Spinal Anesthesia in Total Hip Arthroplasty is Associated With Improved Outcomes in the American Joint Replacement Registry Population

Abstract

Background

Despite previous studies showing benefits of spinal anesthesia (SA) for patients undergoing elective total hip arthroplasty (THA), most THA procedures throughout the United States still utilize general anesthesia (GA). Using the American Joint Replacement Registry data, our study explored outcome difference for patients undergoing THA administered SA vs GA.

Methods

All available THAs were identified using American Joint Replacement Registry data from 2017 to 2020. THA patients were categorized into 2 cohorts by anesthesia type. Demographics, hospital characteristics, and comorbidities were documented for each patient. Outcomes included operative time, length of stay, 30- and 90-day readmission, and 90-day all-cause revision. Chi-square analysis was used to assess categorical variables while multivariable regression analyzed the association between anesthesia type and outcomes of interest.

Results

A total of 217,124 THAs were identified, including 119,425 (55.0%) patients who received GA and 97,699 (45.0%) patients who received SA. Multivariable regression showed that SA was associated with a decreased risk of hospital length of stay >3 days (adjusted odds ratio [aOR] 0.4, 95% confidence interval [CI]: 0.34-0.36, P < .0001) and a lower likelihood of prolonged operative time (aOR 0.8, 95% CI: 0.79-0.82, P < .0001). Additionally, patients who received SA had lower rates of 90-day readmission (aOR 0.7, 95% CI: 0.67-0.78, P < .0001) and a decreased risk of 90-day all-cause revision (aOR 0.5, 95% CI: 0.47-0.54, P < .0001).

Conclusions

Patients receiving SA during THA had shorter operative time, reduced length of stay, and decreased rates of readmission and revision compared to patients who received GA. These findings add to the growing body of literature supporting the benefits of SA over GA for THA patients.

Introduction

Multiple studies have suggested neuraxial spinal anesthesia (SA) may be superior to general anesthesia (GA) for total hip arthroplasty (THA) [[1], [2], [3], [4], [5], [6], [7], [8], [9]]. Reported benefits of SA include lower transfusion rates, fewer thromboembolic events, decreased opiate consumption, shorter hospital length of stay (LOS), lower revision and complication rates, and lower overall hospitalization cost [[1], [2], [3], [4], [5], [6], [7], [8], [9]]. In recent years, there has been a push toward increased utilization of SA during THA, spurred, in part, by inclusion in the Joint Commission’s performance criteria [10]. Despite this, studies have consistently shown that more than 50% of THA procedures in the United States continue to be performed under GA [[11], [12], [13]].

The reasons for continued predominant GA use are likely multifactorial, including both patient and institutional factors [7]. Certain comorbidities common to the THA age group, including coagulopathy, cardiac valve disorders, aortic stenosis, and lumbar spine pathology may also preclude the use of SA in select patients [7,13,14]. Several recent studies utilizing modern GA and postoperative pain protocols have reported equivalent or improved outcomes with GA including successful early discharge [[15], [16], [17]]. As such, GA continues to remain a prevalent option for patients undergoing THA.

While there is an established body of evidence suggesting superior outcomes associated with SA, these studies often rely on institutional databases that may not be generalizable or on older registry data sets with potentially outdated anesthesia protocols. Furthermore, many previous investigations also occurred before the advent of modern recovery protocols, including opioid-sparing anesthesia, multimodal analgesia, and the broad utilization of perioperative medication such as tranexamic acid and decadron. The purpose of our study was to conduct a population-wide contemporary analysis looking at differences in operative time, LOS, as well as readmission and revision rates between patients receiving spinal vs GA.

Material and methods

THA cases were identified using the American Joint Replacement Registry (AJRR) data from 2017 to 2020. The AJRR is the largest registry of THA data in terms of annual procedures reported, collecting data from all 50 states and over 1300 institutions [18]. Patients were categorized according to anesthesia type, SA or GA, defining the 2 comparative cohorts in our study. The AJRR data were linked to the Center for Medicare and Medicaid Services database to capture all patient revisions or readmissions at non-AJRR facilities for patients 65 years of age and older. Cases with incomplete, missing, or combined general and SA were excluded. Demographic characteristics including sex, age, and body mass index (BMI) were documented for each patient. The American Hospital Association Data Survey Fiscal Year 2015 was used to help define hospital size and teaching status while the Charlson Comorbidity Index (CCI) was used to account for any patient comorbidities. [19] As all patient-level data were de-identified in accordance with the Health Insurance Portability and Accountability Act, this study was exempt from institutional review board review.

Outcomes of interest included early revision, 30-day and 90-day readmission, LOS, and operative time. Early revision was defined as a revision procedure performed within 90 days of the index THA. These revisions were identified by matching procedure site, laterality, and case identifier. Cases lasting longer than the overall median surgical time were classified as having longer operative time. Any inpatient stay >3 days was defined as having an extended LOS.

Descriptive statistics were utilized to evaluate the relationship between the independent variables and the binary outcome variable. Chi-square analysis was used to determine statistical differences between categorical variables. Multivariable logistic regression analyses were employed to analyze the association between anesthesia type and the outcomes of interest, accounting for potential confounding covariates such as BMI, gender, age, region, CCI, hospital teaching status, and hospital size. SAS, version 9.4, was utilized to conduct all statistical analyses (SAS Institute, NC, USA).

Patient demographics and cohort characteristics

A total of 217,124 THAs identified, of which 119,425 (55%) received GA, and 97,699 (45%) received SA. Over the study period, there was an increase in the use of SA, with 43% of patients who were operated under SA in 2017 compared to 47% in 2020. (Table 1) Compared to patients receiving SA, patients receiving GA had similar but statistically different ages (GA 66.9 ± 11.8 vs SA 66.5 ± 11.0, P < .0001), BMI (GA 30.2 ± 6.9 vs SA 29.8 ± 6.1, P < .0001), and proportion of women (GA 54.6% vs SA 55.1%, P < .022). However, patients in the GA group had a higher CCI score on average (GA 0.6 ± 1.3 vs SA 0.4 ± 1.0, P < .0001) with a greater proportion of patients having a CCI greater than 3 (GA 6.8% vs SA 3.9%, P < .0001). (Table 1)

Table 1.

Demographics, hospital characteristics, and utilization of spinal vs general anesthesia in total hip arthroplasty^a.

N = 217,124	General		Spinal		Total		P value
Year	N	Row %	N	Row %	N	Row %
2017	16,670	56.8	12,686	43.2	29,356	100	<.0001
2018	34,973	57.6	25,699	42.4	60,672	100
2019	38,146	53.5	33,115	46.5	71,261	100
2020	29,636	53.1	26,199	46.9	55,835	100

	N	Col %	N	Col %	N	Col %
Age
Mean age, y (SD) and 95% confidence limits	66.9 ± 11.8 (66.8, 66.9)		66.5 ± 11.0 (66.4, 66.6)		66.7 ± 11.5 (66.6, 66.7)		<.0001
<50	8353	7.0	6245	6.4	14,598	6.7	<.0001
50-59	21,270	17.8	17,569	18.0	38,839	17.9
60-69	39,490	33.1	34,732	35.6	74,222	34.2
70-79	33,971	28.5	28,208	28.9	62,179	28.6
80-89	13,760	11.5	9779	10.0	23,539	10.8
>90	2581	2.2	1166	1.2	3747	1.7
Sex
Women	65,228	54.6	53,873	55.1	119,101	54.9	.022
Men	51,593	43.2	43,472	44.5	95,065	43.8
Missing	2604	2.2	354	0.4	2958	1.4
Charlson comorbidity index (CCI)
Mean CCI score ±SD and 95% confidence limits	0.6 ± 1.3 (0.6, 0.6)		0.4 ± 1. (0.4, 0.4)		0.5 ± 1.2 (0.5, 0.5)		<.0001
CCI score 0	84,857	71.1	75,777	77.6	160,634	74.0	<.0001
CCI score 1	18,382	15.4	13,092	13.4	31,474	14.5
CCI score 2	8100	6.8	5011	5.1	13,111	6.0
CCI score >3	8086	6.8	3819	3.9	11,905	5.5
Patient BMI
Mean BMI (SD) and 95% confidence limits	30.2 ± 6.9 (30.2, 30.2)		29.8 ± 6.1 (29.7, 29.8)		30.0 ± 6.5 (30.0, 30.0)		<.0001
Underweight	1520	1.3	849	0.9	2369	1.1	<.0001
Normal	22,322	18.7	18,184	18.6	40,506	18.7
Preobese	34,492	28.9	30,239	31.0	64,731	29.8
Obesity class I	27,303	22.9	23,123	23.7	50,426	23.2
Obesity class II	14,865	12.5	11,602	11.9	26,467	12.2
Obesity class III	10,227	8.6	6004	6.2	16,231	7.5
Missing	8696	7.3	7698	7.9	16,394	7.6
REGION
Northeast	20,145	16.9	25,966	26.6	46,111	21.2	<.0001
Midwest	30,723	25.7	22,209	22.7	52,932	24.4
South	35,417	29.7	24,444	25.0	59,861	27.6
West	32,570	27.3	24,385	25.0	56,955	26.2
Missing	570	0.5	695	0.7	1265	0.6
Hospital bed size
Small (1-99)	21,375	17.9	18,503	18.9	39,878	18.4	<.0001
Medium (100-399)	51,521	43.1	38,819	39.7	90,340	41.6
Large (≥400)	38,513	32.3	32,026	32.8	70,539	32.5
Missing	8016	6.7	8351	8.6	16,367	7.5
Teaching status
Major	19,055	16.0	18,569	19.0	37,624	17.3	<.0001
Nonteaching	29,817	25.0	22,953	23.5	52,770	24.3
Minor	63,256	53.0	49,222	50.4	112,478	51.8
Missing	7297	6.1	6955	7.1	14,252	6.6
Length of stayb
Mean length of stay, d (SD) and 95% confidence limits	2.2 ± 2.8 (2.2, 2.3)		1.6 ± 1.4 (1.6, 1.6)		2.0 ± 2.3 (1.9, 2.0)		<.0001
0 d	5838	4.9	6213	6.4	12,051	5.6	<.0001
1 d	54,601	45.7	54,133	55.4	108,734	50.1
2 d	28,097	23.5	22,717	23.3	50,814	23.4
3 d	13,844	11.6	9125	9.3	22,969	10.6
>4 d	16,477	13.8	5128	5.3	21,605	10.0
Missing	568	0.5	383	0.4	951	0.4
Operative timec
Mean operative time, mins (SD) and 95% confidence limits	96.5 ± 47.7 (96.2, 96.8)		87.3 ± 37.6 (87.1, 87.6)		92.3 ± 43.7 (92.2, 92.5)		<.0001
<64 mins	25,745	21.6	24,137	24.7	49,882	23.0	<.0001
65-83 mins	26,785	22.4	25,126	25.7	51,911	23.9
84-109 mins	25,734	21.6	23,265	23.8	48,999	22.6
>110mins	30,847	25.8	17,809	18.2	48,656	22.4
Missing	10,314	8.6	7362	7.5	17,676	8.1

^aStatistical testing performed on known variables only.

^bLength of Stay calculated as (discharge date – admission date) with accepted range of 0-90 days.

^cMedian operative time, mins [IQR] for general (85 mins [49]), spinal (80 mins [39], and total (83 mins [45] where interquartile range = upper quartile – lower quartile = Q_-3 – Q_-1.

Hospital characteristics

Hospitals in the Northeast (GA 16.9% vs SA 26.6%, P < .0001) were more likely to utilize SA over GA. In contrast, hospitals in the Midwest, South, and West were more likely to administer GA compared to SA. Additionally, patients receiving SA were more likely to be seen at smaller and larger hospitals, while patients who received GA were more likely to be seen at medium-sized hospitals. Nonteaching and minor teaching hospitals were more likely to use GA, while major teaching hospitals utilized SA more frequently (GA 16.0% vs SA 19.0%, P < .0001) (Table 1).

Results

Length of stay

Patients who received GA experienced a longer average LOS (GA 2.2 ± 2.8 vs SA 1.6 ± 1.4 days, P < .0001). Additionally, a greater proportion of SA patients stayed for 0 or 1 days postoperatively while a greater proportion of GA patients stayed for 2, 3, and greater than 4 days. (Table 1) After adjusting for confounding factors, patients in the SA group had a lower likelihood of staying in the hospital for more than 3 days (adjusted odds ratio [aOR] 0.3, 95% confidence interval [CI]: 0.34-0.36, P < .0001) (Table 2).

Table 2.

Logistic regression assessing factor association to longer length of stay (LOS)^a in THAs (N = 183,802).

Covariate	Point estimate (odds ratio)	95% lower confidence limit	95% upper confidence limit	P value
Spinal (ref: General)	0.3	0.34	0.36	<.0001
Age	1.1	1.05	1.05	<.0001
Sex (ref: Men)	1.4	1.32	1.41	<.0001
CCI	1.4	1.43	1.46	<.0001
BMI	0.9	0.91	0.93	<.0001
Region: Northeast (ref: West)	0.9	0.87	0.96	.0001
Region: South (ref: West)	1.0	0.92	1.01	.1075
Region: Midwest (ref: West)	1.0	0.92	1.01	.1017
Hospital bed size: Small (ref: medium)	1.1	1.05	1.14	<.0001
Hospital bed size: Large (ref: medium)	0.5	0.44	0.49	<.0001
Hospital teaching status: Major (ref: minor)	1.6	1.58	1.72	<.0001
Hospital teaching status: Nonteaching (ref: minor)	0.8	0.76	0.83	<.0001

^aLonger LOS is defined as more than 3 days.

Operative time

Compared to patients who received SA, those who were administered GA had longer operative times (GA 96.5 ± 47.7 vs 87.3 ± 37.6 minutes, P < .0001). Additionally, a greater proportion of GA patients had operative times greater than or equal to 110 minutes, whereas SA patients had operative times less than 110 minutes. (Table 1) After accounting for potential confounding variables, patients with SA had a reduced risk of having an operative time greater than 83 minutes (aOR 0.8, 95% CI: 0.79-0.82, P < .0001) (Table 2, Table 3).

Table 3.

Logistic regression table assessing factor association to longer operative time^a in THAs (N = 170,931).

Covariate	Point estimate (odds Ratio)	95% lower confidence limit	95% upper confidence limit	P value
Spinal (ref: general)	0.8	0.79	0.82	<.0001
Age	0.9	0.99	0.99	<.0001
Sex (ref: men)	0.9	0.86	0.90	<.0001
CCI	1.0	1.03	1.05	<.0001
BMI	1.2	1.15	1.17	<.0001
Region: Northeast (ref: West)	1.0	0.93	0.98	.0011
Region: South (ref: West)	0.9	0.88	0.93	<.0001
Region: Midwest (ref: West)	1.1	0.98	1.03	.6082
Hospital bed size: Small (ref: medium)	1.0	1.02	1.07	.001
Hospital bed size: Large (ref: medium)	0.9	0.90	0.95	<.0001
Hospital teaching status: Major (ref: minor)	1.5	1.44	1.53	<.0001
Hospital teaching status: Nonteaching (ref: minor)	0.9	0.92	0.97	<.0001

^aLonger operative time is defined as more than the overall median time of 83 minutes.

Unplanned readmission

After accounting for confounding variables, patients who received SA were found to be at a lower risk of readmission within both 30 days (aOR 0.6, 95% CI: 0.56-0.69, P < .0001) and 90 days (aOR 0.7, 95% CI: 0.67-0.78, P < .0001) (Table 4, Table 5). Of interest, patients with a higher BMI had increased rates of 90-day (aOR 1.1, 95% CI: 1.04-1.10, P < .0001) and 30-day readmission (aOR 1.1, 95% CI: 1.06-1.14, P < .0001).

Table 4.

Logistic regression assessing factor association to 90-day readmission in THAs (N = 184,202).

Covariate	Point estimate (odds ratio)	95% lower confidence limit	95% upper confidence limit	P value
Spinal (ref: general)	0.7	0.67	0.78	<.0001
Age	1.0	1.01	1.01	<.0001
Sex (ref: men)	1.0	0.90	1.03	.3012
CCI	1.2	1.22	1.26	<.0001
BMI	1.1	1.04	1.10	<.0001
Region: Midwest (ref: West)	0.5	0.41	0.50	<.0001
Region: Northeast (ref: West)	0.2	0.17	0.22	<.0001
Region: South (ref: West)	0.2	0.17	0.21	<.0001
Hospital bed size: Large (ref: medium)	1.0	0.94	1.12	.5805
Hospital bed size: Small (ref: medium)	0.9	0.83	1.00	.0561
Hospital teaching status: Major (ref: minor)	1.5	1.30	1.64	<.0001
Hospital teaching status: Nonteaching (ref: minor)	1.2	1.15	1.36	<.0001

Table 5.

Logistic regression assessing factor association to 30-day readmission in THAs (N = 184,202).

Covariate	Point estimate (odds ratio)	95% lower confidence limit	95% upper confidence limit	P value
Spinal (ref: general)	0.6	0.56	0.69	<.0001
Age	1.0	1.02	1.02	<.0001
Sex (ref: men)	1.0	0.91	1.11	.8983
CCI	1.3	1.23	1.29	<.0001
BMI	1.1	1.06	1.14	<.0001
Region: Midwest (ref: West)	0.5	0.42	0.54	<.0001
Region: Northeast (ref: West)	0.2	0.17	0.24	<.0001
Region: South (ref: West)	0.2	0.17	0.23	<.0001
Hospital bed size: Large (ref: Medium)	0.9	0.84	1.07	.3748
Hospital bed size: Small (ref: medium)	0.8	0.73	0.96	.0098
Hospital teaching status: Major (ref: minor)	1.5	1.28	1.75	<.0001
Hospital teaching status: Nonteaching (ref: minor)	1.2	1.04	1.31	.0111

Revision surgery

When accounting for potential confounding factors, patients who received SA had a lower risk of all-cause revision within 90 days of THA (aOR 0.5 95% CI: 0.47-0.54, P < .0001) (Table 6). Notably, an increased rate of revision was found for patients with increased age (aOR 1.1, 95% CI: 1.05-1.05, P < .0001), CCI (aOR 1.1, 95% CI: 1.07-1.11, P < .0001), and BMI (aOR 1.0, 95% CI: 1.02-1.07, P = .0007).

Table 6.

Logistic regression assessing factor association to 90-day revision^a in THAs (N = 184,223).

Covariate	Point estimate (odds ratio)	95% lower confidence limit	95% upper confidence limit	P value
Spinal (ref: general)	0.5	0.47	0.54	<.0001
Age	1.1	1.05	1.05	<.0001
Sex (ref: men)	1.1	1.02	1.14	.0119
CCI	1.1	1.07	1.11	<.0001
BMI	1.0	1.02	1.068	.0007
Region: Midwest (ref: West)	0.7	0.67	0.79	<.0001
Region: Northeast (ref: West)	0.5	0.49	0.58	<.0001
Region: South (ref: West)	0.7	0.66	0.77	<.0001
Hospital bed size: Large (ref: medium)	1.0	0.94	1.08	.9243
Hospital bed size: Small (ref: medium)	0.7	0.63	0.76	<.0001
Hospital teaching status: Major (ref: minor)	1.5	1.37	1.61	<.0001
Hospital teaching status: Nonteaching (ref: minor)	0.8	0.71	0.83	<.0001

^a90-Day revision is defined as a case having a revision procedure within 90 days of a primary procedure matching patient ID, procedure site, and laterality.

Discussion

In our study containing 217,124 THAs, we found that patients who received SA had significantly shorter LOS, reduced operative times, lower rates of 30-day and 90-day readmissions, and lower rates of revision compared to patients who received GA. This further adds to the existing body of literature suggesting SA may be the optimal form of anesthesia for patients undergoing THA [2,3,5,7,12,13,17,[20], [21], [22], [23], [24]].

SA has long been shown to be associated with shorter LOS following THA [11,20,25]. In a single-institutional retrospective study of 500 THAs, Kelly et al. found there was a significant difference in the LOS for SA and GA patients, with patients who received SA having a significantly lower average LOS (SA 32.7 ± 14.8 hours vs GA 38.1 ± 24.0 hours, P = .003) [20]. Furthermore, Harsten et al. found that neuraxial anesthesia reduced LOS (SA 26 hours vs GA 30 hours, P = .004) in their randomized controlled trial of 120 THA patients [17]. Our study adds to these findings, demonstrating that SA patients have a 0.3 times reduced risk of having a prolonged hospital LOS greater than 3 days compared to GA patients. This further emphasizes the added benefits of using SA as longer LOS for THA patients have been found to be associated with higher readmission rates, increased risk of complications, and higher overall costs to the healthcare system [20,26,27].

Our study also showed significant differences in operative times: patients receiving GA had an average of 9 minutes longer surgery time compared to patients receiving neuraxial anesthesia. These findings are similar to those reported by Basques et al., who conducted a retrospective study involving 20,936 THA patients using the ACS-NSQIP database. Basques showed that patients receiving SA spent 12 minutes fewer in the operating room (SA 87 ± 38 minutes vs GA 99 ± 40 minutes, P < .001), with the authors concluding that these differences were due to the time needed for patient extubation [7]. Furthermore, unlike GA, SA can be administered in induction rooms before entry to the operating room, which reduces time spent in the operating room [28]. Of note, however, even though SA does not need to be administered in the operative room, neuraxial anesthesia does take longer to induce, which must be taken into consideration during general preoperative planning [29,30]. In addition to differences in operative times, Sowers et al. reported that with GA there is also a significantly higher total non-operative time and time to transfer patients postoperatively compared to patients who receive neuraxial anesthesia [28]. Thus, the benefits of SA may extend beyond reduced operative time and including reduced perioperative times, as well.

Data in the current literature comparing readmission rates of general vs SA patients in THA offer mixed results [7,13,23]. In a single-institutional review of 13,730 THAs, Owen et al. reported no significant differences between 30- and 90-day readmission risk when comparing patients receiving general and SA. [23] Warren et al. conducted a NSQIP database study examining 110,963 THA patients, of which 45,871 received SA and 65,092 received GA. After accounting for confounding variables, the authors found that patients receiving GA had a significantly increased risk of 30-day readmission (aOR 1.1, 95% CI: 1.03-1.12, P = .004). [13]. Our study supports the findings of Warren et al., demonstrating that patients receiving SA had a 0.6-fold reduced risk of 30-day readmission and a 0.7-fold reduced risk of 90-day readmission.

When examining 90-day revision rates, our study found that patients who received SA had a 0.5-fold reduced risk of revision compared to GA patients. This is in contrast to the findings of Owen et al., who found no differences in 90-day revision rates between SA and GA patients [23]. Whereas Owen’s study included 13,730 patients, all of whom were from a single institution with dates of surgery ranging from 2001 to 2016, our cohort had 217,124 patients included from a broad sample of institutions who had a THA between a narrower time window. Thus, not only was our cohort more sizable, but given changes in anesthesia techniques over time, our cohort likely more accurately reflects current anesthesia practices. While our study provides promising findings regarding the impact of SA on reducing revision risks, there is limited literature on this topic, emphasizing the need for further research to assess this relationship.

Our study has several limitations. AJRR relies on accurate data provided by a wide scope of institutions, which may be prone to errors during data extraction and transmission from disparate electronic health record systems. However, AJRR conducts internal audits to ensure the highest data accuracy provided by all institutions. Additionally, the AJRR recently acquired claims data from the Centers for Medicare and Medicaid Services to supplement missing and incomplete data for at least Medicare-eligible patients in the registry, maximizing both data completeness and accuracy. Therefore, it is highly unlikely that any potential inaccuracies in the source data would influence the results of this study by skewing one group preferentially more than the other, given that AJRR provides the largest sample of primary elective THAs in the United States [18]. Additionally, this study is vulnerable to potential confounding factors given its retrospective nature. While multivariable analyses were used to account for differences in cohorts' baseline characteristics such as BMI and CCI, there are still potential unobserved differences that we were unable to account for including anesthesia and surgeon selection bias. For example, surgical approach may influence anesthesia choice depending on an anesthesiologist’s comfort level intubating a patient in a particular position. Furthermore, we were unable to control for factors such as type of SA administered or the specific GA technique employed, including whether patients were intubated with paralysis or managed with a laryngeal mask airway. While our study can help establish an association between the use of SA and improved outcomes such as reduced operative times, shorter LOS, and decreased rates of revision and readmission, we cannot establish any causal relationship. Finally, certain patients, such as those with lumbar degenerative disease or with various coagulopathies, may not have the option of receiving SA and may naturally be at higher risk for longer operative times, extended hospital stay, and greater risk of readmission and revision. These factors that can affect the choice of anesthesia may not have been identified accurately in our study. Given the unobserved differences and selection biases, coupled with our inability to provide a causative explanation for the differences in outcomes, we must acknowledge that it is plausible that THA patients administered GA may have comparable 30-day and 90-day outcomes to THA patients administered SA and that the differences in outcomes observed in the present study may be related to unobservable confounding factors. As such, orthopaedic surgeons should interpret the data presented within the context of these limitations that are inherent to any retrospective observational registry study.

Conclusions

In the linked AJRR and Center for Medicare and Medicaid Services data set, SA was associated with improved outcomes, including shorter LOS, decreased operative time, and reduced risk of readmission and revision compared to GA. These findings add to the existing body of literature supporting the use of SA for patients undergoing primary THA. Future multicenter prospective studies should be conducted to evaluate whether the administration of SA has a causal effect on THA outcomes reported in this study.

Conflicts of interest

The authors declare there are no conflicts of interest.

For full disclosure statements refer to https://doi.org/10.1016/j.artd.2024.101566.

CRediT authorship contribution statement

Sagar Telang: Writing – original draft, Investigation, Formal analysis, Data curation. Nathanael D. Heckmann: Writing – original draft, Investigation, Formal analysis, Data curation, Conceptualization. Adam Olsen: Writing – original draft, Validation, Investigation, Formal analysis, Data curation, Conceptualization. Ayushmita De: Writing – original draft, Validation, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Jeffrey B. Stambough: Writing – original draft, Investigation, Formal analysis, Data curation, Conceptualization.