Outpatient shoulder arthroplasty in the COVID-19 era: 90-day complications and risk factors

Abstract

Background

With the COVID-19 pandemic placing an increased burden on health care systems, shoulder arthroplasties are more commonly being performed as outpatient procedures. The purpose of this study was to characterize the 90-day episode-of-care complications of consecutive shoulder arthroplasties defaulted for outpatient surgery without using a prior algorithm for patient selection and to assess for their risk factors. We hypothesized that outpatient shoulder arthroplasty would be a safe procedure for all patients, regardless of patient demographics and comorbidities.

Methods

A retrospective review of consecutive patients who underwent planned outpatient anatomic or reverse total shoulder arthroplasty between March 2020 and January 2022 with 3-month follow-up was performed. All patients were scheduled for outpatient surgery regardless of medical comorbidities. Patient demographics; pre/postoperative patient-reported outcomes including visual analog scale, subjective shoulder value, and American Shoulder and Elbow Surgeons score; pre/postoperative range of motion; and complications were collected from medical chart review. Multivariate logistic regression was used to identify predictors of the following outcomes: 1. Unplanned overnight hospital stay, 2. 90-day unplanned emergency department (ED)/clinic visit, 3. 90-day hospital readmission, 4. 90-day complications requiring revision.

Results

One hundred twenty-seven patients (47% male, 17% tobacco users, 18% diabetics) with a mean age 69 ± 9 years were identified, of whom 92 underwent reverse total shoulder arthroplasty and 35 underwent anatomic total shoulder arthroplasty. All patient-reported outcomes and range of motion were significantly improved at 3 months. There were 15 unplanned overnight hospital stays (11.8%) after the procedure. Within 90 days postoperatively, there were 17 unplanned ED/clinic visits (13.4%), 7 hospital readmissions (5.5%), and 4 complications requiring revision (3.1%). Factors predictive of unplanned overnight stay included age above 70 years (odds ratio [OR], 36.80 [95% confidence interval [CI], 2.20-615.49]; P = .012), tobacco use (OR, 12.90 [95% CI, 1.23-135.31]; P = .033), and American Society of Anesthesiologists status of 3 (OR, 13.84 [95% CI, 1.22-156.57]; P = .034). The only factor predictive of unplanned ED/clinic visit was age over 70 years old (OR, 7.52 [95% CI, 1.26-45.45]; P = .027). No factors were predictive of 90-day hospital readmission or revision.

Conclusion

Outpatient shoulder arthroplasty is a safe procedure with excellent outcomes and low rates of readmissions and can be considered as the default plan for all patient undergoing shoulder arthroplasty. Patients who are above 70 years of age, use tobacco, and have ASA score of 3, however, may be less suitable for outpatient arthroplasty and should be counseled regarding the higher risk of unplanned overnight hospitalization.

Both anatomic and reverse total shoulder arthroplasty have been shown to have favorable outcomes for a variety of indications including glenohumeral osteoarthritis.^{17 , 18 , 22 , 26 , 29} These procedures have gained tremendous popularity over recent decades, given expanding indications, advancement in implant technology and surgical technique, as well as improvement in outcomes.^{6 , 13 , 32} Between 2011 and 2017, the number of primary shoulder arthroplasties performed increased by 103.7%.³² As shoulder arthroplasty has gained popularity in recent years, there has been a push towards performing these procedures in the outpatient setting. Recently in 2021, the Center for Medicare and Medicaid Services removed the shoulder arthroplasty Current Procedural Terminology codes from their inpatient-only list.²⁸ This transition has been largely driven by an emphasis on reducing health care costs and optimizing resource allocation.^{11 , 16}

Our current understanding of outpatient shoulder arthroplasty is that it is generally safe and efficacious for “appropriately selected patients”.^{1 , 7 , 14 , 25} Many of the landmark studies of outpatient shoulder arthroplasty have proposed stringent patient selection algorithms, often excluding patients with common comorbidities including obesity, chronic obstructive pulmonary disease (COPD), history of myocardial infarction (MI), and history of deep vein thrombosis (DVT) or pulmonary embolism (PE).^{4 , 7 , 9 , 14} However, with the COVID-19 pandemic placing an increased burden on health care systems, it has become more pressing than ever to perform elective procedures in the outpatient setting. Thus, over the past 2 years, more outpatient shoulder arthroplasties have been performed with less stringent preoperative selection criteria, expanding to include patients with increased age and comorbidities.^{23 , 27 , 33}

Very few studies, however, have assessed outcomes of outpatient shoulder arthroplasty in the COVID-19 era with less stringent patient selection criteria. Thus, the purpose of this study was to characterize the 90-day episode-of-care complications of outpatient shoulder arthroplasty without using a prior algorithm for patient selection and to assess for their risk factors. We hypothesized that outpatient shoulder arthroplasty may be a safe procedure for all patients, regardless of patient demographics and comorbidities.

Methods

Study design and patient collection

This was a retrospective cohort study that reviewed electronic medical records of patients who underwent planned outpatient anatomic or reverse total shoulder arthroplasty by one fellowship-trained orthopedic shoulder surgeon between March 2020 and January 2022.

Inclusion criteria for the study were patients above 30 years of age who underwent planned anatomic or reverse total shoulder arthroplasty with a minimum of 3-month postoperative follow-up data. During the study period, all patients were scheduled for outpatient surgery regardless of medical comorbidities due to the COVID-19 pandemic. Indications for shoulder arthroplasty in this study included patient's glenohumeral osteoarthritis, rotator cuff arthropathy, and humeral fracture. Exclusion criteria included history of prior ipsilateral shoulder arthroplasty, prior infection of the ipsilateral shoulder, preoperative brachial plexus dysfunction, and postoperative follow-up less than 3 months.

Data acquisition: demographic and intraoperative data

Baseline demographic variables of age, body mass index BMI, sex, handedness, and occupation were recorded, along with several comorbidities including tobacco/alcohol use, diabetes, history of MI, history of DVT/PE, and COPD. These parameters were used to calculate patient's Charlson Comorbidity Index (CCI), a well-validated tool used to determine postoperative mortality and complication risk.¹⁰ Additionally, American Society of Anesthesiologists (ASA) physical status class was collected from the preoperative anesthesiology note in the electronic medical record.

Intraoperative variables were also collected including surgery type, surgery indication, anesthesia used, implant type, length of procedure, intraoperative complications, and postoperative discharge medications (pain medication and antibiotics).

Data acquisition: 90-day outcomes

Primary outcomes were 90-day episode-of-care complications including unplanned hospital stay on the first night postoperatively, unplanned emergency department (ED) or clinic visits, unplanned hospital readmissions, need for revision, 90-day mortality, request for additional pain meds, surgical site infection, and DVT/PE.

All preoperative secondary outcomes were determined at the last clinic visit before surgery and all postoperative secondary outcomes were determined at the 3-month follow-up clinic visit postoperatively. Secondary outcomes included preoperative and postoperative (3-month) patient-reported outcomes including visual analog scale (VAS), subjective shoulder value (SSV), and American Shoulder and Elbow Surgeons (ASES) Shoulder score as well as range of motion (ROM) in forward flexion (FF), and external rotation (ER). ROM testing was performed via physical exam by a fellowship trained shoulder surgeon. FF and ER ROM were evaluated in their respective planes.

Operative details

All procedures were done with the patient in the beach chair position through a standard deltopectoral approach with subscapularis peel (rTSA) or lesser tuberosity osteotomy (aTSA). Preoperative prophylactic antibiotics were administered within an hour of skin incision.

After surgery, patients were evaluated in the postanesthesia care unit. Once vital signs were stable, pain control was adequate, and patients could ambulate and void without assistance, they were discharged home. Patients with unstable vital signs, severe pain or nausea, and inability to void or ambulate were admitted to be monitored overnight and typically discharged the following day.

All patients received a single shot interscalene nerve block (preoperative for aTSA and postoperative for rTSA). Medications on discharge were standardized for all patients and included a multimodal pain regimen including acetaminophen, celecoxib, gabapentin, and oxycodone as well as cephalexin for 24-hour antibiotic prophylaxis (only given to those that did not require unplanned overnight hospital stay).

Postoperative rehabilitation protocol

Postoperative rehabilitation was standardized for all patients. All patients were fitted with a sling for 4 weeks postoperatively. Formal physical therapy guided by a physical therapist in a rehabilitation center was initiated at 2 weeks postoperatively starting with passive ROM exercises, including passive FF, passive external rotation, and pendulums as tolerated. Active ROM exercises as well as active-assisted exercises were gradually initiated at 6 weeks postoperatively as tolerated. Finally, shoulder strengthening exercises were introduced starting at 12 weeks.

Statistical analysis

Univariate analyses were conducted using 2-tailed paired samples t test. Multivariate logistic regression modeling was used to model the relationship between hypothesized risk factors and the following postoperative complications: 1. Unplanned overnight hospital stay, 2. 90-day unplanned ED/clinic visit, 3. 90-day hospital readmission, 4. 90-day complications requiring revision. Hypothesized risk factors included in the model were age, BMI, tobacco use, diabetes, history of MI, history of COPD, history of DVT/PE as well as CCI and preoperative ASA status. Covariate selection from a list of a priori variables was validated using univariate analyses. The final multivariate logistic regression model was validated via log likelihood ratio tests. All statistical analyses were conducted using SPSS software (version 28; IBM, Armonk, NY, USA). Significance level was set to P < .05.

Results

Study cohort

A total of 169 patients who underwent planned outpatient anatomic or reverse total shoulder arthroplasty were identified, of whom 127 had 3-month follow-up data and met the remaining inclusion criteria to be included in the final analysis. There were 92 rTSA (72.4%) and 35 aTSA (27.6%) procedures for the following indications: 63 glenohumeral osteoarthritis (49.6%), 54 rotator cuff arthropathy (42.5%), and 10 humerus fractures (7.9%) (Table I ). The final 127 patient cohort had an average age of 68.7 ± 9.1 years with an average BMI of 30.0 ± 6.0. Of the 127 patients, 59 were male (46.5%) and 21 used tobacco products (16.5%). Additional co-morbidities are shown in Table I. With regards to ASA status, 78 patients had a score of 2 (61.4%), indicating mild systemic illness, and 49 patients had a score of 3 (38.6%), indicating severe systemic illness (Table I). Finally with regards to CCI, 44 patients had a score of 1-2 (34.6%), 47 patients had a score of 3-4 (37.0%), and 36 patients had a score of 5+ (28.3%) (Table I).

Table I.

Demographic and treatment data for study cohort

Cohort variable	Outpatient arthroplasty (n = 127)
Procedure	92 rTSA (72.4%)
	35 aTSA (27.6%)
Indication	63 glenohumeral OA (49.6%)
	54 RC arthropathy (42.5%)
	10 humerus fracture (7.9%)
Age	68.7 ± 9.1 yr
BMI	30.0 ± 6.0
Sex (male)	59 (46.5%)
Diabetes	23 (18.1%)
Tobacco use	21 (16.5%)
History of MI	10 (7.9%)
History of COPD	10 (7.9%)
History of DVT/PE	10 (7.9%)
ASA physical status	2 – Mild systemic illness: 78 (61.4%)
	3 – Severe systemic illness: 49 (38.6%)
CCI	1-2 – Mild: 44 patients (34.6%)
	3-4 – Moderate: 47 patients (37.0%)
	≥5 – Severe: 36 patients (28.3%)

rTSA, reverse total shoulder arthroplasty; aTSA, anatomic total shoulder arthroplasty; OA, osteoarthritis; RC, rotator cuff; BMI, body mass index; MI, myocardial infarction; COPD, chronic obstructive pulmonary disease; DVT, deep vein thrombosis; PE, pulmonary embolism; ASA, American Society of Anesthesiologists; CCI, Charlson comorbidity index.

Ninety-day complications and risk factors

Of the 127 patients who underwent planned outpatient shoulder arthroplasty, 15 patients required an unplanned overnight stay (11.8%) as shown in Table II . All but one of these patients requiring overnight stay was discharged the following day after clearing medical evaluation. The one patient requiring longer stay had hypoxia secondary to atelectasis that subsequently developed into pneumonia (Table II). Factors predictive of unplanned overnight stay included age over 70 years (odds ratio [OR], 36.8 [95% confidence interval [CI], 2.2-615.5]; P = .012), tobacco use (OR, 12.9 [95% CI, 1.2-135.3]; P = .03), and ASA status of 3 (OR, 13.8 [95% CI, 1.2-156.6]; P = .034) (Table III ).

Table II.

90-Day complications following planned outpatient arthroplasty

90-Day complications	Incidence (n = 127)	Indications
Unplanned overnight stay	15 (11.8%)	6 for pain control
		4 for postoperative hypoxia
		3 for fever
		2 for severe nausea
Unplanned ED/clinic visit	17 (13.4%)	4 falls onto operative extremity
		3 DVT
		2 PE
		2 postoperative altered mental status
		2 operative arm swelling – not DVT
		1 urinary retention
		1 surgical site infection
		1 postoperative pneumonia
		1 trouble ambulating postoperatively
Hospital readmission	7 (5.5%)	2 PE (1 followed by MI)
		1 postoperative sepsis
		1 postoperative transient global amnesia
		1 postoperative pneumonia
		1 GI bleed i/s/o NSAID use
		1 irrigation and debridement
Need for revision	4 (3.1%)	2 dislocations
		1 fracture after fall
		1 secondary to surgical site infection

ED, emergency department; DVT, deep vein thrombosis; PE, pulmonary embolism; MI, myocardial infarction; GI, gastrointestinal; NSAID, non-steroidal anti-inflammatory drug.

Table III.

Odds of complications by demographics and comorbidities

Variable (n = 127)	Odds ratio for unplanned overnight stay	95% CI	P value
Age ≥ 70 yr	36.80	2.20-615.49	.012
BMI	1.09	0.97-1.21	.153
Tobacco use	12.90	1.23-135.31	.033
Diabetes	1.92	0.32-15.00	.427
COPD	5.21	0.70-55.23	.178
History of MI	3.76	0.25-67.26	.323
History of DVT/PE	3.13	0.28-34.85	.354
ASA physical status (3 as opposed to 2)	13.84	1.22-156.57	.034
CCI ≥ 5 (as opposed to CCI 1-2)	3.22	0.76-55.19	.245
CCI 3-4 (as opposed to CCI 1-2)	4.15	0.33-76.13	.374

Variable (n = 127)	Odds ratio for 90-day unplanned ED/clinic visit	95% CI	P value
Age ≥ 70 yr	7.90	1.26-45.45	.027
BMI	1.04	0.95-1.13	.448
Tobacco use	1.18	0.21-6.8	.856
Diabetes	3.08	0.47-20.22	.240
COPD	1.18	0.11-13.26	.891
History of MI	2.33	0.23-11.43	.418
History of DVT/PE	1.85	0.29-11.82	.517
ASA physical status (3 as opposed to 2)	1.74	0.39-7.68	.467
CCI ≥ 5 (as opposed to CCI 1-2)	2.07	0.43-9.93	.362
CCI 3-4 (as opposed to CCI 1-2)	1.52	0.18-13.21	.705

Variable (n = 127)	Odds ratio for 90-day hospital readmission	95% CI	P value
Age ≥ 70-yr	4.01	0.89-16.78	.824
BMI	1.08	0.95-1.33	.614
Tobacco use	1.41	0.14-14.54	.772
Diabetes	1.11	0.08-15.31	.936
COPD	0.87	0.40-1.62	.840
History of MI	9.05	0.59-137.5	.213
History of DVT/PE	2.13	0.27-30.21	.444
ASA physical status (3 as opposed to 2)	1.01	0.071-14.25	.995
CCI ≥ 5 (as opposed to CCI 1-2)	3.78	0.20-70.57	.373
CCI 3-4 (as opposed to CCI 1-2)	2.29	0.04-118.80	.682

Variable (n = 127)	Odds ratio for 90-day need for revision	95% CI	P value
Age ≥ 70-yr	2.11	0.94-4.22	.678
BMI	1.17	0.90-1.54	.234
Tobacco use	12.06	0.54-271	.117
Diabetes	1.66	0.11-19.22	.993
COPD	Insufficient data	Insufficient data	N/A
History of MI	Insufficient data	Insufficient data	N/A
History of DVT/PE	Insufficient data	Insufficient data	N/A
ASA physical status (3 as opposed to 2)	13.40	0.43-275.73	.111
CCI ≥ 5 (as opposed to CCI 1-2)	1.44	0.09-12.66	.998
CCI 3-4 (as opposed to CCI 1-2)	0.87	0.16-25.53	.997

BMI, body mass index; COPD, chronic obstructive pulmonary disease; MI, myocardial infarction; DVT, deep vein thrombosis; PE, pulmonary embolism; ASA, American Society of Anesthesiolog; CCI, Charlson comorbidity index; ED, emergency department.

Significance set at P value < .05 (bold).

In the 90-day period postoperatively, there were 17 unplanned visits to either the ED or to the clinic (13.4%) as shown in Table II. Of these 17 ED/clinic visits, 2 visits were on postoperative day one (due to urinary retention and postoperative hypoxia) and another 4 visits occurred within one week postoperatively. Seven visits were between one week and one month postoperatively while 4 visits were beyond one month postoperatively. The only factor predictive of unplanned ED/clinic visit was age over 70 years old (OR, 7.5 [95% CI, 1.3-45.5]; P = .027) (Table III).

Additionally, there were 7 hospital admissions related to the surgery (5.5%) as shown in Table II. All hospital admissions occurred after ED visits. Only 1 hospital admission was on postoperative day 1 (due to postoperative hypoxia). No factors were predictive of hospital readmission (Table III).

Within 90 days postoperatively, 4 patients required a revision surgery (3.1%) as shown in Table II. No factors were predictive of need for revision (Table III).

There were no episode-of-care mortalities in the 90-day postoperative period. Finally, 29 patients requested additional narcotic pain medication (22.8%). Total incidence of surgical site infection was 1 (0.8%), DVT was 3 (2.4%), and PE was 2 (1.6%).

Patient-reported outcomes and ROM

At the 3-month clinic visit, VAS was 1.7 ± 1.9, significantly decreased compared to the preoperative VAS of 8.2 ± 6.2 by an average of 6.5 points (P < .001); 92% of patients met the minimal clinically important difference (MCID) for VAS at this time³¹ (Table IV ). The 3-month SSV was 76 ± 18%, significantly improved from the preoperative SSV of 37 ± 22% by an average of 39% (P < .001); 70% of patients met the MCID for SSV at this time³¹ (Table IV). The 3-month ASES was 72 ± 15, significantly increased from the preoperative ASES of 35 ± 15 by an average of 37 points (P < .001); 91% of the patients met the MCID for ASES at this time³¹ (Table IV).

Table IV.

Comparison between outcomes preoperatively vs. 3-months postoperatively after planned outpatient shoulder arthroplasty

Outcome	Preoperative (n = 127)	3-Month (n = 127)	% Achieving MCID	P value
VAS	8.2 ± 6.2	1.7 ± 1.9	92	<.001
SSV (%)	37 ± 22	76 ± 18	70	<.001
ASES	35 ± 15 (n = 81)	72 ± 15 (n = 81)	91	<.001
ROM FF (deg)	105 ± 38	129 ± 30	N/A	<.001
ROM ER (deg)	26 ± 15	34 ± 15	N/A	<.001

VAS, visual analog scale; SSV, subjective shoulder value; ASES, American Shoulder and Elbow Surgeons score; ROM, range of motion; FF, forward flexion; ER, external rotation; deg, degrees; MCID, minimal clinically important difference.

Significance set at P value < .05 (bold).

FF at 3-months was 129° ± 30°, significantly improved from the preoperative baseline of 105° ± 38° (P < .001) (Table IV); 55% of the patients met the MCID for FF ROM at this time.³⁰ ER at 3 months was 34° ± 15°, significantly improved from the preoperative baseline of 26° ± 15° (P < .001) (Table IV); 59% of the patients met the MCID for ER ROM at this time.³⁰

Discussion

The COVID-19 pandemic has placed an increased burden on health care systems in the United States and elsewhere, limiting inpatient resources. As a result, more shoulder arthroplasties are being performed as outpatient procedures, with a subsequent loosening of patient selection criteria.^{23 , 27}

To our knowledge, very few studies have assessed clinical outcomes of outpatient shoulder arthroplasty without a preoperative patient selection criterion based on patient comorbidities. All the patients in this study were planned for outpatient arthroplasty regardless of preoperative comorbidities. The results of this study indicate that outpatient shoulder arthroplasty is a safe procedure with excellent 90-day outcomes and low rates of readmission regardless of comorbidities. We found that no preoperative comorbidities are predictive of unplanned ED/clinic visit, hospital readmission, and revisions within the first 90 days postoperatively. However, patients who are older, use tobacco, and have a preoperative ASA physical status of 3 as opposed to 2 were more likely to require overnight hospital stay postoperatively. Importantly, we found that CCI, BMI, diabetes, COPD, history of MI, and history of DVT/PE were not associated with complications in the 90-day postoperative period.

The incidence of unplanned overnight stay after planned outpatient arthroplasty in this cohort was 11.8%. Most commonly, this was due to inadequate pain control and postoperative hypoxia, suggesting that proper pain counseling and postoperative spirometry should be incorporated routinely for outpatient shoulder arthroplasty patients. However, the overall rate of unplanned admission for both pain and hypoxia remained relatively low, and all but 1 patient in our cohort were discharged on postoperative day 1. The incidence of 90-day unplanned ED/clinic visit and hospital readmission in this cohort was 13.4% and 5.5%, respectively. Most common indications for ED/clinic visits included trauma to the operative extremity, DVT/PE, and postoperative swelling (not due to DVT) that resolved. Thus, outpatient arthroplasty patients should be provided appropriate anticoagulation based on preoperative patient risk and should also be given preoperative counseling about routine swelling not indicative of pathology. Hospital readmissions were largely due to PE, sepsis, pneumonia, and surgical site infection, further emphasizing the need for proper anticoagulation and postoperative spirometry. The 90-day revision rate was 3.1% and was largely driven by dislocation/fracture after falls or trauma. Of all these complications, only 2 ED visits and 1 hospital readmission occurred on postoperative day 1, and therefore, most may not have been prevented by default inpatient stay.

Fournier et al outlined a decision tree for selecting patients for outpatient shoulder arthroplasty. They recommend inpatient surgery for patients over age 70, severe pulmonary comorbidities, and cardiac comorbidities. Our study found that for those above the age of 70, the odds of requiring overnight stay were nearly 37 times that of the odds for patients below 70. Likewise, the odds of 90-day unplanned ED/clinic visit for those above age 70 were nearly 8 times that of patients below 70. Age as a risk factor for worse outcomes after outpatient shoulder arthroplasty is also outlined by Willenbring et al, which found that each 1-year increase in age above 65 years increased the odds of postoperative complications by 14%.³³ Additionally, our study found that tobacco users and those with an ASA score of 3 as opposed to 2 also have higher odds of unplanned overnight stay by 13 and 14 times, respectively. These findings align with those of Biron et al who used a machine learning algorithm to determine that patients above 70 and ASA score above 2 were likely to require longer hospital stay after shoulder arthroplasty.⁵ Unlike Fournier et al and Biron et al, however, our study found no association between history of COPD, myocardial infarction, or DVT/PE and risk of unplanned overnight stay, or any other 90-day complications, suggesting that these may play a less vital role in patient selection criteria. Our results suggest that age above 70, tobacco use, and higher ASA score may be more influential when creating an outpatient shoulder arthroplasty selection algorithm (rather than BMI, MI, COPD, and DVT/PE history). Although these patients are at increased risk of overnight stay, these may not serve as absolute contraindications for outpatient arthroplasty. Rather, patients with these risk factors may be counseled that there is a higher risk for unplanned overnight stay. Even for patients with unplanned overnight hospitalization, our study suggests safe discharge to home for most patients by postoperative day 1. In addition, most ED visits and readmissions occurred well beyond postoperative day 1 suggesting that the common default practice of inpatient overnight stay with next day discharge may have little influence on preventing ED evaluation or readmission.

Many prior studies assessing the safety and efficacy of outpatient shoulder arthroplasty excluded patients based on strict patient selection criteria, considered comorbidities like COPD, obstructive sleep apnea, obesity, history of cardiac stenting or cardiac surgery, and history of DVT/PE.¹⁴ As a result, the patient cohorts in these studies may have consisted of relatively healthy patients not reflective of the current demographics of patients undergoing outpatient shoulder arthroplasty in the COVID-19 era. A recent systematic review further concluded that many of the studies comparing outcomes of inpatient to outpatient shoulder arthroplasty are subject to selection bias, as those undergoing ambulatory surgery were more likely to be younger males with a lower BMI and ASA status compared to their matched inpatient cohorts.¹ Perhaps unsurprisingly, all but 2 studies^{2 , 15} in the pre-COVID era have found that patients undergoing outpatient shoulder arthroplasty have similar or lower readmission and complication rates compared to inpatient shoulder arthroplasty.^{3 , 7 , 8 , 20 , 24} Reported rates of 90-day ED visits for outpatient arthroplasty in these studies vary with rates between 2.4%-16.8%, which are similar to the incidence of 13.4% found in this study.^{4 , 19 , 21 , 24 , 33} Finally, the reported rates of 90-day readmissions for outpatient arthroplasty in these studies vary with rates between 2.85%-9.3%, which are similar to the incidence of 5.5% found in this study.^{3 , 8 , 19} The similar incidences in this study may suggest that preoperative patient selection algorithms may be less critical than once believed.

Regardless of preoperative comorbidities and postoperative complications, however, outpatient shoulder arthroplasty demonstrated excellent patient-reported outcomes and ROM at the 3-month postoperative visit. Pain scores, SSV, and ASES all significantly improved compared to preoperatively, and most patients met MCID for all 3 outcomes. ROM in FF and ER significantly improved compared to the preoperative baseline as well. These findings align with those in the literature assessing patient satisfaction rates after outpatient shoulder arthroplasty. Erickson et al found that ASES, VAS, and SANE scores improved significantly in a cohort of patients who underwent outpatient shoulder arthroplasty at 1 year.¹² Bean et al found a significant improvement in pain scores at 3 months by nearly 7 points, similar to this study with an improvement of 6.5 points.⁴ Furthermore, 97% of patients reported a good to excellent outpatient shoulder arthroplasty experience in a questionnaire by Leroux et al.²¹ Finally, Charles et al found that all 50 patients in their cohort who underwent outpatient shoulder arthroplasty had significant improvements in their ROM and functional scores.⁹ While their outcomes were at a mean of 9.3 months postoperatively, they had similar improvements in FF of about 25° (compared to 24° in this study) but greater improvements in ER of about 18° (compared to 8° in this study). At 9.3 months, their postoperative ASES and VAS scores were 80 and 1.5, which were similar to the 3-month outcomes of this study at 72 and 1.7, respectively.

This study is subject to the limitations of a retrospective study design without a comparison group of inpatient surgery patients. As such, there may be some inherent heterogeneity between patients that cannot be accounted for, including differences in physical therapy regimens due to our inability to standardize all patients to the same physical therapist. As an observational study, no inferences about causation can be made and the findings should be interpreted accordingly. Additionally, due to the relatively low number of revision surgeries, the study may have been underpowered to detect associations between hypothesized risk factors and this outcome. Furthermore, our primary outcomes, including readmissions and ED/clinic visits, may be under-reported if patients sought care outside our institution, although our system has a large conglomerate of hospitals over a large region that casts a wide net for patient care. While this study reflects complications in the 90-day period during the COVID-19 pandemic, it possible some patients may not have reported to the ED/clinic or hospital despite having complications. Finally, our cohort of patients underwent procedures at a high-volume surgery center by a single surgeon and thus, the results may not generalize well to other settings.

Conclusion

Outpatient shoulder arthroplasty is a safe procedure with excellent outcomes and low rates of readmissions and can be considered as the default plan for all patient undergoing shoulder arthroplasty. Patients who are above 70 years of age, use tobacco, and have ASA score of 3, however, may be less suitable for outpatient arthroplasty and should be counseled regarding the higher risk of unplanned overnight hospitalization.

Disclaimers

Funding: No funding was used for this study.

Conflicts of interest: Albert Lin is a paid consultant for Arthrex and Wright Medical. All the other authors, their immediate families, and any research foundation with which they are affiliated have not received any financial payments or other benefits from any commercial entity related to the subject of this article.