Ankle arthrodesis is the most commonly performed operative treatment for tibiotalar arthritis in high-demand patients due to its predictable reproducibility of outcomes compared with ankle replacement.1–3 Arthroscopic technique seems to improve patient outcomes4 while decreasing hospital-related costs remarkably.5
Surprisingly, the incidence of ankle fusion nonunions has been reported to range from 5.7%6,7 to 22.6%.8 Frey et al9 described a high rate of nonunion for posttraumatic arthritis with the initial diagnosis of talus fractures (67%), plafond fractures (59%), and combined talus/plafond fractures (75%). Smokers are 4 times more likely to experience a nonunion.10
Neuromuscular pathology has also been reported to interfere with union.11 Adjacent fusion of the subtalar joint (risk ratio, 3.89; P<.05) as well as preoperative varus alignment of greater than 10° have both been reported to increase rates of nonunion.12 Rheumatoid arthritis with the concurrent use of immunosuppressive medication has a similar nonunion rate.13,14 A nonunion in the lower extremity is an unfortunate challenge that involves a prolonged period of disability, burden to the patient, and significant use of health care resources.15
Although surgeons can choose the most suitable fixation method and the type of implant, few techniques involving biological adjuncts to increase the rate of successful unions have been established. Wheeler et al8 reported on locally generated bone slurry to increase the bridging at the 6- and 12-week marks. However, Crosby et al16 found that the ankle fusion rate could not be improved by bone grafting. Autograft from sites such as the ipsilateral iliac crest includes substantial morbidity17 of the operative site and prolongs inpatient stay.
On the other hand, the pharmaceutical industry offers suitable products to achieve union that differ in terms of their efficacy but commonly involve high cost. As outlined by Obremskey et al,18 individual surgeons are then left with the decision whether to use costly orthobiologics (bone morphogenetic protein, platelet-rich plasma) for improved outcomes while at the same time being under the constraints of limited health care resources in a single-payer system.
Preventing nonunion is an obvious priority for patients and surgeons, but at what cost? Although the cost of autogenous bone graft and its added morbidity has been previously examined,19 the current authors recognized that they had little information on the actual cost of an ankle fusion nonunion to their hospital system.
The authors set out to define a cohort of patients needing a revision ankle fusion for symptomatic nonunion to calculate the cost of all related institutional expenses. They also set out to determine at what point buying orthobiologics to decrease nonunions would become cost-effective. They hypothesized that the cost and skin-to-skin time of the revision would be greater than those of the index surgery and that orthobiologics would need to be less expensive than currently marketed to be cost-effective.
Materials and Methods
Study Design and Population
This study, which received institutional review board approval, was designed as a multicenter retrospective case-control chart review of all patients who had a symptomatic ankle fusion nonunion that required revision.
Patients were included in this study if their first tibiotalar arthritis procedure was performed at the authors' centers between 2004 and 2017 and was an ankle arthrodesis. The revision surgery for symptomatic tibiotalar nonunion also had to be performed at the authors' centers. Accordingly, patients who had their first procedure elsewhere and underwent revision at the authors' centers were excluded because the authors could not capture every expenditure until a referral was placed. The authors' outlined follow-up was defined by a group shared clinical practice pattern consisting of follow-up at 2 weeks, 6 weeks, 3 months, 6 months, 1 year, and beyond if clinically required.
The relevant electronic medical record systems were searched for the period between 2004 and 2017 by 2 groups (4+1) of fellowship-trained foot and ankle surgeons (A.N.V., M.G., M.P., K.W., A.S.E.Y.) working in 2 referral centers with large catchment areas within 2 states. The electronic records of all patients who had a procedure code for tibiotalar fusion (694) were reviewed, leading to a list of patients who received a second surgery for nonunion being generated.
Data
Of the above described patients with 2 or more tibiotalar fusion codes on the same extremity, exclusions were applied to 1 who had a total ankle replacement as the initial index surgery, 2 who had had their index surgery elsewhere, 1 who received an elective ankle fusion revision for varus malunion, 1 who had a revision for progressing Charcot arthropathy, and 3 who had takedown ankle fusion to a replacement and subtalar fusion. Once these 8 patients were excluded, 24 remained.
Electronic medical records, hospital charts, and radiology software were reviewed for the 24 included patients. All data collection was performed by an independent surgeon (O.J.G.) observer who had privileges at both centers. Attention was focused on the patients' smoking status, type of arthritis, body mass index (BMI), age, diabetes status, and workers' compensation claim status.
Statistical Analysis
To determine the extra cost associated with a revision of a symptomatic non-union, every clinic visit, examination, revision surgery, and postrevision in-hospital stay were recorded for further analysis. The period following the index fusion was considered standard care and thus was not included in the calculation until a computed tomography (CT) scan was performed to investigate a potential nonunion. If a CT scan was absent, another milestone was established at the 1-year mark after the index procedure. Any intervention conducted after reaching 1 of the 2 milestones was considered outside of standard care and thus was included in the cost calculation. For every visit, the surgeon's billing fee was added to the hospital cost of any clinic visit as well as the cost of interventions (ie, radiograph, CT scan) performed. For the revision procedure, the surgeon's and the anesthesiologist's fees along with the hospital-derived bundled cost of the operating room, nursing staff, facilities, intraoperative fluoroscopy, implants, postanesthetic recovery unit, and inpatient hospital stay were added. Those calculations required the evaluation of the detailed records of the duration of the procedure as well as the inpatient stay. Postrevision clinic visits were added and calculated as mentioned above until the last follow-up related to the tibiotalar procedure was completed (Figure 1).
Hospital billing codes, practitioners' billing fees, and references were used to calculate the total costs related to ankle fusion nonunion. These were acquired from the institution's finance department and from the centralized payment schedule.20 The total costs were reported in US dollars and also converted into equivalent inpatient overnight stays as a more tangible unit for advocacy.
The Student's t test and linear regression were used to evaluate the statistical significance of the collected data.
Results
The 24 patients identified with a non-union meant that the incidence of ankle fusions that required a revision for nonunion was 3.4%. Cases of nonunion that did not undergo further surgery were not captured.
As summarized in Table 1, the mean age at the time of surgery was 64 years (95% confidence interval [CI], 60–69), and 6 patients were female. The mean BMI was 30 kg/m2 (95% CI, 26–34). Nine (38%) patients had diabetes mellitus and 5 (21%) had inflammatory arthritis. Few of these patients (3 of 14) were involved in a workers' compensation claim. Only 2 patients (8%) were active smokers. Ten (42%) were former smokers who had ceased smoking more than 12 months ago. The majority (12; 50%) had never smoked. The index surgeries were performed arthroscopically in 13 cases (54%) and open in 11 cases (46%) through a lateral approach. Most patients had a screws-only construct. Two (8%) patients received a hindfoot nail.
Of the described cohort, 20 of 24 cases had a CT scan to assess nonunion and delayed union a mean of 290 days (95% CI, 192–387) after index surgery. The mean time between the CT-confirmed diagnosis of nonunion and revision surgery was 205 days (95% CI, 127–282). For the entire cohort, the mean time between index and revision surgery was 482 days (95% CI, 381–583). Open revision surgery was performed in all cases. In terms of intra-operative use of autograft, 10 cases had local autograft from the joint preparation, 5 cases had iliac crest bone graft, and 2 cases had proximal tibia graft (Table 2). As for bone substitutes, 3 cases received OSTEOSET T (Wright Medical Group, Arlington, Tennessee), 8 cases received AUGMENT (Wright Medical Group), 6 cases received Accell CONNEXUS (IsoTis Inc, Irvine, California), and 1 case received bone morphogenetic protein. Three cases did not use any adjuncts in the revision. Fourteen patients received a combination of 2 adjuncts listed above. The mean duration of a case was 126 minutes (95% CI, 112–179). The mean tourniquet time was 109 minutes (95% CI, 96–122).
After revision surgery, patients remained in the hospital a mean of 2 days (range, 1–13 days); 1 patient was discharged the same day. The mean number of clinic encounters after revision was 6 (range, 1–12), with the last follow-up occurring a mean of 475 days after the revision. One patient with multiple comorbidities and thromboembolic history died 3 days after discharge from the hospital. None of the patients had a repeat revision ankle fusion.
The first period, between diagnosis of nonunion and revision surgery, had a mean cost of $627 (95% CI, 366–888). The total cost of surgery was $3026 (95% CI, 2616–3437), consisting of $898 (95% CI, 761–1035) for the surgeon's billing fees, $256 (95% CI, 236–276) for anesthetic services, and $1883 (95% CI, 1531–2235) for the hospital-bundled operative cost. The subsequent inpatient stay cost a mean of $3432 (95% CI, 2219–4645). The mean cost of follow-up after revision was $1754 (95% CI, 1399–2110). The costs of radiological diagnostics from diagnosis of nonunion to last follow-up totaled $1061 (95% CI, 722–1400). If the case were to be performed in the ambulatory surgical center, the cost of inpatient stay would not apply.
A mean total cost of $9683 (95% CI, 8217–11,148) was incurred. For ambulatory surgical center cases, the total cost would be $6251 (95% CI, 3572–8929). The cost of a repeat fusion for nonunion differs significantly (P<.05) from the previously reported cost of a primary ankle fusion ($4104±$373).21 Because the cost of 1 overnight stay at the treating institution was $1129, it was concluded that the mean sum of the institutional costs of a nonunion revision would be equivalent to 9 inpatient overnight stays or 6 if the surgery was performed in an ambulatory surgical center.
Discussion
This study is the first of its kind to examine the clinical outcome of nonunion in ankle fusions and quantify its holistic institutional health system cost. This study offers crucial information to the surgeon advocating for tools to prevent this and opens the conversation with research and innovation industry teams. By encompassing 4 surgeons' practices, this study had a large number (N=24) of patients with a relatively rare condition.
The rate of revision for nonunion of 3.4% was lower than the 8.6% cited in a recent review by Abicht and Roukis.6 However, the current case-control study was not designed to report the incidence of nonunion but rather the rate of revision for nonunion. The authors' number was also only based on cases that underwent revision and did not capture patients with nonunion who chose to delay revision and continue with observation for health-related or social reasons. However, the revision rate is useful because the authors' centers are the only institutions in a large geographic region that take on revisions for ankle fusion nonunions. Furthermore, patients are unlikely to travel to another part of the country for their revision due to restrictions of their health care plan and the relatively limited access to a specialist's clinic. The authors' rate does not represent ankle fusion nonunions that led to below-the-knee amputations.
The current cohort's demographics were compared with the larger arthrodesis subgroup analysis of the parent prospective database previously reported by Daniels et al.1 Age, sex, and side matched those of the current cohort and were not identified to be different. The current non-union cohort had 9 of 24 (38%) patients with diabetes, whereas the arthrodesis subgroup had 18 of 107 (17%). Therefore, the current patients undergoing ankle fusion revision were twice as likely to have diabetes compared with the index procedure cohort.
The current nonunion cohort had a mean BMI of 30.4±8.32 kg/m2, whereas the arthrodesis subgroup had a mean BMI of 28.7±5.6 kg/m2, which was not statistically significant (P<.1506). This association has been reported previously.9,22
One of the striking findings of the current study was the impact that postoperative inpatient stay had on total cost. As seen in Figure 2, close to 40% of the overall additional cost due to nonunion was attributed to inpatient hospital stay. The total cost of added institutional interventions ($9683) was more than twice the cost of the index primary ankle fusion ($4104±$373).21 If the surgeries were performed at an ambulatory surgical center, the total cost would be $6251.
Using allograft instead of autograft has been shown to facilitate earlier discharge, as allograft avoids the pain burden from the second operative site.23–25 The field of orthopedic surgery has seen an increase in orthobiologic products during the past 20 years, and different practices exist regarding their use in the index procedure. The current study highlighted that the higher the cost of a nonunion, the more valuable its prevention becomes. The correct use of an orthobiologic agent at index surgery, technical considerations of joint preparation, compression and fixation, and patient education and compliance all drive the incidence of nonunion.
Intangible costs that were not represented in this study can be substantial, even when standards of care are met.26 For reference, the public relations department handling various litigations and lawsuits with the use of legal counsel, the attribution of large compensation payout, and third-party risk management consultants represented £36 million for foot and ankle alone in the United Kingdom during a 17-year period.27 Although no causality could be concluded, the current study indicated a trend with certain well-known risk factors (increased BMI and diabetes). Because these risk factors increase the probability of a nonunion,28 surgeons should consider delaying operations until both conditions have been optimized.
On designing an intervention that would address a nonunion, one has to recognize the different causes from local biology, fixation, and patient factors. Given that the number needed to treat for this study was 29 and that the cost of a non-union is now known, the authors can infer that no more than $335 should be spent on an adjunct during every index surgery prior to outweighing the cost of ankle fusion nonunion. In the setting of an intervention or orthobiologic that is only 50% efficacious, that cost would then have to be $167 or less. If the surgery does not require postoperative admission in the ambulatory surgical center setting, no more than $216 should be spent on an adjunct during every index surgery or $108 if the intervention is 50% efficacious.
Against the authors' initial estimation, the mean skin-to-skin time did not differ (P>.05) between index (114 minutes) and revision (126 minutes) surgery. This is most likely attributable to the fact that all revisions were performed as open procedures.
Conclusion
Ankle fusion nonunion creates a financial burden on the health care system. The total institutional cost represents, on average, the equivalent of a 9-day inpatient stay in an acute care hospital, with approximately 40% of the overall expenses being related to the inpatient stay. The total cost in an ambulatory surgical center represents the equivalent of a 6-day inpatient stay. Responsible use of public funds for adjuncts in primary ankle fusion seems prudent for patients with diabetes and obesity. Additional efforts should be directed toward shortening postoperative admissions and developing integrated models of pain management from the intraoperative delivery of care to the community services. Further research is needed regarding orthobiologic products and their efficacy in reducing the rate of symptomatic nonunion, their harm reduction potential, and their cost-effectiveness.
References
- Daniels TR, Younger ASE, Penner M, et al. Intermediate-term results of total ankle replacement and ankle arthrodesis: a COFAS multicenter study. J Bone Joint Surg Am. 2014;96(2):135–142. doi:10.2106/JBJS.L.01597 [CrossRef] PMID:24430413
- Terrell RD, Montgomery SR, Pannell WC, et al. Comparison of practice patterns in total ankle replacement and ankle fusion in the United States. Foot Ankle Int. 2013;34(11):1486–1492. doi:10.1177/1071100713494380 [CrossRef] PMID:23774468
- Mendicino SS, Kreplick AL, Walters JL. Open ankle arthrodesis. Clin Podiatr Med Surg. 2017;34(4):489–502. doi:10.1016/j.cpm.2017.05.006 [CrossRef] PMID:28867055
- Townshend D, Di Silvestro M, Krause F, et al. Arthroscopic versus open ankle arthrodesis: a multicenter comparative case series. J Bone Joint Surg Am. 2013;95(2):98–102. doi:10.2106/JBJS.K.01240 [CrossRef] PMID:23235956
- Peterson KS, Lee MS, Buddecke DE. Arthroscopic versus open ankle arthrodesis: a retrospective cost analysis. J Foot Ankle Surg. 2010;49(3):242–247. doi:10.1053/j.jfas.2010.02.019 [CrossRef]
- Abicht BP, Roukis TS. Incidence of non-union after isolated arthroscopic ankle arthrodesis. Arthroscopy. 2013;29(5):949–954. doi:10.1016/j.arthro.2012.12.001 [CrossRef] PMID:23395470
- Haddad SL, Coetzee JC, Estok R, Fahrbach K, Banel D, Nalysnyk L. Intermediate and long-term outcomes of total ankle arthroplasty and ankle arthrodesis: a systematic review of the literature. J Bone Joint Surg Am. 2007;89(9):1899–1905. doi:10.2106/JBJS.F.01149 [CrossRef] PMID:17768184
- Wheeler J, Sangeorzan A, Crass SM, Sangeorzan BJ, Benirschke SK, Hansen ST Jr, . Locally generated bone slurry accelerated ankle arthrodesis. Foot Ankle Int. 2009;30(7):686–689. doi:10.3113/FAI.2009.0686 [CrossRef] PMID:19589317
- Frey C, Halikus NM, Vu-Rose T, Ebramzadeh E. A review of ankle arthrodesis: predisposing factors to nonunion. Foot Ankle Int. 1994;15(11):581–584. doi:10.1177/107110079401501102 [CrossRef]. PMID:7849972
- Cobb TK, Gabrielsen TA, Campbell DC II, Wallrichs SL, Ilstrup DM. Cigarette smoking and nonunion after ankle arthrodesis. Foot Ankle Int. 1994;15(2):64–67. doi:10.1177/107110079401500202 [CrossRef]. PMID:7981802
- Jain SK, Tiernan D, Kearns SR. Analysis of risk factors for failure of arthroscopic ankle fusion in a series of 52 ankles. Foot Ankle Surg. 2016;22(2):91–96. doi:10.1016/j.fas.2015.05.007 [CrossRef]
- Chalayon O, Wang B, Blankenhorn B, et al. Factors affecting the outcomes of uncomplicated primary open ankle arthrodesis. Foot Ankle Int. 2015;36(10):1170–1179. doi:10.1177/1071100715587045 [CrossRef]. PMID:25994833
- Rabinovich RV, Haleem AM, Rozbruch SR. Complex ankle arthrodesis: review of the literature. World J Orthop. 2015;6(8):602–613. doi:10.5312/wjo.v6.i8.602 [CrossRef]. PMID:26396936
- Mäenpää H, Lehto MU, Belt EA. Why do ankle arthrodeses fail in patients with rheumatic disease?Foot Ankle Int. 2001;22(5):403–408. doi:10.1177/107110070102200508 [CrossRef]. PMID:11428759
- Heckman JD, Sarasohn-Kahn J. The economics of treating tibia fractures: the cost of delayed unions. Bull Hosp Jt Dis. 1997;56(1):63–72. PMID:9063607
- Crosby LA, Yee TC, Formanek TS, Fitzgibbons TC. Complications following arthroscopic ankle arthrodesis. Foot Ankle Int. 1996;17(6):340–342. doi:10.1177/107110079601700608 [CrossRef] PMID:8791081
- Calori GM, Colombo M, Mazza EL, Mazzola S, Malagoli E, Mineo GV. Incidence of donor site morbidity following harvesting from iliac crest or RIA graft. Injury. 2014;45(suppl 6):S116–S120. doi:10.1016/j.injury.2014.10.034 [CrossRef]. PMID:25457330
- Obremskey WT, Marotta JS, Yaszemski MJ, Churchill LR, Boden SD, Dirschl DR. The introduction of biologics in orthopaedics: issues of cost, commercialism, and ethics. J Bone Joint Surg Am. 2007;89(7):1641–1649. doi:10.2106/JBJS.F.01185 [CrossRef] PMID:17606804
- St John TA, Vaccaro AR, Sah AP, et al. Physical and monetary costs associated with autogenous bone graft harvesting. Am J Orthop (Belle Mead NJ).2003;32(1):18–23. PMID:12580346
- BC-MoH. Medical Services Commission Payment Schedule. Victoria, British Columbia, Canada: British Columbia Ministry of Health; 2018.
- Younger AS, MacLean S, Daniels TR, et al. Initial hospital-related cost comparison of total ankle replacement and ankle fusion with hip and knee joint replacement. Foot Ankle Int. 2015;36(3):253–257. doi:10.1177/1071100714558844 [CrossRef]. PMID:25367250
- Collman DR, Kaas MH, Schuberth JM. Arthroscopic ankle arthrodesis: factors influencing union in 39 consecutive patients. Foot Ankle Int. 2006;27(12):1079–1085. doi:10.1177/107110070602701214 [CrossRef] PMID:17207436
- Huang YC, Chen CY, Lin KC, et al. Comparing morbidities of bone graft harvesting from the anterior iliac crest and proximal tibia: a retrospective study. J Orthop Surg Res. 2018;13(1):115. doi:10.1186/s13018-018-0820-3 [CrossRef]. PMID:29769090
- Sasso RC, LeHuec JC, Shaffrey CSpine Interbody Research Group. Iliac crest bone graft donor site pain after anterior lumbar interbody fusion: a prospective patient satisfaction outcome assessment. J Spinal Disord Tech. 2005;18(suppl):S77–S81. doi:10.1097/01.bsd.0000112045.36255.83 [CrossRef] PMID:15699810
- Tricot M, Deleu PA, Detrembleur C, Leemrijse T. Clinical assessment of 115 cases of hindfoot fusion with two different types of graft: allograft+DBM+bone marrow aspirate versus autograft+DBM. Orthop Traumatol Surg Res. 2017;103(5):697–702. doi:10.1016/j.otsr.2017.03.014 [CrossRef] PMID:28416462
- Metcalfe CW, Harrison WD, Nayagam S, Narayan B. Negligence claims following non-union and malunion of long bone fractures: an analysis of 15 years of data. Injury. 2016;47(10):2312–2314. doi:10.1016/j.injury.2016.07.017 [CrossRef]. PMID:27461778
- Ring J, Talbot CL, Clough TM. Clinical negligence in foot and ankle surgery: a 17-year review of claims to the NHS Litigation Authority. Bone Joint J. 2014;96-b(11):1510–1514. doi:10.1302/0301-620x.96b11.33963 [CrossRef]
- Thevendran G, Wang C, Pinney SJ, Penner MJ, Wing KJ, Younger AS. Nonunion risk assessment in foot and ankle surgery: proposing a predictive risk assessment model. Foot Ankle Int. 2015;36(8):901–907. doi:10.1177/1071100715577789 [CrossRef]. PMID:25810460
Patient Characteristics
| Characteristic | Value |
|---|
| Demographics | |
| Patients, No. | 24 |
| Age at surgery, mean, y | 64 (95% CI, 60–69) |
| Sex, female:male, No. | 6:18 |
| Body mass index, mean, kg/m2 | 30.4 (95% CI, 26–34) |
| Side, right:left, No. | 10:14 |
| Diabetes mellitus, No. | 9 (38%) |
| Smoker, No. | |
| At any time | 12 (50%) |
| Nonsmoker | 12 (50%) |
| Inflammatory arthritis, No. | 5 (21%) |
| Workers' compensation case, No/total no. | 3/14 (21%) |
| Index surgery | |
| Open, No. | 11 |
| Arthroscopic, No. | 13 |
| Skin-to-skin time, mean, hours:minutes | 01:54 (95% CI, 01:41–02:06) |
| Construct, No. | |
| Fibular plate and screws | 8 |
| 2 screws | 7 |
| 3 screws | 5 |
| 4 screws | 0 |
| 5 screws | 2 |
| Hindfoot fusion nail | 2 |
| Same day discharge, No. | 4 |
| Length of stay, mean, d | 3.05 (95% CI, 2–4) |
Use of Bone Graft and Orthobiologics for Both Revision and Index Surgery
| Type | No. |
|---|
|
|---|
| Index Surgery | Revision Surgery |
|---|
| Adjunct | | |
| Autograft | | |
| Iliac crest | 1 | 5 |
| Proximal tibia | 0 | 2 |
| Joint preparation slurry local graft | 10 | 10 |
| Allograft | | |
| Femoral head | 1 | 0 |
| Bone substitute | | |
| OSTEOSET T bone grafta | 2 | 3 |
| AUGMENT bone grafta | 4 | 8 |
| Bone morphogenetic protein | 0 | 1 |
| Accell CONNEXUSb | 9 | 6 |
| Dynagraftb | 1 | 0 |
| No adjunct | 5 | 3 |