OpenLB
Open Library of Bioscience
Advanced Search
Archives of orthopaedic and trauma surgery
Jan 17, 2025;145 (1): 131.

Machine learning to predict periprosthetic joint infections following primary total hip arthroplasty using a national database.

Mehdi S Salimy, Anirudh Buddhiraju, Tony L-W Chen, Ashish Mittal, Pengwei Xiao, Young-Min Kwon
Author Information
  1. Mehdi S Salimy: Bioengineering Laboratory, Department of Orthopaedic Surgery, Massachusetts General Hospital, Harvard Medical School, 55 Fruit Street, Boston, MA, 02114, USA.
  2. Anirudh Buddhiraju: Bioengineering Laboratory, Department of Orthopaedic Surgery, Massachusetts General Hospital, Harvard Medical School, 55 Fruit Street, Boston, MA, 02114, USA.
  3. Tony L-W Chen: Bioengineering Laboratory, Department of Orthopaedic Surgery, Massachusetts General Hospital, Harvard Medical School, 55 Fruit Street, Boston, MA, 02114, USA.
  4. Ashish Mittal: Bioengineering Laboratory, Department of Orthopaedic Surgery, Massachusetts General Hospital, Harvard Medical School, 55 Fruit Street, Boston, MA, 02114, USA.
  5. Pengwei Xiao: Bioengineering Laboratory, Department of Orthopaedic Surgery, Massachusetts General Hospital, Harvard Medical School, 55 Fruit Street, Boston, MA, 02114, USA.
  6. Young-Min Kwon: Bioengineering Laboratory, Department of Orthopaedic Surgery, Massachusetts General Hospital, Harvard Medical School, 55 Fruit Street, Boston, MA, 02114, USA. ymkwon@mgh.harvard.edu. ORCID
PMID: 39820648 DOI: 10.1007/s00402-025-05757-4

Abstract

INTRODUCTION: Periprosthetic joint infection (PJI) following total hip arthroplasty (THA) remains a devastating complication for patients and surgeons. Given the implications of these infections and the current paucity of risk calculators utilizing machine learning (ML), this study aimed to develop an ML algorithm that could accurately identify risk factors for developing a PJI following primary THA using a national database.
MATERIALS AND METHODS: A total of 51,053 patients who underwent primary THA between 2013 and 2020 were identified using the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database. Demographic, preoperative, intraoperative, and immediate postoperative outcomes were collected. Five ML models were created. The receiver operating characteristic curves, the area under the curve (AUC), calibration plots, slopes, intercepts, and Brier scores were evaluated.
RESULTS: The histogram-based gradient boosting (HGB) model demonstrated good PJI discriminatory ability with an AUC of 0.88. The test-specific metrics supported the model's performance and validation in predicting PJI (calibration curve slope: 0.79; intercept: 0.32; Brier score: 0.007). The top five predictors of PJI were the length of stay (> 3 days), patient weight at the time of surgery (> 94.3 kg), an American Society of Anesthesiologists (ASA) class of 4 or higher, preoperative platelet count (< 249,890/mm3), and preoperative sodium (< 139.5 mEq/L).
CONCLUSION: This study developed a highly specific ML model that could predict patient-specific PJI development following primary THA. Considering the feature importance of the top predictors of infection, surgeons should counsel at-risk patients to optimize resource utilization and potentially improve surgical outcomes.

Keywords

Artificial intelligence Histogram-based gradient boosting Machine learning Periprosthetic joint infection Surgical site infection Total hip arthroplasty

References

  1. Shan L, Shan B, Graham D, Saxena A (2014) Total hip replacement: a systematic review and meta-analysis on mid-term quality of life. Osteoarthritis Cartilage 22:389–406. https://doi.org/10.1016/j.joca.2013.12.006 [DOI: 10.1016/j.joca.2013.12.006]
  2. Belmont PJ, Goodman GP, Hamilton W, Waterman BR, Bader JO, Schoenfeld AJ (2014) Morbidity and mortality in the thirty-day period following total hip arthroplasty: risk factors and incidence. J Arthroplasty 29:2025–2030. https://doi.org/10.1016/j.arth.2014.05.015 [DOI: 10.1016/j.arth.2014.05.015]
  3. Florschutz AV, Fagan RP, Matar WY, Sawyer RG, Berrios-Torres SI (2015) Surgical site infection risk factors and risk stratification. J Am Acad Orthop Surg 23:S8–11. https://doi.org/10.5435/JAAOS-D-14-00447 [DOI: 10.5435/JAAOS-D-14-00447]
  4. Centers for Disease Control and Prevention (CDC) National Healthcare Safety Network (NHSN) (2022) Surg Site Infect Event SSI 2022. https://www.cdc.gov/nhsn/pdfs/pscmanual/9pscssicurrent.pdf
  5. Poultsides LA, Ma Y, Della Valle AG, Chiu Y-L, Sculco TP, Memtsoudis SG (2013) In-hospital surgical site infections after primary hip and knee arthroplasty–incidence and risk factors. J Arthroplasty 28:385–389. https://doi.org/10.1016/j.arth.2012.06.027 [DOI: 10.1016/j.arth.2012.06.027]
  6. Garrido-Gómez J, Arrabal-Polo MA, Girón-Prieto MS, Cabello-Salas J, Torres-Barroso J, Parra-Ruiz J (2013) Descriptive analysis of the economic costs of periprosthetic joint infection of the knee for the public health system of Andalusia. J Arthroplasty 28:1057–1060. https://doi.org/10.1016/j.arth.2013.02.012 [DOI: 10.1016/j.arth.2013.02.012]
  7. Wolford HM, Hatfield KM, Paul P, Yi SH, Slayton RB (2018) The projected burden of complex surgical site infections following hip and knee arthroplasties in adults in the United States, 2020 through 2030. Infect Control Hosp Epidemiol 39:1189–1195. https://doi.org/10.1017/ice.2018.184 [DOI: 10.1017/ice.2018.184]
  8. Kurtz S, Ong K, Lau E, Mowat F, Halpern M (2007) Projections of primary and revision hip and knee arthroplasty in the United States from 2005 to 2030. J Bone Joint Surg Am 89:780–785. https://doi.org/10.2106/JBJS.F.00222 [DOI: 10.2106/JBJS.F.00222]
  9. Jin X, Luxan BG, Hanly M, Pratt NL, Harris I, de Steiger R et al (2022) Estimating incidence rates of periprosthetic joint infection after hip and knee arthroplasty for osteoarthritis using linked registry and administrative health data. Bone Jt J 104–B:1060–1066. https://doi.org/10.1302/0301-620X.104B9.BJJ-2022-0116.R1 [DOI: 10.1302/0301-620X.104B9.BJJ-2022-0116.R1]
  10. Premkumar A, Kolin DA, Farley KX, Wilson JM, McLawhorn AS, Cross MB et al (2021) Projected economic burden of periprosthetic joint infection of the hip and knee in the United States. J Arthroplasty 36:1484-1489e3. https://doi.org/10.1016/j.arth.2020.12.005 [DOI: 10.1016/j.arth.2020.12.005]
  11. Kurtz SM, Lau E, Watson H, Schmier JK, Parvizi J (2012) Economic burden of periprosthetic joint infection in the United States. J Arthroplasty 27:61–65e1. https://doi.org/10.1016/j.arth.2012.02.022 [DOI: 10.1016/j.arth.2012.02.022]
  12. Amanatullah D, Dennis D, Oltra EG, Marcelino Gomes LS, Goodman SB, Hamlin B et al (2019) Hip and knee section, diagnosis, definitions: proceedings of international consensus on orthopedic infections. J Arthroplasty 34:S329–S337. https://doi.org/10.1016/j.arth.2018.09.044 [DOI: 10.1016/j.arth.2018.09.044]
  13. Humphrey TJ, Salimy MS, Melnic CM, Bedair HS (2022) Dislocation after debridement, antibiotics, and implant retention for periprosthetic joint infections of the hip. J Arthroplasty. https://doi.org/10.1016/j.arth.2022.08.029 [DOI: 10.1016/j.arth.2022.08.029]
  14. Corona PS, Vicente M, Carrera L, Rodríguez-Pardo D, Corró S (2020) Current actual success rate of the two-stage exchange arthroplasty strategy in chronic hip and knee periprosthetic joint infection. Bone Jt J 102–B:1682–1688. https://doi.org/10.1302/0301-620X.102B12.BJJ-2020-0792.R1 [DOI: 10.1302/0301-620X.102B12.BJJ-2020-0792.R1]
  15. Browne JA, Cancienne JM, Novicoff WM, Werner BC (2017) Removal of an infected hip arthroplasty is a high-risk surgery: putting morbidity into context with other major nonorthopedic operations. J Arthroplasty 32:2834–2841. https://doi.org/10.1016/j.arth.2017.03.061 [DOI: 10.1016/j.arth.2017.03.061]
  16. Parvizi J, Shohat N, Gehrke T (2017) Prevention of periprosthetic joint infection: new guidelines. Bone Jt J 99–B:3–10. https://doi.org/10.1302/0301-620X.99B4.BJJ-2016-1212.R1 [DOI: 10.1302/0301-620X.99B4.BJJ-2016-1212.R1]
  17. Alamanda VK, Springer BD (2019) The prevention of infection: 12 modifiable risk factors. Bone Jt J 101–B:3–9. https://doi.org/10.1302/0301-620X.101B1.BJJ-2018-0233.R1 [DOI: 10.1302/0301-620X.101B1.BJJ-2018-0233.R1]
  18. Gallo J, Nieslanikova E (2020) Prevention of prosthetic joint infection: from traditional approaches towards quality improvement and data mining. J Clin Med 9:2190. https://doi.org/10.3390/jcm9072190 [DOI: 10.3390/jcm9072190]
  19. Bozic KJ, Ong K, Lau E, Berry DJ, Vail TP, Kurtz SM et al (2013) Estimating risk in medicare patients with THA: an electronic risk calculator for periprosthetic joint infection and mortality. Clin Orthop 471:574–583. https://doi.org/10.1007/s11999-012-2605-z [DOI: 10.1007/s11999-012-2605-z]
  20. Wingert NC, Gotoff J, Parrilla E, Gotoff R, Hou L, Ghanem E (2016) The ACS NSQIP risk calculator is a fair predictor of acute periprosthetic joint infection. Clin Orthop 474:1643–1648. https://doi.org/10.1007/s11999-016-4717-3 [DOI: 10.1007/s11999-016-4717-3]
  21. Cabitza F, Locoro A, Banfi G (2018) Machine learning in orthopedics: a literature review. Front Bioeng Biotechnol 6:75. https://doi.org/10.3389/fbioe.2018.00075 [DOI: 10.3389/fbioe.2018.00075]
  22. Haeberle HS, Helm JM, Navarro SM, Karnuta JM, Schaffer JL, Callaghan JJ et al (2019) Artificial intelligence and machine learning in lower extremity arthroplasty: a review. J Arthroplasty 34:2201–2203. https://doi.org/10.1016/j.arth.2019.05.055 [DOI: 10.1016/j.arth.2019.05.055]
  23. Murphy MP, Brown NM (2021) CORR synthesis: when should the orthopaedic surgeon use artificial intelligence, machine learning, and deep learning? Clin Orthop 479:1497–1505. https://doi.org/10.1097/CORR.0000000000001679 [DOI: 10.1097/CORR.0000000000001679]
  24. Klemt C, Laurencin S, Uzosike AC, Burns JC, Costales TG, Yeo I et al (2022) Machine learning models accurately predict recurrent infection following revision total knee arthroplasty for periprosthetic joint infection. Knee Surg Sports Traumatol Arthrosc off J ESSKA 30:2582–2590. https://doi.org/10.1007/s00167-021-06794-3 [DOI: 10.1007/s00167-021-06794-3]
  25. Yeo I, Klemt C, Robinson MG, Esposito JG, Uzosike AC, Kwon Y-M (2022) The use of artificial neural networks for the prediction of surgical site infection following TKA. J Knee Surg. https://doi.org/10.1055/s-0041-1741396 [DOI: 10.1055/s-0041-1741396]
  26. Shiloach M, Frencher SK, Steeger JE, Rowell KS, Bartzokis K, Tomeh MG et al (2010) Toward robust information: data quality and inter-rater reliability in the American college of surgeons national surgical quality improvement program. J Am Coll Surg 210:6–16. https://doi.org/10.1016/j.jamcollsurg.2009.09.031 [DOI: 10.1016/j.jamcollsurg.2009.09.031]
  27. Darst BF, Malecki KC, Engelman CD (2018) Using recursive feature elimination in random forest to account for correlated variables in high dimensional data. BMC Genet 19:65. https://doi.org/10.1186/s12863-018-0633-8 [DOI: 10.1186/s12863-018-0633-8]
  28. Klemt C, Laurencin S, Alpaugh K, Tirumala V, Barghi A, Yeo I et al (2022) The utility of machine learning algorithms for the prediction of early revision surgery after primary total hip arthroplasty. JAAOS—J Am Acad Orthop Surg 30:513. https://doi.org/10.5435/JAAOS-D-21-01039 [DOI: 10.5435/JAAOS-D-21-01039]
  29. Steyerberg EW, Vickers AJ, Cook NR, Gerds T, Gonen M, Obuchowski N et al (2010) Assessing the performance of prediction models: a framework for traditional and novel measures. Epidemiol Camb Mass 21:128–138. https://doi.org/10.1097/EDE.0b013e3181c30fb2 [DOI: 10.1097/EDE.0b013e3181c30fb2]
  30. de Hond AAH, Steyerberg EW, van Calster B (2022) Interpreting area under the receiver operating characteristic curve. Lancet Digit Health 4:e853–e855. https://doi.org/10.1016/S2589-7500(22)00188-1 [DOI: 10.1016/S2589-7500(22)00188-1]
  31. Huang Y, Li W, Macheret F, Gabriel RA, Ohno-Machado L (2020) A tutorial on calibration measurements and calibration models for clinical prediction models. J Am Med Inf Assoc 27:621–633. https://doi.org/10.1093/jamia/ocz228 [DOI: 10.1093/jamia/ocz228]
  32. Vickers AJ, van Calster B, Steyerberg EW (2019) A simple, step-by-step guide to interpreting decision curve analysis. Diagn Progn Res 3:18. https://doi.org/10.1186/s41512-019-0064-7 [DOI: 10.1186/s41512-019-0064-7]
  33. Olthof M, Stevens M, Bulstra SK, van den Akker-Scheek I (2014) The association between comorbidity and length of hospital stay and costs in total hip arthroplasty patients: a systematic review. J Arthroplasty 29:1009–1014. https://doi.org/10.1016/j.arth.2013.10.008 [DOI: 10.1016/j.arth.2013.10.008]
  34. Bozic KJ, Kurtz SM, Lau E, Ong K, Vail TP, Berry DJ (2009) The epidemiology of revision total hip arthroplasty in the United States. J Bone Joint Surg Am 91:128–133. https://doi.org/10.2106/JBJS.H.00155 [DOI: 10.2106/JBJS.H.00155]
  35. Pulido L, Ghanem E, Joshi A, Purtill JJ, Parvizi J (2008) Periprosthetic joint infection: the incidence, timing, and predisposing factors. Clin Orthop 466:1710–1715. https://doi.org/10.1007/s11999-008-0209-4 [DOI: 10.1007/s11999-008-0209-4]
  36. Bhanushali A, Tran L, Nairne-Nagy J, Bereza S, Callary SA, Atkins GJ, Ramasamy B, Solomon LB (2024) Patient-related predictors of treatment failure after two-stage total hip arthroplasty revision for periprosthetic joint infection: a systematic review and meta-analysis. J Arthroplasty 39(9):2395-2402e14. https://doi.org/10.1016/j.arth.2024.04.053 [DOI: 10.1016/j.arth.2024.04.053]
  37. Carender CN, Fruth KM, Lewallen DG, Berry DJ, Abdel MP, Bedard NA (2024) Obesity and primary total hip arthroplasty: the Absolute versus relative risk of periprosthetic joint infection at 15 years. J Arthroplasty 39(9S2):S436-S443.e1. https://doi.org/10.1016/j.arth.2024.03.033 [DOI: 10.1016/j.arth.2024.03.033]
  38. Zhong J, Wang B, Chen Y, Li H, Lin N, Xu X, Lu H (2020) Relationship between body mass index and the risk of periprosthetic joint infection after primary total hip arthroplasty and total knee arthroplasty. Ann Transl Med 8(7):464. https://doi.org/10.21037/atm.2020.03.112 [DOI: 10.21037/atm.2020.03.112]
  39. Ong KL, Kurtz SM, Lau E, Bozic KJ, Berry DJ, Parvizi J (2009) Prosthetic joint infection risk after total hip arthroplasty in the medicare population. J Arthroplasty 24:105–109. https://doi.org/10.1016/j.arth.2009.04.027 [DOI: 10.1016/j.arth.2009.04.027]
  40. Molloy IB, Martin BI, Moschetti WE, Jevsevar DS (2017) Effects of the length of stay on the cost of total knee and total hip arthroplasty from 2002 to 2013. J Bone Joint Surg Am 99:402–407. https://doi.org/10.2106/JBJS.16.00019 [DOI: 10.2106/JBJS.16.00019]
  41. Grosso MJ, Neuwirth AL, Boddapati V, Shah RP, Cooper HJ, Geller JA (2019) Decreasing length of hospital stay and postoperative complications after primary total hip arthroplasty: a decade analysis from 2006 to 2016. J Arthroplasty 34:422–425. https://doi.org/10.1016/j.arth.2018.11.005 [DOI: 10.1016/j.arth.2018.11.005]
  42. Peters CL, Shirley B, Erickson J (2006) The effect of a new multimodal perioperative anesthetic regimen on postoperative pain, side effects, rehabilitation, and length of hospital stay after total joint arthroplasty. J Arthroplasty 21:132–138. https://doi.org/10.1016/j.arth.2006.04.017 [DOI: 10.1016/j.arth.2006.04.017]
  43. Sutton JCI, Antoniou J, Epure LM, Huk OL, Zukor DJ, Bergeron SG (2016) Hospital discharge within 2 days following total hip or knee arthroplasty does not increase major-complication and readmission rates. JBJS 98:1419. https://doi.org/10.2106/JBJS.15.01109 [DOI: 10.2106/JBJS.15.01109]
  44. Dorr LD, Thomas DJ, Zhu J, Dastane M, Chao L, Long WT (2010) Outpatient total hip arthroplasty. J Arthroplasty 25:501–506. https://doi.org/10.1016/j.arth.2009.06.005 [DOI: 10.1016/j.arth.2009.06.005]
  45. DeMik DE, Carender CN, Kohler JG, An Q, Brown TS, Bedard NA (2022) Recent increases in outpatient total hip arthroplasty have not increased early complications. J Arthroplasty 37:325-329e1. https://doi.org/10.1016/j.arth.2021.11.003 [DOI: 10.1016/j.arth.2021.11.003]
  46. Tanenbaum JE, Bomberger TT, Knapik DM, Fitzgerald SJ, Shah NS, Wera GD (2020) Preoperative hyponatremia is a risk factor for adverse 30-day outcomes following total hip arthroplasty. J Hip Surg 04:007–014. https://doi.org/10.1055/s-0039-1701005 [DOI: 10.1055/s-0039-1701005]
  47. Baker CM, Goh GS, Tarabichi S, Sherman MB, Khan IA, Parvizi J (2023) Hyponatremia is an overlooked sign of trouble following total joint arthroplasty. JBJS 105:744. https://doi.org/10.2106/JBJS.22.00928 [DOI: 10.2106/JBJS.22.00928]
  48. Bini SA (2018) Artificial intelligence, machine learning, deep learning, and cognitive computing: what do these terms mean and how will they impact health care? J Arthroplasty 33:2358–2361. https://doi.org/10.1016/j.arth.2018.02.067 [DOI: 10.1016/j.arth.2018.02.067]
  49. Karlin EA, Lin CC, Meftah M, Slover JD, Schwarzkopf R (2023) The impact of machine learning on total joint arthroplasty patient outcomes: a systemic review. J Arthroplasty 38:2085–2095. https://doi.org/10.1016/j.arth.2022.10.039 [DOI: 10.1016/j.arth.2022.10.039]
  50. Alsoof D, McDonald CL, Kuris EO, Daniels AH (2022) Machine learning for the orthopaedic surgeon: uses and limitations. JBJS 104:1586. https://doi.org/10.2106/JBJS.21.01305 [DOI: 10.2106/JBJS.21.01305]

MeSH Term

Humans
Arthroplasty, Replacement, Hip
Machine Learning
Male
Female
Prosthesis-Related Infections
Middle Aged
Aged
Databases, Factual
Risk Factors
United States
Journal Article

Word Cloud

Created with Highcharts 10.0.0PJIinfectionfollowingTHAMLprimary0jointtotalhiparthroplastypatientslearningusingdatabasepreoperativePeriprostheticsurgeonsinfectionsriskstudynationalAmericanSurgicaloutcomescurveAUCcalibrationBriergradientboostingmodeltoppredictorspredictMachineINTRODUCTION:remainsdevastatingcomplicationGivenimplicationscurrentpaucitycalculatorsutilizingmachineaimeddevelopalgorithmaccuratelyidentifyfactorsdevelopingMATERIALSANDMETHODS:51053underwent20132020identifiedCollegeSurgeonsNationalQualityImprovementProgramACS-NSQIPDemographicintraoperativeimmediatepostoperativecollectedFivemodelscreatedreceiveroperatingcharacteristiccurvesareaplotsslopesinterceptsscoresevaluatedRESULTS:histogram-basedHGBdemonstratedgooddiscriminatoryability88test-specificmetricssupportedmodel'sperformancevalidationpredictingslope:79intercept:32score:007fivelengthstay> 3dayspatientweighttimesurgery> 943 kgSocietyAnesthesiologistsASAclass4higherplateletcount< 249890/mm3sodium< 1395 mEq/LCONCLUSION:developedhighlyspecificpatient-specificdevelopmentConsideringfeatureimportancecounselat-riskoptimizeresourceutilizationpotentiallyimprovesurgicalperiprostheticArtificialintelligenceHistogram-basedsiteTotal

Similar Articles

Cited By

No available data.