EURURO Vol. 72 No. 6

882 1030

effects in contemporary treatments may be slightly lower.
Another discrepancy noted is that while some studies
report worse declines in urinary function after surgery
others report that surgery resulted in more
incontinence but less irritating/obstructive symptoms than
EBRT This variation could be explained by the
use of different PROMs. Studies reporting urinary function
decline after RP use the UCLA-PCI tool, which focuses
primarily on urinary incontinence, while studies reporting
less irritative/obstructive symptoms use the EPIC tool. EPIC
addresses irritative and obstructive voiding symptoms, and
provides a more comprehensive assessment of urinary QoL
. Urinary irritation symptoms are sometimes said to be
worse with EBRT; however, this was not confirmed by
ProtecT trial .
Regarding the effect of AS on QoL, as recently highlighted
there is a lack of data. We were able to identify only one
NRCS including patients undergoing AS, which
similarly to ProtecT trial reported no major perturba-
tions to their cancer-specific QoL.
With respect to BT cancer-specific QoL outcomes, the
best available evidence comes from one small RCT . The
authors compared post-treatment QoL scores for patients
undergoing BT and NSRP with their baseline scores only
(there is no comparison between groups available), and
they reported that at 1 yr, BT had a negative impact on
urinary irritative symptomatology. This result is consistent
among all observational studies that use the EPIC tool,
which addresses irritative–obstructive symptomatology
Unexpectedly, authors also reported that
both BT and NSRP had no significant impact on QoL 5 yr
after treatment. Conversely, the SPIRIT , which directly
compared QoL outcomes for BT and RP found a statistically
significant difference in favour of BT in the urinary and
sexual domains. In that trial though, the small difference in
the overall mean scores in the urinary domain may have a
questionable clinical significance.
As only a small proportion of patients with early-stage
PCa progress to metastatic disease and die from cancer
within 10–15 yr , understanding the long-term impact
of treatment on disease-specific QoL is critical. This
systematic review revealed an important knowledge gap
in the evidence base, as we were able to identify only one
NRCS that reported QoL outcomes at a follow-up of
>10 yr. Interestingly, data from the PCOS showed that
there were no significant differences in the adjusted odds of
urinary incontinence, bowel dysfunction, or erectile dys-
function between RP and EBRT at 15 yr. PCOS provided two
further important observations: firstly, at the end of follow-
up, the prevalence of erectile dysfunction was very high
( 80%) in both treatment arms and secondly, patients had
significant declines in sexual and urinary function over the
duration of follow-up. These observations have also been
reported for patients undergoing RP and watchful waiting
(WW), in the most recent publication of the Scandinavian
Prostate Cancer Group-4 trial, regarding HRQoL outcomes
, for a median follow-up of 12.2 yr. While it would be
difficult to determine whether these declines are a
consequence of treatment, advancing age, or both, data
from the Prostate, Lung, Colorectal, and Ovarian Cancer
Screening trial comparing a sample of screened PCa
survivors with a sample of screened noncancer controls
suggested that these persistent symptoms were due to
treatment .
3.5.2. Strength and limitations of the review
The strengths of this review are the systematic, transparent,
and robust approach taken to examine the evidence base,
including the use of Cochrane reviewmethodology, RoB and
confounding assessment, and adherence to PRISMA guide-
lines. The clinical question was prioritised by a multidisci-
plinary panel of clinical experts, methodologists, and
patient representatives (EAU Prostate Cancer Guideline
Panel), and the work was undertaken as part of the panel’s
clinical practice guideline update for 2017. The inclusion
criteria restricted the review to studies reporting data on
cancer-specific QoL outcomes, measured by validated
PROMs only. This approach ensured a comprehensive
review of the literature, while maintaining methodological
rigour.
It is important for the authors of this review to
acknowledge several limitations. The number of RCTs
providing level 1 evidence is limited (ie, three), two of
which recruited small cohorts. The quality of the evidence
obtained from observational studies is problematic in
relation to high risk of selection, performance, and detection
biases, and the minority of the studies accounting for our a
priori identified confounders. There is large methodological
heterogeneity among studies (as different PROMS being
used to measure the same outcome, along with outcomes
being measured at different time points), as well as
heterogeneity regarding outcome reporting. It should also
be noted that thresholds for clinically meaningful differ-
ences are not available for all cancer-specific PROM
outcomes. Some tools such as FACT-P and the EPIC short-
form questionnaires suggest reference ranges for minimally
important differences in global QoL scores or symptom
subscales (ie, six to 10 points and four to 12 points,
respectively) . This is an important point for
researchers to consider when designing their trials that
they should be cognizant of what a clinically meaningful
difference in their outcomes would be (indeed this is
essential for sample size estimates). For those tools that
cannot provide such information, caution should be taken
before using such questionnaires in research designs, as
interpretation of ‘‘significant results’’ can be challenging.
3.5.3. How does this systematic review compare with other
systematic reviews?
To our knowledge, this is the first systematic review
comparing the impact on cancer-specific QoL of different
primary treatments for men with clinically localised PCa,
using outcomes measured by validated cancer-specific
PROMs only. The most recent systematic review
including mostly single-treatment cohorts is over 4 yr
out of date (searches up to January 2013) and highlighted
the lack of sufficient quality data to make recommendations
to patients about QoL outcomes. Similar were the results of
E U R O P E A N U R O L O G Y 7 2 ( 2 0 1 7 ) 8 6 9 – 8 8 5 882