1.Study Selection and Characteristics
The initial literature search obtained 682 total studies. After removal of 287 duplicates, the rest articles were screened based on inclusion and exclusion criteria. Full texts of 376 unique studies were evaluated and assessed for eligibility. Ultimately sixteen articles[10, 13-27] were selected for our meta-analysis, including thirteen observational studies[13-19, 21-25, 27] and three RCTs(Randomized Controlled Trials) studies[10, 20, 26]. The search strategy is presented in Figure.1.
A total of 953 patients were included in this meta-analysis, 447 patients were performed RFA combined stent placement compared with 506 controls only given stent placement. Among those studies, eight studies[10, 14, 16, 17, 20, 24-26] via ERCP (Endoscopic retrograde Cholangiopancreatography), seven studies[13, 15, 18, 19, 21, 22, 27] used percutaneous transhepatic cholangiography (PTC) to implant stent, ERCP and PTC were mixed used in one research[23]. Types of biliary stents were mentioned in 15 of 16 studies. Metal biliary stents placement were performed in 12 studies[10, 13-16, 19-21, 23-25, 27], among those studies, uncovered self-expanding metallic stents (USEMS) were used in 8 studies[10, 13-15, 18, 19, 24, 27], USEMS and covered self-expanding metallic stents (SEMS) were mixed performed in 3 studies[21, 23, 25]. Plastic stents were chosen in 2 studies[20, 26], both of them were RCTs. Metal and plastic stents were mixed used in 2 studies[16, 22]. The details of selected process were presented in Table.1.
Table 1: study characteristics
Author
Year country
|
RCT
|
Num
|
Age
(Mean ± SD)
|
Male
|
Etiology
|
Method
|
Stent
|
Prior bilirubin
(mean±SD) μmol/L
|
RFA
|
Control
|
RFA
|
Control
|
Liang
2015 China
|
NO
|
76
|
67.5 ± 2.1
|
63.1± 1.6
|
22:26
|
Extrahepatic CCA
|
PTC/ERCP
|
SEMS/USEMS
|
198.4±23.4
|
212.9±18.9
|
Li, T. F.
2015 CHINA
|
NO
|
26
|
53*
|
60*
|
7:8
|
Mixed
|
PTC
|
USEMS
|
287.2± 123.5
|
254.2± 108.5
|
Kallis, Y.
2015 UK
|
NO
|
69
|
68.9± 9.0
|
69.8± 9.9
|
12:24
|
Pancreatic Cancer
|
ERCP
|
USEMS
|
244.1±48.4
|
203.9±149.9
|
Hu
2016 China
|
YES
|
63
|
71.9±11.5
|
71.0±10.7
|
NA
|
Extrahepatic CCA
|
ERCP
|
Plastic
|
153.8±107.8
|
174.5±156.2
|
Wang
2016 China
|
NO
|
36
|
56. 6 ±12.0
|
58.3 ±10.7
|
14:12
|
Extrahepatic MBO
|
PTC
|
USEMS
|
93.1±32.4
|
99.2±38.6
|
Kadayifci, A.
2016 USA
|
NO
|
50
|
65.4±13.1
|
62.4±11.7
|
8:14
|
Mixed
|
ERCP
|
USEMS
|
-
|
-
|
Sampath K
2016 USA
|
NO
|
25
|
73.3*
|
67.3*
|
6:9
|
Hilar CCA
|
ERCP
|
Plastic/Metal
|
-
|
-
|
Dutta, A. K.
2017 UK
|
NO
|
31
|
78*
|
76*
|
7:8
|
Mixed
|
ERCP
|
USEMS/p-SEMS
|
-
|
-
|
Ma
2017 China
|
NO
|
80
|
58.1±12.6
|
55.1±12.2
|
26:27
|
Mixed
|
PTC
|
USEMS
|
253.7±76.1
|
246.2±77.2
|
Wu
2017 China
|
NO
|
71
|
59.2±7.5
|
56.5±8.75
|
26:25
|
Distal MBO
|
PTC
|
SEMS/USEMS
|
246.7±120.1
|
-
|
Yang
2018 China
|
YES
|
55
|
62±7.7
|
64.5±3.4
|
15:18
|
Extrahepatic MBO
|
ERCP
|
Plastic
|
266.8±88.5
|
245.9±76.2
|
Teoh, A. Y.
2018 HK, China
|
YES
|
47
|
67.2±11.9
|
77.2±10.9
|
15:12
|
Distal MBO
|
ERCP
|
USEMS
|
-
|
-
|
Bokemeye A
2019 Germany
|
NO
|
42
|
68±2.0
|
66
|
-
|
Hilar MBO
|
PTC
|
Plastic/ SEMS
|
-
|
-
|
Cui W
2019 China
|
NO
|
163
|
-
|
-
|
-
|
Mixed
|
PTC
|
USEMS
|
-
|
-
|
Buerlein R
2019 USA
|
NO
|
49
|
-
|
-
|
-
|
Hilar MBO
|
ERCP
|
-
|
218.88*
|
123.12*
|
Yu, T.
2020 CHINA
|
No
|
70
|
64.5*
|
64*
|
15:25
|
Mixed
|
PTC
|
USEMS
|
247.7±90.9
|
241.6±81.3
|
*mean the median. RCT, Randomized Controlled Trials. RFA, Radiofrequency ablation combined with stent placement. Control, only stent placement. CCA, Cholangiocarcinoma. MBO: malignant biliary obstruction. Mixed, the etiology includes more than 2 tumor types. PTC, percutaneous transhepatic cholangial drainage. ERCP, Endoscopic retrograde cholangiopancreatography. SEMS: self-expanding metal stents. USEMS, uncovered SEMS. p- SEMS, partially covered SEMS
2.Methodologic Quality and Risk of Bias Assessment
Seven observational studies[14, 15, 19, 22-25] were of high quality and six[13, 16-18, 20, 27] were judged as moderate quality in Newcastle-Ottawa Scale assessment. In three RCTs: two studies[10, 20] could not be assessed bias risk for no mention of experimental methodology, one study[26] was confined as low risk by the Cochrane tool. The overall heterogeneity was P = .12 and I2 = 32% and considered to be low. Egger’s test showed P =.166, meaning there’s no public bias. Risk of bias assessment was provided in Figure.2.
3.Baseline Characteristics for Patients
The baseline characteristics and patient demographics were showed in Table 2. There was no statistical significance between the RFA-treated and stent alone group in gender proportion, age, adjuvant chemotherapy or radiotherapy, types of stent, preoperative bilirubin level and primary stenosis length.
Table 2:baseline characteristics in RFA + Stent group and Stent alone
|
RFA+ Stent
|
Stent
|
P value
|
Gender (male%)
|
59.43
|
57.8
|
.61
|
Age(y)
|
64.58±10.72
|
63.75±10.74
|
.73
|
Chemotherapy (%)
|
44.30
|
47.25
|
0.32
|
Cholangiocarcinoma(%)
|
72.56
|
69.69
|
0.45
|
Metastases (%)
|
36.97
|
38.89
|
.75
|
Metal stent (%)
|
55.80
|
60.97
|
.14
|
Uncovered Stent (%)
|
87.5
|
87.12
|
.87
|
Complications
|
|
|
.75
|
Abdominal pain (%)
|
43.83
|
47.42
|
.70
|
Pancreatitis (%)
|
4.94
|
1.86
|
.34
|
Cholangitis (%)
|
29.01
|
29.81
|
.36
|
Pro-Bilirubin(μmol/L)
|
241.57
±79.0
|
235.60
±83.3
|
.88
|
Initial stenosis length(cm)
|
3.48
|
3.45
|
.58
|
4.Survival Analysis
The pooled median survival time was significantly longer in the intervention groups with RFA (255 days, 95% CI, 237-272) compared with control groups (178 days, 95% CI,168-188, P <0.01). The reconstructed Kaplan-Meier analyses showed improved survival in the intervention groups with RFA (hazard ratio, 1.53; 95% CI, 1.34-1.75, P <.001) (Figure.3). The mean difference between the RFA-treated group and the controls was 54.87 days (95%CI,34.6-75.14, P <.001), favoring patients receiving RFA.
The median time of RFA-treated groups intervened by ERCP was 292 days (95% CI,248-336),exhibiting a prolonged survival than PTC groups (213 days, 95% CI,189-237, P<.001). The reconstructed Kaplan-Meier analyses showed improved survival in the intervention groups with ERCP groups (HR= 1.39; 95% CI, 1.15-1.7). Subgroup analysis revealed the mean difference of the ERCP groups was 87.15 days (95% CI,37.2-137.28, P <.001, Cochran Q test I² =35%, P =.17), the PTC groups was 43.14 days (95% CI,18.73-67.55, P <.001 Cochran Q test I² =49%, P =.08), but there was no significant difference between the two groups, P=0.12. (Figure 4). Since most of the included studies cause MBO as extrahepatic factors, another subgroup analysis was grouped according to different extrahepatic obstruction sites. The result showed that the survival benefit of the hilar cholangiocarcinoma was 129.84 days (95%CI: 59.76-199.92, p<.001), however, there was no survival benefit observed in the distal cholangiocarcinoma group(p=.53) Figure 5.
5.Stent patency
In the studies mentioning stent patency time, the overall heterogeneity was low (Cochran Q test I² =43.0%, P =.12), and there was no publication bias in the Egger’s (P =.9).
Overall patency time was 216 days (95% CI, 197-235) in the RFA groups compared to 156 days (95% CI, 144-168; P< .001) in the controls, showing a great improvement in the RFA combining with stent implantation. The pooled overall patency time from reconstructed Kaplan-Meier analyses showed improved time in patients receiving RFA compared with patients undergoing biliary stent placement alone (hazard ratio, 1.63; 95% CI, 1.35-1.96, P < .001) (Figure.6). The mean difference of patency time between the RFA and control groups was 42.88 days (95% CI,34.02-51.37).
Subgroup analysis was conducted based on the approaches of stent implantation and etiology. Among the subgroup of procedures used for biliary drainage, the mean difference in the PTC was 54.81 days (95% CI 39.89-69.73, P <.001, Cochran Q test I² =0%, P =.42). In the ERCP group, the mean difference was 39.88 days (95% CI 21.51-58.24,p<.001, Cochran Q test I² =16%, P =.30). There was no statistical difference between the two groups (P =.22)(Figure 7). Another subgroup analysis was based on different types of stents. The overall benefit of patency time in the uncovered SEMS was 53.81 days (95% CI 39.76-67.87, P <.001, Cochran Q test I² = 0) was longer than that of the plastic stent group 33.00 days (P= .03).
We divided the studies into 3 groups according to the different etiology of MBO. The three groups were the Cholangiocarcinoma studies, the pancreatic cancer studies, and the mixed studies (Figure 8). Subgroup analysis showed the mean difference of pure- Cholangiocarcinoma groups was 51.09 days (95% CI,24.53-77.66, Cochran Q test I²=62%, P =.05), the pure pancreatic cancer groups were 90.16 days (95% CI,43.19-137.12, Cochran Q test I²=10%, P =.33), the group of the mixed etiology of MBO was 51.77 days (95% CI,35.66-67.87, Cochran Q test I²=0%, P =.72).
6.Adverse Events
Total procedure-related complications referred from each study included abdominal pain, cholecystitis, pancreatitis and hemobilia[20, 23, 28-31],abdominal pain was the most common complications(range from 10%~77.14%), but it did not reach statistical significance between RFA+ stent group and stent alone group(P =.75), the same with the acute cholecystitis (P =.36) and pancreatitis(P =.34). All postoperative pancreatitis only existed in the ERCP-related groups. Cholangitis existed in ERCP group and PTC group.