Machine Learning Algorithm May Predict Postfracture Infection
Infection after surgery to treat a tibial shaft fracture can have devastating consequences, with significant associated costs and burdens. Although research has identified general risk factors that increase the likelihood of infection (including complexity of injury and fracture patterns and patient-related factors such as smoking and diabetes), predicting risks for individual patients remains difficult.
In a recent study in The Journal, investigators from the Machine Learning Consortium reported on an algorithm they developed to predict the risk of infection in specific patients who receive operative treatment for a tibial shaft fracture. To develop their model, the researchers used high-quality data from the SPRINT (Study to Prospectively Evaluate Reamed Intramedullary Nails in Patients with Tibial Fractures) and FLOW (Fluid Lavage of Open Wounds) randomized controlled trials.
The Australian researchers “trained” 5 machine learning algorithms and tested them against various performance measures to evaluate 1,822 fractures, including 170 (9%) that developed an infection. Based on predictive performance in that derivation portion of the study, 3 algorithms were validated and 1 prediction model was found to be superior. In that model, Gustilo-Anderson Type IIIA and IIIB fractures, age, AO/OTA type 42C3 fractures, crush injuries, and falls were the strongest predictors of infection.
Researchers have made their model available in an online, open-access prediction tool. Although the authors emphasize that this preliminary tool is intended for research and not for widespread clinical use, I think it has profoundly positive potential. Being able to risk-stratify a patient with a tibial shaft fracture at or near the time of admission could allow surgeons to closely monitor—and intervene sooner—in fracture cases at risk for infection, thereby possibly preventing devastating complications. This prediction tool certainly needs external validation prior to “prime-time” adoption, but when it comes to exploring artificial intelligence and machine learning in orthopaedics, the future is now.
Matthew R. Schmitz, MD
JBJS Deputy Editor for Social Media
Revision Shoulder Arthroplasty: IV or Oral Antibiotics?
Surgeons performing revision shoulder arthroplasty typically order postoperative antibiotics to be administered while they wait for results from intraoperative cultures. Based on their index of suspicion from preoperative exams and intraoperative observations, they order either intravenous (high suspicion of infection) or oral (low suspicion) antibiotics during the waiting period. In the June 3, 2020 issue of JBJS, Yao et al. report on a retrospective review of 175 patients who underwent revision shoulder arthroplasty, finding that surgeons’ presumptive choice of antibiotic type matched the culture results in 75% of the cases.
Among the 175 patients in the study, IV antibiotics were initiated in 62, while 113 patients received oral antibiotics. Cultures from 49 of the 62 patients started on IV antibiotics came back positive, and cultures from 83 of the 113 patients started on oral antibiotics came back negative. Treatment of patients whose initial antibiotic regimen did not match culture results was modified accordingly.
After multivariate analysis Yao et al. found that male sex, prior ipsilateral infection, and intraoperative presence of a humeral membrane were 3 independent predictors of surgeons initiating IV antibiotics. Antibiotic-related adverse events (including GI, dermatologic, and allergic reactions) occurred in 19% of the patients. Not surprisingly, the rate of these complications was highest among those receiving IV antibiotics.
Although the surgeons’ empirical initiation of antibiotic administration route was “correct” 75% of the time, that still left 25% of the patients needing modification of therapy based on culture results. While the authors observe that their study was not designed “to report the relative effectiveness of the 2 antibiotic protocols in minimizing the risk of recurrent infection,” their findings confirm that preoperative and intraoperative observations can help surgeons select the “right” type of antibiotic without culture results—and that is heartening.
What’s New in Pediatric Orthopaedic Surgery 2020
Every month, JBJS publishes a review of the most pertinent and impactful studies published in the orthopaedic literature during the previous year in 13 subspecialties. Click here for a collection of OrthoBuzz summaries of these “What’s New” articles. This month, co-author Kelly Vanderhave, MD selected the 5 most clinically compelling findings from the more than 50 studies summarized in the February 19, 2020 “What’s New in Pediatric Orthopaedic Surgery.”
ACL Reconstruction
—ACL reconstruction in pediatric patients continues to receive research attention. A recent review of >560 cases showed that soft-tissue grafts used in this population were twice as likely to fail (13%) as patellar tendon grafts (6%) (p <0.001).1
Septic Arthritis of the Hip
—A multicenter study identified the following independent risk factors for a repeat surgical procedure after initial arthrotomy for septic arthritis of the hip: presenting CRP of >10 mg/dL and ESR of >40 mm/hr, and the presence of osteomyelitis and MRSA.2
Adolescent Idiopathic Scoliosis
—A minimum 20-year follow-up of a cohort study evaluating 180 patients after observation, bracing, or surgical management of adolescent idiopathic scoliosis found the following:
- In the observation cohort, 5 of 36 patients underwent a scoliosis surgical procedure as an adult.
- In the bracing cohort, only 1 of 41 patients required an additional spinal surgical procedure.
- In the surgical cohort, 7 of 103 patients required a revision surgical procedure.
At a mean follow-up of 30 years, there were no significant differences in patient-reported outcomes between the 3 cohorts.3
Infection after Spinal Deformity Surgery
—A retrospective study of >600 pediatric patients who underwent spinal deformity surgery identified 2 independent risk factors among 11 cases of deep surgical site infection that occurred >3 months after the procedure:
- Nonidiopathic scoliosis (e.g., neuromuscular, congenital, and syndromic etiologies)
- High volume of crystalloid administered during surgery (mean of 3.3 ±1.2 L in the group with surgical site infections vs 2.4 ±1.0 L in the infected group)
Redosing antibiotics intraoperatively after 3 hours did not significantly influence the risk of infection.4
Hip Dislocations in Infants with CP
—Among 11 patients (15 hips) with spastic cerebral palsy whose preoperative mean acetabular index was 29°, surgical hip reconstruction (a combination of open reduction, adductor tenotomy, femoral osteotomy, and/or pelvic osteotomy) yielded the following results at a mean follow-up of 40 months:
- Mean migration index of 7%
- Mean acetabular index of 22°
- No instances of osteonecrosis
- 90% achievement and maintenance of hip reduction in those who underwent open reduction with or without pelvic or femoral osteotomy.5
References
- Ho B, Edmonds EW, Chambers HG, Bastrom TP, Pennock AT. Risk factors for early ACL reconstruction failure in pediatric and adolescent patients: a review of 561 cases. J Pediatr Orthop. 2018 Aug;38(7):388-92.
- Murphy RF, Plumblee L, Barfield WB, Murphy JS, Fuerstenau N, Spence DD, Kelly DM, Dow MA, Mooney JF 3rd. Septic arthritis of the hip-risk factors associated with secondary surgery. J Am Acad Orthop Surg. 2019 May 1;27(9):321-6.
- Larson AN, Baky F, Ashraf A, Baghdadi YM, Treder V, Polly DW Jr, Yaszemski MJ. Minimum 20-year health-related quality of life and surgical rates after the treatment of adolescent idiopathic scoliosis. Spine Deform. 2019 May;7(3):417-27.
- Du JY, Poe-Kochert C, Thompson GH, Son-Hing JP, Hardesty CK, Mistovich RJ. Risk factors for early infection in pediatric spinal deformity surgery: a multivariate analysis. Spine Deform. 2019 May;7(3):410-6.
- Refakis CA, Baldwin KD, Spiegel DA, Sankar WN. Treatment of the dislocated hip in infants with spasticity. J Pediatr Orthop. 2018 Aug;38(7):345-9.
What’s New in Adult Reconstructive Knee Surgery 2020
Every month, JBJS publishes a review of the most pertinent and impactful studies published in the orthopaedic literature during the previous year in 13 subspecialties. Click here for a collection of OrthoBuzz summaries of these “What’s New” articles. This month, author Michael J. Taunton, MD selected the 5 most clinically compelling findings from the more than 130 studies summarized in the January 15, 2020 “What’s New in Adult Reconstructive Knee Surgery.”
Unicompartmental Knee Arthroplasty (UKA)
—A prospective cohort study of 1,000 Oxford cementless UKAs indicated by standard Kozinn and Scott criteria found that revision-free survivorship at 10 years was 97%. Progression of lateral osteoarthritis and dislocation of the bearing were the most common reasons for revision.1
Pain Management
—Authors of a double-blinded, prospective, randomized study assigned 60 primary total knee arthroplasty (TKA) patients to receive either a continuous adductor canal block or a single-injection adductor canal block with adjuvant agents. They found no between-group differences in pain scores up to 42 hours postoperatively.2
Post-TKA Physical Therapy (PT)
—A prospective, randomized, noninferiority trial demonstrated that 290 post-TKA patients who were randomized to either outpatient PT, unsupervised web-based PT at home, or unsupervised printed-instruction-based PT at home had no difference in knee range of motion or in patient-reported outcomes at 4 to 6 weeks or 6 months postoperatively.3
Infection Prevention
—In a retrospective review of 29,695 total joint arthroplasties, preoperative penicillin allergy testing led to a 1.19% higher rate of infection-free survival at 10 years, principally by allowing more routine use of the prophylactic antibiotic cefazolin.4
Revision TKA
—A retrospective case series found that patients undergoing revision TKA at an age of < 50 years had a survivorship free of re-revision of 66% at 10 years. Regardless of the reason for revision, this population also had a higher risk of mortality than the general population at 10 years.5
References
- Campi S, Pandit H, Hooper G, Snell D, Jenkins C, Dodd CAF, et al. Ten-year survival and seven-year functional results of cementless Oxford unicompartmental knee replacement: A prospective consecutive series of our first 1000 cases. Knee. 2018 Dec;25(6):1231-7. Epub 2018/08/29.
- Turner JD, Dobson SW, Henshaw DS, Edwards CJ, Weller RS, Reynolds JW, et al. Single-Injection Adductor Canal Block With Multiple Adjuvants Provides Equivalent Analgesia When Compared With Continuous Adductor Canal Blockade for Primary Total Knee Arthroplasty: A Double-Blinded, Randomized, Controlled, Equivalency Trial. J Arthroplasty. 2018 Oct;33(10):3160-6 e1. Epub 2018/06/16.
- Fleischman AN, Crizer MP, Tarabichi M, Smith S, Rothman RH, Lonner JH, et al. 2018 John N. Insall Award: Recovery of Knee Flexion With Unsupervised Home Exercise Is Not Inferior to Outpatient Physical Therapy After TKA: A Randomized Trial. Clin Orthop Relat Res. 2019 Jan;477(1):60-9. Epub 2019/02/23.
- Wyles CC, Hevesi M, Osmon DR, Park MA, Habermann EB, Lewallen DG, et al. 2019 John Charnley Award: Increased risk of prosthetic joint infection following primary total knee and hip arthroplasty with the use of alternative antibiotics to cefazolin: the value of allergy testing for antibiotic prophylaxis. Bone Joint J. 2019 Jun;101-B(6_Supple_B):9-15. Epub 2019/05/31.
- Chalmers BP, Pallante GD, Sierra RJ, Lewallen DG, Pagnano MW, Trousdale RT. Contemporary Revision Total Knee Arthroplasty in Patients Younger Than 50 Years: 1 in 3 Risk of Re-Revision by 10 Years. J Arthroplasty. 2019 Jul;34(7S):S266-S70. Epub 2019/03/03.
Which Germs Cause Pelvic Prosthesis Infections?
Prosthetic infections involving total hip or knee implants are bad enough, but infections involving pelvic endoprostheses following tumor resection can be particularly devastating, often necessitating multiple surgical interventions. And, such infections are disconcertingly common, affecting an estimated 11% to 53% of pelvic endoprostheses.
Findings from a retrospective multisite cohort study by Sanders et al. in the May 1, 2019 issue of The Journal of Bone & Joint Surgery reveal more about the specific microorganisms underlying those infections—and may offer insight into how to prevent them.
The authors analyzed 70 patients who underwent pelvic endoprosthetic reconstruction following a tumor resection. Eighteen patients (26%) developed an infection, and in 14 out of those 18 cases, the infection was determined to be polymicrobial. Cultures from 12 of the 18 patients (67%) were positive for a member of the Enterobacteriaceae family of gram-negative bacteria, which includes Escherichia coli. More generally, microorganisms associated with intestinal flora appeared 32 out of the 42 times that any microorganism was isolated.
At the latest follow-up (median follow-up was 66 months), 9 of the 18 patients still had the original implant, although 2 of those patients had a fistula and another 2 were receiving suppressive antibiotics. Of the remaining 9 patients who had the original implant removed, 3 had a second implant in situ.
The authors emphasize how different these pelvic endoprosthetic infections are from infections related to joint arthroplasty. The close proximity of incisions for periacetabular tumor resection to the gut and other highly colonized areas might contribute to these infections, they speculate. Sanders et al. say the findings of this study may prompt surgeons to employ additional surgical-site antiseptic measures before and during these surgeries and “may justify the use of a broader spectrum of [systemic] antibiotic prophylaxis aimed at gram-negative bacteria.” They also suggest that investigations into “selective gut decontamination” might yield additional information about how to prevent infections in this surgical setting.
Revisiting INR Targets Prior to TKA
An elevated International Normalized Ratio (INR)—a standardized gauge for how long it takes blood to clot—is rarely a good sign when someone is about to undergo an elective orthopaedic procedure. This is especially true for larger surgeries such as total hip or knee arthroplasty, in which there are already concerns about perioperative bleeding. Excessive surgery-related blood loss can lead to wound complications, increased length of hospital stay, and higher mortality rates. But what precisely constitutes an “elevated” INR? While some recommendations suggest that elective procedures be performed only when a patient’s INR is ≤1.5, the evidence supporting this recommendation, especially in the setting of total knee arthroplasty (TKA), is sparse at best.
In the March 20, 2019 issue of The Journal, Rudasill et al. use the National Surgical Quality Improvement Program (NSQIP) database to help define what “elevated” should mean in the context of TKA. They evaluated data from >21,000 patients who underwent a TKA between 2010 and 2016 and who also had an INR level reported within one day before their joint replacement. They stratified these patients based on their INR levels (≤1, >1 to 1.25, >1.25 to 1.5, and >1.5). Using multivariate regression analysis to adjust for patient demographics and comorbidities, the authors found a progressively increasing risk of bleeding requiring transfusion for each group with an INR >1 (odds ratios of 1.19, 1.29 and 2.02, respectively). Relative to patients with an INR of ≤1, Rudasill et al. also found a significantly increased risk of infection in TKA patients with an INR >1.5 (odds ratio 5.34), and an increased risk of mortality within 30 days of surgery among patients with an INR >1.25 to 1.5 (odds ratio 3.37). Lastly, rates of readmission and the length of stay were significantly increased in patients with an INR >1.25.
While there are certainly weaknesses inherent in using the NSQIP dataset, this study is the first to carefully evaluate the impact of slight INR elevations on post-TKA morbidity and mortality. While I was not surprised that increasing INR levels were associated with increased bleeding events, I was impressed by the profound differences in length of stay, infection, and mortality between patients with an INR ≤1 and those with an INR >1.25. I agree with the authors’ conclusion that “current guidelines for a target INR of <1.5 should be reconsidered for patients undergoing TKA.” Further, based on the risks highlighted in this study, prospective or propensity matched cohort studies should be performed to help determine whether anyone with an INR >1 should undergo a TKA.
Chad A. Krueger, MD
JBJS Deputy Editor for Social Media
What’s New in Reconstructive Knee Surgery 2019
Every month, JBJS publishes a review of the most pertinent and impactful studies published in the orthopaedic literature during the previous year in one of 13 subspecialties. Click here for a collection of all OrthoBuzz subspecialty summaries. This month, Michael J. Taunton, MD, author of the January 16, 2019 “What’s New in Adult Reconstructive Knee Surgery,” selected the five most compelling findings from among the more than 100 noteworthy studies summarized in the article.
Cementless vs Cemented TKA Fixation
—A matched case-control study of 400 primary total knee arthroplasties (TKAs) found that cementless TKAs had a 0.5% rate of aseptic loosening over a mean follow-up of 2.5 years, while cemented TKAs had an aseptic loosening rate of 2.5%.1
TKA Component Size in Obese Patients
—Among 35 revision-TKA patients with a varus collapse of the tibia, 29 weighed >200 lbs. Fehring et al. found that patients with implants at the small end of the range of the manufacturer’s tibial size offering and with >5° of preoperative varus were at increased risk of tibial-component failure.2
Outpatient TKA
—A retrospective multivariate analysis of >4,300 patients who underwent outpatient TKA and >128,900 patients who underwent inpatient TKA found that, within 1 year, those who had outpatient procedures were more likely to experience a tibial and/or femoral component revision due to a noninfectious cause, irrigation and debridement, explantation of the prosthesis, and stiffness requiring manipulation under anesthesia.
Infection Prevention
—In a randomized trial of patients undergoing TKA, one group received 15 mg/kg of systemic intravenous vancomycin, and a second group received intraosseous regional administration of 500 mg vancomycin into the tibia. Mean tissue concentrations of the antibiotic were 34.4 mg/g in the intraosseous group and 6.1 mg/g in the intravenous group, suggesting that intraosseous administration provides a significantly higher tissue concentration of that antibiotic. 3
TKA Anesthesia Protocol
—A retrospective review of 156 consecutive patients who underwent primary TKA found that procedures performed with mepivacaine spinal anesthesia led to fewer episodes of urinary catheterization and shorter mean length of stay compared with procedures performed with bupivacaine spinal anesthesia.4
References
- Miller AJ, Stimac JD, Smith LS, Feher AW, Yakkanti MR, Malkani AL. Results of cemented vs cementless primary total knee arthroplasty using the same implant design. J Arthroplasty.2018 Apr;33(4):1089-93. Epub 2017 Dec
- Fehring TK, Fehring KA, Anderson LA, Otero JE, Springer BD. Catastrophic varus collapse of the tibia in obese total knee arthroplasty. J Arthroplasty.2017 May;32(5):1625-9. Epub 2017 Jan 30.
- Chin SJ, Moore GA, Zhang M, Clarke HD, Spangehl MJ, Young SW. The AAHKS Clinical Research Award: intraosseous regional prophylaxis provides higher tissue concentrations in high BMI patients in total knee arthroplasty: a randomized trial. J Arthroplasty.2018 Jul;33(7S):S13-8. Epub 2018 Mar 15.
- Mahan MC, Jildeh TR, Tenbrunsel TN, Davis JJ. Mepivacaine spinal anesthesia facilitates rapid recovery in total knee arthroplasty compared to bupivacaine. J Arthroplasty.2018 Jun;33(6):1699-704. Epub 2018 Jan 16.
Routine Diagnostic Tests for Periprosthetic Joint Infection Demonstrate a High False-Negative Rate
What’s New in Musculoskeletal Basic Science 2018
Every month, JBJS publishes a Specialty Update—a review of the most pertinent and impactful studies published in the orthopaedic literature during the previous year in 13 subspecialties. Click here for a collection of all OrthoBuzz Specialty Update summaries.
This month, Matthew J. Allen, VetMB, PhD, author of the December 5, 2018 Specialty Update on Musculoskeletal Basic Science, focuses on the five most compelling findings from among the more than 60 noteworthy studies summarized in the article.
Gene Editing in Orthopaedics
–Gene-editing tools such as CRISPR-Cas9 have great potential as a means of introducing therapeutic genes into mesenchymal stem cells that can then be targeted to tissues in vivo. These researchers1 reported on genetically modified stem cells that have the potential to differentiate into chondrocytes encoding a natural inhibitor of interleukin-1, providing an opportunity for localized release of immunomodulatory factors.
Managing Orthopaedic Infections
–A novel study2 in which transmission electron microscopy was used to identify viable bacteria deep within the canalicular structure of cortical bone, remote from the site of an infected implant, suggests that effective debridement requires the removal of not just necrotic tissue, but also of adjacent, apparently unaffected bone.
Computational Modeling of Human Movement
–This report3 presented a human musculoskeletal model that provided extremely accurate predictions of ground reaction forces during simulated walking and squatting. As similar models are developed and validated, surgeons will have improved tools for evaluating patients, planning surgery, and making decisions about which procedure/implant is most appropriate for an individual patient.
Sex-Related Differences
–This report4 demonstrated sexually dimorphic regulation of gene-expression profiles in bone marrow osteoprogenitor cells that could partly explain clinical observations in sex differences in peak bone mass, bone remodeling, and immunomodulation.
Biological Enhancement of Ligament Healing
–Among several basic science papers focused on the optimal healing and durable fixation of tendons and ligaments, this notable work5 reported on the translation of bridge-enhanced ligament repair for the anterior cruciate ligament.
References
- Brunger JM, Zutshi A, Willard VP, Gersbach CA, Guilak F. CRISPR/Cas9 editing of murine induced pluripotent stem cells for engineering inflammation-resistant tissues. Arthritis Rheumatol.2017 May;69(5):1111-21. Epub 2017 Mar 31.
- de Mesy Bentley KL, Trombetta R, Nishitani K, Bello-Irizarry SN, Ninomiya M, Zhang L, Chung HL, McGrath JL, Daiss JL, Awad HA, Kates SL, Schwarz EM. Evidence of Staphylococcus aureus deformation, proliferation, and migration in canaliculi of live cortical bone in murine models of osteomyelitis. J Bone Miner Res.2017 May;32(5):985-90. Epub 2017 Jan 26.
- Jung Y, Koo YJ, Koo S. Simultaneous estimation of ground reaction force and knee contact force during walking and squatting. Int J Precis Eng Manuf.2017;18(9):1263-8.
- Kot A, Zhong ZA, Zhang H, Lay YE, Lane NE, Yao W. Sex dimorphic regulation of osteoprogenitor progesterone in bone stromal cells. J Mol Endocrinol.2017 Nov;59(4):351-63. Epub 2017 Sep 4.
- Perrone GS, Proffen BL, Kiapour AM, Sieker JT, Fleming BC, Murray MM. Bench-to-bedside: bridge-enhanced anterior cruciate ligament repair. J Orthop Res.2017 Dec;35(12):2606-12. Epub 2017 Jul 9.
Physical Parameters Beyond BMI Affect TKA Outcomes
It is well established that obese patients who undergo total joint arthroplasty have increased risks of complications and infections. But what about folks who are not obese, but are just generally large? Do they also have increased post-arthroplasty complications, compared to their smaller counterparts? That is the question Christensen et al. explored in a registry-based study in the November 7, 2018 edition of JBJS.
In addition to BMI, the authors examined 3 other physical parameters—body surface area, body mass, and height—to determine whether these less-studied characteristics (all contributing to “bigness”) were associated with an increased rate of various adverse outcomes, including mechanical failure and infection, after primary total knee arthroplasty (TKA). They evaluated data from more than 22,000 TKAs performed at a single institution and found that the risk of any revision procedure or revision for a mechanical failure was directly associated with every 1 standard deviation increase in BMI (Hazard Ratio [HR], 1.19 and 1.15, respectively), body surface area (HR, 1.37 and 1.35, respectively), body mass (HR, 1.30 and 1.27, respectively), and height (HR, 1.22 and 1.23, respectively). In this study, 1 standard deviation was equivalent to 6.3 kg/m2 for BMI, 0.3 m2 for body surface area, 20 kg for body mass, and 10.5 cm for height.
These findings, while not all that surprising, are enlightening nonetheless. The study shows that increasing height has a greater negative impact on TKA outcomes than previously thought. While I spend a lot of time counseling patients with high BMIs about the increased risks of undergoing a TKA (and while such patients can take certain actions to lower their BMI prior to surgery), I do not spend nearly as much time counseling patients who are much taller than normal about their increased risks (and height is not a modifiable risk factor). Nor do I spend much time thinking about a patient’s overall body mass or body surface area in addition to their BMI. This study will remind me not to overlook these less commonly examined physical parameters when discussing TKA with patients in the future.
Chad A. Krueger, MD
JBJS Deputy Editor for Social Media
Try Clinical Classroom
Recent Comments