|Array CGH in patients with learning disability (mental retardation) and congenital anomalies: updated systematic review and meta-analysis of 19 studies and 13,926 subjects
|Sagoo G S, Butterworth A S, Sanderson S, Shaw-Smith C, Higgins J P, Burton H
This review found that array-based comparative genomic hybridisation could be used to identify genetic abnormalities in patients with learning disability or congenital anomalies in whom cytogenetic tests were negative, but there was a risk of false-positive results. The reliability of these conclusions is uncertain due to some unclear reporting, potential publication bias, and failure to appropriately consider study quality.
To determine the diagnostic and false-positive yield of array-based comparative genomic hybridisation (array CGH) in patients with learning disability and congenital anomalies.
MEDLINE, EMBASE and Web of Science were searched to March 2008. Full details of the search were reported in a separate publication (see Other Publications of Related Interest). No language restrictions were applied. Reference lists of primary studies were screened to identify additional studies.
Cohort studies and case series were eligible for inclusion if they used array CGH to identify genetic abnormalities in patients with learning disability and congenital anomalies, in whom conventional cytogenetic analysis was negative.
Studies were conducted with patients based in the USA, Europe, Japan, or combinations of these. All studies included sampling of control DNA as part of their protocol. Studies used a one mega base pairs (Mb) array for investigating the whole genome, targeted arrays and oligonucleotide arrays with 30 to 35 kilo base pairs (kb) resolution, 100kb array, and tiling bacterial artificial chromosome array. Most studies used control samples from healthy people, but one used samples from other patients in the cohort. Some studies used the de Vries clinical severity score for patient selection.
The authors did not state how the papers were selected for the review, or how many reviewers performed the selection.
Assessment of study quality
Study quality was assessed based on the following criteria: clear description of setting and population; description of selection criteria; evidence of appropriate pre-testing; inclusion and description of control samples; description of array CGH methodology; description of steps taken to identify and exclude known copy number polymorphisms using genome databases; appropriate follow-up testing; and clear interpretation of array CGH results. The authors did not state how many reviewers performed the quality assessment.
Data were extracted using a standardised form. Diagnostic yield was defined as the number of patients, who had variants detected by array CGH that were judged to be causal, divided by the total number of patients tested. False-positive yield was defined as the number of patients, who had variants detected by array CGH that were judged to be non causal or of unknown significance, divided by the total number of patients tested. Diagnostic and false-positive yields were calculated for each study.
Two reviewers independently performed the data extraction and disagreements were resolved through discussion.
Methods of synthesis
Diagnostic and false-positive yields were pooled using random-effects models after transforming these values onto a logit scale. The number needed to test to obtain one patient with a causal variant was calculated as the inverse of the diagnostic yield. Heterogeneity was assessed using the χ2 and I2 statistics. Meta-regression was used to investigate the effects of the following covariates: study size (<100, 100 to 499, or ≥500), patient source (genetic laboratory or clinical setting), geographic location (Europe, North America, Japan, or combined), array resolution (<1Mb, 1Mb, or targeted array) and use of de Vries clinical score (yes or no). Publication bias was assessed visually using funnel plots and statistically using the Egger test.
Results of the review
Nineteen studies (13,926 patients) were included in the review.
Diagnostic yield: This ranged from 6% to 35% with a pooled value of 10% (95% CI: 8 to 12; 19 studies). The number needed for testing to identify one new causal variant was 10% (95% CI: 8 to 13). There was strong evidence of heterogeneity (p<0.001; I2=72%). Meta-regression analysis showed that sample size and array resolution accounted for much of the observed heterogeneity. The pooled diagnostic yield from the three largest studies (over 1,000 patients) was 7% (95% CI: 7 to 8). There was strong evidence of publication bias (Egger test p=0.002).
False-positive yield: This ranged from 1% to 67% across studies with a pooled value of 7% (95% CI: 5 to 10), based on 18 studies. There was strong evidence of heterogeneity (p<0.001; I2=91%). There was no evidence of publication bias (p=0.796).
This review supported the use of array CGH in investigating patients with learning disability and congenital anomalies in whom conventional cytogenetic tests were negative. As this technology also identified similar numbers of false positives to causal variants it should be used with caution in clinical practice.
This review addressed a focused question supported by broadly defined inclusion criteria. The literature search was adequate for published studies, but specific attempts were not made to locate unpublished studies. Publication bias was assessed in the review, but the validity of the methods used for this type of study is unclear. Appropriate steps were taken to minimise bias and error in the extraction of data, but it was unclear whether these steps were also taken in the selection of studies and assessment of study quality. Study quality was assessed using appropriate criteria, but the results were not clearly reported or considered in the analysis. Appropriate methods were used to pool data and heterogeneity was assessed and investigated.
The authors conclusions are supported by the data presented, but their reliability is unclear due to uncertainties in the reporting of the review process, the possibility of publication bias, and failure to appropriately consider study quality.
Implications of the review for practice and research
Practice: The authors stated that these methods should be used with caution in clinical practice due to the possibility of false-positive results.
Research: The authors stated that specific cytogenetic regions could be targeted for greater coverage in future array designs to allow a more accurate estimation of the size and characteristics of any rearrangements identified. The use of array CGH as a first-line investigation in all patients with learning disabilities should be further evaluated in large prospective studies.
UKGTN; PHG foundation; National Institute for Health Research; and MRC grant (U.1052.00.011).
Sagoo G S, Butterworth A S, Sanderson S, Shaw-Smith C, Higgins J P, Burton H. Array CGH in patients with learning disability (mental retardation) and congenital anomalies: updated systematic review and meta-analysis of 19 studies and 13,926 subjects Genetics in Medicine 2009; 11(3): 139-146
Other publications of related interest
Subramonia-Iyer S, Sanderson S, Sagoo G, et al. Array-based comparative genomic hybridization for investigating chromosomal abnormalities in patients with learning disability: systematic review meta-analysis of diagnostic and false-positive yields. Genet Med 2007;9:74-9.
Subject indexing assigned by NLM
Comparative Genomic Hybridization /methods; Congenital Abnormalities /diagnosis /genetics; Humans; Intellectual Disability /diagnosis /genetics; Reproducibility of Results; Sensitivity and Specificity
Database entry date
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.