Orthopedics

Prior reviews have demonstrated a strong link between surgeon and hospital TSA volume and outcomes of the procedure.^8–10 Somerson and colleagues¹¹ investigated fellowship training among surgeons performing TSA in 2012 and found that only 28% had completed a shoulder and elbow fellowship. In addition to prior analyses2, ¹², Somerson and colleagues confirmed a persistent geographic variation in utilization of TSA.¹¹ In conjunction with the evolution of shoulder arthroplasty, dedicated shoulder and elbow fellowship training has expanded. With a shift toward specialization in care, nearly 90% of orthopedic surgery residents plan to pursue shoulder and elbow fellowships, comprising 4.6% of (42/897) of available positions.¹³  

What remains unknown is the specialization of surgeons performing TSA, the regularity of their arthroplasty volume, and trends in TSA specialization over time. Therefore, this study aims to (a) identify surgeons performing shoulder arthroplasty and cohort changes over time, (b) determine the case profile of surgeons consistently performing shoulder arthroplasty, and (c) establish the characteristics of shoulder arthroplasty surgeons with a specific focus on fellowship training. 

METHODS

Prior to collecting surgeon-specific data, we identified surgeons performing TSA through the Centers for Medicare and Medicaid Services’ public release of “Medicare Provider Utilization and Payment Data: Physician and Other Supplier.”¹⁴ Datasets from 2012, 2013, and 2014 were used to identify all surgeons performing >10 TSAs (Current Procedural Terminology [CPT] Code 23472) during at least 1 of those years. This dataset provides the name, identification number, address, and all billing (by volume) for each unique CPT code submitted ≥10 times in a calendar year.

Once the cohort of surgeons had been generated, the number of surgeons consistently performing TSA year-over-year was determined. This allowed for an analysis of the consistency with which surgeons are performing moderate- to high-volume TSA. To form a case profile of surgeons performing TSA and observe how this shifted over time, a count and a description of each CPT code submitted by each surgeon was identified. To maintain patient privacy, only those claims made >10 times are reported for a provider (both physicians and physician-extenders are included in this dataset). First, all CPT codes were reviewed and tagged as surgical or non-surgical events. Then, every procedural CPT code identified was reviewed and categorized based upon anatomic location and procedure (eg, total knee arthroplasty [TKA]). It is important to note that all claims in this dataset are limited to those patients participating in Medicare’s fee-for-service program. 

Specialization was defined as the number of categorized procedures as a percentage of all procedures performed on Medicare patients. The trends for national, regional, and individual specialization of TSA, arthroplasty (major joint), and shoulder procedures were determined.

Continue to: To investigate the characteristics of surgeons...

To investigate the characteristics of surgeons consistently performing TSA, all surgeons performing a minimum of 11 TSA in Medicare fee-for-service beneficiaries in all years between 2012 and 2014 were identified. Such surgeons were defined as consistent TSA surgeons. Investigation of this cohort included a web-based search of their self-reported post-graduate fellowship training and year of graduation from medical school. Using these data, the percentage of surgeons performing TSA who underwent formal shoulder and elbow training was determined. In addition, the impact of fellowship training on shoulder specialization and practice location was determined. Surgeons who had completed multiple fellowships were categorized under all of them. As such, there may be some duplication of surgeons in the comparisons. In addition, other potential characteristics of shoulder and elbow fellowship-trained surgeons were investigated: number of regional shoulder surgeons, urban area, total number of Medicare beneficiaries, average reimbursement for TSA, ethnicity of Medicare beneficiaries, and percentage of Medicare patients eligible for Medicaid. Geographic regions were defined by the Dartmouth Atlas and assigned by hospital referral region.¹⁵ These defined regions were used to assess the beneficiaries (number and characteristics) that individual surgeons were likely serving. The United States Census Bureau characterization of zip code-based regions as urban areas (population >50,000), urban clusters (2500 to 50,000), and rural region (<2500) was used to categorize practice location.¹⁶ 

Descriptive statistics were used initially to report these findings. To analyze predictors of utilization and specialization, comparative statistics were undertaken. For comparison of binomial variables between groups, a χ2analysis was utilized. For continuous variables, data normality was assessed. A skewness and kurtosis <2 and 12, respectively, was considered to represent parametric data. For parametric data, the mean was reported; conversely, the median is reported for non-parametric data. To assess continuous variables between groups, a t test or a Wilcoxon rank-sum test was used for parametric and non-parametric distributions, respectively. 

RESULTS

Between 2012 and 2014, 1374 surgeons (39 females [2.8%]) performed >10 TSA in Medicare patients in at least 1 year, for a combined total of 71,973 TSAs (Table 1). In 2012, only 834 surgeons (13 females [1.6%]) performed a minimum of 10 TSA in Medicare patients (21,137 arthroplasties; 25.3 per surgeon). This increased to 1078 surgeons (33 females [3.1%]; P = .04) performing 26,865 TSA (24.92 per surgeon) in 2014. Utilization of non-physician assistants in TSA also increased significantly over this period, with 307 assisting in 6885 TSAs (22.4 per provider) in 2012 and 465 assisting 10,433 TSA (22.4 per provider) in 2014. When all procedures were considered, including those performed at outpatient visits, 1319 physicians (95.9% of cohort) were active in 2012—providing either surgical procedures or outpatient consults to the Medicare population. Yet, only 63.2% performed >10 TSA in Medicare patients. The number of active surgeons performing TSA increased to 79.6% (1078/1353) in 2014 (P < .001). 

Table 1. Trends in the Number of Providers Performing TSA between 2012 and 2014*

* Included are the number of arthroplasties and total procedures over time among this cohort. The number of active providers, determined by billing Medicare for office or surgical procedures within that year, is reported.

Abbreviation: TSA, total shoulder arthroplasty.  

In addition to TSA, this cohort of surgeons submitted 240 unique CPT codes with case volumes >10 annually over the 3-year study period. Of these, 80.2% (1102/1374) of surgeons performed non-arthroplasty shoulder procedures on Medicare patients, for a combined total of 202,335 procedures over the 3-year study period (Table 2). A significant proportion of these procedures were arthroscopic debridement (60,014 procedures performed by 908 surgeons) and arthroscopic rotator cuff repair (47,089 procedures performed by 809 surgeons). Just under half (49.1%; 674/1374) of surgeons performing TSA also performed TKA during this period (77,873 arthroplasties). Fewer surgeons (27.8%; 382/1374) performed total hip arthroplasty during this period (27,322 arthroplasties). Other procedure types that this group of surgeons routinely performed on Medicare patients were repairs of traumatic injuries (29.8%), non-arthroplasty knee surgeries (27.2%), elbow procedures (19.6%), and hand surgery (15.4%). By case load, non-arthroplasty shoulder procedures consisted of 43% of Medicare volume over the study period (Figure 1). Between 2012 and 2014, the average proportion of Medicare cases that were shoulder arthroplasties increased from 13.8% (21,137/152,862) to 16.7% (26,865/160,851; P = .001). Shoulder arthroplasty constituted 100% of the Medicare surgical case volume for 67 (4.9%; 67/1374) of the surgeons.    

Table 2. Case Volumes over Time with All Procedures Categorized by Anatomic Region and Arthroplasty vs Non-arthroplasty*

* Procedures of interest with high case volumes are reported individually.

Only 44.3% (609/1374) of surgeons performed TSA in a minimum of 11 Medicare patients in all 3 years of the study period (consistent providers of TSA), providing a total of 55,538 TSA (77.2%; 55,538/71,973). When fellowship training was evaluated, 191 (31.4%; 191/609) of these surgeons were shoulder and elbow fellowship trained (21,444 TSA; 38.6%; Table 3). More than one-third (36.6%; 223/609) had completed a sports surgery fellowship (18,899 TSA; 34.0%). Surgeons trained in hand surgery (12.5%; 76/609) and adult reconstruction (5.3%; 32/610) also made contributions to meeting the TSA demand with 6971 (12.6%) and 2485 (4.5%) TSA, respectively. One-fifth of this cohort (18.1%; 110/609) had unknown fellowship training: they either reported no fellowship (13.6%; 83/609) or did not specify the type of training (4.4%; 27/609). Shoulder and elbow fellowship-trained surgeons performed more TSA (median: 89.0 TSA per surgeon between 2012 and 2014) than surgeons without shoulder and elbow fellowship training (median: 67.0 TSA per surgeon; P < 0.001). More than one-third (37%) of shoulder and elbow fellowship-trained surgeons’ surgical case volume was comprised of TSA, with an additional 35% from non-arthroplasty shoulder procedures (Figure 2). In order for the current supply of shoulder and elbow fellowship-trained surgeons to meet the Medicare TSA demand, each fellowship graduate would have to perform 140.6 TSA in Medicare patients annually. Shoulder and elbow fellowship-trained surgeons were more likely to practice in referral regions with an increased Medicare population (P < .001), an increased number of surgeons performing TSA (P < .001), and a higher proportion of Medicaid-eligible patients (P = .01; Table 4). Shoulder and elbow fellowship-trained surgeons (18.7 years post-medical school graduation) were also earlier in their careers than other consistent TSA surgeons (23.1 years post-graduation; P < .001). 

Table 3. A Representation of Fellowships Among TSA Surgeons and Their Shoulder Arthroplasty Case Load*

* Not all fellowships (eg, oncology) included due to small numbers. Also, many surgeons performed multiple fellowships.

Abbreviations: SA, shoulder arthroplasty; TSA, total shoulder arthroplasty.

Table 4. Breakdown of Geographic Characteristics of Orthopedic Surgeons Consistently

Performing TSA between 2012 and 2014 Stratified by Fellowship Training

Abbreviations: HRR, hospital referral region; TSA, total shoulder arthroplasty.

Continue to: DISCUSSION...

DISCUSSION

Utilization of TSA has continued to rise; however, access to this cost-effective procedure was recently demonstrated to be limited.¹¹ In a separate analysis, we established the continued rise in use of TSA in the Medicare population, coupled with an increase in the number of surgeons routinely performing TSA.² Multiple analyses have demonstrated the importance of high-volume surgeons and hospitals familiar with the intricacies of shoulder arthroplasty concepts in minimizing complications, improving the quality and decreasing the cost of TSA.^6,10,17 Specifically, Singh and colleagues18 demonstrated from a multi-center registry that surgeons and hospitals with greater shoulder arthroplasty volumes had decreased intra-operative blood loss, operative time, and hospital length of stay. As the demand for TSA, both anatomic and reverse, continues to rise, it is imperative that the healthcare delivery system is optimized to provide the best possible care. Before we can determine whether specialized training in shoulder arthroplasty influences surgical outcomes, characteristics and training of surgeons performing TSA should be described. 

The number of surgeons performing >10 TSA in the Medicare population rose significantly between 2012 and 2014 (29.3%). However, the number of TSAs per surgeon over this time period remained consistent (approximately 25 per surgeon). Furthermore, the increase in the number of surgeons performing a reportable volume of TSA by 2014 was from the addition of already active surgeons (ie, the growth in TSA was not from the addition of newly trained arthroplasty surgeons but originated from the existing orthopedic surgeon workforce). In a recently published analysis, Somerson and colleagues, ¹¹ using this same dataset, demonstrated persistent limitations in access to high-volume TSA surgeons. In a more recent analysis, we showed that while still lacking for some patients, access to a high-volume TSA surgeon has improved significantly over the past 3 years, with 96.9% of the United States population residing within 200 kilometers of a high-volume TSA surgeon (>20 Medicare cases).² This analysis validates those findings, with the caveat that the average annual volume per surgeon is not increasing. What remains unknown, due to limitations of this dataset, is how many surgeons are not identified because they are performing ≤10 TSA each year or are performing TSA in non-Medicare patients. 

With the specialization of healthcare delivery, specifically in orthopedics, it is imperative that mechanisms for providing specialty-focused care be established. However, the proportion of their practice that surgeons dedicate to TSA was unknown. This study demonstrates that this proportion is increasing. Including non-arthroplasty procedures, more than half (58%) of the procedures performed by this surgeon cohort were shoulder-specific. Furthermore, this analysis demonstrates that surgeons performing TSA have significant case diversity, including nearly half of the cohort performing TKA. Repeated evidence has demonstrated the effect of case volume on improved outcomes following orthopedic procedures.^8,19–21 The pre-existing location-based model for delivering orthopedic care supports case diversity; however, this model continues to be challenged with high-volume centers of excellence and patient travel.^22–24 Hip and knee arthroplasty experienced a similar surge in demand, with a subsequent shift in care to high-volume surgeons and centers.²⁵ Shoulder and elbow fellowship-trained surgeons would need to nearly quadruple their current Medicare TSA volume to meet the entire current demand for TSA in the Medicare population (and this does not account for TSA performed by very low-volume surgeons not included in this cohort). With increased utilization of TSA, policymakers and the orthopedic community must determine the structure of delivery (centers of excellence or medium-volume disseminated throughout the country) that is optimal. 

For those surgeons consistently performing TSA over the study period, fellowship training was diverse. While the current focus in orthopedics is on case volume, research in other specialties, namely general surgery, has provided repeated evidence that surgical specialization (more so than high case volume) provides improved outcomes.^26–29Furthermore, Leopold and colleagues³⁰ demonstrated an inverse relationship between competency in performing a procedure and confidence in one’s ability to do so. In their study, educational intervention provided improved competency in the procedure. Less than one-third (29.8%) of TSA in this cohort were performed by a shoulder and elbow fellowship-trained surgeon consistently performing this procedure. Approximately another quarter (26.2%) were performed by consistent TSA surgeons trained in sports surgery. Meanwhile, 34.6% of TSA in this study cohort were performed by a surgeon who did not consistently meet the minimum threshold in all study years (16,435 TSA; 22.8%) or by a surgeon performing TSA without fellowship training (8,489 TSA; 11.8%). There has been a trend toward orthopedic subspecialty training with an increased demand for fellowship-trained surgeons.³¹ Despite this and the complexities of TSA, many continue to be performed by surgeons with an inconsistent volume and those without arthroplasty-specific fellowship training. The available evidence supports a push toward the fellowship-trained, high-volume TSA surgeon in providing reproducible high-quality shoulder arthroplasty care. For now, that surgeon is more likely to be earlier in his/her career and reside in large, referral-based centers surrounded by other surgeons performing TSA.

These findings must be considered in the light of the study limitations. First, this is a large publicly available database. While this type of database provides a unique opportunity to assess the geographic distributions and characteristics of orthopedic surgeons, specifically those performing TSA, it completely prevents any assessment of the relationship between these findings and quality. As such, while the reader may generate hypotheses regarding the implications of our findings on the quality of TSA delivery, the true effects cannot be determined. In the same vein, for the purpose of privacy, surgeons performing ≤10 TSA were not included in this dataset. This limitation prevents the identification of low-volume TSA surgeons. Also, it is likely that the observed increase in surgeons over time is likely a reflection of small increases in volume for surgeons already performing TSA. Lastly, a web-based search was undertaken to identify surgeons’ self-reported fellowship training. The results of this web-based search could not be validated, and it is possible that fellowship training, or the lack thereof, was mischaracterized and simply not obtainable through a web-based search. Furthermore, it is not possible to fully assess the extent of high-quality TSA training in these various fellowships.

CONCLUSION

In just the past decade, the utilization of TSA in the Medicare population has increased significantly. However, this increase was not achieved by the addition of highly specialized, high-volume surgeons but by the addition of many surgeons performing lower numbers of TSA surgeries. Furthermore, for those performing this cost-effective procedure, TSA constitutes a relatively small proportion of the surgeries they perform. Shoulder and elbow fellowship-trained surgeons currently account for a low percentage of the overall number of surgeons performing TSA. The implications of these findings must be considered and investigated.

Prior reviews have demonstrated a strong link between surgeon and hospital TSA volume and outcomes of the procedure.^8–10 Somerson and colleagues¹¹ investigated fellowship training among surgeons performing TSA in 2012 and found that only 28% had completed a shoulder and elbow fellowship. In addition to prior analyses2, ¹², Somerson and colleagues confirmed a persistent geographic variation in utilization of TSA.¹¹ In conjunction with the evolution of shoulder arthroplasty, dedicated shoulder and elbow fellowship training has expanded. With a shift toward specialization in care, nearly 90% of orthopedic surgery residents plan to pursue shoulder and elbow fellowships, comprising 4.6% of (42/897) of available positions.¹³  

What remains unknown is the specialization of surgeons performing TSA, the regularity of their arthroplasty volume, and trends in TSA specialization over time. Therefore, this study aims to (a) identify surgeons performing shoulder arthroplasty and cohort changes over time, (b) determine the case profile of surgeons consistently performing shoulder arthroplasty, and (c) establish the characteristics of shoulder arthroplasty surgeons with a specific focus on fellowship training. 

METHODS

Prior to collecting surgeon-specific data, we identified surgeons performing TSA through the Centers for Medicare and Medicaid Services’ public release of “Medicare Provider Utilization and Payment Data: Physician and Other Supplier.”¹⁴ Datasets from 2012, 2013, and 2014 were used to identify all surgeons performing >10 TSAs (Current Procedural Terminology [CPT] Code 23472) during at least 1 of those years. This dataset provides the name, identification number, address, and all billing (by volume) for each unique CPT code submitted ≥10 times in a calendar year.

Once the cohort of surgeons had been generated, the number of surgeons consistently performing TSA year-over-year was determined. This allowed for an analysis of the consistency with which surgeons are performing moderate- to high-volume TSA. To form a case profile of surgeons performing TSA and observe how this shifted over time, a count and a description of each CPT code submitted by each surgeon was identified. To maintain patient privacy, only those claims made >10 times are reported for a provider (both physicians and physician-extenders are included in this dataset). First, all CPT codes were reviewed and tagged as surgical or non-surgical events. Then, every procedural CPT code identified was reviewed and categorized based upon anatomic location and procedure (eg, total knee arthroplasty [TKA]). It is important to note that all claims in this dataset are limited to those patients participating in Medicare’s fee-for-service program. 

Specialization was defined as the number of categorized procedures as a percentage of all procedures performed on Medicare patients. The trends for national, regional, and individual specialization of TSA, arthroplasty (major joint), and shoulder procedures were determined.

Continue to: To investigate the characteristics of surgeons...

To investigate the characteristics of surgeons consistently performing TSA, all surgeons performing a minimum of 11 TSA in Medicare fee-for-service beneficiaries in all years between 2012 and 2014 were identified. Such surgeons were defined as consistent TSA surgeons. Investigation of this cohort included a web-based search of their self-reported post-graduate fellowship training and year of graduation from medical school. Using these data, the percentage of surgeons performing TSA who underwent formal shoulder and elbow training was determined. In addition, the impact of fellowship training on shoulder specialization and practice location was determined. Surgeons who had completed multiple fellowships were categorized under all of them. As such, there may be some duplication of surgeons in the comparisons. In addition, other potential characteristics of shoulder and elbow fellowship-trained surgeons were investigated: number of regional shoulder surgeons, urban area, total number of Medicare beneficiaries, average reimbursement for TSA, ethnicity of Medicare beneficiaries, and percentage of Medicare patients eligible for Medicaid. Geographic regions were defined by the Dartmouth Atlas and assigned by hospital referral region.¹⁵ These defined regions were used to assess the beneficiaries (number and characteristics) that individual surgeons were likely serving. The United States Census Bureau characterization of zip code-based regions as urban areas (population >50,000), urban clusters (2500 to 50,000), and rural region (<2500) was used to categorize practice location.¹⁶ 

Descriptive statistics were used initially to report these findings. To analyze predictors of utilization and specialization, comparative statistics were undertaken. For comparison of binomial variables between groups, a χ2analysis was utilized. For continuous variables, data normality was assessed. A skewness and kurtosis <2 and 12, respectively, was considered to represent parametric data. For parametric data, the mean was reported; conversely, the median is reported for non-parametric data. To assess continuous variables between groups, a t test or a Wilcoxon rank-sum test was used for parametric and non-parametric distributions, respectively. 

RESULTS

Between 2012 and 2014, 1374 surgeons (39 females [2.8%]) performed >10 TSA in Medicare patients in at least 1 year, for a combined total of 71,973 TSAs (Table 1). In 2012, only 834 surgeons (13 females [1.6%]) performed a minimum of 10 TSA in Medicare patients (21,137 arthroplasties; 25.3 per surgeon). This increased to 1078 surgeons (33 females [3.1%]; P = .04) performing 26,865 TSA (24.92 per surgeon) in 2014. Utilization of non-physician assistants in TSA also increased significantly over this period, with 307 assisting in 6885 TSAs (22.4 per provider) in 2012 and 465 assisting 10,433 TSA (22.4 per provider) in 2014. When all procedures were considered, including those performed at outpatient visits, 1319 physicians (95.9% of cohort) were active in 2012—providing either surgical procedures or outpatient consults to the Medicare population. Yet, only 63.2% performed >10 TSA in Medicare patients. The number of active surgeons performing TSA increased to 79.6% (1078/1353) in 2014 (P < .001). 

Table 1. Trends in the Number of Providers Performing TSA between 2012 and 2014*

* Included are the number of arthroplasties and total procedures over time among this cohort. The number of active providers, determined by billing Medicare for office or surgical procedures within that year, is reported.

Abbreviation: TSA, total shoulder arthroplasty.  

In addition to TSA, this cohort of surgeons submitted 240 unique CPT codes with case volumes >10 annually over the 3-year study period. Of these, 80.2% (1102/1374) of surgeons performed non-arthroplasty shoulder procedures on Medicare patients, for a combined total of 202,335 procedures over the 3-year study period (Table 2). A significant proportion of these procedures were arthroscopic debridement (60,014 procedures performed by 908 surgeons) and arthroscopic rotator cuff repair (47,089 procedures performed by 809 surgeons). Just under half (49.1%; 674/1374) of surgeons performing TSA also performed TKA during this period (77,873 arthroplasties). Fewer surgeons (27.8%; 382/1374) performed total hip arthroplasty during this period (27,322 arthroplasties). Other procedure types that this group of surgeons routinely performed on Medicare patients were repairs of traumatic injuries (29.8%), non-arthroplasty knee surgeries (27.2%), elbow procedures (19.6%), and hand surgery (15.4%). By case load, non-arthroplasty shoulder procedures consisted of 43% of Medicare volume over the study period (Figure 1). Between 2012 and 2014, the average proportion of Medicare cases that were shoulder arthroplasties increased from 13.8% (21,137/152,862) to 16.7% (26,865/160,851; P = .001). Shoulder arthroplasty constituted 100% of the Medicare surgical case volume for 67 (4.9%; 67/1374) of the surgeons.    

Table 2. Case Volumes over Time with All Procedures Categorized by Anatomic Region and Arthroplasty vs Non-arthroplasty*

* Procedures of interest with high case volumes are reported individually.

Only 44.3% (609/1374) of surgeons performed TSA in a minimum of 11 Medicare patients in all 3 years of the study period (consistent providers of TSA), providing a total of 55,538 TSA (77.2%; 55,538/71,973). When fellowship training was evaluated, 191 (31.4%; 191/609) of these surgeons were shoulder and elbow fellowship trained (21,444 TSA; 38.6%; Table 3). More than one-third (36.6%; 223/609) had completed a sports surgery fellowship (18,899 TSA; 34.0%). Surgeons trained in hand surgery (12.5%; 76/609) and adult reconstruction (5.3%; 32/610) also made contributions to meeting the TSA demand with 6971 (12.6%) and 2485 (4.5%) TSA, respectively. One-fifth of this cohort (18.1%; 110/609) had unknown fellowship training: they either reported no fellowship (13.6%; 83/609) or did not specify the type of training (4.4%; 27/609). Shoulder and elbow fellowship-trained surgeons performed more TSA (median: 89.0 TSA per surgeon between 2012 and 2014) than surgeons without shoulder and elbow fellowship training (median: 67.0 TSA per surgeon; P < 0.001). More than one-third (37%) of shoulder and elbow fellowship-trained surgeons’ surgical case volume was comprised of TSA, with an additional 35% from non-arthroplasty shoulder procedures (Figure 2). In order for the current supply of shoulder and elbow fellowship-trained surgeons to meet the Medicare TSA demand, each fellowship graduate would have to perform 140.6 TSA in Medicare patients annually. Shoulder and elbow fellowship-trained surgeons were more likely to practice in referral regions with an increased Medicare population (P < .001), an increased number of surgeons performing TSA (P < .001), and a higher proportion of Medicaid-eligible patients (P = .01; Table 4). Shoulder and elbow fellowship-trained surgeons (18.7 years post-medical school graduation) were also earlier in their careers than other consistent TSA surgeons (23.1 years post-graduation; P < .001). 

Table 3. A Representation of Fellowships Among TSA Surgeons and Their Shoulder Arthroplasty Case Load*

* Not all fellowships (eg, oncology) included due to small numbers. Also, many surgeons performed multiple fellowships.

Abbreviations: SA, shoulder arthroplasty; TSA, total shoulder arthroplasty.

Table 4. Breakdown of Geographic Characteristics of Orthopedic Surgeons Consistently

Performing TSA between 2012 and 2014 Stratified by Fellowship Training

Abbreviations: HRR, hospital referral region; TSA, total shoulder arthroplasty.

Continue to: DISCUSSION...

DISCUSSION

Utilization of TSA has continued to rise; however, access to this cost-effective procedure was recently demonstrated to be limited.¹¹ In a separate analysis, we established the continued rise in use of TSA in the Medicare population, coupled with an increase in the number of surgeons routinely performing TSA.² Multiple analyses have demonstrated the importance of high-volume surgeons and hospitals familiar with the intricacies of shoulder arthroplasty concepts in minimizing complications, improving the quality and decreasing the cost of TSA.^6,10,17 Specifically, Singh and colleagues18 demonstrated from a multi-center registry that surgeons and hospitals with greater shoulder arthroplasty volumes had decreased intra-operative blood loss, operative time, and hospital length of stay. As the demand for TSA, both anatomic and reverse, continues to rise, it is imperative that the healthcare delivery system is optimized to provide the best possible care. Before we can determine whether specialized training in shoulder arthroplasty influences surgical outcomes, characteristics and training of surgeons performing TSA should be described. 

The number of surgeons performing >10 TSA in the Medicare population rose significantly between 2012 and 2014 (29.3%). However, the number of TSAs per surgeon over this time period remained consistent (approximately 25 per surgeon). Furthermore, the increase in the number of surgeons performing a reportable volume of TSA by 2014 was from the addition of already active surgeons (ie, the growth in TSA was not from the addition of newly trained arthroplasty surgeons but originated from the existing orthopedic surgeon workforce). In a recently published analysis, Somerson and colleagues, ¹¹ using this same dataset, demonstrated persistent limitations in access to high-volume TSA surgeons. In a more recent analysis, we showed that while still lacking for some patients, access to a high-volume TSA surgeon has improved significantly over the past 3 years, with 96.9% of the United States population residing within 200 kilometers of a high-volume TSA surgeon (>20 Medicare cases).² This analysis validates those findings, with the caveat that the average annual volume per surgeon is not increasing. What remains unknown, due to limitations of this dataset, is how many surgeons are not identified because they are performing ≤10 TSA each year or are performing TSA in non-Medicare patients. 

With the specialization of healthcare delivery, specifically in orthopedics, it is imperative that mechanisms for providing specialty-focused care be established. However, the proportion of their practice that surgeons dedicate to TSA was unknown. This study demonstrates that this proportion is increasing. Including non-arthroplasty procedures, more than half (58%) of the procedures performed by this surgeon cohort were shoulder-specific. Furthermore, this analysis demonstrates that surgeons performing TSA have significant case diversity, including nearly half of the cohort performing TKA. Repeated evidence has demonstrated the effect of case volume on improved outcomes following orthopedic procedures.^8,19–21 The pre-existing location-based model for delivering orthopedic care supports case diversity; however, this model continues to be challenged with high-volume centers of excellence and patient travel.^22–24 Hip and knee arthroplasty experienced a similar surge in demand, with a subsequent shift in care to high-volume surgeons and centers.²⁵ Shoulder and elbow fellowship-trained surgeons would need to nearly quadruple their current Medicare TSA volume to meet the entire current demand for TSA in the Medicare population (and this does not account for TSA performed by very low-volume surgeons not included in this cohort). With increased utilization of TSA, policymakers and the orthopedic community must determine the structure of delivery (centers of excellence or medium-volume disseminated throughout the country) that is optimal. 

For those surgeons consistently performing TSA over the study period, fellowship training was diverse. While the current focus in orthopedics is on case volume, research in other specialties, namely general surgery, has provided repeated evidence that surgical specialization (more so than high case volume) provides improved outcomes.^26–29Furthermore, Leopold and colleagues³⁰ demonstrated an inverse relationship between competency in performing a procedure and confidence in one’s ability to do so. In their study, educational intervention provided improved competency in the procedure. Less than one-third (29.8%) of TSA in this cohort were performed by a shoulder and elbow fellowship-trained surgeon consistently performing this procedure. Approximately another quarter (26.2%) were performed by consistent TSA surgeons trained in sports surgery. Meanwhile, 34.6% of TSA in this study cohort were performed by a surgeon who did not consistently meet the minimum threshold in all study years (16,435 TSA; 22.8%) or by a surgeon performing TSA without fellowship training (8,489 TSA; 11.8%). There has been a trend toward orthopedic subspecialty training with an increased demand for fellowship-trained surgeons.³¹ Despite this and the complexities of TSA, many continue to be performed by surgeons with an inconsistent volume and those without arthroplasty-specific fellowship training. The available evidence supports a push toward the fellowship-trained, high-volume TSA surgeon in providing reproducible high-quality shoulder arthroplasty care. For now, that surgeon is more likely to be earlier in his/her career and reside in large, referral-based centers surrounded by other surgeons performing TSA.

These findings must be considered in the light of the study limitations. First, this is a large publicly available database. While this type of database provides a unique opportunity to assess the geographic distributions and characteristics of orthopedic surgeons, specifically those performing TSA, it completely prevents any assessment of the relationship between these findings and quality. As such, while the reader may generate hypotheses regarding the implications of our findings on the quality of TSA delivery, the true effects cannot be determined. In the same vein, for the purpose of privacy, surgeons performing ≤10 TSA were not included in this dataset. This limitation prevents the identification of low-volume TSA surgeons. Also, it is likely that the observed increase in surgeons over time is likely a reflection of small increases in volume for surgeons already performing TSA. Lastly, a web-based search was undertaken to identify surgeons’ self-reported fellowship training. The results of this web-based search could not be validated, and it is possible that fellowship training, or the lack thereof, was mischaracterized and simply not obtainable through a web-based search. Furthermore, it is not possible to fully assess the extent of high-quality TSA training in these various fellowships.

CONCLUSION

In just the past decade, the utilization of TSA in the Medicare population has increased significantly. However, this increase was not achieved by the addition of highly specialized, high-volume surgeons but by the addition of many surgeons performing lower numbers of TSA surgeries. Furthermore, for those performing this cost-effective procedure, TSA constitutes a relatively small proportion of the surgeries they perform. Shoulder and elbow fellowship-trained surgeons currently account for a low percentage of the overall number of surgeons performing TSA. The implications of these findings must be considered and investigated.

Prior reviews have demonstrated a strong link between surgeon and hospital TSA volume and outcomes of the procedure.^8–10 Somerson and colleagues¹¹ investigated fellowship training among surgeons performing TSA in 2012 and found that only 28% had completed a shoulder and elbow fellowship. In addition to prior analyses2, ¹², Somerson and colleagues confirmed a persistent geographic variation in utilization of TSA.¹¹ In conjunction with the evolution of shoulder arthroplasty, dedicated shoulder and elbow fellowship training has expanded. With a shift toward specialization in care, nearly 90% of orthopedic surgery residents plan to pursue shoulder and elbow fellowships, comprising 4.6% of (42/897) of available positions.¹³  

What remains unknown is the specialization of surgeons performing TSA, the regularity of their arthroplasty volume, and trends in TSA specialization over time. Therefore, this study aims to (a) identify surgeons performing shoulder arthroplasty and cohort changes over time, (b) determine the case profile of surgeons consistently performing shoulder arthroplasty, and (c) establish the characteristics of shoulder arthroplasty surgeons with a specific focus on fellowship training. 

METHODS

Prior to collecting surgeon-specific data, we identified surgeons performing TSA through the Centers for Medicare and Medicaid Services’ public release of “Medicare Provider Utilization and Payment Data: Physician and Other Supplier.”¹⁴ Datasets from 2012, 2013, and 2014 were used to identify all surgeons performing >10 TSAs (Current Procedural Terminology [CPT] Code 23472) during at least 1 of those years. This dataset provides the name, identification number, address, and all billing (by volume) for each unique CPT code submitted ≥10 times in a calendar year.

Once the cohort of surgeons had been generated, the number of surgeons consistently performing TSA year-over-year was determined. This allowed for an analysis of the consistency with which surgeons are performing moderate- to high-volume TSA. To form a case profile of surgeons performing TSA and observe how this shifted over time, a count and a description of each CPT code submitted by each surgeon was identified. To maintain patient privacy, only those claims made >10 times are reported for a provider (both physicians and physician-extenders are included in this dataset). First, all CPT codes were reviewed and tagged as surgical or non-surgical events. Then, every procedural CPT code identified was reviewed and categorized based upon anatomic location and procedure (eg, total knee arthroplasty [TKA]). It is important to note that all claims in this dataset are limited to those patients participating in Medicare’s fee-for-service program. 

Specialization was defined as the number of categorized procedures as a percentage of all procedures performed on Medicare patients. The trends for national, regional, and individual specialization of TSA, arthroplasty (major joint), and shoulder procedures were determined.

Continue to: To investigate the characteristics of surgeons...

To investigate the characteristics of surgeons consistently performing TSA, all surgeons performing a minimum of 11 TSA in Medicare fee-for-service beneficiaries in all years between 2012 and 2014 were identified. Such surgeons were defined as consistent TSA surgeons. Investigation of this cohort included a web-based search of their self-reported post-graduate fellowship training and year of graduation from medical school. Using these data, the percentage of surgeons performing TSA who underwent formal shoulder and elbow training was determined. In addition, the impact of fellowship training on shoulder specialization and practice location was determined. Surgeons who had completed multiple fellowships were categorized under all of them. As such, there may be some duplication of surgeons in the comparisons. In addition, other potential characteristics of shoulder and elbow fellowship-trained surgeons were investigated: number of regional shoulder surgeons, urban area, total number of Medicare beneficiaries, average reimbursement for TSA, ethnicity of Medicare beneficiaries, and percentage of Medicare patients eligible for Medicaid. Geographic regions were defined by the Dartmouth Atlas and assigned by hospital referral region.¹⁵ These defined regions were used to assess the beneficiaries (number and characteristics) that individual surgeons were likely serving. The United States Census Bureau characterization of zip code-based regions as urban areas (population >50,000), urban clusters (2500 to 50,000), and rural region (<2500) was used to categorize practice location.¹⁶ 

Descriptive statistics were used initially to report these findings. To analyze predictors of utilization and specialization, comparative statistics were undertaken. For comparison of binomial variables between groups, a χ2analysis was utilized. For continuous variables, data normality was assessed. A skewness and kurtosis <2 and 12, respectively, was considered to represent parametric data. For parametric data, the mean was reported; conversely, the median is reported for non-parametric data. To assess continuous variables between groups, a t test or a Wilcoxon rank-sum test was used for parametric and non-parametric distributions, respectively. 

RESULTS

Between 2012 and 2014, 1374 surgeons (39 females [2.8%]) performed >10 TSA in Medicare patients in at least 1 year, for a combined total of 71,973 TSAs (Table 1). In 2012, only 834 surgeons (13 females [1.6%]) performed a minimum of 10 TSA in Medicare patients (21,137 arthroplasties; 25.3 per surgeon). This increased to 1078 surgeons (33 females [3.1%]; P = .04) performing 26,865 TSA (24.92 per surgeon) in 2014. Utilization of non-physician assistants in TSA also increased significantly over this period, with 307 assisting in 6885 TSAs (22.4 per provider) in 2012 and 465 assisting 10,433 TSA (22.4 per provider) in 2014. When all procedures were considered, including those performed at outpatient visits, 1319 physicians (95.9% of cohort) were active in 2012—providing either surgical procedures or outpatient consults to the Medicare population. Yet, only 63.2% performed >10 TSA in Medicare patients. The number of active surgeons performing TSA increased to 79.6% (1078/1353) in 2014 (P < .001). 

Table 1. Trends in the Number of Providers Performing TSA between 2012 and 2014*

* Included are the number of arthroplasties and total procedures over time among this cohort. The number of active providers, determined by billing Medicare for office or surgical procedures within that year, is reported.

Abbreviation: TSA, total shoulder arthroplasty.  

In addition to TSA, this cohort of surgeons submitted 240 unique CPT codes with case volumes >10 annually over the 3-year study period. Of these, 80.2% (1102/1374) of surgeons performed non-arthroplasty shoulder procedures on Medicare patients, for a combined total of 202,335 procedures over the 3-year study period (Table 2). A significant proportion of these procedures were arthroscopic debridement (60,014 procedures performed by 908 surgeons) and arthroscopic rotator cuff repair (47,089 procedures performed by 809 surgeons). Just under half (49.1%; 674/1374) of surgeons performing TSA also performed TKA during this period (77,873 arthroplasties). Fewer surgeons (27.8%; 382/1374) performed total hip arthroplasty during this period (27,322 arthroplasties). Other procedure types that this group of surgeons routinely performed on Medicare patients were repairs of traumatic injuries (29.8%), non-arthroplasty knee surgeries (27.2%), elbow procedures (19.6%), and hand surgery (15.4%). By case load, non-arthroplasty shoulder procedures consisted of 43% of Medicare volume over the study period (Figure 1). Between 2012 and 2014, the average proportion of Medicare cases that were shoulder arthroplasties increased from 13.8% (21,137/152,862) to 16.7% (26,865/160,851; P = .001). Shoulder arthroplasty constituted 100% of the Medicare surgical case volume for 67 (4.9%; 67/1374) of the surgeons.    

Table 2. Case Volumes over Time with All Procedures Categorized by Anatomic Region and Arthroplasty vs Non-arthroplasty*

* Procedures of interest with high case volumes are reported individually.

Only 44.3% (609/1374) of surgeons performed TSA in a minimum of 11 Medicare patients in all 3 years of the study period (consistent providers of TSA), providing a total of 55,538 TSA (77.2%; 55,538/71,973). When fellowship training was evaluated, 191 (31.4%; 191/609) of these surgeons were shoulder and elbow fellowship trained (21,444 TSA; 38.6%; Table 3). More than one-third (36.6%; 223/609) had completed a sports surgery fellowship (18,899 TSA; 34.0%). Surgeons trained in hand surgery (12.5%; 76/609) and adult reconstruction (5.3%; 32/610) also made contributions to meeting the TSA demand with 6971 (12.6%) and 2485 (4.5%) TSA, respectively. One-fifth of this cohort (18.1%; 110/609) had unknown fellowship training: they either reported no fellowship (13.6%; 83/609) or did not specify the type of training (4.4%; 27/609). Shoulder and elbow fellowship-trained surgeons performed more TSA (median: 89.0 TSA per surgeon between 2012 and 2014) than surgeons without shoulder and elbow fellowship training (median: 67.0 TSA per surgeon; P < 0.001). More than one-third (37%) of shoulder and elbow fellowship-trained surgeons’ surgical case volume was comprised of TSA, with an additional 35% from non-arthroplasty shoulder procedures (Figure 2). In order for the current supply of shoulder and elbow fellowship-trained surgeons to meet the Medicare TSA demand, each fellowship graduate would have to perform 140.6 TSA in Medicare patients annually. Shoulder and elbow fellowship-trained surgeons were more likely to practice in referral regions with an increased Medicare population (P < .001), an increased number of surgeons performing TSA (P < .001), and a higher proportion of Medicaid-eligible patients (P = .01; Table 4). Shoulder and elbow fellowship-trained surgeons (18.7 years post-medical school graduation) were also earlier in their careers than other consistent TSA surgeons (23.1 years post-graduation; P < .001). 

Table 3. A Representation of Fellowships Among TSA Surgeons and Their Shoulder Arthroplasty Case Load*

* Not all fellowships (eg, oncology) included due to small numbers. Also, many surgeons performed multiple fellowships.

Abbreviations: SA, shoulder arthroplasty; TSA, total shoulder arthroplasty.

Table 4. Breakdown of Geographic Characteristics of Orthopedic Surgeons Consistently

Performing TSA between 2012 and 2014 Stratified by Fellowship Training

Abbreviations: HRR, hospital referral region; TSA, total shoulder arthroplasty.

Continue to: DISCUSSION...

DISCUSSION

Utilization of TSA has continued to rise; however, access to this cost-effective procedure was recently demonstrated to be limited.¹¹ In a separate analysis, we established the continued rise in use of TSA in the Medicare population, coupled with an increase in the number of surgeons routinely performing TSA.² Multiple analyses have demonstrated the importance of high-volume surgeons and hospitals familiar with the intricacies of shoulder arthroplasty concepts in minimizing complications, improving the quality and decreasing the cost of TSA.^6,10,17 Specifically, Singh and colleagues18 demonstrated from a multi-center registry that surgeons and hospitals with greater shoulder arthroplasty volumes had decreased intra-operative blood loss, operative time, and hospital length of stay. As the demand for TSA, both anatomic and reverse, continues to rise, it is imperative that the healthcare delivery system is optimized to provide the best possible care. Before we can determine whether specialized training in shoulder arthroplasty influences surgical outcomes, characteristics and training of surgeons performing TSA should be described. 

The number of surgeons performing >10 TSA in the Medicare population rose significantly between 2012 and 2014 (29.3%). However, the number of TSAs per surgeon over this time period remained consistent (approximately 25 per surgeon). Furthermore, the increase in the number of surgeons performing a reportable volume of TSA by 2014 was from the addition of already active surgeons (ie, the growth in TSA was not from the addition of newly trained arthroplasty surgeons but originated from the existing orthopedic surgeon workforce). In a recently published analysis, Somerson and colleagues, ¹¹ using this same dataset, demonstrated persistent limitations in access to high-volume TSA surgeons. In a more recent analysis, we showed that while still lacking for some patients, access to a high-volume TSA surgeon has improved significantly over the past 3 years, with 96.9% of the United States population residing within 200 kilometers of a high-volume TSA surgeon (>20 Medicare cases).² This analysis validates those findings, with the caveat that the average annual volume per surgeon is not increasing. What remains unknown, due to limitations of this dataset, is how many surgeons are not identified because they are performing ≤10 TSA each year or are performing TSA in non-Medicare patients. 

With the specialization of healthcare delivery, specifically in orthopedics, it is imperative that mechanisms for providing specialty-focused care be established. However, the proportion of their practice that surgeons dedicate to TSA was unknown. This study demonstrates that this proportion is increasing. Including non-arthroplasty procedures, more than half (58%) of the procedures performed by this surgeon cohort were shoulder-specific. Furthermore, this analysis demonstrates that surgeons performing TSA have significant case diversity, including nearly half of the cohort performing TKA. Repeated evidence has demonstrated the effect of case volume on improved outcomes following orthopedic procedures.^8,19–21 The pre-existing location-based model for delivering orthopedic care supports case diversity; however, this model continues to be challenged with high-volume centers of excellence and patient travel.^22–24 Hip and knee arthroplasty experienced a similar surge in demand, with a subsequent shift in care to high-volume surgeons and centers.²⁵ Shoulder and elbow fellowship-trained surgeons would need to nearly quadruple their current Medicare TSA volume to meet the entire current demand for TSA in the Medicare population (and this does not account for TSA performed by very low-volume surgeons not included in this cohort). With increased utilization of TSA, policymakers and the orthopedic community must determine the structure of delivery (centers of excellence or medium-volume disseminated throughout the country) that is optimal. 

For those surgeons consistently performing TSA over the study period, fellowship training was diverse. While the current focus in orthopedics is on case volume, research in other specialties, namely general surgery, has provided repeated evidence that surgical specialization (more so than high case volume) provides improved outcomes.^26–29Furthermore, Leopold and colleagues³⁰ demonstrated an inverse relationship between competency in performing a procedure and confidence in one’s ability to do so. In their study, educational intervention provided improved competency in the procedure. Less than one-third (29.8%) of TSA in this cohort were performed by a shoulder and elbow fellowship-trained surgeon consistently performing this procedure. Approximately another quarter (26.2%) were performed by consistent TSA surgeons trained in sports surgery. Meanwhile, 34.6% of TSA in this study cohort were performed by a surgeon who did not consistently meet the minimum threshold in all study years (16,435 TSA; 22.8%) or by a surgeon performing TSA without fellowship training (8,489 TSA; 11.8%). There has been a trend toward orthopedic subspecialty training with an increased demand for fellowship-trained surgeons.³¹ Despite this and the complexities of TSA, many continue to be performed by surgeons with an inconsistent volume and those without arthroplasty-specific fellowship training. The available evidence supports a push toward the fellowship-trained, high-volume TSA surgeon in providing reproducible high-quality shoulder arthroplasty care. For now, that surgeon is more likely to be earlier in his/her career and reside in large, referral-based centers surrounded by other surgeons performing TSA.

These findings must be considered in the light of the study limitations. First, this is a large publicly available database. While this type of database provides a unique opportunity to assess the geographic distributions and characteristics of orthopedic surgeons, specifically those performing TSA, it completely prevents any assessment of the relationship between these findings and quality. As such, while the reader may generate hypotheses regarding the implications of our findings on the quality of TSA delivery, the true effects cannot be determined. In the same vein, for the purpose of privacy, surgeons performing ≤10 TSA were not included in this dataset. This limitation prevents the identification of low-volume TSA surgeons. Also, it is likely that the observed increase in surgeons over time is likely a reflection of small increases in volume for surgeons already performing TSA. Lastly, a web-based search was undertaken to identify surgeons’ self-reported fellowship training. The results of this web-based search could not be validated, and it is possible that fellowship training, or the lack thereof, was mischaracterized and simply not obtainable through a web-based search. Furthermore, it is not possible to fully assess the extent of high-quality TSA training in these various fellowships.

CONCLUSION

In just the past decade, the utilization of TSA in the Medicare population has increased significantly. However, this increase was not achieved by the addition of highly specialized, high-volume surgeons but by the addition of many surgeons performing lower numbers of TSA surgeries. Furthermore, for those performing this cost-effective procedure, TSA constitutes a relatively small proportion of the surgeries they perform. Shoulder and elbow fellowship-trained surgeons currently account for a low percentage of the overall number of surgeons performing TSA. The implications of these findings must be considered and investigated.

The dilemma for the treating surgeon is deciding which patients to treat with RCR and who to treat with rTSA. Predicting which surgical procedure will provide a better functional result is difficult and controversial.⁷ The RCR method is a bone-conserving procedure with relatively low surgical risk and allows the option for rTSA to be performed as a salvage surgery should repair fail. It also may be less costly in the appropriate population.⁸ However, large rotator cuff tears in elderly patients have low healing potential, and the prospect of participating in a lengthy rehabilitation after an operation that may not prove successful can be deterring.^9,10 In the elderly population, rTSA may be a reliable option, as tendon healing of the cuff is not necessary to restore function. However, rTSA does not conserve bone, provides a non-anatomic solution, and has had historically high complication rates.^4,5

In an effort to aid in the decision-making process when considering these 2 surgical options, we compared RCR and rTSA performed at a single institution for MCT in patients >65 years. Our aim was to identify preoperative patient variables that influence a surgeon’s decision to proceed with 1 of the 2 procedures. Moreover, we evaluated clinical outcomes in these 2 patient populations. We hypothesized that (1) patients selected for rTSA would have worse preoperative function, less range of motion, more comorbidities, more evidence of radiographic subluxation, and a higher likelihood of having undergone previous RCR than those selected for RCR, and (2) both RCR and rTSA would be successful and result in improved clinical outcomes with high patient satisfaction.

MATERIALS AND METHODS

PATIENT SELECTION

We performed a retrospective chart review using our practice database of all patients undergoing arthroscopic RCR and rTSA for any indication by the senior author (M.A.F.) between January 2004 and April 2015. A total of 1503 RCRs and 1973 rTSAs were conducted during the study period. Patient medical records were reviewed, and those meeting the following criteria were included in the study: >65 years at the time of surgery, MCT, no preoperative glenohumeral arthritis, minimum follow-up of 12 months, functional deltoid muscle on physical examination, and no prior shoulder surgery except for RCR or diagnostic arthroscopy. A total of 92 patients who underwent arthroscopic RCR and 89 patients who underwent rTSA met the inclusion criteria. For patients with bilateral shoulder surgery, we measured each shoulder independently. Three patients underwent bilateral rTSA, and 3 patients underwent bilateral RCR, leaving 95 shoulders in the RCR group and 92 in the rTSA group. The Western Institutional Review Board determined this study to be exempt from review.

RADIOGRAPHIC EVALUATION

All patient charts included a radiology report and documented interpretation of the images by the treating surgeon prior to surgery. Radiographs were assessed to assure the absence of preoperative glenohumeral osteoarthritis. The images were also graded based on the Hamada classification.¹¹ Stage 1 is associated with minimal radiographic change with an acromiohumeral interval (AHI) >6 mm; stage 2 is characterized by narrowing of the AHI <6 mm; and Stage 3 is defined by narrowing of the AHI with radiographic changes of the acromion. Stages 4 and higher include arthritic changes to the glenohumeral joint, and they were not included in the study population. The AHI measurements and the presence or absence of glenohumeral subluxation were documented.

Continue to: MASSIVE CUFF TEAR DETERMINATION...

We defined MCT on the basis of previously described criteria of tears involving ≥2 tendons or tears measuring ≥5 cm in greatest dimension.^12,13 Patient charts were screened, and those whose clinical notes or radiology reports indicated an absence of MCT were excluded. Preoperative imaging of the remaining patients was then evaluated by 3 fellowship-trained shoulder surgeons to confirm MCT in all patients with a clinically documented MCT, as well as to assess those who had insufficient documentation of tear size in the notes.

Advanced imaging was evaluated for fatty atrophy of the rotator cuff musculature, and Goutallier classification was assigned.^14,15 Length of retraction was measured from the tendon end to the medial aspect of the footprint on coronal imaging, and the subscapularis and teres minor were assessed and documented as torn or intact.^16,17

DATA COLLECTION

We reviewed clinical charts and patient questionnaire forms from both the preoperative and follow-up visits. Clinical data collected included gender, age at surgery, active range of motion (forward elevation, abduction, external rotation, and internal rotation), comorbidities, smoking status, BMI, history of shoulder surgery, and any postoperative complications or need for secondary surgery. All patients completed patient-centered questionnaires regarding shoulder pain and dysfunction at each visit or via telephone communication with clinic staff. Outcome measurements used for analysis included American Shoulder and Elbow Surgeons (ASES) Score, simple shoulder test (SST), visual analog score (VAS) pain scale, and patient-reported satisfaction (Graded 1-10; 1 = poor outcome; 4 = satisfactory outcome; 7 = good outcome; 10 = excellent outcome).

DATA ANALYSIS AND STATISTICAL METHODS

Statistical tests were selected based on the result of Shapiro–Wilk test for normality. Continuous variables were evaluated with either independent t test or Mann–Whitney U test. Dependent t test was used to evaluate outcome variables. For categorical variables, either Pearson’s χ²or Fisher’s exact test was performed depending on the sample size. Alpha was set at P =.05.

Continue to: RESULTS...

PREOPERATIVE CHARACTERISTICS

Of the 187 shoulders in the study group, 95 had RCR and 92 had rTSA. Demographic information and preoperative variables for both groups are summarized in Table 1 and Table 2. Patients in the RCR group had greater preoperative forward elevation, abduction, and external rotation and higher preoperative functional scores than those in the rTSA group. Patients in the rTSA group were older and more likely to be female than those in the RCR group. More patients in the rTSA group had undergone prior RCR compared with those in the RCR group. Each of these differences was statistically significant. Subjective pain scores, BMI, and comorbidities were similar between the 2 groups.

Table 1. Patient demographics

Abbreviations: BMI, body mass index; RCR, rotator cuff repair; rTSA, reverse total shoulder arthroplasty.

Table 2. Preoperative variables

Abbreviations: AB, abduction; ASES, American Shoulder and Elbow Society score; ER, external rotation; FE, forward elevation; IR, internal rotation; preop, preoperative; SST, simple shoulder test.

Radiographically, patients selected to undergo rTSA had a smaller AHI (4.8 vs 8.7, P < .0001) and more evidence of superior subluxation (50.6% vs 14.1%, P < .0001) than those in the RCR group. Average Hamada grade was 1.4 ± 0.7 and 2.2 ± 0.7 for the RCR and rTSA groups, respectively (P < .0001). Average Goutallier grade was similar between the groups (2.2 ± 0.6 for RCR vs 2.2 ± 0.8 for rTSA, P =.227), and 25.5% of the RCR group had Grade 3 or 4 atrophy compared with 28.1% of the rTSA group.

POSTOPERATIVE OUTCOMES

The average follow-up time was 44 months for RCR and 47 months for rTSA. Patients in the RCR and rTSA groups were highly satisfied with the surgery (8.5 ± 2.6 vs 8.2 ± 2.6, P = .461) and had significantly increased range of motion in all planes and improved functional scores (Table 3). The rTSA group had greater net improvement in forward elevation, abduction, and external rotation than the RCR group. Both groups demonstrated similar improvement in ASES, SST, and VAS pain scores.

Table 3. Postoperative outcomes

Abbreviations: AB, abduction; ASES, American Shoulder and Elbow Society score; ER, external rotation; FE, forward elevation; IR, internal rotation; SST, simple shoulder test; VAS – visual analog score.

In the RCR group, 5 patients (5.3%) required reoperation: 3 patients underwent conversion to rTSA, 1 patient underwent biceps tenotomy with subacromial decompression, and 1 patient underwent arthroscopic irrigation and debridement for a postoperative Propionibacterium acnes infection. In the rTSA group, 2 patients (2.2%) required reoperation: 1 patient underwent open reduction internal fixation for a scapula fracture that failed conservative management, and 1 patient had an open irrigation and debridement with polyethylene exchange for an acute postoperative infection of unknown source.

DISCUSSION

Massive, retracted rotator cuff tears are a common and difficult problem.¹ The treatment options are numerous and depend on a variety of preoperative factors including patient-specific characteristics and factors specific to the tear. For certain patients, nonoperative management may be a reasonable first step, as an MCT does not necessarily preclude painless, functional shoulder motion. Elderly, lower demand individuals have been shown to do well with physical rehabilitation.¹⁸ Similarly, for the same category of elderly patients who do not respond to conservative measures, arthroscopic tendon debridement with or without subacromial decompression and/or biceps tenotomy may be effective.^1,19 This technique has been described as “limited goals surgery;” despite some mixed results in the literature, multiple studies have reported symptomatic and functional improvement after simple debridement.^2,19–21The consensus among several authors has been that this procedure continues to play a role for elderly, low-demand patients whose functional goals are limited and whose primary complaint is pain.^1,2,20

For the majority of patients with MCT who desire pain relief and a restoration of shoulder function, RCR remains the gold standard of treatment and should be the primary aim if feasible. Complete RCR has consistently outperformed both partial repair and debridement in multiple studies in terms of pain relief and functional improvement.^10,21,22However, elderly patients with chronic, massive tears, particularly in the setting of muscle atrophy, are at high risk of failure with attempted cuff repair.^9,23 Novel techniques such as superior capsular reconstruction and subacromial balloon spacer implantation may offer a minimally invasive method of re-centering the humeral head and stabilizing the glenohumeral joint; however, these new treatment options lack any long-term data in the literature to support their widespread use.^24–26 Alternatively, rTSA has been shown to be a reliable option to restore shoulder function in the setting of a massive irreparable rotator cuff tear, even in the absence of arthritis.^5,27-31

Continue to: The decision-making process...

The decision-making process for selecting RCR or rTSA in the setting of MCT without arthritis in the older population (age >65 years) remains challenging. We attempted to quantify the data of a high-volume surgeon and identify the differences and similarities between those patients selected for either procedure. At our institution, we generally performed rTSA on patients with low preoperative range of motion, poor function based on SST and ASES scores, small AHI, and strong evidence of superior subluxation. We were also more likely to perform rTSA if the patient had a history of rotator cuff surgery. There was a predilection for older age and female gender in those who underwent rTSA.

For our study, we elected to focus on patients >65 years. In our experience, the choice of which surgical procedure to perform is generally easier in younger patients. Most surgeons appropriately opt for an arthroscopic procedure or tendon transfer to preserve bone and maintain the option of rTSA as a salvage procedure if necessary in the future. Studies have reported that <60 years is a predictor of poor outcome with rTSA, and patients <65 years who undergo rTSA have been shown to have high complication rates.^30-32 Furthermore, the longevity of the implant in young patients is a significant concern, and revision surgery after rTSA is technically demanding and known to result in poor functional outcomes.^32,33

Although the indications for rTSA are expanding, attempts at RCR in the setting of MCT remain largely appropriate. Preserved preoperative anterior elevation >90° has been associated with loss of motion after rTSA and poor satisfaction, and one should exercise caution when considering rTSA in this setting.³ The current study confirmed that even older patients with MCT may be very satisfied with arthroscopic RCR (Figure 1). Both range of motion and function significantly improved, and patients were largely satisfied with the procedure with an average self-reported outcome of good to excellent. At the time of final follow-up for this study, only 3 shoulders in the RCR group had undergone conversion to rTSA. This number may be expected to rise with long follow-up periods, and we feel that prolonging the time before arthroplasty is generally in the best interest of the patient.

Our results were consistent with several reported studies in which RCR has been shown to be successful in the setting of MCT.^34–37 Henry and colleagues³⁶ performed a systematic review that evaluated 954 patients who underwent partial or complete anatomic RCR for MCT. Although the average age was 63 years (range, 37–87), functional outcome scores, VAS pain score, and overall range of motion consistently and significantly improved.

rTSA may be a “more reliable” option than RCR in treating MCT in the older population because it does not rely on tendon healing. However, the relationship between tendon healing and clinical outcomes after RCR is unclear. The aforementioned systematic review reported re-tear rates to be as high as 79%, but several studies have reported high satisfaction even in the setting of retear.³⁶ Yoo and colleagues³⁸ and Chung and colleagues⁹ reported re-tear rates of 45.5% and 39.8%, respectively, but both studies noted that there was no difference in outcome measures between those patients with and without re-tears. In particular, for patients who have had no prior rotator cuff surgery, an attempt at arthroscopic repair may be a prudent option with relatively low risk.

Although certain patients may clinically improve despite suffering a re-tear (or inability to heal in the first place), others continue to experience pain and dysfunction that negatively affect their quality of life.^39–41 These patients are more often appropriate candidates for rTSA. Indeed, several studies have demonstrated a higher re-tear rate in patients with a history of surgery than in those without.^23,31,38,42 Shamsudin and colleagues⁴³ found revision arthroscopic RCR, even in a younger age group with tears of all sizes, to be twice as likely to re-tear. Notably, re-tear after revision repair may be more likely to be symptomatic, as these re-tears are routinely associated with pain, stiffness, and loss of function. Even in the hands of experienced surgeons in a younger population, revision repair has only been able to reverse pseudoparalysis in 43% of patients, leading to only 39% return to sport or full activity.⁴⁴ In examining our data, we were much less likely to perform an RCR in patients who had a history of cuff repair surgery than in those without this history.

Continue to: Overall, those patients selected for rTSA...

Overall, those patients selected for rTSA in our study population performed well postoperatively (Figure 2 and Figure 3). Vast improvements were noted in range of motion, function, and pain scores at final follow up. Moreover, no patients in the study group required revision arthroplasty during the follow-up period. Although the average follow-up period was only 47 months, these results suggested that elderly patients with MCT without arthritis may be particularly ideal candidates for rTSA with regard to implant survival and anticipated revision rate when chosen appropriately.

Several weaknesses were noted within this paper. First, the study was retrospective, precluding randomization of treatment groups and standardization of data collection and follow-up. The outcomes of RCR and rTSA could not be compared directly due to the inherent selection bias. The groups clearly differed in many respects, and these preoperative factors likely played a role in postoperative outcomes. However, the primary goal of this study was not to compare outcomes of the treatment groups but to analyze the patterns of patient selection by an experienced treating surgeon and contribute to published data that each surgery can be successful in this patient population when chosen appropriately.

Second, our data were based on a single surgeon’s decisions, and results may not be generalizable. Furthermore, the senior author has had a longstanding interest in reverse shoulder arthroplasty and has published data illustrating successful outcomes for rTSA in patients with MCT. For this reason, one could presume that there may have been some bias toward treating patients with rTSA. However, we feel that the senior author’s unique and longstanding experience in treating MCT allows for a thorough evaluation and comparison of preoperative variables and outcomes declared within this study. Indeed, many patients included in this study were referred from outside institutions specifically for rTSA but instead were deemed more appropriate candidates for RCR and underwent successful arthroscopic repair, a common scenario which served as an impetus for this study.

CONCLUSION

RCR and rTSA are both viable options for patients >65 years with MCT without arthritis. Treatment must be individualized for each patient with careful consideration of a number of preoperative variables and patient characteristics. At our institution, patients with previous RCR, decreased range of motion, poor function, and strong radiographic evidence of subluxation are more likely to undergo rTSA. When chosen appropriately, both RCR and rTSA can result in improved range of motion, function, and high patient satisfaction in this patient population.

The dilemma for the treating surgeon is deciding which patients to treat with RCR and who to treat with rTSA. Predicting which surgical procedure will provide a better functional result is difficult and controversial.⁷ The RCR method is a bone-conserving procedure with relatively low surgical risk and allows the option for rTSA to be performed as a salvage surgery should repair fail. It also may be less costly in the appropriate population.⁸ However, large rotator cuff tears in elderly patients have low healing potential, and the prospect of participating in a lengthy rehabilitation after an operation that may not prove successful can be deterring.^9,10 In the elderly population, rTSA may be a reliable option, as tendon healing of the cuff is not necessary to restore function. However, rTSA does not conserve bone, provides a non-anatomic solution, and has had historically high complication rates.^4,5

In an effort to aid in the decision-making process when considering these 2 surgical options, we compared RCR and rTSA performed at a single institution for MCT in patients >65 years. Our aim was to identify preoperative patient variables that influence a surgeon’s decision to proceed with 1 of the 2 procedures. Moreover, we evaluated clinical outcomes in these 2 patient populations. We hypothesized that (1) patients selected for rTSA would have worse preoperative function, less range of motion, more comorbidities, more evidence of radiographic subluxation, and a higher likelihood of having undergone previous RCR than those selected for RCR, and (2) both RCR and rTSA would be successful and result in improved clinical outcomes with high patient satisfaction.

MATERIALS AND METHODS

PATIENT SELECTION

We performed a retrospective chart review using our practice database of all patients undergoing arthroscopic RCR and rTSA for any indication by the senior author (M.A.F.) between January 2004 and April 2015. A total of 1503 RCRs and 1973 rTSAs were conducted during the study period. Patient medical records were reviewed, and those meeting the following criteria were included in the study: >65 years at the time of surgery, MCT, no preoperative glenohumeral arthritis, minimum follow-up of 12 months, functional deltoid muscle on physical examination, and no prior shoulder surgery except for RCR or diagnostic arthroscopy. A total of 92 patients who underwent arthroscopic RCR and 89 patients who underwent rTSA met the inclusion criteria. For patients with bilateral shoulder surgery, we measured each shoulder independently. Three patients underwent bilateral rTSA, and 3 patients underwent bilateral RCR, leaving 95 shoulders in the RCR group and 92 in the rTSA group. The Western Institutional Review Board determined this study to be exempt from review.

RADIOGRAPHIC EVALUATION

All patient charts included a radiology report and documented interpretation of the images by the treating surgeon prior to surgery. Radiographs were assessed to assure the absence of preoperative glenohumeral osteoarthritis. The images were also graded based on the Hamada classification.¹¹ Stage 1 is associated with minimal radiographic change with an acromiohumeral interval (AHI) >6 mm; stage 2 is characterized by narrowing of the AHI <6 mm; and Stage 3 is defined by narrowing of the AHI with radiographic changes of the acromion. Stages 4 and higher include arthritic changes to the glenohumeral joint, and they were not included in the study population. The AHI measurements and the presence or absence of glenohumeral subluxation were documented.

Continue to: MASSIVE CUFF TEAR DETERMINATION...

We defined MCT on the basis of previously described criteria of tears involving ≥2 tendons or tears measuring ≥5 cm in greatest dimension.^12,13 Patient charts were screened, and those whose clinical notes or radiology reports indicated an absence of MCT were excluded. Preoperative imaging of the remaining patients was then evaluated by 3 fellowship-trained shoulder surgeons to confirm MCT in all patients with a clinically documented MCT, as well as to assess those who had insufficient documentation of tear size in the notes.

Advanced imaging was evaluated for fatty atrophy of the rotator cuff musculature, and Goutallier classification was assigned.^14,15 Length of retraction was measured from the tendon end to the medial aspect of the footprint on coronal imaging, and the subscapularis and teres minor were assessed and documented as torn or intact.^16,17

DATA COLLECTION

We reviewed clinical charts and patient questionnaire forms from both the preoperative and follow-up visits. Clinical data collected included gender, age at surgery, active range of motion (forward elevation, abduction, external rotation, and internal rotation), comorbidities, smoking status, BMI, history of shoulder surgery, and any postoperative complications or need for secondary surgery. All patients completed patient-centered questionnaires regarding shoulder pain and dysfunction at each visit or via telephone communication with clinic staff. Outcome measurements used for analysis included American Shoulder and Elbow Surgeons (ASES) Score, simple shoulder test (SST), visual analog score (VAS) pain scale, and patient-reported satisfaction (Graded 1-10; 1 = poor outcome; 4 = satisfactory outcome; 7 = good outcome; 10 = excellent outcome).

DATA ANALYSIS AND STATISTICAL METHODS

Statistical tests were selected based on the result of Shapiro–Wilk test for normality. Continuous variables were evaluated with either independent t test or Mann–Whitney U test. Dependent t test was used to evaluate outcome variables. For categorical variables, either Pearson’s χ²or Fisher’s exact test was performed depending on the sample size. Alpha was set at P =.05.

Continue to: RESULTS...

PREOPERATIVE CHARACTERISTICS

Of the 187 shoulders in the study group, 95 had RCR and 92 had rTSA. Demographic information and preoperative variables for both groups are summarized in Table 1 and Table 2. Patients in the RCR group had greater preoperative forward elevation, abduction, and external rotation and higher preoperative functional scores than those in the rTSA group. Patients in the rTSA group were older and more likely to be female than those in the RCR group. More patients in the rTSA group had undergone prior RCR compared with those in the RCR group. Each of these differences was statistically significant. Subjective pain scores, BMI, and comorbidities were similar between the 2 groups.

Table 1. Patient demographics

Abbreviations: BMI, body mass index; RCR, rotator cuff repair; rTSA, reverse total shoulder arthroplasty.

Table 2. Preoperative variables

Abbreviations: AB, abduction; ASES, American Shoulder and Elbow Society score; ER, external rotation; FE, forward elevation; IR, internal rotation; preop, preoperative; SST, simple shoulder test.

Radiographically, patients selected to undergo rTSA had a smaller AHI (4.8 vs 8.7, P < .0001) and more evidence of superior subluxation (50.6% vs 14.1%, P < .0001) than those in the RCR group. Average Hamada grade was 1.4 ± 0.7 and 2.2 ± 0.7 for the RCR and rTSA groups, respectively (P < .0001). Average Goutallier grade was similar between the groups (2.2 ± 0.6 for RCR vs 2.2 ± 0.8 for rTSA, P =.227), and 25.5% of the RCR group had Grade 3 or 4 atrophy compared with 28.1% of the rTSA group.

POSTOPERATIVE OUTCOMES

The average follow-up time was 44 months for RCR and 47 months for rTSA. Patients in the RCR and rTSA groups were highly satisfied with the surgery (8.5 ± 2.6 vs 8.2 ± 2.6, P = .461) and had significantly increased range of motion in all planes and improved functional scores (Table 3). The rTSA group had greater net improvement in forward elevation, abduction, and external rotation than the RCR group. Both groups demonstrated similar improvement in ASES, SST, and VAS pain scores.

Table 3. Postoperative outcomes

Abbreviations: AB, abduction; ASES, American Shoulder and Elbow Society score; ER, external rotation; FE, forward elevation; IR, internal rotation; SST, simple shoulder test; VAS – visual analog score.

In the RCR group, 5 patients (5.3%) required reoperation: 3 patients underwent conversion to rTSA, 1 patient underwent biceps tenotomy with subacromial decompression, and 1 patient underwent arthroscopic irrigation and debridement for a postoperative Propionibacterium acnes infection. In the rTSA group, 2 patients (2.2%) required reoperation: 1 patient underwent open reduction internal fixation for a scapula fracture that failed conservative management, and 1 patient had an open irrigation and debridement with polyethylene exchange for an acute postoperative infection of unknown source.

DISCUSSION

Massive, retracted rotator cuff tears are a common and difficult problem.¹ The treatment options are numerous and depend on a variety of preoperative factors including patient-specific characteristics and factors specific to the tear. For certain patients, nonoperative management may be a reasonable first step, as an MCT does not necessarily preclude painless, functional shoulder motion. Elderly, lower demand individuals have been shown to do well with physical rehabilitation.¹⁸ Similarly, for the same category of elderly patients who do not respond to conservative measures, arthroscopic tendon debridement with or without subacromial decompression and/or biceps tenotomy may be effective.^1,19 This technique has been described as “limited goals surgery;” despite some mixed results in the literature, multiple studies have reported symptomatic and functional improvement after simple debridement.^2,19–21The consensus among several authors has been that this procedure continues to play a role for elderly, low-demand patients whose functional goals are limited and whose primary complaint is pain.^1,2,20

For the majority of patients with MCT who desire pain relief and a restoration of shoulder function, RCR remains the gold standard of treatment and should be the primary aim if feasible. Complete RCR has consistently outperformed both partial repair and debridement in multiple studies in terms of pain relief and functional improvement.^10,21,22However, elderly patients with chronic, massive tears, particularly in the setting of muscle atrophy, are at high risk of failure with attempted cuff repair.^9,23 Novel techniques such as superior capsular reconstruction and subacromial balloon spacer implantation may offer a minimally invasive method of re-centering the humeral head and stabilizing the glenohumeral joint; however, these new treatment options lack any long-term data in the literature to support their widespread use.^24–26 Alternatively, rTSA has been shown to be a reliable option to restore shoulder function in the setting of a massive irreparable rotator cuff tear, even in the absence of arthritis.^5,27-31

Continue to: The decision-making process...

The decision-making process for selecting RCR or rTSA in the setting of MCT without arthritis in the older population (age >65 years) remains challenging. We attempted to quantify the data of a high-volume surgeon and identify the differences and similarities between those patients selected for either procedure. At our institution, we generally performed rTSA on patients with low preoperative range of motion, poor function based on SST and ASES scores, small AHI, and strong evidence of superior subluxation. We were also more likely to perform rTSA if the patient had a history of rotator cuff surgery. There was a predilection for older age and female gender in those who underwent rTSA.

For our study, we elected to focus on patients >65 years. In our experience, the choice of which surgical procedure to perform is generally easier in younger patients. Most surgeons appropriately opt for an arthroscopic procedure or tendon transfer to preserve bone and maintain the option of rTSA as a salvage procedure if necessary in the future. Studies have reported that <60 years is a predictor of poor outcome with rTSA, and patients <65 years who undergo rTSA have been shown to have high complication rates.^30-32 Furthermore, the longevity of the implant in young patients is a significant concern, and revision surgery after rTSA is technically demanding and known to result in poor functional outcomes.^32,33

Although the indications for rTSA are expanding, attempts at RCR in the setting of MCT remain largely appropriate. Preserved preoperative anterior elevation >90° has been associated with loss of motion after rTSA and poor satisfaction, and one should exercise caution when considering rTSA in this setting.³ The current study confirmed that even older patients with MCT may be very satisfied with arthroscopic RCR (Figure 1). Both range of motion and function significantly improved, and patients were largely satisfied with the procedure with an average self-reported outcome of good to excellent. At the time of final follow-up for this study, only 3 shoulders in the RCR group had undergone conversion to rTSA. This number may be expected to rise with long follow-up periods, and we feel that prolonging the time before arthroplasty is generally in the best interest of the patient.

Our results were consistent with several reported studies in which RCR has been shown to be successful in the setting of MCT.^34–37 Henry and colleagues³⁶ performed a systematic review that evaluated 954 patients who underwent partial or complete anatomic RCR for MCT. Although the average age was 63 years (range, 37–87), functional outcome scores, VAS pain score, and overall range of motion consistently and significantly improved.

rTSA may be a “more reliable” option than RCR in treating MCT in the older population because it does not rely on tendon healing. However, the relationship between tendon healing and clinical outcomes after RCR is unclear. The aforementioned systematic review reported re-tear rates to be as high as 79%, but several studies have reported high satisfaction even in the setting of retear.³⁶ Yoo and colleagues³⁸ and Chung and colleagues⁹ reported re-tear rates of 45.5% and 39.8%, respectively, but both studies noted that there was no difference in outcome measures between those patients with and without re-tears. In particular, for patients who have had no prior rotator cuff surgery, an attempt at arthroscopic repair may be a prudent option with relatively low risk.

Although certain patients may clinically improve despite suffering a re-tear (or inability to heal in the first place), others continue to experience pain and dysfunction that negatively affect their quality of life.^39–41 These patients are more often appropriate candidates for rTSA. Indeed, several studies have demonstrated a higher re-tear rate in patients with a history of surgery than in those without.^23,31,38,42 Shamsudin and colleagues⁴³ found revision arthroscopic RCR, even in a younger age group with tears of all sizes, to be twice as likely to re-tear. Notably, re-tear after revision repair may be more likely to be symptomatic, as these re-tears are routinely associated with pain, stiffness, and loss of function. Even in the hands of experienced surgeons in a younger population, revision repair has only been able to reverse pseudoparalysis in 43% of patients, leading to only 39% return to sport or full activity.⁴⁴ In examining our data, we were much less likely to perform an RCR in patients who had a history of cuff repair surgery than in those without this history.

Continue to: Overall, those patients selected for rTSA...

Overall, those patients selected for rTSA in our study population performed well postoperatively (Figure 2 and Figure 3). Vast improvements were noted in range of motion, function, and pain scores at final follow up. Moreover, no patients in the study group required revision arthroplasty during the follow-up period. Although the average follow-up period was only 47 months, these results suggested that elderly patients with MCT without arthritis may be particularly ideal candidates for rTSA with regard to implant survival and anticipated revision rate when chosen appropriately.

Several weaknesses were noted within this paper. First, the study was retrospective, precluding randomization of treatment groups and standardization of data collection and follow-up. The outcomes of RCR and rTSA could not be compared directly due to the inherent selection bias. The groups clearly differed in many respects, and these preoperative factors likely played a role in postoperative outcomes. However, the primary goal of this study was not to compare outcomes of the treatment groups but to analyze the patterns of patient selection by an experienced treating surgeon and contribute to published data that each surgery can be successful in this patient population when chosen appropriately.

Second, our data were based on a single surgeon’s decisions, and results may not be generalizable. Furthermore, the senior author has had a longstanding interest in reverse shoulder arthroplasty and has published data illustrating successful outcomes for rTSA in patients with MCT. For this reason, one could presume that there may have been some bias toward treating patients with rTSA. However, we feel that the senior author’s unique and longstanding experience in treating MCT allows for a thorough evaluation and comparison of preoperative variables and outcomes declared within this study. Indeed, many patients included in this study were referred from outside institutions specifically for rTSA but instead were deemed more appropriate candidates for RCR and underwent successful arthroscopic repair, a common scenario which served as an impetus for this study.

CONCLUSION

RCR and rTSA are both viable options for patients >65 years with MCT without arthritis. Treatment must be individualized for each patient with careful consideration of a number of preoperative variables and patient characteristics. At our institution, patients with previous RCR, decreased range of motion, poor function, and strong radiographic evidence of subluxation are more likely to undergo rTSA. When chosen appropriately, both RCR and rTSA can result in improved range of motion, function, and high patient satisfaction in this patient population.

The dilemma for the treating surgeon is deciding which patients to treat with RCR and who to treat with rTSA. Predicting which surgical procedure will provide a better functional result is difficult and controversial.⁷ The RCR method is a bone-conserving procedure with relatively low surgical risk and allows the option for rTSA to be performed as a salvage surgery should repair fail. It also may be less costly in the appropriate population.⁸ However, large rotator cuff tears in elderly patients have low healing potential, and the prospect of participating in a lengthy rehabilitation after an operation that may not prove successful can be deterring.^9,10 In the elderly population, rTSA may be a reliable option, as tendon healing of the cuff is not necessary to restore function. However, rTSA does not conserve bone, provides a non-anatomic solution, and has had historically high complication rates.^4,5

In an effort to aid in the decision-making process when considering these 2 surgical options, we compared RCR and rTSA performed at a single institution for MCT in patients >65 years. Our aim was to identify preoperative patient variables that influence a surgeon’s decision to proceed with 1 of the 2 procedures. Moreover, we evaluated clinical outcomes in these 2 patient populations. We hypothesized that (1) patients selected for rTSA would have worse preoperative function, less range of motion, more comorbidities, more evidence of radiographic subluxation, and a higher likelihood of having undergone previous RCR than those selected for RCR, and (2) both RCR and rTSA would be successful and result in improved clinical outcomes with high patient satisfaction.

MATERIALS AND METHODS

PATIENT SELECTION

We performed a retrospective chart review using our practice database of all patients undergoing arthroscopic RCR and rTSA for any indication by the senior author (M.A.F.) between January 2004 and April 2015. A total of 1503 RCRs and 1973 rTSAs were conducted during the study period. Patient medical records were reviewed, and those meeting the following criteria were included in the study: >65 years at the time of surgery, MCT, no preoperative glenohumeral arthritis, minimum follow-up of 12 months, functional deltoid muscle on physical examination, and no prior shoulder surgery except for RCR or diagnostic arthroscopy. A total of 92 patients who underwent arthroscopic RCR and 89 patients who underwent rTSA met the inclusion criteria. For patients with bilateral shoulder surgery, we measured each shoulder independently. Three patients underwent bilateral rTSA, and 3 patients underwent bilateral RCR, leaving 95 shoulders in the RCR group and 92 in the rTSA group. The Western Institutional Review Board determined this study to be exempt from review.

RADIOGRAPHIC EVALUATION

All patient charts included a radiology report and documented interpretation of the images by the treating surgeon prior to surgery. Radiographs were assessed to assure the absence of preoperative glenohumeral osteoarthritis. The images were also graded based on the Hamada classification.¹¹ Stage 1 is associated with minimal radiographic change with an acromiohumeral interval (AHI) >6 mm; stage 2 is characterized by narrowing of the AHI <6 mm; and Stage 3 is defined by narrowing of the AHI with radiographic changes of the acromion. Stages 4 and higher include arthritic changes to the glenohumeral joint, and they were not included in the study population. The AHI measurements and the presence or absence of glenohumeral subluxation were documented.

Continue to: MASSIVE CUFF TEAR DETERMINATION...

We defined MCT on the basis of previously described criteria of tears involving ≥2 tendons or tears measuring ≥5 cm in greatest dimension.^12,13 Patient charts were screened, and those whose clinical notes or radiology reports indicated an absence of MCT were excluded. Preoperative imaging of the remaining patients was then evaluated by 3 fellowship-trained shoulder surgeons to confirm MCT in all patients with a clinically documented MCT, as well as to assess those who had insufficient documentation of tear size in the notes.

Advanced imaging was evaluated for fatty atrophy of the rotator cuff musculature, and Goutallier classification was assigned.^14,15 Length of retraction was measured from the tendon end to the medial aspect of the footprint on coronal imaging, and the subscapularis and teres minor were assessed and documented as torn or intact.^16,17

DATA COLLECTION

We reviewed clinical charts and patient questionnaire forms from both the preoperative and follow-up visits. Clinical data collected included gender, age at surgery, active range of motion (forward elevation, abduction, external rotation, and internal rotation), comorbidities, smoking status, BMI, history of shoulder surgery, and any postoperative complications or need for secondary surgery. All patients completed patient-centered questionnaires regarding shoulder pain and dysfunction at each visit or via telephone communication with clinic staff. Outcome measurements used for analysis included American Shoulder and Elbow Surgeons (ASES) Score, simple shoulder test (SST), visual analog score (VAS) pain scale, and patient-reported satisfaction (Graded 1-10; 1 = poor outcome; 4 = satisfactory outcome; 7 = good outcome; 10 = excellent outcome).

DATA ANALYSIS AND STATISTICAL METHODS

Statistical tests were selected based on the result of Shapiro–Wilk test for normality. Continuous variables were evaluated with either independent t test or Mann–Whitney U test. Dependent t test was used to evaluate outcome variables. For categorical variables, either Pearson’s χ²or Fisher’s exact test was performed depending on the sample size. Alpha was set at P =.05.

Continue to: RESULTS...

PREOPERATIVE CHARACTERISTICS

Of the 187 shoulders in the study group, 95 had RCR and 92 had rTSA. Demographic information and preoperative variables for both groups are summarized in Table 1 and Table 2. Patients in the RCR group had greater preoperative forward elevation, abduction, and external rotation and higher preoperative functional scores than those in the rTSA group. Patients in the rTSA group were older and more likely to be female than those in the RCR group. More patients in the rTSA group had undergone prior RCR compared with those in the RCR group. Each of these differences was statistically significant. Subjective pain scores, BMI, and comorbidities were similar between the 2 groups.

Table 1. Patient demographics

Abbreviations: BMI, body mass index; RCR, rotator cuff repair; rTSA, reverse total shoulder arthroplasty.

Table 2. Preoperative variables

Abbreviations: AB, abduction; ASES, American Shoulder and Elbow Society score; ER, external rotation; FE, forward elevation; IR, internal rotation; preop, preoperative; SST, simple shoulder test.

Radiographically, patients selected to undergo rTSA had a smaller AHI (4.8 vs 8.7, P < .0001) and more evidence of superior subluxation (50.6% vs 14.1%, P < .0001) than those in the RCR group. Average Hamada grade was 1.4 ± 0.7 and 2.2 ± 0.7 for the RCR and rTSA groups, respectively (P < .0001). Average Goutallier grade was similar between the groups (2.2 ± 0.6 for RCR vs 2.2 ± 0.8 for rTSA, P =.227), and 25.5% of the RCR group had Grade 3 or 4 atrophy compared with 28.1% of the rTSA group.

POSTOPERATIVE OUTCOMES

The average follow-up time was 44 months for RCR and 47 months for rTSA. Patients in the RCR and rTSA groups were highly satisfied with the surgery (8.5 ± 2.6 vs 8.2 ± 2.6, P = .461) and had significantly increased range of motion in all planes and improved functional scores (Table 3). The rTSA group had greater net improvement in forward elevation, abduction, and external rotation than the RCR group. Both groups demonstrated similar improvement in ASES, SST, and VAS pain scores.

Table 3. Postoperative outcomes

Abbreviations: AB, abduction; ASES, American Shoulder and Elbow Society score; ER, external rotation; FE, forward elevation; IR, internal rotation; SST, simple shoulder test; VAS – visual analog score.

In the RCR group, 5 patients (5.3%) required reoperation: 3 patients underwent conversion to rTSA, 1 patient underwent biceps tenotomy with subacromial decompression, and 1 patient underwent arthroscopic irrigation and debridement for a postoperative Propionibacterium acnes infection. In the rTSA group, 2 patients (2.2%) required reoperation: 1 patient underwent open reduction internal fixation for a scapula fracture that failed conservative management, and 1 patient had an open irrigation and debridement with polyethylene exchange for an acute postoperative infection of unknown source.

DISCUSSION

Massive, retracted rotator cuff tears are a common and difficult problem.¹ The treatment options are numerous and depend on a variety of preoperative factors including patient-specific characteristics and factors specific to the tear. For certain patients, nonoperative management may be a reasonable first step, as an MCT does not necessarily preclude painless, functional shoulder motion. Elderly, lower demand individuals have been shown to do well with physical rehabilitation.¹⁸ Similarly, for the same category of elderly patients who do not respond to conservative measures, arthroscopic tendon debridement with or without subacromial decompression and/or biceps tenotomy may be effective.^1,19 This technique has been described as “limited goals surgery;” despite some mixed results in the literature, multiple studies have reported symptomatic and functional improvement after simple debridement.^2,19–21The consensus among several authors has been that this procedure continues to play a role for elderly, low-demand patients whose functional goals are limited and whose primary complaint is pain.^1,2,20

For the majority of patients with MCT who desire pain relief and a restoration of shoulder function, RCR remains the gold standard of treatment and should be the primary aim if feasible. Complete RCR has consistently outperformed both partial repair and debridement in multiple studies in terms of pain relief and functional improvement.^10,21,22However, elderly patients with chronic, massive tears, particularly in the setting of muscle atrophy, are at high risk of failure with attempted cuff repair.^9,23 Novel techniques such as superior capsular reconstruction and subacromial balloon spacer implantation may offer a minimally invasive method of re-centering the humeral head and stabilizing the glenohumeral joint; however, these new treatment options lack any long-term data in the literature to support their widespread use.^24–26 Alternatively, rTSA has been shown to be a reliable option to restore shoulder function in the setting of a massive irreparable rotator cuff tear, even in the absence of arthritis.^5,27-31

Continue to: The decision-making process...

The decision-making process for selecting RCR or rTSA in the setting of MCT without arthritis in the older population (age >65 years) remains challenging. We attempted to quantify the data of a high-volume surgeon and identify the differences and similarities between those patients selected for either procedure. At our institution, we generally performed rTSA on patients with low preoperative range of motion, poor function based on SST and ASES scores, small AHI, and strong evidence of superior subluxation. We were also more likely to perform rTSA if the patient had a history of rotator cuff surgery. There was a predilection for older age and female gender in those who underwent rTSA.

For our study, we elected to focus on patients >65 years. In our experience, the choice of which surgical procedure to perform is generally easier in younger patients. Most surgeons appropriately opt for an arthroscopic procedure or tendon transfer to preserve bone and maintain the option of rTSA as a salvage procedure if necessary in the future. Studies have reported that <60 years is a predictor of poor outcome with rTSA, and patients <65 years who undergo rTSA have been shown to have high complication rates.^30-32 Furthermore, the longevity of the implant in young patients is a significant concern, and revision surgery after rTSA is technically demanding and known to result in poor functional outcomes.^32,33

Although the indications for rTSA are expanding, attempts at RCR in the setting of MCT remain largely appropriate. Preserved preoperative anterior elevation >90° has been associated with loss of motion after rTSA and poor satisfaction, and one should exercise caution when considering rTSA in this setting.³ The current study confirmed that even older patients with MCT may be very satisfied with arthroscopic RCR (Figure 1). Both range of motion and function significantly improved, and patients were largely satisfied with the procedure with an average self-reported outcome of good to excellent. At the time of final follow-up for this study, only 3 shoulders in the RCR group had undergone conversion to rTSA. This number may be expected to rise with long follow-up periods, and we feel that prolonging the time before arthroplasty is generally in the best interest of the patient.

Our results were consistent with several reported studies in which RCR has been shown to be successful in the setting of MCT.^34–37 Henry and colleagues³⁶ performed a systematic review that evaluated 954 patients who underwent partial or complete anatomic RCR for MCT. Although the average age was 63 years (range, 37–87), functional outcome scores, VAS pain score, and overall range of motion consistently and significantly improved.

rTSA may be a “more reliable” option than RCR in treating MCT in the older population because it does not rely on tendon healing. However, the relationship between tendon healing and clinical outcomes after RCR is unclear. The aforementioned systematic review reported re-tear rates to be as high as 79%, but several studies have reported high satisfaction even in the setting of retear.³⁶ Yoo and colleagues³⁸ and Chung and colleagues⁹ reported re-tear rates of 45.5% and 39.8%, respectively, but both studies noted that there was no difference in outcome measures between those patients with and without re-tears. In particular, for patients who have had no prior rotator cuff surgery, an attempt at arthroscopic repair may be a prudent option with relatively low risk.

Although certain patients may clinically improve despite suffering a re-tear (or inability to heal in the first place), others continue to experience pain and dysfunction that negatively affect their quality of life.^39–41 These patients are more often appropriate candidates for rTSA. Indeed, several studies have demonstrated a higher re-tear rate in patients with a history of surgery than in those without.^23,31,38,42 Shamsudin and colleagues⁴³ found revision arthroscopic RCR, even in a younger age group with tears of all sizes, to be twice as likely to re-tear. Notably, re-tear after revision repair may be more likely to be symptomatic, as these re-tears are routinely associated with pain, stiffness, and loss of function. Even in the hands of experienced surgeons in a younger population, revision repair has only been able to reverse pseudoparalysis in 43% of patients, leading to only 39% return to sport or full activity.⁴⁴ In examining our data, we were much less likely to perform an RCR in patients who had a history of cuff repair surgery than in those without this history.

Continue to: Overall, those patients selected for rTSA...

Overall, those patients selected for rTSA in our study population performed well postoperatively (Figure 2 and Figure 3). Vast improvements were noted in range of motion, function, and pain scores at final follow up. Moreover, no patients in the study group required revision arthroplasty during the follow-up period. Although the average follow-up period was only 47 months, these results suggested that elderly patients with MCT without arthritis may be particularly ideal candidates for rTSA with regard to implant survival and anticipated revision rate when chosen appropriately.

Several weaknesses were noted within this paper. First, the study was retrospective, precluding randomization of treatment groups and standardization of data collection and follow-up. The outcomes of RCR and rTSA could not be compared directly due to the inherent selection bias. The groups clearly differed in many respects, and these preoperative factors likely played a role in postoperative outcomes. However, the primary goal of this study was not to compare outcomes of the treatment groups but to analyze the patterns of patient selection by an experienced treating surgeon and contribute to published data that each surgery can be successful in this patient population when chosen appropriately.

Second, our data were based on a single surgeon’s decisions, and results may not be generalizable. Furthermore, the senior author has had a longstanding interest in reverse shoulder arthroplasty and has published data illustrating successful outcomes for rTSA in patients with MCT. For this reason, one could presume that there may have been some bias toward treating patients with rTSA. However, we feel that the senior author’s unique and longstanding experience in treating MCT allows for a thorough evaluation and comparison of preoperative variables and outcomes declared within this study. Indeed, many patients included in this study were referred from outside institutions specifically for rTSA but instead were deemed more appropriate candidates for RCR and underwent successful arthroscopic repair, a common scenario which served as an impetus for this study.

CONCLUSION

RCR and rTSA are both viable options for patients >65 years with MCT without arthritis. Treatment must be individualized for each patient with careful consideration of a number of preoperative variables and patient characteristics. At our institution, patients with previous RCR, decreased range of motion, poor function, and strong radiographic evidence of subluxation are more likely to undergo rTSA. When chosen appropriately, both RCR and rTSA can result in improved range of motion, function, and high patient satisfaction in this patient population.

Anatomic factors that contribute to PFI include ligament laxity, trochlear dysplasia, patella alta, excessive femoral anteversion or tibial torsion, genu valgum, and increased tibial tubercle-to-trochlear groove distance.^1,3,8-11 When considering surgical treatment for PFI, all anatomic factors should be considered, with an emphasis on, and understanding of, the role of residual growth and development. One must also consider balancing static soft tissue, dynamic soft tissue, and osseous constraints of the patellofemoral joint to optimize the overall health and balance of the patella (Figure 1).

CHALLENGES IN TREATING PEDIATRIC PFI

A primary challenge in the treatment of PFI in pediatric patients is accounting for the impact of anatomic changes occurring secondary to growth and maturity. From birth to adulthood, the collagen composition of soft tissue changes from an elastic type III collagen to a stiffer type I collagen.¹² These physiologic changes may influence the rigidity of the soft tissue restraints around the patellofemoral joint during periods of rapid growth. Longitudinal growth and rotational changes can also occur at the distal femoral and proximal tibial physes. The position of the tibial tubercle and the magnitude of femoral anteversion may also change after growth in adolescents.

Developing effective technical analogs of surgical procedures performed in mature patients with PFI for use in skeletally immature patients has been a second challenge. For example, a varus-producing distal femoral osteotomy to address genu valgum and PFI¹³ is contraindicated in immature patients, when a hemiepiphysiodesis for guided growth should be considered.¹⁴ Similarly, a periosteal sleeve medialization of the insertion of the patellar tendon may be used instead of a tibial tubercle transfer.¹⁵

If a medial patellofemoral ligament (MPFL) reconstruction is considered in immature patients, careful consideration of the position of the MPFL in relation to the distal femoral physis is paramount. Shea and colleagues¹⁶ originally described the position of the MPFL, based on Schottle’s point¹⁷ on lateral radiographs, to be proximal to the distal femoral physis (Figure 2). However, due to the undulation of the physis, the lateral projection may falsely demonstrate Schottle’s point to be on or just proximal to the physis. Other researchers have evaluated the position of the distal femoral physis to be proximal to the origin of the MPFL by a range of 2.9 mm to 8.5 mm on AP radiograph and MRI, respectively.^18,19 More recent cadaveric studies have demonstrated the origin of the MPFL in pediatric specimens in relation to the physis to be variable.²⁰ Although we believe that a relative anatomic femoral position of the MPFL origin can be achieved without disrupting the physis using Schottle’s point, some have indicated concerns that this may produce a nonisometric position, which has not been our experience.²¹

Continue to: TREATMENT OPTIONS...

The evaluation of a skeletally immature patient may differ from that of an adult patient.  (Table 1) represents a standard evaluation of a skeletally immature patient with PFI. The injury, physical examination findings, activity level, presence of an osteochondral fracture, and severity of bony dysplasia may indicate surgical treatment and influence selection of the surgical technique. Indications for surgical treatment include recurrent patellar instability, symptomatic patellar instability, the presence of a chondral or an osteochondral fracture, and a primary patella dislocation in patients with a high risk for recurrent dislocations. Table 2 represents a list of the possible surgical treatment options for the skeletally immature patient with PFI.^21,22 Variable outcomes and recurrence rates are noted with each procedure.^21,22

MPFL REPAIR VS RECONSTRUCTION

Although there is increased consensus for restoring normal anatomy, continued controversy exists regarding the utility of an MPFL repair versus reconstruction. An MPFL repair to the patella has been reported in a large series and was noted to have a 72% success rate.²³ Although a perceived benefit of the procedure is the absence of risk to the physis, concerns exist regarding the unacceptable rate of continued patellar instability after a repair of native tissue.

Several studies have demonstrated that reconstruction of the MPFL yields lower rates of recurrent dislocation, improved knee function, and an ability to return to prior levels of activity with little to no functional impairment.^1,2,24-26 Studies have also shown that MPFL reconstruction can be performed safely in skeletally immature patients, with little risk of growth impairment, good to excellent results, and low re-rupture rates.^27,28

We propose that MPFL reconstruction for the skeletally immature patient with PFI should be the primary surgical treatment. Any patient with atypical or severe bony dysplasia may warrant additional intervention. Additional correction of valgus, in the form of guided growth, should be considered in conjunction with an MPFL reconstruction if the patient demonstrates a valgus of grade >2 or a lateral distal femoral angle of <84°.

GRAFT OPTIONS

When an MPFL reconstruction is indicated, graft options include autologous hamstring, quadriceps, adductor, or patellar tendon grafts. Allograft tendon may be an acceptable choice, and use of synthetic graft has also been described. A recent systematic review²⁹ concluded that there was no difference in recurrence rates or outcomes based on graft type. However, in this study, the overall complication rates were higher in patients who had autologous graft than in those who had allograft. Although the use of allograft has not been specifically reported in a pediatric cohort, allografts have been successfully used in this age group. Obviating graft harvest eliminates an additional or extended incision, limits postoperative weakness, and may speed early recovery.

Continue to: PREFERRED SURGICAL TECHNIQUE FOR MPFL RECONSTRUCTION WITH ALLOGRAFT...

Surgery should be scheduled and performed as an outpatient procedure. Preoperative instructions may include crutch training, when indicated, and postoperative education regarding expected early therapy and pain management strategy may be employed. Chlorhexidine wipes are provided to patients preoperatively with instruction to use daily starting 3 days prior to surgery. A home exercise program with a focus on quadriceps activation and range of motion is provided and requested to be performed postoperatively.

A standard operating room setup for knee arthroscopy is employed (preference list noted in Table 3; items bolded may be unique to an MPFL reconstruction). Regional, single-shot anesthesia is employed using a sensory-only adductor canal block. A universal surgical timeout is performed before beginning the surgery to include verification of prophylactic antibiotics (Ancef or clindamycin in penicillin-allergic patients) and consideration of the use of a sequential compression device in children aged >10 years when ≥1 venous thromboembolic risk factors are detected in preoperative screening.

Nonirradiated allografts (semitendinosus or tibialis anterior), preferably from a donor <30 years, are our preferred graft choice. The minimum length of the graft should be 240 mm and the doubled thickness should be 5 to 6 mm. After thawing, the graft is lavaged with a mixture of antibiotic saline consisting of 50,000 U of bacitracin in 1 l of normal saline. Tension is then placed on the graft using the graft preparation board, and a whip stitch is placed on each side of the graft using a #2 FiberWire (Arthrex). The graft is sized with care to ensure that it is not thicker than 6.5 mm with an optimal goal of 6 mm.

A standard knee arthroscopy is performed with an emphasis on evaluating the patellofemoral anatomy. Although insufflation can cause distortion and, often, lateralize the patella, the surgeon should consider an estimate of when the patella engages the trochlea during knee flexion. This position of knee flexion will determine the appropriate position of the knee during graft fixation. A bipolar device (90°Arthrocare wand) may be used to debride synovial folds (ie, plica) when affecting the patellar tracking. To maximize visualization of the patella within the knee, we recommend switching the arthroscope from the anterolateral portal to the anteromedial portal. This allows for improved visualization of the lateral femoral condyle and the inferomedial patellar facet, common locations for chondral and osteochondral damage, as well as optimal visualization of patellar engagement within the trochlea. During an arthroscopic dynamic exam of the patella tracking, a lateralized patella may be observed. If the patella tilts upon manual medial translation toward the trochlear groove during the dynamic exam, the lateral retinaculum may be tight or constraining the patella laterally. If this occurs, a partial or complete lateral release may be indicated with a bipolar wand. Visualization of the posteromedial and the posterolateral compartments is required as loose chondral bodies may be present in these locations.     

Any osteochondral or chondral injuries are addressed during the arthroscopy. Large osteochondral fractures of the lateral femoral condyle or the patella are repaired using headless compression screws (Acutrak Screws, Acumed) or headed low-profile screws (1.5-mm screws, DePuy Synthes). Headless screws can be buried 2 to 3 mm below the chondral surface when the fragment has an appropriate thickness, whereas thin fragments may necessitate headed screws for adequate fixation and subsequent implant removal in 8 to 10 weeks. Defects in the patella most often require an open arthrotomy to evert the patella 90° to allow adequate exposure for treatment. Chronic chondral fragmentation may be debrided using the motorized chondrotome or removed and indicated for microfracture (<2 cm² in surface area and <4 mm in depth) or other types of chondral replacement. Chondral lesions over the inferior medial facet are typically not symptomatic and often require minimal treatment.

Continue to: There are 3 methods to identify...

There are 3 methods to identify the insertion of the MPFL into the patella. During the diagnostic arthroscopy, an 18-gauge needle can be used to mark the insertion of the MPFL as visualization of the ligament arthroscopically is often possible (Figure 3). Another useful technique is to follow the inferior aspect of the distal insertion of the vastus medialis oblique (VMO) into the patella. The typical insertion point of the MPFL is immediately distal to the insertion of the inferior aspect of the VMO (Figure 4). It is also helpful to note that the center point of the insertion of the MPFL is at the junction of the proximal one-third and distal two-thirds of the palpable osseous patella. The MPFL origin has been noted to be at the exact midpoint of the chondral surface of the patella or 5 mm proximal (41% of the length of the patella) to the midpoint of the osseous patella.³⁰

Following arthroscopic examination and treatment, a linear incision is made at the superior two-thirds of the patella, 1-finger breadth medial to the patella (Figure 5). During the subcutaneous dissection, the goal is to visualize the fascia overlying the VMO. Once this is identified, the dissection, in this layer, is carried over to expose the anterior and central surface of the patella. Army/navy retractors are used to retract the skin, and the assistant will place manual pressure on the patella to stabilize it for preparation of the patellar surface.

A bovie cautery or a knife is used to mark the insertion of the MPFL, which is immediately distal to the inferior border of the VMO. An incision over the medial surface of the patella creates a T-incision with elevation of the subsequent flaps proximal and distal. This allows exposure of the superficial and medial surface of the patella. During medial exposure of the patella, care is taken to avoid an arthrotomy by leaving the synovial lining attachment. A rongeur is used to decorticate the medial patella and the superficial surface (Figure 6). If an MPFL avulsion is present, it is often embedded within the soft tissues adjacent to the medial patellar. The MPFL avulsion can be exposed and removed if small. During an MPFL reconstruction, a repair of the avulsion is typically not performed. A double-loaded suture anchor is inserted at the site identified as the insertion of the native MPFL (3.0 Biocomposite SutureTak, Arthrex). A single suture anchor is used instead of an interosseous tunnel or double tunnels to avoid creating a large defect that may increase the risk of fracture in a small, skeletally immature patella.³¹

A hemostat is used to identify the layer between the medial retinaculum and the synovium over the medial soft tissue of the knee. The MPFL is a well-defined thickening of the medial retinaculum, and the ideal placement of the reconstruction is immediately inferior to this layer. Once this layer has been identified, a blunt hemostat is inserted to mark the end of a blind pouch that is apparent in this layer. This blind pouch is marked on the surface of the skin as the origin of the MPFL.

Fluoroscopy is now used to identify the origin of the MPFL on the medial femoral condyle. A perfect lateral is obtained by lining up the posterior condyle. Often, existent trochlear dysplasia will modify the normal appearance of the anterior structures of the condyles. Schottle’s point¹⁷ is USED to locate the origin of the MPFL. This is defined by drawing a line from the posterior cortex distally through the condyles (Figure 7). Two perpendicular lines are drawn, one to the extension of the posterior cortex at the level of the posterior extent of Blumensaat’s line and a second at the metaphyseal flare. Schottle’s point is located just anterior to the posterior cortex and in between these perpendicular lines. In a skeletally immature patient, this point appears to be at the level of the physis, on the lateral projection.

Continue to: An anteroposterior and/or a notch view projection...

An anteroposterior and/or a notch view projection is then performed to confirm the location of Schottle’s point, distal to the physis. Because the origin of the MPFL is posterior to the condyle, the notch view should be used to visualize Schottle’s point in relation to the physis in the coronal plane. This location has been confirmed to approximate the origin of the MPFL by published anatomic studies. If the graft measures 6 mm, an ideal distance of 5 to 8 mm distal to the physis will ensure a tunnel location with a proximal margin 2 to 4 mm distal to the physis. In our experience, Schottle’s point is approximately 5 to 7 mm distal to the physis, and rarely do we adjust our position. The guide wire is placed on the medial femoral condyle at this point and then angled 30° proximal to distal to allow tunnel trajectory distal to the undulating physis (Figure 8).

A 6-mm reamer is used over the guide wire to a depth of 20 mm. The newly created tunnel is now exposed and visualized. The graft is folded over at the center aspect and marked 20 mm from the tip of a Bio-tenodesis screwdriver (Arthrex). A 6.25 × 15 mm Bio-tenodesis screw is used. The graft and the screw are inserted into the condyle, and the appropriate graft fixation is confirmed by longitudinal tension placed upon the graft.

The hemostat or the passing suture previously placed into the blind pouch below the native MPFL is used to pass the graft immediately superficial to the synovial lining, in an extra-articular location. The ends of the graft are now exposed through the incision adjacent to the patella. The knee is confirmed to be flexed to approximately 45° on the surgical positioning triangleor at the flexion position determined during diagnostic arthroscopy at which the patella engages the trochlea. The graft is set to length alongside the medial surface of the patella at the position of the suture anchor, and a single stitch is placed through both ends of the graft at this adjacent position using a free needle (Davis Tonsil ½ Circle Taper Point; Vessel Medical). A hemostat is placed on the single suture throw to provisionally secure the graft in this trial position. This allows the surgeon to trial and examine the MPFL graft in extension, 45° of flexion, and terminal flexion. The goal is to provide a check rein to lateral translation of the patella and provide a firm endpoint to avoid further dislocation without overtension of the graft that may lead to increased contact pressures in the patellofemoral joint.

Once appropriate graft position is confirmed, using the double-loaded suture, we prefer to secure each limb separately to the decorticated medial patella. One end of the suture is threaded through the graft once and will act as a post. The other end of the suture is threaded through the graft 3 times in a modified Mason-Allen stitch. With the knee in that same position of flexion, the knot is tied and the graft is secured to the medial side of the patella. For secondary fixation, the remaining ends of the graft are passed below the periosteum on the anterior surface of the patella. Using a 0 Vicryl (Ethicon US), the periosteal flaps are sutured on top of the graft, incorporating the residual graft. The ends are cut and a repeat dynamic examination is performed to confirm the position of the patella and the patellar tracking and to ensure that overtensioning did not occur. Following irrigation, a standard closure is performed. The patient is placed in a hinged knee brace and cryotherapy is applied (Polar Care, Breg Inc).

POSTOPERATIVE PROTOCOL

For the initial or the acute postoperative phases, an emphasis on edema control, early quadriceps activation, and range of motion is recommended. We recommend weight-bearing as tolerated with the leg locked in extension until adequate quadriceps control is achieved. The patient must be able to perform 10 straight leg raises without an extension loss to be cleared to weight-bear as tolerated without motion restriction in the brace. Full motion is allowed immediately.

OUTCOMES

In our experience using isolated MPFL reconstruction in the skeletally immature patient, we have had no evidence of physeal arrest, leg-length inequality, or angular deformity, and only11.4% of patients have had recurrent instability. The mean Kujala score in this cohort was 90.4, with a mean Tegner activity of 7, after the procedure. All failures in our cohort had had severe trochlear dysplasia.

Anatomic factors that contribute to PFI include ligament laxity, trochlear dysplasia, patella alta, excessive femoral anteversion or tibial torsion, genu valgum, and increased tibial tubercle-to-trochlear groove distance.^1,3,8-11 When considering surgical treatment for PFI, all anatomic factors should be considered, with an emphasis on, and understanding of, the role of residual growth and development. One must also consider balancing static soft tissue, dynamic soft tissue, and osseous constraints of the patellofemoral joint to optimize the overall health and balance of the patella (Figure 1).

CHALLENGES IN TREATING PEDIATRIC PFI

A primary challenge in the treatment of PFI in pediatric patients is accounting for the impact of anatomic changes occurring secondary to growth and maturity. From birth to adulthood, the collagen composition of soft tissue changes from an elastic type III collagen to a stiffer type I collagen.¹² These physiologic changes may influence the rigidity of the soft tissue restraints around the patellofemoral joint during periods of rapid growth. Longitudinal growth and rotational changes can also occur at the distal femoral and proximal tibial physes. The position of the tibial tubercle and the magnitude of femoral anteversion may also change after growth in adolescents.

Developing effective technical analogs of surgical procedures performed in mature patients with PFI for use in skeletally immature patients has been a second challenge. For example, a varus-producing distal femoral osteotomy to address genu valgum and PFI¹³ is contraindicated in immature patients, when a hemiepiphysiodesis for guided growth should be considered.¹⁴ Similarly, a periosteal sleeve medialization of the insertion of the patellar tendon may be used instead of a tibial tubercle transfer.¹⁵

If a medial patellofemoral ligament (MPFL) reconstruction is considered in immature patients, careful consideration of the position of the MPFL in relation to the distal femoral physis is paramount. Shea and colleagues¹⁶ originally described the position of the MPFL, based on Schottle’s point¹⁷ on lateral radiographs, to be proximal to the distal femoral physis (Figure 2). However, due to the undulation of the physis, the lateral projection may falsely demonstrate Schottle’s point to be on or just proximal to the physis. Other researchers have evaluated the position of the distal femoral physis to be proximal to the origin of the MPFL by a range of 2.9 mm to 8.5 mm on AP radiograph and MRI, respectively.^18,19 More recent cadaveric studies have demonstrated the origin of the MPFL in pediatric specimens in relation to the physis to be variable.²⁰ Although we believe that a relative anatomic femoral position of the MPFL origin can be achieved without disrupting the physis using Schottle’s point, some have indicated concerns that this may produce a nonisometric position, which has not been our experience.²¹

Continue to: TREATMENT OPTIONS...

The evaluation of a skeletally immature patient may differ from that of an adult patient.  (Table 1) represents a standard evaluation of a skeletally immature patient with PFI. The injury, physical examination findings, activity level, presence of an osteochondral fracture, and severity of bony dysplasia may indicate surgical treatment and influence selection of the surgical technique. Indications for surgical treatment include recurrent patellar instability, symptomatic patellar instability, the presence of a chondral or an osteochondral fracture, and a primary patella dislocation in patients with a high risk for recurrent dislocations. Table 2 represents a list of the possible surgical treatment options for the skeletally immature patient with PFI.^21,22 Variable outcomes and recurrence rates are noted with each procedure.^21,22

MPFL REPAIR VS RECONSTRUCTION

Although there is increased consensus for restoring normal anatomy, continued controversy exists regarding the utility of an MPFL repair versus reconstruction. An MPFL repair to the patella has been reported in a large series and was noted to have a 72% success rate.²³ Although a perceived benefit of the procedure is the absence of risk to the physis, concerns exist regarding the unacceptable rate of continued patellar instability after a repair of native tissue.

Several studies have demonstrated that reconstruction of the MPFL yields lower rates of recurrent dislocation, improved knee function, and an ability to return to prior levels of activity with little to no functional impairment.^1,2,24-26 Studies have also shown that MPFL reconstruction can be performed safely in skeletally immature patients, with little risk of growth impairment, good to excellent results, and low re-rupture rates.^27,28

We propose that MPFL reconstruction for the skeletally immature patient with PFI should be the primary surgical treatment. Any patient with atypical or severe bony dysplasia may warrant additional intervention. Additional correction of valgus, in the form of guided growth, should be considered in conjunction with an MPFL reconstruction if the patient demonstrates a valgus of grade >2 or a lateral distal femoral angle of <84°.

GRAFT OPTIONS

When an MPFL reconstruction is indicated, graft options include autologous hamstring, quadriceps, adductor, or patellar tendon grafts. Allograft tendon may be an acceptable choice, and use of synthetic graft has also been described. A recent systematic review²⁹ concluded that there was no difference in recurrence rates or outcomes based on graft type. However, in this study, the overall complication rates were higher in patients who had autologous graft than in those who had allograft. Although the use of allograft has not been specifically reported in a pediatric cohort, allografts have been successfully used in this age group. Obviating graft harvest eliminates an additional or extended incision, limits postoperative weakness, and may speed early recovery.

Continue to: PREFERRED SURGICAL TECHNIQUE FOR MPFL RECONSTRUCTION WITH ALLOGRAFT...

Surgery should be scheduled and performed as an outpatient procedure. Preoperative instructions may include crutch training, when indicated, and postoperative education regarding expected early therapy and pain management strategy may be employed. Chlorhexidine wipes are provided to patients preoperatively with instruction to use daily starting 3 days prior to surgery. A home exercise program with a focus on quadriceps activation and range of motion is provided and requested to be performed postoperatively.

A standard operating room setup for knee arthroscopy is employed (preference list noted in Table 3; items bolded may be unique to an MPFL reconstruction). Regional, single-shot anesthesia is employed using a sensory-only adductor canal block. A universal surgical timeout is performed before beginning the surgery to include verification of prophylactic antibiotics (Ancef or clindamycin in penicillin-allergic patients) and consideration of the use of a sequential compression device in children aged >10 years when ≥1 venous thromboembolic risk factors are detected in preoperative screening.

Nonirradiated allografts (semitendinosus or tibialis anterior), preferably from a donor <30 years, are our preferred graft choice. The minimum length of the graft should be 240 mm and the doubled thickness should be 5 to 6 mm. After thawing, the graft is lavaged with a mixture of antibiotic saline consisting of 50,000 U of bacitracin in 1 l of normal saline. Tension is then placed on the graft using the graft preparation board, and a whip stitch is placed on each side of the graft using a #2 FiberWire (Arthrex). The graft is sized with care to ensure that it is not thicker than 6.5 mm with an optimal goal of 6 mm.

A standard knee arthroscopy is performed with an emphasis on evaluating the patellofemoral anatomy. Although insufflation can cause distortion and, often, lateralize the patella, the surgeon should consider an estimate of when the patella engages the trochlea during knee flexion. This position of knee flexion will determine the appropriate position of the knee during graft fixation. A bipolar device (90°Arthrocare wand) may be used to debride synovial folds (ie, plica) when affecting the patellar tracking. To maximize visualization of the patella within the knee, we recommend switching the arthroscope from the anterolateral portal to the anteromedial portal. This allows for improved visualization of the lateral femoral condyle and the inferomedial patellar facet, common locations for chondral and osteochondral damage, as well as optimal visualization of patellar engagement within the trochlea. During an arthroscopic dynamic exam of the patella tracking, a lateralized patella may be observed. If the patella tilts upon manual medial translation toward the trochlear groove during the dynamic exam, the lateral retinaculum may be tight or constraining the patella laterally. If this occurs, a partial or complete lateral release may be indicated with a bipolar wand. Visualization of the posteromedial and the posterolateral compartments is required as loose chondral bodies may be present in these locations.     

Any osteochondral or chondral injuries are addressed during the arthroscopy. Large osteochondral fractures of the lateral femoral condyle or the patella are repaired using headless compression screws (Acutrak Screws, Acumed) or headed low-profile screws (1.5-mm screws, DePuy Synthes). Headless screws can be buried 2 to 3 mm below the chondral surface when the fragment has an appropriate thickness, whereas thin fragments may necessitate headed screws for adequate fixation and subsequent implant removal in 8 to 10 weeks. Defects in the patella most often require an open arthrotomy to evert the patella 90° to allow adequate exposure for treatment. Chronic chondral fragmentation may be debrided using the motorized chondrotome or removed and indicated for microfracture (<2 cm² in surface area and <4 mm in depth) or other types of chondral replacement. Chondral lesions over the inferior medial facet are typically not symptomatic and often require minimal treatment.

Continue to: There are 3 methods to identify...

There are 3 methods to identify the insertion of the MPFL into the patella. During the diagnostic arthroscopy, an 18-gauge needle can be used to mark the insertion of the MPFL as visualization of the ligament arthroscopically is often possible (Figure 3). Another useful technique is to follow the inferior aspect of the distal insertion of the vastus medialis oblique (VMO) into the patella. The typical insertion point of the MPFL is immediately distal to the insertion of the inferior aspect of the VMO (Figure 4). It is also helpful to note that the center point of the insertion of the MPFL is at the junction of the proximal one-third and distal two-thirds of the palpable osseous patella. The MPFL origin has been noted to be at the exact midpoint of the chondral surface of the patella or 5 mm proximal (41% of the length of the patella) to the midpoint of the osseous patella.³⁰

Following arthroscopic examination and treatment, a linear incision is made at the superior two-thirds of the patella, 1-finger breadth medial to the patella (Figure 5). During the subcutaneous dissection, the goal is to visualize the fascia overlying the VMO. Once this is identified, the dissection, in this layer, is carried over to expose the anterior and central surface of the patella. Army/navy retractors are used to retract the skin, and the assistant will place manual pressure on the patella to stabilize it for preparation of the patellar surface.

A bovie cautery or a knife is used to mark the insertion of the MPFL, which is immediately distal to the inferior border of the VMO. An incision over the medial surface of the patella creates a T-incision with elevation of the subsequent flaps proximal and distal. This allows exposure of the superficial and medial surface of the patella. During medial exposure of the patella, care is taken to avoid an arthrotomy by leaving the synovial lining attachment. A rongeur is used to decorticate the medial patella and the superficial surface (Figure 6). If an MPFL avulsion is present, it is often embedded within the soft tissues adjacent to the medial patellar. The MPFL avulsion can be exposed and removed if small. During an MPFL reconstruction, a repair of the avulsion is typically not performed. A double-loaded suture anchor is inserted at the site identified as the insertion of the native MPFL (3.0 Biocomposite SutureTak, Arthrex). A single suture anchor is used instead of an interosseous tunnel or double tunnels to avoid creating a large defect that may increase the risk of fracture in a small, skeletally immature patella.³¹

A hemostat is used to identify the layer between the medial retinaculum and the synovium over the medial soft tissue of the knee. The MPFL is a well-defined thickening of the medial retinaculum, and the ideal placement of the reconstruction is immediately inferior to this layer. Once this layer has been identified, a blunt hemostat is inserted to mark the end of a blind pouch that is apparent in this layer. This blind pouch is marked on the surface of the skin as the origin of the MPFL.

Fluoroscopy is now used to identify the origin of the MPFL on the medial femoral condyle. A perfect lateral is obtained by lining up the posterior condyle. Often, existent trochlear dysplasia will modify the normal appearance of the anterior structures of the condyles. Schottle’s point¹⁷ is USED to locate the origin of the MPFL. This is defined by drawing a line from the posterior cortex distally through the condyles (Figure 7). Two perpendicular lines are drawn, one to the extension of the posterior cortex at the level of the posterior extent of Blumensaat’s line and a second at the metaphyseal flare. Schottle’s point is located just anterior to the posterior cortex and in between these perpendicular lines. In a skeletally immature patient, this point appears to be at the level of the physis, on the lateral projection.

Continue to: An anteroposterior and/or a notch view projection...

An anteroposterior and/or a notch view projection is then performed to confirm the location of Schottle’s point, distal to the physis. Because the origin of the MPFL is posterior to the condyle, the notch view should be used to visualize Schottle’s point in relation to the physis in the coronal plane. This location has been confirmed to approximate the origin of the MPFL by published anatomic studies. If the graft measures 6 mm, an ideal distance of 5 to 8 mm distal to the physis will ensure a tunnel location with a proximal margin 2 to 4 mm distal to the physis. In our experience, Schottle’s point is approximately 5 to 7 mm distal to the physis, and rarely do we adjust our position. The guide wire is placed on the medial femoral condyle at this point and then angled 30° proximal to distal to allow tunnel trajectory distal to the undulating physis (Figure 8).

A 6-mm reamer is used over the guide wire to a depth of 20 mm. The newly created tunnel is now exposed and visualized. The graft is folded over at the center aspect and marked 20 mm from the tip of a Bio-tenodesis screwdriver (Arthrex). A 6.25 × 15 mm Bio-tenodesis screw is used. The graft and the screw are inserted into the condyle, and the appropriate graft fixation is confirmed by longitudinal tension placed upon the graft.

The hemostat or the passing suture previously placed into the blind pouch below the native MPFL is used to pass the graft immediately superficial to the synovial lining, in an extra-articular location. The ends of the graft are now exposed through the incision adjacent to the patella. The knee is confirmed to be flexed to approximately 45° on the surgical positioning triangleor at the flexion position determined during diagnostic arthroscopy at which the patella engages the trochlea. The graft is set to length alongside the medial surface of the patella at the position of the suture anchor, and a single stitch is placed through both ends of the graft at this adjacent position using a free needle (Davis Tonsil ½ Circle Taper Point; Vessel Medical). A hemostat is placed on the single suture throw to provisionally secure the graft in this trial position. This allows the surgeon to trial and examine the MPFL graft in extension, 45° of flexion, and terminal flexion. The goal is to provide a check rein to lateral translation of the patella and provide a firm endpoint to avoid further dislocation without overtension of the graft that may lead to increased contact pressures in the patellofemoral joint.

Once appropriate graft position is confirmed, using the double-loaded suture, we prefer to secure each limb separately to the decorticated medial patella. One end of the suture is threaded through the graft once and will act as a post. The other end of the suture is threaded through the graft 3 times in a modified Mason-Allen stitch. With the knee in that same position of flexion, the knot is tied and the graft is secured to the medial side of the patella. For secondary fixation, the remaining ends of the graft are passed below the periosteum on the anterior surface of the patella. Using a 0 Vicryl (Ethicon US), the periosteal flaps are sutured on top of the graft, incorporating the residual graft. The ends are cut and a repeat dynamic examination is performed to confirm the position of the patella and the patellar tracking and to ensure that overtensioning did not occur. Following irrigation, a standard closure is performed. The patient is placed in a hinged knee brace and cryotherapy is applied (Polar Care, Breg Inc).

POSTOPERATIVE PROTOCOL

For the initial or the acute postoperative phases, an emphasis on edema control, early quadriceps activation, and range of motion is recommended. We recommend weight-bearing as tolerated with the leg locked in extension until adequate quadriceps control is achieved. The patient must be able to perform 10 straight leg raises without an extension loss to be cleared to weight-bear as tolerated without motion restriction in the brace. Full motion is allowed immediately.

OUTCOMES

In our experience using isolated MPFL reconstruction in the skeletally immature patient, we have had no evidence of physeal arrest, leg-length inequality, or angular deformity, and only11.4% of patients have had recurrent instability. The mean Kujala score in this cohort was 90.4, with a mean Tegner activity of 7, after the procedure. All failures in our cohort had had severe trochlear dysplasia.

Anatomic factors that contribute to PFI include ligament laxity, trochlear dysplasia, patella alta, excessive femoral anteversion or tibial torsion, genu valgum, and increased tibial tubercle-to-trochlear groove distance.^1,3,8-11 When considering surgical treatment for PFI, all anatomic factors should be considered, with an emphasis on, and understanding of, the role of residual growth and development. One must also consider balancing static soft tissue, dynamic soft tissue, and osseous constraints of the patellofemoral joint to optimize the overall health and balance of the patella (Figure 1).

CHALLENGES IN TREATING PEDIATRIC PFI

A primary challenge in the treatment of PFI in pediatric patients is accounting for the impact of anatomic changes occurring secondary to growth and maturity. From birth to adulthood, the collagen composition of soft tissue changes from an elastic type III collagen to a stiffer type I collagen.¹² These physiologic changes may influence the rigidity of the soft tissue restraints around the patellofemoral joint during periods of rapid growth. Longitudinal growth and rotational changes can also occur at the distal femoral and proximal tibial physes. The position of the tibial tubercle and the magnitude of femoral anteversion may also change after growth in adolescents.

Developing effective technical analogs of surgical procedures performed in mature patients with PFI for use in skeletally immature patients has been a second challenge. For example, a varus-producing distal femoral osteotomy to address genu valgum and PFI¹³ is contraindicated in immature patients, when a hemiepiphysiodesis for guided growth should be considered.¹⁴ Similarly, a periosteal sleeve medialization of the insertion of the patellar tendon may be used instead of a tibial tubercle transfer.¹⁵

If a medial patellofemoral ligament (MPFL) reconstruction is considered in immature patients, careful consideration of the position of the MPFL in relation to the distal femoral physis is paramount. Shea and colleagues¹⁶ originally described the position of the MPFL, based on Schottle’s point¹⁷ on lateral radiographs, to be proximal to the distal femoral physis (Figure 2). However, due to the undulation of the physis, the lateral projection may falsely demonstrate Schottle’s point to be on or just proximal to the physis. Other researchers have evaluated the position of the distal femoral physis to be proximal to the origin of the MPFL by a range of 2.9 mm to 8.5 mm on AP radiograph and MRI, respectively.^18,19 More recent cadaveric studies have demonstrated the origin of the MPFL in pediatric specimens in relation to the physis to be variable.²⁰ Although we believe that a relative anatomic femoral position of the MPFL origin can be achieved without disrupting the physis using Schottle’s point, some have indicated concerns that this may produce a nonisometric position, which has not been our experience.²¹

Continue to: TREATMENT OPTIONS...

The evaluation of a skeletally immature patient may differ from that of an adult patient.  (Table 1) represents a standard evaluation of a skeletally immature patient with PFI. The injury, physical examination findings, activity level, presence of an osteochondral fracture, and severity of bony dysplasia may indicate surgical treatment and influence selection of the surgical technique. Indications for surgical treatment include recurrent patellar instability, symptomatic patellar instability, the presence of a chondral or an osteochondral fracture, and a primary patella dislocation in patients with a high risk for recurrent dislocations. Table 2 represents a list of the possible surgical treatment options for the skeletally immature patient with PFI.^21,22 Variable outcomes and recurrence rates are noted with each procedure.^21,22

MPFL REPAIR VS RECONSTRUCTION

Although there is increased consensus for restoring normal anatomy, continued controversy exists regarding the utility of an MPFL repair versus reconstruction. An MPFL repair to the patella has been reported in a large series and was noted to have a 72% success rate.²³ Although a perceived benefit of the procedure is the absence of risk to the physis, concerns exist regarding the unacceptable rate of continued patellar instability after a repair of native tissue.

Several studies have demonstrated that reconstruction of the MPFL yields lower rates of recurrent dislocation, improved knee function, and an ability to return to prior levels of activity with little to no functional impairment.^1,2,24-26 Studies have also shown that MPFL reconstruction can be performed safely in skeletally immature patients, with little risk of growth impairment, good to excellent results, and low re-rupture rates.^27,28

We propose that MPFL reconstruction for the skeletally immature patient with PFI should be the primary surgical treatment. Any patient with atypical or severe bony dysplasia may warrant additional intervention. Additional correction of valgus, in the form of guided growth, should be considered in conjunction with an MPFL reconstruction if the patient demonstrates a valgus of grade >2 or a lateral distal femoral angle of <84°.

GRAFT OPTIONS

When an MPFL reconstruction is indicated, graft options include autologous hamstring, quadriceps, adductor, or patellar tendon grafts. Allograft tendon may be an acceptable choice, and use of synthetic graft has also been described. A recent systematic review²⁹ concluded that there was no difference in recurrence rates or outcomes based on graft type. However, in this study, the overall complication rates were higher in patients who had autologous graft than in those who had allograft. Although the use of allograft has not been specifically reported in a pediatric cohort, allografts have been successfully used in this age group. Obviating graft harvest eliminates an additional or extended incision, limits postoperative weakness, and may speed early recovery.

Continue to: PREFERRED SURGICAL TECHNIQUE FOR MPFL RECONSTRUCTION WITH ALLOGRAFT...

Surgery should be scheduled and performed as an outpatient procedure. Preoperative instructions may include crutch training, when indicated, and postoperative education regarding expected early therapy and pain management strategy may be employed. Chlorhexidine wipes are provided to patients preoperatively with instruction to use daily starting 3 days prior to surgery. A home exercise program with a focus on quadriceps activation and range of motion is provided and requested to be performed postoperatively.

A standard operating room setup for knee arthroscopy is employed (preference list noted in Table 3; items bolded may be unique to an MPFL reconstruction). Regional, single-shot anesthesia is employed using a sensory-only adductor canal block. A universal surgical timeout is performed before beginning the surgery to include verification of prophylactic antibiotics (Ancef or clindamycin in penicillin-allergic patients) and consideration of the use of a sequential compression device in children aged >10 years when ≥1 venous thromboembolic risk factors are detected in preoperative screening.

Nonirradiated allografts (semitendinosus or tibialis anterior), preferably from a donor <30 years, are our preferred graft choice. The minimum length of the graft should be 240 mm and the doubled thickness should be 5 to 6 mm. After thawing, the graft is lavaged with a mixture of antibiotic saline consisting of 50,000 U of bacitracin in 1 l of normal saline. Tension is then placed on the graft using the graft preparation board, and a whip stitch is placed on each side of the graft using a #2 FiberWire (Arthrex). The graft is sized with care to ensure that it is not thicker than 6.5 mm with an optimal goal of 6 mm.

A standard knee arthroscopy is performed with an emphasis on evaluating the patellofemoral anatomy. Although insufflation can cause distortion and, often, lateralize the patella, the surgeon should consider an estimate of when the patella engages the trochlea during knee flexion. This position of knee flexion will determine the appropriate position of the knee during graft fixation. A bipolar device (90°Arthrocare wand) may be used to debride synovial folds (ie, plica) when affecting the patellar tracking. To maximize visualization of the patella within the knee, we recommend switching the arthroscope from the anterolateral portal to the anteromedial portal. This allows for improved visualization of the lateral femoral condyle and the inferomedial patellar facet, common locations for chondral and osteochondral damage, as well as optimal visualization of patellar engagement within the trochlea. During an arthroscopic dynamic exam of the patella tracking, a lateralized patella may be observed. If the patella tilts upon manual medial translation toward the trochlear groove during the dynamic exam, the lateral retinaculum may be tight or constraining the patella laterally. If this occurs, a partial or complete lateral release may be indicated with a bipolar wand. Visualization of the posteromedial and the posterolateral compartments is required as loose chondral bodies may be present in these locations.     

Any osteochondral or chondral injuries are addressed during the arthroscopy. Large osteochondral fractures of the lateral femoral condyle or the patella are repaired using headless compression screws (Acutrak Screws, Acumed) or headed low-profile screws (1.5-mm screws, DePuy Synthes). Headless screws can be buried 2 to 3 mm below the chondral surface when the fragment has an appropriate thickness, whereas thin fragments may necessitate headed screws for adequate fixation and subsequent implant removal in 8 to 10 weeks. Defects in the patella most often require an open arthrotomy to evert the patella 90° to allow adequate exposure for treatment. Chronic chondral fragmentation may be debrided using the motorized chondrotome or removed and indicated for microfracture (<2 cm² in surface area and <4 mm in depth) or other types of chondral replacement. Chondral lesions over the inferior medial facet are typically not symptomatic and often require minimal treatment.

Continue to: There are 3 methods to identify...

There are 3 methods to identify the insertion of the MPFL into the patella. During the diagnostic arthroscopy, an 18-gauge needle can be used to mark the insertion of the MPFL as visualization of the ligament arthroscopically is often possible (Figure 3). Another useful technique is to follow the inferior aspect of the distal insertion of the vastus medialis oblique (VMO) into the patella. The typical insertion point of the MPFL is immediately distal to the insertion of the inferior aspect of the VMO (Figure 4). It is also helpful to note that the center point of the insertion of the MPFL is at the junction of the proximal one-third and distal two-thirds of the palpable osseous patella. The MPFL origin has been noted to be at the exact midpoint of the chondral surface of the patella or 5 mm proximal (41% of the length of the patella) to the midpoint of the osseous patella.³⁰

Following arthroscopic examination and treatment, a linear incision is made at the superior two-thirds of the patella, 1-finger breadth medial to the patella (Figure 5). During the subcutaneous dissection, the goal is to visualize the fascia overlying the VMO. Once this is identified, the dissection, in this layer, is carried over to expose the anterior and central surface of the patella. Army/navy retractors are used to retract the skin, and the assistant will place manual pressure on the patella to stabilize it for preparation of the patellar surface.

A bovie cautery or a knife is used to mark the insertion of the MPFL, which is immediately distal to the inferior border of the VMO. An incision over the medial surface of the patella creates a T-incision with elevation of the subsequent flaps proximal and distal. This allows exposure of the superficial and medial surface of the patella. During medial exposure of the patella, care is taken to avoid an arthrotomy by leaving the synovial lining attachment. A rongeur is used to decorticate the medial patella and the superficial surface (Figure 6). If an MPFL avulsion is present, it is often embedded within the soft tissues adjacent to the medial patellar. The MPFL avulsion can be exposed and removed if small. During an MPFL reconstruction, a repair of the avulsion is typically not performed. A double-loaded suture anchor is inserted at the site identified as the insertion of the native MPFL (3.0 Biocomposite SutureTak, Arthrex). A single suture anchor is used instead of an interosseous tunnel or double tunnels to avoid creating a large defect that may increase the risk of fracture in a small, skeletally immature patella.³¹

A hemostat is used to identify the layer between the medial retinaculum and the synovium over the medial soft tissue of the knee. The MPFL is a well-defined thickening of the medial retinaculum, and the ideal placement of the reconstruction is immediately inferior to this layer. Once this layer has been identified, a blunt hemostat is inserted to mark the end of a blind pouch that is apparent in this layer. This blind pouch is marked on the surface of the skin as the origin of the MPFL.

Fluoroscopy is now used to identify the origin of the MPFL on the medial femoral condyle. A perfect lateral is obtained by lining up the posterior condyle. Often, existent trochlear dysplasia will modify the normal appearance of the anterior structures of the condyles. Schottle’s point¹⁷ is USED to locate the origin of the MPFL. This is defined by drawing a line from the posterior cortex distally through the condyles (Figure 7). Two perpendicular lines are drawn, one to the extension of the posterior cortex at the level of the posterior extent of Blumensaat’s line and a second at the metaphyseal flare. Schottle’s point is located just anterior to the posterior cortex and in between these perpendicular lines. In a skeletally immature patient, this point appears to be at the level of the physis, on the lateral projection.

Continue to: An anteroposterior and/or a notch view projection...

An anteroposterior and/or a notch view projection is then performed to confirm the location of Schottle’s point, distal to the physis. Because the origin of the MPFL is posterior to the condyle, the notch view should be used to visualize Schottle’s point in relation to the physis in the coronal plane. This location has been confirmed to approximate the origin of the MPFL by published anatomic studies. If the graft measures 6 mm, an ideal distance of 5 to 8 mm distal to the physis will ensure a tunnel location with a proximal margin 2 to 4 mm distal to the physis. In our experience, Schottle’s point is approximately 5 to 7 mm distal to the physis, and rarely do we adjust our position. The guide wire is placed on the medial femoral condyle at this point and then angled 30° proximal to distal to allow tunnel trajectory distal to the undulating physis (Figure 8).

A 6-mm reamer is used over the guide wire to a depth of 20 mm. The newly created tunnel is now exposed and visualized. The graft is folded over at the center aspect and marked 20 mm from the tip of a Bio-tenodesis screwdriver (Arthrex). A 6.25 × 15 mm Bio-tenodesis screw is used. The graft and the screw are inserted into the condyle, and the appropriate graft fixation is confirmed by longitudinal tension placed upon the graft.

The hemostat or the passing suture previously placed into the blind pouch below the native MPFL is used to pass the graft immediately superficial to the synovial lining, in an extra-articular location. The ends of the graft are now exposed through the incision adjacent to the patella. The knee is confirmed to be flexed to approximately 45° on the surgical positioning triangleor at the flexion position determined during diagnostic arthroscopy at which the patella engages the trochlea. The graft is set to length alongside the medial surface of the patella at the position of the suture anchor, and a single stitch is placed through both ends of the graft at this adjacent position using a free needle (Davis Tonsil ½ Circle Taper Point; Vessel Medical). A hemostat is placed on the single suture throw to provisionally secure the graft in this trial position. This allows the surgeon to trial and examine the MPFL graft in extension, 45° of flexion, and terminal flexion. The goal is to provide a check rein to lateral translation of the patella and provide a firm endpoint to avoid further dislocation without overtension of the graft that may lead to increased contact pressures in the patellofemoral joint.

Once appropriate graft position is confirmed, using the double-loaded suture, we prefer to secure each limb separately to the decorticated medial patella. One end of the suture is threaded through the graft once and will act as a post. The other end of the suture is threaded through the graft 3 times in a modified Mason-Allen stitch. With the knee in that same position of flexion, the knot is tied and the graft is secured to the medial side of the patella. For secondary fixation, the remaining ends of the graft are passed below the periosteum on the anterior surface of the patella. Using a 0 Vicryl (Ethicon US), the periosteal flaps are sutured on top of the graft, incorporating the residual graft. The ends are cut and a repeat dynamic examination is performed to confirm the position of the patella and the patellar tracking and to ensure that overtensioning did not occur. Following irrigation, a standard closure is performed. The patient is placed in a hinged knee brace and cryotherapy is applied (Polar Care, Breg Inc).

POSTOPERATIVE PROTOCOL

For the initial or the acute postoperative phases, an emphasis on edema control, early quadriceps activation, and range of motion is recommended. We recommend weight-bearing as tolerated with the leg locked in extension until adequate quadriceps control is achieved. The patient must be able to perform 10 straight leg raises without an extension loss to be cleared to weight-bear as tolerated without motion restriction in the brace. Full motion is allowed immediately.

OUTCOMES

In our experience using isolated MPFL reconstruction in the skeletally immature patient, we have had no evidence of physeal arrest, leg-length inequality, or angular deformity, and only11.4% of patients have had recurrent instability. The mean Kujala score in this cohort was 90.4, with a mean Tegner activity of 7, after the procedure. All failures in our cohort had had severe trochlear dysplasia.

It is interesting to note that the typical ACL injury pattern occurs during a normal sport-specific movement, yet most rehabilitation protocols fail to recognize the potential causes of the aberrant movement pattern and how to best modify it so that the risk of repeated stress to the ACL can be minimized. It should be understood that movement occurs through the interaction of 3 discrete factors: the individual, the task being performed, and the environment in which it is performed.¹ All of these factors will play a role in how the final movement pattern is produced. For example, a soccer player (individual) may backpedal and pivot to the left 60° and accelerate to sprint after a player moving towards the touchline (task) while receiving instructions from teammates and monitoring the movements of opposing players (environment). A small variance in any 1 of these factors could significantly impact the movement pattern as the player completes the task.

In most rehabilitation programs, each of these factors may be treated in a singular, non-specific manner, but if these factors are not coordinated effectively throughout the program to produce the desired sport-specific movement, a faulty pattern may persist, leaving the player at risk for injury. Current rehabilitation programs seem to have a strong focus on creating stability, mobility, and strength, but these are trained in silos, with an internal focus of control, which only solves the biomechanical equation. Often, it is difficult for the player to coordinate good biomechanics into an efficient, protective movement pattern that is specific to the tasks performed on the field during the normal course of play. The missing link here is the central nervous system (CNS).

Limitations to the current ACL protocols may be that they rely heavily on musculoskeletal rehabilitation and that they have limited emphasis on neurological rehabilitation. As will be discussed later, the CNS has a large impact on the final movement selected by the player. In fact, cognition, perception, and action are the three factors that comprise the individual’s part of the movement paradigm,¹ yet rarely are these factors addressed in most ACL rehabilitation programs. These elements are a large part of the movement equation, so it is easy to understand how failing to address these features can lead to poor movement quality and subsequent ACL re-injury.

In addition to central neurological factors, cognitive issues may play a role in the player’s ability to return to sport. Determining optimal readiness for return to play is a difficult task for the medical community, with many variables to consider. Previous research studies have assessed the variability in return to play for various sports, including football, rugby, soccer, skiing, running, and tennis, with return-to-play rates ranging from 18% to 100%.^2,3,4-10 The risk of secondary injury may cast doubt and fear on athletes as they contemplate their successful return to play.^8,11 Although robust functional testing has become commonplace for determining athlete readiness after injury,^12-20 the assessment of psychological readiness, persistent fear, and loss of confidence are often neglected and not as commonly integrated into the return-to-play algorithm.^21-24 The purpose of this paper is to assess the various cognitive and central neural factors affecting a soccer player’s ability to recover from an ACL injury and offer suggestions for integrating treatments into the protocols to address these issues.

Continue to: CENTRAL NERVOUS SYSTEM NEUROPLASTICITY... 

Despite the vast amount of attention and research focused on the ACL, the re-injury rate still remains quite high. It has been reported that rehabilitation programs that employ traditional neuromotor training produce a re-injury rate as high as 30% after the athlete returns to sport.^25-28 The overall rate of sustaining a second ACL injury is 15%11 in all patient populations. For the general population <25 years of age, the re-injury rate is 21%, and for athletes <25 years of age, the re-injury rate rises to 23%.¹¹ With re-injury rates at this level, it is certainly fair to consider and be critical of the current rehabilitation methods being used with this population. One opportunity for improvement lies in the general approach used to rehabilitate ACL-injured patients. Therapy for this injury is protocol-driven, and the fact remains that most protocols prioritize restoration of peripheral systems, with minimal thought given to the cortical control necessary to manage those systems.^29,30 When neural factors are considered, it is usually within the context of increasing strength, balance, power, and biomechanical control,^31-34 which are certainly important but peripheral factors nonetheless. The missing element in many ACL protocols may be how to best manage the central neural components and cognitive factors associated with this injury.

If the CNS were to receive more consideration in ACL protocols, the opportunity for improved outcomes could be substantial because the CNS has been proven to be a very malleable system, as long as it receives the correct input. The CNS demonstrates neuroplasticity,³⁵ which means that it is capable of reorganization, based on the stimuli that it receives, whether internal or external.³⁶

This is an important consideration in ACL rehabilitation because the ACL graft, while restoring the biomechanical properties to the knee, is not fully capable of producing the same neurosensory properties of the original ACL.^37-42This is an important concept to understand because an ACL tear does indeed cause deafferentation in the ascending pathways to the brain.^37-40,42-46 This can lead to CNS reorganization and subsequent alterations in efferent output to the periphery.^37-40,42-46 Therefore, if a protocol with traditional musculoskeletal principles was used, then the mechanical function of the knee may certainly be remediated, but the neurosensory function will remain in a maladaptive state,^47-50 potentially leading to aberrant, non-protective movement strategies and a higher risk of re-injury.

The process of CNS reorganization may begin with the initial ACL injury. A peripheral musculoskeletal injury creates an inflammatory response that results in the arrival of chemical mediators such as histamine, substance P, calcitonin, and calcitonin gene-related peptide at the site of injury.⁵¹ As edema accumulates in the joint, tension is applied to the capsule, which may adversely affect proprioception from the receptors located within.⁴⁵ The interruption of consistent input from the peripheral mechanoreceptors may lead to long-term differentiation of the ascending pathways.⁵² This information is synthesized at 3 different levels of the CNS (spinal cord, brain stem, and motor cortex) to produce motor output.^53-56 Differentiation in the ascending circuitry can cause inhibition of motor neurons at the spinal cord.⁴⁵Animal research has shown that this differentiation can cause a breakdown in the cuneate nucleus of the brainstem,⁵⁷which provides sensory information from the upper body, while the gracile nucleus does the same for the lower body. These structures transfer proprioceptive input to the ventral posterior lateral nucleus in the thalamus, where it is then sent to the primary somatosensory cortex.⁵⁷ In general, the somatosensory, visual, and vestibular systems interpret afferent inputs to control movement, balance, and stability.^58,59 In a sport like soccer, where the movement tasks are dynamic and unpredictable, it is easy to see why even a slight deficit in somatosensory processing could disrupt a movement. Valeriani and colleagues^42,46 showed that somatosensory-evoked potentials were indeed altered in a cohort of ACL reconstruction (ACLR) subjects, indicating reorganization within the CNS.  Additionally, the deafferentation could not be changed by other afferent input coming from the knee or by the new ACL graft placed in the knee.^42,46The primary motor cortex has been found to have a substantial network of connectivity with the primary somatosensory cortex, which supports the theory that the motor cortex has a very strong linkage with the peripheral receptors in the joint.⁶⁰  The ligaments in the joint contain Ruffini, Pacinian, and Golgi receptors, all of which react to changes in the collagen fibers and send information regarding tension, length, speed, acceleration, position, and movement back to the CNS.^61-64 Unfortunately, the ascending pathway deafferentation can cause reorganization within the CNS, which makes the feedback provided from the periphery less effective in motor planning.

Ward and colleagues⁶⁵ have reported that reorganization within the motor cortex is the primary cause of chronic neuromuscular movement deficits in peripheral joint injuries. Researchers have used functional magnetic resonance imaging, transcranial magnetic stimulation, and electroencephalography in ACL patients to demonstrate changes in cortical activity and subsequent CNS reorganization.⁶⁵ Kapreli and colleagues⁴¹ reported that subjects with an ACL injury demonstrated higher cortical activation in the pre-supplementary motor area (pre-SMA). This is a region that is responsible for more complex motor planning.^66,67 This area becomes active before the primary motor cortex and is responsible for preparing the final movement pattern that the motor cortex executes.⁴¹ As the task becomes more complex, activity in the pre-SMA will increase.⁴¹ Additionally, they found that the posterior secondary somatosensory area and posterior inferior temporal gyrus showed increased cortical activity compared with controls.⁴¹ Visual planning is processed in the posterior inferior temporal gyrus, and so, it appears that the difficulty in processing somatosensory information due to ascending pathway deafferentation places an increased reliance on the visual system for movement planning.^68-70 This was observed while ACL-injured subjects performed a simple knee flexion-extension movement encompassing 40°, indicating the need to incorporate higher central levels of planning for a very simple movement pattern.⁴¹ Baumeister and colleagues^37,38 also showed that subjects with ACLR had higher levels of cortical activation in the areas of the brain that require attention and that process sensory input. They theorized that this occurred because of reduced efficiency of neural processing at lower levels in the CNS. Despite the higher levels of cortical activity observed, they found that subjects with an ACLR demonstrated proprioceptive testing that was deficient compared with that of controls. Heroux and Tremblay⁷¹ also demonstrated that subjects with an ACLR had increased resting motor cortex activity. They believed that this occurred as the motor cortex attempted to maintain neuromotor output to the periphery in the face of diminished afferent input.

Continue to: The reorganization that results in movement planning...

The reorganization that results in movement planning, transitioning from subcortical levels to cortical levels, is a phenomenon that researchers believe can lead to deficiencies even as the athlete has returned to sport. Grooms et al⁷²revealed in a case report that a subject with an ACLR showed higher levels of activity in the crus region of the cerebellum. This area contains corticobulbar and corticospinal tracts that transmit neural input to maintain balance and coordination.⁷³ These changes in the cerebellum, combined with increased motor cortex activity, are thought to be indicative of a global neural strategy that uses higher levels of the CNS, as opposed to subcortical processing.⁷²

The current research makes a clear and compelling argument for the importance of CNS reorganization after an ACL injury, placing increased reliance on higher cortical levels of control, as well as the visual system to coordinate balance and movement. It is thought that this reorganized method of neural transmission can then become imprinted within the CNS, if not corrected.^35,74 If this is the case, then traditional strength programs may not be sufficient to restore these connections to their pre-injury level. If the CNS has the ability to reorganize based on the aberrant input that it receives from the periphery, then it also certainly has the potential to adapt to more specific structured input via the ascending afferent pathways.^41,45 The rehabilitation program, however, needs to be structured specifically to target the reorganized regions of the brain. There needs to be an emphasis on rehabilitating not only the peripheral neuromotor structures but also the CNS.⁷⁵

CENTRAL NERVOUS SYSTEM REHABILITATION PRINCIPLES

For a neurological rehabilitation to be successful, the interventions need to be repetitive and task-specific, involve learning, employ whole and part practice, and transition from using an internal to an external focus of control.⁷⁶Movements that are repetitive, but which lack structured learning and skill, have been shown to have no effect on inducing neuroplastic changes in the primary motor cortex.^77,78 However, using neurological rehabilitation techniques that facilitate the acquisition of new motor skills by the CNS have been shown to cause neuroplastic adaptation in the motor cortex.^79-85 This occurs because neuroplasticity is determined by experience and practice.⁷⁸ The CNS operates on cues received in the ascending tracts by mechanoreceptors in the joint. If a new movement pattern is being learned by the athlete, then this new afferent input received from the periphery will start to initiate reorganization in the higher learning centers. If this occurs with optimal repetition and precision, then a positive reorganization can take place within the CNS that results in a higher percentage of motor planning and control being filtered down to a subcortical level. Essentially, the movements become instinctive, which is crucial in athletics, where attention in higher cortical areas is frequently diverted to external aspects of the competition and not solely used to focus on movement.

This is why shifting neurological rehabilitation from an internal focus of control to an external focus of control is paramount. While using an internal focus of control is required early in rehabilitation to enable the athlete to understand the specific tasks required in a composite movement, a gradual transition to an external focus of control is necessary as the athlete begins to perform tasks that are more soccer-specific. This autonomous stage of motor learning is crucial because it transfers the burden of motor planning from higher to lower levels of the CNS and frees up the pre-SMA and primary motor cortex to handle more complex patterns.^58,86-88

Continue to: ANTERIOR CRUCIATE LIGAMENT RISK POTENTIAL IN SOCCER PLAYERS...

If a comprehensive neuromotor rehabilitation program is to be used effectively with soccer athletes, then the first priority is to define how the players should move, so that they can demonstrate protective kinematics with all soccer-specific tasks and minimize stress to the ACL. As the ideal movement pattern becomes autonomous, then it should be trained within the context of a dynamic environment; remembering that environmental changes have a large impact on the final movement pattern selected by the individual. Brophy et al⁸⁹ evaluated videos of non-contact ACL injuries in male and female soccer players and determined that 45% occurred while cutting, 25% while landing, and 16% during deceleration. These 3 patterns represent 86% of the ACL injuries observed and offer an opportunity for evaluation and treatment with specific central neuromotor rehabilitation techniques.

The foundational movement patterns for the soccer player should focus on producing leverage that minimizes stress to the ACL during the 3 primary tasks outlined above. To achieve this, it is necessary to reduce posterior ground reaction forces at the hip and knee joint during these movements. There is a high correlation between the magnitude of the posterior ground reaction force, and anterior tibial shear, and subsequent displacement.^90,91 This stress can be reduced by increasing the hip and knee flexion angles during soccer-specific movements that involve pivoting, decelerating, and landing from a jump in a unipedal stance.⁹²

This phenomenon can be explained by observing changes in the ACL elevation angle, hamstring insertion angle, and patella tendon-tibial tuberosity insertion angle. As the knee moves into flexion, the ACL takes on a more parallel orientation to the tibia, and its fibers are better able to resist elastic deformation accompanied by a posterior ground reaction force.^93,94 The quadriceps will produce less anterior translation on the tibia because the patella tendon insertion angle is reduced relative to the longitudinal axis of the tibia, and the mechanical advantage of the quadriceps is decreased.⁹⁵ Lastly, the hamstrings will be able to provide better leverage posteriorly because the resultant force trends toward a more parallel orientation to the tibial plateau, which enables the player to counter, more effectively, the posterior ground reaction force and the anterior pull directed by the quadriceps.⁹⁵

This theory is supported by the work of Li and colleagues,⁹⁶ who showed that there is an inverse relationship between knee flexion angle and ACL loading. In their study, they applied a constant quadriceps force of 200 N at 15°, 30°, and 60° angles. The anterior shear force was obviously the highest at 15° and reduced by 20% at 30° and 60% at 60°. When hamstrings co-contraction was added, there was an additional 30% reduction in anterior shear at 15° and 50% at 30° and 60°. From a more flexed position, the hamstrings can increase joint compression and reduce the anterior translation by allowing the concave medial tibial plateau to limit the anterior drawer effect and absorb the forces that occur with excessive anterior shear, internal rotation, and valgus loads.⁹⁷ As the knee flexion angle approximates 60°, the hamstring leverage is increased, and the quadriceps leverage is diminished to the point where its ability to produce anterior tibial translation is neutralized.⁹⁸ Daniel and colleagues98 referred to this as the quadriceps neutral angle.

For soccer-specific movements that are potentially injurious to the ACL, it may then be beneficial to create a default movement pattern at the knee that approximates this value. In keeping with the information presented in this paper, it will be important to have the player reproduce this angle consistently during activities that involve pivoting, decelerating, and landing from a jump within the context of match play. This will certainly require that segments located both proximal and distal to the knee are able to function within specific parameters so that a cohesive protective synergy is produced throughout the lower quarter which minimizes posterior ground reaction forces and is protective of the ACL. This is where structured neuromotor training that is able to modulate networks within the CNS may be beneficial.

Continue to: CENTRAL NERVOUS SYSTEM TREATMENT TECHNIQUES FOR THE SOCCER PLAYER...

CENTRAL NERVOUS SYSTEM TREATMENT TECHNIQUES FOR THE SOCCER PLAYER

The ultimate goal is to create a foundational movement pattern that optimizes leverage and is protective of the ACL during decelerating, pivoting, and landing in a unipedal stance from a jump. The composite segments that are necessary to achieve this include local core stability to create lumbopelvic stiffness, and global core activation to enhance posterior chain stability. This should enable the player to feel more balanced when placing the pelvis in a more posterior and inferior position while still maintaining the trunk in a position that is parallel with the tibia, as the knee is flexed to an approximate 60° angle (Figure 1). From a frontal plane perspective, the acetabulum should bisect the malleoli of the stance leg, with a neutral tibiofemoral joint alignment (Figure 1).

The neuromotor training for the composite segments of this movement can begin in the “acute postoperative phase” (Table 1).   Because the surgical repair will limit the player’s capabilities in this stage, this is a good time to break down the foundational movement pattern into its component parts and ensure that the CNS receives a high number of quality repetitions of parts. In this phase, the player may begin an isolated training for the transversus abdominis, multifidus, and pubococcygeus. This can start in a supine position using biofeedback with an isometric contraction, progress to a standing position, and incorporate deep core activation with stance-phase gait training, mini squats, and lunge variations. This phase will require an abundance of visual and verbal feedback with an internal focus of control as the player gets used to activating the deep core and quad/hip synergy during functional lower extremity movements. Even in this early phase, the player should look to minimize anterior and/or posterior pelvic tilting and maintain a stiff thoracolumbar segment that remains parallel with the tibia during all functional movements.

Table 1. Adjunct CNS Treatment Principles for ACL Reconstruction in Soccer Athletes

Abbreviations: ACL, anterior cruciate ligament; CNS, central nervous system; CIMT, constraint-induced movement therapy.

As the player moves into the “subacute postoperative phase” (Table 1).   He or she will continue to use an internal focus of control to activate the local/global core synergy with functional movements progressing from double-leg, to single-leg positions. Partial practice, instead of whole practice, is still the predominant theme of the neural training process. In this phase, the knee and hip flexion angles can increase, and the player’s trunk and pelvic position should be critiqued in a single-leg position so that the trunk remains parallel with the tibia in the sagittal plane with a slight forward hip hinge. The pelvis should remain level throughout single-leg stance to ensure adequate activation of the lateral hip stabilizers. This is the stage where the player can learn to isolate closed kinetic chain hip rotation for pivoting, and so, single-leg hip internal and external rotation drills are useful, both with and without resistance. Skill acquisition is crucial in this phase because the patterns that the CNS adopts will form the foundation for more dynamic patterns that will occur in the later soccer-specific stages. Higher levels of cortical planning are still needed in this phase. For this reason, it is important that poor quality repetitions are recognized by the player and clinician so that he or she can learn to perform them correctly, albeit still with an internal focus of control. This is also a good time to begin to employ constraint-induced movement therapy as the player is able to replicate the desired pattern with more precision. For example, by eliminating the use of the upper extremities as a source of balance, the CNS is forced to program alternate synergies such as the lumbopelvic, and foot and ankle segments to maintain the desired alignment.

The “lower quarter static stability phase” (Table 1) marks a point where it may be useful to use direct strategies that have the capability to change CNS efferent owutput from a primary reliance on the visual processing areas in the posterior-inferior temporal gyrus back to the somatosensory area. It is critical that the player is able to make this transition in cortical reorganization and control, because ACL-injured subjects have been shown to have balance scores similar to healthy controls when they are able to use their vision, but this is reduced when vision is taken away.^99-105  Their balance will diminish even further if vision is modulated during more complex landing and pivoting maneuvers.^99-105 This may certainly explain why defending is a riskier task for ACL-injured players⁸⁹ as their visual system is focused more on tracking a player than attending to precision with movement planning.

To enhance this cortical reorganization within the context of soccer-specific movements, it is useful to start from a foundational single-leg position, with the knee approximating 60° of flexion and the trunk parallel to the tibia. In the frontal plane, the pelvis should be level, the trunk vertical, and the acetabulum bisecting the malleoli of the stance leg (Figure 1).   The player may initially work on getting into this foundational position with vision either partially obstructed using stroboscopic eyewear or completely obstructed if this equipment is unavailable.^106-108 The pattern can be progressed by constraining the upper extremities to force reliance on the lumbopelvic and foot/ankle strategies for balance. Head and/or trunk turns can be added to simulate the external focus of control that is required with movement in soccer. These should progress from slow to fast and anticipated to unanticipated as the player demonstrates competence in maintaining stability at each segment within the foundational stance position. Once this is in play, a ball should be introduced into the drills. As the player maintains the foundational position with vision diminished and upper extremities constrained, they should attempt to reach for or trap a ball from this position. If vision is completely obstructed, then the player can be instructed to open his/her eyes just as the ball arrives to induce a reactive response. Again, quality repetitions are essential for learning to occur, and subsequent skill acquisition to take place in the CNS; thus, close scrutiny should be paid to the qualitative essence of the movement patterns to ensure pristine biomechanics during this phase.

The “lower quarter dynamic stability phase” (Table 1) should continue with the same neuromotor training principles employed in the previous phase, except that the drills will now involve plyometrics. The player should ultimately progress from double-leg, to single-leg jumps and then linear to diagonal. Vision should still be obstructed and upper extremities constrained to channel the lumbopelvic region for force production and balance. Movement quality in the foundational position remains paramount with these drills to ensure that skill acquisition is occurring and injury risk is being mitigated. An external focus of control can be introduced by applying an unanticipated perturbation during a jump. Additional learning opportunities should include unanticipated head and trunk turns while landing in a unipedal stance from a jump. The task can be made more specific by having the player trap a pass while doing linear or diagonal single-leg hop progressions. In this manner, the player’s CNS can become reorganized to program the requisite synergies to maintain a protective foundational position on the stance leg, as the contralateral limb is required to perform work that is far outside the player’s base of support.

Continue to: The final segment of the CNS neuromotor rehabilitation program...

The final segment of the CNS neuromotor rehabilitation program is the “lower quarter dynamic agility phase” (Table 1), when the player will learn to perform an unanticipated directional change in a foundational position for the pivot leg. The player can begin this phase by initiating sprint-deceleration-pivot efforts, progressing at 45°, 90°, 120°, and 180° turns. This should be trained in both a forward and backpedal position. Close attention should be paid to the deceleration phase of the sprint-pivot effort, as this will set the player up to demonstrate protective kinematics during the pivot phase of the task. In this phase, the center of mass should become lower and move posteriorly, so that a deeper knee and hip flexion angle, supported by posterior chain synergies, can occur at the pivot point. This is an important skill for the player to acquire, as Cortes and colleagues⁹² have reported that female collegiate soccer players tend to perform a pivoting task with a more erect trunk position. In the same cohort, they also measured the mean knee flexion angle at initial contact during pivoting to be 24°.⁹² Movement patterns that reflect an elevated center of mass, with arms abducted away from the trunk, should be discouraged here. The drills can be progressed to have the player react to a command and perform unanticipated pivots within a 5 × 5-meter box to simulate defending. This should be progressed from eyes open and arms unconstrained, to vision disrupted and arms constrained. From here, an external focus of control can be added by playing a ball to the athlete. Vision should be withheld until the instant that the ball arrives at the player, when he/she is required to play the ball to an unanticipated spot. As is the case in all other phases of the neuromotor training, the quality of movement is the most important parameter to critique with each drill. From a qualitative standpoint, the player should demonstrate stiffness throughout the thoracolumbar region and power and control through the pelvis with each directional change. In addition, he or she should maintain a low and posteriorly oriented center of mass to optimize leverage in the hamstrings/gluteals compared with the quadriceps and reduce posterior ground reaction forces.

PSYCHOLOGICAL READINESS FOR RETURN TO PLAY

After an injury is sustained, an athlete is often subjected to a range of psychological responses in addition to the functional impairment, including stress, hesitancy, alterations in self-esteem, depression, fear of re-injury, and anxiety.^43,109-111 The aforementioned responses are often at their height in the time immediately following an injury and generally subside over time during the rehabilitation process.¹¹⁰ The rates at which athletes experience psychological distress following an injury range between 5% and 19%; the levels are comparable with patients receiving treatment for mental health illness.⁴³ However, these elements may persist, or even increase, in the later stages of the rehabilitation process as the topic of return to play is deliberated.^112,113 If these fears are left unresolved, then a significant delay can be incurred during the rehabilitation process, which might ultimately jeopardize the successful return to play.^114,115

When athletes have been cleared to return to sport, fear tends to be the most common reason for their decision to not return to play.^21,116 The persistence of fear has clinical implications and warrants close monitoring to ensure that the athlete feels adequately supported in the decision to return to sport.^117,118 Building the athlete’s confidence by addressing hesitation, lack of confidence, heightened awareness of risk or re-injury, and safe reintegration into athletic participation are important themes identified to encourage a safe return to play.⁴³ A variety of validated tools  can be integrated into an existing return-to-play decision-making algorithm (Table 2).^118-120 

Abbreviations: ACL, anterior cruciate ligament; CNS, central nervous system; CIMT, constraint-induced movement therapy.

Table 2. Self-Report Measurement Tools to Integrate into Return-to-Play Decision-Making Algorithm

By integrating the necessary screening of patients for kinesiophobia and assessing patient expectations after enduring an ACL injury, clinicians may be able to identify patients who are at risk for poorer functional outcomes. A consideration of psychosocial elements such as activity avoidance, fear of movement and re-injury, loss of confidence and expectations/assumptions during the continuum of the rehabilitation process, and the decision to return to play may favorably impact the individual’s ability to safely return to sport. It is critical to address both the physical and psychosocial factors during the rehabilitation process to more optimally transition individuals back to their prior level of athleticism.

CONCLUSION

Psychosocial factors may play a role in determining a player’s readiness to return to sport, as well as a potential for re-injury. A number of tests are available for use with this patient population to identify mental deficits that may impact player performance upon return. Additionally, the CNS should be considered as a source of impairment in players with ACL injuries. Current protocols may not fully appreciate the CNS’ impact on the player’s functional outcome. Therefore, an approach that includes CNS neuromotor training with traditional musculoskeletal rehabilitation, which also incorporates cognitive and psychosocial factors, may define an improved paradigm for treating soccer athletes following an ACL injury and assessing return-to-play capability.