Sphenopalatine block for headache: more bupivacaine buys nothing

Design

randomised, four-armed, factorial dose-finding trial in two emergency departments; unblinded, no sham procedure

Population

220 adults with moderate or severe primary or secondary headache, randomised from 2,494 screened

Primary outcome

sustained relief — mild or no headache by two hours, maintained to 48 hours with no rescue analgesia

Effect

31% (3 ml BL), 34% (3 ml UL), 35% (1 ml BL), 19% (1 ml UL); 3 ml BL vs 1 ml UL difference 13% (95% CI -4% to 29%)

Two emergency departments randomised 220 adults with moderate or severe headache to a transnasal sphenopalatine ganglion block using 0.5% bupivacaine, in a factorial design crossing unilateral against bilateral and 1 ml against 3 ml per side — so patients received between 1 ml and 6 ml in total. The primary outcome was demanding: headache mild or none within two hours and still mild or none at 48 hours, with no rescue analgesia at any point.

No dose-response appeared. Sustained relief was 31% with 3 ml bilateral, 34% with 3 ml unilateral, 35% with 1 ml bilateral and 19% with 1 ml unilateral. Against the smallest dose, neither 3 ml bilateral (difference 13%, 95% CI -4% to 29%) nor 3 ml unilateral (15%, 95% CI -2% to 33%) reached significance, and both intervals sit below the 15 to 20% difference the investigators had defined as clinically important. Minor procedure-related adverse events affected up to 10% per arm.

Two things temper this. The study was unblinded with no sham, so the absolute relief rates carry a placebo component of unknown size — but that limitation affects all four arms equally, which is what the dose comparison relies on. And the recruitment figure is the most quotable number in the paper: of 2,494 patients screened, 1,612 refused because they wanted something other than a nasal route. Whatever the block's efficacy, two-thirds of eligible patients did not want it.

• Use the smaller volume if you offer this block — the larger one did not deliver more sustained relief
• Offer it as one option among several rather than a default; most patients screened preferred another route
• Judge success against a realistic bar: even the best arm gave sustained 48-hour relief to about a third
• Absolute response rates come from an unblinded trial without a sham and are not a clean efficacy estimate
• Consent should mention minor nasal discomfort and local adverse effects, seen in up to 10%

The statistics, in plain English

The confidence intervals for the dose comparisons (-4% to 29% and -2% to 33%) cross zero, so a real benefit from the larger dose has not been excluded — this is an inconclusive comparison rather than proof of equivalence. What makes it decision-useful is that the trial pre-specified 15 to 20% as the smallest difference worth having, and the point estimates sit at or below that threshold with wide intervals. With 44 to 65 patients per arm the study could only have detected a large dose effect, so a small one may exist and would not matter much clinically.

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.