Permethrin cream beat oral ivermectin for scabies in a trial that ran for six years

Scabies is an infestation by the mite Sarcoptes scabiei affecting an estimated 400 million people worldwide [s2]. It is one of the few common conditions where treatment has drifted toward the more convenient option without a head-to-head trial establishing that the convenient option works as well.

Oral ivermectin is a tablet. Topical permethrin is a cream that must be applied over the whole body, left on, and repeated. A cluster randomised trial published in The BMJ on 6 January compared them directly, and found the cream better [s1].

The design

The trial ran across 28 French hospitals from 19 January 2016 to 16 December 2021 [s1]. Index cases were adults and children weighing more than 15 kg with scabies confirmed by dermoscopy — not by clinical impression alone [s1].

The randomisation unit was the household. Index cases were assigned 1:1 to ivermectin or permethrin, and every member of the household received the same treatment as the index case [s1]. That design reflects how scabies is actually treated: an untreated household member reinfests everyone else.

507 participants in 142 households received ivermectin; 568 participants in 147 households received permethrin [s1]. The primary outcome was clinical cure of the whole cluster at day 28, meaning disappearance of clinical signs and symptoms in every member [s1]. Assessment was by dermatologists masked to treatment allocation [s1].

Both regimens were given on days 0 and 10 [s1].

The result

Cluster-level cure rates were 71.8% with ivermectin and 88.5% with permethrin — a difference of −16.7 percentage points (95% CI −26.3 to −7.1) [s1].

The secondary analyses ran the same way. For index cases, 76.6% against 91.5%, a difference of −14.9 points (95% CI −23.6 to −6.2) [s1]. At the individual participant level, 85.3% against 94.2%, a difference of −9.2 points (−14.9 to −3.5) [s1].

Cutaneous adverse events occurred in 11.9% of participants treated with ivermectin and 15.6% of those treated with permethrin [s1] — the one measure on which the cream came off worse, which is unsurprising for a whole-body topical treatment.

The authors state the outcome precisely. The trial did not show non-inferiority of oral ivermectin compared with 5% permethrin cream; conversely, it showed the statistical superiority of permethrin [s1].

Intraclass correlation coefficients were higher for permethrin than ivermectin across all clusters (0.68, 95% CI 0.61 to 0.75, against 0.46, 0.37 to 0.56) [s1] — meaning outcomes within a permethrin household were more alike, consistent with a treatment whose success depends heavily on whether the household as a whole applied it properly.

Why the setting matters for how far this travels

This trial was conducted in French hospitals [s1]. That is the main constraint on generalising it.

Permethrin's practical disadvantage is delivery rather than efficacy: the trial regimen required enough cream for every household member, whole-body application, and a repeat ten days later [s1]. A tablet given on the same two days is far easier to distribute at population scale. A trial showing permethrin is more effective in hospital-attending French households therefore does not settle which drug is the better tool in a programme treating a whole community at once.

What it does establish is that the efficacy gap is real and reasonably large, which changes the calculation. A 16.7 percentage point difference in household cure is not a rounding error to be traded away for convenience without saying so.

The transmission picture underneath

A study published in Nature Communications in November 2025 provides the epidemiological context that explains why household-level cure matters so much [s2].

It reports that scabies affects 400 million people worldwide and that European incidence is rising, while fundamental epidemiological characteristics have remained unknown [s2]. Using symptom onset data from four outbreaks and surveillance data from the Netherlands from 2011 to 2023, the authors estimated a serial interval — the gap between symptom onset in successive cases — ranging from 98 to 167 days, with a pooled estimate of 123 days (95% CI 91 to 153) [s2].

They estimated an annual growth rate of 0.25 cases per 1,000 people (95% CI 0.20 to 0.30) and a basic reproduction number of 1.09 (95% CI 1.07 to 1.11) [s2]. Transmission peaked in July [s2].

A reproduction number just above 1 is the reason the treatment comparison is consequential. The authors calculate that preventing 8% of secondary infections through public awareness and targeted measures could control epidemic growth [s2]. At an R₀ of 1.09, an epidemic sits close to the threshold, and a 16.7 percentage point difference in household cure rates is well inside the margin that determines whether transmission grows or shrinks.

The serial interval of roughly four months also explains why scabies outbreaks are hard to recognise while they are happening — the chain of transmission unfolds far too slowly to look like an outbreak.

Limits

The trial is one country, hospital-based, and used dermoscopy-confirmed classic scabies, so it says nothing about crusted scabies or about settings where diagnosis is clinical. The epidemiological estimates rest largely on Dutch surveillance and four outbreaks, and the authors themselves call for better transmission data [s2]. Neither paper is a basis for any individual treatment decision.

What to watch

Whether WHO or national guidelines revise their positioning of oral ivermectin as a first-line individual treatment, and whether a comparable trial is run in a mass drug administration setting where the logistics, rather than the efficacy, decide the outcome.

Sources

1. [s1] Oral ivermectin versus 5% permethrin cream to treat children and adults with classic scabies: multicentre, assessor blinded, cluster randomised clinical trial. BMJ, 6 January 2026. https://doi.org/10.1136/bmj-2025-086277
2. [s2] Estimation of the epidemiological characteristics of scabies. Nature Communications, 26 November 2025. https://doi.org/10.1038/s41467-025-65544-y