Abstract
Background
Routine blood tests following total shoulder arthroplasty (TSA) cost the British national health system over £65,000 in 2018 without definite evidence of their impact on patients’ outcomes. This study aimed to ascertain whether routine laboratory tests are necessary after elective TSA.
Methods
A retrospective review of electronic records was completed for 274 patients that underwent TSA in the past six years. In total, 216 patients were eligible for analysis. The primary outcome was intervention rate for abnormal postoperative blood tests. Secondary outcomes were hospital length of stay (LOS), and readmissions at 30 and 90 days.
Results
In total, 193 patients underwent 216 TSAs; 72% were female, and the mean cohort age was 78 ± 7.2 years. Overall, 136 patients (63%) had an abnormal postoperative blood test, however; only 8 (3.7%) required intervention. The average postoperative drop in haemoglobin (Hb) was 19g/l. In four patients (1.8%) Hb levels dropped to < 80g/l; however, only two (0.9%) were symptomatic and received a red blood cell transfusion. Six patients (2.8%) developed acute kidney injury and did not require haemofiltration. The mean LOS was 3.2 ± 2.9 days. Five patients (2.3%) were readmitted within 30 days and six patients (2.8%) within 90 days. Univariate analysis only showed association between patients with abnormal creatinine and LOS (p < 0.05), and all of these patients had raised preoperative creatinine levels.
Conclusions
Routine postoperative blood tests are not required as they did not change the clinical outcome in this cohort or the readmission rate, causing unnecessary costs. It is recommended to request postoperative blood tests for patients with raised preoperative creatinine levels.
Keywords: Total shoulder arthroplasty, Laboratory tests, Day-case arthroplasty
Introduction
Total shoulder arthroplasty (TSA) is a common procedure performed primarily for shoulder arthritis. Reverse total shoulder arthroplasty (rTSA) has allowed greater versatility in managing a broad range of shoulder pathology with an exponential rise in its popularity worldwide over recent years. The UK National Joint Registry has recorded over 37,000 primary TSAs since the start of data collection in April 2012. A total of 5,487 TSAs were performed in 2018 alone.1 As the number of arthroplasty procedures continues to rise, healthcare providers strive to achieve excellent outcomes, being mindful of cost burdens on the public health system.
It is common practice to perform routine postoperative haematological and biochemical investigations as part of the postoperative management of arthroplasty patients. At a cost of £12 per patient per set of routine laboratory tests,2 the amount spent nationally in 2018 performing routine postoperative laboratory investigations for TSAs was over £65,000. However, with advances in preoperative optimisation, such as correcting anaemia and perioperative improvements including the use of tranexamic acid, the need for these investigations for all patients undergoing arthroplasty procedures has been called into question.3–6
Following an extensive literature review, little evidence was found regarding the efficacy and cost-effectiveness of performing routine postoperative laboratory tests in patients undergoing elective TSAs.
The aim of this study is to determine whether the postoperative investigations altered the course of postoperative management of patients undergoing elective TSAs.
Methods
This is a retrospective study of all consecutive patients that underwent elective TSA, including revision TSA (revTSA) under the care of the senior surgeon (OWD) on an inpatient basis between February 2014 and January 2020 (6 years).
Prospectively collected data from the coding department of electronic hospital records produced a total of 274 TSAs. Of these, 24 were trauma, 18 had incomplete preoperative or postoperative laboratory analysis and 16 had incomplete electronic medical records. All of these were excluded, resulting in 216 cases for analysis (Figure 1). These were then grouped into three different cohorts, depending on the type of arthroplasty received (anatomical TSA (aTSA), rTSA or revision TSA) (Table 1).
Figure 1 .
Flowchart indicating exclusion criteria for patients included in the analysis
Table 1 .
Types of arthroplasties performed
| Type of arthroplasty performed | No. |
|---|---|
| Reverse total shoulder arthroplasty (rTSA) | 134 |
| Anatomical total shoulder arthroplasty (aTSA) | 64 |
| Revision arthroplasty (revTSA) | 18 |
The procedures were performed under general anaesthetic, with the patients positioned in the beach chair position. Prior to the skin incision, the site was infiltrated with local anaesthetic and diluted adrenaline. The access to the glenohumeral joint was gained through the deltopectoral approach, and the intraoperative estimated blood loss (EBL) was 300ml (100–400ml). All patients underwent the same rehabilitation programme initially, which was then modified as per the patients’ needs.
Variables tabulated and analysed included age, sex, American Society of Anaesthesia (ASA) score,7 postoperative laboratory tests including haemoglobin (Hb, g/l), serum urea and electrolytes (UEs), which included sodium (Na, mmol/l), potassium (K, mmol/l), urea (Ur, mmol/l) and creatinine (Cr, μmol/l), interventions (transfusion of red blood cells, intravenous (IV) fluids, as well as review of patients’ drug charts) to correct laboratory abnormalities, length of stay (LOS, days) and readmission within 30 and 90 days from discharge.
Following an international consensus statement, this study considered abnormal Hb for both men and women as <109g/l.8 If a patient had symptomatic severe postoperative anaemia (Hb < 80g/l), they received a red blood cell transfusion (RBCt).
Postoperative acute kidney injury (AKI) was defined as an increase of serum Cr ≥ 26µmol/l, or an increase of serum Cr of 1.5 times the baseline.9 AKI was assessed using an AKI management hospital protocol, which could include IV fluid administration. Although not directly affecting the primary outcome, this study also captured all other complications. These were treated as per current best practice guidelines.
The primary outcome was intervention rate for abnormal postoperative blood tests. Secondary outcomes were hospital LOS, as well as readmissions at 30 and 90 days.
Statistical analysis was performed using SPSS Statistics Software 25.0 (IBM, Armonk, NY, USA). The alpha value for comparisons was set at p < 0.05. A univariate analysis was conducted between variables and primary and secondary outcomes.
Results
The demographic characteristics of this cohort are listed in Table 2. Two hundred and sixteen TSAs were performed on 193 patients, of which the majority were female (72%), with an average age of 78 ± 7.2 years (Table 2). In this cohort, a third of the patients were ASA 3, with over 50% being ASA 2. One hundred and thirty-six (63%) patients had blood tests outside the normal range; however, only eight (3.7%) patients received an intervention to correct the abnormality.
Table 2 .
Demographics of the analysed cohort
| Variable | All TSAs (216) | rTSA (134, 62%) | aTSA (64, 29.6%) | revTSA (18, 8.4%) |
|---|---|---|---|---|
| Age, years | 78 ± 7.2 | 79.4 ± 6.1 | 72.3 ± 7.2 | 75.2 ± 6.5 |
| Sex, male:female (%) | 60:156 (28:72) | 33:101 (25,75) | 23:41 (36,64) | 4:14 (22,78) |
| ASA | 2.3 ± 0.6 | 2.4 ± 0.6 | 2 ± 0.6 | 2.2 ± 0.4 |
| 1 | 15 (6.9) | 6 (4.5) | 9 (14) | 0 (0) |
| 2 | 131 (60.8) | 76 (56.7) | 42 (65.7) | 13 (72.2) |
| 3 | 66 (30.5) | 48 (35.8) | 13 (20.3) | 5 (27.8) |
| 4 | 4 (1.8) | 4 (3) | 0 (0) | 0 (0) |
| Normal postoperative laboratory tests | 80 (37) | 39 (29.1) | 34 (53.1) | 7 (38.9) |
| Abnormal postoperative Hb | 78 (36) | 59 (44) | 12 (18.8) | 7 (38.9) |
| No transfusion | 76 (97.4) | 57 (96.6) | 12 (100) | 7 (100) |
| Transfusion | 2 (2.6) | 2 (3.4) | 0 (0) | 0 (0) |
| Abnormal postoperative UE | 60 (27.8) | 42 (31.3) | 12 (18.8) | 6 (33.3) |
| No intervention, no repeat bloods | 49 (81.7) | 33 (78.6) | 10 (83.3) | 6 (100) |
| No intervention, repeat bloods | 5 (8.3) | 3 (7.1) | 2 (16.6) | 0 (0) |
| Intervention | 6 (10) | 6 (14.3) | 0 (0) | 0 (0) |
| LOS | 3.2 ± 2.9 | 3.8 ± 3.7 | 2.4 ± 1.9 | 2.2 ± 1.1 |
| R30 | 5 | 4 (3) | 1 (1.5) | 0 (0) |
| R90 | 6 | 3 (2.2) | 3 (4.7) | 0 (0) |
Data represented as mean ± standard deviation or n, number and (%)
TSAs = total shoulder arthroplasty; rTSA = reverse total shoulder arthroplasty; aTSA = anatomical total shoulder arthroplasty; revTSA = revision total shoulder arthroplasty; ASA = American Society of Anaesthesia; Hb = haemoglobin; UE = urea and electrolytes; LOS = length of stay; R30 = readmission within 30 days; R90 = readmission within 90 days
The average postoperative drop in Hb was 19g/l (p < 0.05), with 94 (43.5%) patients having an abnormal postoperative Hb. However, only four (1.8%) patients developed severe postoperative anaemia (Hb < 80g/l),8 of which two (0.9%) received RBCt. One of the cases was a primary TSA, with an intraoperative EBL of 300ml, and the other one was a rev TSA with intraoperative EBL of 400ml. Only six (2.8%) patients developed postoperative AKI, all of whom received IV fluids until resolution of renal impairment.
The majority of postoperative Hb and UE abnormalities did not receive an intervention, and a letter was sent to their general practitioner for further outpatient evaluation. Therefore, routine postoperative laboratory tests did not alter patient management in 208 (96.3%) patients.
The mean LOS was 3.2 ± 2.9 days; however, this was increased because of excessive outliers, as shown by a median LOS of 2 days. The patients with abnormal postoperative Hb had an aggregate mean LOS of 4.4 ± 3.2 days; however, after removal of outliers, the LOS was 2.7 ± 1.5 days. Patients who developed postoperative AKI and received IV fluids as part of their management had a mean LOS of 4 days; however, this was not felt to be meaningful as this mean was only obtained from six cases of AKI.
Five (2.3%) patients were readmitted within 30 days. Two patients had oesophagitis, one had cholecystitis and two had urinary tract infections. Looking back at these patients’ postoperative blood tests, two had normal postoperative bloods, two had abnormal Hb (one of which received a RBCt postoperatively) and one had abnormal UEs with no intervention.
Six (2.8%) patients were readmitted within 90 days. Three had increasing confusion, two had a fall and one had upper gastrointestinal bleed. Two had normal postoperative bloods, three had abnormal Hb with no intervention and one had abnormal UEs with no intervention.
There was no statistical association between readmission and incidence of abnormal postoperative blood tests. Of the 11 patients that were readmitted at 30 and 90 days, 7 received rTSA, 4 received aTSA and none of the revTSA cohort were readmitted. Further investigation of the readmission cases revealed it was unlikely that the episode leading to readmission was related to their postoperative laboratory abnormality.
Univariate analysis did not show statistical association between age (p = 0.287), ASA (p = 0.289), postoperative Hb (p = 0.230) and LOS nor readmission at 30 or 90 days. However, there was statistical association following univariate analysis between raised Cr and LOS (p < 0.05).
Apart from the aforementioned biochemical and haematological complications, there was also one case of surgical site infection, another of urinary tract infection and another of hospital acquired pneumonia, all of which resolved with oral antibiotics. There were two cases of deep vein thrombosis treated with novel oral anticoagulants, and one case of immune thrombocytopenic purpura, which resolved after four days in the intensive care unit (Table 3).
Table 3 .
Summary of complications found in the analysed cohort
| Complication | n (%) |
|---|---|
| Anaemia | |
| Moderate (<109g/l) | 90 (41.8) |
| Severe (<80g/l) | 4 (1.8) |
| AKI | 6 (2.8) |
| Hyponatraemia (<134mmol/l) | 95 (44.1) |
| ITP | 1 (0.5) |
| SSI | 1 (0.5) |
| UTI | 1 (0.5) |
| HAP | 1 (0.5) |
| Lower limb DVT | 2 (1) |
Data represented as n = number and %; AKI = acute kidney injury; ITP = immune thrombocytopenic purpura; SSI = surgical site infection; UTI = urinary tract infection; HAP = hospital acquired pneumonia; DVT = deep vein thrombosis
Discussion
This is the first study showing the association between interventions performed for abnormal postoperative blood tests and LOS, as well as readmission at 30 and 90 days for patients that underwent elective aTSAs, rTSAs and revTSAs.
In this cohort, 136 (63%) patients had an abnormal postoperative blood test; however, the intervention rate to correct the abnormality was only 3.7%. This result is comparable to a retrospective study of 319 total knee arthroplasties, where 95.6% of laboratory results did not change the course of care, even though 89 patients had abnormal postoperative laboratory tests.3 The same conclusion was drawn from a similar study in total hip arthroplasties.4 In a study reviewing postoperative laboratory tests in uni-compartmental knee arthroplasties, a very low rate of laboratory-associated interventions was found, thus determining that performing routine laboratory investigations is not necessary and is cost ineffective.5
This study had a postoperative RBCt rate of 0.9%, which is significantly lower than similar studies. Anthony et al,10 Padegimas et al11 and Kandil et al12 reported transfusion rates of 4% to 6% following TSA. Paynter et al6 reported a 10.3% transfusion rate post rTSA. These differences between the available literature and this study could be multifactorial. The authors believe that having a preoperative haemoglobin optimisation protocol helped reduce the RBCt rate. At the centre where this study was conducted, if a patient was found to be anaemic at preoperative assessment (male Hb < 130g/l, female Hb < 120g/l), they would have haematinics and transferrin saturation investigations. If the patient had iron deficiency, they would receive either oral or IV iron replacement therapy and reassessment in six weeks. If the anaemia was not secondary to iron deficiency, the patient was referred to the haematologists for further assessment and management.
Furthermore, the routine use of perioperative tranexamic acid played a favourable role in reducing this rate. It could also be due to different thresholds used at different centres.
Higher age and ASA scores could be causative; however, this study had the highest average age and comparable ASA score. Another study reported rates of postoperative transfusion as high as 19.6%; however, 26% of their cohort were TSAs performed in patients with acute traumatic proximal humerus fractures.13 There is a varying degree of consensus regarding risk factors associated with higher RBC transfusion requirements following TSA procedures, for example, older age, low preoperative Hb level, a higher ASA score and higher Charlson comorbidity index (CCI).6,10–16 Most are non-modifiable risk factors; however, some would benefit from preoperative optimisation, as well as close postoperative monitoring.
The literature also identified some of the risk factors to developing AKI following total joint arthroplasty surgeries. These factors include high body mass index (BMI), vascular disease, hypertension, diabetes mellitus, congestive heart failure, chronic obstructive pulmonary disease and older age group.17–21 There is also a reported association between the use of angiotensin-converting enzyme inhibitors and angiotensin receptor blockers with a higher risk of postoperative AKI.22 Other studies report chronic kidney disease being a major risk factor for developing postoperative AKI.23–25 Although our study did not include these factors in its analysis, the authors reviewed the case notes of the six patients that developed AKI to assess their medical comorbidities. Three patients had hypertension, all of whom took angiotensin-converting enzyme inhibitors, and one had tablet-controlled diabetes. This group’s average BMI was 28.
A recent published study of primary reverse TSAs reported that 88% of patients had abnormal postoperative laboratory results, 12% of which required interventions.6 Despite the lower cohort age and the exclusion of revision and reverse TSAs for proximal humerus fractures, this study reported a higher rate of blood transfusion (10.3%) compared with our study (0.9%). This may be due to tranexamic acid not being used. In addition, there is no information available to indicate whether they performed preoperative optimisation. Furthermore, in this study there was an association between abnormal postoperative blood tests requiring interventions and a lower BMI, a higher CCI and a lower preoperative Hb.
The mean LOS was 3.2 ± 2.9 days; however, this was increased because of excessive outliers, as shown by a median LOS of 2 days. Following univariate analysis, the only association found was between postoperative Cr and LOS. Further analysis of the patients with a high postoperative Cr did not show any predictive factors. The average age of these patients was 77 ± 5.1 years, and the average ASA grade was 2.25. Only one patient of this subcohort had a high preoperative Cr. Therefore, this study could not find association between the studied variables and LOS, apart from Cr, which did not have any association with risk factors. Furthermore, univariate analysis did not find any statistically significant associations between abnormal postoperative blood tests and reattendance at 30 and 90 days.
The low intervention rate of 3.7% following postoperative blood investigations indicates that they add a cost and may delay patient discharge with no clear benefit to patient care. Stopping the routine postoperative blood tests, among other modifications, has opened the door for day-case TSA for elective cases with prior optimisation in the US.26,27 Brolin et al reported no significant difference between TSA performed as outpatient day cases and inpatient cases and argued that outpatient TSA could be a safe alternative to hospital admission in appropriately selected patients.28
This study had some limitations. The data for this cohort were reviewed retrospectively. This centre did not routinely make a note of the patients’ CCI, and therefore comparison with similar articles may be incomplete.
Conclusion
This study shows a low intervention rate for abnormal postoperative blood tests, where low postoperative haemoglobin did not show statistically significant association with LOS nor readmission at 30 or 90 days. The only statistically significant univariate association was found between raised postoperative creatinine and LOS, however no risk factors could be identified.
The authors do not recommend doing routine postoperative blood tests for all patients undergoing elective TSA, and rather advocate for preoperative surgical assessment and optimisation and requesting blood tests on a per patient basis. In addition, the use of perioperative tranexamic acid might play a key role in reducing blood loss, and thus transfusion rate. The authors also recommend that postoperative blood tests are requested for patients with raised preoperative creatinine levels.
References
- 1.National Joint Registry for England and Wales. 16th annual report. http://www.njrreports.org.uk/Portals/0/PDFdownloads/NJR%2016th%20Annual%20Report%202019.pdf (cited May 2020).
- 2.National Clinical Guideline Centre. Routine preoperative tests for elective surgery (cited May 2020).
- 3.Halawi MJ, Lyall V, Cote MP. Re-evaluating the utility of routine postoperative laboratory tests after primary total knee arthroplasty. J Clin Orthop Trauma 2020; 11(Suppl 2): S219–S222. 10.1016/j.jcot.2019.01.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Halawi MJ, Plourde JM, Cote MP. Routine postoperative laboratory tests are not necessary after primary total hip arthroplasty. J Arthroplasty 2019; 34: 538–541. 10.1016/j.arth.2018.11.037 [DOI] [PubMed] [Google Scholar]
- 5.Shaner JL, Karim AR, Casper Det al. Routine postoperative laboratory tests are unnecessary after partial knee arthroplasty. J Arthroplasty 2016; 31: 2764–2767. 10.1016/j.arth.2016.05.052 [DOI] [PubMed] [Google Scholar]
- 6.Paynter JW, Raley JA, Kyrkos JGet al. Routine postoperative laboratory tests are unnecessary after primary reverse shoulder arthroplasty. J Shoulder Elbow Surg 2020. pii: S1058-2746(20)30013-6. 10.1016/j.jse.2019.12.023 [DOI] [PubMed] [Google Scholar]
- 7.Saklad M. Grading of patients for surgical procedures. Anesthesiology 1941; 2: 281–284. 10.1097/00000542-194105000-00004 [DOI] [Google Scholar]
- 8.Muñoz M, Acheson AG, Bisbe Eet al. An international consensus statement on the management of postoperative anaemia after major surgical procedures. Anaesthesia 2018; 73: 1418–1431. 10.1111/anae.14358 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Bellomo R, Ronco C, Kellum JAet al. Acute renal failure – definition, outcome measures, animal models, fluid therapy and information technology needs: the second international consensus conference of the acute dialysis quality initiative (ADQI) group. Crit Care 2004; 8: R204–R212. 10.1186/cc2872 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Anthony CA, Westermann RW, Gao Yet al. What are risk factors for 30-day morbidity and transfusion in total shoulder arthroplasty? A review of 1922 cases. Clin Orthop Relat Res 2015; 473: 2099–2105. 10.1007/s11999-014-4107-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Padegimas EM, Clyde CT, Zmistowski BMet al. Risk factors for blood transfusion after shoulder arthroplasty. Bone Joint J 2016; 98-B: 224–228. 10.1302/0301-620X.98B2.36068 [DOI] [PubMed] [Google Scholar]
- 12.Kandil A, Griffin JW, Novicoff WMet al. Blood transfusion after total shoulder arthroplasty: which patients are at high risk? Int J Shoulder Surg 2016; 10: 72–77. 10.4103/0973-6042.180719 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Malcherczyk D, Hack J, Klasan Aet al. Differences in total blood loss and transfusion rate between different indications for shoulder arthroplasty. Int Orthop 2019; 43: 653–658. 10.1007/s00264-018-4047-z [DOI] [PubMed] [Google Scholar]
- 14.Gruson KI, Accousti KJ, Parsons BOet al. Transfusion after shoulder arthroplasty: an analysis of rates and risk factors. J Shoulder Elb Surg 2009; 18: 225–230. 10.1016/j.jse.2008.08.005 [DOI] [PubMed] [Google Scholar]
- 15.Hardy JC, Hung M, Snow BJet al. Blood transfusion associated with shoulder arthroplasty. J Shoulder Elb Surg 2013; 22: 233–239. 10.1016/j.jse.2012.04.013 [DOI] [PubMed] [Google Scholar]
- 16.Schumer RA, Chae JS, Markert RJet al. Predicting transfusion in shoulder arthroplasty. J Shoulder Elbow Surg 2010; 19: 91–96. 10.1016/j.jse.2009.05.001 [DOI] [PubMed] [Google Scholar]
- 17.Weingarten TN, Gurrieri C, Jarett PDet al. Acute kidney injury following total joint arthroplasty: retrospective analysis. Can J Anesth 2012; 59: 1111–1118. 10.1007/s12630-012-9797-2 [DOI] [PubMed] [Google Scholar]
- 18.Kimmel LA, Wilson S, Janardan JDet al. Incidence of acute kidney injury following total joint arthroplasty: a retrospective review by RIFLE criteria. Clin Kidney J 2014; 7: 546–551. 10.1093/ckj/sfu108 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Hassan BK, Sahlstrom A, Dessau RBC. Risk factors for renal dysfunction after total hip joint replacement; a retrospective cohort study. J Orthop Surg Res 2015; 10: 158. 10.1186/s13018-015-0299-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Jafari SM, Huang R, Joshi Aet al. Renal impairment following total joint arthroplasty. Who is at risk? J Arthroplasty 2010; 25(Suppl 6): 49–53. 10.1016/j.arth.2010.04.008 [DOI] [PubMed] [Google Scholar]
- 21.Bell S, Dekker FW, Vadiveloo Tet al. Risk of postoperative acute kidney injury in patients undergoing orthopaedic surgery-development and validation of a risk score and effect of acute kidney injury on survival: observational cohort study. BMJ 2015; 351. 10.1136/bmj.h5639 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Nielson E, Hennrikus E, Lehman Eet al. Angiotensin axis blockade, hypotension, and acute kidney injury in elective major orthopedic surgery. J Hosp Med 2014; 9: 283–288. 10.1002/jhm.2155 [DOI] [PubMed] [Google Scholar]
- 23.Nowicka A, Selvaraj T. Incidence of acute kidney injury after elective lower limb arthroplasty. J Clin Anesth 2016; 34: 520–523. 10.1016/j.jclinane.2016.06.010 [DOI] [PubMed] [Google Scholar]
- 24.Perregaard H, Damholt MB, Solgaard Set al. Renal function after elective total hip replacement. Acta Orthop 2016; 87: 235–238. 10.3109/17453674.2016.1155130 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Kildow BJ, Karas V, Howell Eet al. The utility of basic metabolic panel tests after total joint arthroplasty. J Arthroplasty 2018; 33: 2752–2758. 10.1016/j.arth.2018.05.003 [DOI] [PubMed] [Google Scholar]
- 26.Leroux TS, Basques BA, Frank RMet al. Outpatient total shoulder arthroplasty: a population-based study comparing adverse event and readmission rates to inpatient total shoulder arthroplasty. J Shoulder Elbow Surg 2016; 25: 1780–1786. 10.1016/j.jse.2016.04.006 [DOI] [PubMed] [Google Scholar]
- 27.Basques BA, Erickson BJ, Leroux Tet al. Comparative outcomes of outpatient and inpatient total shoulder arthroplasty: an analysis of the Medicare dataset. Bone Joint J 2017; 99-B: 934–938. 10.1302/0301-620X.99B7.BJJ-2016-0976.R1 [DOI] [PubMed] [Google Scholar]
- 28.Brolin TJ, Mulligan RP, Azar FMet al. Outpatient total shoulder arthroplasty in an ambulatory surgery centre is a safe alternative to inpatient total shoulder arthroplasty in a hospital: a matched cohort study. J Shoulder Elbow Surg 2016. 10.1016/j.jse.2016.07.011 [DOI] [PubMed] [Google Scholar]