Two paywalls cracked: OFACAR and ESPB versus paravertebral block

Both items had been on the outstanding list for over a week — OFACAR blocked on three separate runs. Europe PMC turned out to serve both abstracts, since abstracts are deposited independently of the paywalled full text. The scheduled run on this date failed to publish (it produced a brief but could not push to the repository), so this entry records what was delivered in conversation instead.

1. OFACAR — opioid-free anesthesia in cardiac surgery

Journal · Published: Anesthesiology 2026;145, first published 22 May 2026 — ~3 months old, sent now because it was blocked on every previous attempt

Randomised, controlled, blinded, multicentre superiority trial across two French tertiary university hospitals, Aug 2021–Dec 2023. 320 adults undergoing elective cardiac surgery with cardiopulmonary bypass, randomised to opioid-free anaesthesia (n=159 — intravenous ketamine, dexamethasone, lidocaine and magnesium sulfate) or opioid anaesthesia with sufentanil (n=161). Primary outcome: a composite of postoperative neurologic, respiratory, cardiovascular or renal complications and/or death within 30 days.

Among 320 participants (mean age 66.4 ± 10.5 yr, 247 men [77.2%]):

• Primary outcome 75.4% vs 84.5% — RR 0.90 (95% CI 0.80–0.99), P=0.049
• Fragility index 1
• Cardiovascular complications 64.2% vs 75.2% (RR 0.86, 0.74–0.99, P=0.031), particularly postoperative myocardial damage (61.6% vs 72.1%)
• Digestive complications 2.5% vs 11.2% (RR 0.22, 0.08–0.65, P=0.007)
• No deaths in the OFA group versus six in the control group (P=0.014)

The authors’ own conclusion: opioid-free anaesthesia “may reduce composite postoperative complications,” but the fragility index means the result “should be considered hypothesis generating and warrant confirmation in larger trials.”

Interpretation. Unusually candid authors, and worth taking them at their word. A fragility index of 1 means a single patient changing outcome flips the primary result to null. Hold that against the mortality signal, which reads dramatically but rests on six events. The digestive-complication difference is the most robust finding and the most mechanistically plausible — it is what you would predict from removing opioids. Note also the 84.5% event rate in the control arm: a composite that common is easy to move without moving anything a patient would notice.

Read the paper · PMID 42190101

2. ESPB versus thoracic paravertebral block after median sternotomy

Journal · Published: British Journal of Anaesthesia 2026;136:687–694, first published 24 November 2025 — ~9 months old

Citation note: the previous run log recorded this as Br J Anaesth 2025;135:764–71. That is wrong, and is probably why the paper stayed unfindable across several runs.

Prospective, patient- and assessor-blinded noninferiority RCT. Adults undergoing median sternotomy for planned cardiac surgery received ultrasound-guided single-shot erector spinae plane block or thoracic paravertebral block, 20 ml of ropivacaine 0.375% per side, under general anaesthesia in the left lateral decubitus position. Primary outcome: pain during forced expiration on the numerical rating scale at 6 h. NCT04546113.

• 74 participants
• At 6 h, mean (SD) NRS 3.5 (2.6) for ESPB vs 3.0 (2.1) for TPVB, 95% CI −0.48 to 1.34
• The upper bound exceeded the predefined 1-point margin, so noninferiority of ESPB could not be demonstrated
• Opioid consumption, NRS at 24 and 48 h, and ICU and hospital length of stay were comparable
• More intraoperative hypertension with ESPB; no serious block-related complications

Interpretation. A failed noninferiority trial is not a demonstration that TPVB is superior. With n=74 the confidence interval is wide enough to contain both “equivalent” and “meaningfully worse,” so the honest reading is that this trial could not answer its question. It sits awkwardly against the 2026 ASA regional analgesia guideline, which gave fascial plane blocks a strong first-line recommendation in open cardiothoracic surgery — that guideline argued from pooled effect sizes, while this is a head-to-head against the older standard failing to clear its bar.

Read the paper · PMID 41290467