Review Article

Cyanoacrylate Injection Compared with Band Ligation for Acute Gastric Variceal Hemorrhage: A Meta-Analysis of Randomized Controlled Trials and Observational Studies

Table 4

Results of subgroup analyses.

VariableNumber of patientsCombined results (95% CI)Model (%)P value for heterogeneity

Hemostasis of active bleeding
Study design
 RCT4 (232)2.64 (1.15, 6.05)Fixed effect0.00.407
 Retrospective2 (167)1.81 (0.59, 5.60)Fixed effect0.00.456
GV type
 Overall1 (30)1.00 (0.06, 17.62)
 GOV1+21 (26)7.80 (1.16, 52.35)
 GOV14 (337)2.34 (1.17, 4.69)Fixed effect0.00.599
Proportion of HCC
 Higher (>median)3 (241)3.87 (1.11, 13.52)Fixed effect0.00.500
 Lower (<median)2 (312)1.89 (0.86, 4.51)Fixed effect0.00.562

GV rebleeding
Study design
 RCT3 (215)0.43 (0.20, 0.92)Random effects56.00.103
 Prospective1 (37)0.28 (0.12, 0.68)
GV type
 Overall2 (157)0.32 (0.19, 0.55)Fixed effect7.10.300
 GOV1+22 (95)0.52 (0.14, 1.99)Random effects70.00.068
Proportion of HCC
 Higher (>median)2 (157)0.32 (0.19, 0.55)Fixed effect7.10.300
 Lower (<median)1 (37)0.28 (0.12, 0.68)

Variceal obliteration
GV type
 Overall2 (157)1.02 (0.54, 1.95)Fixed effect0.00.530
 GOV1+21 (58)0.91 (0.29, 2.88)
 GOV11 (150)0.19 (0.02, 1.66)
Proportion of HCC
 Higher (>median)2 (157)1.02 (0.54, 1.95)Fixed effect0.00.530
 Lower (<median)1 (150)0.19 (0.02, 1.66)

Treatment sessions
Study design
 RCT3 (307)−0.46 (−1.58, 0.66)Random effects95.30.000
 Prospective1 (37)−0.86 (−1.53, −0.18)
GV type
 Overall2 (157)0.07 (−0.63, 0.77)Random effects78.50.031
 GOV1+21 (37)−0.86 (−1.53, −0.18)
 GOV11 (150)−1.51 (−1.87, −1.15)
Proportion of HCC
 Higher (>median)2 (157)0.05 (−0.64, 0.73)Random effects78.40.032
 Lower (<median)2 (187)−0.99 (−1.20, −0.79)Fixed effect0.00.852

Complications (overall)1
Study design
 RCT3 (215)0.58 (0.32, 1.06)Fixed effect0.00.386
 Prospective1 (37)0.93 (0.22, 3.96)
 Retrospective2 (246)0.95 (0.51, 1.79)Fixed effect0.00.804
GV type
 Overall2 (157)0.68 (0.32, 1.41)Fixed effect28.60.263
 GOV1+22 (95)0.56 (0.24, 1.31)Fixed effect0.00.393
 GOV12 (246)0.95 (0.51, 1.79)Fixed effect0.00.804
Proportion of HCC
 Higher (>median)3 (241)0.70 (0.37, 1.35)Fixed effect0.00.485
 Lower (<median)2 (199)0.98 (0.52, 1.84)Fixed effect0.00.938

Complications (ulcers/ulcer bleeding)
Study design
 RCT2 (210)0.92 (0.03, 33.71)Random effects77.90.033
 Prospective1 (37)0.17 (0.01, 3.78)
 Retrospective2 (246)0.30 (0.14, 0.65)Fixed effect0.00.994
GV type
 Overall1 (60)0.18 (0.03, 0.94)
 GOV1+21 (37)0.17 (0.01, 3.78)
 GOV13 (396)0.37 (0.18, 0.76)Fixed effect51.20.129
Proportion of HCC
 Higher (>median)2 (144)0.21 (0.05, 0.86)Fixed effect0.00.759
 Lower (<median)3 (349)0.36 (0.17, 0.74)Fixed effect53.60.116

Complications (infections)
Study design
 RCT4 (365)1.11 (0.56, 2.18)Fixed effect15.40.315
 Prospective1 (37)1.33 (0.25, 7.01)
 Retrospective2 (246)0.72 (0.35, 1.49)Fixed effect0.00.742
GV type
 Overall2 (157)0.88 (0.38, 2.03)Fixed effect0.00.941
 GOV1+22 (95)1.07 (0.37, 3.06)Fixed effect0.00.735
 GOV13 (396)0.92 (0.46, 1.83)Fixed effect54.10.113
Proportion of HCC
 Higher (>median)3 (241)0.89 (0.42, 1.88)Fixed effect0.00.995
 Lower (<median)3 (349)0.98 (0.50, 1.94)Fixed effect55.70.104

GVO: cyanoacrylate injection; GVL: band ligation; GV: gastric varices; GOV: gastroesophageal varices; HCC: hepatocellular carcinoma; RCT: randomized controlled trial; CI: confidence interval; OR: odds ratio.
1Study by El Amin et al. [12] was excluded owing to heterogeneity.