Use of ethanol in the treatment of distal tarsal joint osteoarthritis: 24 cases



Reasons for performing this study: Intra-articular ethanol has been described to promote distal tarsal joint ankylosis. Its use and results in clinical cases affected by osteoarthritis (OA) have not been reported.

Objectives: To describe and evaluate the results of treatment of distal tarsal joint OA by facilitated ankylosis stimulated by intra-articular ethanol injection.

Methods: Twenty-four horses met the inclusion criteria of tarsometatarsal and centrodistal joint OA diagnosed by a positive response to intra-articular analgesia, radiographic evaluation and recurrence of lameness ≤4 months after intra-articular medication with a corticosteroid. Horses were sedated and, following a radiographic contrast study of the tarsometatarsal joint, medication with 2–4 ml of either 100% pure ethanol (G100) or a 70% ethanol (G70) solution was applied. Horses were classified as improved based on a 50% reduction from initial lameness grade combined with an increase in exercise level.

Results: Of the 24 horses included in this study, 20 had the treatment performed bilaterally and 4 unilaterally. All horses were available for initial follow-up examination and 21 for a second one 6–9 months after treatment. This represented a total of 44 treated limbs and 35 available for long-term follow-up. Of these, 21/35 (60%) were considered improved, which corresponds to 11/21 horses (52%). Of 21 horses, 4 (19%) deteriorated and 2 of these developed significant complications related to treatment.

Conclusions: Distal tarsal joint ankylosis with ethanol should be considered a safe and economic treatment in cases of distal tarsal joint OA that fail to show long-term improvement with intra-articular corticosteroid treatment.

Potential relevance: Ethanol should be considered in the treatment of certain cases of distal tarsal joint OA. The importance of performing an adequate radiographic contrast study of the tarsometatarsal joint prior to treatment is highlighted.


Osteoarthritis (OA) of the distal tarsal joints is one of the most common causes of hindlimb lameness in the horse [1]. Diagnosis is made by a combination of procedures, which include physical and lameness examinations, diagnostic analgesia and radiography [2]. Nuclear scintigraphic evaluation may be useful in certain cases. Magnetic resonance imaging and histology have also been used to further characterise the pathological processes and changes associated with this condition [3–5].

In spite of significant diagnostic advances, treatment options available to the general practitioner are still limited. Those available can be divided into nonsurgical and surgical [6]. Nonsurgical treatments involve systemic and local medication, which, combined with exercise management, attempt to reduce inflammation and pain [6]. Surgical options include procedures aiming to stimulate ankylosis of the small tarsal joints [7,8], to desensitise the area by tibial neurectomy [9] or to reduce pressure on the tarsus by cunean tenectomy [6]. Several methods of stimulating ankylosis by physical or chemical cartilage destruction have been described. These include drilling across the joint surfaces, with or without additional internal fixation, the use of a laser and controlled injection of monoiodoacetate (MIA) [10]. Results from the use of MIA vary greatly between studies [8,11] and up to 3 injections may be required to achieve ankylosis [10].

Recently, Shoemaker et al. [12], described the effects of intra-articular infiltration of ethanol into the tarsometatarsal joints of sound horses. Importantly, this report described histological evidence of ankylosis 12 months after injection and reported no associated significant complications or discomfort. The results of its use in 11 clinical cases of distal tarsal OA were reported in 2009 with 100% success for reduction in lameness at 12 months [13] and, in a separate study, 16 of 16 horses demonstrated short to medium term improvement [14].

The authors have used ethanol since 2007 as a treatment for distal tarsal joint OA in cases that failed or ceased to respond to intra-articular corticosteroids. Our aims in this study were: 1) to evaluate the clinical response to treatment with intra-articular ethanol in horses that had received and not responded satisfactorily to intra-articular corticosteroid treatment; and 2) to describe and critically analyse the technique of tarsometatarsal joint (TMTJ) intra-articular treatment with ethanol. To our knowledge this is the largest published report on the effects of this treatment in clinical cases of distal tarsal joint OA to date.

Materials and methods

Inclusion criteria

All horses included in this study met the following inclusion criteria: 1) having chronic (>4 months) hindlimb lameness which showed significant improvement (>50%) in the most lame limb following intra-articular analgesia of the tarsometatarsal joint with mepivacaine hydrochloride performed within one month prior to ethanol treatment; 2) recurrence of lameness within 4 months of intra-articular corticosteroids (triamcinolone acetate or methylprednisolone acetate); and 3) having one or more radiographic changes (narrowing of joint spaces, subchondral bone sclerosis, ‘subchondral bone lysis’ and ‘periarticular remodelling’) consistent with OA of the small tarsal joints based on subjective visual assessment of 4 standard radiographic views of the tarsus.

Case details

Age, breed, sex and use were recorded for all horses.

Lameness examinations and grading

Horses were assessed in-hand walking and trotting in straight lines and when lunged in either direction on a soft and a firm surface. Lameness was graded on a scale of 0–10 (0 = sound, 10 = nonweightbearing [15] for each of the above conditions and given an overall score.

Cases were classified as having uni- or bilateral lameness. The limb showing the most severe lameness was defined as the ‘lamest limb’ and, for bilaterally affected cases the contralateral limb was conversely designated as the ‘least lame limb’. The lamest limb was graded during initial examination and a score for the least lame limb assigned after analgesia of the lamest limb had been performed or when the horse was evaluated on the lunge (limbs loaded unevenly).

At follow-up examinations, lameness was assessed and graded under the same conditions as above.

Response to treatment

For individual limbs response to treatment was classified as improved if there was a 50% or more reduction in lameness score at follow-up examination. Horses were overall classified as ‘improved’ if they met both of the following criteria: 1) a minimally 50% reduction in lameness in at least one of the limbs and, if bilaterally lame, no increase in lameness score seen in the contralateral limb; and 2) an increase in the exercise level of the horse according to the clinical history provided by the owner.

Conversely, limbs that did not meet the criteria for improvement were classified as ‘not improved’. Of these, if the inclusion criteria were not met and lameness score had increased at re-examinations, limbs were subclassified as ‘deteriorated’. Horses were classified as having ‘no significant response’ if they did not meet the above criteria or ‘deteriorated’ if the lameness grade in one of the limbs had increased or if there was a reduction in their exercise level.

Contrast study

A contrast study was performed of all joints medicated prior to treatment with ethanol under standing sedation. Following subcutaneous infiltration with 1 ml of 2% mepivacaine (Intra-Epicaine)a at the injection site the TMTJ was approached via a lateroplantar approach immediately proximal to the fourth metatarsal bone [16]. A 20 or 21 gauge needle was used to enter the joint and initial joint penetration was assessed by feel and the presence of synovial fluid at the hub of the needle. A 3 way plastic stop-tap was then attached to the needle and between 2 and 5 ml of iohexol (Omnipaque)b at 240 mg/ml injected until moderate resistance was felt. The contrast was kept within the joint space using the 3 way stop-tap while radiographs (lateromedial and dorsoplantar views) were taken to assess communication with adjacent synovial structures. The stop-tap was then detached but the needle stayed in the joint. The contrast agent was then allowed to drip out of the joint and negative pressure was applied to remove as much as possible from the joint space. Radiographs were assessed for communication with the proximal intertarsal joint (PITJ) and centrodistal tarsal joint (CDTJ) at the time by the veterinary surgeon carrying out treatment and later by 2 blinded independent observers.

Ethanol treatment

Joints were treated if there was no evidence of communication with the PITJ. The same needle was kept in the joint space and 70% ethanol (Ethanol 99.8%)c solution in sterile water (G70) or 100% (G100) ethanolc was injected until minor resistance was felt (2–4 ml). The choice of solution concentration was at the clinician's discretion.

Post treatment care

Although the details of post treatment care varied slightly, all horses started a controlled exercise programme after treatment. The recommended post treatment exercise regime consisted of: box rest and in-hand walking for short distances for 1–2 weeks followed by paddock rest and ridden walk for 6 weeks. Walking and trotting for a further 6 weeks was started if there was improvement seen at lameness re-examination. Horses resumed normal exercise after this period.

Follow-up examination

Horses were hospitalised for 24 h for monitoring and then discharged to the owner's care.

Lameness examinations were performed at 6–12 weeks after treatment and a second examination at 6–9 months after treatment. All examinations were performed by a veterinary surgeon familiar with the case history.

Owners' assessment

A telephone questionnaire was carried out at least 9 months after treatment and owners were asked the following questions: 1) do you think your horse has improved with this treatment? 2) Would you repeat this treatment in the same circumstances to another horse of yours?

Data analysis

Data were recorded and analysed on GraphPad Prism and Excel 2004 software. The difference in limb lameness scores before and after treatment was compared using a Wilcoxon's signed ranks test. A Fisher's exact test was used 1) to compare the effect of TMTJ and CDTJ communication and outcome; and 2) to compare the effect of ethanol concentration on outcome. Statistical significance was defined as P<0.05, and confidence intervals (CIs) were also reported.


Case details

A total of 24 horses met the inclusion criteria for this study. Of these, 20 were geldings and 4 were mares. Their age range was 4–20 years (median 11.5 years). All were riding horses of relatively low level in their disciplines or had a mixed use. Bilateral lameness was diagnosed in 21 horses (87%). The lameness scores ranged from 1 to 5/10 (median 2) in a straight line. Results are presented in Table S1.

Contrast study

A total of 45 separate TMTJ contrast studies were performed and 41 were available for independent evaluation by 2 observers for purposes related to this study. Of these 41, 29 (70%) showed evidence of communication between the TMTJ and CDTJ (Fig 1). In one case there was evidence of communication with the PITJ, and in another there was distal spreadingof the contrast agent in a metatarsal vessel, but no adverse reactions were seen. A Fisher's exact test revealed no relationship between joint communication and outcome (P = 0.71).

Figure 1.

Example of a dorsoplantar radiographic following contrast study showing contrast agent in the TMTJ and CDTJ.

Ethanol treatment

Treatment with ethanol was completed in 44 of 45 lame limbs. One limb was not treated because of evidence of communication with the PITJ. There were 37 limbs in G70 and 7 in G100. Fisher's exact test showed a significant difference in limb improvement between groups, with P = 0.037. Results are summarised in Tables 1 and 2.

Table 1. Results of lameness examinations per limb
 NumberG70 improvedG100 improvedG70+G100 improvedDeteriorated
  1. G70 = limbs treated with 70% ethanol; G100 = limbs treated with 100% ethanol.

6–12 weeks4429/37 (78%)0/729/44 (66%)3/44 (9%)
6–9 months3520/32 (63%)1/321/35 (60%)5/35 (14%)
Table 2. Results of lameness examinations per horse
 NumberG70 improvedG100 improvedG70+G100 improvedDeteriorated
  1. G70 = limbs treated with 70% ethanol; G100 = limbs treated with 100% ethanol.

6–12 weeks2413/20 (65%)0/413/24 (54%)3/24 (13%)
6–9 months2111/17 (65%)1/411/21 (52%)4/21 (19%)


First examination at 6–12 weeks post treatment:

All horses were evaluated within this time interval. Of the 44 treated limbs, 29 (66%) showed a significant improvement. None of the limbs in G100 showed improvement at this stage but 29 of 37 (78%) limbs in G70 were improved (Table 1). This corresponded to an overall improvement in 13 of 20 horses (65%) in G70 (Table 2). Three limbs showed deterioration in lameness score during this period. Wilcoxon's signed ranks test comparing pre- and post treatment lameness scores demonstrated a significant difference between values P<0.0001; 95% CI -0.5, -1.5.

Second examination at 6–9 months post treatment:

Nineteen horses (79%) were available for examination within this time period. Two horses were subjected to euthanasia between 12 weeks and 9 months owing to ongoing multiple limb lameness and were therefore classified as having no significant response to treatment. A further 3 horses were lost to follow-up in the same period. When limbs were classified individually, a total of 35 treated limbs were assessed and 20/32 limbs (63%) in G70 were improved, whereas only one of 3 limbs in G100 was considered improved (Table 1). This corresponded to 11/17 (65%) horses improved in G70 but none in G100, giving a total of 11/21 horses (52%) classified as improved (Table 2). Five limbs considered improved at the first evaluation had deteriorated at this stage. Results of a Wilcoxon's signed ranks test showed a significant difference between lameness scores before and after treatment at P = 0.0006, CI 95% -0.5, -1.5.


Mild swelling at the site of injection was seen in 3 of 24 horses but resolved within 7 days. Severe swelling with secondary cellulitis and fibrosis was seen in one horse. This required rest for 60 days and oral NSAIDs for 15 days and, although a moderate thickening in the area remained, there was no associated lameness. At the second examination, in a total of 5 limbs the lameness had deteriorated and in 2 the lameness was severe and considered to be caused by the development of PITJ OA based on radiographic examination of the area.

Owner questionnaire

Seventy-five percent of owners felt that their horse had improved; the same number answered that they would consider treating another of their horses by the same method.



The terms alcohol, ethyl alcohol and ethanol are used interchangeably, but the correct chemical nomenclature for C2H5OH is ethanol [17]. The mechanism of action and the therapeutic effects of ethanol in joint disease are not fully elucidated but it has been shown that ethanol causes neurolysis and proteinolysis, which lead to cell death [18–20]. Shoemaker et al. [12] demonstrated by pathological and histological examination of joints injected with ethanol that ankylosis is achieved consistently. This provides the most reliable information to date on the mode of action following intra-articular ethanol injection in horses.

The use of ethanol at 2 different concentrations was based on findings from Shoemaker et al. [12] where a 95% concentration of ethanol appeared to produce a higher rate of joint fusion than 70%. This was in contrast to superior results obtained with 70% solutions in clinical cases as reported by Carmalt et al. [13]. Our poor initial results obtained using 100% ethanol made us discontinue its use. We did not set out to compare results using 100% and 70% solutions: not only were case numbers very different for the 2 groups, there was no attempt at randomisation or blinding. Nonetheless, our results using 70% solutions were statistically superior to those of 100% ethanol. It is difficult to envisage a logical mechanism whereby better results would be obtained with a lower concentration. It may be that alcohol persists within the joint longer when injected as a 70% solution, giving it longer contact time with articular cells. Higher concentrations of ethanol could draw extracellular water more quickly in to the joint by osmosis, simultaneously reducing the water content and effectiveness of the ethanol on the cells immediately in contact with the mixture. It would be a laudable aim of future research to analyse the relative effectiveness of different ethanol solutions to identify the ideal mixture.

The injection protocol followed in this study was adapted from that suggested by J. Carmalt (personal communication). The only difference was that we infiltrated the subcutaneous tissues with 1 ml of 2% mepivacaine to allow for a safer and more accurate needle placement in the joint. A small number of cases showed swelling consistent with tissue irritation. Subjectively we found that the more attempts to enter the joint were made, the greater the post treatment subcutaneous swelling. Reasons for this include increased bleeding and inflammation or, more likely, by creating multiple leakage routes for ethanol and/or iodine contrast agent into the subcutaneous tissues. Accurate initial needle placement is therefore essential and, if >2 or 3 attempts at entering the joint are needed, consideration should be given to postponing treatment temporarily.

Although ethanol is not classified as a sterilising agent, high concentrations do not support microbiological growth [21]. We did not perform any supplementary sterilisation of commercial ethanol before injection and no cases of iatrogenic synovial sepsis occurred. Ethanol is a plasticising solvent and therefore contact with plastic polymer syringes should be minimised and the use of syringes with additional soft rubber plungers be avoided.


Ankylosis is difficult to judge radiographically [22]. In our experience radiographic changes consistent with degenerative joint disease further complicate this assessment. For this reason we did not attempt to correlate radiographic changes with response to treatment, as the first variable would not have been obtained via an accurate method.

Although it was assumed that ankylosis was the therapeutic mechanism of action, other factors may also play a role. A rapid improvement in lameness was seen in all evaluated animals the day after injection, and 5 limbs classified as improved initially went on to deteriorate in the long term. We speculate that these findings suggest that neurolysis also plays a role.

The lack of signs of discomfort following treatment was seen as a significant advantage over MIA induced ankylosis. Although a recent publication stated that the majority of horses treated with MIA showed only mild or no signs of discomfort [11], this is neither our experience nor that reported in several other publications [10,23], where severe post MIA injection pain was observed. Other factors in favour of ethanol include its safety, wide availability and low cost.

Tarsal disease

Our study clearly has some limitations in its design and these should be taken into consideration. There was no attempt at randomisation or to establish a control group. Furthermore, as those involved in both treatment and follow-up examinations were not blinded to the study, it is possible that some bias might have been introduced. Finally, the 2 limbs of a bilaterally lame horse are not completely separate entities. This led to the presentation of our results in 2 ways: improvement seen in lameness scores of each limb and overall improvement of each horse. This separation can be justified by the localised nature of the treatment and by the fact that if the response in individual limbs was not analysed some of the effects of treatment would have been overlooked. On the other hand, the difficulties and pitfalls of grading bilateral hindlimb lameness [24] must be taken into account. Indeed, this might have introduced some degree of variation in our assessments. However, the strict criteria we used to classify outcome reduced this by combining a grading system with changes in exercise level.

Our criteria for diagnosis of distal tarsal OA were in line with those of other publications [2,25]. However, in cases of bilateral hindlimb lameness (87% of cases) intra-articular analgesia was not always performed on the contralateral lame limb for financial and safety reasons. The condition was considered to be bilateral if radiographic changes were present in both limbs. The overall long-term success in this series was 52% for horses and 60% for individual limbs. The improvement rate seen in this series was lower than that reported by Dowling et al. [11] of 82% following treatment with MIA but comparable to other studies using corticosteroids [2,25]. Apart from the obvious differences between studies, the inclusion criteria stipulated in our study were that the diagnosis had to be achieved by a combination of positive response to intra-articular analgesia of the TMTJ and, following this, to have failed to respond to intra-articular treatment with corticosteroids. The latter feature potentially places our cases in a more problematic subset which might be less responsive to ankylosis.

Since the publication of Shoemaker's report [12], ethanol has been used to treat this condition, and the presentations by Bell et al. [14] and Carmalt and Wilson [13] reported much better outcomes than those obtained in our study. We essentially followed the same treatment protocol, but a difference in population size, type, clinical history and the longer-term follow-up of our cases might account for this difference. Further comparisons between those presentations and our study can be made once more details of their work are available.

Although reasons for unsuccessful long-term response to corticosteroids are not fully understood, it is likely that more complex pathology is occurring in these joints [5]. Another reason may be the inconsistency in communication between CDTJ and TMTJs (71% were seen to communicate) and that ethanol may not diffuse as readily between the TMTJ and CDTJ as has been reported for mepivacaine [26] and methylprednisolone [27]. Further work is necessary to characterise how the method of assessing communication and the presence of joint pathology might influence the results obtained. Performing a second injection into the CDTJ joint would abolish some of these factors; however, we opted not to do this as we find it difficult to avoid partial subcutaneous injection when infiltrating this joint under pressure. This would make a contrast study difficult to interpret and subcutaneous or cunean bursa injection of ethanol likely to occur. In support of our decision are the Fisher's exact test results confirming the hypothesis that there was no relationship apparent between joint communication and outcome.

Deterioration in lameness was seen in 4/35 (11%) limbs at the 6–9 months evaluation. In 2 cases this was considered mild but, in the other 2 PITJ OA developed and they became severely lame. This complication has also occurred following MIA treatment [10,24]. This could be due to either a change in tarsal biomechanics, with distal ankylosis increasing stress in the proximal joints, or possible contamination of the proximal joint with ethanol at the time of treatment. Regardless of its cause, long-term deterioration should be seen as a potential complication and explained to owners. As has been found in many previous studies, client satisfaction was higher than objective results at reassessment. This is likely to be attributable to a placebo effect resulting from clients having paid for a new treatment, therefore having a more sympathetic view of the results achieved.

The findings of this study support the use of 70% ethanol solution for treatment of distal tarsal joint OA in cases that do not respond to corticosteroid medication. Careful case selection and accurate contrast-facilitated injection technique are mandatory to minimise potential complications of this treatment.

Authors' declaration of interests

No conflicts of interest have been declared.

Sources of funding



We would like to thank all the Veterinary Surgeons and staff who work at our clinics and hospitals and at the Minster Equine Clinic; Mr Nick Freeman for supplying a case to this study, and Dr James Carmalt for advice and providing us with his initial experiences and treatment protocol.

Manufacturers' addresses

a Dechra, Stoke-on-Trent, Staffordshire, UK.

b GE Healthcare Limited, Little Chalfont, Buckinghamshire, UK.

c Fisher Scientific UK Ltd, Loughborough, Leicestershire, UK.