants with high‐risk neuroblastoma were evaluated. In Kohler 2000, the exact number of relevant participants was not reported. 13‐cis‐retinoic acid (CRA, isotretinoin) was applied in 30 studies. 4‐hydroxyphenylretinamide (fenretinide) was applied in three studies. In 22 of the 33 studies, the authors reported participants with mild to severe organ toxicities associated with retinoic acid treatment. Most participants had mild and transient symptoms such as dry skin, cheilitis, rash, and bone pain. Retinoic acid was applied after a series of highly toxic chemotherapeutic regimens such as induction chemotherapy, myoablative chemotherapy, and autologous HSCT. It is difficult to distinguish adverse events associated with one agent rather than another, and it is difficult to draw causal inferences. Nevertheless, Khan 1996 reported a correlation between peak serum levels of retinoic acid and grade 3 to 4 toxicity such as involvement of skin, liver, as well as hypercalcaemia. We did not extract grade 1 to 2 toxicities. Maurer 2013 reported two participants with probable attribution of an elevation of serum alanine transaminase and aspartate transaminase and reported one participant with definite attribution of diarrhoea. However, six of 32 evaluated participants did not have a prior HDCT followed by autologous HSCT and the assignment of individual participants was not reported. We identified 13 studies that specifically reported an association of distinct adverse events with retinoic acid treatment (Clarke 2003; Cross 2009; Haysom 2005; Inamo 1999; Kohler 2000; Kreissman 2013; Marmor 2008; Mugishima 2008; Nishimura 1997; Rayburg 2009; Turman 1999; Villablanca 1993; Villablanca 1995). It should be noted that this was not a complete search for toxicity data and we did not assess the risks of bias in these studies. As a result, we can draw no definitive conclusions from these data.

5. Toxicity possibly associated with retinoic acid after high‐dose chemotherapy followed by autologous haematopoietic stem cell transplantation.

Study	Study design	Type of HSCT	Type of RA	Dose1	Pat2	Type of adverse event (N of affected participants)
Clarke 2003	CR	PBSCT	CRA	160	1	Pneumocystis carinii pneumonia (1)
Cross 2009	CR	NR	CRA	160	1	Hypercalcaemia (1), osteoblastic lesions (1)
De Kraker 2008	SA‐IS	BMT or PBSCT	CRA	160	44	NR
Granger 2012	SA‐IS	PBSCT	CRA	160	33	NR
Hamidieh 2012	SA‐IS	PBSCT	CRA	120 to 160	14	NR
Haysom 2005	CR	BMT	CRA	160	2	Bone marrow transplant nephropathy (2)
Inamo 1999	CR	BMT	CRA	33 to 1023	1	Growth failure (1)
Khan 1996	SA‐IS	BMT	CRA	100 to 200	31	Grade 3/4 toxicity of skin, liver and hypercalcaemia correlated with peak serum levels of CRA
Kletzel 2002	SA‐IS	PBSCT	CRA	160	12	Ataxia (1)
Kogner 2004	RCT	BMT	CRA	NR	12	NR
Kohler 2000	RCT	NR	CRA	15 to 224	NR	Dry skin (47), cheilitis (24), bone pain (16), other (13)
Kreissman 2013	RCT	PBSCT	CRA	160	192	Grade 3 toxic effects: hypertension (4), hematuria (2), elevated serum creatinine (2), proteinuria (3); purged and non‐purged transplantation group combined
Kushner 2003a	CS	NR	CRA	160	1	Cheilitis (1)
Laskin 2011	CS	NR	CRA	NR	20	NR
Marabelle 2009	CR	NR	CRA	160	3	Hypercalcaemia (3)
Marmor 2008	CR	NR	Fenretinide	666 to 20515	2	Rod electroretinogram suppression (2)
Mastrangelo 2011	SA‐IS	PBSCT	CRA	160	8	NR
Matthay 2006	SA‐IS	BMT	CRA	160	22	NR
Mugishima 2008	CR	BMT	CRA	130 to 400	2	Hypercalcaemia (2)
Nishimura 1997	CR	BMT	CRA	33 to 1023	1	Generalised metaphyseal modification (1)
Park 2011	SA‐IS	PBSCT	CRA	160	30	NR
Rayburg 2009	CR	NR	Fenretinide	2210	1	Langerhans cell histiocytosis (1)
Saarinen‐Pihkala 2012	SA‐IS	BMT or PBSCT	CRA	NR	36	NR
Simon 2011	SA‐IS	NR	CRA	160	75	NR
Sirachainan 2008	CR	NR	CRA	140	1	NR
Sung 2007	SA‐IS	PBSCT	CRA	125	44	Skin eruption, particularly face
Turman 1999	CR	BMT	CRA	160	2	Bone marrow transplant nephropathy (2)
Veal 2007	SA‐IS	NR	CRA	160	28	Mild skin toxicity (9), cheilitis (1), hypercalcaemia (2)
Veal 2013	SA‐IS	NR	CRA	160	103	Grade 3 ‐ 4 skin toxicity or cheilitis (5)
Villablanca 1993	SA‐IS	BMT	CRA	100 to 200	49	Dose‐limiting toxicity of hypercalcaemia (3), arthralgia and myalgia (1), grade 1 to 3 hypercalcaemia (9)
Villablanca 1995	SA‐IS	BMT	CRA	200	51	Hypercalcaemia (3), rash (2)
Villablanca 2011	SA‐IS	NR	Fenretinide	1800 to 2475	626	Rash (1), diarrhoea (1), nausea (2), vomiting (1), nyktalopia (1), abdominal pain (4)
Yu 2010	RCT	NR	CRA	160	108	"Few toxic effects"

Notice: Studies presented in this table were not eligible for inclusion in this review.

¹Dose in mg/m²/day. The unit mg/kg may be transformed to mg/m². The average body weight, body length, and body surface of a 6‐month‐old child may be 8 kg, 67 cm, and 0.39 m², thus 8 kg divided by 0.39 m² is roughly equalto a conversion factor of 20 (CDC 2000a). This factor increases continuously with age. The average body weight, body length, body surface of an 11‐year‐old child may be 36 kg, 143 cm, and 1.20 m², thus 36 kg divided by 1.20 m² is roughly equal to a conversion factor of 30 (CDC 2000b). The unit mg per day may be transformed to m² per day. For example, 300 mg per day may vary on average between 769 mg/m² (300 mg/0.39 m²) and 250 m² (300 mg/1.20 m²)

²Participants treated with retinoic acid after HDCT followed by autologous HSCT acid and evaluated for toxicity.

³Inamo 1999; Nishimura 1997: Patients received retinoic acid at a dose of 40 mg per day. The daily dose may vary on average between 102 mg/m² (40 mg/0.39 m²) and 33 m² (40 mg/1.20 m²)

⁴Kohler 2000: Participants received retinoic acid at a dose of 0.75 mg/kg per day. The daily dose may vary on average between 15 mg/m² (0.75 mg/kg * 20) and 22 mg/m² (0.75 mg/kg * 30).

⁵Marmor 2008: Participants received retinoic acid at a dose of 800 mg per day. The daily dose may vary on average between 2051 mg/m² (800 mg/0.39 m²) and 666 m² (800 mg/1.20 m²).

⁶Villablanca 2011: 51 of the 62 participants received HSCT.

BMT: bone marrow transplantation; CR: case report; CS: case series; CRA: 13‐cis‐retinoic acid; HDCT: high‐dose chemotherapy; HSCT: haematopoietic stem cell transplantation; N: number of participants; NR: not reported; PBSCT: peripheral blood stem cell transplantation; RA: retinoic acid; SA‐IS: single‐arm intervention study such as phase‐1 or phase‐2 clinical trial

We have only included RCTs, since it is widely recognised that an RCT is the only study design which can be used to obtain unbiased evidence on the use of different treatment options, provided that the design and execution of the RCT are adequate. However, even though RCTs are the highest level of evidence, it should be recognised that data from non‐randomised studies are available.

Quality of the evidence

Separate analyses of transplanted patients treated with or without additional retinoic acid were not reported in the included study. Nevertheless, the data were expressed in the survival curves presented in the study report. We extrapolated survival data from these graphs and estimated a HR to compare the two treatment groups. However, these estimates may be somewhat imprecise because the deduction and estimation process may deviate from the original data. Nevertheless, even considerable deviations would not be expected to change the observed difference in survival estimates.

The risk of bias in the included study was difficult to assess, in part due to a lack of reporting. As a result, we could not rule out the possible presence of selection bias, performance bias, attrition bias, and other bias. However, this remains the best available evidence from RCTs comparing the efficacy of additional retinoic acid versus control treatment after HDCT followed by autologous HSCT in high‐risk neuroblastoma patients.

Based on the GRADE assessments, in which we have looked at the five GRADE considerations (study limitations, consistency of effect, imprecision, indirectness, and publication bias), we judged the quality of the evidence for the two outcomes for which data were available as low; we downgraded the quality of the evidence due to risks of bias and imprecision.

Potential biases in the review process

One of the strengths of this review is the breadth of the search strategy, such that study retrieval bias is very unlikely. Nevertheless, there remains a slight possibility that an unknown number of studies were not registered and not published. Duplicate publication bias is very unlikely, because we searched for follow‐up papers of a single study to ensure that we included the updated version. Also, we excluded secondary analyses of registers or databases, which may use data that have been published previously by individual contributing study centres. Overall, the possibility of reporting bias seems to be low.

Authors' conclusions

Implications for practice.

We identified one RCT that evaluated additional 13‐cis‐retinoic acid after HDCT followed by autologous HSCT versus no additional retinoic acid in people with high‐risk neuroblastoma without progressive disease. The difference in overall survival and event‐free survival between both treatment alternatives was not statistically significantly different. This could be the result of low power; the included study was not designed to identify a superior treatment combination. Information on other outcomes, like toxicity, were not available. Also, this trial was performed between 1991 and 1996. Since then many changes in, for example, treatment and risk classification have occurred. Therefore, based on the currently available evidence, we are uncertain about the effects of additional retinoic acid after HDCT followed by autologous HSCT in people with high‐risk neuroblastoma.

Implications for research.

More research is needed for a definitive conclusion. Future trials on the use of retinoic acid (either with or without other treatments like anti‐GD2) after HDCT followed by autologous HSCT for people with high‐risk neuroblastoma should be RCTs focusing on overall survival, adverse effects, and quality of life. RCTs should be performed in homogeneous study populations (like stage of disease) and have a long‐term follow‐up. The number of included participants should be sufficient for the power needed for the results to be reliable. Different risk groups, using the most recent definitions, should be taken into account. For example, particular subgroups of participants may benefit from retinoic acid.

What's new

Date	Event	Description
24 November 2016	New search has been performed	We updated the search for eligible studies to 24 November 2016.
24 November 2016	New citation required but conclusions have not changed	We could include no new studies in this update; we found an erratum of the already‐included study, but the information provided did not change the conclusions.

History

Protocol first published: Issue 7, 2013
 Review first published: Issue 1, 2015

Date	Event	Description
14 April 2016	Amended	Contact details updated.

Appendices

Appendix 1. Search strategy for CENTRAL (The Cochrane Library)

1. For Retinoic acid the following text words were used:

retinoic acid OR retinoic acids OR Retinoid* OR Retinoid OR Retinoids OR tretinoin OR Vitamin A Acid OR trans‐Retinoic Acid OR trans Retinoic Acid OR all‐trans‐Retinoic Acid OR all trans Retinoic Acid OR beta‐all‐trans‐Retinoic Acid OR beta all trans Retinoic Acid OR 13‐cis‐RA OR 13‐cis‐retinoic acid OR 4759‐48‐2 OR Retin‐A OR Retin A OR Vesanoid OR isotretinoin OR ATRA OR Accutane OR Airol OR Dermairol

2. For Neuroblastoma the following text words were used:

neuroblastoma OR neuroblastomas OR neuroblast* OR ganglioneuroblastoma OR ganglioneuroblastomas OR ganglioneuroblast* OR neuroepithelioma OR neuroepitheliomas OR neuroepitheliom* OR esthesioneuroblastoma OR esthesioneuroblastomas OR esthesioneuroblastom* OR schwannian

Final search 1 and 2
 The search was performed in title, abstract or keywords

[*=zero or more characters]

Appendix 2. Search strategy for MEDLINE (PubMed)

1. For Retinoic acid the following MeSH headings and text words were used:

retinoic acid OR retinoic acids OR Retinoid* OR Retinoid OR Retinoids OR tretinoin OR Vitamin A Acid OR Acid, Vitamin A OR trans‐Retinoic Acid OR Acid, trans‐Retinoic OR trans Retinoic Acid OR all‐trans‐Retinoic Acid OR Acid, all‐trans‐Retinoic OR all trans Retinoic Acid OR beta‐all‐trans‐Retinoic Acid OR beta all trans Retinoic Acid OR 3‐cis‐RA OR 13‐cis‐retinoic acid OR 4759‐48‐2 OR Retin‐A OR Retin A OR Vesanoid OR isotretinoin OR ATRA OR Accutane OR Airol OR Dermairol

2. For Neuroblastoma the following MeSH headings and text words were used:

neuroblastoma OR neuroblastomas OR neuroblast* OR ganglioneuroblastoma OR ganglioneuroblastomas OR ganglioneuroblast* OR neuroepithelioma OR neuroepitheliomas OR neuroepitheliom* OR esthesioneuroblastoma OR esthesioneuroblastomas OR esthesioneuroblastom* OR schwannian

3. For RCTs and CCTs the following MeSH headings and text words were used:

((randomized controlled trial[pt]) OR (controlled clinical trial[pt]) OR (randomized[tiab]) OR (placebo[tiab]) OR (drug therapy[sh]) OR (randomly[tiab]) OR (trial[tiab]) OR (groups[tiab])) AND (humans[mh])

Final search 1 and 2 and 3

[pt = publication type; tiab = title, abstract; sh = subject heading; mh = MeSH term; *=zero or more characters; RCT = randomized controlled trial; CCT = controlled clinical trial]

Appendix 3. Search strategy for Embase (OVID)

1. For Retinoic acid the following Emtree terms and text words were used:

1. exp retinoic acid/
 2. (retinoic acid or retinoic acids).mp.
 3. (retinoid* or retinoid or retinoids).mp.
 4. tretinoin.mp.
 5. Vitamin A Acid.mp.
 6. (trans‐retinoic Acid or trans retinoic Acid or all‐trans‐Retinoic Acid or all trans Retinoic Acid).mp.
 7. (beta‐all‐trans‐retinoic acid or beta all trans retinoic acid or 3‐cis‐RA or 13‐cis‐retinoic acid).mp.
 8. 4759‐48‐2.rn.
 9. (Retin‐A or Retin A or Vesanoid or isotretinoin or ATRA or Accutane or Airol or Dermairol).mp.
 10. or/1‐9

2. For Neuroblastoma the following Emtree terms and text words were used:

1. exp neuroblastoma/
 2. (neuroblastoma or neuroblastomas or neuroblast$).mp.
 3. (ganglioneuroblastoma or ganglioneuroblastomas or ganglioneuroblast$).mp.
 4. (neuroepithelioma or neuroepitheliomas or neuroepitheliom$).mp.
 5. exp esthesioneuroblastoma/
 6. (esthesioneuroblastoma or esthesioneuroblastomas or esthesioneuroblastoma$).mp.
 7. schwannian.mp.
 8. or/1‐7

3. For RCTs and CCTs the following Emtree terms and text words were used:

1. Randomized Controlled Trial/
 2. Controlled Clinical Trial/
 3. randomized.ti,ab.
 4. placebo.ti,ab.
 5. randomly.ti,ab.
 6. trial.ti,ab.
 7. groups.ti,ab.
 8. drug therapy.sh.
 9. or/1‐8
 10. Human/
 11. 9 and 10

Final search 1 and 2 and 3

[mp = title, abstract, subject headings, heading word, drug trade name, original title, device manufacturer, drug manufacturer name]

Data and analyses

Comparison 1. Retinoic acid versus no further therapy.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Overall survival	1		Hazard Ratio (Random, 95% CI)	0.87 [0.46, 1.63]
2 Event‐free survival	1		Hazard Ratio (Random, 95% CI)	0.86 [0.50, 1.49]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Matthay 1999.

Methods	Setting Multicentre study denoted CCG‐3891. The abbreviation CCG derives from the Children's Cancer Group that merged in 2000 with other groups to form the Children's Oncology Group United States Duration of enrollment From 1991 to 1996 Randomisation A permuted‐block design was used for the random assignment of approximately equal numbers of participants from each of 2 strata (those with and those without metastatic disease) The study design included 2 randomisations. The 2nd randomisation was similarly balanced with respect to the numbers of participants from each group of the first randomisation and non‐randomised patients who were ineligible for transplantation First randomisation: Participants without progressive disease after 2 of 5 cycles of initial chemotherapy were randomly allocated to cytotoxic therapy group 1 that consisted of myeloablative therapy and autologous bone marrow transplantation or to cytotoxic therapy group 2 that consisted of 3 cycles of continuation chemotherapy. Participants with progressive disease were non‐randomly assigned to continuation chemotherapy 2nd randomisation: Participants without progressive disease after consolidation therapy were randomly assigned to 13‐cis‐retinoic acid or to no further therapy Median follow‐up time After diagnosis: median of 43 months (range 2 to 89) for 539 participants eligible for initial chemotherapy; not reported separately for the 98 participants eligible for this review From transplantation to start of 13‐cis‐retinoic acid median of 97 days; no information on time from transplantation to 2nd randomisation for the 98 participants eligible for this review Length of follow‐up from 2nd randomisation until end of study not reported for the 98 participants eligible for this review
Participants	Eligibility criteria Newly‐diagnosed high‐risk neuroblastoma patients; high‐risk neuroblastoma was defined as: stage IV neuroblastoma; stage III disease with 1 or more of the following: amplification of the MYCN oncogene, a serum ferritin level of at least 143 ng per mL, and unfavourable histopathological findings; stage II disease with amplification of MYCN (age > 1 year); stage I or II disease with bone metastases before therapy other than surgery; and stage IV disease with MYCN amplification for < 1 year. Staging was done using the Evans staging criteria (Evans 1971). Unfavourable histopathological findings were based on the Shimada classification (Shimada 1984) 1 to 18 years of age Number of patients eligible for this review 98 participants after high‐dose chemotherapy followed by bone marrow transplantation: 50 received retinoic acid and 48 received no further therapy Age Median and range not reported for 98 participants eligible for this review Gender Not reported Stage of disease Not reported for 98 participants eligible for this review Remission status None of the 98 participants eligible for this review had progressive disease at the time of the 2nd randomisation; no further information provided Compliance with randomisation Not reported for the 2nd randomisation Previous treatment, except initial and consolidation chemotherapy Not reported Comorbidity Not reported
Interventions	All participants Initial chemotherapy for 5 cycles at 28‐day intervals. 1 cycle consisted of: cisplatin (60 mg/m2 total dose), doxorubicin (30 mg/m2 total dose), etoposide (200 mg/m2 total dose), and cyclophosphamide (2000 mg/m2 total dose). Following induction chemotherapy patients with gross residual disease received surgery and radiotherapy (number of participants not reported) First randomisation BMT arm (patients in the continuation chemotherapy arm were not eligible for this review) Carboplatin (1000 mg/m2 total dose), etoposide (640 mg/m2 total dose), melphalan (210 mg/m2 total dose), total‐body irradiation (1000 cGy), purged bone marrow harvested at end of cycle 2, and granulocyte‐macrophage colony‐stimulating factor (250 microgram/m2 per day, number of days not reported) 2nd randomisation Retinoic acid arm 50 of 98 participants after high‐dose chemotherapy followed by autologous bone marrow transplantation were randomised to the retinoic acid arm to receive 6 cycles of retinoic acid. 1 cycle consists of: 13‐Cis retinoic acid at a dose of 160 mg/mg/m2 per day for 14 consecutive days in a 28‐day cycle (14 days on, 14 days off), resulting in a cumulative dose after 28 days of 2240 mg/m2 Control arm 48 of 98 participants after high‐dose chemotherapy followed by autologous bone marrow transplantation were randomised to the no‐further‐therapy arm
Outcomes	Primary outcomes Overall survival (definition not provided, but measured from 2nd randomisation) Secondary outcomes Event‐free survival. According to Matthay 1999: "The primary end point, prespecified by the protocol, was event‐free survival calculated from the time of second randomisation. The events considered were relapse, disease progression, death from any cause, and a second neoplasm, whichever occurred first."
Notes	No competing interest reported, funding, grants, and awards received from not‐for‐profit organizations. Supported in part by the National Cancer Institute; Neil Bogart Memorial Laboraties of the T.J. Martell Foundation for Leukemia, Cancer, and AIDS Research; American Institute for Cancer Research. The authors of Matthay 1999 concluded in their follow‐up paper in 2009: "Myeloablative therapy and autologous hematopoietic cell rescue result in significantly better 5‐year EFS and OS than nonmyeloablative chemotherapy; cis‐RA given after consolidation independently results in significantly improved OS." We have questioned the results of the statistical analyses and the authors eventually initiated a revision and published an erratum (Erratum 2014) and revised their conclusion as follows: "Myeloablative therapy and autologous hematopoietic cell rescue result in significantly better 5‐year EFS than nonmyeloablative chemotherapy; neither myeloablative therapy with autologous hematopoietic cell rescue nor cis‐RA given after consolidation therapy significantly improved OS.”
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	A permuted‐block design was used for the random assignment of approximately equal numbers of participants from each of 2 strata (those with and those without metastatic disease) to transplantation or continuation chemotherapy. The 2nd randomisation was similarly balanced with respect to the numbers of participants from each group of the first randomisation and non‐randomised participants who were ineligible for transplantation
Allocation concealment (selection bias)	Unclear risk	Allocation concealment was not described
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Blinding of participants, physicians and nurses was not reported
Blinding of outcome assessment (detection bias) All outcomes	Low risk	The study committee and investigators were unaware of participants' treatment assignments, and the study was monitored by an independent committee according to a group sequential monitoring plan
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	It was not reported if all reported outcomes for all participants were assessed
Selective reporting (reporting bias)	Low risk	Reporting was in agreement with the protocol with regard to the outcome measures. We were concerned with the possibility that data for the same participants were included in several publications. However, we did not include duplicate data in the review
Other bias	Unclear risk	Analysis was conducted using the data of the randomised participants, which should be consistent with the ITT principle. It was unclear if all 98 participants received the treatment to which they were randomised. Also, not all participants treated with transplantation in the first randomisation and without progressive disease afterwards were included in the 2nd randomisation. It is unclear how many eligible patients did not undergo the 2nd randomisation. The consequences of 2 randomisations in 1 study for the risk of bias are unclear.

AIDS: acquired immunodeficiency syndrome; BMT: bone marrow transplantation; cGy: centi‐Gray; CR: complete response; N: number.

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Adamson 1997	Not intervention of interest: not HDCT followed by autologous HSCT
Adamson 2007	Not comparator of interest
Aksoylar 2013	Not comparator of interest
Anderson 2005	Not intervention of interest: not HDCT followed by autologous HSCT, only title/abstract available
Atra 1996	Not intervention of interest: not retinoic acid
Bagatell 2014	Not comparator of interest: all participants received retinoic acid
Bauters 2011	Not publication type of interest
Castel 2004	Not publication type of interest
Chan 2007	Not intervention of interest: not HDCT followed by autologous HSCT
Chen 2015	Not an intervention of interest: not a HDCT followed by autologous HSCT
Clarke 2003	Not comparator of interest: no comparator
Cross 2009	Not comparator of interest: no comparator
De Kraker 2008	Not comparator of interest
Di Bella 2009	Not intervention of interest: not HDCT followed by autologous HSCT, only title/abstract available
Dmitrovsky 2004	Not publication type of interest
Elimam 2006	Not intervention of interest: not retinoic acid, only title/abstract available
Finklestein 1992	Not intervention of interest: not HDCT followed by autologous HSCT
Formelli 2008	Not intervention of interest: not HDCT followed by autologous HSCT
Formelli 2010	Not intervention of interest: not HDCT followed by autologous HSCT
Fouladi 2010	Not outcome of interest or not reported separately
French 2013	Not intervention of interest: not retinoic acid
Frgala 2007	Not comparator of interest, only title/abstract available
Garaventa 2003	Not comparator of interest
Granger 2012	Not comparator of interest
Grissom 1996	Not comparator of interest: no comparator
Gyorfy 2003	Not comparator of interest: no comparator
Hamidieh 2012	Not comparator of interest: all participants received retinoic acid
Haysom 2005	Not comparator of interest: no comparator
Hoefer‐Janker 1969	Not diagnosis of interest, only title/abstract available
Inamo 1999	Not comparator of interest: no comparator
Kazanowska 2008	Not intervention of interest: not HDCT followed by autologous HSCT
Khan 1996	Not comparator of interest: pharmacokinetic study
Kletzel 2002	Not publication type of interest: not RCT
Kogner 2004	Not comparator of interest
Kohler 2000	Not intervention of interest: 28% (49/175) patients did not receive high‐dose chemotherapy and bone marrow transplantation confirmed by author inquiry and after contacting the first author of this study, it became clear that separate data on the eligible participants were not available
Kreissman 2013	Not comparator of interest
Kushner 1994	Not intervention of interest: not consolidation therapy
Kushner 2001a	Not intervention of interest: not HDCT followed by autologous HSCT
Kushner 2001b	Not publication type of interest: not RCT
Kushner 2003a	Not comparator of interest: no comparator
Kushner 2003b	Not comparator of interest: no comparator
Kushner 2015	Not publication type of interest: not RCT
Kushner 2016	Not a comparator of interest: both arms received retinoic acid
Ladenstein 2004	Not comparator of interest
Ladenstein 2014	Not comparator of interest
Laskin 2011	Not publication type of interest: not RCT
Levin 2006	Not diagnosis of interest
Lie 1993	Not publication type of interest: narrative review
Marabelle 2009	Not comparator of interest: no comparator
Maris 2000	Not comparator of interest
Marmor 2008	Not comparator of interest: no comparator
Mastrangelo 2011	Not publication type of interest: not RCT
Matthay 1995	Not publication type of interest: narrative review
Matthay 1999a	Not intervention of interest, only title/abstract available
Matthay 2000	Not publication type of interest: editorial
Matthay 2006	Not intervention of interest
Matthay 2013	Not publication type of interest: comment
Maurer 2013	Not comparator of interest
Maurer 2014	Not comparator of interest
McCann 1993	Not diagnosis of interest, only title/abstract available
Mora 2015	Not publication type of interest: not RCT
Mostoufi‐Moab 2016	Not publication type of interest: not RCT
Mugishima 1995	Not intervention of interest, only title/abstract available
Mugishima 2008	Not comparator of interest: no comparator
Nishimura 1997	not comparator of interest: no comparator
Olgun 2008	Not intervention of interest: not retinoic acid, only title/abstract available
Ozkaynak 2014	Not comparator of interest
Park 2005	Not publication type of interest: duplicate data
Park 2011	Not comparator of interest
Pearson 2008	Not comparator of interest: all participants received retinoic acid
Rayburg 2009	Not comparator of interest: no comparator
Reed 1999	Not diagnosis of interest: not high‐risk neuroblastoma
Reynolds 2001	Not publication type of interest: narrative review
Richtig 2005	Not diagnosis of interest: melanoma
Rustin 1982	Not diagnosis of interest: not neuroblastoma
Saarinen‐Pihkala 2012	Not comparator of interest
Sato 2012	Not comparator of interest: no comparator
Seeger 2000	Not outcome of interest or not reported separately: pharmacokinetics
Simon 2005	Not intervention of interest: not retinoic acid
Simon 2011	Not comparator of interest: all participants received retinoic acid
Sirachainan 2008	Not comparator of interest: no comparator
Sung 2007	Not comparator of interest
Tang 2006	Not comparator of interest
Turman 1999	Not comparator of interest: no comparator
Veal 2007	Not outcome of interest or not reported separately: pharmacokinetics
Veal 2013	Not outcome of interest or not reported separately: pharmacokinetics
Villablanca 1992	Not outcome of interest or not reported separately: pharmacokinetics
Villablanca 1993	Not comparator of interest: no comparator
Villablanca 1995	Not comparator of interest: no comparator
Villablanca 2006	Not comparator of interest: no comparator
Villablanca 2011	Not comparator of interest: retinoic acid independent of HDCT followed by autologous HSCT
Weitman 1997	Not comparator of interest: neuroblastoma not reported separately
Ye 2010	Not comparator of interest: retinoic acid independent of not HDCT followed by autologous HSCT, only title/abstract available
Yu 2010	Not comparator of interest: all participants received retinoic acid
Yung 1996	Not diagnosis of interest: glioma
Zhu 2010	Not intervention of interest: not retinoic acid, only title/abstract available

HDCT: high‐dose chemotherapy; HSCT: hematopoietic stem cell transplantation; RCT: randomised controlled trial

Characteristics of studies awaiting assessment [ordered by study ID]

Reynolds 1998.

Methods	RCT
Participants	Patients with neuroblastoma
Interventions	13‐cis‐retinoic acid after intensive consolidation therapy
Outcomes	Event‐free survival
Notes	Title of an abstract of the annual meeting of the American Society of Clinical Oncology in 1998. Presumably associated with the study Matthay 1999.

Reynolds 2002.

Methods	RCT
Participants	Patients with high‐risk neuroblastoma
Interventions	13‐cis‐retinoic acid following myeloablative therapy
Outcomes	Overall survival
Notes	Title of an abstract of the annual meeting of the American Society of Clinical Oncology in 2002. Presumably associated with the study Matthay 1999.

Differences between protocol and review

In the protocol, we stated that we would present five outcomes in the 'Summary of findings' table: overall survival, treatment‐related mortality, early toxicity, late toxicity, and health‐related quality of life. In the 'Summary of findings' table, up to seven outcomes may be chosen for inclusion that should represent the most important outcomes to patients. Progression‐free survival and event‐free survival lead the list of genuinely important secondary outcomes in the protocol. We added those two outcomes to the 'Summary of findings' table to present more completely the seven most important outcomes. The types of primary and secondary outcomes in the review remain unchanged from the protocol.

Differences between the previous version and this version of the review

The search terms for the ongoing trials database ClinicalTrials.gov have been slightly changed. We specified the search to find study types 'interventional studies' and clinical study phase '2' or '3' to improve precision. In the protocol, we stated that we would search for abstracts presented at the last five consecutive annual meetings of the American Society of Clinical Oncology (ASCO), the International Society of Paediatric Oncology (SIOP) and the Advances in Neuroblastoma Research (ANR). We did not do this for the original version of the review, but did so during the update for ASCO, SIOP, and the meetings of the American Society of Hematology (ASH), but not for ANR.

Contributions of authors

All review authors approved the final version of the review.

FP: concept, selecting and appraising studies (both for the original version of the review and the update), extracting and analysing data, interpretation of results, primary manuscript preparation.

ECVD: appraising studies (for the original version), extracting and analysing data, interpretation of results, manuscript review.

HE: selecting and appraising studies (for the update), manuscript review.

FB: selecting and appraising studies (for the original version), providing a clinical perspective, interpretation of results, manuscript review.

Sources of support

Internal sources

• University of Cologne, Germany.

  Provision of the full‐text articles

External sources

• Stichting Kinderen Kankervrij (KiKa), Netherlands.

  The salary of ECVD is paid by KiKa; KiKa was not involved in the design and execution of this Cochrane Review.

Declarations of interest

FP: none known

ECVD: none known

HE: none known

FB: none known

New search for studies and content updated (no change to conclusions)