In a meta-analysis of 11 studies involving children born after assisted reproductive technologies, no significant increased risk of cancer was found. Limitations in the review process, the lack of a validity assessment, and substantial heterogeneity between the studies render the validity of the results questionable.

To investigate the relationship between assisted reproductive technologies (ART) and cancer in children.

MEDLINE and EMBASE were searched to February 2005. A further search of the bibliographies of retrieved articles, review articles and citation indices was performed. The search was limited to papers published in English. The keywords used were listed.

Study designs of evaluations included in the review

The inclusion criteria stated that cohort studies were eligible for inclusion. Most of the included studies were cohort studies, two were based on a hypothetical cohort and one was based on national statistical data without a systematic follow-up. The duration of follow-up ranged from 1 to 13 years.

Specific interventions included in the review

Studies in which ART was used were eligible for inclusion. No further details of the type of technique (e.g. in vitro fertilisation, gamete intra-fallopian transfer or intra-cytoplasmic sperm injection) used in the included studies were reported.

Participants included in the review

Studies that involved children born after ART were eligible for inclusion. No further details of the children in the included studies were given.

Outcomes assessed in the review

Studies in which cancer (of any type) was reported as an outcome were eligible for inclusion. The outcomes reported in the review were standardised cancer incidence ratios (SIRs).

How were decisions on the relevance of primary studies made?

Two experts independently screened the titles and abstracts of the papers. The authors did not state how any disagreements were resolved.

The authors did not state that they assessed validity.

The authors did not state how the data were extracted for the review, or how many reviewers performed the data extraction. The observed number of cases was extracted for each study, while the expected numbers were either extracted directly or calculated if not reported in the original study. Details of the data used to derive the expected numbers were reported. SIRs were calculated using the observed/expected number of cancer cases.

How were the studies combined?

The studies were combined using either a fixed-effect or random-effects meta-analysis, depending on the test for heterogeneity. An overall SIR adjusted for study was calculated. Further details of the statistical methods used were not provided. Publication bias was investigated by examining a funnel plot and using Egger's test.

How were differences between studies investigated?

Heterogeneity across the studies was assessed using Cochran's test. Sensitivity analyses were conducted by excluding the three studies of non-cohort design.

Eleven studies (approximately 40,815 children) were included in the review. The number of exposed children was not known for one of the included studies and was approximate for another. The 8 cohort studies included 36,530 children.

None of the 8 cohort studies found a significant increase in cancer among children born following ART compared with the general population. Of the 3 non-cohort designs, one found a significant increase of neuroectodermal cancer, one of retinoblastoma and one of all cancers combined. The observed number of cancer cases per study ranged from none to 11.

The meta-analysis of 11 studies found no significant increase in cancer among children born after ART; the study-adjusted SIR was 1.33 (95% confidence interval: 0.62, 2.85). There was no evidence of publication bias (P=0.70), but there was significant heterogeneity between the studies (P<0.001).

When restricting the analysis to the 8 cohort studies, there was a non significant decrease in cancer among children born after ART; the study-adjusted SIR was 0.77 (95% confidence interval: 0.41, 1.42). There was significant heterogeneity (P=0.003), but this was not present when the 2 small studies with no observed cancer cases were excluded (P=0.76).

There appeared to be no increased risk of cancer in children born after ART, although a firm conclusion was not possible given the limited number of studies.

The research question was clearly stated, as were inclusion criteria for the review. The search strategy was not reported in sufficient detail to allow replication. Efforts were made to locate other published studies, but language restrictions and no attempts to identify unpublished material increased the chance of some studies being missed. No details of how the data were extracted were reported, making it difficult to comment on efforts made to reduce errors and bias. The validity of the included studies does not appear to have been assessed, and limited study and participant details were provided. Hence, the reliability of the synthesised results is questionable, particularly as the risk of confounding factors cannot be assessed.

Studies that did not meet the stated inclusion criteria were included in the review, although sensitivity analyses of only the eligible study designs were conducted. The authors presented results from a meta-analysis, despite the fact that there was significant heterogeneity between the study results and the calculations of the expected cancer rates varied between countries. There was also no discussion of the in vitro fertilisation methods used in each study, or any other clinical or methodological differences which might have contributed to the observed heterogeneity.

This review had several limitations: no validity assessment; considerable between-study heterogeneity; limited reporting of the review methods, study and participant details; and differing lengths of follow-up. The pooled SIR results should be treated with caution. However, the authors' overall conclusion is appropriately cautious given the limited evidence presented.

Practice: The authors did not state any implications for practice.

Research: The authors stated that a pooled analysis of the available data would be beneficial. Further research into the relationship between hormones given to women for infertility problems and childhood cancer, as well as congenital malformation in relation to ART, is needed.

Ministry of Health, grant number 521C/7-I.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.