More bupivacaine does not make a better sphenopalatine block

Design

randomised, unblinded, four-arm dose-finding study across two emergency departments, no sham procedure

Population

220 patients randomised (2,494 screened) with moderate or severe primary or secondary headache

Primary outcome

sustained headache relief — mild or none within 2 hours, maintained 48 hours with no rescue analgesia

Effect

3 mL bilateral 31%, 3 mL unilateral 34%, 1 mL bilateral 35%, 1 mL unilateral 19%; differences vs 1 mL unilateral 13% (95% CI −4 to 29) and 15% (−2 to 33)

Transnasal sphenopalatine ganglion block is used for headache in emergency departments with little agreement on dose or laterality. This four-arm randomised study across two departments assigned 220 patients with moderate or severe headache to unilateral or bilateral block and to 1 mL or 3 mL of 0.5% bupivacaine per side — so total doses ranged from 1 mL to 6 mL. It was unblinded, with no sham.

The primary outcome was demanding: headache reduced to mild or none within two hours and held there for 48 hours with no rescue analgesia at any point. Rates were 31% for 3 mL bilateral, 34% for 3 mL unilateral, 35% for 1 mL bilateral and 19% for 1 mL unilateral. Against the 1 mL unilateral arm, neither high-dose arm produced significantly more relief; the confidence intervals for the differences both crossed zero. Between 58% and 73% of patients in every arm said they would want the same treatment again.

The recruitment figure deserves its own line. Of 2,494 patients screened, 1,612 — 65% — declined because they did not want anything given up their nose. That is not a footnote; it is the main constraint on this procedure in practice, and it dwarfs any dose question.

• Use the smallest dose — 1 mL of 0.5% bupivacaine per side — since larger doses gave no additional benefit.
• Offer it, but expect most patients to decline the nasal route; have an alternative ready rather than a negotiation.
• Sustained 48-hour relief without rescue is a hard endpoint, and roughly a third achieving it is not a poor result for a procedure with minimal systemic effect.
• Minor procedure-related adverse events occurred in up to 10% per arm — tell the patient to expect nasal discomfort and a bitter taste.
• The trial was unblinded with no sham, so the absolute rates include whatever a procedure contributes by being a procedure.

Why it matters

Dose escalation is the reflex when a block underperforms, and here it changed nothing.

Don't overread it

Unblinded, with no sham arm — this compares doses against each other, and says nothing about whether the block beats placebo.

The statistics, in plain English

This was designed to detect a difference of 15 to 20 absolute percentage points, and the observed differences against the reference arm (13% and 15%) came with intervals of −4% to 29% and −2% to 33%. So the trial neither found a difference nor excluded a clinically important one — with about 50 patients per arm, it could not. Note too that the lowest rate came from the smallest total dose (1 mL unilateral, 19%) while the highest came from a mid dose (1 mL bilateral, 35%), which is not a dose-response pattern and is what noise across four small arms looks like.

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.