Backonja 2008.

Study characteristics
Methods	RCT, DB, multicentre, parallel groups, single application, 12‐week duration. Patch applied to painful area, up to 1000 cm2 Oral pain medication continued without change. Transdermal opioids (morphine equivalent ≤ 60 mg/day) permitted, but not topical analgesics Rescue medication: after application participants allowed hydrocodone/paracetamol (5/500 mg) for ≤ 5 days Pain assessed daily (average pain for last 24 hours). PGIC assessed at endpoint. Clinic visits at 4, 8, 12 weeks
Participants	Postherpetic neuropathy with at least moderate pain, ≥ 6 months since vesicle crusting. Exclusion: pain in/around facial area N = 402 M = 190, F = 212 Mean age: 71 years Baseline pain: 30 mm to 90 mm (mean 60 mm)
Interventions	(1) Capsaicin patch 8%, n = 206 (2) Control patch, n = 196 Topical local anaesthetic applied for 60 min, then patch applied for 60 min Control patch contained 0.04% capsaicin to mimic AEs
Outcomes	PI: 11‐point numeric pain rating scale (responder: ≥ 30% and ≥ 2‐point reduction from baseline) PGIC: 7‐point scale (responder: much and very much improved) AEs Withdrawals
Notes	Oxford Quality Score: R1, DB2, W1. Total = 4/5
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not described
Allocation concealment (selection bias)	Low risk	Remote treatment assignment, using unique number on printed labels affixed to outside of patch envelope
Blinding (performance bias and detection bias) All outcomes	Low risk	Low concentration of capsaicin in "identically formulated" control patch to mimic local skin reaction of active treatment
Incomplete outcome data (attrition bias) All outcomes	Low risk	Modified (no details) LOCF analysis for primary outcome, but no imputation for weekly scores. All participants included for safety analysis
Size	Low risk	206 participants in capsaicin arm,196 participants in control arm