ADVERTISEMENT
Is The Keller Arthroplasty An Underappreciated Option For Treating Diabetic Foot Ulcers?
In my experience, our education and training characterizes the Keller arthroplasty as outdated and as little more than an academic mention. While it seems well-known that the Keller procedure carries wide indications for geriatric bunion surgery, there is only rare discussion of performing a Keller for a plantar hallux diabetic ulcer. The most recent edition of the McGlamry textbook does not mention diabetic ulcers as an indication for performing a Keller arthroplasty.1 The third edition of the same textbook includes only one sentence on this topic.2
A comprehensive literature search revealed a total of five published articles that specifically evaluated the Keller arthroplasty for plantar hallux ulcers.3-7 The five articles combined looked at a total of 82 procedures, and the below discussion stems from their findings. Additionally, the article will introduce some controversial topics related to the Keller to highlight a void in the current medical literature.
How Can The Keller Arthroplasty Impact Plantar Hallux Ulcers?
Basic biomechanics of the first ray helps explain the reasoning for why the Keller can be an effective procedure. Patients with plantar hallux ulcers likely have either a structural or functional hallux limitus leading to decreased range of motion at the first MPJ. These ulcers can result from compensation for the hallux limitus pathology distally and transfer of pressure to the the hallux interphalangeal joint.8 The Keller procedure removes the base of the proximal phalanx, leading to increased range of motion at the first MPJ, effectively reducing pressure at the hallux interphalangeal joint.9 The five articles reviewed indicate that of the 82 overall procedures performed on plantar hallux ulcers, all resulted in wound closure, suggesting that the Keller arthroplasty is extremely effective for this purpose.3-7
Common complications reported include: wound dehiscence; delayed healing; postoperative infection; transfer metatarsal ulcer; and ulcer to the tip of the second toe.3-7 The most common complication amongst these is wound dehiscence of the dorsal incision site. One can most likely attribute this to the natural dead space created when performing a Keller and the possible resultant hematoma. Additionally, one can reasonably argue that trading a dorsal wound for a closed plantar ulcer is fair, because a wound on the weight bearing surface can pose unique difficulties compared to a dorsal dehiscence.
None of the articles reviewed cited cases of flail toe after performing a Keller arthroplasty.3-7 One article described a case of “cock up hallux,” and goes on to describe an uneventful correction with extensor hallucis longus tendon lengthening.7 In my observation, there seems to be a fear amongst foot and ankle surgeons that the Keller procedure will destabilize the hallux and lead to a poorly functioning flail toe. While this seems to remain a feared complication, this literature review doesn’t report any cases of flail toe following Keller arthroplasty when performed for plantar hallux ulcers.3-7
Regarding the use of fixation, three articles in this review used no fixation with no advanced soft tissue or capsular repair.3,5,7 These articles all showed resolution of the plantar hallux ulcer with no resultant flail toe, suggesting that no fixation is a reasonable choice.
One study used K-wire fixation in combination with a specific medial capsulorraphy soft tissue repair technique.4 What’s interesting about this study is that they reported five cases out of 13 of transfer metatarsal ulcer. In comparison to the other studies mentioned this is a relatively large number of transfer metatarsal cases.3-7
One might expect transfer ulcers to result from a destabilized hallux not addressed by a K-wire or capsular repair, but this was not the case. It’s likely that destabilizing the hallux has more to do with how much bone one excises from the proximal phalanx rather than fixation and soft tissue repair techniques. The study that found more transfer ulcers describes excising the entire proximal one-third of the proximal phalanx bone in the technique portion of their study.4 Henry and Waugh looked at the concept of overzealous bone resection in their study that showed hallux purchase during weight bearing after Keller arthroplasty worsens as one resects more bone.10
A Closer Look At Special Case Considerations
All of the studies discussed above only performed the Keller arthroplasty for ulcers that did not involve probing or depth to bone. Ruling out osteomyelitis seemed to be a clear feature of each study, described in the materials and methods sections. To my knowledge, no literature exists addressing use of the Keller arthroplasty in situations where osteomyelitis is, or might be, present. Although possibly controversial to some, in my experience, one can perform the Keller arthroplasty on patients with ulcers that probe to bone and/or have magnetic resonance imaging (MRI) results suggesting osteomyelitis. In one such case, I used the Keller arthroplasty in addition to wound debridement and oral antibiotics (see first two photos), along with intraoperative bone and soft tissue cultures. Keep in mind that when a patient has signs of severe infection with active draining purulence this approach is not indicated and, instead, hallux amputation should be the procedure of choice. A Keller is not the best choice for every case, and I encourage readers to utilize common sense as to when amputation may indeed be most appropriate. More research is necessary in this particular area.
Does the Keller arthroplasty work for a plantar first metatarsal head diabetic foot ulcer? To my knowledge, there is also no published literature on this topic. An interesting clinical observation I have noticed is that plantar first metatarsal head diabetic foot ulcers usually do not arise from a plantarflexed first ray, as one might expect. In my experience, most patients with diabetes and cavus deformity will present with plantar fifth metatarsal head ulcers rather than first metatarsal head ulcers. Additionally, I argue that most first metatarsal head diabetic foot ulcers are a result of uncompensated hallux limitus, which in theory should make the Keller arthroplasty a reasonable surgical option for these patients.
The natural progression of a plantar first metatarsal head ulcer that doesn’t heal includes possible development of osteomyelitis and need for partial first ray amputation. Published data on the success rate of partial first ray amputations includes a 44 percent failure rate.11 This is extremely high and truly sheds light on the importance of an organized treatment approach. I’ve found, in my personal experience, that the Keller arthroplasty is not quite as successful for plantar first metatarsal head ulcers as compared to plantar hallux ulcers. Although, when combined with appropriate arch support to offload, and in combination with gastrocnemius tendon recession, reasonable results can be achieved while avoiding the need for partial first ray amputation (see last two photos).
In Conclusion
Overall, the Keller arthroplasty is an extremely effective procedure when used as part of a treatment plan for plantar hallux ulcers. The literature overwhelmingly shows that wound closure is achievable in most patients undergoing this procedure. The Keller procedure addresses the biomechanical source of the ulcer by increasing the range of motion at the first MPJ, which offloads the plantar hallux interphalangeal joint. Complications are common, but manageable. The most common complications include: wound dehiscence; postoperative infection; delayed healing; transfer metatarsal ulcer; and ulcer to the tip of the second toe. Flail toe was not a significant complication in the articles reviewed above. In regards to fixation, there isn’t strong data to suggest that fixation provides any advantage. The literature currently doesn’t address utilizing the Keller arthroplasty for ulcers that have depth to bone with possible underlying osteomyelitis or for plantar first metatarsal head ulcers. More research is necessary in these areas before making evidence-based recommendations.
Dr. Moon is a Fellow of the American College of Foot and Ankle Surgeons. He is Medical Director of the St. Mary Medical Center Wound Clinic and a podiatric surgeon at St. Mary Medical Center, both in Hobart, Ind. Dr. Moon is a faculty member of The Podiatry Institute in Decatur, Ga.
1. Vanore JV, Montross WG, Jimenez AL, et al. First metatarsal phalangeal joint arthroplasty. In: Southerland JT, Boberg JS, Downey MS, et al, (eds). McGlamry’s Comprehensive Textbook of Foot and Ankle Surgery, 4th ed. Philadelphia:Lippincott Williams and Wilkins;2013;362-399.
2. TJ Chang, CA Camasta. Hallux Limitus and Hallux Rigidus. In: Banks AS, Downey MS, Martin DE, Miller SJ, (eds). McGlamry’s Comprehensive Textbook of Foot and Ankle Surgery, 3rd ed. Philadelphia:Lippincott Williams and Wilkins;2001:705.
3. Armstrong DG, Lavery LA, Vazquez JR, et al. Clinical efficacy of the first metatarsophalangeal joint arthroplasty as a curative procedure for hallux interphalangeal joint wound in patients with diabetes. Diabetes Care. 2003;26(12):3284-3287.
4. Berner AB, Sage R, Niemela J. Keller procedure for the treatment of resistant plantar ulceration of the hallux. J Foot Ankle Surg. 2005;44(2):133-136.
5. Downs DM, Jacobs RL. Treatment of resistant ulcers on the plantar surface of the great toe in diabetics. J Bone Joint Surg. 1982;64(6): 930-933.
6. Lin SS, Bono CM, Lee TH. Total contact casting and Keller arthroplasty for diabetic great toe ulceration under the interphalangeal joint. Foot Ankle Int. 2000;21(7):588- 593.
7. Tamir E, Tamir J, Beer Y, et al. Resection arthroplasty for resistant ulcers underlying the hallux in insensate diabetics. Foot Ankle Int. 2015;36(8):969-975.
8. Bingold AC, Collins DH. Hallux rigidus. J Bone Joint Surg Br. 1950;32:214-222.
9. Keller WL. The surgical treatment of bunions and hallux valgus. NY Med J. 1904;80:741.
10. Henry APJ, Waugh W. The use of footprints in assessing the results of operations for hallux valgus. J Bone Joint Surgery Br. 1975;57B(4):478-481.
11. Borkosky SL, Roukis TS. Incidence of repeat amputation after first ray amputation associated with diabetes melliuts and peripheral vascular disease: an 11 year review. J Foot Ankle Surg. 2013;52(3):335-338.
