Submucosal injection solutions are essential for safe and effective colorectal endoscopic mucosal resection (EMR), yet comparative evidence across available agents is limited. Viscous and hypertonic solutions have been proposed to improve technical performance and resection quality. This study evaluates the comparative efficacy of submucosal injectates using a network meta-analysis (NMA) of complete resection and a structured synthesis of secondary technical and safety outcomes.

Nine randomized trials were included in the study. Three colorectal EMR trials with extractable complete resection (R0/CRR) data contributed to the connected NMA network of normal saline (NS), hyaluronic acid (HA), 50% dextrose (DEX), and indigo carmine–mixed saline (IC) (total 564 lesions). A fixed-effect contrast-based NMA was used to generate risk ratios (RRs). Secondary outcomes, including submucosal elevation durability, injection volume, procedure time, and adverse events, were synthesized from all relevant colorectal trials. 

Primary NMA: Compared with NS, HA significantly improved complete resection (RR 1.21; 95% CI 1.02–1.45), DEX showed comparable efficacy (RR 1.26; 95% CI 0.97–1.65), and IC showed no advantage over NS (RR 0.96; 95% CI 0.89–1.05). Treatment ranking favored HA ≈ DEX > NS > IC.

Secondary Outcomes: Direct comparative evidence demonstrated that both hydroxyethyl starch (HES) and 50% dextrose substantially prolonged submucosal elevation and reduced reinjection volume versus NS. HES also modestly shortened procedure time and reduced EMR-related bleeding in large laterally spreading tumors. HA-based solutions exhibited a bleeding risk comparable to NS, whereas indigo-carmine saline and 0.13% HA did not significantly influence bleeding or procedural safety. In a separate comparison unconnected to the saline-based network, Eleview required significantly less injectable volume and improved Sydney Resection Quotient compared with hetastarch, without clear adverse event differences. Across all colorectal trials, perforation was rare, and no injectate demonstrated a consistent safety liability.

In this colorectal EMR NMA, hyaluronic acid was the only agent that showed a statistically significant improvement in complete/curative resection compared to normal saline. Dextrose 50% showed a similar effect but with less precision. Secondary comparative analyses indicate that HES, dextrose, and Eleview each enhance technical performance by improving cushion durability or lowering injection volume, whereas bleeding and perforation rates remain low across agents. These results show that viscous and hypertonic solutions, especially HA and DEX, are good injectants for clinical use. HES and Eleview also have some technical advantages in certain situations. Additional well-designed randomized controlled trials are required to standardize efficacy and safety assessments across all injectable categories.