Prosthetic joint infections: 6 weeks of oral antibiotics results in a low failure rate

Abstract

Background

Need for parenteral administration and total duration of antibiotic therapy for prosthetic joint infection (PJI) are debated. We report our PJI management, in which outpatient care is privileged.

Methods

This was a retrospective multicentre cohort study of PJI managed from January 2017 to Jun 2021. Microbial diagnosis was based on surgical samples. Surgical procedures and antibiotic treatments were reported. Chronic PJI was defined by a course >1 month. Oral antibiotic therapy (OAT) was defined by exclusive use of oral antibiotics or by ≤3 days of parenteral treatments. Management failure was defined by clinical and/or microbial relapse of PJI over 24 months after surgical treatment.

Results

One hundred and seventy-two patients from 13 institutions were included: 103 were male (60%) and mean age was (±SD): 73 ± 12 years. Sites for PJI were mainly hip (50%) and knee (35%), being chronic infections in 70 cases (41%). The main bacterial genus in monomicrobial infections was Staphylococcus spp. (60%). We recorded 41 (24%) implant exchanges. An OAT was prescribed in 76 cases (44%), and the median (range) course for parenteral route was 6 days (4–180) for 96 cases. Median (range) duration of antimicrobials was 42 days (21–180). Management failure was observed in 7/76 (9.2%) cases treated with OAT and 15/96 (15.6%) treated with prolonged parenteral therapy. In multivariate analysis, risk factors for failure were a knee PJI [adjusted OR (95% CI) = 3.27 (1.27–8.40)] and a polymicrobial infection [4.09 (1.46–11.49)].

Conclusions

OAT for 6 weeks for PJI was associated with a low rate of management failure.

Introduction

Prosthetic joint infections (PJIs) are serious diseases with a significant rate of unfavourable outcome and always with a high negative impact on autonomy and well-being.¹ Moreover, treatments are expensive, and both surgery and long duration of parenteral antibiotic therapy are a source of iatrogenic events.^2–4 In a cohort study of 129 patients with bone infections receiving their antibiotic through an implantable port, rates of catheter-related complications and antibiotic adverse effects were 21% and 16%, respectively.⁵ Even administered by the oral route, considering the most frequent bacteria, i.e. Staphylococcus spp., antibiotics are still a source of the iatrogenic events, with a rate between 6% to 15% according to which drugs are used.⁶ In a recent nationwide UK study of bone and joint infections, the rate of catheter complications was 9.4%.⁴

Thus, even limited improvements of care for patients with PJI are welcome. In recent decades, general concepts on antibiotic therapies have evolved towards smaller durations and an administration by the oral route as soon as possible.⁷ Recent studies have suggested that early oral antibiotic therapy (OAT) and reduction of the duration of the antimicrobials were possible for patients with PJI, allowing a reduction from several months to a few weeks of treatment, but contradictory results have emerged recently.^4,8–11 In an open-label randomized controlled study comparing 6 versus 12 weeks of antibiotics for PJI, the rate of persistent infection was 18.1% and 8.7%, respectively, leading to the lengthening of the duration of antimicrobials.¹¹

Based on our experiences of adverse effects of parenteral therapy of antibiotics, as well as the possibility of shortened courses of antimicrobial in PJI, it seemed that we did not observe such high rates of failure in PJI management.^5,6,8 Thus, our aim was to describe the outcomes of our PJI management in our cohort of patients, focusing on the impact of a short course of OAT.

Methods

This is a multicentre cohort study including all adult patients (age ≥18 years) with PJI managed from January 2017 to Jun 2021 in private institutions in France.

The study was conducted in accordance with French legislation: patients or their relatives provided written consent for computerization of their anonymized personal data and must declare their opposition using the latter for retrospective studies. Our study was approved by our institutional review board. Also, the antibiotic audits are sponsored by the French National Health Agency as a tool of antimicrobial stewardship.

As we work in a network, surgeons and infectious diseases (ID) specialists cooperated in daily practice for the management of these patients with PJI. Indeed, for nearly all patients the surgeons communicated with the ID specialists when the clinical diagnosis of PJI was suspected or even already established by the results of surgical samples. Thus, the definition of PJI was based on positive bacteriological samples of the infected joint made through surgical procedure, whatever the reported clinical signs.

The procedure, i.e. debridement, antibiotics and implant retention (DAIR) or one- or two-stage revision was based on comorbid conditions and local anatomical settings,¹ and the ID specialists interacted with the microbiological laboratory for rapid acquisition of all data to optimize the antibiotic therapy.

At least two positive microbial cultures were obtained after incubation for at least 14 days, and both had to give the same result for skin bacteria such as CoNS or Cutibacterium acnes. Antibiotic susceptibility profiles were determined using VITEK 2 or disc diffusion methods according to the guidelines of the Antibiogram Committee of the French Society for Microbiology (CASFM/EUCAST).¹² MDR bacteria were defined by methicillin resistance for Staphylococcus spp. and production of ESBL for Gram-negative bacilli. Polymicrobial infections were defined by at least two different bacterial species found in surgical samples. Blood cultures were realized at the discretion of the physicians in charge.

Epidemiological data, microbiological results and each step of surgical and antibiotic treatments were reported. The duration of the antibiotic treatment was the sum of post-operative antibiotic therapy and of targeted antibiotic therapy. Adverse effects due to antibiotics were defined by the need for drug replacement. Chronic PJI was defined by a duration over 1 month after the initial procedure. OAT was defined by an exclusive use of oral antibiotics or by ≤3 days of parenteral treatments followed by oral treatment.

A management failure was defined by a clinical and/or microbial relapse of the PJI during the 24 month period after antimicrobial cessation, whatever the bacteria involved in relapse. Accordingly, effective cure of PJI was defined by the absence of failure after 2 years of follow-up, and if a recurrence was observed after 24 months, it was considered as a new infection if the bacteria were unrelated to the initial causative strain(s). Thus, we excluded patients who relapsed or died during the antibiotic therapy and those who died before 24 months after antibiotic therapy.

Patients who received suppressive antibiotic therapy were also excluded from the analysis.

Our primary outcome was to determine the impact of short duration of oral antibiotic therapies on the rate of management failure.

Statistical analysis

The data were analysed with StatView software version 5.0, and statistical significance was established at α=0.05. Continuous variables were compared with Student’s t-test or the Mann–Whitney non-parametric test. Proportions were compared with the χ² or Fisher’s exact test when appropriate. Logistic regression was used for multivariate analysis of the risk factors associated with management failure, and the results are presented as adjusted ORs (aORs) with their 95% CIs. Variables were selected for the multivariate analysis based on the level of significance of the univariate association with management failure (P < 0.1).

Results

From January 2017 to June 2021, 205 patients with PJI were included from 13 private institutions, in which three ID specialists worked together with 38 orthopaedic surgeons. Considering our selection criteria, 33 patients (14%) were excluded: 14 who were lost of follow-up before 24 months, 10 who died during the study period without relationship to the PJI, 5 patients for whom the surgeons did not provide required information, and 4 patients for whom a long-term suppressive antibiotic therapy was prescribed.

The main characteristics of the 172 remaining patients are detailed in Table 1. There were 103 male (60%) and 69 female (43%) patients, with mean age (±SD): 72 ± 12 years. Sites for PJI were mainly hip (n = 86; 50%) and knee (n = 65; 35%), being chronic infections in 70 cases (41%). The main bacterial genus was Staphylococcus spp. (n = 69; 40%), including 42 methicillin-resistant strains. Blood cultures were realized in 91 cases (53%), being positive in 15 cases (16%), mainly related to Staphylococcus aureus (n = 6; 40%).

Table 1.

Characteristics of the patients according to the duration of parenteral treatment

Characteristic	≤3d PAT	>3d PAT	P	Total
	n =  76 (44%)	n =  96 (56%)		n = 172
Age (years), mean ± SD	70 ± 12	73 ± 11	0.036	72 ± 12
Sex ratio (male/female)	1.37	1.59	0.63	1.49
Main comorbid conditions
No comorbid conditions	24 (32)	36 (37)	0.418	60 (35)
Cardiovascular diseases	21 (28)	30 (31)	0.605	51 (30)
Obesity (BMI >30 kg/m2)	27 (36)	27 (28)	0.299	54 (31)
Diabetes	9 (12)	18 (19)	0.216	27 (16)
Chronic renal insufficiencya	10 (13)	12 (13)	0.897	22 (13)
Othersb	5 (7)	11 (11)	0.273	16 (9)
Sites of joint infections
Hip	38 (50)	48 (50)	>0.999	86 (50)
Knee	27 (36)	33 (34)	0.875	60 (35)
Shoulder	4 (5)	6 (6)	>0.999	10 (6)
Othersc	7 (9)	9 (9)	0.970	16 (9)
Chronic infections (>1 month)	32 (42)	38 (40)	0.738	70 (41)
Bacterial species before surgery
S. aureus	31 (41)	30 (31)	0.194	61 (35)
CoNS	18 (24)	24 (25)	0.841	42 (24)
Polymicrobial infections	8 (11)	19 (20)	0.097	27 (16)
Gram-negative bacillid	8 (11)	14 (15)	0.428	22 (13)
C. acnes	10 (13)	3 (3)	0.029	13 (8)
Streptococcus or Enterococcus spp.	1 (1)	10 (10)	0.035	11 (6)
MDR strainse	12 (16)	25 (26)	0.104	37 (22)
Positive blood cultures	7/31 (23)	8/60 (13)	0.259	15/91 (16)
Surgical procedures
DAIR	61 (80)	70 (73)	0.261	131 (76)
One-stage prosthetic joint implant exchange	13 (17)	16 (17)	0.939	29 (17)
Two-stage prosthetic joint implant exchange	2 (3)	10 (10)	0.091	12 (7)
Main empirical antibiotic therapies
Daptomycin + piperacillin/tazobactam	12 (16)	30 (31)	0.019	42 (24)
Vancomycin-based therapy	8 (11)	28 (29)	0.002	36 (21)
Other parenteral drugs	16 (21)	38 (40)	0.009	54 (32)
Duration of parenteral antibiotic compounds (days)f	3 (1–3)	6 (4–180)	<0.001	5 (1–180)
Main targeted antibiotic therapies
Fluoroquinolone + rifampicin	44 (58)	40 (42)	0.034	84 (49)
Other fluoroquinolone-based therapies	10 (13)	24 (25)	0.052	34 (20)
Clindamycin-based therapies	11 (14)	7 (7)	0.126	18 (10)
Co-trimoxazole-based therapies	4 (5)	5 (5)	>0.999	9 (5)
Other antibiotic compounds	7 (9)	17 (18)	0.110	24 (14)
Total duration of antibiotic therapies (days)f	42 (37–92)	47 (21–180)	<0.001	42 (21–180)
Total duration of antibiotic therapy ≥50 days	3 (4)	34 (35)	<0.001	37 (22)
Adverse effects of antibiotic compounds	10/40 (25)	13/76 (17)	0.310	23/116 (20)
Management failure at Month 24	7 (9.2)	15 (15.6)	0.211	22 (12.7)
With the same strain (yes/no/unknown)	5/0/2	12/3/0	—	17/3/2

Patients were treated with oral antimicrobials only (n = 40) or with ≤3 days of parenteral antibiotic therapy (n = 36) (≤3d PAT). Univariate analysis.

^aGlomerular filtration rate <30 mL/min/1.73 m².

^bNeurological diseases (n = 6), ongoing cancer (n = 6), hepatic diseases (n = 2), COPD (n = 2).

^cElbow (n = 7), ankle (n = 6), backbone (n = 3).

^dEscherichia coli (n = 10), other Enterobacterales (n = 7), Pseudomonas aeruginosa (n = 5).

^eMethicillin-resistant Staphylococcus spp. (n = 34), ESBL-producing Gram-negative bacilli (n = 3).

^fMedian (range).

PJI management

The surgical management was based primarily on DAIR (n = 131; 76%), including 43/70 patients (61%) with chronic infection.

The antibiotic therapy was mainly initiated by the parenteral route (n = 132; 77%), 3 patients (2%) being exclusively treated by IV infusion. The median (range) duration of parenteral administration of the antimicrobials was 5 (1–180) days, being ≤3 days in 36 cases (27%) and >7 days in 34 cases (26%). The two most frequent parenteral antibiotic combinations were daptomycin + piperacillin/tazobactam (n = 42; 24%) and vancomycin-based treatments (n = 36; 21%).

The antibiotic therapy was immediately by the oral route in 40 cases (23%). The two most frequent enteral antibiotic combinations were fluoroquinolone + rifampicin (n = 84; 49%) and co-trimoxazole-based treatments (n = 20; 12%).

The characteristics of the patients with oral antibiotic administration only or with ≤3 days of parenteral antimicrobials (n = 76; 44%) are described in Table 1, the latter showing no difference in terms of comorbid conditions and site of infection compared with patients with longer durations of parenteral administration of the drugs. However, microbiological data and antimicrobial lines were different in accordance with the availability of enteral drugs for susceptible strains.

Information on adverse effects of the antimicrobials was available for 116 patients (67%), and observed in 23/116 cases (20%), the main one being gastrointestinal disorders related to rifampicin (n = 6; 26%).

Overall, the median (range) duration of the antibiotic treatments was 42 days (21–180), and 135 patients (78%) had <50 days of antibiotics.

Failure of PJI management and risk factors

Failure of care management was observed in 22 cases (12.8%), including 7/76 (9.2%) patients with immediate OAT or ≤3 days of parenteral antibiotics and 15/96 (15.6%) patients with initial parenteral therapy >3 days, P = 0.211. The main characteristics of patients with management failure are described in Table 2.

Table 2.

Characteristics of the patients with management failure (n = 22/172; 12.7%)

Cases	Age	Site	Chronic infection	Bacteria	MDR strains	Surgical treatment	Empirical or targeted therapya	Duration IV therapy (days)	Targeted therapy	OAT	Total duration (days)	Failure (months)
1	47	knee	no	Staphylococcus epidermidis + C. acnes	no	DAIR	—	—	levofloxacin + rifampicin	yes	42	2
2	50	knee	no	S. aureus	no	DAIR	cefazoline	1	ofloxacin + rifampicin	yes	42	1
3	55	hip	no	S. epidermidis	yes	DAIR	—	—	ofloxacin + rifampicin	yes	42	2
4	58	knee	no	S. epidermidis + C. acnes	no	DAIR	—	—	levofloxacin + rifampicin	yes	42	8
5	63	hip	no	S. aureus	yes	DAIR	dapto + pip/taz	5	levofloxacin + rifampicin	no	52	20
6	64	hip	yes	S. epidermidis + Staphylococcus hominis	yes	DAIR	vancomycin + ceftazidime	5	levofloxacin + rifampicin	no	47	4
7	65	ankle	no	S. epidermidis	no	One-stage exchange	—	—	Co-trimoxazole + rifampicin	yes	42	24
8	67	knee	no	S. epidermidis + Staphylococcus haemolyticus	yes	DAIR	dapto + pip/taz followed by dapto alone	26	linezolid	no	55	3
9	68	shoulder	no	S. aureus	no	DAIR	vancomycin	3	levofloxacin + rifampicin	yes	48	2
10	73	knee	yes	S. epidermidis + C. acnes	no	DAIR	dapto + pip/taz	4	levofloxacin + rifampicin	no	46	24
11	74	hip	no	S. aureus	no	DAIR	AMC + gentamicin	3	ofloxacin + rifampicin	yes	42	6
12	76	knee	no	S. aureus	yes	DAIR	vancomycin + ceftriaxone	5	levofloxacin + rifampicin	no	42	4
13	77	knee	no	S. aureus	no	DAIR	vancomycin	6	levofloxacin + rifampicin	no	42	15
14	82	knee	no	Staphylococcus warneri + Enterococcus faecalis + Klebsiella pneumoniae	yes	DAIR	vancomycin + ceftriaxone	42	levofloxacin + rifampicin	no	57	4
15	82	hip	no	P. aeruginosa + Serratia marcescens	no	DAIR	vancomycin + ceftriaxone	36	cefepime + ciprofloxacin	no	90	4
16	83	hip	no	S. aureus	no	DAIR	clindamycin	5	levofloxacin + rifampicin	no	42	7
17	83	knee	no	Enterobacter cloacae + E. coli + K. pneumoniae	yes	DAIR	dapto + pip/taz	6	ciprofloxacin	no	47	4
18	84	knee	yes	S. aureus	no	DAIR	vancomycin	6	ofloxacin + rifampicin	no	90	14
19	85	knee	yes	P. aeruginosa	no	One-stage exchange	dapto + pip/taz	6	cefepime + ciprofloxacin	no	42	6
20	86	knee	no	S. aureus	yes	DAIR	vancomycin	7	doxycycline + rifampicin	no	120	13
21	88	hip	yes	Streptococcus mitis	no	DAIR	amoxicillin	21	amoxicillin	no	50	15
22	89	knee	yes	Streptococcus gordonii	no	DAIR	amoxicillin	18	amoxicillin	no	42	1

MDR strains were defined by methicillin resistance for Staphylococcus spp. and ESBL production for Gram-negative bacilli. OAT was defined by exclusive use of oral antibiotics or by ≤3 days of parenteral treatment followed by oral treatment. Thirteen patients out of 22 had blood cultures, which were positive only for Case 15. We did not observe the emergence of resistant bacteria.

AMC, amoxicillin/clavulanate; dapto + pip/taz, daptomycin plus piperacillin/tazobactam.

^aThe first line of antibiotic therapy could be empirical or targeted right from the start.

The microbiological investigations at the time of relapse showed the same strain in 17 cases (77%), with unchanged susceptibilities to antibiotics, a different strain in 3 cases, and in 2 cases the diagnostic of relapse was based on clinical signs only. Of note, three patients presented with a new infection (i.e. with a new bacterium) of the joint during the follow-up over 24 months.

In a logistic regression (see Table 3), the two risk factors associated with management failure were knee localization of PJI [aOR (95% CI) = 3.27 (1.27–8.40), P = 0.013] and polymicrobial infections [aOR (95% CI) = 4.09 (1.46–11.49), P = 0.007].

Table 3.

Main characteristics of the patients with prosthetic joint infection (PJI) and risk factors of management failure

Characteristic	Success	Failures	P	aOR (95% CI)
	n =  150 (87.2%)	n =  22 (12.8%)
Age (years), mean ±SD	72 ± 11	73 ± 13	0.545
Sex ratio (male/female)	1.58	0.50	0.311
Main comorbid conditions
No comorbid conditions	48 (32)	12 (55)	0.038
Cardiovascular diseases	48 (32)	3 (14)	0.078
Obesity (BMI >30 kg/m2)	49 (33)	5 (23)	0.348
Diabetes	24 (16)	3 (14)	0.776
Chronic renal insufficiency	19 (13)	3 (14)	>0.999
Others	14 (9)	2 (9)	>0.999
Sites of joint infections
Hip	79 (53)	7 (32)	0.067
Knee	47 (31)	13 (59)	0.010	3.27 (1.27–8.40)
Shoulder	9 (6)	1 (5)	>0.999
Others	15 (10)	1 (5)	0.667
Chronic infections (>1 month)	64 (43)	6 (27)	0.169
Bacterial species before surgery
S. aureus	52 (35)	9 (41)	0.567
CoNS	38 (25)	4 (18)	0.465
Polymicrobial infections	19 (13)	8 (36)	0.004	4.09 (1.46–11.49)
Gram-negative bacilli	21 (14)	1 (5)	0.363
C. acnes	13 (9)	0	0.315
Streptococcus or Enterococcus spp.	9 (6)	2 (9)	>0.999
MDR strains	29 (19)	8 (36)	0.069
Positive blood cultures	14/78 (18)	1/13 (8)	0.600
Surgical procedures
DAIR	111 (74)	20 (91)	0.082
One-stage prosthetic joint implant exchange	27 (18)	2 (9)	0.297
Two-stage prosthetic joint implant exchange	12 (7)	0	0.353
Antibiotic therapy
Administration of antimicrobials
Oral antibiotic therapy only	36 (24)	4 (18)	0.546
≤3 days of parenteral administration	33 (29)	3 (17)	0.422
>4 days and ≤7 days of parenteral administration	52 (46)	10 (56)	0.432
>7 days of parenteral administration	29 (25)	5 (28)	0.833
Including parenteral antimicrobials only	3 (2)	0	—
Total duration of parenteral antimicrobials (days)a	5.0 (1–180)	6.0 (1–42)	0.164
Main empirical therapies by parenteral administration
Daptomycin + piperacillin/tazobactam	37 (25)	5 (23)	0.843
Vancomycin-based therapy	28 (19)	8 (36)	0.056
Other parenteral drugs	49 (33)	5 (23)	0.348
Main oral antibiotic therapies
Fluoroquinolone + rifampicin	70 (47)	14 (64)	0.137
Other fluoroquinolone-based therapies	31 (21)	3 (14)	0.626
Clindamycin-based therapies	18 (12)	0	0.178
Co-trimoxazole-based therapies	8 (5)	1 (5)	>0.999
Other antibiotic compounds	20 (13)	4 (18)	0.776
Adverse effects of antibiotic compounds (n = 116)	17/99 (17)	6/17 (35)	0.083
Total duration of antibiotic therapies (days)a	42 (21–180)	44 (42–120)	0.324
Total duration of antibiotic therapies ≥50 days	30 (20)	7 (32)	0.207

Univariate and logistic regression analyses.

^aMedian (range).

Discussion

Our study shows that for patients with PJI, antibiotic therapy exclusively by the oral route or for ≤3 days by IV infusion and for less than 50 days, was associated with a similar rate of management failure with a follow-up of 24 months after surgical treatment (9.2%), compared with other antibiotic therapy options (15.6%). The analysis did not show any relationship between management failure and a 6 week course of antibiotic therapy. Also, the type of surgical procedure (DAIR versus one- or two-stage surgery) as well as the occurrence of adverse effects to the drugs were not related to management failure. Moreover, the risk factors of management failure (knee site and polymicrobial infection) are not modifiable.

Our study has some limitations, the first one being its retrospective nature. However, data needed to study PJI are not numerous and most of the time available in patients’ charts and thus easy to report (except the occurrence of adverse events of antibiotics, for which surgeons and ID physicians do not pay the same attention in our experience). The second major limitation is that our study was not a comparative trial. However, it is hard to establish a strict comparative trial in the field of PJI because there are at least four highly variable elements to integrate in the decision process: the patient, the surgical means, the bacteria and, depending on susceptibilities, the antibiotic therapy. This is the main reason herein to talk about management failure, and not failure of antibiotic treatment or of surgical procedure. Accordingly, in the recent trial in PJI comparing 6 or 12 weeks of antibiotic therapy, some different trends appeared between groups: S. aureus represented 38% versus 30% of the strains, respectively, and surgical procedures were not randomized, with most treatment failures with 6 weeks of antibiotics being associated with DAIR, leading to the question of the true risk factor of failure: surgical procedure or duration of antibiotics?¹¹ This comparative trial also showed a large heterogeneity of antibiotic treatments: 16 different compounds were used for the initial parenteral treatment, the median (IQR) time of IV infusion of antibiotics being 9 days (5–15). Indeed, the most frequently prescribed parenteral molecule was vancomycin, in 51%, which is not the easiest drug to prescribe: it requires a loading dose, pharmacological monitoring for comparison with MIC, and it has a poor cortical bone diffusion.^13–15

The rate of management failure of PJI observed in our study (12.8%) is comparable to several other recent studies, and it was measured between 13% and 17% for total knee arthroplasty in a recent literature review.^10,16,17 Our results are also in accordance with a recent meta-analysis showing a trend toward a better outcome when DAIR was associated with a short duration of antimicrobials compared with implant exchange.¹⁸

This significant efficacy of OAT for PJI is based on basic pharmacokinetic/pharmacodynamic (PK/PD) concepts.^14,15,19 The important thing is to match the local concentration of the antibiotics to the bacterial MIC. Therefore, the route of administration is irrelevant for antibiotics with good bioavailability and good bone diffusion such as rifampicin or fluoroquinolones, for example. Also, we took into consideration local factors such as biofilm, which artificially increases the MIC, and 90/103 patients (87%) with staphylococcal PJI received rifampicin.²⁰ Thus, our oral antibiotic combinations were based on molecules, such as fluoroquinolones, clindamycin or co-trimoxazole, recognized for their high bioavailability and satisfactory bone diffusion. These drug combinations replaced, for several days, other parenteral combinations with poorer bone diffusion such as vancomycin or β-lactams.^13–15 Accordingly, we observed a trend toward the use of vancomycin as empirical parenteral treatment and management failure (see Table 2). In clinical practice, some patients with bacteraemia associated with staphylococcal PJI have received prolonged parenteral antibiotic therapy (14 days) to follow some specific recommendations.²¹ However, for most, the advantages of OAT such as the decrease of catheter-related complications, the reduction of hospital stay and costs, and improvement of patient’s comfort favoured its use.^22,23 Of note, early OAT has been associated with the emergence of rifampicin-resistant bacteria using the combination of vancomycin plus rifampicin for post-osteosynthesis infections.²⁴ But vancomycin plus rifampicin could be a false combination therapy in the first days of the treatment due to a significant difference of bone diffusion between these two drugs. By the way, for the treatment of prosthetic valve endocarditis due to S. aureus, rifampicin is prescribed after a few days of effective therapy once the bacteraemia has been cleared.²⁵

Lastly, among risk factors of management failure described in the literature, knee site was regularly cited, as well as polymicrobial infections, although to a lesser extent.^1–3

In conclusion, our study shows that exclusive OAT or with ≤3 days of IV infusion prescribed for 6 weeks in PJI led to significant efficacy. Thus, except for MDR strains, most courses of parenteral antimicrobials could be reduced to the peri-operative period.

Acknowledgements

We thank the surgeons who took part in patient management: Drs Pierre Petit, Pierre Madezo, Vincent Martinel, Patricio Maltes Fermandois, Clinique Ormeau-Pyrénées, Tarbes; Drs Samia Awtel, Pascal Bréhier, Marc Blaysat, Polyclinique Les Fleurs, Ollioules; Drs Pascal Fargues, Julien Normand, Gilles Mariambourg, Clinique du Sidobre, Castres; Drs Jean Ballestro, Jean-Marc Durand, Emilien Engel, Clinique Guillaume du Varye, St Doulchard; Drs Claude Serra, Marc Perraud, Jean-Charles Grillo, Erick Meyer, Clinique du Cap d’Or, La Seyne-sur-Mer; Dr Sophie Cammas, Clinique Jean Villar, Bruges; Dr Jacques Soyer, Clinique Inkermann, Niort; Dr Pierre Leguilloux, Clinique St Michel, Toulon; Drs Marie-Pierre Mirous, Jean-Yves Simonet, Olivier Peter, Stéphane Naudi, Nicolas Vendemmia, Polyclinique St Roch, Cabestany; Drs Nicolas Maisse, Pierre D’Arzac, Clinique Montréal, Carcassonne; Drs Jean-Christophe Godchaux, Damien Girerd, Jean-Christophe Masset, Clinique Jeanne d’Arc, Arles; and Drs Jean Vaquier, Jérôme Alain, Jacques Fourastier, Yann Maczuk, Stéphane Louisa, Cédric Coste, Virginie Vacquerie, Anthony Dotzis, Polyclinique de Limoges, site Chénieux, Limoges—all in France.

Funding

This study was carried out as part of our routine work.

Transparency declarations

None to declare.

Data availability

The data used during the current study are available from the corresponding author on reasonable request.