Few microRNAs were found consistently dysregulated in type 2 diabetes (T2D) that would gain confidence from Big Pharma to develop diagnostic or therapeutic biomarkers. This study aimed to corroborate evidence from eligible microRNAs-T2D association studies according to stringent quality criteria covering both biological and statistical significance in T2D for biomarker development.
Methods and analyses
Controlled microRNA expression profiling studies on human with T2D will be retrieved from PubMed, ScienceDirect, and Embase for selecting the statistically significant microRNAs according to pre-specified search strategies and inclusion criteria. Multiple meta-analyses with restricted maximum-likelihood estimation and empirical Bayes estimation under the random-effects model will be conducted by metafor package in R. Subgroup and sensitivity analyses further examine the microRNA candidates for their disease specificity, tissue specificity, blood fraction specificity, and statistical robustness of evidence. Biologically relevant microRNAs will then be selected through genomic database corroboration. Their association with T2D is further measured by area under the curve (AUC) of receive operating characteristic (ROC). Meta-analysis of AUC of potential biomarkers will also be conducted. Enrichment analysis on potential microRNA biomarkers and their target genes will be performed by iPathwayGuide and clusterProfiler, respectively. The corresponding reporting guidelines will be used to assess the quality of included studies according to their profiling methods (microarray, RT-PCR, and RNA-Seq).
Ethics and dissemination
No ethics approval is required since this study does not include identifiable personal patient data.
Citation: Zhu H, Leung S-w (2021) MicroRNA biomarkers of type 2 diabetes: A protocol for corroborating evidence by computational genomics and meta-analyses. PLoS ONE 16(4): e0247556. https://doi.org/10.1371/journal.pone.0247556
Editor: Rajvir Dahiya, University of California San Francisco, UNITED STATES
Received: June 3, 2020; Accepted: January 28, 2021; Published: April 6, 2021
Copyright: © 2021 Zhu, Leung. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the manuscript and its Supporting information files.
Funding: The work of HZ and SL has been supported by research grants (MYRG190-Y3-L3-ICMS11-LSW, MYRG2014-00117-ICMS-QRCM, and MYRG2019-00159-ICMS) received from the University of Macau. The authors also thank the University of Edinburgh for their support of publication fee support of this article. The supporters had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Type 2 diabetes (T2D) is characterized by insulin resistance and beta-cell dysfunction , which severity would progress gradually (or alleviate temporarily) for decades. Without appropriate and timely interventions, preferably in pre-diabetes (Pre-T2D) and early T2D, malfunctions would occur in multiple organs, especially the heart, blood vessels, eyes, kidneys, and nerves . The International Diabetes Federation (IDF) estimated that 50% of the adult diabetics were still undiagnosed in 2017 and detectable Pre-T2D are still too late in disease states to manage . Therefore, the demands for even earlier diagnosis of diabetes using molecular medicine are imminent. MicroRNAs represent a kind of promising biomarkers for early diagnosis and monitoring of complex diseases [4, 5] including T2D . Recent studies found that the use of biomarkers in patient-selection exhibited higher overall success probabilities, compared to trials without biomarkers during the drug development [7, 8]. In addition, there have been patent applications regarding the use of microRNAs as disease-specific biomarkers (https://patentscope.wipo.int/search/en/result.jsf), such as (https://patentscope.wipo.int/search/en/detail.jsf?docId=WO2018125019&recNum=3&office=&queryString=FP%3A%28microRNA%29&prevFilter=&sortOption=Pub+Date+Desc&maxRec=3482).
MicroRNAs are small (approximately 22 nucleotides), endogenous, noncoding, highly stable, and gene-regulating RNAs. Their dysregulation were linked to many diseases [9, 10], including T2D [6, 11]. According to our pilot study , the controlled profiling studies on microRNAs in T2D were heterogeneous and inconsistent with one another; thus, they are subject to evaluation by proper meta-analysis which can integrate statistical evidence from multiple studies, estimates overall effect sizes, and assesses the reliability, credibility, and uncertainty of the evidence [13, 14]. Meta-analysis methods have been increasingly used in recent years for evidence-based clinical, biomedical, and social research [15, 16]. Since most of genomic studies have small sample sizes , meta-analysis would increase statistical power by combining multiple studies for identification of differentially expressed genes, biological networks and predictive models of diseases . Our demonstrative study published in early 2015 tested use a proper statistical method, replacing the obsolete vote-counting method, for conducting non-trivial microRNA meta-analyses . Since then, over 100 studies have cited and/or followed our pilot method for studying many other diseases. In the meantime, the present protocol will have been conducting to take stock of all microRNAs that are reliably associated with T2D, representing the largest meta-analytical cross-identification and cross-validation on microRNAs in T2D. According to our PubMed searches, the number of microRNA studies on T2D dramatically increased from 253 to 1404 (i.e. 5.5 folds) between 11 March 2014 and 3 April 2020. And eight systematic reviews [18–25] with various flaws published since our pilot study were found among the 1404 studies. There is a study  conducted a qualitative analysis without quality assessment and statistical meta-analysis. A study  followed our pilot method but it focuses on a single microRNA, namely miR-146a. Another study  replicated the methods and argumentation of our pilot study12. Assmann et al.  performed a qualitative analysis and Zhang et al.  conducted quantitative analysis on type 1 diabetes (T1D) rather than T2D. Gholaminejad et al.  and Zhou et al.  adopted an vote-counting method, which was abandoned by meta-analysts , to analyze diabetic nephropathy and retinopathy, respectively. Park et al.  focused on mixed data from both T1D and T2D nephropathy. Thus, an up-to-date and comprehensive meta-analysis since our publication in 2015  is still missing to fill the gap to confirm which microRNAs are reliably associated with T2D (Table 1). On the other hand, deeper biological relevance should be evaluated after confirming statistical significance and before any large-scale population-based studies in order to establish a genetic testing panel of microRNAs specific for early T2D diagnosis. Those population-based studies that should recruit >10K participants and might cost millions of US dollars would not be justified in terms of cost-effectiveness without a deep meta-analysis to corroborate and cross-validate evidence in terms of both statistical significance and biological relevance of specific microRNAs. Therefore, the deep meta-analysis and genomic analysis as provided in the present study protocol is crucial to evidence-based selection of T2D-associated microRNAs for establishing a genomic testing panel.
Methods and material
Databases including PubMed, ScienceDirect, and Embase will be searched for T2D microRNA expression profiling studies.
Basic search terms include diabetes, microRNA, expression and “profil*” will be used to search title, abstract and/or keywords for selection of studies comparing T2D with controls. For example, PubMed will be searched by following strategy: (‘miRNA’, ‘diabetes’ and ‘expression’ in Title/Abstract) or (‘miRNA’, ‘diabetes’ and ‘profil*’ in Title/Abstract) or (‘microRNA’, ‘diabetes’ and ‘expression’ in Title/Abstract) or (‘microRNA’, ‘diabetes’ and ‘profil*’ in Title/Abstract). Search strategies for the selected databases were tested from Nov 2019 to Jan 2020.
Eligibility criteria and study selection
Eligible studies have to meet the inclusion criteria: (a) they are microRNA expression profiling studies on the participants with T2D; (b) they use diabetic and non-diabetic control samples for comparison; (c) they use microRNA expression arrays; (d) they report cut-off criteria of differentially expressed microRNAs and (e) they report sample sizes. The microRNA profiling studies using saliva or urine of the participants with T2D will be excluded, because we focus on microRNAs in blood and microRNAs in saliva and urine are mostly released from oral cancer  and urinary tract cancer [27–29], respectively. The PRISMA-compliant flow chart for study selection will be presented as shown in Fig 1.
Data extraction and quality assessment
From the full text and S1 Checklist of each expression profiling study, the following eligibility items will be collected and recorded: first author, year of publication, location of study, selection and characteristics of the participants with T2D, microRNA expression profiling platform, sample sizes, tissue types, cut-off criteria of up-regulated and down-regulated microRNAs and the list of differentially expressed microRNAs (Table 2). The extracted microRNAs will be alignment with miRBase version 22  to unify the names before quality assessment. Quality assessment of microarrays will be performed according to the MIAME guideline version 2.0 . Studies adopted qPCR-based microRNA arrays will be assessed according to the MIQE guideline  which should be corresponding to the MIAME guideline (Table 3 and Fig 2). RNA-Seq studies would be assessed according to MINSEQE guideline (Table 4 and Fig 3), proposed by FGED Society in 2012 (http://fged.org/projects/minseqe/). And the quality of transcriptomic parts of included studies will be assessed, including the collection of raw data, actual data processing, sample annotation, experimental design, array annotation, and data processing protocol. Each item is evaluated with low risk, unclear risk and high risk, suggesting high reproducibility, ambiguous reproducibility and low reproducibility, respectively.
Green bars, yellow bars and red bars respectively indicate the items were low risk, unclear risk and high risk.
Green bars, yellow bars and red bars respectively indicate the items were low risk, unclear risk and high risk.
Extracted data will be transferred to R with the metafor package  for meta-analysis under random-effects model. Both restricted maximum-likelihood estimation (REML) and empirical Bayes estimation (EB) will be employed to estimate the outcomes. The outcomes of differential expressed microRNA are presented as loge (odds ratios) (LogOR), based on the numbers of dysregulation events in both the participants with T2D and non-diabetic control samples, with their 95% confidence intervals (CI). Adjusted p values less than 0.05 were considered statistically significant after Bonferroni corrections. The microRNAs statistically identified by both estimation methods (REML and EB) will be considered significant as differentially expressed in this meta-analysis. When compared to non-diabetic control groups, a significant logOR (> 0) indicates microRNA up-regulation. When compared to diabetic group, a significant logOR (> 0) indicates microRNA down-regulation. Differentially expressed microRNAs in the participants with T2D and non-diabetic control samples will be ranked according to the following order of importance (a) p values; (b) the number of consistent reports, and (c) OR values. Area under the curve (AUC) of receive operating characteristic (ROC) for each potential biomarker will be further preprocessed by pROC package  among studies with raw data on blood and blood fractions, before meta-analysis of AUC.
MicroRNAs are differentially expressed among complications or risk factors of T2D, tissue types and blood fractions, with corresponding overall effects and heterogeneities. Subgroup analyses will split the extracted data according to the complications/risk factors of T2D (heart failure, diabetic nephropathy, diabetic foot, hypertension, obesity, etc.), tissue types (blood, muscle, pancreas, liver, etc.) and blood fractions [serum, plasma, peripheral blood mononuclear cells (PBMCs), etc.] to compare microRNA expression profiles among complications/risk factors of T2D, tissue types and blood fractions (i.e. disease specificity, tissue specificity and blood fraction specificity). When examining tissue specificity, studies using serum, plasma, PBMCs will be classified as blood as they are from blood and aimed to investigate circulating microRNAs. And studies that do not report the pancreatic tissue is whole pancreas or pancreatic islets will be pooled with the other studies on pancreatic islets.
Sensitivity analysis will performed on the sample size to test the robustness of findings. Meta-analyses will be repeated after excluding the studies whose sample sizes are less than 25 and 50, respectively. Laterza et al.  and Kosaka et al.  have demonstrated how circulating microRNAs may indicate the physiological state at tissue level. The microRNAs that circulate in the blood are in a stable form and remain stable even after multiple freeze-thaw cycles. They can be detected by least invasive techniques  and are specific to tissue and disease states . If these microRNAs in circulating blood can serve as biomarkers, they would provide a less invasive approach to diagnosing and monitoring T2D. Therefore, the potential circulating biomarker candidates should be statistically significant and robustly up/down-regulated in sensitivity analyses (groups of sample sizes ≥ 50) and detectable in blood or blood fractions.
Funnel plots will be generated to visualize possible publication bias. Begg test  and Egger test  using the package metafor will be performed to evaluate the statistical significance of the publication bias. For selecting microRNAs as biomarkers, we only consider the microRNAs with statistical significant effect sizes that are corrected by the trim-and-fill method.
Genomic data integration
A set of known genes closely associated with a disease are called seed genes. Seed genes of T2D will be retrieved from KEGG (Kyoto Encyclopedia of Genes and Genomes) database , OMIM (Online Mendelian Inheritance in Man) database  and GWAS (Genome-Wide Association Studies) Catalog (URL: https://www.ebi.ac.uk/gwas/; selection criteria: with a significance cutoff of p-value no more than 1*10−10). Interactions among human microRNAs and their experimentally validated target genes will be downloaded from the databases miRTarBase 7.0 , miRecords 4.0 , and TransmiR 2.0 . MicroRNA names will be aligned with miRBase version 22 . Conversion of target gene symbols to gene IDs will be done by bioDBnet (biological DataBase network)  since a single gene may have multiple symbols but one unique ID. T2D seed genes will be used as a filter to identify corresponding microRNAs associated with T2D from above microRNA-target gene interactions. The frequencies of microRNAs, which represent the strength of the association between the microRNA and T2D, will be calculated after the identification. A high frequency indicates that a microRNA would interact with more seed genes or there are more articles reporting such interactions between the microRNA and seed genes. The microRNAs with above-average frequencies will be grouped for category intersection and visualized in a Venn diagram to identify potential circulating biomarker, such that our potential microRNA biomarkers would have both statistical significance and biological importance. The overall design of this protocol and selection of potential biomarkers will be shown in a flow chart (Fig 4). Enrichment analysis will be further performed on selected potential microRNA biomarkers and their target gens (both seed and non-seed genes) using iPathwayGuide (https://www.advaitabio.com/ipathwayguide) and clusterProfiler package  for R, respectively to validate whether those microRNAs are enriched in pathways involved in T2D.
Statistical significant circulating microRNAs will be first identified by meta-analysis, which is filtering by sensitivity analysis to find the robust one. Experiment validated microRNA targets are used to further screen biologically significant microRNA markers according to well-established T2D genes. Number with underline, in italic and in bold indicates the number of statistically significant microRNAs, the number of interactions between microRNAs and targets, and the number of seed genes of T2D, respectivily. ⊕, the union of microRNAs. ⦻, the intersection of microRNAs.
The protocol of this study has been registered in PROSPERO (International Prospective Register of Systematic Reviews) following the PRISMA-P guideline (S1 Checklist) with the registration number CRD42017081659. The meta-analyses and genomic integration will be conducted accordingly, and the final report of the corroborated evidence will be submitted to a peer-reviewed journal for publication.
- 1. Kahn SE. Clinical review 135: The importance of beta-cell failure in the development and progression of type 2 diabetes. J Clin Endocrinol Metab. 2001;86: 4047–58. Available: http://www.ncbi.nlm.nih.gov/pubmed/11549624.
- 2. Winer N, Sowers JR. Epidemiology of diabetes. J Clin Pharmacol. 2004;44: 397–405. pmid:15051748
- 3. International Diabetes Federation. IDF diabetes atlas 8th edition. International Diabetes Federation, Brussels; 2017.
- 4. Meijer LL, Frampton AE, Garajová I, Caparello C, Le Large TYS, Funel N, et al. Circulating microRNAs as dynamic biomarkers of response to treatment with FOLFIRINOX combination therapy in advanced pancreatic ductal adenocarcinoma. Lancet. 2017;389: 68.
- 5. Patel M, Verma A, Aslam I, Pringle H, Singh B. Novel plasma microRNA biomarkers for the identification of colitis-associated carcinoma. Lancet. Elsevier Ltd; 2015;385: S78. pmid:26312900
- 6. Shantikumar S, Caporali A, Emanueli C. Role of microRNAs in diabetes and its cardiovascular complications. Cardiovasc Res. 2012;93: 583–93. pmid:22065734
- 7. Wong CH, Siah KW. Estimation of clinical trial success rates and related parameters. Biostatistics. 2018;00: 1–14. pmid:29394327
- 8. Thomas D, Burns J, Audette J, Carrol A, Dow-Hygelund C, Hay M. Clinical development success rates 2006–2015. Biomedtracker. San Diego: Biomedtracker; 2016.
- 9. Ali AS, Ali S, Ahmad A, Bao B, Philip PA, Sarkar FH. Expression of microRNAs: potential molecular link between obesity, diabetes and cancer. Obes Rev. 2011;12: 1050–1062. pmid:21767342
- 10. Fabian MR, Sonenberg N, Filipowicz W. Regulation of mRNA translation and stability by microRNAs. Annu Rev Biochem. 2010;79: 351–79. pmid:20533884
- 11. Heneghan HM, Miller N, Kerin MJ. Role of microRNAs in obesity and the metabolic syndrome. Obes Rev. 2010;11: 354–361. pmid:19793375
- 12. Zhu H, Leung SW. Identification of microRNA biomarkers in type 2 diabetes: a meta-analysis of controlled profiling studies. Diabetologia. 2015;58: 900–11. pmid:25677225
- 13. Koricheva J, Gurevitch J. Place of meta-analysis among other methods of research synthesis. Handbook of Meta-analysis in Ecology and Evolution. 2013. p. 498.
- 14. Borenstein M, Hedges L V., Higgins JPT, Rothstein HR. Introduction to meta-analysis [Internet]. Chichester, UK: John Wiley & Sons, Ltd; 2009. https://doi.org/10.1002/9780470743386 pmid:29218357
- 15. Evangelou E, Ioannidis JPA. Meta-analysis methods for genome-wide association studies and beyond. Nat Rev Genet. 2013;14: 379–389. pmid:23657481
- 16. Levinson DF. Meta-analysis in psychiatric genetics [Internet]. Current Psychiatry Reports. 2005. pp. 143–151. pmid:15802092
- 17. Tseng GC, Ghosh D, Feingold E. Comprehensive literature review and statistical considerations for microarray meta-analysis. Nucleic Acids Res. 2012;40: 3785–3799. pmid:22262733
- 18. Villard A, Marchand L, Thivolet C, Rome S. Diagnostic value of cell-free circulating microRNAs for obesity and type 2 diabetes: a meta-analysis. J Mol Biomark Diagn. 2015;6: 1–21. pmid:27308097
- 19. Alipoor B, Ghaedi H, Meshkani R, Torkamandi S, Saffari S, Iranpour M, et al. Association of miR-146a expression and type 2 diabetes mellitus: a meta-analysis. Int J Mol Cell Med. 2017;6: 156–163. pmid:29682487
- 20. Liang YZ, Li JJH, Xiao HB, He Y, Zhang L, Yan YX. Identification of stress-related microRNA biomarkers in type 2 diabetes mellitus: a systematic review and meta-analysis. J Diabetes. 2018; pmid:29341487
- 21. Assmann TS, Recamonde-mendoza M, De Souza BM, Crispim D. MicroRNA expression profiles and type 1 diabetes mellitus: systematic review and bioinformatic analysis. Endocr Connect. 2017;6: 773–790. pmid:28986402
- 22. Gholaminejad A, Tehrani HA, Fesharaki MG. Identification of candidate microRNA biomarkers in diabetic nephropathy: a meta-analysis of profiling studies. J Nephrol. Springer International Publishing; 2018;31: 813–831. pmid:30019103
- 23. Park S, Moon SR, Lee K, Park IB, Lee DH, Seungyoon N. Urinary and blood microRNA-126 and -770 are potential noninvasive biomarker candidates for diabetic nephropathy: a meta-analysis. Cell Physiol Biochem. 2018;46: 1331–1340. pmid:29689545
- 24. Zhang L, Wu H, Zhao M, Lu Q. Identifying the differentially expressed microRNAs in autoimmunity: a systemic review and meta-analysis. Autoimmunity. Taylor & Francis; 2020;0: 1–15. pmid:31902241
- 25. Zhou H, Peng C, Huang DS, Liu L, Guan P. MicroRNA expression profiling based on microarray approach in human diabetic retinopathy: a systematic review and meta-analysis. DNA Cell Biol. 2020;39: 441–450. pmid:32101049
- 26. Park N, Zhou H, Elashoff D. Salivary microRNA: discovery, characterization, and clinical utility for oral cancer detection. Clin Cancer Res. 2009;15: 5473–5477. pmid:19706812
- 27. Hanke M, Hoefig K, Merz H, Feller AC, Kausch I, Jocham D, et al. A robust methodology to study urine microRNA as tumor marker: microRNA-126 and microRNA-182 are related to urinary bladder cancer. Urol Oncol. Elsevier Inc.; 2010;28: 655–61. pmid:19375957
- 28. Weber JA, Baxter DH, Zhang S, Huang DY, Huang KH, Lee MJ, et al. The microRNA spectrum in 12 body fluids. Clin Chem. 2010;56: 1733–41. pmid:20847327
- 29. Yamada Y, Enokida H, Kojima S, Kawakami K, Chiyomaru T, Tatarano S, et al. MiR-96 and miR-183 detection in urine serve as potential tumor markers of urothelial carcinoma: correlation with stage and grade, and comparison with urinary cytology. Cancer Sci. 2011;102: 522–9. pmid:21166959
- 30. Kozomara A, Griffiths-Jones S. miRBase: annotating high confidence microRNAs using deep sequencing data. Nucleic Acids Res. 2014;42: 68–73. pmid:24275495
- 31. Brazma A, Hingamp P, Quackenbush J, Sherlock G, Spellman P, Stoeckert C, et al. Minimum information about a microarray experiment (MIAME)-toward standards for microarray data. Nat Genet. 2001;29: 365–371. pmid:11726920
- 32. Bustin SA, Benes V, Garson JA, Hellemans J, Huggett J, Kubista M, et al. The MIQE guidelines: minimum information for publication of quantitative real-time PCR experiments. Clin Chem. 2009;55: 611–22. pmid:19246619
- 33. Viechtbauer W. Conducting meta-analyses in R with the metafor package. J Stat Softw. 2010;36: 1–48.
- 34. Robin X, Turck N, Hainard A, Tiberti N, Lisacek F, Sanchez JC, et al. pROC: An open-source package for R and S+ to analyze and compare ROC curves. BMC Bioinformatics. 2011; pmid:21414208
- 35. Laterza OF, Lim L, Garrett-Engele PW, Vlasakova K, Muniappa N, Tanaka WK, et al. Plasma microRNAs as sensitive and specific biomarkers of tissue injury. Clin Chem. 2009;55: 1977–83. pmid:19745058
- 36. Kosaka N, Iguchi H, Yoshioka Y, Takeshita F, Matsuki Y, Ochiya T. Secretory mechanisms and intercellular transfer of microRNAs in living cells. J Biol Chem. 2010;285: 17442–52. pmid:20353945
- 37. Cho WCS. Circulating microRNAs as minimally invasive biomarkers for cancer theragnosis and prognosis. Front Genet. 2011;2: 1–6. pmid:22303300
- 38. Chen X, Ba Y, Ma L, Cai X, Yin Y, Wang K, et al. Characterization of microRNAs in serum: a novel class of biomarkers for diagnosis of cancer and other diseases. Cell Res. 2008;18: 997–1006. pmid:18766170
- 39. Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics. 1993;50: 1088–1101.
- 40. Egger M, Smith GD, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ. 1997;315: 629–34. Available: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2665616&tool=pmcentrez&rendertype=abstract. pmid:9310563
- 41. Kanehisa M, Sato Y, Kawashima M, Furumichi M, Tanabe M. KEGG as a reference resource for gene and protein annotation. Nucleic Acids Res. 2016;44: D457–D462. pmid:26476454
- 42. Hamosh A, Scott AF, Amberger J, Valle D, McKusick VA. Online Mendelian Inheritance in Man (OMIM). Hum Mutat. 2000;15: 57–61. pmid:10612823
- 43. Chou C-H, Shrestha S, Yang C-D, Chang N-W, Lin Y-L, Liao K-W, et al. miRTarBase update 2018: a resource for experimentally validated microRNA-target interactions. Nucleic Acids Res. Oxford University Press; 2018;46: D296–D302. pmid:29126174
- 44. Xiao F, Zuo Z, Cai G, Kang S, Gao X, Li T. miRecords: an integrated resource for microRNA-target interactions. Nucleic Acids Res. 2009;37: D105–10. pmid:18996891
- 45. Wang J, Lu M, Qiu C, Cui Q. TransmiR: a transcription factor microRNA regulation database. Nucleic Acids Res. 2010;38: 119–122. pmid:19786497
- 46. Mudunuri U, Che A, Yi M, Stephens RM. bioDBnet: The biological database network. Bioinformatics. 2009;25: 555–556. pmid:19129209
- 47. Yu G, Wang LG, Han Y, He QY. ClusterProfiler: An R package for comparing biological themes among gene clusters. OMICS. 2012;16: 284–287. pmid:22455463