Marjan Wouthuyzen-Bakker, Kresimir Crnogaca, Marc W. Nijhof.

Response/Recommendation: Yes, the type of venous thromboembolism (VTE) prophylaxis influences the risk of subsequent periprosthetic joint infection (PJI). The strongest association is observed for vitamin K antagonists (VKA) when compared to acetylsalicylic acid (Aspirin [ASA]).

Strength of Recommendation: Moderate.

Rationale: Impaired hemostasis and bleeding in an arthroplasty wound, resulting in hematoma formation and persistent wound drainage, might favor bacterial growth and the subsequent development of PJI. Therefore, it is reasonable to assume that the risk to develop a PJI is influenced by the type of VTE prophylaxis used. We have conducted a literature search in PubMed and Embase according to the search strategy defined in the appendix. From the total of 107 articles, a final number of 23 articles met the predefined inclusion and exclusion criteria. The details of these studies are summarized in Table 1.

Most of the included studies compared the infection risk between low-molecular-weight heparin (LMWH) and direct oral anticoagulants (DOAC)^1-10. The largest analysis has been performed by Jameson et al., a retrospective multicenter observational analysis in which the authors compared the incidence of several wound complications after total hip arthroplasty (THA) and total knee arthroplasty (TKA) in 12 hospitals in the United Kingdom before and after a change of VTE prophylaxis protocol from LMWH to rivaroxaban³. One of the primary endpoints of the study was deep infection in which early reoperation was necessary. A total of 13,123 patients were included in the study, in which the infection rate was 0.53% in the LMWH group and 0.62% in the rivaroxaban group (no significant difference [NS]). A major limitation of this study was the fact that it was not possible to discriminate between surgical wound irrigation for infection or hematoma. In addition, two randomized trials comparing LMWH and DOAC were performed, in which one was too small a sample size to detect any infection complication⁵. The other, performed by Lassen et al., in patients undergoing THA or TKA (RECORD programme), a post-operative wound infection rate of 0.27% in the LMWH group was observed compared to 0.16% in the rivaroxaban group (NS)⁶. No differences were observed between THA and TKA. Again, no clear definition was provided for post-operative wound infection. Figure 1A depicts a forest plot of all studies comparing LMWH with DOAC with respect to their risk of developing an infection. Only the studies in which the absolute numbers were depicted by the authors are included. The analyzed studies showed low heterogeneity, and no differences between both types of VTE prophylaxis were observed.

The second most common comparison has been made between ASA and VKA^9-14. All of these studies were retrospective analyses. The study performed by Huang et al., was the only one that defined PJI according to the Musculoskeletal Infection Society (MSIS) criteria¹². In this study, the authors divided patients into those with a high-risk (n = 4,898) versus low-risk for VTE (n = 22,751), and consistently demonstrated a significantly lower PJI incidence in the ASA groups vs. the VKA groups (Table 1), with a PJI incidence of 0.18 versus 1.26% of the total cohort, respectively. The studies of Cafri et al., and Singh et al., showed a clear trend towards a lower infection rate in the ASA group, but this difference was NS^10,13. Tan et al., also reported a lower risk for PJI when using ASA, but absolute numbers in this study were not provided¹⁴. In the study from Agaba et al., analyzing different VTE agents⁹, warfarin, (VKA) was the only one significantly associated with the highest PJI risk, in particular in the early post-operative period, with an odds ratio (OR) of 1.44. An international normalized ratio (INR) greater than 1.5 was found to be more prevalent in patients who had post-operative wound complications and subsequent PJI¹⁹. Figure 1B depicts the forest plot of the three studies comparing ASA with VKA, including solely those studies in which the absolute numbers were depicted by the authors. With a high heterogeneity between studies, there was a significant difference observed in infection rate between the ASA and the VKA group in the pooled analysis. The dose of ASA (80 mg vs. 325 mg) does not seem to have any influence on the infection rate, either for THA or TKA when analyzed separately^15-18.

Unfortunately, only a few studies have directly compared infection rates between LMWH vs. ASA. In a retrospective analysis, Agaba et al., evaluated different types of VTE prophylaxis in 72,670 patients undergoing THA⁹. Rivaroxaban (DOAC), ASA, enoxaparin (LMWH), and fondaparinux had a significant protective effect on the development of PJI within 90 days after the index surgery, with OR of 0.27, 0.34, 0.40, and 0.47, respectively. For another DOAC, apixaban, a protective effect was not observed. With overlapping confidence intervals, the PJI risk for LMWH versus ASA was not significantly different. The largest analysis in which LMWH was directly compared to ASA is the study of Tan et al.¹⁴. In this study, 60,467 primary and revision total joint arthroplasties were retrospectively evaluated. The use of ASA was associated with a significantly lower risk for PJI development compared to LMWH and VKA (both p < 0.001). For LMWH vs. VKA, the PJI rate was lower for the high-risk VTE group only (p < 0.001). Unfortunately, no absolute numbers on PJI rates per type of VTE prophylaxis were provided in this study.

In conclusion, based on the literature review, VKA seem to be associated with the highest, and ASA (at least when compared to VKA) with the lowest risk for PJI. For LMWH and DOAC, no significant difference in PJI risk could be identified. Important limitations of the reviewed articles were the lack of a clear and adequate definition for (deep) infection and/or PJI. In addition, few studies performed multivariate analyses in which it remains unclear whether the type of VTE prophylaxis is an independent predictor for PJI.

Figure 1A.      Forest plot. Depicting studies comparing LMWH with DOAC.

LMWH=Low-molecular-weight heparin; DOAC=Direct oral anticoagulants; C.I.=Confidence interval; PJI=Periprosthetic joint infection.

Figure 1B.      Forest plot. Depicting studies comparingASA with VKA.

ASA=Aspirin; VKA=Vitamin K antagonist; C.I.=Confidence interval; PJI=Periprosthetic joint infection.

Table 1.          Overview results selected studies.

Author	Year		Joint	VTE prophylaxis	Dose	Duration	Outcome	Infection rate/odds	p-value	Study design
Agaba et al.9	2017		Hip	ASA (n = 551) enoxaparin (n = 6,791) warfarin (n = 12,008) rivaroxaban (n = 5,403) fondaparinux (n = 876) apixaban (n = 337)   ASA (n = 551) enoxaparin (n = 6,791) warfarin (n = 12,008) rivaroxaban (n = 5.403) fondaparinux (n = 876) apixaban (n = 337)	NP             NP	≤ 30 days             ≤ 30 days	PJI < 30 days1             PJI < 90 days1	OR 0.86 (0.54, 1.38) OR 0.53 (0.44, 0.65) OR 1.44 (1.26, 1.64) OR 0.36 (0.29, 0.46) OR 0.40 (0.24, 0.67) OR 1.58 (0.83, 3.01)   OR 0.47 (0.25, 0.88) OR 0.34 (0.27, 0.44) OR 1.17 (1.01, 1.34) OR 0.27 (0.20, 0.35) OR 0.40 (0.24, 0.67) OR 0.77 (0.31, 1.87)	NP             NP	Retrospective cohort             Retrospective cohort
Brimmo et al.20	2015		Hip, Knee	Rivaroxaban (n = 159) other (n = 480)2	10 – 20 mg OD	≥ 2 weeks	Deep SSI (≥ 2 cultures)	2.5% 0.2%	< 0.015	Retrospective cohort
Cafri et al.10	2017		Knee	ASA (n = 5,124) enoxaparin (n=13,318) fondaparinux (n=3,225) warfarin (n=8,832)	325 mg OD 40 – 60 mg OD 2.5 mg OD INR goal 1.8 – 2.0	NP	SSI: deep infection or revision surgery for infection related reasons < 90 days index procedure	0.39% / 1.00 0.39% / 0.90 (0.48 – 1.67) 0.41% / 0.84 (0.36 – 1.92) 0.46% / 0.80 (0.42 – 1.53)   (OR: vs. ASA)	– 0.732 /0.148 0.674/ 0.172 0.500/ 0.089   (superiority /non-inferiority)	Retrospective cohort
Chahal et al.1	2013		Hip, knee	Enoxaparin (n = 227)           rivaroxaban (n = 160)	40 mg OD           10 mg OD	6 weeks or stopped at discharge and continued on ASA   10 days for knees, 30 days for hips	Infection defined as returning to theatre < 12 months	0.88%           1.88%	NP	Comparison with retrospective cohort after change in protocol
Charters et al.2	2015		Hip, knee	Enoxaparin (n = 1,113)   rivaroxaban (n = 649)	30 mg BID for knees 40 mg OD for hips 10 mg OD for knees 10 mg OD for hips	14 days 21 days 12 days 35 days	Deep infection requiring DAIR	0.9%   0.9%	0.99	Comparison with retrospective cohort after change in protocol
Di Benedetto et al.21	2017		Hip	Rivaroxaban (n = 145) other (n = 60)3	NP	35 days	PJI < 4 weeks	0% 0%	1.00	Retrospective cohort
Feldstein et al.15	2017		Hip, knee	ASA 325 mg BID (n = 282) ASA 81 mg BID (n = 361)	325 mg BID 81 mg BID	1 month	PJI < 1 month	0% 0%	1.00	Prospective cohort
Glassberg et al.22	2019		Hip	Community insured warfarin (n = 12,876) rivaroxaban (n = 10,892)     Medicare warfarin (n = 7,416) rivaroxaban (n = 4,739)	NP NP       NP NP	No info	PJI < 90 days	0.88% 0.62% OR 1.57 (1.16, 2.13)     0.85% 0.49% OR 1.79, (1.14 – 2.81)	0.02         0.02	Retrospective cohort
Huang et al.12	2016		All joints	ASA low risk (n = 4,102) warfarin low-risk (n = 18,649)   ASA high-risk (n = 796) warfarin high-risk (n = 6,723)   warfarin high-risk (n = 6,723)	81 or 325 mg BID INR goal 1.8 -2.0   81 or 325 mg BID INR goal 1.8 – 2.0   INR goal 1.8 – 2.0	4 weeks postop 4 weeks postop   4 weeks postop 4 weeks postop   4 weeks postop	PJI < 90 days (MSIS criteria)	0.2% 1.1%   0.1% 1.7%   An OR 13.7 (1.9, 98.5)	< 0.001     0.001	Retrospective
Huang et al.11	2015		All joints	ASA (n = 1,456) warfarin (n = 1,700)   warfarin (n = 1,700)	325 BID INR goal 1.8 – 2.0   INR goal 1.8 – 2.0	6 weeks postop	PJI < 90 days	0.4% 1.5%   An OR 2.77 (1.19, 6.45)	< 0.001	Comparison with retrospective cohort after change in protocol
Jameson et al.3	2012		Hip, knee	LMWH (n = 10,361) rivaroxaban (n = 2,762)	NP	14 days knees 21 days hips	SSI and PJI requiring return to surgery < 30 days	0.53% 0.62%	0.59	Comparison with retrospective cohort after change in protocol
Jensen et al.4	2011		Hip, knee	LMWH (n = 489) rivaroxaban (n = 559)	4500 U 10 mg OD	28 days 14 days knees 28 days hips	Deep infection requiring DAIR < 30 days	1.0% 2.5%	0.10	Comparison with retrospective cohort after change in protocol
Kim et al.5	2015		Hip	Rivaroxaban (n = 350) enoxaparin (n = 351) placebo (n = 185)	10 mg OD 40 mg OD	7 – 12 days postop	PJI	0% 0% 0%	1.00	Randomized trial
Kulshrestha et al.23	2013		Knee	Routine LMWH (n = 450) Risk stratification (ASA ± LMWH) (n = 450)	40 mg OD   325 BID ± 40 mg OD	2 weeks postop   4 weeks postop ± 2 weeks postop	PJI	0.9%   0.2%		Randomized trial
Lassen et al.6	2012		Hip, Knee	Rivaroxaban (n = 6,183) enoxaparin (n = 6,200)	10 mg OD 40 mg OD or 30 mg BID	10 – 40 days	Wound-infection < 30 days	0.16% 0.27%		Randomized trial
Matharu et al.24	2020		Hip                 Knee	ASA ± LMWH (n = 28,049) direct thrombin inhibitor ± LMWH (n = 28,049)   ASA ± LMWH (n = 19,021) factor Xa inhibitor ± LMWH (n = 19,021)     ASA ± LMWH (n = 34,161) direct thrombin inhibitor ± LMWH (n = 34,161)   ASA ± LMWH (n = 25,114) factor Xa inhibitor ± LMWH (n = 25,114)	NP	NP	SSI < 90 days	OR 1.04 (0.84, 1.28)       OR 0.91 (0.70, 1.17)         OR 1.09 (0.93, 1.27)       OR 0.91 (0.75, 1.11)		Retrospective (national joint registry)
Parvizi et al.16	2017		Hip Knee	ASA 81 mg BID (n = 1,459) ASA 325 mg BID (n = 3,192)	81 mg BID 325 mg BID	4 weeks	PJI < 90 days	0.2% 0.5%	0.28	Prospective crossover study
Singh et al.13	2020		Hip, Knee	ASA 325 mg BID (n = 2,183) warfarin (n = 3,333)	325 mg BID	6 weeks	PJI < 6 months	1.4% 1.8%	0.23	Retrospective
Tan et al.14	2019		All joints4	LMWH (n = 17,554) warfarin (n = 29,303) ASA (13,610)	NP INR goal 1.8 – 2.0 81 mg or 325 mg BID	4 – 6 weeks	PJI < 90 days	No absolute numbers or % reported		Retrospective
Tang et al.17	2020		Knee 5	ASA (n = 435) ASA (n = 1,003)	81 mg BID 325 mg BID	1 month	PJI < 90 days6	0.2%   0.6%	0.36	Retrospective
Tang et al.18	2020		Hip5	ASA (n = 388) ASA (n = 973)	81 mg BID 325 mg BID	1 month	PJI <90 days6	0.77% 1.2%	0.46	Retrospective
Yen et al.7	2014		Knee	Rivaroxaban (n = 61) enoxaparin (n = 52)	10 mg once 20 mg BID	2 weeks	Need for I&D < 90 days	0% 0%	1.00	Retrospective
Zou et al.8	2014		Knee	Rivaroxaban (n = 102) LMWH (n = 112) ASA 100 mg (n = 110)	10 mg/day 0.4 ml/day 100 mg/day	14 days	Wound complications  < 4 weeks	4.9% 2.7% 1.8%	0.027, 0.014	Prospective randomized controlled trial

1. Based on ICD – 9 – CM codes 996.60, 996.66, 996.67, 998.6
2. Other: 322 enoxaparin, 161 ASA, 33 warfarin
3. Other: 25 fondaparinux, 19 warfarin/acenocumarol, 5 calciparin, 3 dabigatran, 8 nadroparin /enoxaparin.
4. High-risk patients for VTE
5. Revisions
6. Acute PJI ICD – 9 (996.6) and ICD – 10 (T84.5)

ASA=Aspirin; n=Number; NP=not presented; PJI=Periprosthetic joint infection; OR=Odds ratio; OD=once daily; mg=milligrams; SSI=Surgical site infection; INR=International normalized ratio; BID=twice daily; DAIR=Debridement antibiotics implant retention; LMWH=Low-molecular-weight heparin; MSIS=Musculoskeletal infection society; U=Units; I&D=Incision and drainage.

References:

1. Chahal GS, Saithna A, Brewster M, Gilbody J, Lever S, Khan WS, and Foguet P. A comparison of complications requiring return to theatre in hip and knee arthroplasty patients taking enoxaparin versus rivaroxaban for thromboprophylaxis. Ortop Traumatol Rehabil. 2013 Mar-Apr;15(2):125-9. doi: 10.5604/15093492.1045953. PMID: 23652532.
2. Charters MA, Frisch NB, Wessell NM, Dobson C, Les CM, and Silverton CD. Rivaroxaban Versus Enoxaparin for Venous Thromboembolism Prophylaxis after Hip and Knee Arthroplasty. J Arthroplasty. 2015 Jul;30(7):1277-80. doi: 10.1016/j.arth.2015.02.009. Epub 2015 Feb 17. PMID: 25724111.
3. Jameson SS, Rymaszewska M, Hui AC, James P, Serrano-Pedraza I, and Muller SD. Wound complications following rivaroxaban administration: a multicenter comparison with low-molecular-weight heparins for thromboprophylaxis in lower limb arthroplasty. J Bone Joint Surg Am. 2012 Sep 5;94(17):1554-8. doi: 10.2106/JBJS.K.00521. PMID: 22832942.
4. Jensen CD, Steval A, Partington PF, Reed MR, and Muller SD. Return to theatre following total hip and knee replacement, before and after the introduction of rivaroxaban: a retrospective cohort study. J Bone Joint Surg Br. 2011 Jan;93(1):91-5. doi: 10.1302/0301-620X.93B1.24987. PMID: 21196550.
5. Kim SM, Moon YW, Lim SJ, Kim DW, and Park YS. Effect of oral factor Xa inhibitor and low-molecular-weight heparin on surgical complications following total hip arthroplasty. Thromb Haemost. 2016 Mar;115(3):600-7. doi: 10.1160/TH15-07-0527. Epub 2016 Jan 21. PMID: 26790579.
6. Lassen MR, Gent M, Kakkar AK, Eriksson BI, Homering M, Berkowitz SD, and Turpie AG. The effects of rivaroxaban on the complications of surgery after total hip or knee replacement: results from the RECORD programme. J Bone Joint Surg Br. 2012 Nov;94(11):1573-8. doi: 10.1302/0301-620X.94B11.28955. PMID: 23109641.
7. Yen SH, Lin PC, Kuo FC, and Wang JW. Thromboprophylaxis after minimally invasive total knee arthroplasty: a comparison of rivaroxaban and enoxaparin. Biomed J. 2014 Jul-Aug;37(4):199-204. doi: 10.4103/2319-4170.125627. PMID: 25116715.
8. Zou Y, Tian S, Wang Y, and Sun K. Administering aspirin, rivaroxaban and low-molecular-weight heparin to prevent deep venous thrombosis after total knee arthroplasty. Blood Coagul Fibrinolysis. 2014 Oct;25(7):660-4. doi: 10.1097/MBC.0000000000000121. PMID: 24695091.
9. Agaba2017. Agaba P, Kildow BJ, Dhotar H, Seyler TM, and Bolognesi M. Comparison of postoperative complications after total hip arthroplasty among patients receiving aspirin, enoxaparin, warfarin, and factor Xa inhibitors. J Orthop. 2017 Aug 14;14(4):537-543. doi: 10.1016/j.jor.2017.08.002. PMID: 28878512; PMCID: PMC5574820.
10. Cafri G, Paxton EW, Chen Y, Cheetham CT, Gould MK, Sluggett J, Bini SA, and Khatod M. Comparative Effectiveness and Safety of Drug Prophylaxis for Prevention of Venous Thromboembolism After Total Knee Arthroplasty. J Arthroplasty. 2017 Nov;32(11):3524-3528.e1. doi: 10.1016/j.arth.2017.05.042. Epub 2017 May 31. PMID: 28634095.
11. Huang R, Buckley PS, Scott B, Parvizi J, and Purtill JJ. Administration of Aspirin as a Prophylaxis Agent Against Venous Thromboembolism Results in Lower Incidence of Periprosthetic Joint Infection. J Arthroplasty. 2015 Sep;30(9 Suppl):39-41. doi: 10.1016/j.arth.2015.07.001. Epub 2015 Jul 7. PMID: 26182982.
12. Huang RC, Parvizi J, Hozack WJ, Chen AF, and Austin MS. Aspirin Is as Effective as and Safer Than Warfarin for Patients at Higher Risk of Venous Thromboembolism Undergoing Total Joint Arthroplasty. J Arthroplasty. 2016 Sep;31(9 Suppl):83-6. doi: 10.1016/j.arth.2016.02.074. Epub 2016 Mar 15. PMID: 27094242.
13. Singh V, Shahi A, Saleh U, Tarabichi S, and Oliashirazi A. Persistent Wound Drainage among Total Joint Arthroplasty Patients Receiving Aspirin vs Coumadin. J Arthroplasty. 2020 Dec;35(12):3743-3746. doi: 10.1016/j.arth.2020.07.004. Epub 2020 Jul 9. PMID: 32788061.
14. Tan TL, Foltz C, Huang R, Chen AF, Higuera C, Siqueira M, Hansen EN, Sing DC, and Parvizi J. Potent Anticoagulation Does Not Reduce Venous Thromboembolism in High-Risk Patients. J Bone Joint Surg Am. 2019 Apr 3;101(7):589-599. doi: 10.2106/JBJS.18.00335. PMID: 30946192.
15. Feldstein MJ, Low SL, Chen AF, Woodward LA, and Hozack WJ. A Comparison of Two Dosing Regimens of ASA Following Total Hip and Knee Arthroplasties. J Arthroplasty. 2017 Sep;32(9S):S157-S161. doi: 10.1016/j.arth.2017.01.009. Epub 2017 Jan 24. PMID: 28214257.
16. Parvizi J, Huang R, Restrepo C, Chen AF, Austin MS, Hozack WJ, and Lonner JH. Low-Dose Aspirin Is Effective Chemoprophylaxis Against Clinically Important Venous Thromboembolism Following Total Joint Arthroplasty: A Preliminary Analysis. J Bone Joint Surg Am. 2017 Jan 18;99(2):91-98. doi: 10.2106/JBJS.16.00147. PMID: 28099298.
17. Tang A, Zak SG, Waren D, Iorio R, Slover JD, Bosco JA, and Schwarzkopf R. Low-Dose Aspirin is Safe and Effective for Venous Thromboembolism Prevention in Patients Undergoing Revision Total Knee Arthroplasty: A Retrospective Cohort Study. J Knee Surg. 2020 Sep 8. doi: 10.1055/s-0040-1716377. Epub ahead of print. PMID: 32898907.
18. Tang A, Zak S, Iorio R, Slover J, Bosco J, and Schwarzkopf R. Low-Dose Aspirin Is Safe and Effective for Venous Thromboembolism Prevention in Patients Undergoing Revision Total Hip Arthroplasty: A Retrospective Cohort Study. J Arthroplasty. 2020 Aug;35(8):2182-2187. doi: 10.1016/j.arth.2020.03.040. Epub 2020 Mar 30. PMID: 32334898.
19. Parvizi J, Ghanem E, Joshi A, Sharkey PF, and Hozack WJ,  Rothman RH. Does excessive anticoagulation predispose to periprosthetic infection? J Arthroplasty. 2007; 22:24-8.
20. Brimmo O, Glenn M, Klika AK, Murray TG, Molloy RM, and Higuera CA. Rivaroxaban Use for Thrombosis Prophylaxis Is Associated With Early Periprosthetic Joint Infection. J Arthroplasty. 2016 Jun;31(6):1295-1298. doi: 10.1016/j.arth.2015.12.027. Epub 2015 Dec 20. PMID: 26796776.
21. Di Benedetto P, Zangari A, De Franceschi D, Di Benedetto ED, Cainero V, Beltrame A, Gisonni R, and Causero A. Rivaroxaban and early periprostethic joint infection: our experience. Acta Biomed. 2017 Oct 18;88(4S):38-42. doi: 10.23750/abm.v88i4-S.6792. PMID: 29083351; PMCID: PMC6357653.
22. Glassberg MB, and Lachiewicz PF. Changing Patterns of Anticoagulation After Total Hip Arthroplasty in the United States: Frequency of Deep Vein Thrombosis, Pulmonary Embolism, and Complications With Rivaroxaban and Warfarin. J Arthroplasty. 2019 Aug;34(8):1793-1801. doi: 10.1016/j.arth.2019.03.057. Epub 2019 Mar 31. PMID: 31005440.
23. Kulshrestha V, and Kumar S. DVT prophylaxis after TKA: routine anticoagulation vs risk screening approach – a randomized study. J Arthroplasty. 2013 Dec;28(10):1868-73. doi: 10.1016/j.arth.2013.05.025. Epub 2013 Jun 21. PMID: 23796558.
24. Matharu GS, Garriga C, Whitehouse MR, Rangan A, andJudge A. Is Aspirin as Effective as the Newer Direct Oral Anticoagulants for Venous Thromboembolism Prophylaxis After Total Hip and Knee Arthroplasty? An Analysis From the National Joint Registry for England, Wales, Northern Ireland, and the Isle of Man. J Arthroplasty. 2020 Sep;35(9):2631-2639.e6. doi: 10.1016/j.arth.2020.04.088. Epub 2020 May 6. PMID: 32532481.