The number of shoulder arthroplasties performed in the United States has increased substantially over the past few decades1. Concomitant with this growth in shoulder arthroplasty procedures is an expected increase in the number of revision procedures and complications such as infection. Shoulder periprosthetic joint infection (PJI) has been reported to occur in about 1% to 2% of cases and not only substantially impacts patient quality of life but also dramatically increases costs to the health-care system2–5. Because the majority of cases of shoulder PJI are caused by Cutibacterium, and because the presentation of a Cutibacterium PJI is typically subtle and delayed, it seems likely that the number of PJI cases is greater than what has been reported6.
Given the rising number of shoulder arthroplasties and the potential societal cost of revision surgeries related to infection, the identification of risk factors for shoulder PJI and the identification of optimal prophylactic measures is of high interest. A number of prophylactic measures can be utilized, including preoperative antibiotics, preoperative topical antimicrobial treatments, antibiotic or Betadine (povidone-iodine) lavage, topical administration of antibiotics, and mechanical prophylaxis. In addition, other means of decreasing bleeding and hematoma formation, such as the use of tranexamic acid (TXA), topical thrombin, and drains, are instituted by some surgeons to decrease the risk of wound drainage and the associated risk of infection.
Currently, much of the available literature is inconclusive regarding the efficacy of many of these infection prophylactic measures7. Therefore, the objective of this study was to study patient characteristics, procedural characteristics, and various infection prophylactic measures and their effects on the risk of shoulder PJI after primary arthroplasty. We hypothesized that specific patient characteristics and infection prophylactic measures would modify the risk of shoulder PJI.
Materials and Methods
This study was approved by the University of Washington institutional review board. This was a retrospective study of a longitudinally maintained institutional database of 1,220 adult patients who underwent primary shoulder arthroplasty between August 2010 and April 2018. Patients without at least 2 years of follow-up were excluded.
Patient Demographics
Patient demographics including age, body mass index (BMI), sex, and marital status as well as smoking status, diabetes status, narcotic and alcohol use, and insurance status were recorded. In addition, a history of ipsilateral shoulder surgery, the preoperative patient-reported Simple Shoulder Test (SST) score, and the American Society of Anesthesiologists (ASA) score were routinely collected when enrolling each patient into the database.
Procedural Characteristics
Procedural characteristics including the indication for surgery, type of surgery, and the use of a postoperative drain were either routinely entered into the database or collected from the electronic medical record.
Infection Prophylaxis
The use of intraoperative intravenous antibiotics, topical antibiotics, antibiotic-containing irrigation solution, topical povidone-iodine, vancomycin-containing graft, and antibiotic-containing cement, were variables collected from the operating room records and electronic medical record.
Hemostatic Measures
The use of TXA, local anesthetic with epinephrine, and thrombin were additional variables collected from the operating room records and electronic medical record. Thrombin was used in cases in which there was localized deep bleeding from bone or tissue bleeding that could not be adequately addressed with electrocautery.
Outcome of Interest
Patients in the database were routinely followed on an annual basis. Any reoperation or revision surgery was logged in the database, and at the time of reoperation, multiple deep-tissue culture samples were taken regardless of the reason for revision. The primary outcome of interest was the development of shoulder PJI. In this study, the definition of shoulder PJI was a modification of the 2018 International Consensus Meeting (ICM) definition of a definite shoulder PJI7. We defined shoulder PJI as the presence of any of the following factors: (1) a sinus tract, (2) gross intra-articular pus, or (3) reoperation with ≥2 positive cultures demonstrating phenotypically identical organisms.
Statistical Analysis
Univariate analyses were conducted using chi-square tests, Fisher exact tests, t tests, and Mann-Whitney U tests, as appropriate. A multivariable logistic-regression analysis was used to determine independent predictors of shoulder PJI. Covariates were selected for inclusion if they had a p value of <0.05 on the univariate analysis. These include age, sex, alcohol use, and procedure type (reverse total shoulder arthroplasty [TSA], ream and run, hemiarthroplasty, anatomic TSA). The final multivariable logistic-regression model was validated using backward stepwise elimination. The level of significance was set at p < 0.05. Statistical analyses were performed with the use of SPSS (version 25.0; IBM).
Source of Funding
The authors acknowledge support for this study from the University of Washington’s Douglas T. Harryman II/Depuy Endowed Chair for Shoulder Research and The Rick and Anne Matsen Endowed Professorship for Shoulder Research.
Results
A total of 1,220 patients were enrolled in the database during the study period, and minimum 2-year follow-up was obtained for 998 (82%). Three of the 1,220 patients were excluded as they had a reoperation at an outside hospital for which culture data were not available. Of the 998 included patients, no patient developed a sinus tract, 2 (0.2%) of the patients presented in the postoperative period with gross intra-articular pus, and 18 (1.8%) underwent reoperation with ≥2 positive cultures with phenotypically identical organisms but without a sinus tract or gross intra-articular pus. Of these 20 total patients, 2 had failure of a subscapularis repair, 1 had mechanical symptoms, and the remaining 17 had a reoperation for persistent pain and stiffness.
The median age of the cohort was 65 years (interquartile range, 57 to 72 years), and 64% of the patients were male. Five percent of the patients were current smokers, and 9% had a diagnosis of diabetes. The majority of patients were ASA class 2 (62%) or class 3 (28%). Osteoarthritis was the most common diagnosis (64%), followed by capsulorrhaphy arthropathy (12%) and cuff tear arthropathy (12%). Thirty-four percent of the patients had prior surgery on the ipsilateral shoulder. Forty-two percent underwent anatomic TSA, 37% underwent ream-and-run arthroplasty, 12% underwent hemiarthroplasty, and 10% underwent reverse TSA.
Of the 20 culture-positive reoperations, Cutibacterium was the causative organism in 19 (95%). The remaining reoperation was culture-positive for Streptococcus mitis. One of the reoperations positive for Cutibacterium was polymicrobial and had the presence of coagulase-negative Staphylococcus (CoNS) and Ralstonia picketti.
Patient Demographics and Procedural Characteristics
On univariate analysis, patients in the shoulder PJI group were younger (median, 56 versus 65 years; p < 0.001) and were more likely to be male (90% versus 63%; p = 0.014). They were also more likely to consume alcohol (85% versus 64%; p = 0.048), to have commercial insurance (55% versus 39%; p = 0.003), and to have a higher proportion of ream-and-run and hemiarthroplasty procedures (p = 0.005). BMI, marital status, smoking status, diabetes, preoperative SST score, prior ipsilateral shoulder surgery, and indication for surgery were not associated with a higher risk of culture-positive revision (p > 0.05) (Table I).
TABLE I -
Comparison of Patient Demographics, Preoperative Characteristics, and Procedure Characteristics in Patients with PJI at Revision Versus Controls (No PJI)
*
| Variable |
All |
Controls |
PJI at Revision |
P Value |
| No. |
998 |
978 |
20 |
— |
| Age (yr)
|
65.1 [56.9-71.7] |
65.3 [57.1-71.8] |
55.8 [49.9-62.4] |
<0.001
|
| BMI (kg/m
2
)
|
28.6 [25.6-32.7] |
28.6 [25.7-32.7] |
27.8 [25.3-32.2] |
0.715 |
| Sex |
|
|
|
0.014
|
|  Male |
637 (63.8%) |
619 (63.3%) |
18 (90.0%) |
|
|  Female |
361 (36.2%) |
359 (36.7%) |
2 (10.0%) |
|
| Marital status |
|
|
|
0.975 |
|  Married |
702 (70.3%) |
688 (70.3%) |
14 (70.0%) |
|
|  Other (e.g., domestic partner, single, divorced) |
296 (29.7%) |
290 (29.7%) |
6 (30.0%) |
|
| Current smoker |
53 (5.3%) |
52 (5.3%) |
1 (5.0%) |
0.949 |
| Diabetes |
90 (9.0%) |
89 (9.1%) |
1 (5.0%) |
0.526 |
| Narcotics use |
242 (24.2%) |
239 (24.4%) |
3 (15.0%) |
0.324 |
| Alcohol use |
638 (63.9%) |
621 (63.5%) |
17 (85.0%) |
0.048
|
| Insurance status |
|
|
|
0.003
|
|  Medicare |
464 (46.5%) |
462 (47.2%) |
2 (10.0%) |
|
|  Medicaid |
46 (4.6%) |
45 (4.6%) |
1 (5.0%) |
|
|  Workers’ Compensation |
58 (5.8%) |
54 (5.5%) |
4 (20.0%) |
|
|  Other |
37 (3.7%) |
35 (3.6%) |
2 (10.0%) |
|
|  Commercial |
393 (39.4%) |
382 (39.1%) |
11 (55.0%) |
|
| Preoperative SST |
3 [1-6] |
3 [1-6] |
5 [2-6] |
0.130 |
| Prior ipsilateral shoulder surgery |
337 (33.8%) |
330 (33.7%) |
7 (35.0%) |
0.911 |
| ASA class |
|
|
|
0.070 |
|  1 |
90 (9.0%) |
85 (8.7%) |
5 (25.0%) |
|
|  2 |
622 (62.3%) |
613 (62.7%) |
9 (45.0%) |
|
|  3 |
274 (27.5%) |
268 (27.4%) |
6 (30.0%) |
|
|  4 |
12 (1.2%) |
12 (1.2%) |
0 (0.0%) |
|
| Indication for surgery |
|
|
|
0.833 |
|  Osteoarthritis |
636 (63.7%) |
624 (63.8%) |
12 (60.0%) |
|
|  Capsulorrhaphy arthropathy |
124 (12.4%) |
122 (12.5%) |
2 (10.0%) |
|
|  Cuff tear arthropathy |
120 (12.0%) |
118 (12.1%) |
2 (10.0%) |
|
|  Fracture/nonunion/PTA |
52 (5.2%) |
50 (5.1%) |
2 (10.0%) |
|
|  Other (ON/RA/chondrolysis) |
66 (6.6%) |
64 (6.5%) |
2 (10.0%) |
|
| Surgery type |
|
|
|
0.005
|
|  Ream and run |
368 (36.9%) |
354 (36.2%) |
14 (70.0%) |
|
|  Anatomic TSA |
415 (41.6%) |
414 (42.3%) |
1 (5.0%) |
|
|  Reverse TSA |
98 (9.8%) |
96 (9.8%) |
2 (10.0%) |
|
|  Hemiarthroplasty |
117 (11.7%) |
114 (11.7%) |
3 (15.0%) |
|
*The values are given as the number, with the percentage in parentheses, except for age, body mass index (BMI), and preoperative Simple Shoulder Test (SST), which are given as the median, with the interquartile range in square brackets. PJI = periprosthetic joint infection, BMI = body mass index, ASA = American Society of Anesthesiologists, PTA = posttraumatic arthritis, ON = osteonecrosis, RA = rheumatoid arthritis, and TSA = total shoulder arthroplasty.
On multivariable analysis, younger age (odds ratio [OR] per year, 0.95; 95% confidence interval [CI], 0.91 to 0.99; p = 0.014) was associated with shoulder PJI. Of the surgical types, reverse TSA (OR, 10.32; 95% CI, 0.92 to 116.33; p = 0.059), hemiarthroplasty (OR, 8.59; 95% CI, 0.86 to 85.50; p = 0.067), and ream-and-run arthroplasty (OR, 6.32; 95% CI, 0.75 to 52.98; p = 0.089) exhibited a marginal association. Sex and use of alcohol were not independently associated with the risk of culture-positive reoperation (p > 0.05) (Table II).
TABLE II -
Multivariable Logistic Regression to Determine Independent Predictors of PJI at Revision Versus Controls (No Presence of PJI)
*
| Variable |
Odds Ratio (95% CI) |
P Value |
| Age, per increase by 1 yr |
0.951 (0.914-0.990) |
0.014
|
| Male sex |
2.775 (0.545-14.135) |
0.219 |
| Use of alcohol |
2.864 (0.807-10.169) |
0.104 |
| Type of surgery, reference = anatomic TSA |
|
|
|  Ream and run |
6.318 (0.753-52.981) |
0.089 |
|  Reverse TSA |
10.321 (0.916-116.326) |
0.059 |
|  Hemiarthroplasty |
8.589 (0.863-85.500) |
0.067 |
*PJI = periprosthetic joint infection, CI = confidence interval, and TSA = total shoulder arthroplasty.
Infection Prophylaxis
On univariate analysis, the type of intraoperative antibiotic, the use of topical vancomycin, the use and type of antibiotic-containing irrigation solution, the use of vancomycin-containing cement, and the use of vancomycin-containing graft were not associated with a lower risk of shoulder PJI (p > 0.05) (Table III).
TABLE III -
Comparison of Antibiotic and Hemostatic Measures in Patients with PJI at Revision Versus Controls (No PJI)
*
| Variable |
All |
Controls |
PJI at Revision |
P Value |
| No. |
998 |
978 |
20 |
— |
| Use of thrombin |
153 (15.3%) |
149 (15.2%) |
4 (20.0%) |
0.558 |
| Use of tranexamic acid |
297 (29.8%) |
292 (29.9%) |
5 (25.0%) |
0.638 |
| Use of topical vancomycin |
539 (54.0%) |
531 (54.3%) |
8 (40.0%) |
0.204 |
| Use of antibiotic-containing irrigation solution |
|
|
|
0.904 |
|  None |
88 (8.8%) |
85 (8.7%) |
3 (15.0%) |
|
|  Vancomycin |
302 (30.3%) |
297 (30.4%) |
5 (25.0%) |
|
|  Cefazolin |
27 (2.7%) |
26 (2.7%) |
1 (5.0%) |
|
|  Ceftriaxone |
3 (0.3%) |
3 (0.3%) |
0 (0.0%) |
|
|  Combination of ≥1 of the above |
577 (57.8%) |
566 (57.9%) |
11 (55.0%) |
|
|  Other |
1 (0.1%) |
1 (0.1%) |
0 (0.0%) |
|
| Intraoperative intravenous antibiotic |
|
|
|
0.632 |
|  Cefazolin |
145 (14.5%) |
140 (14.3%) |
5 (25.0%) |
|
|  Clindamycin |
27 (2.7%) |
27 (2.8%) |
0 (0.0%) |
|
|  Vancomycin |
19 (1.9%) |
19 (1.9%) |
0 (0.0%) |
|
|  Vancomycin/ceftriaxone |
752 (75.4%) |
738 (75.5%) |
14 (70.0%) |
|
|  Combination of ≥1 of the above |
55 (5.5%) |
54 (5.5%) |
1 (5.0%) |
|
| Use of intraoperative vancomycin-containing graft |
675 (67.6%) |
660 (67.5%) |
15 (75.0%) |
0.477 |
| Use of vancomycin-containing cement |
84 (8.4%) |
84 (8.6%) |
0 (0%) |
0.171 |
| Use of local anesthetic |
|
|
|
0.112 |
|  None |
30 (3.0%) |
28 (2.9%) |
2 (10.0%) |
|
|  Bupivacaine without epinephrine |
51 (5.1%) |
51 (5.2%) |
0 (0%) |
|
|  Bupivacaine with epinephrine |
917 (91.9%) |
899 (91.9%) |
18 (90.0%) |
|
| Use of drain |
94 (9.4%) |
90 (9.2%) |
4 (20.0%) |
0.102 |
*The values are given as the number, with the percentage in parentheses. PJI = periprosthetic joint infection.
Hemostatic Measures
On univariate analysis, the use of TXA, the use and type of local anesthetic, and the use of thrombin were not associated with a lower risk of shoulder PJI (p > 0.05) (Table III). Placement of a postoperative drain had a nonsignificant association (20% versus 9%; p = 0.102).
Discussion
This study examined modifiable and nonmodifiable factors associated with a higher risk of shoulder PJI using a large shoulder arthroplasty database. The data suggest that younger patients and patients undergoing arthroplasty types other than anatomic TSA may benefit from counseling on the potentially higher risk of shoulder PJI. We did not identify modifiable factors such as infection prophylactic measures (e.g., intraoperative antibiotics, topical vancomycin) or hemostatic measures (e.g., TXA) that decreased the risk of shoulder PJI.
The type and use of perioperative antibiotics did not have a significant impact. The 2018 ICM strongly recommended intravenous antibiotics that cover gram-positive and gram-negative organisms commonly encountered in periprosthetic shoulder infections7, which include Cutibacterium acnes, Staphylococcus aureus, and other Staphylococcus species8. Traditionally, and consistent with the 2018 ICM, cefazolin is utilized for intravenous prophylaxis7. The findings of Marigi et al.9 support the use of cefazolin: a higher risk of PJI was found with vancomycin and clindamycin. Prior literature10,11, including a clinical study12, has suggested increased resistance of Cutibacterium to clindamycin. It is possible that our study was underpowered to detect any true differences in the effectiveness of antibiotic types.
Topical vancomycin is another modality often utilized around the time of shoulder arthroplasty to decrease the risk of PJI. Literature on the use of antibiotic-containing powders, intrawound agents, and irrigation solution for the shoulder is limited7. The evaluation of vancomycin powder at other body locations has suggested its safety and efficacy13–17. The data from this study did not demonstrate a decreased risk of PJI with vancomycin powder or antibiotic irrigation but may have been underpowered to do so.
Surgical types other than anatomic TSA demonstrated a marginal association with the risk for shoulder PJI. The risk of revision due to infection has been reported to be elevated in reverse arthroplasty compared with anatomic arthroplasty, particularly in males18. These other types of arthroplasty may be associated with an increased risk of shoulder PJI because of a more complex shoulder diagnosis or a diagnosis related to previous surgery (e.g., capsulorrhaphy arthropathy). In addition, surgeries performed for diagnoses other than glenohumeral arthritis and in patients who previously have undergone non-arthroplasty surgery may be associated with a higher revision rate, which increases the opportunity for culture-positive revision compared with anatomic TSA, which has the lowest rate of revision18–20.
This study had several limitations. First, there was a relatively small number of events of shoulder PJI, which may lend itself to instability of the multivariable analysis. Second, interventions that were consistently used throughout the study period could not be statistically analyzed. For example, the use of a chlorhexidine preparation solution prior to surgical incision did not change during the study time period, and therefore, no statistical analysis of this variable was possible. Third, there was also the risk of selection bias given that a surgeon may have instituted certain measures (e.g., topical vancomycin use) for specific patients who were deemed to be at higher risk for infection. The multivariable analysis helps to mitigate this bias but does not eliminate it. Fourth, this study was conducted at a tertiary care center, and therefore, the diagnoses, procedure types, and patient demographics may not be generalizable. Because of selection bias in the treatment of different types of shoulder pathology, they may not represent a typical shoulder arthroplasty practice. In particular, there were a relatively high number of ream-and-run procedures in this cohort. Fifth, revision surgeries done years after index arthroplasty can be culture-positive, and this study does not account for late, indolent infections.
Conclusions
Younger patients and patients undergoing procedures other than anatomic TSA may benefit from counseling regarding the potentially higher risk of shoulder PJI. The majority of culture-positive reoperations were the result of Cutibacterium species, but none of the infection prophylactic measures in this study were found to decrease the risk of shoulder PJI.
Note: The authors thank Susan DeBartolo (University of Washington, Department of Orthopaedics and Sports Medicine) for her editorial work on the manuscript.
References
1. Schwartz BE, Savin DD, Youderian AR, Mossad D, Goldberg BA. National trends and perioperative outcomes in primary and revision total shoulder arthroplasty: Trends in total shoulder arthroplasty. Int Orthop. 2015 Feb;39(2):271-6.
2. Gates S, Nguyen I, Del Core M, Nakonezny PA, Bradley H, Khazzam M. Incidence and predictors of positive intraoperative cultures in primary shoulder arthroplasty following prior ipsilateral shoulder surgery. JSES Int. 2020 Feb 21;4(2):366-71.
3. Kunutsor SK, Beswick AD, Peters TJ, Gooberman-Hill R, Whitehouse MR, Blom AW, Moore AJ. Health care needs and support for patients undergoing treatment for prosthetic joint infection following hip or knee arthroplasty: A systematic review. PLoS One. 2017 Jan 3;12(1):e0169068.
4. Moore AJ, Blom AW, Whitehouse MR, Gooberman-Hill R. Deep prosthetic joint infection: a qualitative study of the impact on patients and their experiences of revision surgery. BMJ Open. 2015 Dec 7;5(12):e009495.
5. Richards J, Inacio MCS, Beckett M, Navarro RA, Singh A, Dillon MT, Sodl JF, Yian EH. Patient and procedure-specific risk factors for deep infection after primary shoulder arthroplasty. Clin Orthop Relat Res. 2014 Sep;472(9):2809-15.
6. Pottinger P, Butler-Wu S, Neradilek MB, Merritt A, Bertelsen A, Jette JL, Warme WJ, Matsen FA 3rd. Prognostic factors for bacterial cultures positive for Propionibacterium acnes and other organisms in a large series of revision shoulder arthroplasties performed for stiffness, pain, or loosening. J Bone Joint Surg Am. 2012 Nov 21;94(22):2075-83.
7. Garrigues GE, Zmistowski B, Cooper AM, Green A. Proceedings from the 2018 International Consensus Meeting on Orthopedic Infections: rationale and methods of the shoulder subgroup. J Shoulder Elbow Surg. 2019 Jun;28(6S):S4-7.
8. Singh JA, Sperling JW, Schleck C, Harmsen W, Cofield RH. Periprosthetic infections after shoulder hemiarthroplasty. J Shoulder Elbow Surg. 2012 Oct;21(10):1304-9.
9. Marigi EM, Bartels DW, Yoon JH, Sperling JW, Sanchez-Sotelo J. Antibiotic prophylaxis with cefazolin is associated with lower shoulder periprosthetic joint infection rates than non-cefazolin alternatives. J Bone Joint Surg Am. 2022 May 18;104(10):872-80.
10. Crane JK, Hohman DW, Nodzo SR, Duquin TR. Antimicrobial susceptibility of Propionibacterium acnes isolates from shoulder surgery. Antimicrob Agents Chemother. 2013 Jul;57(7):3424-6.
11. Smieja M. Current indications for the use of clindamycin: A critical review. Can J Infect Dis. 1998 Jan;9(1):22-8.
12. Yian EH, Chan PH, Burfeind W, Navarro RA, Singh A, Dillon MT. Perioperative Clindamycin use in Penicillin allergic patients is associated with a higher risk of infection after shoulder arthroplasty. J Am Acad Orthop Surg. 2020 Mar 15;28(6):e270-6.
13. Ghobrial GM, Cadotte DW, Williams K Jr, Fehlings MG, Harrop JS. Complications from the use of intrawound vancomycin in lumbar spinal surgery: a systematic review. Neurosurg Focus. 2015 Oct;39(4):E11.
14. Hey HWD, Thiam DW, Koh ZSD, Thambiah JS, Kumar N, Lau LL, Liu KG, Wong HK. Is intraoperative local Vancomycin powder the answer to surgical site infections in spine surgery? Spine. 2017 Feb 15;42(4):267-74.
15. Riesgo AM, Park BK, Herrero CP, Yu S, Schwarzkopf R, Iorio R. Vancomycin Povidone-Iodine protocol improves survivorship of periprosthetic joint infection treated with irrigation and debridement. J Arthroplasty. 2018 Mar;33(3):847-50.
16. Tomov M, Mitsunaga L, Durbin-Johnson B, Nallur D, Roberto R. Reducing surgical site infection in spinal surgery with betadine irrigation and intrawound vancomycin powder. Spine. 2015 Apr 1;40(7):491-9.
17. Yan H, He J, Chen S, Yu S, Fan C. Intrawound application of vancomycin reduces wound infection after open release of post-traumatic stiff elbows: a retrospective comparative study. J Shoulder Elbow Surg. 2014 May;23(5):686-92.
18. Moeini S, Rasmussen JV, Salomonsson B, Domeij-Arverud E, Fenstad AM, Hole R, Jensen SL, Brorson S. Reverse shoulder arthroplasty has a higher risk of revision due to infection than anatomical shoulder arthroplasty: 17 730 primary shoulder arthroplasties from the Nordic Arthroplasty Register Association. Bone Joint J. 2019 Jun;101-B(6):702-7.
19. Schiffman CJ, Hannay WM, Whitson AJ, Neradilek MB, Matsen FA 3rd, Hsu JE. Impact of previous non-arthroplasty surgery on clinical outcomes after primary anatomic shoulder arthroplasty. J Shoulder Elbow Surg. 2020 Oct;29(10):2056-64.
20. Werthel JD, Hatta T, Schoch B, Cofield R, Sperling JW, Elhassan BT. Is previous nonarthroplasty surgery a risk factor for periprosthetic infection in primary shoulder arthroplasty? J Shoulder Elbow Surg. 2017 Apr;26(4):635-40.
