MUMultiple speciesBmoderateT1hypothesis only
Of 107 untreated sarcoids, 1.9 per cent regressed: what links bovine and human papillomaviruses, and where the comparison ends.
Executive summary
The equine sarcoid counts as the most frequently reported skin tumour of the horse, and it traces back to a cattle virus foreign to the species. In one vaccination trial every unvaccinated control horse developed lesions after inoculation, while almost all vaccinated horses stayed clear. The comparison with human papillomaviruses explains the mechanism well. It ends where vaccination programmes, cancer progression and the choice of a treatment begin.
Key points
The equine sarcoid counts as the most frequently reported skin tumour of the horse. Behind it stands a virus that belongs to a different animal species: the bovine papillomavirus. A review from 2003 established what still holds today: most sarcoids contain detectable viral genetic material and produce the main transforming viral protein, known as E5. An appraisal from 2026 names types 1 and 2 and probably type 13, together with a skin injury and a genetic predisposition.
Clinically, six forms are distinguished: occult, verrucous, nodular, fibroblastic, mixed and malignant. A laboratory study from 2024 expressly describes the sarcoid as a skin tumour that does not metastasise: it grows in place and does not spread to other organs.
That the virus triggers the lesion rather than merely accompanying it is shown by a trial from 2017. After immunisation, ten sites on the neck were inoculated with viral particles in every horse. All seven unvaccinated control horses developed a lesion at every single site, while 13 of 14 vaccinated horses stayed clear. These were, however, pseudo-sarcoids, produced by a very large quantity of viral particles, and so an artificial challenge.
The route of transmission remains unexplained. One laboratory study exposed 420 stable flies to sarcoid tissue or bovine papilloma tissue: the viral load on the flies rose measurably, was higher and detectable for longer after contact with the cattle tissue, and fell back to the starting level within a few days. Transmission to a horse was not attempted.
Papillomaviruses usually multiply in maturing skin cells and release finished particles. In the sarcoid this apparently does not happen: the 2003 review found viral genetic material and messenger RNA, but no sign of infectious viral particles. This state is called an abortive infection. The virus is present, it redirects the cell, and it never assembles completely.
Why the immune system does not clear the transformed cells was studied in a cell model. In horse fibroblasts additionally made immortal with human telomerase, the protein E5 suppressed the major histocompatibility complex of class I on two levels: the reading of the heavy chain was slowed, and the finished complex was held back in the Golgi apparatus. Without that marker on the cell surface, defence cells struggle to recognise an altered cell as altered.
The obvious step would be to read clinical behaviour off viral activity. A study of 27 sarcoids and five samples of healthy skin found the opposite: the number of viral copies did not differ between the subtypes and was even higher in horses with fewer tumours. Differences did appear in individual viral genes and in inflammatory messengers. The authors put it cautiously: their results point to differences between tumour formation driven by bovine and by human papillomaviruses.
A Belgian study from 2008 examined 42 horses in four groups. In healthy, unchanged skin, viral genetic material was found in 57 per cent of all animals: in 73 per cent of the affected horses and of those in contact with cattle, in 50 per cent of the animals living alongside affected horses, and still in 30 per cent of the control group with no known contact. In blood, nothing could be detected.
The authors conclude on a latent infection, widely distributed in the horse population. Carrying the virus is therefore ordinary, developing a tumour is the exception. Anyone asking after the cause of the sarcoid is therefore asking not only about the virus, but about what is added in the individual horse.
How strongly such a result depends on the method of measurement is shown by a study of 58 semen samples from clinically healthy stallions. The very sensitive droplet PCR found 34 positive samples, conventional quantitative PCR only five in the same samples. Type 14 was detected in the horse for the first time in the process. Detecting genetic material implies neither a virus able to multiply nor a route of transmission.
In the tumour tissue itself, multiple infections are the rule. Across 93 sarcoids examined, type 1 led with 83 lesions, 66 of these tumours additionally contained genetic material of type 2, six a third and one a fourth viral species. Whether such a double infection goes with a more severe, more treatment-resistant course, the authors consider possible and call for further study; it is not established.
As early as 1995, a Swedish study of 120 affected horses described an individual predisposition linked to leucocyte antigens: certain markers went with an early onset, others with more frequent recurrences after surgery. The first signs appeared on average at three and a half years, and the preferred site was the lower abdominal wall.
Not as a single trait. What was found is a predisposition: in a case-control study, a variant in the region of a gene of the major histocompatibility complex was associated with the sarcoid. The Swedish study of 120 horses had already described differences in leucocyte antigens. Both are associations within particular populations, not a mode of inheritance.
That case-control study from 2016 compared 82 affected with 270 unaffected horses across the whole genome. Of six initially striking regions, only two survived the stricter follow-up analysis: one in a non-coding section of a gene of the major histocompatibility complex class II, the other outside that complex. In this sample the first variant went with roughly a fivefold higher chance, the second with roughly a fourfold one. The authors set the finding beside known links between variants of the same complex and papillomavirus-driven tumours in humans and rabbits. The host thus stands alongside the virus, yet such a clustering describes a particular population and not a risk for an individual horse.
For decades the sentence held that around a third of untreated sarcoids disappear on their own. It comes from the Swedish series of 1995 and counts horses, not tumours: in some of the untreated animals the changes were no longer present later on.
A 2026 appraisal contradicts that reading sharply: spontaneous regression is said to be an exceptional event, confined to mild, occult and at most verrucous forms. The argument is not a rival figure but an objection to the data base. Series with many regressions rest on clinically made diagnoses without histological confirmation, and so without certainty that these were sarcoids at all.
The hardest figure comes from a prospective study with 25 horses and 164 tumours. Of these, 107 were left untreated and followed for an average of 34 months. Two of them regressed, that is 1.9 per cent. This figure has a catch of its own: the untreated tumours sat on horses whose other lesions were being treated locally. That was not a control group set up as such.
The contrast with the experimental model remains striking: the pseudo-sarcoids produced with a very large quantity of viral particles regress within weeks to months. The naturally arisen sarcoid almost never does. Why, is open.
In 2024 the first and so far only systematic review of sarcoid treatment appeared. After a structured search of five databases, ten intervention studies remained, with regression rates of 28 to 100 per cent per lesion and of 9 to 100 per cent per horse. In only 60 per cent of the studies were all lesions confirmed histologically, and the risk of bias ranged from some concerns to critical.
The conclusion is unusually plain: there is not enough evidence to recommend one procedure over another. The authors regard randomised, placebo-controlled studies of adequate size as urgently needed. Anyone who finds a ranking of sarcoid treatments online finds something the literature does not supply.
Orders of magnitude from practice come from a Belgian series over six years with 230 animals and 614 sarcoids. Overall success stood just short of 75 per cent, counted as complete regression without recurrence for at least six months; electrosurgical removal reached almost 87 per cent, and horses with several tumours failed more often. Because the choice of procedure depended on the lesion, the treatments are not comparable with one another.
The Swedish series gives a recurrence rate of around 35 per cent after the first attempt at treatment, most of them within four months. An open study of two topically applied agents reported complete regression in 84.4 and 75.0 per cent of the treated tumours, with recurrence rates of 7.3 and 21.4 per cent. Without random allocation and without blinding, the comparison stays open.
How narrow the data base is, a Swiss example shows. In 2010 a randomised study of a mistletoe extract appeared, in 53 horses with 444 sarcoids: in 13 of 32 treated horses, lesions regressed fully or in part, against 3 of 21 control horses. The authors called the extract a safe and effective treatment of the clinically diagnosed sarcoid. The hardest criterion, the complete disappearance of individual lesions, already fell short of statistical confirmation at the time.
In 2024 the same substance was tested by a randomised, double-blind, placebo-controlled study run by the same institutions. Forty-five horses were spread across three groups, treated for seven months and observed for a further seven. No statistically secure difference between mistletoe and placebo was found, neither for administration by mouth nor for administration under the skin. Oral administration was well tolerated.
| Feature | Study of 2010 | Study of 2024 |
|---|---|---|
| Design | Randomised, control group, administration under the skin | Randomised, double-blind, placebo, administration by mouth and under the skin |
| Size | 53 horses with 444 sarcoids | 45 horses in three groups |
| Observation period | 12 months | 14 months |
| Fully or partly regressed | 13 of 32 treated against 3 of 21 control horses | 6 of 15 by mouth, 4 of 14 under the skin, 4 of 16 on placebo |
| Completely regressed | 27 of 95 against 9 of 68 lesions, difference not secure | 4 of 15, 3 of 14 and 2 of 16 horses |
| Conclusion of the authors | Safe and effective treatment of the clinically diagnosed sarcoid | No secure difference between mistletoe and placebo |
The case is instructive in both directions. The newer study does not refute the older one: with 45 horses across three groups, no effect and a clear effect remain equally compatible with the data. A missing difference is no evidence of a missing effect. What does become visible is the price of the missing blinding in the older work.
In humans, papillomaviruses are the best-studied case of a virus that causes tumours, and several building blocks recur in the horse. In both cases viral oncoproteins redirect the host cell, and in both cases the major histocompatibility complex is in play: the genetic study in the horse points expressly to the corresponding findings in humans and rabbits.
On the clinical side the difference is vast. A Cochrane review from 2018 brought together 26 randomised studies with 73,428 participants and rates the certainty of this evidence as high: vaccination lowers the precursors of cervical cancer in young women. In women virus-negative at the outset, the precursors attributable to types 16 and 18 fell from 164 to 2 per 10,000.
A meta-analysis from 2019 covering 65 studies from 14 countries with data on 60 million people shows the same under everyday conditions: the frequency of types 16 and 18 fell by 83 per cent in girls between 13 and 19 years, genital warts declined by 67 per cent in young women and by 48 per cent in young men of the same age. The decline in unvaccinated groups as well demonstrates herd protection.
In the horse there is exactly one counterpart, and it is experimental. The trial already mentioned, using virus-like particles, protected almost completely against lesions caused by type 1. Cross-protection against type 2 was measurable but incomplete: 11 of 14 horses developed small, transient nodules.
Four differences are fundamental. First the host: in humans it is their own virus that multiplies; in the horse a cattle virus foreign to the species sits in the skin and apparently forms no complete particles there. Second the course: the studies cited describe the sarcoid as a skin tumour that does not metastasise, and a transition into a spreading carcinoma does not appear in them.
Third the direction of time. In humans the infection is mostly transient: a meta-analysis of 86 studies with more than 100,000 women found a mean duration of detection of around ten months, and only persistence of the infection counts as the strongest risk factor for high-grade precursors. In the horse it is the other way round. There the lesion stays, and it almost never regresses.
Fourth the application. In humans preventive vaccination is a programme with a measurable effect at population level, and a meta-analysis from 2022 suggests that vaccination around local surgery could lower the recurrence rate; the authors themselves rate the certainty of this evidence from very low to moderate. For the horse no vaccination programme exists, and on an effect in an already affected animal there are no data.
What remains is a sound comparison at the level of mechanism and a clean cut at the level of the clinic. It helps to understand why a sarcoid stays: because a virus redirects the cell and the defences recognise it poorly. It does not help in choosing a procedure. That question stays with the veterinary practice, and the literature so far has no secure answer to it.
Original analysis
| Material examined | Method | Reported finding | What follows | What does not follow |
|---|---|---|---|---|
| Healthy skin of control horses with no known contact | Swab and biopsy, conventional PCR | 30 per cent positive | Viral genetic material is widespread in the horse population | That these horses will develop a sarcoid |
| Healthy skin of horses with a sarcoid | Swab and biopsy, conventional PCR | 73 per cent positive | Affected horses carry the virus outside the lesion too | That the healthy skin is the source of the lesion |
| Blood of the same horses | conventional PCR | no detection | Spread through the bloodstream is not demonstrated this way | That it would not exist with more sensitive methods |
| Semen of clinically healthy stallions | droplet PCR | 34 of 58 samples positive | Genetic material is detectable in the semen of healthy animals too | That a virus able to multiply or a route at covering is present |
| The same semen samples | conventional quantitative PCR | 5 of 58 samples positive | The result depends strongly on the detection method | That older and newer series are directly comparable |
| Sarcoid tissue | droplet PCR for four viral types | 83 of 93 with type 1, of which 66 also with type 2 | Multiple infections are the rule in tumour tissue | That the double infection determines the course |
| Sarcoid tissue | count of viral copies per subtype | no difference between the subtypes | Viral load does not explain the clinical form | That viral genes are irrelevant to the course |
| Stable flies after tissue contact | measurement of viral load over a week | rise, return to the starting level within days | Flies can carry genetic material briefly | That a fly transmits a sarcoid |
| Lesions in treatment studies | histological confirmation | in 60 per cent of studies for all lesions | Part of the literature rests on clinical diagnosis | That all reported regressions concerned sarcoids |
The route of transmission remains unexplained, and the simple picture of contagion fits the findings poorly. In the sarcoid itself, viral genetic material and messenger RNA could be detected, but no infectious viral particles. At the same time 57 per cent of 42 horses examined carried viral genetic material in healthy skin, including 30 per cent of control animals with no known contact with affected horses or with cattle. Stable flies can carry viral genetic material for a few days after tissue contact; that they transmit a sarcoid in doing so has never been tested. How an individual horse is to be kept in a concrete case is decided by the veterinary practice.
Rarely. In the most informative observation so far, 107 tumours were left untreated and followed for an average of 34 months: two of them regressed, that is 1.9 per cent. A 2026 appraisal calls spontaneous regression an exceptional event and confines it to mild, occult and at most verrucous forms. The older rule of thumb that a third disappears on its own counted horses rather than tumours and rested on clinically made diagnoses without histological confirmation. It should no longer stand as the usual estimate.
Because the virus stays in the skin even when the visible tumour has been removed. The Swedish series of 1995 gave a recurrence rate of around 35 per cent after the first attempt at treatment, most recurrences within four months. In a Belgian series of 230 animals, just under 75 per cent of treatments counted as successful, while horses with several tumours failed markedly more often. In the cell model the viral protein E5 suppresses the markers on the cell surface, which makes a poor immune response plausible. In the living horse this link is not proven.
The evidence is contradictory, and both studies come from Switzerland. In 2010, in a randomised study, lesions regressed fully or in part in 13 of 32 treated horses, against 3 of 21 control horses; the authors called the extract effective. In 2024 a randomised, double-blind, placebo-controlled study in 45 horses found no secure difference from the sham treatment. With three such small groups the uncertainty is large: the result does not refute an effect, but neither does it establish one.
Only in part. Shared are the virus family, the redirecting viral proteins and the involvement of the major histocompatibility complex; a genetic study in the horse points expressly to the corresponding findings in humans. The differences, however, are fundamental: in the horse a cattle virus foreign to the species sits in the skin without forming complete particles there, and the studies cited describe the sarcoid as a skin tumour that does not spread. In humans the infection usually disappears again; in the horse the lesion stays. The comparison explains a mechanism, not a treatment.
No vaccination programme, but one trial. In 2017 horses were immunised with virus-like particles and then inoculated at ten sites on the neck: all seven unvaccinated control horses developed a lesion at every site, 13 of 14 vaccinated horses stayed clear. Cross-protection against the second viral type was incomplete. The direction is what matters: vaccination was preventive and given to healthy horses, and the lesions produced were pseudo-sarcoids. On an effect in a horse that already carries a sarcoid there are no data.
Terms defined in this text
Equine Health
In the placebo-controlled trial 47 per cent of vaccinated horses halved their lesion score, 13 per cent on placebo. Where the 88 per cent figure comes
MUMultiple speciesCweakT1hypothesis only Open access
Comparative Health
Ten of sixteen clinically unremarkable horses stayed under 30 minutes of REM sleep: measured lying times and what sleep loss actually shows.
MUMultiple speciesBmoderateT1hypothesis only Open access
Equine Health
82 per cent of clinically healthy horses reacted to leptospires: what serology, genetics and the autoimmune trail really establish, and what they do n
HOHorseBmoderate Open access
Once a month: new syntheses, checked studies, and what has changed in the state of the evidence.