The first issue raised by Dr. Blumenthal and Dr. Borgeat regarding
differences in concentration and volume may indeed aid in explaining the
differences between their observations and ours; however, we would like to
point out that there were also several other differences between our studies.
We did not administer continuous analgesia at the second operative site, nor
did we use a bolus 30 mL of ropivacaine at the iliac crest graft site prior to
beginning use of the pump. We agree that additional studies are needed to
clarify these issues.
With regard to the "methodological concerns," we agree that the
pain at the surgical recipient site may have influenced the results. We refer
Dr. Blumenthal and Dr. Borgeat to the last three paragraphs of our paper,
where we address this very concern.
Again, as noted in our Discussion, study limitations included the variation
in the graft harvest site, the degree and variation in recipient sites, and
the size of the trap-door osteotomy. Analysis of these variables demonstrated
no statistical differences.
Another interesting issue raised by Dr. Blumenthal and Dr. Borgeat was the
effect of narcotics on activity-induced pain. We concede that opioids, in
contrast to local anesthetics, are weak blockers of the A-sigma fibers, which
are closely linked to movement, and that a visual analog pain score during
motion might differ from one at rest. However, pain scores are certainly not
obtained during motion in most studies, and even when they have been, as they
were in the study by Blumenthal et
al.1, they have
been, as expected, unchanged or worse compared with the pain score at rest.
Thus, we think that obtaining the pain scores during motion would have had
little impact on our study conclusions. However, this may be an area that
warrants additional study in the future.
We also appreciate the comments by Dr. Blumenthal and Dr. Borgeat regarding
our statistical analyses. They correctly point out the possibility that the
loss of two subjects from the study may have decreased the power of our
analyses, thereby limiting the ability to detect the difference specified as
clinically meaningful. However, an analysis of the effect size performed in
our study indicated that it is highly unlikely that the addition of two
subjects would have changed the study results. A power calculation indicates
that, given the small difference in the mean donor-site pain scores found
between the groups, 4800 subjects would be required to detect this difference,
assuming alpha = 0.05 and power = 0.80. As described in the article, such a
small difference was not considered to be clinically relevant, and it is
improbable that the addition of one subject to each group would have produced
significance. Regarding the repetitive pain assessment, Dr. Blumenthal and Dr.
Borgeat were also correct in noting the need for an adjustment of the alpha
value because of the multiple comparisons. We recognized this need and
identified it as a limitation of the study in the Discussion section, in which
we stated that the significant difference at the twenty-four-hour time-point
was likely a chance finding resulting from multiple comparisons. It is notable
that this is the only significant finding reported across the time-points, and
that while the Bonferroni adjustment would have reduced the alpha level
required for significance, it would not have changed the statistical test nor
the overall conclusions of nonsignificance for the other time-points.