Extended Antibiotic Prophylaxis May Be Linked to Lower Peri-prosthetic Joint Infection Rates in High-Risk Patients: An Evidence-Based Review

Abstract

In 2018, Inabathula et al. published the results of a historical control study examining 90-day peri-prosthetic joint infection (PJI) rates at a single center before and after the institution of an extended post-operative oral antibiotic protocol for high-risk total hip arthroplasty (THA) and total knee arthroplasty (TKA) patients. In the study, “Extended Oral Antibiotic Prophylaxis in High-Risk Patients Substantially Reduces Primary Total Hip and Knee Arthroplasty 90-Day Infection Rate” (J Bone Joint Surg Am. 2018;100[24]:2103–2109), the authors considered any one of several patient-related criteria—including diabetes, a body mass index of 35 kg/m² or higher, or active smoking—as sufficient to designate a patient “high risk” for PJI. Before the extended antibiotic therapy protocol was instituted, PJI rates for primary THA and TKA in the high-risk subgroup were 4.3% and 2.1%, respectively. After it was instituted, the respective rates dropped to 1.1% and 0.4%, comparable to those seen in the non-high-risk patients treated using standard peri-operative prophylaxis. After adjusting for patient factors, regression analysis showed that high-risk patients receiving only peri-operative antibiotics were more likely to develop PJI than high-risk patients receiving the extended antibiotic protocol. Although these results suggest possible benefits of extended antibiotic prophylaxis in arthroplasty, methodologic limitations and inadequate discussion of potential drawbacks of widespread adoption of such protocols limit the impact of the findings. Future research is warranted to more narrowly define risk factors for PJI and to demonstrate the safety and efficacy of extended antibiotics in reducing the long-term burden of PJI.

Introduction

Research has projected that by 2030, more than 4 million primary total hip arthroplasty (THA) and total knee arthroplasty (TKA) procedures and 96,700 and 268,000 revision THA and TKA procedures, respectively, will be performed annually in the USA [12]. With more than 25% of revisions attributable to peri-prosthetic joint infection (PJI) [9], reduction of PJI rates is an important target for reducing the costs and morbidity related to arthroplasty.

Decades of laboratory [4] and clinical [5] studies have demonstrated the benefit of peri-incisional antibiotic prophylaxis in surgery. Peri-operative prophylactic intravenous antibiotic therapy is used almost universally as a means of reducing the incidence of surgical site infection and PJI after arthroplasty, with guidelines in favor of appropriate prophylactic antibiotic dosing within 1 h before arthroplasty surgery [14]. However, the US Centers for Disease Control and Prevention and the World Health Organization recommend against prophylaxis after skin closure [1]. Despite these recommendations against it, the American Association of Hip and Knee Surgeons continues to recommend post-operative antibiotics for 24 h after inpatient arthroplasty [18]. Prophylactic antibiotic use longer than 24 h is not known to be part of any major organization’s recommendations.

Although long-term antibiotic therapy after confirmation of PJI is part of standard treatment [2, 17], little is known regarding whether extended courses of prophylactic antibiotics prevent PJI after primary THA and TKA. A 2018 study by Inabathula et al. [8] examined the effect of extended post-operative oral antibiotic prophylaxis for 7 days in lowering rates of PJI after THA and TKA in high-risk patients. The authors based the designation of high risk on several patient-related criteria and examined PJI rates with and without extended antibiotic therapy. In this review, we detail the study methods and evaluate its findings. We also comment on the study’s strengths, weaknesses, and potential influence on practice and research.

Article

“Extended Oral Antibiotic Prophylaxis in High-Risk Patients Substantially Reduces Primary Total Hip and Knee Arthroplasty 90-Day Infection Rate.”

Inabathula A, Dilley JE, Ziemba-Davis M, Warth LC, Azzam KA, Ireland PH, Meneghini RM. J Bone Joint Surg Am. 2018;100(24):2103–2109.

In this retrospective study, Inabathula et al. investigated whether 7 days of oral antibiotics reduces the 90-day risk of PJI after primary THA and TKA in high-risk patients. They considered 2181 inpatient primary THAs and TKAs performed between December 2011 and December 2016 at their tertiary care academic institution in Indiana. Medical clearance and weekly conferences were used to assess all subjects for high-risk criteria (Table 1).

Table 1.

Criteria for high-risk designation^a

BMI ≥ 35

Diabetes mellitusb

Active smoking

Chronic kidney disease

Autoimmune diseasec

Nasal colonization with MSSA or MRSAd

^a≥ 1 criterion designates high risk

^bIdentified using chart, interview, and hemoglobin A_1c

^cIncluded rheumatoid, psoriatic, and inflammatory arthritis; systemic lupus erythematosus; inflammatory bowel disease and ulcerative colitis; multiple sclerosis; psoriasis; Grave disease and hypothyroidism; Hashimoto’s disease and thyroiditis; myasthenia gravis; vasculitis, celiac disease; pernicious anemia; vitiligo; scleroderma and systemic sclerosis; Addison’s disease; and Sjögren’s and sicca syndromes

^dAdded in July 2016

BMI body mass index, MSSA methicillin-sensitive Staphylococcus aureus, MRSA methicillin-resistant S. aureus

Before January 2015, the hospital used standard peri-operative infection-control practices—or standard care (SC)—to minimize the risk of PJI. SC consisted of nasal screening for methicillin-resistant Staphylococcus aureus (MRSA), administration of intravenous antibiotics within 1 h before surgery, povidone-iodine irrigation before wound closure, and 24 h of post-operative antibiotics. Beginning in January 2015, a protocol was instituted in which 7 days of post-discharge oral antibiotics (Table 2) were given to high-risk patients, in addition to SC. Numbers of PJIs were determined from record review.

Table 2.

Oral antibiotic regimens used in the reviewed study

Antibiotic	Dose	Frequency and duration
Cefadroxil	500 mg	Two times per day for 7 days
Clindamycina	300 mg	Three times per day for 7 days
Sulfamethoxazole and trimethoprimb	800 mg/160 mg	Two times per day for 7 days

^aReplaced cefadroxil if patient had a history of anaphylaxis reaction to cephalosporins

^bReplaced cefadroxil if nares swab for methicillin-resistant Staphylococcus aureus before surgery was positive

Three patient groups were examined: group A included patients meeting no high-risk criteria who received SC peri-operatively, group B included high-risk patients who received only SC peri-operatively (those treated before January 2015), and group C included high-risk patients who received both SC and extended oral antibiotics (high-risk patients treated from 2015 on). Results are summarized in Table 3. For high-risk patients treated with SC only (group B), 90-day PJI rates were 4.1% for THA and 2.1% for TKA, four to five times the rates seen in high-risk patients who received both SC and extended oral antibiotics (group C) (THA, p = 0.034; TKA, p = 0.038).

Table 3.

Patient groups examined in the study

Group	Risk status	Treatment	Study period	THA		TKA
				n = 894	PJI cases (%)	n = 1287	PJI cases (%)
A	Not high risk	Standard peri-operative antibiotic therapy	Dec. 2011 through Dec. 2016 (61 months)	336	5 (1.5)	369	1 (0.3)
B	High risk	Standard peri-operative antibiotic therapy	Dec. 2011 through Dec. 2014 (37 months)	282	12 (4.3)	468	10 (2.1)
C	High risk	Standard peri-operative antibiotic therapy plus extended oral antibiotic therapy	Jan. 2015 through Dec. 2016 (24 months)	276	3 (1.1)	450	2 (0.4)

THA total hip arthroplasty, TKA total knee arthroplasty, PJI peri-prosthetic joint infection

Logistic regression models were fit to account for potential confounders in baseline data. Study group was predictive of infection in both THA and TKA (group B vs group C for THA, odds ratio [OR] 4.0; 95% confidence interval [CI] 1.1 to 14.5; for TKA, OR 4.9; 95% CI 1.1 to 22.4). Among TKA subjects, male sex was also predictive of infection (OR 7.2; 95% CI 2.0 to 26.2). Extended antibiotic use was associated with lower PJI rates in some notable subgroups (Table 4); no infections were reported in patients with diabetes mellitus or morbid obesity who were given extended antibiotics. No clinically significant between-group differences were found in antibiotic-related complications.

Table 4.

PJI Rates in selected subgroups

Subgroup	PJI rate after THA		PJI rate after TKA
	Standard peri-operative antibiotic therapy	Standard peri-operative antibiotic therapy plus extended oral antibiotic therapy	Standard peri-operative antibiotic therapy	Standard peri-operative antibiotic therapy plus extended oral antibiotic therapy
BMI ≥ 40	2.1%	0%	11.1%	0%a
Diabetes mellitus	3.6%	0%a	4.5%	0%
Active smoking	1.8%	3.9%	7.9%	0%

^aStatistically significant (p < 0.05)

PJI peri-prosthetic joint infection, THA total hip arthroplasty, TKA total knee arthroplasty, BMI body mass index

Commentary

Debate continues regarding the optimal prophylactic peri-operative antibiotic regimen in THA and TKA, especially because of the high morbidity associated with PJI. Inabathula et al. found significantly lower rates of 90-day PJI among high-risk patients who were given 7 days of post-discharge oral antibiotics. The authors concluded that they hoped their findings would “serve as a catalyst for further study and deliberation regarding extended antibiotic prophylaxis in high-risk (total joint arthroplasty), weighed appropriately against potential adverse consequences [8].” In commenting on this study, we must explore its limitations and determine how much influence it should have.

In the study, rates of tranexamic acid and liposomal bupivacaine use varied substantially between groups, and surgical volume for high-risk patients was higher after the new antibiotic protocol was instituted (11.5 cases per month in group C vs 7.6 cases per month in group B). This suggests that potentially confounding changes in peri-operative practice were not accounted for. Without prospective randomization, the methods used (such as regression) cannot account for unrecognized or unrecorded confounders, leaving the results vulnerable to error. Undercounting of PJIs may also have been an issue because substantially longer courses of antibiotics might have delayed the occurrence of some PJIs until after the 90-day mark. Further, although the antibiotic protocol is defined as a 7-day course, the methods section notes that patients were given antibiotics for a minimum of 7 days after discharge. It is also not clear whether patients continued to receive parenteral antibiotics after 24 h but prior to hospital discharge. These issues could affect the validity of the findings.

The small number of PJIs recorded renders the statistical significance of the findings tenuous. In fact, by reviewing a table in the study called “Rate of Infection Within 90 Days Based on Number of Risk Factors and Receipt of Extended Oral Antibiotic Prophylaxis,” we found that one more PJI in the extended antibiotic group for THA would increase the associated p value from 0.034 to 0.074. Likewise, one more PJI in the extended antibiotic group for TKA would increase the associated p value from 0.038 to 0.092. Therefore, the reliability of the reported findings may be limited. Risk of PJI may also be significantly affected by institution size, environmental factors (such as whether the setting is urban or rural), surgical volume, and surgical complexity.

It should also be noted that the PJI rates under the standard antibiotic protocol in high-risk patients were significantly higher than the 1 to 2% rates cited in the literature [6, 9, 11]. These rates decreased to less than 2% when the extended antibiotic protocol was introduced later in the study period. In identifying the root cause of this phenomenon, it is important to consider whether other undisclosed hospital-based interventions (such as changes in skin preparation, staffing changes, or installation of new operating rooms or sterilizers)—and not just the change in antibiotic prophylaxis—account for this. Overall, this study may be limited not only by a lack of generalizability related to the study environment but also by the flaws inherent in historical control studies.

Readers should also note the study’s 90-day follow-up. Any risk of infection from hematogenous spread or adjacent disease after the acute post-surgical phase is likely not affected by the short course of antibiotics investigated. Recent research estimates that more than half of PJIs after THA and TKA occur after the initial 90 post-operative days [13], suggesting that addressing risk factors for subacute PJI—such as obesity and glucose control—may result in greater cumulative decreases in infection rates. Longer studies would be needed to evaluate whether a short course of post-operative antibiotic prophylaxis actually lowers the 2-year and 10-year PJI rates. This is especially important because extended antibiotic prophylaxis may delay but not prevent some PJIs.

Perhaps the most cited negative consequence to expanded antibiotic prophylaxis is increased antimicrobial resistance [3, 7, 15]. Induction of antimicrobial resistance is of utmost importance, and, as Khatod wrote in an invited commentary on the article by Inabathula et al., “The surgeon must also be a steward of antibiotic usage” [10]. This is important to remember, as the proposed extended antibiotic protocol would indicate a major proportion of arthroplasty patients for extended antibiotics. In fact, using 2030 projections for THA and TKA volume [12] along with the proportions of subjects indicated for extended antibiotics in the reviewed study (62% for THA and 70% for TKA), over 350,000 THA cases and over 3.4 million TKA cases would be expected to be indicated for extended antibiotics in 2030. This would introduce over 50,000 patient-years of oral antibiotics annually in the USA, potentially increasing the pressure for creating multidrug-resistant organisms (one patient–year is an exposure of 1 patient taking antibiotics for 1 year). Benefits in decreased infection rates with extended antibiotics will need to be carefully weighed against potential patient-specific risks, but also against institutional and public-health needs to reduce the spread of multidrug-resistant organisms.

Beyond concerns about antimicrobial resistance, the risk of adverse drug events must be considered. Although Inabathula et al. reported no differences in adverse event rates between groups, their retrospective chart review likely neither fully captured secondary outcomes nor was adequately powered to identify such differences.

Considering that the long-term risk of PJI is reported in the literature to be only 1 to 2% [9], designating 62% of THA and 70% of TKA patients as high risk (as in this study) may not be justified. Future work could more specifically define high-risk subgroups, which may allow for targeted prophylaxis while minimizing the induction of antimicrobial resistance. The development of a risk score may also help in choosing patients for targeted prophylaxis.

We believe it is premature to change practice on the basis of this study’s results. Expanded use of antibiotic prophylaxis is an issue not restricted to surgeons; all physicians must be antimicrobial stewards. This study raises many questions that need careful attention before ushering in widespread extended antibiotic prophylaxis for arthroplasty. Ideally, large, multicenter, randomized trials of arthroplasty patients that include close monitoring for patient- and institution-level complications would yield the quality of data needed to appropriately influence practice across such a broad patient group. This is especially true as new high-quality research continues to find no evidence in favor of antibiotic prophylaxis past 24 h after incision for arthroplasty [16]. When the effects of extended antibiotic prophylaxis are borne out in larger prospective, high-quality studies, physicians can then fully consider the costs and benefits of such practices. We must also better understand and appreciate how widespread adoption of extended antibiotic prophylaxis in high-risk arthroplasty patients would affect the collective antibiogram. Once we have built a larger view of the potential risks and costs of extended antibiotic therapy, we can only then appropriately decide whether it is worthwhile.

Compliance with Ethical Standards

Conflict of Interest

Christopher J. DeFrancesco, MD; Michael C. Fu, MD, MHS; Cynthia A. Kahlenberg, MD; and Andy O. Miller, MD, declare that they have no conflicts of interest. Mathias P. Bostrom, MD, reports personal fees from Smith & Nephew, outside the submitted work.

Human/Animal Rights

N/A

Informed Consent

N/A

Required Author Forms

Disclosure forms provided by the authors are available with the online version of this article.