Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • Letter to the Editor
  • Published:

Issues in genetic association studies: limitations of statistical analysis and biological plausibility

Bone Marrow Transplantation volume 46, pages 906–907 (2011)Cite this article

Subjects

We read with great interest the article of Carvalho et al. reporting the impact of single-nucleotide polymorphisms (SNPs) in genes involved in the IL-23/Th17 pathway on the outcome of SCT. Their comprehensive multivariate analyses showed significant associations between SNPs in IL-17 and the IL-23 receptor (IL-23R) genes and infectious complications, the occurrence of acute GVHD, disease relapse, and overall survival of SCT recipients.1 Genetic association studies have helped enhance our understanding of the pathogenesis of SCT complications and transplant immunology.2, 3 However, establishing the clinical utility of these SNPs using patient and donor material has proven challenging mainly because of the difficulties encountered in performing SNP analysis, such as the lack of reproducibility, inadequate sample sizes, publication bias and the context (genetic background, type of transplantation and so on).4 We would also like to highlight two additional issues concerning SNP analysis, namely, the limitations of statistical analysis and the biological plausibility of genetic associations.

It is vital to appreciate the limitations of the statistical tools used for the analysis when exploring the impact of multiple SNPs among heterogeneous cohorts of limited sample size so as not to claim associations or overestimate the impact when there are none. In their multivariate analysis of acute GVHD, the authors used only those factors that were found to yield a P-value ⩽0.15 in univariate analysis—4 variables in total, namely, phase of disease, IL-17F rs763780 in donor and recipient, and IL-17A rs8193036 in the recipient. However, how is one to interpret the finding of such high hazard ratios for three associations when the event rate—acute GVHD—affected only 12 (6%) of the 201 patients? A fitted regression model is only likely to be reliable when the number of predictors (or candidate predictors if using variable selection) is less than m/10 or even m/20, where m is the ‘limiting sample size’. In the case of survival-time analyses, the ‘limiting sample size’ equals the number of failures, for example, occurrence of GVHD, death or relapse in this case.5 The limiting sample size for analysing the acute GVHD end point appears to be 12; hence, any multivariable analysis taking two or more predictors into account is unlikely to prove reliable and any hazard ratios that are high must be viewed with suspicion. The same applies to the end point of CMV infection, as 20 observed events does not allow proper multivariable analysis starting with five predictors. Given the authors' own comment that the low event rate precludes proper appreciation of their findings, perhaps it would have been better to omit their multivariate analysis altogether, as it neither weakens nor strengthens their message that SNPs in the IL-23/Th17 pathway might be prognostic factors for the SCT outcome.

This is a preview of subscription content, access via your institution

Access options

Buy this article

  • Purchase on Springer Link
  • Instant access to full article PDF
Buy now

Prices may be subject to local taxes which are calculated during checkout

References

  1. Carvalho A, Cunha C, Di Ianni M, Pitzurra L, Aloisi T, Falzetti F et al. Prognostic significance of genetic variants in the IL-23/Th17 pathway for the outcome of T cell-depleted allogeneic stem cell transplantation. Bone Marrow Transplant 2010; 45: 1645–1652.

    Article  CAS  PubMed  Google Scholar 

  2. Holler E, Rogler G, Herfarth H, Brenmoehl J, Wild PJ, Hahn J et al. Both donor and recipient NOD2/CARD15 mutations associate with transplant-related mortality and GvHD following allogeneic stem cell transplantation. Blood 2004; 104: 889–894.

    Article  CAS  PubMed  Google Scholar 

  3. Bochud PY, Chien JW, Marr KA, Leisenring WM, Upton A, Janer M et al. Toll-like receptor 4 polymorphisms and aspergillosis in stem-cell transplantation. N Engl J Med 2008; 359: 1766–1777.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  4. Colhoun HM, McKeigue PM, Davey SG . Problems of reporting genetic associations with complex outcomes. Lancet 2003; 361: 865–872.

    Article  PubMed  Google Scholar 

  5. Harrell FE . Regression Modeling Strategies, 1st edn. Springer-Verlag: New York, 2001, pp 60–61.

    Book  Google Scholar 

  6. Ioannidis JP . Why most published research findings are false. PLoS Med 2005; 2: e124.

    Article  PubMed  PubMed Central  Google Scholar 

  7. Gaffen SL . Structure and signalling in the IL-17 receptor family. Nat Rev Immunol 2009; 9: 556–567.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  8. Langrish CL, McKenzie BS, Wilson NJ, de Waal MR, Kastelein RA, Cua DJ . IL-12 and IL-23: master regulators of innate and adaptive immunity. Immunol Rev 2004; 202: 96–105.

    Article  CAS  PubMed  Google Scholar 

Download references

Author information

Authors and Affiliations

  1. Department of Haematology, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands

    W J F M Van der Velden, W B C Stevens, J P Donnelly & N M A Blijlevens

  2. Department of Epidemiology, Biostatistics and Health Technology Assessment, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands

    T Feuth

Authors
  1. W J F M Van der Velden
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. T Feuth
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. W B C Stevens
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. J P Donnelly
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. N M A Blijlevens
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to W J F M Van der Velden.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Rights and permissions

Reprints and permissions

About this article

Cite this article

Van der Velden, W., Feuth, T., Stevens, W. et al. Issues in genetic association studies: limitations of statistical analysis and biological plausibility. Bone Marrow Transplant 46, 906–907 (2011). https://doi.org/10.1038/bmt.2010.211

Download citation

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/bmt.2010.211

Search

Advanced search

Quick links