EURURO Vol. 72 No. 4

601 656

mesothelial cells (1/34). Stem cells, mostly derived from
adipose tissue, bone marrow or human umbilical cord, were
used in 10 studies.
3.2.2. Clinical studies
Study characteristics of clinical studies are listed in
(see Supplementary data 4 for references of listed studies).
Clinical studies were performed with males, except for one
study ). From 23 studies, 16 used an inlay approach,
two a full circumferential procedure, one used both
approaches, while in four studies the procedure was
unclear. The majority of studies (21/23) used decellularized
templates, while natural and synthetic templates were both
used once. Four studies used cell-seeded templates; 2/23
buccal mucosa keratinocytes and/or fibroblasts and 2/23
bladder smooth muscle cells and/or urothelial cells.
3.3. Quality assessment
The quality of reporting was assessed for 63 preclinical and
13 clinical studies from which outcome data could
sufficiently be extracted for inclusion in the meta-analysis
). Results per study are listed in Supplementary data
5. Reporting of information regarding included animals/
patients, such as species and strain, sex, number of animals/
patients, age/weight, and patient inclusion criteria, were
generally well described.
Overall quality of the experimental setup was poor.
Although the different experimental groups were well
described, hardly any control groups were present, and
randomization and blinding were seldom mentioned in
both preclinical and clinical studies. Also, clinical study
protocols were not published. However, surgical procedure,
composition, size, and preparation of the implants were
clearly described in most studies. Reporting of outcome
measures was good for both preclinical and clinical studies
with respect to the description of outcome measures,
follow-up time, and side effects. The number of drop-outs
was clearly mentioned in clinical studies, but only in half of
the preclinical studies. For preclinical studies, histological
sampling location and representativeness of the results
were poorly described.
3.4. Meta-analysis
3.4.1. Preclinical studies
For full circumferential repair A), the addition of cells
significantly reduced the probability of side effects,
independent of the type of biomaterial used (p = 0.001).
Exact point estimates including CI are given in Supplemen-
tary data 6. Regarding the type of biomaterial, when no cells
were used, estimates show that synthetic materials had a
higher probability for having no side effects compared to
decellularized and natural materials. With cells seeded,
estimated probabilities were similar for all materials. For
functionality and study completion, estimated probabilities
were similar for all study conditions.
For inlay repair ), the addition of cells signifi-
cantly reduced the probability of side effects (p = 0.003),
albeit less than for full repair. Estimated probabilities were
similar for all types of biomaterial regardless of the addition
of cells. For functionality and study completion, estimated
probabilities were similar for all study conditions. It was
impossible to estimate study completion probability per
biomaterial as almost all animals survived inlay repair
(statistical model did not converge).
Although estimated probabilities for dogs and rabbits
were slightly different, differences were not statistically
significant. Consequently, the animal species had only
marginal influence on outcome (data not shown).
3.4.2. Clinical studies
For clinical studies, a similar meta-analysis was performed
C). Only inlay repair using decellularized materials
with or without cells could be analyzed due to the limited
number of other combinations. No statistically significant
differences were found for the inclusion of cells for any of
the outcome measures (p = 0.5 for side effects, p = 0.7 for
functionality, and p = 0.08 for study completion).
When comparing preclinical and clinical estimated
probabilities, point estimates for absence of side effects
after inlay repair seem to be higher in clinical studies for
both acellular and cellular templates. For functionality, the
point estimates were similar. The estimated probability for
study completion was much lower in clinical studies
compared to preclinical studies regardless of the addition
of cells, but these cannot be directly compared due to
distinctive definitions for study completion and differences
in disease status.
4. Discussion
Reconstructive surgery using biomaterials has been studied
as an alternative approach for urethral repair since the early
seventies and efforts along these lines expanded rapidly in
the nineties when the term ‘‘Tissue Engineering’’ was
introduced ) . Nowadays, preclinical studies have
been readily performed, but clinical studies have not
followed this trend. Although many (pre)clinical studies
have been performed, tissue engineering is not used as an
alternative treatment in routine clinical practice, except for
a select patient group with a history of failed repairs
. In this systematic review, all (pre)clinical publications
on urethra tissue engineering until June 2016 were ana-
lyzed to assess the evidence for the efficacy. For clinical
studies, the term ‘‘effectiveness’’ may be more suitable, as
most studies included in this review showed a heteroge-
neous patient population . However, we used the term
‘‘efficacy’’ for preclinical and clinical studies throughout this
systematic review. For both preclinical and clinical studies,
tissue engineering had a high probability for functionality,
defined as voiding with continence. Study completion was
high in preclinical studies, but not in clinical studies. This
may be related to the difference in our definition of study
completion and in study design. In preclinical studies,
animals generally only need to survive for several months to
study the tissue regeneration process, compared with
patients that need to show a good long term outcome
E U R O P E A N U R O L O G Y 7 2 ( 2 0 1 7 ) 5 9 4 – 6 0 6 601