Cold sore antivirals shorten healing by about a day, and nobody knows if that matters

Design

Cochrane systematic review and random-effects meta-analysis, GRADE-rated

Population

22,631 immunocompetent participants across 87 randomised trials of 58 interventions

Primary outcome

healing time and adverse events leading to treatment cessation

Effect

topical aciclovir -0.88 days (95% CI -1.43 to -0.33); topical penciclovir -1.20 days; oral valaciclovir -1.05 days

Eighty-seven randomised trials and 22,631 immunocompetent participants were pooled across 58 interventions for herpes labialis. Most of the alternative therapies the review set out to capture - zinc, honey, photodynamic therapy - produced nothing poolable.

What the main comparisons show is consistent and modest. Topical aciclovir shortened healing by 0.88 days against placebo (95% CI -1.43 to -0.33, I2 = 74%, moderate certainty). Topical penciclovir shortened it by 1.20 days (95% CI -1.80 to -0.60, moderate certainty), and the one head-to-head trial found the two similar. Oral valaciclovir shortened healing by 1.05 days (95% CI -1.40 to -0.70, I2 = 0%). Oral aciclovir, on a single 145-patient trial, showed little or no difference. Adverse events were uncommon and balanced against placebo throughout. No trial compared topical with oral aciclovir.

The honest conclusion is the authors' own: no minimally important difference has ever been defined for cold sore healing time, so a one-day reduction is a number without a yardstick. None of the main comparisons reported pain, lesion severity or patient-rated improvement at all - which is to say the outcomes patients care about were not measured in twenty years of trials. Treatment is cheap and safe; set the expectation at about a day rather than at cure.

• Tell patients to expect roughly one day less of healing, not a shortened episode by much more.
• Topical aciclovir and penciclovir performed similarly in the single head-to-head trial.
• Oral valaciclovir shortened healing; oral aciclovir did not, on limited evidence.
• Start topical treatment at the prodrome - the trials treated early episodes, not established crusts.
• Zinc, honey and photodynamic therapy have no pooled evidence supporting them.

Why it matters

It converts a routinely prescribed treatment from an assumed benefit into a quantified one that patients can decide about.

Don't overread it

Healing time has no defined minimally important difference, so a statistically clear one-day reduction may or may not matter to a patient.

The statistics, in plain English

An I2 of 74% across the seven topical aciclovir trials means they disagree considerably, which is part of why the certainty is rated moderate rather than high. The deeper limitation is not statistical: without an agreed threshold for what counts as a worthwhile reduction in healing time, a confidently estimated 0.88 days cannot be called clinically important or unimportant. Precision and meaning are different things.

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.