Ketamine with midazolam outperformed fentanyl with midazolam for radiologist-led sedation

Design

randomised clinical trial, single academic centre, June 2025 to February 2026

Population

264 image-guided lung biopsies, bone biopsies and abscess drainages in 260 adults; median age 68

Primary outcome

maximum intraprocedural pain, 0-10 numeric rating scale

Effect

mean difference -1.4 points (95% CI -2.0 to -0.8); pain >4 in 2.3% against 17%; desaturation <90% in 2.3% against 9.8%; hallucinations 11.4% against 3.8%

Two hundred and sixty-four image-guided procedures - lung biopsy, bone biopsy and abscess drainage, under ultrasound, CT, CT fluoroscopy or combined guidance - were randomised to ketamine with midazolam or fentanyl with midazolam, both given by the interventional radiologist. Median age was 68. The primary outcome was maximum intraprocedural pain on a 0-10 numeric rating scale.

Ketamine reduced maximum pain by a mean 1.4 points (95% CI -2.0 to -0.8). The distribution matters more than the mean: pain above 4 occurred in 2.3% with ketamine against 17% with fentanyl, an absolute difference of 14 percentage points (95% CI 8-21). Respiratory outcomes moved the same way - nadir oxygen saturation was 1.4% higher (0.6-2.2) and desaturation below 90% occurred in 2.3% against 9.8%. Ketamine gave deeper sedation and higher intraprocedural systolic blood pressure. The cost was hallucinations, in 11.4% against 3.8%, though overall comfort, reduced recall and perceived adequacy of sedation all favoured ketamine, and procedural and sedation complications did not differ.

This is one academic centre and it is open to the objection that sedation given by the operator is hard to blind. But the direction is coherent with what is known about both drugs, and the respiratory finding is the one that should carry weight in a unit where an anaesthetist is not present: the regimen that hurt less also desaturated less. An eleven percent hallucination rate needs warning the patient about beforehand, not discovering afterwards.

• Consider ketamine with midazolam rather than fentanyl with midazolam for painful image-guided procedures under radiologist-led sedation
• Warn the patient about hallucinations and dissociative experiences before the procedure, not during recovery
• Anticipate higher intraprocedural blood pressure and deeper sedation - monitoring requirements do not fall
• Keep a documented sedation policy and trained recovery staff; deeper sedation shifts the risk profile even as desaturation falls
• Note this was tested in lung and bone biopsy and abscess drainage, not in every interventional procedure

Why it matters

It challenges the default opioid-benzodiazepine regimen in exactly the setting where no anaesthetist is in the room.

Don't overread it

A single-centre trial in which the operator gave the sedation and assessed the procedure - blinding was necessarily imperfect.

The statistics, in plain English

A mean pain difference of 1.4 points on a 0-10 scale sits around the usual threshold for a difference patients notice, so on its own it is modest. The more persuasive number is the tail: 17% of fentanyl patients had pain above 4 against 2.3% on ketamine, which is where the clinical experience of a procedure actually lives. The desaturation difference (2.3% against 9.8%) has a confidence interval running from 1.9 to 13.3 percentage points - clearly favouring ketamine, but with the size only loosely pinned down at this sample size.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.