The two letter writers express concerns about our data describing a
protective role of folic acid supplementation in cleft lip. Dr van der
Wouden is correct that an adjusted odds ratio for the highest level of
folic acid is not by itself statistically significant. However, it would
be wrong to suggest that there is anything misleading in these analyses,
which are fully laid out in the paper. Table 2 provides a complete
breakdown of the crude and adjusted odds ratios for three levels of folic
acid. The three crude estimates for the highest level of folic acid are
strongly statistically significant, and statistical adjustment has little
impact on the risk. The pooling of the two lower categories of folic acid
was explicitly described and justified in the paper.
Dr. Cooney asks about further adjustments for family history of
clefts, epilepsy, and anti-convulsant medications. We did collect this
information; there were few affected mothers, and these factors had no
effect on our estimates. Regarding sample size, most experts agree that
once a study is completed, confidence intervals are much better than crude
sample size calculations in describing the power of the study.
We are surprised that the letter writers focus on such small points
and dismiss the bigger picture. If we take into account the supporting
animal studies, the strengths of the present study design (constructed
specifically to test the folic acid hypothesis), and the specificity of
the folic acid association with cleft lip but not cleft palate (consistent
with prior studies), there is a striking convergence of evidence to
suggest that folic acid protects against cleft lip. This convergence seems
worth taking seriously.
Rapid Response:
The authors respond.
The two letter writers express concerns about our data describing a
protective role of folic acid supplementation in cleft lip. Dr van der
Wouden is correct that an adjusted odds ratio for the highest level of
folic acid is not by itself statistically significant. However, it would
be wrong to suggest that there is anything misleading in these analyses,
which are fully laid out in the paper. Table 2 provides a complete
breakdown of the crude and adjusted odds ratios for three levels of folic
acid. The three crude estimates for the highest level of folic acid are
strongly statistically significant, and statistical adjustment has little
impact on the risk. The pooling of the two lower categories of folic acid
was explicitly described and justified in the paper.
Dr. Cooney asks about further adjustments for family history of
clefts, epilepsy, and anti-convulsant medications. We did collect this
information; there were few affected mothers, and these factors had no
effect on our estimates. Regarding sample size, most experts agree that
once a study is completed, confidence intervals are much better than crude
sample size calculations in describing the power of the study.
We are surprised that the letter writers focus on such small points
and dismiss the bigger picture. If we take into account the supporting
animal studies, the strengths of the present study design (constructed
specifically to test the folic acid hypothesis), and the specificity of
the folic acid association with cleft lip but not cleft palate (consistent
with prior studies), there is a striking convergence of evidence to
suggest that folic acid protects against cleft lip. This convergence seems
worth taking seriously.
Competing interests:
None declared
Competing interests: No competing interests