Eligible Studies and Characteristics
A total of 19 studies with 2651 patients were included in this meta-analysis,1369
underwent EBRT and 1282 underwent CTURBT. The characteristics of the included articles
were presented in Table 1. Of all articles, 4 were RCTs[7, 12-14], 3 were prospective
studies[11, 15, 19] and 12 were retrospective studies[6, 8-10, 16-18, 20-24]. Laser,
“button” shape electrode, loop electrode or HybridKnife were used in the EBRT group.
Loop electrode was used in CTURBT group. Fig.1 summarized the inclusion process. We
performed intravesical chemotherapy or Bacille Calmette-Guerin(BCG) for postoperative patients, mitomycin was used in 5 studies[8, 9,15-17], epirubicin
was used in 7 studies[6,7,10,12,14,23,24], pirarubicin was used in 5 studies[13,18,20-22],
BCG was used in 1 study[11] and BCG combined epirubicin[19] was used in 1 study.
Quality Assessment of Included Studies
We described the level of evidence of the 19 articles in Table 1. According to the
risk of bias graph, 4 RCTs were all considered high-quality studies, with >3 kinds
of bias were at low risk (Fig 2). Furthermore, 11 non-randomised studies [6,8,9,11,15-17,20-22,24]
were considered of high quality due to the score ≥7 stars (Table 1).
Operative time, hospitalization time, catheterization time
12 articles reported the HT, 15 articles reported the AT, and 17 articles reported
the OT. Based on 12 included articles[6-9,12,13,15,16,18,21,22,24], the patients treated
with EBRT had significantly decreased HT compared with CTURBT(P<0.00001).The MD was
-1.30, in favour of EBRT[(random effect) 95% confidence interval [Cl], -1.70 to -0.91;
p<0.00001] (Table 2). Similarly, 15 articles[6-10,12,13,15,16,18,19,21-24] provided
evidence that suggested that the difference in AT was significant between the two
groups[MD, -0.97; 95% [Cl], −1.30 to -0.64; p<0.00001] (Table 2). Furthermore, no
significantly difference was found in terms of OT[MD, -0.56;95% [Cl], −2.16 to -1.04;
p=0.49] [6-10,12-16,18,21-23] (Table 2). While there was a high heterogeneity in all
comparisons, which may come from the patient demographics, difference in types of
surgery and tumor characteristics.
Table 2 The main result of this meta-analysis.
Complications
Where reported, the main intraoperative and postoperative side effects included: obturator
nerve reflex[7-10,12,13,15,17-24], bladder perforation[6-10, 13,15,16,18,20-24], bladder
irritation[6,7,9,16], urethral stricture, bleeding and fever[6,7,9,14-16,20,21,24].
According to the pooled articles, the patients treated with EBRT had significantly
decreased the obturator nerve reflex[OR, 0.12;95% [Cl], 0.07 to 0.19; p<0.00001] (Table
2), bladder perforation[OR, 0.17;95% [Cl], 0.09 to 0.35; p<0.00001] (Table 2), bladder
irritation[OR, 0.21;95% [Cl], 0.14 to 0.32; p<0.00001] (Table 2) and postoperative
complications[OR, 0.40;95% [Cl], 0.20 to 0.82; p=0.01] (Table 2) when compared with
ETURBT. There was no significant heterogeneity among the all comparisons.
The residual tumor on the base
The repeated biopsy in the base was performed by Zhang et al.[14, 18] after the tumor
was resected. 2 pooled articles[14, 18] showed that CTURBT had similar residual tumor
rate compared with EBRT[OR, 0.47;95% [Cl], 0.07 to 3.27; p=0.44] (Table 2).
Bladder detrusor muscle
The bladder detrusor muscle in the specimen was showed in 4 articles[14,18,20,21].
Although there were 94% and 86.9% positive rate in EBRT and CTURBT group, respectively,
no significant difference was found between two groups[OR, 3.59;95% [Cl], 0.6 to 21.63;
p=0.16] (Table 2).
12-month recurrence
The 12-month recurrence rate between groups was compared in 8 studies[6,7,10, 14,
16, 18-20]. The patients treated with EBRT had a lower 12-recurrence rate. No significant
difference was found between groups[OR, 0.77;95% [Cl], 0.55 to 1.07; p=0.12].In subgroup
analyses(Fig 3) according to the different therapy of intravesical instillations (Table
3), no difference was found between groups using epirubicin[OR, 0.92;95% [Cl], 0.61
to 1.37; p=0.68] , pirarubicin[OR, 0.79;95% [Cl], 0.29 to 2.18; p=0.65] or epirubicin
combined with BCG[OR, 0.54;95% [Cl], 0.19 to 1.58; p=0.26] for NMIBC. However, based
on 2 articles using mitomycin, the pooled OR is 0.31 [OR (fixed effect) 95% Cl, 0.10
to 0.93; p = 0.04 ], which indicated EBRT had a lower 12-month recurrence rate than
CTURBT. Subgroup was also performed based on the study type. Table 3 summarized the
results.
24-month recurrence
A total of 1559 patients were included, 762 underwent EBRT and 797 underwent CTURBT.
In meta-analysis, 12 pooled studies[6-11,13,15,16,22-24] showed that CTURBT had a
apparently higher 24-month recurrence compared with EBRT[OR, 0.62;95% [Cl], 0.48 to
0.80; p=0.0003]. Subgroup analyses were conducted based on the different therapy of
intravesical instillations. There were no significantly difference between two groups
in the epirubicin subgroup[OR, 0.71;95% [Cl], 0.45 to 1.10; p=0.13] or pirarubicin
subgroup[OR, 0.65;95% [Cl], 0.36 to 1.16; p=0.14]. However, in mitomycin[OR, 0.61;95%
[Cl], 0.41 to 0.90; p=0.01] and BCG subgroups[OR, 0.24;95% [Cl], 0.07 to 0.84; p=0.03]
(Fig 4), the pooled results showed that patients treated with EBRT had a lower 24-month
recurrence. No significant heterogeneity was showed in all comparisons. Subgroup was
also performed based on the study type. Table 3 summarized the results.
36-month recurrence
4 articles were analyzed for 36-month recurrence as the outcome. 4 pooled studies[7,
14, 16, 20] including 203 patients showed that there were 29.6% and 32.4% 36-month
recurrence rate in EBRT and CTURBT group, respectively, but no significant difference
was found between two groups[OR, 0.72;95% [Cl], 0.37 to 1.39; p=0.32] (Table 3). Our
pooled estimate showed significant heterogeneity(I2=58%), which may come from the Cheng’s study. Because only HybridKnife was used for
EBRT by Cheng, laser was used by the others.
Same site recurrence
5 included articles[11, 13, 18, 19, 23] evaluated the same site recurrence rate with
the follow-up time between 12-38 months. Compared with the CTURBT, EBRT had a lower
same site recurrence rate, while no significant difference was found between two groups[OR,
0.49;95% [Cl], 0.21 to 1.14; p=0.10] (Table 3).
Low-risk, intermediate-risk, high-risk
According to the postoperative pathologic examinations, patients were divided into
the low-risk group[6, 9, 13,14], intermediated-risk[6, 9, 13,14] and high -risk group[9,
13,14]. The main evidence for grouping was tumor size, tumor numbers, tumor category
and tumor grade. No significant difference was found in terms of recurrence rate between
12-36 months in the low-risk group[OR, 1.01;95% [Cl], 0.63 to 1.63; p=0.96], intermediated-risk
group[OR, 0.76;95% [Cl], 0.47 to 1.23; p=0.26] and high-risk group[OR, 0.82;95% [Cl],
0.24 to 2.85; p=0.76] (Table 3).
Publication bias
According to the funnel plots, although a publication bias exists in obturator nerve
reflex. No significant publication bias was detected for our other results.