Is the Actual Failure Rate of Hip Arthroscopy Higher Than Most Published Series? An Analysis of a Private Insurance Database

Abstract

Background:

The use of hip arthroscopy (HA) for the management of intra-articular hip pathology has increased greatly, with a 600% increase in utilization from 2006-2010. Studies have demonstrated good to excellent outcomes in patients undergoing hip arthroscopy for treatment of femoroacetabular impingement (FAI) syndrome. However, some patients undergoing primary hip arthroscopy will require revision hip arthroscopy (revision HA) or conversion to total hip arthroplasty (THA). The purpose of the present study was to evaluate the association between hip arthroscopy failure and (1) osteoarthritis, (2) age > 40 years, and (3) psychiatric comorbidities.

Methods:

The Humana Inc. insurance claims database was used to identify patients undergoing hip arthroscopy between 2007 and 2015, with query by CPT (current procedural terminology code) of more than 25 million deidentified insurance and Medicare beneficiary claims. Following primary hip arthroscopy, patients were longitudinally tracked for subsequent ipsilateral hip arthroscopy (revision HA) or total hip arthroplasty (THA) with a minimum of 1-year clinical follow-up from the primary HA procedure. Hip arthroscopy failure (HA failure) was defined specifically as patients who underwent a revision HA or THA with a minimum of 1-year of clinical follow-up from the primary HA procedure. Variables assessed included presence of pre-existing osteoarthritis, age < 40 years or age > 40 years, and presence of preoperatively diagnosed psychiatric comorbidities including depression or anxiety. The relationships between revision HA, THA, or HA failure and these variables were assessed utilizing univariate and multiple logistic regression analysis. Independent predictors of revision ipsilateral hip arthroscopy and subsequent hip arthroplasty were identified using multiple logistic regression.

Results:

In total, 785 patients (64.1% female) underwent primary hip arthroscopy. The overall failure rate with a minimum of 1-year clinical follow-up from the index HA procedure was 18%[140/785; 8% (63/785) revision hip arthroscopy, 10% (82/785) THA]. Multivariable logistic regression analysis identified psychiatric comorbidities (Odds Ratio [OR] 2.8, 95% Confidence Interval [CI] 1.2-6.2, p<0.01) as the only independent predictor of hip arthroscopy failure (revision HA or THA). Independent predictors of revision HA included both psychiatric comorbidity (OR 2.8, 95% CI 1.2-6.2, p<0.01) and age < 40 years (OR 2.6, 95% CI 1.4-5.0, p<0.01), while age > 40 years (OR 3.09, 1.47-7.25, p<0.005), smoking (OR 2.05, 95% CI, 1.68-1.88, p=0.02), and osteoarthritis (OR 3.24, 95% CI 1.98-5.43, p<0.001) predicted conversion to THA.

Conclusion:

The hip arthroscopy failure rate of 18% in the present study is alarmingly high, a figure much higher than reported in previously published series. Patient factors associated with conversion to THA included age > 40 years, smoking, and preexisting osteoarthritis. The presence of psychiatric comorbidities, specifically depression and anxiety, was independently associated with revision HA and overall HA failure.

Level of Evidence: III

Introduction

The use of hip arthroscopy (HA) for the management of intra-articular hip pathology has increased greatly, with more than a 600% increase in utilization from 2006-2010.^1-3 Studies have demonstrated good to excellent outcomes in patients undergoing hip arthroscopy for treatment of femoroacetabular impingement (FAI) syndrome with appropriate bony offset correction and treatment of soft tissue.^2-4 However, a subset of patients undergoing primary hip arthroscopy will eventually require revision hip arthroscopy (revision HA) or conversion to total hip arthroplasty (THA), with reported rates in the literature ranging from 1.2%-8% and 1.7%-5.6%, respectively.^3-6 Further investigation into the causes of revision HA and conversion to THA is therefore of significant clinical interest, with persistent deformity, recurrent or persistent labral tears, chondral lesions, and the presence of osteoarthritis serving as potential factors leading to revision or poor clinical outcomes.⁷

In addition to surgical and technical factors, several patient factors, including preoperative opioid use, smoking, obesity, and psychiatric comorbidities, specifically the presence of depression or anxiety, have been postulated to affect clinical outcomes and increase the risk of subsequent revision HA or conversion to THA.^8-13 Prior literature has suggested that pre-existing hip osteoarthritis^14,15 and age^16,17 may influence success rates following hip arthroscopy, while the association between pre-existing medical comorbidities and hip arthroscopy failure, especially psychiatric comorbidities,^18,19 remain poorly elucidated in the literature.

While intra-articular technical factors leading to failure after HA have been previously reported, there remains a paucity of data on patient demographic variables and psychiatric comorbidities that may significantly influence outcomes. The purpose of the present study was to evaluate the association between osteoarthritis, age > 40 years, and the association between psychiatric disorders, specifically anxiety and depression, with failure after index HA procedure. HA failure was defined as subsequent ipsilateral revision HA or conversion to THA with a minimum of 1-year of clinical follow-up from the index HA procedure using a large database of de-identified Humana and Medicare recipients. We hypothesized that age > 40 years and pre-existing osteoarthritis would be associated with increased conversion to THA following hip arthroscopy. Additionally, it was hypothesized that there would be a positive association between psychiatric diagnoses and failure rates after primary HA, with increased revision HA or THA rates in patients with psychiatric comorbidities diagnosed preoperatively prior to the index procedure.

Methods

Data Source

The Humana Inc., administrative claims dataset was accessed using the PearlDiver Research Program (PearlDiver Inc, Fort Wayne, IN).²⁰ The Humana Inc. administrative claims dataset contains over 25 million de-identified, HIPAA (Health Insurance Portability and Accountability Act) compliant medical billing records from patients across the United States, including commercially insured patients and Medicare advantage beneficiaries. The dataset permits longitudinal tracking of patients, procedures, and medical diagnoses. All data within this database are Health Insurance Portability and Accountability Act compliant and were therefore deemed exempt from Institutional Review Board approval.

Study Design and Cohort Definition

Patients active in the dataset between 2007 and 2015 that underwent primary hip arthroscopy procedures were identified using Common Procedural Terminology (CPT) codes (Appendix 1). Only patients enrolled with the insurance provider for a minimum of three months prior to and one year following their index hip arthroscopy were included. Laterality modifiers were utilized to identify patients who underwent left or right hip arthroscopy. Patients who had bilateral hip arthroscopies during the study period were excluded to prevent multiple procedures from confounding our analysis of preoperative opioid prescriptions. Patients were tracked for the presence of an ipsilateral hip arthroscopy and/or ipsilateral conversion to total hip arthroplasty. Patient demographics and preoperative comorbidities were identified for the entire cohort using ICD-9 codes. Variables recorded included age, sex, presence of a diagnosis (Appendix 2) of anxiety and/or depression, obesity, smoking status, diabetes, pre-existing hip osteoarthritis and preoperative opioid use defined as opioid prescription filling within three months prior to the index hip arthroscopy procedure. Patients were tracked over time for the occurrence of ipsilateral revision HA or ipsilateral THA using CPT code and laterality modifiers to ensure that revision HA or THA was performed on the same hip as the previous hip arthroscopy. Subgroups analysis based on age < 40 years and age > 40 years at the time of primary hip arthroscopy were employed to investigate the association between age group and HA failure. This categorical distinction in age (40 years) was determined based on prior literature.^16,21,22 Basic demographic data (age and sex) of the entire cohort was also identified. During the study period, 2,567 patients underwent primary HA and were available for analysis. Subsequently, filters for active patient follow-up were applied, which yielded 1,436 active patients who had a minimum of three months preoperative and 1-year postoperative follow-up. After the application of laterality modifiers to the hip arthroscopy CPT codes, 785 patients (30.58%, 785/2567) were remaining for the longitudinal analysis of subsequent revision HA, ipsilateral THA, and overall HA failure.

Outcomes

The outcomes of interest included ipsilateral revision HA and conversion to THA. Overall HA failure was defined as a revision HA and/or conversion to THA with a minimum of 1-year of clinical follow-up from the index HA procedure. The impact of each demographic and comorbidity variable on the incidence of the three outcomes of interest was also evaluated.

Statistical Analysis

Univariate analysis was performed to determine the association between patient variables and HA failure. Multivariate logistic regression to account for multiple patient-variables was performed to determine Odds Ratios (OR) and the corresponding 95% Confidence Intervals (CI) for the association of each variable on revision HA, THA, and overall HA failure. Statistical analysis was performed utilizing the R© statistical package in the PearlDiver research program (www.pearldiverinc.com; PearlDiver Inc, Fort Wayne, IN). Statistical significance was defined as p<0.05.

Results

Female (503, 64.1% female) and male (282, 35.9%) patients underwent index HA. The top three largest age groups that underwent primary hip scope were ages 40-44 (11.3%), 45-49 (11.3%), and 50-54 (11.04%) years. Patients most commonly underwent labral resection (CPT code 29862). The overall failure rate was 18% with 1-year minimum of clinical follow-up from the index HA procedure, with revision HA (8%, 63/785) and THA (10%, 82/785) represented. Table 1 depicts the demographics of patients undergoing HA in the present study.

Table 1.

Demographics of Hip Arthroscopy Patients

Demographic Factor	Primary HA	Revision HA	THA	HA Failure
Male	282 (36%)	18 (28.6%)	27	43 (30.7%)
Female	503 (64.1%)	45 (71.4%)	55	97 (69.3%)
Age < 40	265 (33.8%)	33 (52.4%)	*	41 (29.3%)
Age > 40	520 (66.2%)	30 (7.6%)	*	99 (70.7%)
Total	785	63	82	140
% of Total	100%	8%	10%	18%

^*Due to limitations from Pearl Diver, with < 11 patients, value could not be calculated.

Risk factors for revision HA, THA, or overall HA failure were assessed (Table 2). In total, 82 patients underwent subsequent conversion to THA (10%, 82/785). Risk factors for conversion to THA included smoking (OR 2.61, 95% CI 1.46-4.52, p<0.01), diabetes (OR 2.04, 95% CI 1.09-3.65, p=0.02), osteoarthritis (OR 3.24, 95% CI 1.98-5.43, p<0.01), psychiatric comorbidities (OR 2.09, 0.98-4.09, p=0.04), and age > 40 years (OR 4.74, 2.4-11.1, p<0.01). In the present study, 18% of patients (140/785) required either ipsilateral revision HA or ipsilateral THA, or both procedures (HA failure). Risk factors for HA failure included age < 40 years (OR 1.33, 95% CI 0.87-2.04, p=0.19), psychiatric comorbidities (OR 2.72, 95% CI 1.50-4.79, p<0.01), smoking (OR 1.94, 95% CI 1.16-3.15, p<0.01), and osteoarthritis (OR 1.59, 95% CI 1.08-2.34, p=0.02).

Table 2.

Risk Factors for HA Failure

Variable	Odds Ratio (OR), (95% Confidence Interval, p-value)
	Revision HA	THA	HA Failure
Preoperative Narcotic Use	0.89 (0.49-1.57,0.69)	1.54, (0.95-2.48,0.07)	1.27 (0.86-1.87,0.22)
Smoking	1.2 (0.51-2.5,0.65)	2.61, (1.46-4.52,<0.01)	1.94 (1.16-3.16,<0.01)
Diabetes	0.74 (0.25-1.74,0.53)	2.04, (1.09-3.65,0.02)	1.43 (0.82-2.41,0.19)
Obesity	1.43 (0.61-3.0,0.37)	1.32, (0.63-2.52,0.43)	1.35 (0.75-2.32,0.29)
Osteoarthritis	0.59 (0.313-1.07, 0.09)	3.24, (1.98-5.43,<0.01)	1.59 (1.08-2.34,0.02)
Psychiatric Comorbidities	2.61(1.14-5.46,<0.01)	2.09, (0.98-4.09,0.04)	2.72 (1.50-4.79,<0.01)
Age < 40	2.58 (1.46-4.59,<0.01)	4.74, (2.4-11.1,<0.01)	1.33 (0.87-2.04,0.19)
Male Sex	0.82, (0.44-1.47,0.52)	0.75, (0.45-1.24,0.28)	0.79 (0.50-1.14,0.19)

Independent predictive variables associated with failure after HA were identified, including revision HA, conversion to THA, or overall HA failure (Table 3). Psychiatric comorbidities (OR 2.81, 95% CI 1.18-6.15, p<0.01) and age < 40 (OR 2.58, 95% CI 1.46-4.59, p<0.01) were independently associated with revision HA. Independent variables associated with conversion to THA included smoking (OR 2.61, 1.46-4.52, p<0.01), osteoarthritis (OR 3.24, 1.98-5.43, p<0.01) and age > 40 years (OR 3.09, 1.47-7.25, p<0.01). Psychiatric comorbidities were (2.72, 95% CI 1.50-4.79, p<0.01) found to be independently associated with hip scope failure while controlling for other factors frequently associated with failure using the multiple logistic regression model.

Table 3.

Variables Independently Associated with HA Failure

Variable	Odds Ratio (OR), (95% Confidence Interval, p-value)
	Revision HA	THA	HA Failure
Preoperative Narcotic Use	0.921 (0.5-1.67,0.79)	1.13 (0.04-0.131,0.64)	1.01 (0.72-1.61,0.72)
Smoking	1.20 (0.49-2.67,0.67)	2.05 (1.68-1.88,0.02)	1.60 (0.93-2.70,0.08)
Diabetes	0.76 (0.23-2.10,0.63)	1.33 (1.09-3.78,0.40)	1.14 (0.61-2.1,0.67)
Obesity	1.91 (0.74-4.48.0.15)	0.731 (0.326-1.51, 0.42)	1.02 (0.54-1.87,0.95)
Osteoarthritis	0.71 (0.35-1.37,0.32)	2.42 (1.43-4.17,<0.01)	1.46 (0.96-2.22,0.08)
Psychiatric Comorbidities	2.81 (1.18-6.15,0.01)	1.42 (0.64-2.95,0.36)	2.21 (1.19-3.98,0.01)
Age < 40	2.59 (1.38-4.96,<0.01)	3.09 (1.47-7.25,<0.01)	1.03 (0.65-1.67,0.90)
Male Sex	0.76 (0.40-1.40.0.39)	0.825 (0.48-1.39,0.48)	0.79 (0.51-1.20,0.27)

Discussion

The purpose of the present study was to evaluate factors associated with revision HA, conversion to THA, and overall HA failure with a minimum of 1-year clinical follow-up from the primary hip arthroscopy procedure. In the current study, results indicated an 18% failure rate following HA with a minimum of 1-year of clinical follow-up from the index HA procedure using a large, national administrative claims dataset of 785 patients undergoing an index hip arthroscopy procedure. In total, 8% of patients underwent revision HA and 10% underwent conversion to total hip arthroplasty (THA) with 1-year minimum of clinical follow-up from the index HA procedure. Revision hip arthroscopy was associated with age < 40 years and preoperative psychiatric comorbidities. The presence of pre-existing osteoarthritis, age > 40 years, and smoking significantly increased the risk of conversion to THA. Additionally, the presence of psychiatric comorbidities at the time of the index HA, particularly depression and anxiety, was independently associated with overall hip arthroscopy failure, despite controlling for additional variables frequently associated with HA failure.

Risk factors for revision hip arthroscopy have previously been reported in the literature and include female sex, age under 40 years, absence of a labral tear, and index procedure performed by a low volume surgeon.¹⁶ While the majority of patients undergoing hip arthroscopy are younger than age 40 years, primary HA in older patients > age 40 years is still commonly performed. A recent systematic review and meta-analysis of outcomes after hip arthroscopy in femoroacetabular impingement by Minkara et al.²³ demonstrated that while patients with a pooled mean age of 29.9 ±1.9 years were undergoing hip arthroscopy across 1911 patients (1981 hips), a wide range of ages were represented, with patients undergoing hip arthroscopy into their 4^th, 5^th, and 6^th decades. Scott et al.²⁴ demonstrated that 3,320 patients > age 65 years underwent primary HA from 2005-2014, with Medicare beneficiaries > 60 years of age undergoing primary HA at a rate of 200% increase from 2007-2011.¹ The current study demonstrated that age < 40 years served as an independent predictive variable leading to increased rates of revision hip arthroscopy after index HA procedure, which has previously been reported by Kester et al.¹⁶ One explanatory possibility that we postulate for this finding is the increasing tendency to undergo revision HA for younger patients < 40 years of age in the case of primary HA failure, with conversion to THA more commonly presented as a surgical alternative for patients of advanced age (> 40 years) in the setting of primary HA failure and persistent symptoms. As the growing body of literature expands to identify factors associated with positive outcomes following primary HA with mid-term and long-term follow-up, age along with patient-reported outcomes (PROs), intraoperative, or surgical factors may explain these disparate revision HA rates and offer opportunities to guide surgical management following primary HA across a more well-defined spectrum of patient ages and risk factors.^25-28

Rates for conversion to THA after primary hip arthroscopy have been reported to range from 1.7%-5.6%, with rates higher in the setting of pre-existing osteoarthritis and age > 50 years.^6,13,16 Bogunovic et al. identified hip osteoarthritis as the second most common reason for failure after HA.²⁹ Radiographic evidence of osteoarthritis has been correlated with higher failure rates after HA for FAI, with increased conversion to total hip arthroplasty.^30,31 Larson et al. further demonstrated a 52% failure rate with >50% loss of joint space or <2mm of joint space.³² Given a rapid increase in hip arthroscopy rates in an increasingly older patient population, further characterization of OA radiographically may aid in selecting patients for operative intervention and predicting failure rates after HA. Importantly, our results suggest that the presence of OA is independently associated with subsequent conversion to THA and should be taken into consideration when indicating patients for hip arthroscopy. Patient selection has been highlighted as an important factor in surgical indications, with age > 50 years suggested by Phillipon et al. to greatly decrease hip survival outcomes after hip arthroscopy procedures.⁴ The results of the current study indicate that older patients (age > 40 years) are at an increased risk of poor outcomes following primary HA, an earlier age than typically reported. Our results further suggest that preoperative smoking may influence conversion rates to THA. Westermann et al. previously reported that smoking may influence hip outcomes postoperatively in the setting of femoroacetabular impingement¹⁸and Kamath et al. reported that smoking was a negative predictor of good or excellent outcomes following hip surgery.¹¹ Preoperative demographic and pathological risk factors including increased age, preoperative smoking, and presence of osteoarthritis should therefore be discussed as part of a shared decision-making process between the patient and the surgeon regarding whether or not to proceed with hip arthroscopy if these risk factors are present.³³

Furthermore, our results suggest that the presence of pre-existing psychiatric comorbidities may also influence rates of revision hip arthroscopy. Aside from a limited number of studies, there is a paucity of literature regarding the influence of psychiatric pathologies on hip arthroscopy outcomes.^18,19 Lansdown et al. reported that the presence of a mental health disorder was associated with lower patient-reported outcomes (PROs) before and after surgical management of FAI.³⁴ Data from the Military Health System Data Repository indicated that the severity of mental health disorders may also intensify after arthroscopic hip surgery³⁵ and may predispose these patients to subsequent HA failure. Recognizing that the presence of psychiatric comorbidities may be independently associated with HA failure is necessary in order to alert providers to the importance of adequate screening and patient selection during the shared decision-making process for operative intervention. Identification of patient risk factors, such as depression and anxiety, has been noted to improve outcomes after orthopedic procedures.³⁶ Additional resources directed to screening and increased recognition of mental health comorbidities may lead to improved surgical outcomes.^37-39

The present study has several limitations. Perhaps the largest limitation inherent in the current study is the inability to control for surgical quality. The retrospective, observational nature of the cohort may narrow generalizability. The nature of the cohort from the Humana database limited the authors’ ability to account and control for all confounders. In addition, the diagnoses in this study were subject to the accuracy of the documentation and coding practices across all participating institutions.⁴⁰ More specific delineation of groups with participants <11 could not be specified due to limitations from the PearlDiver Database to ensure compliance with HIPPA. Further, the presence of osteoarthritis may not always be clinically diagnosed and documented and the use of the ICD-9 code as a proxy for the presence of OA may have underrepresented the true proportion of OA among patients in this analysis. ICD-9 codes are unable to delineate laterality when assessing the database for revision hip scope procedure and pre-existing osteoarthritis. In addition, the present study is descriptive and inferential and did not report PROs (Patient-Reported Outcomes) or functional outcomes after primary HA. The results represented in the current work were limited to patients active within the Humana claims database at the time of the initial query, with a minimum of 1-year clinical follow-up from the index HA procedure. Patients may have undergone HA failure, revision HA, or conversion to THA with less than 1-year of clinical follow-up available in the database, introducing the possibility that the rates presented here are an underrepresentation of true HA failure rates. Further investigations with additional mid-term and long-term follow-up from the primary HA procedure may be warranted to further delineate failure rates following primary HA.

Conclusion

The hip arthroscopy failure rate of 18% in the present study is alarmingly high, a figure much higher than reported in previously published series. Patient factors associated with conversion to THA included age > 40 years, smoking, and pre-existing osteoarthritis. The presence of psychiatric comorbidities, specifically depression and anxiety, was independently associated with revision HA and overall HA failure.

Appendix 1: Hip Arthroscopy and Additional Procedures by CPT Code¹

Procedure	CPT Codes
Primary Hip Arthroscopy	29860, 29861, 29862, 29863, 29914, 29915, 29916
Revision Hip Arthroscopy (HA)	29860, 29861, 29862, 29863, 29914, 29915, 29916
Total Hip Arthroplasty (THA)	27130

¹CPT Code: Current Procedural Terminology Code.

Appendix 2: Preoperative Comorbidities for Patients Undergoing Hip Arthroscopy

Preoperative Comorbidities	ICD 9/10 Codes1
Depression	266.20-266.26,296.30-296.36
Anxiety	300.02
Obesity	27800, V853, V8530, V8539, V27801, V854, V8541, V8545
Tobacco Use	3051
Diabetes	25000-25003,25010-25013,25020-25023, 25030-25033 25040-25043, 25050-25053, 25060-25063,25070-25073,25080-25083,25090-25093
Pre-existing Hip Osteoarthritis (OA)	715.15, 715,25, 715.35, 715.95

¹ICD 9/10 Code: International Classification of Disease Codes; 9th and 10th edition codes are represented.