Through the preliminary search, 560 literatures were obtained, and 18 literatures were obtained by reading the title, abstract and full text of the articles and referring to the inclusion and exclusion criteria [12–29]. The total number of cases was 1,648, including 835 in the PVBT group and 813 in the TURBT group. The specific search process is shown in Fig. 1.
According to the original data obtained from the literature, baseline data, such as surgical grouping, age, gender, tumor staging/grading, tumor number, and position, were plotted and compared, and all the studies were consistent with baseline data (Table 1). Tumor staging in different articles at different time period was based on different AJCC staging versions which are 5th, 6th, and 7th, respectively. Nevertheless, the staging of superficial bladder cancer was consistent in these tumor staging criteria. Therefore, the AJCC staging doesn’t affect our results.
Table 1
List of basic characteristics of the incorporated literature
Inclusive study | Number of cases PVBT/TURBT | Average age (years) PVBT/TURBT | Male patient PVBT/TURBT | Female patient PVBT/TURBT | Base-line comparability | Follow-up outcome |
Xu et al. 2015 [29] | 99/94 | 63.06/62.82 | 80/76 | 19/18 | Yes | ①②④⑤⑦⑧⑨ |
Li et al. 2017 [12] | 34/30 | 49.6/50.7 | 20/16 | 14/14 | Yes | ①②④⑤⑦⑧⑨ |
Zhang et al. 2016 [13] | 43/43 | 56.41/52.29 | 28/26 | 15/17 | Yes | ④⑤⑦⑧⑨ |
Wang et al. 2016 [14] | 60/60 | 57.23/57.65 | 33/32 | 27/28 | Yes | ①②④⑤⑥⑦⑧⑨ |
Liu 2015 [15] | 43/43 | 62.3/60.5 | 31/29 | 12//14 | Yes | ①②④⑤⑧⑨ |
Han et al. 2015 [16] | 59/59 | 65.29/65.76 | 32/31 | 27/28 | Yes | ①②⑤⑥⑦⑧ |
Liu et al. 2015 [17] | 31/30 | 53.20/51.93 | 23/21 | 8//9 | Yes | ①②④⑤⑧⑨ |
Shen et al. 2015 [18] | 120/120 | 54.2* | 108* | 132* | Yes | ①②③④⑤⑥⑦⑧⑨ |
Wen et al. 2014 [19] | 30/30 | 56/57 | 23/22 | 7//8 | Yes | ①②④⑤⑥⑦⑧⑨ |
Ge et al. 2011 [20] | 24/24 | 46.5* | 27* | 21* | Yes | ①②③④ |
Luo et al. 2011 [21] | 28/28 | 60.5* | 43* | 13* | Yes | ①②④⑤⑥⑦⑧⑨ |
Cao and Cao 2011 [22] | 46/47 | 56/58 | 25/26 | 21/21 | Yes | ①②④⑤⑥ |
Li et al. 2010 [23] | 35/35 | 65/62 | 26/25 | 9//10 | Yes | ①②④⑤⑥⑦⑧⑨ |
Wang et al. 2009 [24] | 44/51 | 59.71/58.6 | 29/42 | 15//9 | Yes | ①②④ |
Huo et al. 2008 [25] | 35/32 | 56/52 | 26/26 | 9//6 | Yes | ①②④⑤⑥⑦⑧⑨ |
Deng et al. 2008 [26] | 42/42 | 62/58 | 23/24 | 19/18 | Yes | ①②④⑥⑧ |
Liu and Li 2007 [27] | 20/20 | 65.5/62.5 | 15/15 | 5/5 | Yes | ①②④⑤⑥ |
Jiang et al. 2006 [28] | 42/25 | 65* | 57* | 10* | Yes | ①②③④⑤⑧⑨ |
Note: (1) *data is recorded in the index document but not grouped; (2) Outcome indicators: ① obturator nerve reflex; ② bladder perforation; ③ urethral stricture; ④ recurrence; ⑤ time required for surgery; ⑥ surgical bleeding volume; ⑦ hospitalization time; ⑧ catheter indwelling time; ⑨ bladder irrigation time. Abbreviations: PVBT, photoselective vaporization of bladder tumor; TURBT, transurethral resection of bladder tumor. |
The quality of 18 included articles was evaluated according to the modified Jadad scale. Among them, there were 2 high quality documents and 16 general quality documents (Table 2).
Table 2
List of quality evaluation of the incorporated literature
Inclusion study | Generation of random sequences | Randomized hiding | Blind method | Withdrawal and loss of follow-up | Score |
Yansheng Xu,2015 | Computer generated randomly | Sealed envelope | Single blind | Described and processed | 7 |
Yijian Li,2017 | Random test without describing method | No clear | Not described | Not described | 2 |
Jianchao Zhang,2016 | Computer generated randomly | computer control | Not described | Not described | 4 |
Zhancheng Wang,2016 | Random test without describing method | No clear | Not described | Not described | 2 |
Kun Liu,2015 | Random number table method | No clear | Not described | Not described | 3 |
Qianhe Han,2015 | Random number table method | No clear | Not described | Not described | 2 |
Zhifeng Liu,2015 | Random test without describing method | No clear | Not described | Not described | 2 |
Yizhen Shen,2015 | Random number table method | No clear | Not described | Not described | 3 |
Yongan Wen,2014 | Random test without describing method | No clear | Not described | Not described | 2 |
Guangcheng Ge,2011 | Random test without describing method | No clear | Single blind | Not described | 3 |
Bin Luo,2011 | Random test without describing method | No clear | Not described | Not described | 2 |
Shiyi Cao,2011 | Random test without describing method | No clear | Single blind | Not described | 3 |
Jian Li,2010 | Random test without describing method | No clear | Not described | Not described | 2 |
Li Wang,2009 | Random test without describing method | No clear | Not described | Describe the loss of follow-up | 3 |
Lizhi Huo,2008 | Random test without describing method | No clear | Not described | Not described | 2 |
Gang Deng,2008 | Random test without describing method | No clear | Single blind | Not described | 3 |
Weijun Liu,2007 | Random test without describing method | No clear | Not described | Not described | 2 |
Shaobo Jiang,2006 | Random test without describing method | No clear | Not described | Not described | 2 |
Note: 1–3 points are considered general quality and 4–7 points are considered high quality. |
|
There were 14 publications [11, 13–17, 19, 20, 22–24, 26–28] met the criteria of obturator nerve reflex inclusion. The heterogeneity analysis results (I2 = 0, P = 0.72) indicated good homogeneity among various studies. Meta-analysis of them revealed that the obturator nerve reflex in the operation was lower in the PVBT group than in the TURBT group and the result was statistical significant (RR = 0.09, 95% CI [0.04, 0.18], Z = 6.91, P < 0.00001, Fig. 2A). Funnel plot showed no publication bias (Fig. 4A).
Quantitative analysis of 14 researches [15–18, 20–29] uncovered the probability of bladder perforation in the PVBT group was lower than in the TURBT group (RR = 0.14, 95% CI [0.07, 0.28], Z = 5.35, P < 0.00001, Fig. 2B) in a statistically relevant manner. These articles were in good homogeneity (I2 = 0, P = 1.00) and no publication bias was found (Fig. 4B).
Quantitative analysis of the basic data of 13 articles [12–19, 21, 23, 25, 27] (MD = − 2.46, 95% CI [− 5.37, 0.46], Z = 1.65, P = 0.10) indicated that the difference was not statistically significant, and there was no difference in the time required for surgery between the PVBT group and TURBT group (Fig. 2C). There was a strong heterogeneity among included articles for surgical time (I2 = 91%,P < 0.00001) and publication bias was found (Fig. 4C). Sensitivity analysis revealed that the results of removal single study and change effect model are inconsistent, which indicated that the results and conclusions of the index is of poor reliability.
Quantitative analysis of 8 studies [14, 16, 18, 19, 21, 23, 25, 27] (MD = − 17.27, 95% CI [− 24.73, − 9.81], Z = 4.54, P < 0.00001) showed that the amount of surgical bleeding of PVBT group was significantly less than TURBT group (Fig. 2D). Since there was a strong heterogeneity (I2 = 99%, P < 0.00001), sensitivity analysis was performed and no difference between results of removal single study and change effect model was found, suggesting that the analysis of intraoperative blood loss was reliable.
Of the 18 articles included in the study, only three specifically described the urethral stricture after surgery [18, 20, 28]. Quantitative analysis of the data regarding urethral stricture post-surgery in these three papers was performed [18, 20, 28] (RR = 0.53, 95% CI [0.15, 1.83], Z = 1.00, P = 0.32, Fig. 3A). Hence, there was no difference in the incidence of postoperative urethral stricture between the PVBT group and the TURBT group.
Quantitative analysis of the basic data of 17 RCT studies [12–15, 17–30] that met the criteria (RR = 0.52, 95% CI [0.40, 0.67], Z = 5.11, P < 0.0001) demonstrated statistical significance. The rate of recurrence of bladder tumors after PVBT was lower than that of the TURBT group (Fig. 3B). The included studies represented a good homogeneity (I2 = 0, P = 0.66) and no publication bias (Fig. 4D).
Quantitative analysis of the length of hospital stay of the 10 works [12–14, 16, 18, 19, 21, 23, 25, 30] (MD = − 2.80, 95% CI [− 3.82, − 1.87], Z = 5.37, P < 0.00001) illustrated that the length of hospital stay of PVBT group was significantly less than TURBT group (Fig. 3C). The heterogeneity analysis results (I2 = 98%, P < 0.00001) indicated there was strong heterogeneity existed. However, sensitivity analysis suggested the meta-analysis was reliable as there was no difference between results of removal single study and change effect model. Funnel plot showed that the distribution of each point was diffuse and uneven, so publication bias was considered (Fig. 4E).
Quantitative analysis of the underlying data of 12 studies [12–19, 21, 23, 25, 30] (MD = − 2.60, 95% CI [− 3.30, − 1.90], Z = 7.29, P < 0.00001) indicated that the duration of catheter indwelling of PVBT group was significantly less than TURBT group (Fig. 3D). There was a strong heterogeneity depending on the heterogeneity analysis results (I2 = 98%, P < 0.00001). The sensitivity analysis suggested there was no difference between results of removal single study and change effect model. A considered publication bias was shown in funnel plot (Fig. 4F).
Postoperative bladder irrigation time data from the remaining 7 studies [12, 14, 17–19, 21, 25] were analyzed (MD = − 0.95, 95% CI [− 1.49, − 0.42], Z = 3.48, P < 0.0005), and the difference between the PVBT group and the TURBT group in the postoperative bladder irrigation time was statistically significant. That is, postoperative bladder irrigation time was significantly shorter in the PVBT group than in the TURBT group (Fig. 3E). Heterogeneity was considered according to the heterogeneity analysis results (I2 = 99%, P < 0.00001). After sensitivity analysis, no difference between results of removal single study and change effect model was shown, so the analysis was still reliable.
Through analysis, it was found that obturator nerve reflex, bladder perforation, urethral stricture, and tumor recurrence all showed good homogeneity. However, the amount of surgical bleeding, the duration of operation, length of hospital stay, the duration of catheter indwelling, postoperative bladder irrigation time showed greater heterogeneity. The inconsistency of measurement tools, measurement units and measurement accuracy should be regarded as the source of heterogeneity.
Sensitivity analysis of the operation time were inconsistent, so the reliability of its related results and conclusions should be doubted. The amount of surgical bleeding, length of hospital stay, catheter indwelling time and bladder irrigation time all showed high heterogeneity. But after sensitivity analysis, the results were still statistically significant, indicating that the results were reliable.