Hip arthroscopy trends and early postoperative outcomes: A National Surgical Quality Improvement Program study from 2014-2023

Abstract

BACKGROUND

Hip arthroscopy has become increasingly popular due to its minimally invasive nature and favorable outcomes compared to open procedures. Despite its growth, national data on surgical volume and complication trends over the past decade remain limited.

AIM

To report the case proportion trends of hip arthroscopies from 2014 to 2023 using the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database, risk of any adverse events, and patient-related risk factors that increase perioperative complications.

METHODS

A retrospective analysis was conducted using the ACS-NSQIP database from 2014 to 2023. Patients who underwent hip arthroscopy were identified via current procedural terminology codes. Demographic, clinical, and perioperative variables were extracted. Complications within 30 days postoperatively were categorized as major or minor. Descriptive statistics and regression analyses were performed to identify predictors of adverse events.

RESULTS

A total of 3349 hip arthroscopy cases were analyzed. The overall 30-day complication rate was 5.58%, with 1.88% major and 3.70% minor complications. The most common major complication was return to the operating room (0.48%), while bleeding requiring transfusion (2.12%) was the most frequent minor complication. Multivariate analysis revealed that preoperative transfusion, partial dependence in functional status, higher American Society of Anesthesiologists (ASA) class, and regional anesthesia were significantly associated with increased complication risk. Male sex was associated with decreased risk. Surgical volume showed an overall upward trend from 2014 to 2022, with a slight decline in 2023.

CONCLUSION

Hip arthroscopy utilization has fluctuated over the past decade. While the procedure remains relatively safe, identifying risk factors such as ASA class, anesthesia type, and functional status is critical for optimizing outcomes. These findings provide updated benchmarks and highlight areas for targeted improvement in perioperative care.

Key Words: Hip arthroscopy; surgical volume trends; postoperative complications; orthopedic surgery

Core Tip: Over the past decade hip arthroscopy utilization has continued to increase, though at a slower rate compared to earlier periods. The 30-day complication rate was 5.58%, but this likely reflects improved case capture rather than a true increase in procedural risk. Overall, hip arthroscopy remains a relatively safe procedure with favorable outcomes.

INTRODUCTION

Hip arthroscopy has evolved to now provide outstanding visualization of the articular surfaces of the hip joint, as well as peritrochanteric and extra-articular spaces around the hip[1]. Compared to traditional open surgical techniques for hip pathology, such as femoroacetabular impingement syndrome, hip arthroscopies have demonstrated significantly higher Nonarthritic Hip Scores and lower reoperation rates[2]. Moreover, it has demonstrated greater patient reported outcomes scores, lower rates of conversion to total hip arthroplasties, higher patient satisfaction and return to daily activities and sport[3-6]. With these benefits, it is no surprise that the popularity of hip arthroscopies has grown exponentially over the last two decades[1,7].

Recent systematic reviews and meta-analyses have demonstrated the 30-day reoperation rate following hip arthroscopy is 5.5% and the risk of surgical complications is 1.7%, with the risk of surgical complications increasing to 10% for bilateral hip arthroscopies[8,9] Postoperative complications include heterotopic ossification, transient neurapraxia, adhesions, and infection[9]. While several contemporary reviews have reported the rates of complication of hip arthroscopies, there has not been a large scale study on the topic across the United States in the past decade: The most recent large scale study documented the hip arthroscopy surgical volume trends and 30-day postoperative complications over the 2006 to 2013 period[10].

Given there has been no recent study, the purpose of this study is to report the case proportion trends of hip arthroscopies from 2014 to 2023 using the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database, risk of any adverse events, and patient-related risk factors that increase perioperative complications.

MATERIALS AND METHODS

This is a retrospective study of data collected prospectively through the ACS-NSQIP database. The NSQIP registry includes detailed demographic, comorbidity, and perioperative data, as well as 30-day postoperative outcomes such as complications, readmissions, and reoperations. Patients are identified using current procedural terminology (CPT) and International Classification of Diseases, Ninth Revision codes. Data are collected by trained surgical clinical reviewers at each participating hospital, ensuring high data fidelity through standardized protocols and regular audits. The NSQIP program includes over 400 hospitals across the United States, excluding Veterans Affairs and ambulatory surgery centers. Quality assurance measures include biweekly audits and inter-rater reliability assessments, with reported disagreement rates below 1.8%.

Patients who underwent hip arthroscopy between January 1, 2014, and December 31, 2023, were identified using CPT codes 29860, 29861, 29862, 29863, 29914, 29915, and 29916. Patient records were screened across both primary and secondary procedure fields to account for concurrent procedures and multiple CPT codes. Records were excluded from analysis if they contained missing, null, or unknown values. Additional exclusions include invalid operating times (< 0 minutes), total length of stay (< 0 days), missing weight (≤ 0 kgs), or negative values for preoperative hematocrit.

Demographic and clinical variables were extracted, including age, sex, body mass index (BMI), diabetes status, functional status, operative time, and anesthesia type. Age, BMI, and operative time were categorized for analysis. Complications within 30 days postoperatively were classified as major or minor. Major complications included deep infection, sepsis, septic shock, wound dehiscence, pulmonary embolism, unplanned intubation, acute renal failure, cardiac arrest, myocardial infarction, cerebrovascular accident, return to the operating room, and death. Minor complications included superficial infection, pneumonia, urinary tract infection, deep vein thrombosis, transfusion-requiring bleeding, and renal insufficiency. Orthopedic-specific complications were not available in the NSQIP dataset (Table 1).

Table 1 Patient demographic data and surgical variables.

	Mean	Standard deviation	%
Age (years)	43.9	17.7
< 30 years			25.5
30-60 years			53.1
> 60 years			21.4
Operative time (minutes)	102.7	65.7
> 90			47.4
< 90			52.6
ASA class	2.04	0.79
I			25.8
II			47.9
III			23.0
IV			3.3
BMI (kg/m2)	28.3	6.3
< 18.5			2.2
18.5-25			28.2
25-30			36.8
> 30			32.8
Sex
Female			55.4
Male			44.6
Race
Asian			1.9
Black			5.5
Native American			0.2
White			55.3
Unknown/unreported			37.1
Anesthesia
General			84.2
Spinal			2.1
MAC			12.7
Regional			0.7
Epidural			0.3
Smoking
No			86.0
Yes			14.0
Steroid use
No			97.6
Yes			2.4

ASA: American Society of Anesthesiologists; MAC: Monitored anesthesia care; BMI: Body mass index.

Descriptive statistics were calculated for all variables. Continuous variables were summarized as means with standard deviations, and categorical variables as frequencies with percentages. The annual frequency of hip arthroscopy procedures was calculated as a proportion of total surgical cases in the NSQIP database from 2014 to 2023. Trends in case proportions were assessed using Pearson correlation coefficients. Univariate and multivariate regression analyses were performed to evaluate associations with the outcome of any adverse event. Primary outcomes include annual hip arthroscopy case proportion. Secondary outcomes include the occurrence of any 30-day postoperative adverse event, defined as the presence of at least one adverse event listed in Table 2. Each patient was counted once in the compounded outcomes regardless of the number of adverse events experienced. All clinically relevant variables and those with univariate association of P < 0.05 were included in the multivariate model. Multicollinearity was assessed using variance inflation factors (< 5) and heteroskedasticity-robust standard errors (HC3) were implemented to account for error structure in linear probability modeling.

Table 2 Complications experienced.

Complication	n	%
Major complications
Return to operating room	16	0.48
Sepsis	13	0.39
Myocardial infarction	11	0.33
Pulmonary embolism	8	0.24
Dehiscence	6	0.18
Deep infection	5	0.15
Death	2	0.06
Cerebrovascular accident	1	0.03
Acute renal failure	1	0.03
Unplanned intubation	0	0
Cardiac arrest	0	0
Total major (including death)	63	1.88
Minor complications
Bleeding	71	2.12
Superficial infection	18	0.54
Deep venous thrombosis	18	0.54
Urinary tract infection	14	0.42
Pneumonia	2	0.06
Renal insufficiency	1	0.03
Total minor	124	3.70
Overall complications	187	5.58

Predictors of overall 30 days adverse events were analyzed using univariate and multivariate linear probability regression models. Categorical variables were evaluated relative to pre-specified reference groups (age < 30 years, female sex, independent functional status, American Society of Anesthesiologists (ASA) Class 1, general anesthesia, White race, and non-diabetic status). Sparse-data categories, including rare anesthesia types and low frequency comorbidities such as chronic dialysis, were examined for wide confidence intervals, and sensitivity analyses were performed using Firth’s penalized likelihood logistic regression. Sensitivity considerations included evaluating composite adverse events across minor and major complication categories. All statistical analyses were conducted using Python version 3.11. A significance threshold of P < 0.05 was used for all analyses.

RESULTS

A total of 3349 patients who underwent hip arthroscopy between 2014 and 2023 were included in the analysis. The mean age was 43.9 years (SD 17.7), with 53.1% aged 30-60 years, 25.5% under 30, and 21.4% over 60. The average operative time was 102.7 minutes (SD 65.7), with 47.4% of procedures lasting more than 90 minutes. The mean BMI was 28.3 kg/m² (SD 6.3), with 36.8% classified as overweight (25-30 kg/m²) and 32.8% as obese (> 30 kg/m²). Most patients were ASA class II (mean 2.04, SD 0.79). Females comprised 55.4% of the cohort. Racial distribution was predominantly White (55.3%), though 37.1% of patients had unreported or unknown race. General anesthesia was used in 84.2% of cases, followed by monitored anesthesia care (MAC) in 12.7%. Additionally, 86.0% of patients were non-smokers, and 97.6% were not on chronic steroid therapy.

Among the 3349 patients analyzed, the 30-day risk of any adverse event following hip arthroscopy was 5.58%. Major complications occurred in 1.88% of patients, with the most common being return to the operating room (0.48%), sepsis (0.39%), and myocardial infarction (0.33%). Other major complications included pulmonary embolism (0.24%), wound dehiscence (0.18%), deep infection (0.15%), acute renal failure (0.03%), and cerebrovascular accident (0.03%). There were no reported cases of unplanned intubation or cardiac arrest. Minor complications were more frequent, affecting 3.70% of patients. The most common minor complication was bleeding requiring transfusion (2.12%), followed by superficial infection (0.54%) and deep venous thrombosis (0.54%). Urinary tract infections occurred in 0.42%, pneumonia in 0.06%, and renal insufficiency in 0.03% of patients. The 30-day mortality rate was 0.06%, with two deaths reported.

Multivariate regression analysis identified several significant predictors of 30-day postoperative complications following hip arthroscopy. Age between 30 and 60 years old was associated with a lower risk of complications (β = -0.0178, 95%CI: -0.035 to -0.001, P = 0.043) while age > 60 showed a significant association in univariate analysis only. Male sex was associated with a lower risk of complications (β = -0.0165, 95%CI: -0.030 to -0.003, P = 0.021). Preoperative transfusion (β = 0.4000, 95%CI: 0.256-0.544, P < 0.001) and partial dependence in functional status (β = 0.2849, 95%CI: 0.183-0.386, P < 0.001) were strongly associated with increased risk. Higher ASA class, particularly class 3 (β = 0.0349, 95%CI: 0.011-0.058, P = 0.004) and class 4 (β = 0.1930, 95%CI: 0.149-0.237, P < 0.001), also predicted greater complication rates. Regional anesthesia was associated with increased risk (β = 0.1425, 95%CI: 0.057-0.228, P = 0.001), while MAC/IV sedation showed a modest but significant association in univariate analysis only. Other variables such as race, diabetes status, and steroid use were not significantly associated with complications in multivariate models (Table 3).

Table 3 Univariate and multivariate analysis, n (%).

	Total (n)	Events	Univariate			Multivariate
			Coefficient	95%CI	P value	Coefficient	95%CI	P value
Age (years)
< 30	853	35 (4.10)	-0.0096	-0.025 to 0.006	0.237
30-60	1778	50 (2.81)	-0.0346	-0.049 to -0.020	< 0.0001	-0.0178	-0.035 to -0.001	0.043
> 60	718	69 (9.61)	0.0606	0.044-0.077	< 0.0001	0.0002	-0.026 to 0.027	0.990
Sex (male)	1492	52 (3.49)	-0.0201	-0.034 to -0.006	0.006	-0.0165	-0.030 to -0.003	0.021
Smoking	469	20 (4.36)	-0.0039	-0.024 to 0.017	0.710
Steroid use	79	7 (8.86)	0.0437	-0.003 to 0.090	0.067
Hypertension meds	709	55 (7.76)	0.0401	0.023 to 0.057	< 0.001	0.0115	-0.008 to 0.031	0.251
Hx of CHF	13	2 (15.38)	0.1083	-0.006 to 0.222	0.063
Hx of COPD	40	5 (12.50)	0.0800	0.015 to 0.145	0.016	-0.0037	-0.068 to 0.061	0.910
Dialysis	2	1 (50.00)	0.4543	0.164 to 0.744	0.002	0.2167	-0.070 to 0.503	0.138
Preop transfusion	8	5 (62.50)	0.5804	0.436 to 0.724	< 0.001	0.4000	0.256 to 0.544	< 0.001
Functional status
Independent	3331	146 (4.38)	Ref
Partially dependent	17	8 (46.06)	0.4268	0.328-0.526	< 0.001	0.2849	0.183-0.386	< 0.001
Totally dependent	1	0 (0.00)	-0.0460	-0.457-0.365	0.826
Race
White	1844	74 (4.07)	Ref
Black	184	9 (4.89)	0.0031	-0.028 to 0.034	0.845
Asian	63	1 (1.59)	-0.0307	-0.083 to 0.022	0.249
American Indian	13	0 (0.00)	-0.0462	-0.160-0.068	0.428
Anesthesia
General	2819	116 (4.11)	Ref
Spinal	71	6 (8.45)	0.0394	-0.010 to 0.089	0.117
MAC/IV sedation	425	28 (6.59)	0.0228	0.001 to 0.044	0.036	-0.0092	-0.032 to 0.014	0.430
Regional	22	4 (18.18)	0.1367	0.049-0.224	0.002	0.1425	0.057 to 0.228	0.001
Epidural	11	0 (0.00)	-0.0461	-0.170 to 0.078	0.466
ASA class
1	863	21 (2.43)	Ref
2	1604	50 (3.12)	-0.0284	-0.043 to -0.014	< 0.001	0.0016	-0.015 to 0.019	0.856
3	771	54 (7.00)	0.0312	0.014 to 0.048	< 0.001	0.0349	0.011 to 0.058	0.004
4	109	28 (25.69)	0.2180	0.179 to 0.257	< 0.001	0.1930	0.149-0.237	< 0.001
Diabetes
No	3137	134 (4.27)	Ref
Non-insulin dependent	149	13 (8.72)	0.0432	0.009 to 0.078	0.014	0.0075	-0.028 to 0.043	0.677
Insulin dependent	63	7 (11.11)	0.0664	0.014 to 0.119	0.013	0.0040	-0.048 to 0.056	0.879

ASA: American Society of Anesthesiologists; MAC: Monitored anesthesia care; Hx of CHF: History of congestive heart failure; Hx of COPD: History of chronic obstructive pulmonary disease.

Figure 1 illustrates the annual trend in the proportion of filtered cases per 10000 total cases from 2014 to 2023. The proportion increased from 3.86 in 2014 (n = 286 out of 750937 total cases) to a peak of 4.24 in 2022 (n = 408 out of 1011899 total cases), reflecting a general upward trend over the study period. Although minor fluctuations were observed year to year, the overall trajectory showed growth. In 2023, the proportion slightly declined to 3.41 per 10000 cases (n = 332 out of 994314 total cases). Across the entire 2014-2023 period, there was no statistically significant linear trend in annual case proportion (Pearson r = 0.01, P = 0.969).

Figure 1 Annual proportion of filtered cases per 10000 total cases in the National Surgical Quality Improvement Program dataset, 2014-2023. NSQIP: National Surgical Quality Improvement Program.

DISCUSSION

This study analyzed hip arthroscopies performed between 2014 to 2023 in the ACS-NSQIP database to examine the case proportion trends and 30-day postoperative complication rates. Hip arthroscopy proportion decreased by 38% between 2016 to 2018, then increased by 65% from 2018 to 2022. The temporal pattern fluctuated over the study period, with an increase in 2018 to 2022 followed by a slight decline in 2023. The overall 30-day postoperative complication rate was 5.58% with the most common complication being bleeding that requires transfusion (2.2%). Regional anesthesia/MAC was associated with a significant increase in risk of any adverse event, while MAC/IV sedation demonstrated a modest but significant association. The incidence of hip arthroscopy complications including DVT, infection, PE, readmission, reoperation, and death in the perioperative period over the last 10 years have not been reported.

The current study demonstrates that the rate of the use of hip arthroscopies has fluctuated over the 2014 to 2023 period. A previous database study analyzing the volume of hip arthroscopies between the 2015 to 2020 period demonstrated a notable decline in the number of arthroscopies performed, which further supports the fluctuations demonstrated in our results[11]. Hip arthroscopy was primarily used as a diagnostic tool in the 1980s-1990s, while in the early 2000s, its use expanded to a therapeutic use, accompanied by a better understanding of hip pathologies like femoroacetabular impingement syndrome and labral tears[1] Its expanded use for both diagnostic and therapeutic purposes became more common between 2006 and 2013, which coincides with a surge in procedures[10]. The continued increase in the use of arthroscopies up to 2016 may reflect to a late majority phase where adoption has reached saturation. The following fluctuations starting in 2019 may be attributed to the early stages of the coronavirus pandemic and the increase the following years may be operating rooms reopening and the resumption of elective cases backlogged[12]. The observed fluctuations, including a decline in 2023 may suggest a plateau in the adoption of hip arthroscopy.

Conversely, a TriNetX study observing the use of hip arthroscopies from 2015 to 2023 demonstrated a steady increase in hip arthroscopies[13]. Differences in results may be due to the TriNetX study including around eight times the amount of total procedures. Nevertheless, TriNetX relies heavily on billing codes and accurate documentation which vary across institutions, so its results are overall less standardized and reliable compared to NSQIP studies[14].

Hip arthroscopy is a relatively safe procedure with small risks of complications that are mostly preventable. The reported rates of complications across literature have been variable, ranging from 1%-8%[15]. A previous NSQIP database study assessing the complications of hip arthroscopy from 2011 to 2016 reported 0.9% of cases had unplanned readmission to the hospital[16]. High wound class and perioperative blood transfusions were associated with the unplanned readmission rates[16]. This aligns with our findings that perioperative blood transfusions significantly increase risk of any adverse event. Furthermore, the most common complications reported in our study were bleeding, superficial infection, and deep venous thrombosis, which are all minor complications, and correspond with the most frequent reasons for admission in the prior study[16]. While minor complication rates are low, it is still crucial to identify these potential risks to optimize patient outcomes given the elective nature of hip arthroscopies.

Comparing trends over time offers valuable insight into how hip arthroscopies may have evolved. The growth in the proportion of hip arthroscopies has decreased and the complication rates of hip arthroscopies from 2014 to 2023 have increased compared to those of 2006 to 2013[10]. While the earlier period observed a 25-fold increase in volume, the most recent decade has shown a 1.2 times increase. The diminishing growth in case proportion may again be attributed to the changes in clinical practice and a better understanding of its indications and a natural decline in innovation adoption. It is also of note that the current study includes three times as many cases compared to Cvetanovich et al’s study[10]. With the greater number of procedures, the absolute number of major and minor complications has risen. Earlier reports document no cases of many of the categories of major complications, including deep infection, sepsis, septic shock, pulmonary embolism, ventilator use for more than 48 hours, acute renal failure, cardiac arrest, cerebrovascular accident, and coma[10]. The increase in major complication rates in the last decade compared to the earlier decade may instead reflect the greater number of surgeries performed rather than a higher complication rate per case. The absolute and rates of minor complications in the last decade are similar when compared to the previous decade. This highlights a gap in surgical complication management and reveals a target for improvement.

A partially dependent functional status was also strongly associated with increased risk for complications. It is difficult to compare our results with previous literature like Cvetanovich et al’s study, because NSQIP transitioned from a two-category system (independent and dependent) to the three-category system (independent, partially dependent, dependent)[10]. Even so, the new categories have helped reveal that a new patient population, partially dependent patients, are at increased risk for population, highlighting the need for functional assessment testing during pre-operative examination in order to minimize risk and optimize function prior to surgery.

Our results also demonstrated that a high ASA score (> 3) is predictive of greater 30-day complication rates. These results align with the current literature, including studies assessing hip fracture surgery outcomes and total hip arthroplasty readmission rates[17,18]. Higher ASA scores have been associated with increased incidence of complications, mortality, length of stay, and readmission rates. Thus, the ASA rating scale remains a valuable tool to identify those who are at increased risk of complications, including in hip arthroscopy.

Sex based differences were also observed in the context of complication rates for hip arthroscopies. Males were found to be associated with lower risk of complications, which aligns with the results from recent systematic reviews assessing hip arthroscopies for femoracetabular impingement[19,20]. Differences between sexes may be linked to anatomical differences or heterogeneity in hip pathologies.

Limitations

This study relies on the ACS-NSQIP database, which captures procedures in hospital-based settings, including hospital-affiliated outpatient surgery centers or surgery centers attached to hospitals, but does not include freestanding ambulatory surgical centers (ASCs)[21]. There has been a recent shift towards utilizing ASCs for hip arthroscopies. As reported by Tiao et al[22], the proportion of hip arthroscopies performed in ASCs increased from approximately 27% to 32.4% between 2013 to 2017. Consequently, our findings may not fully reflect all hip arthroscopy outcomes across all hospital settings. However, despite this increase for ASCs, its utilization remains lower than that of hospital-based centers, where the majority of hip arthroscopies are still performed[22]. Selection bias and confounding by indication may influence the observed complication rates, as those with greater comorbidity burden may be preferentially treated at hospital-affiliated outpatient surgery centers while healthier patients are selected for ASCs[23]. However, this distinction has not been well characterized in the NSQIP database, so it is difficult to determine whether the observed increase in complication rates over the past decade reflects a true trend or rather differences in underlying health. Therefore changes in complication rates should not necessarily be interpreted as evidence of changes in procedural safety. Acknowledging recent trends and nuances in the location of these procedures should be considered for future database research on hip arthroscopies.

Complications beyond the 30-day postoperative window are also not recorded, limiting the ability to evaluate long-term outcomes and complications. Additionally, NSQIP data does not identify surgeon factors, such as level of training, that may influence the complication rates. Hip arthroscopy procedural volume is low among orthopedic residents, and with the rising need for these procedures, there may be a steep learning curve, which may contribute to complication rates[24]. Current literature suggests proficiency may be achieved after around 30 hip arthroscopy cases, though up to 500 cases may be needed to be considered experienced, with complication rates generally declining as experience increases[25,26]. Sparse-data bias may also affect estimates for uncommon complications or small subgroups, potentially contributing to adjusted estimates. Finally, like any NSQIP database study, this study relied on the accuracy and completeness of the data and CPT coding. Concurrent diagnoses that may have contributed to complications are not able to be assessed and could potentially confound results. Accordingly, the observed associations should not be interpreted as causal and further research incorporating comprehensive clinical and long-term outcome data is warranted.

CONCLUSION

Over the last decade, the utilization of hip arthroscopy has fluctuated. The 30-day risk of any adverse event was 5.58%, with most being minor complications (3.70%). The most common major complication was return to the operating room and the most common minor complication was bleeding requiring transfusion. Preoperative transfusion, partial dependence in functional status, higher ASA class, and regional anesthesia were associated with increased risks of complications, while male sex was associated with decreased risks of complications. Long-term outcomes beyond the 30-day postoperative window were not recorded and results were limited by hospital-based case capture of NSQIP, so these associations should be interpreted with caution.