- Email: [email protected]
Recent National Trends and Outcomes in Total Ankle Arthroplasty in the United States
ARTICLE IN PRESS The Journal of Foot & Ankle Surgery ■■ (2018) ■■–■■
Contents lists available at ScienceDirect
The Journal of Foot & Ankle Surgery j o u r n a l h o m e p a g e : w w w. j f a s . o r g
Original Research
Recent National Trends and Outcomes in Total Ankle Arthroplasty in the United States Danil Rybalko, MD, MS 1, Garrett Schwarzman, MD 1, Vincent Moretti, MD 2 1Resident,
Department of Orthopedics, University of Illinois at Chicago, Chicago, IL Surgeon, Department of Orthopedic Surgery, Drexel University, Philadelphia, PA
2
A R T I C L E
I N F O
Level of Clinical Evidence: 3
Keywords: national hospital discharge survey outcomes total ankle arthroplasty trends usage
A B S T R A C T
Total ankle arthroplasty (TAA) is an evolving option for treating ankle arthritis. We assessed the national trends in usage and perioperative outcomes of TAA in the United States. International Classification of Diseases, 9th revision (ICD-9), codes were used to search the National Hospital Discharge Survey database for TAA from 1997 to 2010. Patient demographics, comorbidities, hospitalization length, discharge disposition, blood transfusion, lower extremity deep vein thrombosis, pulmonary embolism, and mortality data were gathered. Trends were evaluated using linear regression with Pearson’s correlation coefficient, and statistical comparisons were performed using Student’s t test and z-test for proportions with significance at p = .05. We identified 120 patients with TAA. TAA demonstrated a positive correlation with time (r = 0.57), significantly increasing from 2.4 cases per 100,000 admissions from 1997 to 2003 to 3.5 cases per 100,000 from 2004 to 2010 (p = .04). The mean age was 57.8 (range 19 to 83) years. The mean number of comorbidities was 4.5 (range 1 to ≥7). Although patient age remained stable (p = .21), the mean number of comorbidities significantly increased from 4.0 from 1997 to 2003 to 4.8 from 2004 to 2010 (p = .02); 8 patients (6.7%) had diabetes, 71 (59.2%) had primary osteoarthritis, and 35 (29.2%) had posttraumatic arthropathy. The mean length of stay significantly decreased from 3.1 to 2.3 days (p = .03). Three patients (2.5%) required a blood transfusion. No deep vein thrombosis or PE was diagnosed. No patients died during the operative admission; 95 patients (87%) were discharged home and 14 (13%) required a skilled rehabilitation facility. Discharge patterns showed no significant change with time (p = .59). Usage of TAA in the United States has increased nearly 50% over the past 14 years. TAA was associated with shorter hospitalization, infrequent rehabilitation facility requirements, and few perioperative complications. © 2018 by the American College of Foot and Ankle Surgeons. All rights reserved.
Total ankle arthroplasty (TAA) is one of the few operative treatments available for end-stage osteoarthritis of the ankle. Historically, ankle arthrodesis has been the standard treatment (1–4). However, with long-term studies showing progression of arthrosis in adjacent joints, TAA has provided an alternative to address the pain associated with ankle arthritis and avoid placing additional wear and stress on the joints of the mid- and forefoot (1–3,5). Although first-generation TAAs had inferior outcomes, secondgeneration systems have been reported to result in good to excellent clinical outcomes in 88% to 90% of patients (1). Nevertheless, TAA continues to be a technically challenging procedure. Additionally, a number of perioperative complications have been reported, including delayed
Financial Disclosure: None reported. Conflict of Interest: None reported. Address correspondence to: Garrett Schwarzman, MD, Department of Orthopaedic Surgery, University of Illinois at Chicago, Room E270 MSS, Chicago, IL 60612. E-mail address: [email protected] (G. Schwarzman).
wound healing, infection, fracture, tendon injury, sensory deficits, postoperative stiffness, reflex sympathetic dystrophy, amputation, and pulmonary embolism (PE) (1,2,5). Given the improved outcomes of the new-generation TAA systems, few studies have considered the trends of TAA usage and the associated perioperative outcomes during the past decade. Our review identified only 4 studies of the trends of TAA usage in the United States over time, with each focusing to a varying degree on patient characteristics and perioperative outcomes. Three of these studies used the Nationwide Inpatient Sample (NIS) database (6–8). The fourth study used a Medicare Provider Analysis and Review (MedPAR) database (9). Our study focused on the trends of TAA usage in the United States, patient characteristics, and perioperative outcomes and complications from 1997 to 2010 using the data set from the National Center for Health Statistics National Hospital Discharge Survey (NHDS) (10). To the best of our knowledge, ours is the only study that has conducted this type of analysis using the NHDS data to identify trends and outcomes in TAA usage.
1067-2516/$ - see front matter © 2018 by the American College of Foot and Ankle Surgeons. All rights reserved. https://doi.org/10.1053/j.jfas.2018.03.037
ARTICLE IN PRESS D. Rybalko et al. / The Journal of Foot & Ankle Surgery ■■ (2018) ■■–■■
2
Materials and Methods Data for the present study were obtained from NHDS and imported into Microsoft Office Excel 2010 (Microsoft, Seattle, WA) for statistical analysis (10). The NHDS collects data on patient discharges from a range of representative hospitals across the United States, excluding federal, military, and Veterans Affairs hospitals (10). The NHDS includes data from hospitals from all 50 states and the District of Columbia, exclusive of military and Department of Veteran Affairs hospitals (10). These hospitals must have ≥6 beds to be included in the survey (10). Each year, data from ~1% of total hospital admissions from throughout the United States are collected, abstracted, and weighted to provide nationwide estimates (10). Information collected from each sampled hospital record includes age, sex, race/ethnicity, marital status, discharge month, discharge status, days of care, hospital location, hospital size, hospital type (proprietary, government, or church/nonprofit) and ≤15 discharge diagnoses and ≤8 procedures performed during that admission (10). International Classification of Diseases, 9th revision (ICD-9), procedure codes were used to search the NHDS for all patients admitted to the surveyed U.S. hospitals after a primary TAA for each year from 1997 to 2010. Data regarding patient demographics, hospitalization length of stay, discharge disposition, lower extremity deep vein thrombosis (DVT), PE, and mortality were all gathered from the NHDS. The identified cases were divided into 2 groups. The first group included cases from 1998 to 2003 and second group, cases from 2004 to 2010. Statistical analysis was performed using linear regression with Pearson’s correlation coefficient, Student’s t test, z-test for proportions, and χ2 analysis. An αlevel of 0.05 was used for significance determinations.
Table 2 Temporal trends in postoperative length of stay and disposition to home versus skilled nursing facility after total ankle arthroplasty (N = 120) Variable
Mean length of stay (days) Discharge (n [%]) Home Skilled nursing facility
Temporal Trend
Difference (%)
p Value
2.3 ± 1.2
−1.2
.02 .58
55 (89) 7 (11)
4 −4
1999 to 2003 (n = 52)
2004 to 2010 (n = 68)
3.1 ± 2.5 40 (85) 7 (15)
Data in parentheses are percentages.
mean length of stay after TAA was 2.6 (range 1 to 18) days. The length of stay decreased significantly with time, from 3.1 days in 1997 to 2003 to 2.3 days in 2004 to 2010 (p = .03). Of the 120 patients, 95 (87%) were discharged home after their inpatient stay; only 14 patients (13%) required admission or transfer to an inpatient rehabilitation or skilled nursing facility (Table 2). This discharge disposition pattern showed no significant change with time (p = .59). Discussion
Results From 1997 to 2010, 120 patients admitted for TAA were identified. The mean age of the TAA patients was 57.8 (range 19 to 83) years. The group included 51 males and 69 females. The mean number of medical comorbidities was 4.5 (range 1 to ≥7). Although patient age remained stable across the study period (p = .21), the mean number of comorbidities significantly increased from 4.0 from 1997 to 2003 to 4.8 from 2004 to 2010 (p = .02). Of the 120 patients, 8 (6.7%) had diabetes, 71 (59.2%) had ankle osteoarthritis, and 35 (29.2%) had posttraumatic ankle arthropathy, with most of the remaining 11.6% of patients having rheumatoid arthritis (Table 1). The use of TAA demonstrated a moderately positive correlation with time (r = 0.57), with a significant increase from a mean of 2.4 cases per 100,000 hospital admissions from 1997 to 2003 to 3.5 cases per 100,000 hospital admissions from 2004 to 2010 (p = .04). Three patients (2.5%) required a blood transfusion. No patient developed DVT or PE during the initial hospital stay, and no patient died during the operative admission. The
Table 1 Demographic characteristics of patients undergoing total ankle arthroplasty (N = 120) Demographic Data Sex (n [%]) Male Female Mean age (y) 1999 to 2003 2004 to 2010 Race (n [%]) White Black Native American Asian Islander Other Multiple Not stated Underlying diagnosis (%) Primary osteoarthritis Traumatic arthroplasty Other Mean comorbidities 1999 to 2003 2004 to 2010
Patients
p Value
51 (42.5) 69 (57.5) .21 56.1 59.1 80 (66.7) 10 (8.33) 0 0 0 3 (2.5) 0 27 (22.5) 59.2 29.2 11.6 .02 4.0 ± 2.0 4.8 ± 1.9
Based on literature review, only 4 studies have examined the trend of TAA usage rates in the United States (6–9), with the present study the only one to use the NHDS database. The studies by Jiang et al (6), Raikin et al (7), and Singh and Ramachandran (8) relied on data from the NIS database, and the study by Pugely et al (9) used the MedPAR database. Of the 4 studies, the studies by Jiang et al (6) and Singh et al (7) were the only ones to consider the TAA length of hospital stay and whether patients were discharged to home or a rehabilitation facility. Jiang et al (6) also analyzed perioperative complications such as DVT, PE, and the need for transfusions. The study by Pugely et al (9) considered TAA usage in a Medicare population only and excluded patients aged <65 years. Our study has provided an opportunity to compare and contrast results from the 3 different national databases to gain a more comprehensive perspective on the trends in TAA usage and its immediate perioperative outcomes. This study found that the usage rate of TAA had increased from 2.4 to 3.5 cases per 100,000 admissions between the 2 study periods, 1997 to 2003 and 2004 to 2010. Singh and Ramachandran (8) reported a sevenfold increase in the usage of TAA from 1998 to 2010 from 0.13 to 0.84 per 100,000 persons, extrapolated to the entire U.S. population, in contrast to the total number of admissions, such as we used. When comparing the increase in usage between 2 periods, 1998 to 2003 and 2004 to 2010, to better resemble our analysis, the increase in usage was 40% in the study by Singh and Ramachandran (8). This was similar to our finding of 46%. Pugely et al (9) also reported an increase in TAA usage from 0.2 to 1.9 cases per 100,000 Medicare enrollees from 1991 to 2010 and a relative per capita increase of 670%. Raikin et al (7) reported a flat rate of TAAs performed from 2000 to 2006, with a steady increase from 2006 to 2010. Jiang et al (6) similarly concluded that the number of TAAs performed was stable from 2002 to 2004, with a decrease in 2005 and 2006, followed by a steady increase from 2007 to 2011. The 2-year decrease was attributed to the poor results reported with the Agility Total Ankle System (DePuy Orthopaedics, Warsaw, IN) (6). Subsequent studies showing positive outcomes using new-generation TAA implants have corresponded with the recent increase, which will likely continue to result in increases in the number of TAAs performed (11,12). Comparing the analysis of demographic data from our study with that of the studies using the NIS data, the mean age was 57.8 years compared with 61 years reported by Jiang et al (6), Raikin et al (7),
ARTICLE IN PRESS D. Rybalko et al. / The Journal of Foot & Ankle Surgery ■■ (2018) ■■–■■
and Singh and Ramachandran (8). The NIS studies observed an increase in the mean age from 61 to 62 years (6–8). We observed a stable mean age for the patients receiving TAA, with a nonsignificant trend toward an increase in age from 56 to 59 years between our 2 study periods. With respect to the underlying diagnosis, our study found the leading diagnoses to be primary osteoarthritis (59.2%) and traumatic arthropathy (29%) compared with primary osteoarthritis (76.8%) and traumatic arthritis (28.4%) in the study by Jiang et al (6), primary osteoarthritis (62%) and traumatic arthritis (18%) in the study by Singh and Ramachandran (8), and primary osteoarthritis (63%) in the study by Raiken et al (7). The remainder of patients in our study had a primary diagnosis of rheumatoid arthritis. The male to female distribution was also similar, with our study showing a 43% to 57% male to female ratio compared with 46% to 54% breakdown reported by the studies using the NIS database (6–8). When analyzing the number of comorbidities, our findings were consistent with those from the previous studies, with a trend toward an increasing number of comorbidities in patients undergoing TAA. Pugely et al (9) reported an increase in comorbidities from 1.3 to 1.8, especially diabetes mellitus and obesity. Jiang et al (6) reported the mean number of comorbid conditions to be 1.3 for the entire sample period. However, they did not report comorbidities with regard to trends over time (6). Singh and Ramachandran (8) reported an increase in the Deyo-Charlson scores with time. Raikin et al (7) also noted an increase in the number of comorbidities, most notably obesity and diabetes mellitus. Although all the studies showed a trend toward an increased number of comorbidities, our study reported the largest mean number of comorbid conditions (4.5) in patients undergoing TAA, with most prevalent conditions including hypertension, gastroesophageal reflux disease, chronic obstructive pulmonary disease/asthma, and obesity. The number of comorbidities increased from 4.0 to 4.8 from the 1997 to 2003 period to the 2004 to 2010 period. This discrepancy might have resulted from differences in reporting between the different data sets. Studies of the trends in hip and knee joint replacements have noted a similar increase in the number of comorbid conditions, specifically obesity (13). Fehring et al (14) performed a retrospective review of total joint arthroplasty (TJA) patients undergoing total knee and total hip replacements at their institutions from 1990 to 2005. They found that the percentage of patients undergoing TJA with a body mass index (BMI) >30 kg/m2 increased from 30.4% to 52.1% during their study period. The mean BMI of their patients increased from 27.8 to 31.3 kg/m2 (14). With the increase in BMI in the general population and, more importantly, in patients undergoing TJA, it was not surprising that the number of comorbidities in TAA patients has been increasing, including hypertension, diabetes, and hyperlipidemia. However, the increased number of comorbidities results in greater risks of complications and associated medical costs (15). According to Hustedt et al (15), an increasing number of cumulative patient comorbidities was associated with an increased risk of postoperative complications and an increase in overall hospital costs. For example, for a patient with 4 comorbidities, the relative risk of a complication might be as great as 10 and the marginal increased hospital cost might be close to $3000 (15). Thus, although our findings showed a low risk of complications with TAA during the surgical admission, surgeons should be mindful of the increased risks and costs associated with the cumulative number of comorbidities. We found that the mean length of stay decreased over time. Singh and Ramachandran (8) reported that the length of stay had decreased from 3 to 2.5 days compared with a decrease from 3.1 to 2.3 in our study. Jiang et al (6) reported a mean length of stay of 3.2 days, and Raikin et al (7) reported a median stay of 2 days. These findings match the recent trend toward earlier discharge after total knee and
3
total hip arthroplasty (16,17). In part fueled by bundled payments, Froemke et al (18) found an 18% decrease in the length of stay with this type of reimbursement scheme. Additionally, recent studies have shown good outcomes with TJA, with discharge to home within 1 day of surgery showing no increase in the 30- or 90-day readmission rate for well-selected patients, further contributing to this trend (16,19). Finally, better pain control with the use of multimodal anesthesia and aggressive physical therapy have made earlier discharge feasible (16). Regarding discharge disposition, Singh and Ramachandran (8) reported the proportion of patients discharged to an inpatient rehabilitation or skilled nursing facility ranged from 12% to 14%. This result corresponds with our finding of 13% of patients requiring admission to an inpatient rehabilitation or a skilled nursing facility and 87% discharged to home. The low rate of discharge to skilled nursing facilities might have resulted from the more aggressive physical therapy protocols, the use of objective physical therapy milestones in determining clearance for discharge, and the increased availability of home health and home physical therapy (16). Perioperative complications were low in all the studies. Jiang et al (6) reported that TAA was associated with a decreased risk of blood transfusions and overall complications compared with ankle arthrodesis. The rates of blood transfusion (1.3%), DVT (0.3%), and PE (0.2%) reported by Jiang et al (6) were in line with our findings, which revealed a transfusion rate of 2.5% and no DVT or PE. Singh and Ramachandran (8) reported mortality rates from 0% to 0.6%, and our study identified no deaths during the hospital stay. The low rate of complications in our study might have been because the data only represented information from the primary surgical admission. The data did not capture readmissions. Thus, some postoperative complications might have been missed, which was a limitation of our study. However, a more recent analysis of 65 TAAs completed by Borenstein et al (20) confirmed the finding of low perioperative complications. In their study, only 1 patient developed wound dehiscence, which was attributed to an unrelated popliteal artery thrombosis months after surgery. Also, only 1 case of wound infection developed, which occurred in a patient with diabetes, rheumatoid arthritis, obesity, and hypertension (20). Although the NHDS database used in the present study showed similar trends to those reported in studies using the NIS and MedPAR data sets, a question remains regarding what differences might exist among the 3 databases (6–9). A study by Bekkers et al (21) evaluated the degree of difference between the NHDS and NIS databases regarding arthroplasty-related research, specifically focusing on total hip arthroplasty. Bekkers et al (21) found that with respect to epidemiologic data, the 2 databases were similar. However, differences between the 2 databases >10% were found to include the length of stay, wound complications, DVT, and transfusions (21). Their analysis showed a trend toward a greater number of comorbidities and adverse events reported in the NIS database compared with those reported in the NHDS (21). This differed from the findings from our study, which identified a greater number of comorbidities in the NHDS database compared with the reports using the NIS system (6–8). The NHDS surveys a smaller proportion of all hospitalizations in the United States, 1% compared with 20% with the NIS data set (21). However, the 2 databases use different algorithms to arrive at an estimated representation of all hospitalizations in the United States (21). Nevertheless, it remains unclear which data set is more representative of the U.S. hospital admissions and outcomes. The present study had several limitations. Current Procedural Terminology codes, which can provide more procedural details, are not available in the NHDS. The effect of ICD-9 coding changes that occurred during the study period was difficult to ascertain. Furthermore, the database is used to estimate a representation of the U.S. population’s usage of medical resources; however, its accuracy has not been
ARTICLE IN PRESS D. Rybalko et al. / The Journal of Foot & Ankle Surgery ■■ (2018) ■■–■■
4
determined. Additionally, the NHDS database tracks admissions and discharges only. Information regarding patients’ postoperative course after hospital discharge cannot be obtained, and readmissions are recorded as new patient encounters. Thus, intermediate- and longterm outcome information is not available, which could result in an underrepresentation in the rate of PE, DVT, amputation, and mortality after TAA. In conclusion, our findings are in agreement with those from previous reports, demonstrating an increase in usage of TAA in U.S. hospitals. These trends are also in line with the increase in usage of TAA by the international community, as reported by Roukis and Prissel (22) in a study of international registries, including Australia, New Zealand, Sweden, Norway, Finland, and Britain. TAA is most commonly performed for primary ankle osteoarthritis and was associated with short hospital stays, infrequent postoperative rehabilitation requirements, and few perioperative complications. The length of stay has continued to decrease, and the need for postadmission rehabilitation or skilled nursing has remained low and stable. References 1. Chou LB, Coughlin MT, Hansen S, Haskell A, Lundeen G, Saltzman CL, Mann RA. Osteoarthritis of the ankle: the role of arthroplasty. J Am Acad Orthop Surg 16:249–259, 2008. 2. Daniels TR, Younger ASE, Penner M, Wing K, Dryden PJ, Wong H, Glazebrook M. Intermediate-term results of total ankle replacement and ankle arthrodesis. J Bone Joint Surg 96:135–142, 2014. 3. Haddad SL, Coetzee JC, Estok R, Fahrbach K, Banel D, Nalysnyk K. Intermediate and long-term outcomes of total ankle arthroplasty and ankle arthrodesis: a systematic review of the literature. J Bone Joint Surg 89:1899–1905, 2007. 4. Neufeld SK, Lee TH. Total ankle arthroplasty: indications, results, and biomechanical rationale. Am J Orthop (Belle Mead NJ) 29:593–602, 2000. 5. Wood PLR, Prem H, Sutton C. Total ankle replacement: medium-term results in 200 Scandinavian total ankle replacements. J Bone Joint Surg Br 90:605–609, 2008. 6. Jiang JJ, Schipper ON, Whyte N, Koh JL, Toolan BC. Comparison of perioperative complications and hospitalization outcomes after ankle arthrodesis versus total ankle arthroplasty from 2002 to 2011. Foot Ankle Int 36:360–368, 2015.
7. Raikin SM, Rasouli MR, Espandar R, Maltenfort MG. Trends in treatment of advanced ankle arthropathy by total ankle replacement or ankle fusion. Foot Ankle Int 35:216–224, 2014. 8. Singh JA, Ramachandran R. Time trends in total ankle arthroplasty in the USA: a study of the National Inpatient Sample. Clin Rheumatol 35:239–245, 2016. 9. Pugely AJ, Lu X, Amendola A, Callaghan JJ, Martin CT, Cram P. Trends in the use of total ankle replacement and ankle arthrodesis in the United States Medicare population. Foot Ankle Int 35:207–215, 2014. 10. Centers for Disease Control and Prevention. NHDS—National Hospital Discharge Survey Homepage. 2013. Available at https://www.cdc.gov/nchs/nhds/ about_nhds.htm. Accessed January 3, 2018. 11. Hofmann KJ, Shabin ZM, Ferkel E, Jockel J, Slovenkai MP. Salto Talaris total ankle arthroplasty. J Bone Joint Surg 98:2036–2046, 2016. 12. Pangrazzi GJ, Baker EA, Shaheen PJ, Okeagu CN, Fortin PT. Single-surgeon experience and complications of a fixed-bearing total ankle arthroplasty. Foot Ankle Int 39:46–58, 2018. 13. Sing DC, Feeley BT, Tay B, Vail TP, Zhang AL. Age-related trends in hip arthroscopy: a large cross-sectional analysis. Arthrosc J Arthrosc Relat Surg 31:2307–2313, 2015. 14. Fehring TK, Odum SM, Griffin WL, Mason JB, McCoy TH. The obesity epidemic: its effect on total joint arthroplasty. J Arthroplasty 22:71–76, 2007. 15. Hustedt JW, Goltzer O, Bohl DD, Fraser JF, Lara NJ, Spangehl MJ. Calculating the cost and risk of comorbidities in total joint arthroplasty in the United States. J Arthroplasty 32:355–361, 2017. 16. Klingenstein GG, Schoifet SD, Jain RK, Reid JJ, Porat MD, Otegbeye MK. Rapid discharge to home after total knee arthroplasty is safe in eligible Medicare patients. J Arthroplasty 32:3308–3313, 2017. 17. Lovald ST, Ong KL, Malkani AL, Lau EC, Schmier JK, Kurtz SM, Manley MT. Complications, mortality, and costs for outpatient and short-stay total knee arthroplasty patients in comparison to standard-stay patients. J Arthroplasty 29:510–515, 2014. 18. Froemke CC, Wang L, DeHart ML, Williamson RK, Ko LM, Duwelius PJ. Standardizing care and improving quality under a bundled payment initiative for total joint arthroplasty. J Arthroplasty 30:1676–1682, 2015. 19. Courtney PM, Rozell JC, Melnic CM, Lee G-C. Who should not undergo short stay hip and knee arthroplasty? Risk factors associated with major medical complications following primary total joint arthroplasty. J Arthroplasty 30:1–4, 2015. 20. Borenstein TR, Anand K, Li Q, Charlton TP, Thordarson DB. A review of perioperative complications of outpatient total ankle arthroplasty. Foot Ankle Int 39:143–148, 2017. 21. Bekkers S, Bot AGJ, Makarawung D, Neuhaus V, Ring D. The national hospital discharge survey and nationwide inpatient sample: the databases used affect results in THA research. Clin Orthop Relat Res 472:3441–3449, 2014. 22. Roukis TS, Prissel MA. Registry data trends of total ankle replacement use. J Foot Ankle Surg 52:728–735, 2013.
Contents lists available at ScienceDirect
The Journal of Foot & Ankle Surgery j o u r n a l h o m e p a g e : w w w. j f a s . o r g
Original Research
Recent National Trends and Outcomes in Total Ankle Arthroplasty in the United States Danil Rybalko, MD, MS 1, Garrett Schwarzman, MD 1, Vincent Moretti, MD 2 1Resident,
Department of Orthopedics, University of Illinois at Chicago, Chicago, IL Surgeon, Department of Orthopedic Surgery, Drexel University, Philadelphia, PA
2
A R T I C L E
I N F O
Level of Clinical Evidence: 3
Keywords: national hospital discharge survey outcomes total ankle arthroplasty trends usage
A B S T R A C T
Total ankle arthroplasty (TAA) is an evolving option for treating ankle arthritis. We assessed the national trends in usage and perioperative outcomes of TAA in the United States. International Classification of Diseases, 9th revision (ICD-9), codes were used to search the National Hospital Discharge Survey database for TAA from 1997 to 2010. Patient demographics, comorbidities, hospitalization length, discharge disposition, blood transfusion, lower extremity deep vein thrombosis, pulmonary embolism, and mortality data were gathered. Trends were evaluated using linear regression with Pearson’s correlation coefficient, and statistical comparisons were performed using Student’s t test and z-test for proportions with significance at p = .05. We identified 120 patients with TAA. TAA demonstrated a positive correlation with time (r = 0.57), significantly increasing from 2.4 cases per 100,000 admissions from 1997 to 2003 to 3.5 cases per 100,000 from 2004 to 2010 (p = .04). The mean age was 57.8 (range 19 to 83) years. The mean number of comorbidities was 4.5 (range 1 to ≥7). Although patient age remained stable (p = .21), the mean number of comorbidities significantly increased from 4.0 from 1997 to 2003 to 4.8 from 2004 to 2010 (p = .02); 8 patients (6.7%) had diabetes, 71 (59.2%) had primary osteoarthritis, and 35 (29.2%) had posttraumatic arthropathy. The mean length of stay significantly decreased from 3.1 to 2.3 days (p = .03). Three patients (2.5%) required a blood transfusion. No deep vein thrombosis or PE was diagnosed. No patients died during the operative admission; 95 patients (87%) were discharged home and 14 (13%) required a skilled rehabilitation facility. Discharge patterns showed no significant change with time (p = .59). Usage of TAA in the United States has increased nearly 50% over the past 14 years. TAA was associated with shorter hospitalization, infrequent rehabilitation facility requirements, and few perioperative complications. © 2018 by the American College of Foot and Ankle Surgeons. All rights reserved.
Total ankle arthroplasty (TAA) is one of the few operative treatments available for end-stage osteoarthritis of the ankle. Historically, ankle arthrodesis has been the standard treatment (1–4). However, with long-term studies showing progression of arthrosis in adjacent joints, TAA has provided an alternative to address the pain associated with ankle arthritis and avoid placing additional wear and stress on the joints of the mid- and forefoot (1–3,5). Although first-generation TAAs had inferior outcomes, secondgeneration systems have been reported to result in good to excellent clinical outcomes in 88% to 90% of patients (1). Nevertheless, TAA continues to be a technically challenging procedure. Additionally, a number of perioperative complications have been reported, including delayed
Financial Disclosure: None reported. Conflict of Interest: None reported. Address correspondence to: Garrett Schwarzman, MD, Department of Orthopaedic Surgery, University of Illinois at Chicago, Room E270 MSS, Chicago, IL 60612. E-mail address: [email protected] (G. Schwarzman).
wound healing, infection, fracture, tendon injury, sensory deficits, postoperative stiffness, reflex sympathetic dystrophy, amputation, and pulmonary embolism (PE) (1,2,5). Given the improved outcomes of the new-generation TAA systems, few studies have considered the trends of TAA usage and the associated perioperative outcomes during the past decade. Our review identified only 4 studies of the trends of TAA usage in the United States over time, with each focusing to a varying degree on patient characteristics and perioperative outcomes. Three of these studies used the Nationwide Inpatient Sample (NIS) database (6–8). The fourth study used a Medicare Provider Analysis and Review (MedPAR) database (9). Our study focused on the trends of TAA usage in the United States, patient characteristics, and perioperative outcomes and complications from 1997 to 2010 using the data set from the National Center for Health Statistics National Hospital Discharge Survey (NHDS) (10). To the best of our knowledge, ours is the only study that has conducted this type of analysis using the NHDS data to identify trends and outcomes in TAA usage.
1067-2516/$ - see front matter © 2018 by the American College of Foot and Ankle Surgeons. All rights reserved. https://doi.org/10.1053/j.jfas.2018.03.037
ARTICLE IN PRESS D. Rybalko et al. / The Journal of Foot & Ankle Surgery ■■ (2018) ■■–■■
2
Materials and Methods Data for the present study were obtained from NHDS and imported into Microsoft Office Excel 2010 (Microsoft, Seattle, WA) for statistical analysis (10). The NHDS collects data on patient discharges from a range of representative hospitals across the United States, excluding federal, military, and Veterans Affairs hospitals (10). The NHDS includes data from hospitals from all 50 states and the District of Columbia, exclusive of military and Department of Veteran Affairs hospitals (10). These hospitals must have ≥6 beds to be included in the survey (10). Each year, data from ~1% of total hospital admissions from throughout the United States are collected, abstracted, and weighted to provide nationwide estimates (10). Information collected from each sampled hospital record includes age, sex, race/ethnicity, marital status, discharge month, discharge status, days of care, hospital location, hospital size, hospital type (proprietary, government, or church/nonprofit) and ≤15 discharge diagnoses and ≤8 procedures performed during that admission (10). International Classification of Diseases, 9th revision (ICD-9), procedure codes were used to search the NHDS for all patients admitted to the surveyed U.S. hospitals after a primary TAA for each year from 1997 to 2010. Data regarding patient demographics, hospitalization length of stay, discharge disposition, lower extremity deep vein thrombosis (DVT), PE, and mortality were all gathered from the NHDS. The identified cases were divided into 2 groups. The first group included cases from 1998 to 2003 and second group, cases from 2004 to 2010. Statistical analysis was performed using linear regression with Pearson’s correlation coefficient, Student’s t test, z-test for proportions, and χ2 analysis. An αlevel of 0.05 was used for significance determinations.
Table 2 Temporal trends in postoperative length of stay and disposition to home versus skilled nursing facility after total ankle arthroplasty (N = 120) Variable
Mean length of stay (days) Discharge (n [%]) Home Skilled nursing facility
Temporal Trend
Difference (%)
p Value
2.3 ± 1.2
−1.2
.02 .58
55 (89) 7 (11)
4 −4
1999 to 2003 (n = 52)
2004 to 2010 (n = 68)
3.1 ± 2.5 40 (85) 7 (15)
Data in parentheses are percentages.
mean length of stay after TAA was 2.6 (range 1 to 18) days. The length of stay decreased significantly with time, from 3.1 days in 1997 to 2003 to 2.3 days in 2004 to 2010 (p = .03). Of the 120 patients, 95 (87%) were discharged home after their inpatient stay; only 14 patients (13%) required admission or transfer to an inpatient rehabilitation or skilled nursing facility (Table 2). This discharge disposition pattern showed no significant change with time (p = .59). Discussion
Results From 1997 to 2010, 120 patients admitted for TAA were identified. The mean age of the TAA patients was 57.8 (range 19 to 83) years. The group included 51 males and 69 females. The mean number of medical comorbidities was 4.5 (range 1 to ≥7). Although patient age remained stable across the study period (p = .21), the mean number of comorbidities significantly increased from 4.0 from 1997 to 2003 to 4.8 from 2004 to 2010 (p = .02). Of the 120 patients, 8 (6.7%) had diabetes, 71 (59.2%) had ankle osteoarthritis, and 35 (29.2%) had posttraumatic ankle arthropathy, with most of the remaining 11.6% of patients having rheumatoid arthritis (Table 1). The use of TAA demonstrated a moderately positive correlation with time (r = 0.57), with a significant increase from a mean of 2.4 cases per 100,000 hospital admissions from 1997 to 2003 to 3.5 cases per 100,000 hospital admissions from 2004 to 2010 (p = .04). Three patients (2.5%) required a blood transfusion. No patient developed DVT or PE during the initial hospital stay, and no patient died during the operative admission. The
Table 1 Demographic characteristics of patients undergoing total ankle arthroplasty (N = 120) Demographic Data Sex (n [%]) Male Female Mean age (y) 1999 to 2003 2004 to 2010 Race (n [%]) White Black Native American Asian Islander Other Multiple Not stated Underlying diagnosis (%) Primary osteoarthritis Traumatic arthroplasty Other Mean comorbidities 1999 to 2003 2004 to 2010
Patients
p Value
51 (42.5) 69 (57.5) .21 56.1 59.1 80 (66.7) 10 (8.33) 0 0 0 3 (2.5) 0 27 (22.5) 59.2 29.2 11.6 .02 4.0 ± 2.0 4.8 ± 1.9
Based on literature review, only 4 studies have examined the trend of TAA usage rates in the United States (6–9), with the present study the only one to use the NHDS database. The studies by Jiang et al (6), Raikin et al (7), and Singh and Ramachandran (8) relied on data from the NIS database, and the study by Pugely et al (9) used the MedPAR database. Of the 4 studies, the studies by Jiang et al (6) and Singh et al (7) were the only ones to consider the TAA length of hospital stay and whether patients were discharged to home or a rehabilitation facility. Jiang et al (6) also analyzed perioperative complications such as DVT, PE, and the need for transfusions. The study by Pugely et al (9) considered TAA usage in a Medicare population only and excluded patients aged <65 years. Our study has provided an opportunity to compare and contrast results from the 3 different national databases to gain a more comprehensive perspective on the trends in TAA usage and its immediate perioperative outcomes. This study found that the usage rate of TAA had increased from 2.4 to 3.5 cases per 100,000 admissions between the 2 study periods, 1997 to 2003 and 2004 to 2010. Singh and Ramachandran (8) reported a sevenfold increase in the usage of TAA from 1998 to 2010 from 0.13 to 0.84 per 100,000 persons, extrapolated to the entire U.S. population, in contrast to the total number of admissions, such as we used. When comparing the increase in usage between 2 periods, 1998 to 2003 and 2004 to 2010, to better resemble our analysis, the increase in usage was 40% in the study by Singh and Ramachandran (8). This was similar to our finding of 46%. Pugely et al (9) also reported an increase in TAA usage from 0.2 to 1.9 cases per 100,000 Medicare enrollees from 1991 to 2010 and a relative per capita increase of 670%. Raikin et al (7) reported a flat rate of TAAs performed from 2000 to 2006, with a steady increase from 2006 to 2010. Jiang et al (6) similarly concluded that the number of TAAs performed was stable from 2002 to 2004, with a decrease in 2005 and 2006, followed by a steady increase from 2007 to 2011. The 2-year decrease was attributed to the poor results reported with the Agility Total Ankle System (DePuy Orthopaedics, Warsaw, IN) (6). Subsequent studies showing positive outcomes using new-generation TAA implants have corresponded with the recent increase, which will likely continue to result in increases in the number of TAAs performed (11,12). Comparing the analysis of demographic data from our study with that of the studies using the NIS data, the mean age was 57.8 years compared with 61 years reported by Jiang et al (6), Raikin et al (7),
ARTICLE IN PRESS D. Rybalko et al. / The Journal of Foot & Ankle Surgery ■■ (2018) ■■–■■
and Singh and Ramachandran (8). The NIS studies observed an increase in the mean age from 61 to 62 years (6–8). We observed a stable mean age for the patients receiving TAA, with a nonsignificant trend toward an increase in age from 56 to 59 years between our 2 study periods. With respect to the underlying diagnosis, our study found the leading diagnoses to be primary osteoarthritis (59.2%) and traumatic arthropathy (29%) compared with primary osteoarthritis (76.8%) and traumatic arthritis (28.4%) in the study by Jiang et al (6), primary osteoarthritis (62%) and traumatic arthritis (18%) in the study by Singh and Ramachandran (8), and primary osteoarthritis (63%) in the study by Raiken et al (7). The remainder of patients in our study had a primary diagnosis of rheumatoid arthritis. The male to female distribution was also similar, with our study showing a 43% to 57% male to female ratio compared with 46% to 54% breakdown reported by the studies using the NIS database (6–8). When analyzing the number of comorbidities, our findings were consistent with those from the previous studies, with a trend toward an increasing number of comorbidities in patients undergoing TAA. Pugely et al (9) reported an increase in comorbidities from 1.3 to 1.8, especially diabetes mellitus and obesity. Jiang et al (6) reported the mean number of comorbid conditions to be 1.3 for the entire sample period. However, they did not report comorbidities with regard to trends over time (6). Singh and Ramachandran (8) reported an increase in the Deyo-Charlson scores with time. Raikin et al (7) also noted an increase in the number of comorbidities, most notably obesity and diabetes mellitus. Although all the studies showed a trend toward an increased number of comorbidities, our study reported the largest mean number of comorbid conditions (4.5) in patients undergoing TAA, with most prevalent conditions including hypertension, gastroesophageal reflux disease, chronic obstructive pulmonary disease/asthma, and obesity. The number of comorbidities increased from 4.0 to 4.8 from the 1997 to 2003 period to the 2004 to 2010 period. This discrepancy might have resulted from differences in reporting between the different data sets. Studies of the trends in hip and knee joint replacements have noted a similar increase in the number of comorbid conditions, specifically obesity (13). Fehring et al (14) performed a retrospective review of total joint arthroplasty (TJA) patients undergoing total knee and total hip replacements at their institutions from 1990 to 2005. They found that the percentage of patients undergoing TJA with a body mass index (BMI) >30 kg/m2 increased from 30.4% to 52.1% during their study period. The mean BMI of their patients increased from 27.8 to 31.3 kg/m2 (14). With the increase in BMI in the general population and, more importantly, in patients undergoing TJA, it was not surprising that the number of comorbidities in TAA patients has been increasing, including hypertension, diabetes, and hyperlipidemia. However, the increased number of comorbidities results in greater risks of complications and associated medical costs (15). According to Hustedt et al (15), an increasing number of cumulative patient comorbidities was associated with an increased risk of postoperative complications and an increase in overall hospital costs. For example, for a patient with 4 comorbidities, the relative risk of a complication might be as great as 10 and the marginal increased hospital cost might be close to $3000 (15). Thus, although our findings showed a low risk of complications with TAA during the surgical admission, surgeons should be mindful of the increased risks and costs associated with the cumulative number of comorbidities. We found that the mean length of stay decreased over time. Singh and Ramachandran (8) reported that the length of stay had decreased from 3 to 2.5 days compared with a decrease from 3.1 to 2.3 in our study. Jiang et al (6) reported a mean length of stay of 3.2 days, and Raikin et al (7) reported a median stay of 2 days. These findings match the recent trend toward earlier discharge after total knee and
3
total hip arthroplasty (16,17). In part fueled by bundled payments, Froemke et al (18) found an 18% decrease in the length of stay with this type of reimbursement scheme. Additionally, recent studies have shown good outcomes with TJA, with discharge to home within 1 day of surgery showing no increase in the 30- or 90-day readmission rate for well-selected patients, further contributing to this trend (16,19). Finally, better pain control with the use of multimodal anesthesia and aggressive physical therapy have made earlier discharge feasible (16). Regarding discharge disposition, Singh and Ramachandran (8) reported the proportion of patients discharged to an inpatient rehabilitation or skilled nursing facility ranged from 12% to 14%. This result corresponds with our finding of 13% of patients requiring admission to an inpatient rehabilitation or a skilled nursing facility and 87% discharged to home. The low rate of discharge to skilled nursing facilities might have resulted from the more aggressive physical therapy protocols, the use of objective physical therapy milestones in determining clearance for discharge, and the increased availability of home health and home physical therapy (16). Perioperative complications were low in all the studies. Jiang et al (6) reported that TAA was associated with a decreased risk of blood transfusions and overall complications compared with ankle arthrodesis. The rates of blood transfusion (1.3%), DVT (0.3%), and PE (0.2%) reported by Jiang et al (6) were in line with our findings, which revealed a transfusion rate of 2.5% and no DVT or PE. Singh and Ramachandran (8) reported mortality rates from 0% to 0.6%, and our study identified no deaths during the hospital stay. The low rate of complications in our study might have been because the data only represented information from the primary surgical admission. The data did not capture readmissions. Thus, some postoperative complications might have been missed, which was a limitation of our study. However, a more recent analysis of 65 TAAs completed by Borenstein et al (20) confirmed the finding of low perioperative complications. In their study, only 1 patient developed wound dehiscence, which was attributed to an unrelated popliteal artery thrombosis months after surgery. Also, only 1 case of wound infection developed, which occurred in a patient with diabetes, rheumatoid arthritis, obesity, and hypertension (20). Although the NHDS database used in the present study showed similar trends to those reported in studies using the NIS and MedPAR data sets, a question remains regarding what differences might exist among the 3 databases (6–9). A study by Bekkers et al (21) evaluated the degree of difference between the NHDS and NIS databases regarding arthroplasty-related research, specifically focusing on total hip arthroplasty. Bekkers et al (21) found that with respect to epidemiologic data, the 2 databases were similar. However, differences between the 2 databases >10% were found to include the length of stay, wound complications, DVT, and transfusions (21). Their analysis showed a trend toward a greater number of comorbidities and adverse events reported in the NIS database compared with those reported in the NHDS (21). This differed from the findings from our study, which identified a greater number of comorbidities in the NHDS database compared with the reports using the NIS system (6–8). The NHDS surveys a smaller proportion of all hospitalizations in the United States, 1% compared with 20% with the NIS data set (21). However, the 2 databases use different algorithms to arrive at an estimated representation of all hospitalizations in the United States (21). Nevertheless, it remains unclear which data set is more representative of the U.S. hospital admissions and outcomes. The present study had several limitations. Current Procedural Terminology codes, which can provide more procedural details, are not available in the NHDS. The effect of ICD-9 coding changes that occurred during the study period was difficult to ascertain. Furthermore, the database is used to estimate a representation of the U.S. population’s usage of medical resources; however, its accuracy has not been
ARTICLE IN PRESS D. Rybalko et al. / The Journal of Foot & Ankle Surgery ■■ (2018) ■■–■■
4
determined. Additionally, the NHDS database tracks admissions and discharges only. Information regarding patients’ postoperative course after hospital discharge cannot be obtained, and readmissions are recorded as new patient encounters. Thus, intermediate- and longterm outcome information is not available, which could result in an underrepresentation in the rate of PE, DVT, amputation, and mortality after TAA. In conclusion, our findings are in agreement with those from previous reports, demonstrating an increase in usage of TAA in U.S. hospitals. These trends are also in line with the increase in usage of TAA by the international community, as reported by Roukis and Prissel (22) in a study of international registries, including Australia, New Zealand, Sweden, Norway, Finland, and Britain. TAA is most commonly performed for primary ankle osteoarthritis and was associated with short hospital stays, infrequent postoperative rehabilitation requirements, and few perioperative complications. The length of stay has continued to decrease, and the need for postadmission rehabilitation or skilled nursing has remained low and stable. References 1. Chou LB, Coughlin MT, Hansen S, Haskell A, Lundeen G, Saltzman CL, Mann RA. Osteoarthritis of the ankle: the role of arthroplasty. J Am Acad Orthop Surg 16:249–259, 2008. 2. Daniels TR, Younger ASE, Penner M, Wing K, Dryden PJ, Wong H, Glazebrook M. Intermediate-term results of total ankle replacement and ankle arthrodesis. J Bone Joint Surg 96:135–142, 2014. 3. Haddad SL, Coetzee JC, Estok R, Fahrbach K, Banel D, Nalysnyk K. Intermediate and long-term outcomes of total ankle arthroplasty and ankle arthrodesis: a systematic review of the literature. J Bone Joint Surg 89:1899–1905, 2007. 4. Neufeld SK, Lee TH. Total ankle arthroplasty: indications, results, and biomechanical rationale. Am J Orthop (Belle Mead NJ) 29:593–602, 2000. 5. Wood PLR, Prem H, Sutton C. Total ankle replacement: medium-term results in 200 Scandinavian total ankle replacements. J Bone Joint Surg Br 90:605–609, 2008. 6. Jiang JJ, Schipper ON, Whyte N, Koh JL, Toolan BC. Comparison of perioperative complications and hospitalization outcomes after ankle arthrodesis versus total ankle arthroplasty from 2002 to 2011. Foot Ankle Int 36:360–368, 2015.
7. Raikin SM, Rasouli MR, Espandar R, Maltenfort MG. Trends in treatment of advanced ankle arthropathy by total ankle replacement or ankle fusion. Foot Ankle Int 35:216–224, 2014. 8. Singh JA, Ramachandran R. Time trends in total ankle arthroplasty in the USA: a study of the National Inpatient Sample. Clin Rheumatol 35:239–245, 2016. 9. Pugely AJ, Lu X, Amendola A, Callaghan JJ, Martin CT, Cram P. Trends in the use of total ankle replacement and ankle arthrodesis in the United States Medicare population. Foot Ankle Int 35:207–215, 2014. 10. Centers for Disease Control and Prevention. NHDS—National Hospital Discharge Survey Homepage. 2013. Available at https://www.cdc.gov/nchs/nhds/ about_nhds.htm. Accessed January 3, 2018. 11. Hofmann KJ, Shabin ZM, Ferkel E, Jockel J, Slovenkai MP. Salto Talaris total ankle arthroplasty. J Bone Joint Surg 98:2036–2046, 2016. 12. Pangrazzi GJ, Baker EA, Shaheen PJ, Okeagu CN, Fortin PT. Single-surgeon experience and complications of a fixed-bearing total ankle arthroplasty. Foot Ankle Int 39:46–58, 2018. 13. Sing DC, Feeley BT, Tay B, Vail TP, Zhang AL. Age-related trends in hip arthroscopy: a large cross-sectional analysis. Arthrosc J Arthrosc Relat Surg 31:2307–2313, 2015. 14. Fehring TK, Odum SM, Griffin WL, Mason JB, McCoy TH. The obesity epidemic: its effect on total joint arthroplasty. J Arthroplasty 22:71–76, 2007. 15. Hustedt JW, Goltzer O, Bohl DD, Fraser JF, Lara NJ, Spangehl MJ. Calculating the cost and risk of comorbidities in total joint arthroplasty in the United States. J Arthroplasty 32:355–361, 2017. 16. Klingenstein GG, Schoifet SD, Jain RK, Reid JJ, Porat MD, Otegbeye MK. Rapid discharge to home after total knee arthroplasty is safe in eligible Medicare patients. J Arthroplasty 32:3308–3313, 2017. 17. Lovald ST, Ong KL, Malkani AL, Lau EC, Schmier JK, Kurtz SM, Manley MT. Complications, mortality, and costs for outpatient and short-stay total knee arthroplasty patients in comparison to standard-stay patients. J Arthroplasty 29:510–515, 2014. 18. Froemke CC, Wang L, DeHart ML, Williamson RK, Ko LM, Duwelius PJ. Standardizing care and improving quality under a bundled payment initiative for total joint arthroplasty. J Arthroplasty 30:1676–1682, 2015. 19. Courtney PM, Rozell JC, Melnic CM, Lee G-C. Who should not undergo short stay hip and knee arthroplasty? Risk factors associated with major medical complications following primary total joint arthroplasty. J Arthroplasty 30:1–4, 2015. 20. Borenstein TR, Anand K, Li Q, Charlton TP, Thordarson DB. A review of perioperative complications of outpatient total ankle arthroplasty. Foot Ankle Int 39:143–148, 2017. 21. Bekkers S, Bot AGJ, Makarawung D, Neuhaus V, Ring D. The national hospital discharge survey and nationwide inpatient sample: the databases used affect results in THA research. Clin Orthop Relat Res 472:3441–3449, 2014. 22. Roukis TS, Prissel MA. Registry data trends of total ankle replacement use. J Foot Ankle Surg 52:728–735, 2013.