Ten RCTs (4 full reports, 6 abstracts) with 653 participants were included.
The methodological quality of the studies was variable, but studies published as full reports achieved a better score.
Primary haemostasis (3 full reports and 3 abstracts): this was only assessed in patients with active bleeding at the time of randomisation. There was no significant difference in rates of primary haemostasis between the 2 treatments (OR: 1.295% CI: 0.5, 2.8), test for heterogeneity non-significant.
Rebleeding (3 full reports, 5 abstracts): the rebleeding rate was significantly higher following sclerotherapy compared to banding ligation (OR: 1.695% CI: 1.1, 2.4: p<0.05), test for heterogeneity non-significant.
Early mortality (3 full reports, 5 abstracts): no significant difference between the 2 treatments (OR: 1.395% CI: 0.8, 1.9), test for heterogeneity non-significant.
Variceal eradication (4 full reports, 2 abstracts): no significant difference between the two treatments (OR: 0.895% CI: 0.6., 1.1), test for heterogeneity non-significant.
Complication rates (4 full reports, 5 abstracts): complications rate was higher after sclerotherapy (OR: 2.695% CI: 1.8, 3.9: p<0.05), however, there was significant between-study heterogeneity (p<0.01).
Number of treatment sessions required for eradication of varices was less with endoscopic banding ligation (3.8 +/- 1.6 versus 5.8 +/1 2.2: p<0.05).
The subgroup analysis of full reports only was not different from the main meta-analysis.