EURURO Vol. 72 No. 4

E91 656

Letter to the Editor
Re: Richard J. Sylvester, Steven E. Canfield, Thomas B[3_TD$DIFF].L.
Lam, et al. Conflict of Evidence: Resolving Discrepancies
When Findings from Randomized Controlled Trials
and Meta-analyses Disagree. Eur Urol [4_TD$DIFF]2017;71:811–9
The article by Sylvester and members of the European
Association of Urology guidelines leadership team
addresses the intriguing issue of conflicting evidence from
individual, large, and well-designed randomized controlled
trials (RCTs) on the one hand and systematic reviews (SRs)
on the other . While we agree with most of the
statements made, we are concerned that some issues are
not laid out as clearly as they could be. We therefore wish to
highlight a few important take home messages.
First, developers of evidence-based clinical practice
guidelines are required to develop their recommendations
on the basis of rigorous SRs of the entire body of evidence.
This reflects a widely accepted principle of evidence-based
medicine (EBM) and a mandate by the Institute of Medicine,
now the National Academy of Medicine, in its publication
Clinical Practice Guidelines We Can Trust . Second, the
phenomenon of conflicting results from different studies
addressing the same clinical question is common and one
that authors of rigorous SRs are well equipped to deal with
through a priori planned subgroup analyses that seek to
investigate the source of unexplained heterogeneity. It is
noteworthy that the results of the SUSPEND trial on the
value of medical expulsive therapy (MET) in patients with
ureteral colic are not that different from the findings of
54 other trials addressing the same question when
combined in a meta-analysis As derived from a
predefined subgroup analysis, MET may in fact be beneficial
in patients with large stones, an analysis that the SUSPEND
trial was not designed or powered to adequately address
. Third, guidance exists as to when SR authors should look
beyond evidence from RCTs alone. Rather than indiscrimi-
nately pooling across RCTs and observational studies, SR
authors should first scrutinize the evidence from RCTs
alone. If study limitations and other issues affecting the
quality of evidence according to GRADE force them to lower
the confidence in the estimates of effect for a given
comparison and outcome, SR authors should draw in
observational data This makes the most sense when
the quality of evidence is low, as is the case for the EORTC
trial, which was marred by lack of blinding, substantial
crossover, and imprecision
Finally, it is important to emphasize that the work of
guideline developers is not limited to interpreting the
evidence. In fact, evidence alone is never enough, which is a
guiding principle of EBM. Instead, guideline developers
need to consider other factors such as the relationship of
benefit to harm, patient values and preferences related to
the question at hand, resource utilization, and issues of
equity, acceptability, and feasibility that are part of an
evidence-to-decision framework Arriving at a common
understanding among urological guideline developers
regarding how to rate the quality of evidence and move
from evidence to recommendations will go a long way
towards improving guideline quality and, ultimately,
promoting high-quality evidence-based care.
Conflicts of interest: The authors have nothing to disclose.
References
[1]
[9_TD$DIFF] .
[2]
.
[3]
.
[4]
.
[5]
.
[6]
.
[7]
ava ilable at
journal homepage:
DOI of original article: .
0302-2838/Published by Elsevier B.V. on behalf of European Association of Urology.