Risk factors for and timing of adverse events after revision total shoulder arthroplasty

Abstract

Background

Despite increasing rates of revision total shoulder arthroplasty (RTSA), there is a paucity of literature on optimizing perioperative outcomes. The purposes of this study were to identify risk factors for unplanned readmission and perioperative complications following RTSA, risk-stratify patients based on these risk factors, and assess timing of complications.

Methods

Bivariate and multivariate analyses of risk factors were assessed on RTSA patients from the ACS-NSQIP database from 2011 to 2015. Patients were risk-stratified and timing of severe adverse events and cause of readmission were evaluated.

Results

Of 809 RTSA patients, 61 suffered a perioperative complication or readmission within 30 days of discharge. Multivariate analysis identified operative time, BMI > 40, infection etiology, high white blood cell count, and low hematocrit as significant independent risk factors for 30-day complications or readmission after RTSA (p ≤ 0.05). Having at least one significant risk factor was associated with 2.71 times risk of complication or readmission within 15 days compared to having no risk factors (p < 0.001). The majority of unplanned readmission, return to the operating room, open/deep wound infection, and sepsis/septic shock occurred within two weeks of RTSA.

Discussion

Patients at high risk of complications and readmission after RTSA should be identified and optimized preoperatively to improve outcomes and lower costs.

Introduction

Total shoulder arthroplasty (TSA) provides pain relief, improved function, and better quality of life for patients with end stage shoulder arthritis.¹ Several studies have shown that the incidence of primary TSA in the United States is on the rise with an increase of 192–322% over the past decade.^2,3 With an aging population and greater number of primary TSAs, the burden of revision will increase as well. Data from the Kaiser Permanente Shoulder Arthroplasty Registry (KPSAR) from 2005 to 2013⁴ revealed a 2.1% revision rate at 3-year follow-up for various etiologies including deep infection, rotator cuff tear, and glenoid component failure, while studies from other countries have demonstrated revision rates of 10% at 5-year follow-up and 22% by 10-year follow-up.^3,5 As physicians, payers, and policymakers focus on improving the value of complex procedures such as revision total shoulder arthroplasty (RTSA), it will be necessary to understand (1) predictors of revision after primary TSA, (2) risk factors associated with adverse events in RTSA patients, and (3) the timing of adverse events after RTSA.

Although prior literature has described complications leading to and predictors of revision after primary TSA, there has been very little discussion about specific risk factors for complications and unplanned readmission after RTSA.^5,6 In a cohort of 78 RTSAs, Dines et al.⁷ found that clinical outcomes of RTSA can be predicted based on the indication for the revision. He reported that component revision provides the best results while soft-tissue reconstruction provides worse results overall. Another study of 230 revisions of reverse TSA by Wagner et al.⁸ found that female gender, history of instability, and prior hemiarthroplasty increased risk for post-revision complications. Saltzman et al.⁹ compared complication rates for 111 primary and 26 RTSAs and found body mass index (BMI) to predict medical complications after primary and RTSA.

We build upon previous work using a large, nationally representative database to analyze perioperative complications after RTSA.^5–9 The objectives of this study were to use the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) database to (1) determine 30-day perioperative complication rates after RTSA, (2) assess risk factors for and timing of unplanned readmission and severe adverse events (SAEs) after RTSA, and (3) risk-stratify RTSA patients based on these risk factors.

Materials and methods

The ACS-NSQIP is a national surgical database that prospectively collects patient data from over 370 participating institutions. All data are validated with strict adherence guidelines including routine audits to ensure high-quality data. Data from medical records, operative reports, and patient interviews are collected up to 30 days post-operatively by trained clinical reviewers. Additionally, NSQIP provides patient demographics such as age, sex, race, smoking status, and functional status among others, as well as patient medical comorbidities including diabetes, cardiac, pulmonary, renal, cancer, and American Society of Anesthesiologists (ASA). Preoperative and intraoperative variables including lab values, days from admission to operation, operative time, type of anesthesia, and days from operation to discharge are also included. Complications occurring within 30 days of operation are tracked by NSQIP and were classified into the following categories for analysis: SAEs and unplanned readmission. SAEs included death, myocardial infarction, cerebrovascular accident, renal failure, pulmonary embolism, venous thromboembolism, sepsis, septic shock, unplanned intubation, paraplegia, deep wound infection, organ/space infection, and return to operating room. Minor adverse events included superficial wound infection, urinary tract infection, and pneumonia. Infectious complications including deep wound infection, superficial wound infection, organ/space infection, and sepsis or septic shock were also compiled for separate analysis. Though ACS-NSQIP data collection goes back to 2007, readmission was first captured in 2011 and cause for readmission in 2012. Because this study used previously de-identified data from ACS-NSQIP, the protocol was determined to be exempt from the institutional review board at our tertiary medical center.

We queried the ACS-NSQIP database to identify all patients that underwent RTSA from 2011 to 2015 using the common procedural terminology codes (CPT) for (1) RTSA (23473/23474) that were created in 2014 and (2) for TSA (23472) with removal of implant (23331/23332) before 2014 (online Appendix Table 1). Patients with incomplete data or post-operative ICD-9/10 diagnosis codes indicative of primary TSA etiology were excluded from the analysis. Etiology of RTSA was classified using post-operative ICD-9/10 diagnosis codes corresponding to: infection, mechanical failure/complication, fracture/dislocation, and other/unknown etiologies (online Appendix Table 2). In order to perform exploratory analysis comparing patient characteristics between primary and RTSA patients, we queried the ACS-NSQIP database to identify all patients that underwent primary anatomic or reverse TSA using the corresponding CPT code (23472).

Table 1.

Comparison of patient demographic and procedure characteristics among RTSA patients.

	No complication	Severe adverse event and/or unplanned readmission	p
	748 (100%)	61 (100%)
Age (mean)	66.6 (SD: 11.2)	69.6 (SD: 10.1)	0.31
Male gender	336 (44.9%)	23 (37.7%)	0.28
Dependent functional status	33 (4.4%)	4 (6.6%)	0.39
BMI > 40	64 (8.6%)	11 (18.0%)	0.01
History of smoking	127 (17.0%)	11 (18.0%)	0.83
History of diabetes	140 (18.7%)	7 (11.5%)	0.16
History of pulmonary disease	55 (7.4%)	8 (13.1%)	0.11
History of chronic heart failure	2 (0.27%)	2 (3.3%)	0.03
Hypertension	462 (61.8%)	42 (68.9%)	0.27
History of renal disease	4 (0.53%)	0 (0.0%)	0.99
Steroids for chronic condition	44 (5.9%)	2 (3.28%)	0.57
Bleeding-causing disorders	18 (2.4%)	2 (3.3%)	0.66
ASA class > 2	430 (57.5%)	33 (54.1%)	0.61
Regional anesthesia	38 (5.1%)	1 (1.6%)	0.35
Operative time (mean)	127.2 (SD: 65.4)	141.1 (SD: 61.8)	0.59
Hospital LOS (mean)	2.1 (SD: 1.5)	2.9 (SD: 2.5)	<0.001
Revision procedure etiology			0.18
Infection	41 (5.5%)	9 (15%)
Mechanical failure/complication	368 (49%)	28 (46%)
Fracture/dislocation	127 (17%)	10 (16%)
Other/unknown	212 (28%)	14 (23%)
Laboratory results within 90 days preop. (%)
Low WBC count (<4500/µL)	41 (5.5%)	6 (9.8%)	0.16
High WBC count (>10,000/µL)	79 (10.6%)	12 (19.7%)	0.03
Low hematocrit (<30%)	17 (2.3%)	4 (6.6%)	0.07
Low platelets (<150,000/µL)	40 (5.4%)	5 (8.2%)	0.35
High INR (>1.1)	30 (4.0%)	2 (3.3%)	0.99
Low sodium (<135 mEq/L)	56 (7.5%)	5 (8.2%)	0.84
High sodium (>145 mEq/L)	5 (0.67%)	0 (0.0%)	0.99
High creatinine (>1.3 mg/dL)	55 (7.4%)	4 (6.6%)	0.98
High blood urea nitrogen (>30 mg/dL)	25 (3.3%)	2 (3.3%)	0.99
High bilirubin (>1.9 mg/dL)	0 (0.0%)	0 (0.0%)	–
Low albumin (<3.4 g/dL)	20 (2.7%)	4 (6.6%)	0.10

WBC: white blood cell; LOS: length of stay; INR: international normalized ratio; BMI: body mass index; ASA class: American Society of Anesthesiology Classification System.

Table 2.

Risk factors for readmission or severe adverse events among RTSA patients.

	Unplanned readmission (n = 32)		Severe adverse event (n = 46)
Risk factors	Odds ratio (95% CI)	p value	Odds ratio (95% CI)	p
Age	1.01 (0.98–1.05)	0.46	1.02 (0.99–1.04)	0.27
Male gender	0.56 (0.26–1.19)	0.13	0.59 (0.31–1.11)	0.10
Dependent functional status	3.22 (1.07–6.72)	0.05	–	–
BMI > 40	2.37 (0.94–5.95)	0.06	1.83 (0.79–4.26)	0.15
History of smoking	1.38 (0.58–3.26)	0.46	1.20 (0.56–2.54)	0.64
History of diabetes	0.63 (0.22–1.83)	0.49	0.30 (0.09–1.01)	0.06
History of pulmonary disease*	1.74 (0.59–5.12)	0.30	2.27 (0.97–5.30)	0.05
History of chronic heart failure	8.32 (0.84–82.31)	0.15	5.63 (0.57–55.21)	0.21
Hypertension	1.35 (0.63–2.88)	0.44	1.14 (0.61–2.13)	0.67
History of renal disease	–	–	–	–
Steroids for chronic condition	–	–	0.74 (0.17–3.16)	0.99
Bleeding-causing disorders	1.29 (0.17–9.92)	0.56	0.87 (0.11–6.65)	0.99
ASA class > 2	0.96 (0.47–1.96)	0.91	0.73 (0.40–1.33)	0.31
General or regional anesthesia?	0.95 (0.93–0.97)	0.40	0.42 (0.05–3.16)	0.72
Operative time	1.00 (0.98–1.01)	0.08	1.00 (0.99–1.01)	0.10
Hospital LOS	1.15 (0.99–1.33)	0.08	1.19 (1.05–1.35)	0.01
Infection etiology	3.84 (1.50–9.83)	0.01	2.45 (1.01–6.13)	0.05
Fracture/dislocation etiologya	0.69 (0.24–2.00)	0.63	1.03 (0.47–2.27)	0.84
Other/unknown etiologya	0.85 (0.38–1.93)	0.71	0.61 (0.29–1.29)	0.24
Laboratory results within 90 days preop. (%)
Low WBC count (<4500/µL)	2.44 (0.81–7.27)	0.11	1.59 (0.55–4.65)	0.33
High WBC count (>10,000/µL)	1.49 (0.56–3.97)	0.42	2.68 (1.31–5.49)	0.01
Low hematocrit (<30%)	1.22 (0.16–9.39)	0.58	4.18 (1.35–12.97)	0.03
Low platelets (<150,000/µL)	2.56 (0.86–7.66)	0.10	1.20 (0.36–4.02)	0.74
High INR (>1.1)	0.78 (0.10–5.87)	0.99	0.52 (0.07–3.93)	0.99
Low sodium (<135 mEq/L)*	1.28 (0.38–4.34)	0.73	1.53 (0.59–4.05)	0.38
High sodium (>145 mEq/L)	0.99 (0.99–1.01)	0.99	0.99 (0.99–0.99)	0.99
High creatinine (>1.3 mg/dL)	1.33 (0.39–4.51)	0.50	0.27 (0.04–2.00)	0.24
High blood urea nitrogen (>30 mg/dL)	0.93 (0.12–7.89)	0.99	0.63 (0.08–4.75)	0.99
High bilirubin (>1.9 mg/dL)	–	–	–	–
Low albumin (<3.4 g/dL)	2.29 (0.51–10.18)	0.25	2.47 (0.71–8.59)	0.15

WBC: white blood cell; LOS: length of stay; INR: international normalized ratio; BMI: body mass index; ASA class: American Society of Anesthesiology Classification System.

^aRelative to osteoarthritis etiology.

^*Significance at p ≤ 0.05.

Statistical analysis was conducted using SAS (version 9.3) with a two-tailed alpha of 0.05. Bivariate analysis was conducted to compare demographics, comorbidities, intraoperative variables, and 30-day outcomes between RTSA patients who suffered 30-day unplanned readmission or SAE versus those who did not. Categorical analysis was conducted with chi-square and Fisher’s exact test where appropriate. Continuous variables were analyzed using Student t test or Mann–Whitney U test after testing for normality and equal variance. Multivariate logistic regression models only included predictors that yielded a p value of 0.20 or less from bivariate analysis. All variables were assessed for confounding and interaction where appropriate. Final models were assessed for goodness of fit using the Hosmer–Lemeshow test.

Results

We identified 809 RTSA patients, of which, 61 (7.5%) suffered a complication or readmission within 30-days of discharge. Those who suffered adverse events were more likely to have BMI > 40, chronic heart failure, high white blood cell (WBC) count, and longer length of stay compared to controls (p < 0.05, Table 1). In bivariate analysis, patients undergoing RTSA with infection etiology (relative to aseptic mechanical failure) were more likely to suffer SAEs (OR: 2.45, p = 0.05) and unplanned readmission (OR: 3.84, p = 0.01) (Table 2). Dependent functional status (OR: 3.22, p = 0.05) was also a risk factor for unplanned readmission, while history of chronic obstructive pulmonary disease (COPD) (OR: 2.27, p = 0.05), high WBC (OR: 2.68, p = 0.01), and low hematocrit (OR: 4.18, p = 0.03) were associated with increased complication risk.

Leading causes for 30-day complications (total 46) included unplanned return to the operating room (41.0%), organ/space infection (13.1%), and deep wound infection (11.5%) (Table 3). The most common etiologies for unplanned readmission were deep wound infection (13%) and infectious disease acquired in-hospital (7%) with 13 cases (21%) of unknown etiology.

Table 3.

Breakdown of severe adverse events and unplanned readmission among RTSA patients.

Severe adverse event or unplanned readmission	61 (100%)
Severe adverse eventa	46 (75.4%)
Deep wound infection	7 (11.5%)
Organ/space infection	8 (13.1%)
Wound dehiscence	3 (4.9%)
Thrombolic event (DVT/PE)	2 (3.3%)
Renal failure	1 (1.6%)
Stroke/CVA	1 (1.6%)
Sepsis	5 (8.2%)
Septic shock	1 (1.6%)
Unplanned return to operating room	25 (41%)
Death	2 (3.3%)
Unplanned readmissionb	32 (52.5%)
Medical	8 (13.1%)
Infectious disease acquired in-hospital  (including sepsis)c	4 (6.6%)
Pneumonia	1 (1.6%)
Pulmonary embolism	1 (1.6%)
Stevens-Johnson syndrome–toxic   epidermal necrolysis (ICD 10: L51.3)	1 (1.6%)
Abdominal pain	1 (1.6%)
Surgical	11 (18%)
Deep incisional SSI	8 (13.1%)
Superficial incisional SSI	1 (1.6%)
Hematoma	1 (1.6%)
Mechanical complication of prosthesis	1 (1.6%)
Etiology unknown	13 (21.3%)

DVT: deep vein thrombosis; PE: pulmonary embolism; CVA: cerebrovascular accident; ICD: International Statistical Classification of Diseases and Related Health Problems; SSI: surgical site infection.

^aIndicates unique number of patients with severe adverse event.

^bIndicates unique number of patients with unplanned readmission.

^cIncludes intestinal infection, sepsis, urinary tract infection.

Multivariate analysis controlling for patient characteristics and comorbidities identified operative time (OR: 1.01), BMI > 40 (OR: 2.74), infection etiology (OR: 2.76), high WBC count (OR: 3.58), and low hematocrit (OR: 5.24) as significant independent risk factors for 30-day SAEs or unplanned readmission (p ≤ 0.05) (Table 4). Using these factors, we stratified patients into those with 0 risk factors (593) and those with ≥1 risk factor (216). Having at least one significant risk factor was associated with 2.71 times risk of SAE or unplanned readmission within 15 days compared to controls (3.0% vs. 7.4% event rate, p = 0.01) (Table 5). There was no significant difference in the rate of adverse events within 16–30 days post-surgery among the two groups (1.2% vs. 2.8% event rate, p = 0.12).

Table 4.

Risk factors for readmission or severe adverse events among RTSA patients.

	30-Day unplanned readmission or severe adverse event
Risk factors	Odds ratio (95% CI)	p
Operative time	1.01 (1.00–1.01)	0.04
BMI > 40	2.74 (1.08–6.95)	0.03
High WBC count (>10,000/µL)	3.58 (1.66–7.71)	0.01
Low hematocrit (<30%)	5.24 (1.49–6.14)	0.01
Infection etiologya	2.76 (1.07–7.10)	0.04

Note: Factors included in multivariate analysis: male gender, BMI > 40, smoking, diabetes, COPD, hypertension, high WBC, low hematocrit, ASA class, infection etiology, and fracture/dislocation etiology.

BMI: body mass index; COPD: chronic obstructive pulmonary disease; WBC: white blood cell.

^aRelative to mechanical failure/complication etiology.

Table 5.

Comparison of rates of 30-day unplanned readmission and severe adverse events by timing of complication (number of days after RTSA).

Number of significant risk factorsa	Number of patients (100%)	30-Day unplanned readmissions or severe adverse event
		Total	0–15 days	Odds ratio (95% CI)b	p value	16–30 days	Odds ratio (95% CI)b	p
0 Risk factors	593	24 (4.1%)	17 (3.0%)	–	–	7 (1.2%)	–	–
≥1 Risk factor	216	22 (10%)	16 (7.4%)	2.71 (1.34–5.47)	0.01	6 (2.8%)	2.39 (0.79–7.20)	0.12

^aMultivariate analysis (Table 4) identified four significant (p < 0.05) risk factors for 30-day unplanned readmission: BMI > 40, high WBC count (>10,000/µL), low hematocrit (<30%), and infection etiology.

^bOdds ratio relative to patient population with 0 risk significant risk factors for 30-day unplanned readmission or severe adverse events.

Analysis of timing of adverse event types revealed that the majority of unplanned readmission (58%), return to the operating room (72%), open/deep wound infection (53%), and sepsis/septic shock (83%) occurred within the two weeks after RTSA (Table 6).

Table 6.

Comparison of timing of adverse events among RTSA patients.

Post-operative adverse event	Number of patients who suffered the SAE	0–7 days	8–15 days	16–22 days	23–30 days
Unplanned readmission	32 (100%)	37%	21%	28%	14%
SAE	46 (100%)	52%	15%	28%	5%
Unplanned return to OR	25 (100%)	65%	7%	26%	2%
Organ, deep, wound infectiona	18 (100%)	40%	13%	47%	0%
Sepsis/septic shock	6 (100%)	83%	0%	0%	17%
Thromboembolic event (DVT/PE)	2 (100%)	50%	50%	0%	0%
Death	2 (100%)	0%	50%	50%	0%

SAE: severe adverse event; DVT: deep vein thrombosis; PE: pulmonary embolism; OR: odds ratio.

^aIncludes organ/space infection, deep wound infection, and wound dehiscence.

In bivariate analysis comparing primary TSA (8949) and RTSA (809) patients, the latter were more likely to be older, functionally dependent, smokers, and ASA class 3/4 (p < 0.03, Table 7). RTSA patients were also more likely to have longer operative time, higher rates of high WBC, and low hematocrit (p < 0.03).

Table 7.

Comparison of patient demographic and procedure characteristics among primary and RTSA patients.

	Primary TSA	RTSA	p
	8949 (100%)	809 (100%)
Age (mean)	69.5 (SD: 9.8)	69.6 (SD: 10.1)	<0.001*
Male gender	3903 (44%)	359 (44%)	0.68
Dependent functional status	245 (2.7%)	37 (4.5%)	0.003*
BMI > 40	913 (10%)	75 (9.3%)	0.40
History of smoking	914 (10%)	138 (17%)	<0.001*
History of diabetes	1537 (17%)	147 (18%)	0.47
History of pulmonary disease	573 (6.4%)	63 (7.8%)	0.13
History of chronic heart failure	41 (0.5%)	4 (0.5%)	0.88
Hypertension	6013 (67%)	504 (62%)	0.01
History of renal disease	45 (0.5%)	4 (0.5%)	0.97
Steroids for chronic condition	427 (4.8%)	46 (5.7%)	0.25
Bleeding-causing disorders	262 (2.9%)	20 (2.5%)	0.46
ASA class > 2	4761 (53%)	463 (57%)	0.03*
Regional anesthesia	341 (3.8%)	39 (4.8%)	0.16
Operative time (mean)	112.9 (SD: 45.5)	141.1 (SD: 61.8)	<0.001*
Hospital LOS (mean)	2.1 (SD: 2.3)	2.9 (SD: 2.5)	0.21
Laboratory results within 90 days preop. (%)
Low WBC count (<4500/µL)	449 (5.0%)	47 (5.8%)	0.33
High WBC count (>10,000/µL)	710 (7.9%)	91 (11%)	0.001*
Low hematocrit (<30%)	139 (1.6%)	21 (2.6%)	0.03*
Low platelets (<150,000/µL)	462 (5.2%)	45 (5.6%)	0.62
High INR (>1.1)	396 (4.4%)	32 (4.0%)	0.53
Low sodium (<135 mEq/L)	574 (6.4%)	61 (7.5%)	0.21
High sodium (>145 mEq/L)	80 (0.9%)	5 (0.6%)	0.42
High creatinine (>1.3 mg/dL)	636 (7.1%)	59 (7.3%)	0.84
High blood urea nitrogen (>30 mg/dL)	417 (4.7%	27 (3.3%)	0.08
High bilirubin (>1.9 mg/dL)	20 (0.2%)	–	0.18
Low albumin (<3.4 g/dL)	207 (2.3%)	24 (3.0%)	0.24

BMI: body mass index; LOS: length of stay; INR: international normalized ratio; ASA class: American Society of Anesthesiology Classification System; TSA: total shoulder arthroplasty.

^*p ≤ 0.05.

Discussion

Total shoulder arthroplasty (TSA) is a successful treatment option to restore function and provide pain relief for patients with shoulder arthritis.¹ With an aging population and increased incidence of primary TSA, the volume of revision TSA (RTSA) will correspondingly increase.^2,3 Given this trend, we queried the ACS-NSQIP database to compile the largest RTSA cohort in the literature to examine short-term outcomes following RTSA. The study aimed to (1) identify predictors of revision, (2) identify risk factors associated with adverse events, and (3) assess the timing of unplanned readmission in order to understand and improve the value of RTSA. The findings from this analysis may inform perioperative management of RTSA patients and help to identify patients at high risk of complications and readmission.

Risk factors for adverse events and unplanned readmission

Risk factors for adverse events and readmission for primary TSA have been well characterized; however, this is less so for RTSA.^10–14 Given that complications after RTSA are more frequent, difficult to manage, and more resource intensive as compared to those after primary TSA,¹⁵ it is all the more critical to understand factors associated with 30-day readmission and complications in RTSA. Prior studies have focused on identifying risk factors for complications such as intraoperative humeral fracture, early dislocation, increased transfusion, infection, and nerve injury.¹⁵ Specifically, Wagner et al.⁸ analyzed 230 RTSA cases and found risk of intraoperative periprosthetic fractures was increased by female gender (OR: 2.41, p = 0.03), history of instability (OR: 2.65, p = 0.02), and prior hemiarthroplasty (OR: 2.34, p = 0.03). Two studies of 73 and 32 RTSA for glenoid loosening found that patients with a full-thickness rotator cuff tear at time of revision surgery had a higher risk of re-operation (p = 0.03), and identified instability, rotator cuff tears, and malunion of the greater tuberosity as risk factors for poor outcomes.^16,17 Ahmadi et al. analyzed 566 consecutive RTSA patients and found older age (OR per 10 years: 1.5, p = 0.002), diabetes (OR: 2.3, p = 0.01), low hemoglobin level (OR: 0.4, p < 0.001), and cardiac disease (OR: 2.7, p < 0.001) to be associated with need for blood transfusions due to blood loss.¹⁸ Lastly, in primary TSA, across a pooled cohort of 27,290 patients multiple authors have identified demographics (age, male gender) and clinical characteristics (ASA class, cardiac disease, inflammatory arthritis, and functional status) as predictors of 30- or 90-day readmissions following primary TSA.^19–22 Beyond the prior analyses by Wagner et al.⁸ and Ahmadi et al.,¹⁸ our analysis identified infection etiology, BMI > 40, high white blood cell (WBC) count, low hematocrit, and longer operative time as independent risk factors for SAEs or unplanned readmission.

Infection etiology

Other studies on RTSA have noted infection rates as high as 15.4% and found that infections in the early post-operative period (Type 1 infections¹⁵) can occur via intraoperative bacterial seeding of the implant or via direct inoculation through a wound that has not yet healed.^23,24 Once infection has been confirmed, the orthopedic surgeon and patient must decide among a spectrum of treatment strategies ranging from antibiotic suppression to tissue debridement without prostheses exchange, to prostheses exchange, to arthrodesis – all of which vary considerably in terms of risk, success rate, effect on long-term joint function/mobility, and cost of treatment.^24,25 Antibiotic suppression and debridement without prosthesis exchange have resulted in high failure rates (>50%),²⁶ while several authors have reported higher success rates with prosthesis of antibiotic-loaded acrylic cement (PROSTALAC) implant and two-stage shoulder implantation.^27,28 In our study, two of the nine patients who suffered readmissions due to wound infection had preoperative high WBC and infectious etiology for RTSA.

Obesity

Several studies have shown that morbidly obese patients undergoing upper extremity procedures have a higher risk for complications as compared to controls.^9,29–32 Potential strategies that have been implemented include the creation of a perioperative “surgical home” or partnering with a multidisciplinary obesity treatment center to optimize patients before and well-after surgery. As noted by Benotti et al.,³³ preoperative weight loss is associated with fewer complications after gastric bypass surgery. It is possible that such intentional weight loss would reduce complications after elective orthopedic surgery as well. Obese RTSA patients should be counseled appropriately before surgery of their increased risk for complications and of the potential to mitigate these risks via enrolling in a preoperative weight-loss program, when feasible.²⁹

High WBC count

Preoperative WBC count is a crude estimate of the patient’s baseline inflammatory state and has been shown to be predictive of in-hospital complications (e.g., bleeding, mortality) and 30-day readmissions across procedures from several surgical fields.^34–38 We sought to understand the relationship between infection etiology and high WBC count, both of which were identified as independent risk factors for 30-day readmission in our analysis. As noted by Eichinger and Galvin,³⁹ TSA patients with joint infection rarely present with elevated WBC, C-reactive protein, or erythrocyte sedimentation rate. This is consistent with our cohort in which only 8% of RTSAs with infection etiology also had an elevated WBC count prior to surgery. As such, although circulating WBC count is not a surrogate for infection etiology, our results suggest that shoulder surgeons should incorporate this marker of inflammation prior to surgery in risk models predicting 30-day readmission following RTSA.

Low hematocrit

Established protocols for the detection and management of anemia in orthopedic patients—such as identifying anemia for nutritional deficiencies, renal insufficiency, or chronic inflammation within 28 days of surgery, and correcting nutritional deficiencies or providing erythropoiesis-stimulating agents when appropriate—have been shown to reduce post-operative morbidity and mortality in orthopedic patients.^40,41

Incidence of and cause for adverse events and unplanned readmission

Multiple studies have shown that complications after RTSA are more frequent and more difficult to manage as compared to those after primary TSA.²⁵ In a comparative analysis of 111 primary TSA and 26 RTSAs, Saltzman et al.⁹ found RTSAs to have higher rates of 90-day surgical complications (62% vs. 18%, p < 0.001), but similar rates of 90-day medical complications (13% vs. 15%, p = 0.17) such as myocardial infraction or thromboembolic event. The leading causes for surgical complications in the RTSA cohort from their analysis were dislocation, superficial infection, hematoma, and nerve injury. Our analysis of ACS-NSQIP revealed higher rates of 30-day SAEs (5.7% vs. 2.5%, p < 0.001) and unplanned readmission (4.0% vs. 2.7%, 0.05) among revision versus primary TSA patients. Similar to Saltzman et al.,⁹ we observed higher rates of surgical SAEs such as wound infection (2.2% vs. 0.2%, p < 0.001), sepsis (0.7% vs. 0.2%, p = 0.02), and unplanned return to the operating room (3.0% vs. 1.0%, p < 0.001) in revision versus primary TSA. We also observed a greater proportion of readmissions driven by surgical site infection (28% vs. 5.1%, p < 0.001) in the RTSA cohort.

Risk stratification and timing of adverse events (Tables 4 to 6)

After stratifying patients into those with 0 or ≥1 risk factors for adverse events (described above), we found the latter to have a 2.74 times odds of SAEs/unplanned readmission within 15 days post-surgery compared to controls. By 16–30 days post-surgery, there was no difference in the incidence of adverse events between the two groups. Our results demonstrate that the first 14 days after surgery is a crucial period for high-risk patients where greater care surveillance and resources help mitigate preventable adverse events/readmission. In particular, 58% of unplanned readmissions and 68% of SAEs occurred during this period. This is consistent with findings in primary TSA as Belmont et al.⁴² found that 55% of all hospital readmissions occurred by 14 days post-surgery.

While identifying high-risk patients along with the critical time period is an initial step, understanding which adverse events are most likely to occur is required to inform post-acute care interventions. In our cohort, 52% of deep wound infection, 83% of sepsis/septic shock, and 100% of thromboembolic cases occurred within 14 days post-surgery. Precautions during and after surgery to prevent infection include: preoperative site preparation, debridement of the surgical site, prophylactic antibiotics, and sampling of synovial fluid prior to surgery.⁴³ Aggressive prophylaxis for thromboembolism can be provided in the two-week period after RTSA with special attention to asymptomatic DVT that can occur in the lower extremities.⁴⁴

Factors associated with RTSA

Using the Norwegian Arthroplasty Register, Fevang et al.⁵ analyzed 69 reverse TSAs and found that risk of revision was lower in women (OR: 0.26) and that the most common cause for RTSA was aseptic loosening. Using the Mayo Clinic Joint Registry, Singh et al.⁴⁵ analyzed 212 RTSAs and found that male gender (OR: 1.72) and rotator cuff disease (OR: 4.71) were associated with greater risk for RTSA (p < 0.05). We similarly observed mechanical failure (without infection) to be the most common cause of RTSA; however, we found no difference in the proportion of females between primary and RTSA cohorts (56% for both). Using the Medicare PearlDiver database, Werner et al.²³ identified 2059 patients who underwent early (within 1 year) RTSA. Age > 65, smoking, and morbid obesity were all associated with 1.4–1.9 times odds of early RTSA (p < 0.001 for all); however, female gender was not protective. They also found that dislocation etiology (OR: 3.2) and stiffness (OR: 1.4) were associated with early revision (p < 0.04 for all). Our analysis also found RTSA patients to be older and more likely to be smokers. Beyond these studies, we found RTSA patients were more likely to be functionally dependent and have ASA class >2, high WBC count (within 90 days of surgery), and low hematocrit (within 90 days of surgery) as compared to primary TSA patients. While most of the factors mentioned above such as functional dependence, male gender, and age are non-modifiable, patient factors such as smoking and morbid obesity should be addressed before primary TSA and continually monitored in order to minimize revision risk. As noted by Werner et al.,²³ some common etiologies for RTSA such as mechanical failure/instability or dislocation can be driven by provider-related factors such as improper implant positioning, component malpositioning, or failure of subscapularis repair after primary TSA.^10,12

Limitations

There are several limitations to our analysis worth noting. Most importantly, ACS-NSQIP only captures adverse events and unplanned readmission with 30 days of RTSA as opposed to the ideal 90-day or 1-year post-procedure follow-up period. As noted by Wagner et al.,⁸ this is one advantage of claims-based databases such as PearlDiver and MEDPAR that can provide a longitudinal view of RTSA patients; however, multiple studies have shown that claims often misreport patient comorbidities, inaccurately capture procedure etiology, and underreport clinical complications as compared to data derived from electronic health records (such as ACS-NSQIP).¹³ Secondly, since NSQIP is based off electronic health records data collected by the reporting hospital/health system, readmissions or adverse events treated at facilities outside the health system would not be captured. Other valuable data, such as anatomical variables, case notes, radiographic measurements, and post-operative patient-reported outcomes are also not collected in this database. A third important limitation of NSQIP is that we could not differentiate whether our RTSA patients initially underwent total or reverse total shoulder arthroplasty, and whether the initial TSA procedure involved unconstrained versus semiconstrained/constrained TSA which has been shown to have higher rates of aseptic loosening, instability, infection, and periprosthetic fractures.¹⁴ Lastly, for 20% of 30-day unplanned readmissions the etiology was unknown or not reported, which is another limitation of using the ACS-NSQIP database.

Conclusion

Patients at high risk of complications and readmission after RTSA should be identified and optimized preoperatively to improve outcomes and lower costs. Particular attention should be paid to those with BMI > 40, infection etiology, high WBC count, and low hematocrit. High-risk RTSA patients face a greater risk of adverse events and readmission within two weeks after surgery and therefore may benefit from aggressive post-acute care surveillance and management.

Supplementary Material

Supplementary material is available at: http://journals.sagepub.com/doi/suppl/10.1177/1758573218780517

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.