|Home | About | Journals | Submit | Contact Us | Français|
The aims of this study are (i) to report on the rates of subsequent surgery following hip arthroscopy and (ii) to identify prognostic variables associated with revision surgery, survival rates and complication rates. The Statewide Planning and Research Cooperative System database, a census of hospital admissions and ambulatory surgery in New York State, was used to identify cases of primary hip arthroscopy. Demographic information and rates of subsequent revision hip arthroscopy or arthroplasty were collected. The risks were modeled with use of age, sex, procedure and surgeon volume as risk factors. Survival analyses were also performed, and 30-day complication was recorded. We identified 8267 procedures in 7836 patients from 1998 to 2012. Revision surgery occurred in 1087 cases (13.2%) at a mean of 1.7±1.6 (mean±SD) years. Revision arthroscopy accounted for 311 cases (3.8%), and arthroplasty for 796 (9.7%) cases. Survival analysis showed a 2-year survival rate of 88.1%, 5-year of 80.7% and 10-year of 74.9%. Regression analysis revealed that age>50years [hazard ratio (HR) 2.09; confidence interval (CI) 1.82–2.39, P<0.01] and a diagnosis of osteoarthritis (HR 2.72; CI 2.21–3.34, P<0.01) were associated with increased risk of re-operation. Labral repair was associated with a lower risk of re-operation (HR 0.71; CI 0.54–0.93, P=0.01). Finally, higher surgeon volume (>164cases/year) resulted in a lower risk of re-operation versus lower volume (<102cases/year) (HR 0.42; CI 0.32–0.54, P<0.01). The 30-day complication rate was 0.2%. Older age and pre-existing osteoarthritis increased the likelihood of re-operation following hip arthroscopy, whereas performing a labral repair and having the procedure performed by a higher-volume surgeon lowered the risk of re-operation.
Hip arthroscopy utilization has significantly increased over the past decade, with annual rates increasing as much as 25-fold over that time period [1–4]. Despite improvements in equipment and training, it remains a challenging procedure, with some studies reporting revision surgery rates of anywhere from 6.3%  to 16.9% [6–8]. However, many of the available studies are limited to single-institution designs with limited sample size and follow-up duration. Presently, there is limited available information on the long-term need for subsequent surgery from a population standpoint.
Similarly, smaller, retrospective studies have identified factors associated with the need for revision surgery. Revision hip arthroscopy has most commonly been associated with residual impingement, while conversion to hip arthroplasty has been associated with advanced patient age, prolonged pre-operative symptoms, acetabular dysplasia, and osteoarthritis (OA) (Tönnis grade 2 or 3) [9–15], although these results have not all been corroborated in larger sample studies Additionally, while studies have demonstrated that complication rates are closely associated with surgeon volume [16–19], the impact of surgeon volume on re-operation rates has not been previously explored.
The purpose of this study was therefore 2-fold: (i) to report on population-based rates of subsequent surgery following hip arthroscopy, including revision hip arthroscopy and conversion to arthroplasty and (ii) to identify prognostic variables associated with the need for revision surgery, including the impact of procedural volume, and to identify survival rates and complication rates associated with hip arthroscopy. Our hypothesis was that revision arthroscopy and arthroplasty would occur at a higher rate in older patients and those treated by lower-volume surgeons.
The Statewide Planning and Research Cooperative System database, a census of all non-federal acute care hospital admissions and ambulatory surgery in New York State, was used to identify cases of outpatient primary hip arthroscopy between 1998 and 2012. Access to the database was granted via request (Request #1107-04) from the Data Governance Committee of the New York State Department of Health. Patients were identified by use of Current Procedural Terminology (CPT) codes specific for hip arthroscopy (Table I).
Demographic information was collected for these patients including patient age, sex, race/ethnicity, insurance status, discharge disposition, and year of surgery. International Classification of Diseases, Ninth Edition (ICD-9) diagnostic codes were also used to identify a diagnosis of OA (codes 715.15, 715.35 and 715.95). After case identification, unique patient identifiers were used to track patients for repeat surgical intervention. Revision ipsilateral hip arthroscopy was identified using the same CPT codes, while total hip arthroplasty (THA) and resurfacing were identified with procedure-specific ICD-9-clinical modification (CM) codes (Table II). ICD-9 diagnostic codes were also used to determine whether patients carried a diagnosis of hip OA (codes 715.15, 715.35 and 715.95). Unique patient identifiers were then used to track patients for 30-day complications requiring re-admission (Table III). The list of procedural complications was derived from a more extensive list of ‘Hospital Acquired Conditions’ on the Center for Medicare & Medicaid Services website (https://www.cms.gov/medicare/medicare-fee-for-service-payment/hospitalacqcond/hospital-acquired_conditions.html).
Descriptive statistics were calculated as means with standard deviations, while categorical variables were presented as frequency counts and percentages. Kaplan–Meier survival analysis, with end points of revision hip arthroscopy or conversion to THA, was performed and revision hazards were measured at two, five and ten years following the index hip arthroscopy procedure.
Surgeon volume was analysed by looking at different volume strata. These were determined by creating a stratum-specific likelihood ratio (SSLR) threshold-analysis model, looking at the endpoint of revision hip arthroscopy, hip resurfacing or THA within 5years of the index hip arthroscopy procedure. The methodology for creating this model was adapted from a previous study . To summarize, physician license numbers were utilized to determine career surgeon volume. The cohort was first partitioned into as many volume strata as possible based on career volume. Subsequently, a risk ratio was calculated for each group for the endpoint of clinical failure (revision scope, resurfacing or THA). These ratios were then reviewed to identify specific points were there was a difference in rates of endpoints. Utilizing these points, the individual strata were then combined to produce larger groups until a significant difference in risk ratios was observed between adjacent groups with a P < 0.05. If multiple thresholds were found in close proximity to each other, we selected the threshold that had the largest risk ratio differences.
A Cox regression analyses was performed to determine the effect of age , sex , race , insurance status , diagnosis of OA  and surgeon volume as potential risk factors. If a patient did not experience additional surgery, he/she was censored at the end of study timeframe (31 December 2012), and the time to additional hip surgery was calculated from the date of index hip scope to the time of revision. Results were reported as hazard ratios with 95% confidence intervals (CI).
All analyses were performed using the SAS System 9.3 (Cary, NC, USA). SSLR analyses were performed in Microsoft Excel 2010 (Microsoft, Redmond, WA, USA).
We identified 8267 hip arthroscopy cases in 7836 patients from 1998 to 2012. This included 23 simultaneous bilateral procedures and 408 staged bilateral procedures. Procedures were performed by 295 surgeons in 137 different surgical centers (Fig. 1). Males represented 46.1% of patients, with 80.1% carrying private insurance (Table IV). Annual hip arthroscopy rates increased 88-fold over the observation period, with a 750% increase over the last 10years (Fig. 2).
Revision surgery (scope or arthroplasty) was required in 1087 cases (13.2%) at a mean of 1.7±1.6years, with 1028 (12.4%) of the revisions occurring within 5years. More specifically, revision hip arthroscopy was required in 311 cases (3.8%) at a mean of 1.8±1.6years after the index procedure, while conversion to resurfacing or THA was required in 796 (9.6%) cases at an average of 1.7±1.7years. Twenty patients (0.2%) had both a revision hip arthroscopy and conversion to hip arthroplasty. A Kaplan–Meier analysis was performed to determine the survivorship of primary hip arthroscopy (Fig. 3). From this, we identified a 2-year survival rate of 88.1%, 5-year of 80.7% and 10-year of 74.9% (Table V). Exclusively reporting on endpoint of conversion to arthroplasty, the survival rate was 91.1% at 2years, 85.6% at 5years and 80.7% at 10years (Table V).
Cox proportional hazard analysis revealed that age>50years was associated with a significantly increased risk of re-operation (HR 2.11; CI 1.84–2.41, P<0.01). Additionally, patients diagnosed with OA had a higher risk of re-operation compared with those without (HR 2.72; CI 2.21–3.34, P<0.01). Patients who had either an isolated diagnostic scope, or a labral repair, had a lower risk of subsequent surgery (HR 0.42; CI 0.23–0.79, P=0.01 and HR 0.71; CI 0.54–0.93, P=0.01). Patients that underwent a removal of loose body had a higher risk of subsequent surgery (HR 1.61; CI 1.25–2.06, P<0.01) (Table VI).
Finally, surgical volume was also associated with an increased risk of re-operation. Patients of surgeons with annual case volumes of higher than 164cases/year had a lower risk of re-operation compared with those operated on by surgeons with a case volume of <102cases/year (HR 0.41; CI 0.32–0.54, P<0.01) (Table VI). The cut-points from the SSLR analysis were used to create 5-year survival curves with four volume stratas (Fig. 4). The survival rate at 2, 5 and 10years was also determined for each volume strata (Table VII).
Complications included both medical (myocardial infarction, ileus, pneumonia, sepsis) and surgical (mechanical complication, hardware failure, DVT/PE, wound infection, dislocation/iatrogenic instability, major bleed). The 30-day procedural complication rate, excluding revision surgery, was 0.2%. The 30-day all-cause re-admission rate was 0.7%.
This study reports on the survival rates following primary hip arthroscopy in a population-based cohort. From 1998 to 2012, we identified a rapid and consistent increase in hip arthroscopy utilization, with a cumulative volume of 8244 primary hip arthroscopy cases. Subsequent surgery was required in 13.4% of patients. More specifically, revision hip arthroscopy was required following 311 cases (3.8%) and conversion to hip arthroplasty was required following 796 cases (9.6%). Older age (>50years old) and a pre-operative diagnosis of OA were associated with a higher risk of failure, while performing a labral repair and having the procedure performed by a higher volume surgeon were associated with lower risk of failure.
While the trends of increased utilization of hip arthroscopy identified with this study are not unique [2–4, 22], this study provides population-level information on the survival of primary hip arthroscopy. Our results are similar to those recently reported in the literature. Sing et al.  performed a review of a private payer health insurance population to assess trends in arthroscopic volume, as well as conversion rates. A total of 8227 cases were identified, from which the authors noted a similar overall conversion rate to THA of 8.7% at 24-month follow-up. Malviya et al.  conducted a similar review, examining all hip arthroscopy cases identified from an administrative hospital database for National Health Service patients in England between 2005 and 2013. From 6395 cases of inpatient hip arthroscopy, they reported comparable revision arthroscopy rates of 4.5% at a mean of 1.7years and conversion rates to hip resurfacing or THA of 10.6% at a mean of 1.4years. Gupta et al.  reported on a series of 595 patients treated with hip arthroscopy by a single surgeon at a high-volume surgical center. They, too, noted similar rates of revision hip arthroscopy (7.7%) and conversion to hip arthroplasty (9.2%), within 30months of their index procedure, although the mean duration until failure was not reported for these cases. Finally, Schairer et al.  reported on utilization trends and arthroplasty conversion rates following hip arthroscopy utilizing California and Florida statewide databases. They identified a similar increase in utilization, with a conversion rate of 11.7% within 2years of primary hip arthroscopy.
Similar to Malviya et al , we performed a survivorship analysis following primary hip arthroscopy finding a cumulate survival rate of 82.6% at 8years. In our study, survival analysis revealed similar rates of 85.6% and 80.7% of patients that did not require THA at the 5- and 10-year follow-up. Additionally, we added revision arthroscopy to our survival analysis, which reduced the overall survival (i.e. no further revision surgery) to 80.7% and 74.9% at 5 and 10years, respectively.
Beyond looking at rates of repeat surgery, further analysis was performed to identify prognostic variables associated with the need for revision hip arthroscopy or conversion to arthroplasty. Consistent with the findings presented here, these previously referenced studies also identified that increasing age at the time of surgery was associated with a greater risk for conversion to THA [7, 21, 23]. Malviya et al.  found age>50years was associated with a 4.7 times higher risk of conversion compared with patients younger than 50years, whereas Sing et al.  found a 17% conversion rate in patients over the age of 50years. This is likely attributable to more severe chondral damage or advanced joint degeneration in these patients, which several studies have demonstrated as a risk factor for conversion to THA following hip arthroscopy [10–12, 21, 24]. The age of 50years was chosen as a cut-off point for our regression analysis based on prior studies indicating that arthroscopic results in patients above the age of 50years were inferior to younger patients, with higher rates of conversion to hip arthroplasty [12, 24, 25].
Interestingly, female sex was not associated with an increased risk of revision surgery. This is in contrast to the results reported by Malviya et al.  and Gupta et al. , where they identified that female sex was associated with an increased risk of revision surgery. This discrepancy is likely multifactorial and is difficult to explain, particularly as all studies have limited information regarding the underlying diagnosis and reason for surgery. Additionally, surgical indications may have also differed based on regional practices, which could contribute to this discrepancy.
A pre-operative diagnosis of hip OA was also associated with a higher risk of subsequent surgery. This is consistent with the findings of Schairer et al. , who identified that a diagnosis of OA was associated with a hazard ratio of 2.3 for requiring subsequent conversion to THA. This was also corroborated in a two recent systematic reviews, where patients with a Tönnis grade of 1 or greater, or less than two mm of available joint space, where are higher risk of poor outcomes and subsequent conversion to hip arthroplasty [10, 26].
Labral repair was also found to be associated with a lower risk of requiring subsequent surgery compared with patients that did not have a labral repair. While this has not been reported in prior studies, it is consistent with prior reports that labral repair resulted in improved clinical outcomes compared with selective labral debridement [27–29]. It may relate to the fact that the labral tissue appeared healthy or was preserved, indicating the procedure was performed early in the disease process limiting further intra-articular damage, contributing to the lower revision risk. Similarly, patients that underwent a diagnostic scope were found to have lower risk of conversion versus those who did not. This cohort may reflect those without significant intra-articular pathology that did not require osteochondroplasty or labral repair, potentially explaining this discrepancy, although this is largely speculative. Conversely, patients that underwent removal of loose bodies were found to have higher rates of revision arthroscopy or conversion to arthroplasty. This may be reflective of a cohort with more extensive intra-articular damage and chondral delamination producing loose bodies, explaining the higher rate of subsequent surgery. This is supported by the studies mentioned earlier, reporting higher conversion with OA [10, 22] and additional studies reporting higher conversion rates with more advanced intra-operative chondral damage at the time of hip arthroscopy [23, 30].
Unique to our study, the SSLR threshold-analysis model and subsequent Cox proportional hazard analysis identified that individual surgeon volume correlated with the need for revision arthroscopy or conversion to arthroplasty. The need for revision surgery was higher with lower surgeon volume (<102cases/year). This is the first study to our knowledge to assess individual surgeon volume as it relates to surgical outcomes. Schairer et al.  provided the only other assessment of the effect of surgical volume on outcomes, however their results report on center volume, rather than individual surgeon volume. They reported that patients who had hip arthroscopy in lower volume surgical centers (<10cases/year), had a higher risk of requiring conversion to THA compared with both moderate- (11–49cases/year) and high-volume (>50cases/year) surgical centers. The paucity of available data on the volume–outcome relationship in hip arthroscopy highlights the importance of the present findings. To date, the only related literature has reported on the associated ‘learning curve’ of hip arthroscopy [16–18, 31]. In the majority of these studies, roughly 30 cases has been identified as the threshold beyond which the rate of procedural complications are noted to decrease [16–18]. However, these results do not define competency, but rather focus on safety as the analyses center on a reduction in associated surgical complications, rather than the effect on patient outcomes. This study provides the first attempt to quantify volume-based competency, although further study and validation of statistical methodology is required.
Finally, the complication rates identified in our study were found to be comparable to those reported by Malviya et al. . They reported a 30-day re-admission rate of 0.5%, comparable to our rate of 0.7%. Further comparisons to studies are limited because of the restrictions of using an administrative database for event identification. Complications reported in our study reflect only those events requiring re-admission, which have historically been considered as ‘major complications’ in other studies. These primarily include deep infection, deep vein thrombosis, pulmonary embolism, vascular injury, fracture, dislocation and death. Similar rates of major complications have previously been reported in systematic reviews by both Harris et al. (0.58%)  and Kowalczuk et al. (0.3%) .
Limitations of this study include that it was based on data obtained from administrative databases tracking outpatient surgery. Limited data was available for each procedure, limiting the generalizability of these results. Conclusions derived from CPT codes alone may be prone to errors due to miscoding or incorrect assumptions, and should be interpreted with caution. Additionally, complication data primarily captured more severe complications requiring re-admission, rather than the more common complications (lateral femoral cutaneous nerve palsy, heterotopic ossification, superficial wound infection) that are often dealt with in an outpatient setting. As a result, the true procedural complication rate may be underreported, although the rates provided are consistent with other study reports of ‘major’ complications. Similarly, patients having revision or subsequent procedures out of state are also not captured and may not be adequately reflected in the results. Finally, while laterality is available for outpatient procedures, including primary and revision hip arthroscopy, it is not routinely listed for the THA or resurfacing procedures. Based on expert opinion (B.T.K.), it was felt that it would be an exceedingly uncommon clinical scenario where a patient receives an ipsilateral hip arthroscopy and contralateral hip arthroplasty within 2years. As a result, assumptions were made that these procedures were performed on the same side as the primary hip arthroscopy. As a result, the rate of conversion may be falsely elevated by assuming a worst-case scenario for each case; however, the conversion rates appear consisted with figures published in comparable studies.
In conclusion, following hip arthroscopy, Kaplan–Meier survival analysis showed a 2-year survival rate of 88.1%, 5-year of 80.7% and 10-year of 74.9%. Age greater than 50, years and a diagnosis of OA increased risk of re-operation, while performing a labral repair and having the procedure performed by a higher volume surgeon was associated with a lower risk of re-operation.
ICMJE criteria for authorship
R.M.D., T.J.P., B.C., N.M., P.D.C., D.H.N., S.L.: substantial contributions to the conception or design of the work; or the acquisition, analysis, or interpretation of data for the work; R.M.D., T.J.P., B.C., N.M., P.D.C., A.S.R., B.T.K., D.H.N., S.L.: drafting the work or revising it critically for important intellectual content; final approval of the version to be published; and agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
R.M.D., T.J.P., B.C., N.M., P.D.C., A.S.R., D.H.N. and S.L. do not have any relevant disclosures. B.T.K. does not have disclosures relevant to the current study, but does report personal fees from Arthrex and A-3 surgical.
Research reported in this publication was supported by the National Institute of Arthritis and Musculoskeletal and Skin Diseases of the National Institutes of Health under Award Number R01AR066069. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Institutional review board approval was not required for the current study as this is publicly accessible data. A copy of the approved data request form will be uploaded.