Meta-analysis of MTHFR 677C→ T polymorphism and coronary heart disease: does totality of evidence support causal role for homocysteine and preventive potential of folate?
I admire authors' conciseness in their report [1]-- they did not
appear to include any other data than a general list of refereces to the
studies they analyzed. I also admire their courage to report just another
negative finding without providing any factual background nor any actual
proof to the reader.
Meta-analyses are, typically, done by collecting all of the
publication data into on large table and then by analyzing all of the
interactions of interest. In many cases, authors are normally provide this
intermediate table (slang: 'meta-table') allowing the reader to judge how
accurate their analysis actually was (for instance, see [2] and the
supplementary table there). In the present study [1] , which appears to
report a negative association, no evidence of a careful analysis being
done is presented to the interested reader.
Thus, in [2], the authors were very kind to provide most of the meta-
table they used in their meta-analysis. This allowed me, among other
readers, to find several strong confounders and to re-analyze again their
thorough, albeit incomplete work [3]. After the analysis, it became clear
that the authors'
1. neglect of the considerable bias from the very small studies,
2. neglect of the accurate ethnicity report and
3. neglect of the type of the study (case control vs
nested vs prospective etc)
resulted in reporting of 4 negative associations [2] where as, in
facts, these were statistically strong associations with P well below
0.001 [3].
Unfortunately, the daring courage of the authors of the present study
[1] entirely prevents me (or any other interested reader of BMJ, for that
matter) to take a better look on the problem.
Although it may seem a bit foolhardy, I would recommend BMJ editorial
staff to require supplementary tables (of the kind given in [2]) for all
meta-analyses accepted for publication in the journal. Otherwise, we are
dealing with sheer opinions, biases, wrangling and not with scientific
research.
REFERENCES
[1] Sarah J Lewis, Shah Ebrahim, and George Davey Smith
Meta-analysis of MTHFR 677CT polymorphism and coronary heart disease: does
totality of evidence support causal role for homocysteine and preventive
potential of folate?
BMJ 2005; 331: 1053
[2] Ye Z, Liu EH, Higgins JP, Keavney BD, Lowe GD, Collins R, Danesh
J. Seven haemostatic gene polymorphisms in coronary disease: meta-analysis
of 66,155 cases and 91,307 controls. Lancet. 2006 Feb 25;367(9511):651-8.
See the web-extras at http://www.thelancet.com/journals/lancet/article/PIIS0140673606682639/ab...,
especially the table mmc2.pdf
[3] Torshin I.Y. Bioinformatics in post-genomic biology: physiology
and medicine, Nova Sci, NY, in press (2007-2008).
Rapid Response:
to perish, to publish, to haste, too...
SIR
I admire authors' conciseness in their report [1]-- they did not
appear to include any other data than a general list of refereces to the
studies they analyzed. I also admire their courage to report just another
negative finding without providing any factual background nor any actual
proof to the reader.
Meta-analyses are, typically, done by collecting all of the
publication data into on large table and then by analyzing all of the
interactions of interest. In many cases, authors are normally provide this
intermediate table (slang: 'meta-table') allowing the reader to judge how
accurate their analysis actually was (for instance, see [2] and the
supplementary table there). In the present study [1] , which appears to
report a negative association, no evidence of a careful analysis being
done is presented to the interested reader.
Thus, in [2], the authors were very kind to provide most of the meta-
table they used in their meta-analysis. This allowed me, among other
readers, to find several strong confounders and to re-analyze again their
thorough, albeit incomplete work [3]. After the analysis, it became clear
that the authors'
1. neglect of the considerable bias from the very small studies,
2. neglect of the accurate ethnicity report and
3. neglect of the type of the study (case control vs
nested vs prospective etc)
resulted in reporting of 4 negative associations [2] where as, in
facts, these were statistically strong associations with P well below
0.001 [3].
Unfortunately, the daring courage of the authors of the present study
[1] entirely prevents me (or any other interested reader of BMJ, for that
matter) to take a better look on the problem.
Although it may seem a bit foolhardy, I would recommend BMJ editorial
staff to require supplementary tables (of the kind given in [2]) for all
meta-analyses accepted for publication in the journal. Otherwise, we are
dealing with sheer opinions, biases, wrangling and not with scientific
research.
REFERENCES
[1] Sarah J Lewis, Shah Ebrahim, and George Davey Smith
Meta-analysis of MTHFR 677CT polymorphism and coronary heart disease: does
totality of evidence support causal role for homocysteine and preventive
potential of folate?
BMJ 2005; 331: 1053
[2] Ye Z, Liu EH, Higgins JP, Keavney BD, Lowe GD, Collins R, Danesh
J. Seven haemostatic gene polymorphisms in coronary disease: meta-analysis
of 66,155 cases and 91,307 controls. Lancet. 2006 Feb 25;367(9511):651-8.
See the web-extras at
http://www.thelancet.com/journals/lancet/article/PIIS0140673606682639/ab...,
especially the table mmc2.pdf
[3] Torshin I.Y. Bioinformatics in post-genomic biology: physiology
and medicine, Nova Sci, NY, in press (2007-2008).
Competing interests:
None declared
Competing interests: No competing interests