775 860
URS), sex (for children only), type of URS (ie, semirigid vs
flexible), stent versus no stent prior to treatment (for both
interventions), and number of SWL/URS sessions. Stone
size, BMI, and stone composition were considered the most
likely to influence outcomes and are reported in the risk of
bias graph.
2.7.
Quality of evidence assessment
The Grading of Recommendations Assessment, Develop-
ment, and Evaluation (GRADE) tool
[19]was used to assess
the quality of evidence for URS versus SWL. Quality of
evidence for
critical
and
important
outcomes for decision
making were rated on study design, risk of bias, directness,
consistency, and precision
[19]. The SFR outcome was rated
as
critical
for decision making. Four weeks were considered
to be the most appropriate time point on which to GRADE
the quality of evidence because it is a fair trade-off between
an
immediate
SFR in the URS arms and allows for a
package
of care approach in the extracorporeal SWL arms. Clavien
Grade 3, and need for secondary procedure were rated as
important
for decision-making.
2.8.
Data analysis
For dichotomous benefit or harm outcomes (eg, SFR or
complications), we report risk ratios and 95% confidence
intervals (CIs) in forest plots but we did not meta-analyse
these estimates due to heterogeneity in one or several of the
following factors: the study designs, the intervention
schedules and utilisations in both the experimental and
control groups, and the outcome definitions and time-
points of measurement. In the results tables, we report
categorical data (SFR) at available time points up to 3 mo
and reported
p
values where available. For other categorical
outcomes (eg, adverse events) we report proportions. For
continuous outcomes, we report mean difference with
standard deviation and/or range and corresponding 95% CIs,
where available.
The primary analysis was per participant randomised.
For studies with more than two intervention groups, only
the intervention groups relevant to the review were
selected, or where possible, groups were combined to
enable a single pair-wise comparison.
We conducted an intention-to-treat analysis, if data
were available; otherwise an available case analysis was
performed. We did not impute missing data. In the case of
incompletely reported data, attempts were made to contact
the relevant authors where possible.
3.
Evidence synthesis
3.1.
Search results
The study selection process is outlined in the Preferred
Reporting Items for Systematic Reviews and Meta-analyses
[15]diagram (Supplementary Fig. 1). The initial search
returned 5380 abstracts of which 387 were scrutinized for
eligibility. No study on children fulfilling the inclusion
criteria was found. A total of 47 studies were eligible for
final inclusion; 22 of these compared URS with SWL. Of
these, 18 were full text articles
[5,9,11,20–34]and four were
conference abstracts
[35–38] .3.2.
Study and patient characteristics
Of the 22 final studies comparing SWL with URS, four were
RCTs
[11,20,21,35] ,one was a QRCT
[22]and 17
[5,9,23– 34,36–38]were comparative NRSs. Of the final 25 included
comparative intragroup (intra-SWL and intra-URS) studies,
24 were full text articles
[7,39–61] ,and one was a
conference abstract
[62]. There was heterogeneity in study
designs and outcomes measured. Details of all studies and
baseline characteristics of the participants in the included
studies are outlined in Supplementary Table 1.
3.3.
Risk of bias assessment of the included studies
The risk of bias assessment for RCTs and NRSs can be viewed
in Supplementary Figure 2 and Supplementary Figure 3,
respectively. In general, there was a low risk of bias for
intergroup RCTs/QRCTs, low to moderate risk of bias for
intragroup RCTs/QRCTs, while for both intergroup and
intragroup NRSs there was a moderate to high risk of bias
for included studies. As for confounding factors BMI, stone
size, and stone composition the NRSs had in general high to
very high risks of bias.
3.4.
Comparisons of interventions results
Principal results can be viewed in
Table 1, Supplementary
Table 2, and in the Forest plots in
Figs. 1 and 2 .3.5.
Evidence synthesis for intergroup comparative studies
3.5.1.
SFR SWL versus URS
All included studies reported SFRs; however, the time points
at which these were measured were heterogeneous. ‘‘Stone-
free’’ was defined for the purposes of this review as no stone
fragments remaining at the reported time point. Stone-free
status was measured immediately in four studies
[11,29,31, 33], at 1wk inone study
[20] ,at 2wk in two studies
[21,22] ,at
3 wk in one study
[9] ,at 4 wk in four studies
[24,25,30,32], at
6wk inone study
[27], at 3mo in six studies
[5,9,24,28,30,35],
and was unclear in six studies
[23,26,34,36–38]. In three of
the studies two different SFR time points were measured
within each study
[9,24,30] .As summarised in
Fig. 1 ,achievement of stone-free
status favouring URS reached statistical significance in nine
studies
[22,24,25,29–33,36]of which four were significant
at 4 wk
[24,25,30,32] .In the remaining 13 studies
[5,9,11, 20,21,23,26–28,34,35,37,38]no significant difference was
found between SWL and URS for the treatment of proximal
ureteric stones. There was significant heterogeneity in the
lithotripter devices used, their power settings, the modali-
ties used for SFR assessment, and the number of shock
waves delivered among those patients treated with SWL.
The SFR also varied widely in the studies reporting on
E U R O P E A N U R O L O G Y 7 2 ( 2 0 1 7 ) 7 7 2 – 7 8 6
775