Two meta-analyses published weeks apart disagree about whether robotic gallbladder surgery injures more bile ducts. Read the confidence intervals

Robotic cholecystectomy has moved from novelty to routine option without ever passing the test that would establish it is better. Its share of all gallbladder removals rose from under 1 percent to between 3 and 26 percent depending on the region by 2024, while laparoscopic surgery still accounts for 85 to 95 percent of the global total.

In the space of about six weeks either side of the new year, two systematic reviews attempted to say whether the machine is safer. They reached opposite conclusions on the outcome that matters most.

The disagreement, stated plainly#

The first concluded that robotic cholecystectomy is a safe and effective alternative with equivalent complication profiles. The second concluded that it is associated with an increased rate of bile duct injury and that its safety profile needs clarifying. Same procedure, same question, overlapping literature, and headlines that point in opposite directions.

Why both can be right#

The resolution is in the interval rather than the estimate. The larger review's confidence interval for bile duct injury runs from 0.43 to 4.66. That range is wide enough to contain a substantial protective effect, no effect at all, and the tripled risk the second review reported. A result like that does not say the two procedures are equally safe. It says the analysis could not tell.

This is the ordinary shape of evidence about a rare event. Bile duct injury happens in well under 1 percent of operations, so even 1.77 million patients yield few enough events that the pooled estimate wobbles. Reporting the odds ratio without the interval converts an inconclusive result into a reassuring one, which is how a review that could not answer the question gets cited as having answered it.

What both reviews agree on#

• Conversion to open surgery is less likely with the robot: odds ratios of 0.35 (0.31–0.41) and 0.44 (0.32–0.61).
• Length of stay is essentially the same, at 1.4 to 2.7 days in the newer review and a pooled difference of 0.02 days in the older one.
• Overall and postoperative complication rates are not meaningfully different, at odds ratios of 0.96 and 0.99 in the older review.
• The robot costs more: roughly 5,000 to 6,000 US dollars per case against 2,000 to 3,000, and 2,088 euro against 1,726 in European centres.

Operative time is the one practical measure that splits by geography. Western centres in the newer review took longer with the robot, 75 minutes against 60; some Asian institutions reported the reverse, 22 minutes against 33. The older review's pooled figure was 12.65 minutes longer for the robotic approach.

The Korean evidence is small and careful#

A team at Korea University College of Medicine compared single-port robotic cholecystectomy against conventional laparoscopic surgery in patients operated on between November 2020 and April 2023. From 1,184 patients they matched 340 pairs on age, sex, body mass index, ASA score, stones, polyps, acute cholecystitis, diabetes, hypertension, preoperative ERCP and — unusually and importantly — the operating surgeon.

Operative time was significantly shorter in the robotic group, 40.0 minutes against 45.0. Their conclusion is deliberately modest: comparable safety and feasibility, with prospective studies needed before broader claims.

Korea also has a domestically developed surgical robot, the Revo-i. The published gallbladder experience with it, from Yonsei University College of Medicine, is a case series of nine patients, all completed without complication or conversion. That is a feasibility report and its authors describe it as one.

What would actually settle this#

1. Both reviews draw heavily on national administrative databases, where an operative approach and a complication are both codes rather than observations, and where the robotic cohort is weighted toward the years when the technique was new.
2. Patients are not randomised to a robot. Where the newer review found a technical advantage — obesity, complex anatomy — is also where surgeons choose the robot, so selection runs in both directions and is not removed by pooling.
3. The Korean comparison is one centre and 340 matched pairs, which is enough to detect a difference in operative time and nowhere near enough to detect a difference in bile duct injury.
4. A trial large enough to resolve a difference between 0.23 and 0.72 percent would need tens of thousands of randomised patients. Nobody has run it, and the cost difference means nobody is obviously motivated to.