Design: multicentre, double-blind, randomised, placebo-controlled trial at 11 Swedish clinics
Population: 370 nulliparous individuals aged 18–31 having a small-diameter levonorgestrel or copper IUD placed
Primary outcome: clinician-adjusted pain during placement on a 100 mm visual analogue scale
Effect: 43.8 mm vs 58.6 mm; mean difference 14.8 mm (95% CI 10.0–19.6), P<0.001; tolerable pain 98.3% vs 91.4% (NNT 15)

Fear of pain is one of the main reasons an intrauterine device is not chosen, and the profession's record on taking that pain seriously is poor. This double-blind trial across 11 Swedish clinics randomised 370 nulliparous women aged 18 to 31 to 10 mL of mepivacaine 20 mg/mL or saline, instilled through a hydrosonography catheter two minutes before placement of a small-tube levonorgestrel or copper device.

Mean pain during placement on a 100 mm visual analogue scale was 43.8 mm with mepivacaine and 58.6 mm with placebo, a mean difference of 14.8 mm (95% CI 10.0 to 19.6, P < 0.001), essentially unchanged after adjusting for the individual clinician (14.2 mm, 9.7 to 18.6). Pain was reported as tolerable by 98.3% of the mepivacaine group against 91.4% of controls (relative risk 1.08, 1.02 to 1.13; number needed to treat 15).

The absolute numbers are worth sitting with. Even with anaesthetic, mean pain was 44 mm — this reduces a substantial pain, it does not abolish it. But it is a cheap, two-minute addition using a catheter most services already stock, and the difference between 91% and 98% reporting the procedure tolerable is the difference between a routine offer and a procedure with a reputation.

Discuss analgesia for IUD placement before the appointment, not at the couch
Consider intrauterine mepivacaine instillation for nulliparous patients as a default rather than a rescue
Record the pain score — services that do not measure it cannot show they have improved it
Do not promise a painless procedure; mean pain remained around 44 mm on a 100 mm scale

Why it matters

It converts a procedure many patients decline out of fear into one that nearly all find tolerable.

The statistics, in plain English

A 14.8 mm difference on a 100 mm scale is usually taken as clinically meaningful for procedural pain, and the confidence interval of 10.0 to 19.6 keeps it meaningful throughout. Adjusting for clinician barely changed it, which suggests the effect is the drug rather than who was doing the procedure. The trial enrolled only nulliparous women having small-tube devices, so it does not speak to replacements or larger devices.