Management of M. abscessus subsp. abscessus early-onset prosthetic joint infection: Case report and literature review

Abstract

Nontuberculous mycobacteria are a rare but still emerging cause of difficult-to-treat prosthetic joint infection. To our knowledge only 17 cases of M. abscessus complex prosthetic joint infection are reported in literature, of which only 1 is by M. abscessus subps. abscessus. No guidelines are available for this clinical scenario.

We describe a 68-years-old female patient with an early-onset M. abscessus subsp. abscessus prosthetic joint infection, successfully treated with a tailored medical-surgical strategy, and present an overview of cases currently available in the literature to assist physicians in the management of these uncommon infections.

1. Introduction

The incidence of prosthetic joint infections (PJI) has raised in recent years as the number of patients undergoing total joint arthroplasty has increased [1]. The cumulative incidence of PJI after total hip arthroplasty (THA) and total knee arthroplasty (TKA) remains unclear; however, it is believed to be approximately 1.5 infections per 1000 person-joint-years [2]. The majority of PJI cases are attributed to gram-positive cocci, especially staphylococci (50–60%), the remaining cases are ascribed to gram-negative bacteria (6%), anaerobes (4%), fungi (1%), polymicrobial (20%) and 7% of cases are culture-negative [3], [4]. Mycobacterium sp is a rare, but still emerging cause of PJI. The most recent systematic review described 115 mycobacterial PJI in the past 30 years, of which M. tuberculosis was detected in 43% of the cases and 16 species of nontuberculous mycobacteria (NTM) were detected in the remaining 57% of the cases [5]. M. abscessus complex (MABC) is a member of the rapidly growing mycobacteria (RGM) and actually comprises three subspecies: M. abscessus subsp. abscessus (from now M. abscessus), M. abscessus subsp. massiliense and M. abscessus subsp. bolletii [6]. To our knowledge, 17 cases of MABC PJI are reported in the literature, of which only 1 is from M. abscessus [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22] (Table 1). There is no standardized treatment or strength recommendation correlating in vitro RGM susceptibility and clinical response [23]. Even though prosthesis removal has been associated with a favorable outcome at 6 and 12 months after diagnosis, the optimal surgical strategy for the treatment of RGM PJI is still unknown [24].

Table 1.

Cases of PJI due to M. abscessus complex (MABC) reported in the literature.

Author's reference	Age/sex	MABC subspecies	Type of prosthesis	Antibiotic regimen (weeks)*	Surgical intervention§	Follow-up	Outcome
Spanyer JM et al.[7]	61/female	M. abscessus subsp. massiliense	Knee	AMK + CFX + CLR (5) AMK + CFX + TGC (10)	REA. Reimplantation at 6 months	52 months	No relapse
Goldstein N et al.[8]	48/female	Not specified	Hip	IMP + AMK + AZM + CFZ (60) AZM + CFZ (20)	Debridement and PE insert exchange. Hemipelvectomy	14 months	Cured
Kim M et al.[9]	83/female	Not specified	Knee	AMK + CFX + MOX + CLR (not reported) AMK (28) + CFX + AZM (40) AZM + RIF + CIP (12) AMK + CFX + AZM (8) RIF + CIP (24)	Multiple debridements. PE insert exchange. REA. Reimplantation at 8 months. Debridement and PE insert exchange 10 months after reimplantation	51 months	No relapse
Kim M et al.[9]	71/female	Not specified	Knee	CLR + LEV + RIF (1) CFX + AMK + CLR (2) AMK (16) + CLR (24) AMK (2) + TGC + CLR (4) CLR + CIP (40)	Multiple debridements. REA. Reimplantation at 15 months	24 months	No relapse
Henry MW et al.[10]	65/male	Not specified	Left hip	TGC + CLR (6) CLR (30)	REA. Reimplantation at 3 months	72 months	No relapse
		Not specified	Right hip	TGC + LNZ (6) AZM (not reported) AMK + IMP + AZM + CFZ (6) CFZ + AZM (not reported) AMK + IMP + AZM (20)	Multiple debridements. PE insert exchange. REA. Reimplantation at 2 ½ after the latest revision surgery	48 months	No relapse
Wang SX et al.[11]	72/female	Not specified + M. fortuitum	Knee	CFX + AMK + CLR (not reported) DOX + CIP + CLR (12) DOX + CIP + CLR + AMK (8) DOX + CIP + CLR (20)	REA. Reimplantation at 4 months	10 months	No relapse
Eid AJ et al.[12]	71/female	Not specified	Knee	CFX + CLR (2)	REA. Palliative care	3 weeks	Death
		Not specified	Elbow	CFX + CLR (2)	REA. Palliative care	3 weeks	Death
Malhotra R et al.[13]	78/male	Not specified	Right knee	AMK + RFB + CLR (24) LEV + RFB + CLR (32)	REA. Debridements. Arthrodesis	18 months	No relapse
		Not specified	Left knee	AMK + RFB + CLR (24) LEV + RFB + CLR (32)	REA. Debridement. Reimplantation at 2 ½ months	18 months	No relapse
Amit P et al.[14]	71/female	Not specified	Knee	CLR + LEV + AMK (3) CLR + LEV + IMP (6) CLR + LEV (13)	REA. Debridement. Reimplantation at 6 ½ months	24 months	No relapse
Pace V et al.[15]	72/female	M. abscessus subsp. abscessus	Hip	IMP + CLR + AMK + LNZ (2) IMP + CLR + AMK (12) IMP + CLR (1) TGC + AMK (2) IMP + AMK (2) IMP + AMK + CLR (1) IMP + CLR (1) TGC + CLR (1) CLR (2)	REA. Reimplantation at 4 months	6 months	No relapse
Nengue L et al.[16]	82/male	M. abscessus subsp. massiliense	Knee	AZM + CFX (2) AZM + TGC (12) CIP (4) AZM (8) AZM + AMK + TGC (2) AZM + AMK + CFX (2) AZM + AMK + LNZ (14) AZM + LNZ (8)	Debridement. REA. Reimplantation at 11 months	6 months	No relapse
Tsuruyama Y et al.[17]	74/male	M. abscessus subsp. massiliense	Knee	IMP + AMK (5) IMP + AMK + AZM (25) IMP + AMK + AZM (not reported) AMK + AZM (not reported)	Multiple arthrocentesis	6 months	No relapse
Genovese N et al.[18]	79/male	M. abscessus subsp. massiliense S. capitis C. glucuronolyticum A. baumannii C. tropicalis S. epidermidis	Hip	VAN (6) AZM + CFX (not reported) AZM + AMK + TGC (not reported) AZM + AMK + TGC + COL + FCZ (not reported) AZM + TGC + CAZ-AZI + CPT (not reported) AZM + ERV + CAZ-AVI + CPT (not reported) AZM + CFZ (32) AZM + CFZ + TGC (not reported) AZM + CFZ + ERV (12) AZM + CFZ (56) AZM (not reported)	Multiple debridements. REA	20 months	No relapse
Petrosoniak A et al.[19]	70/female	Not specified	Hip	MEM + CLR (4) CFX + CLR (12)	REA. Reimplantation at 4 months	18 months	No relapse
Ma Q et al.[20]	73/male	Not specified	Knee	CFX + IMP + AMK (24) CLR + LNZ (56)	Debridement. REA. Reimplantation	Not reported	Cured
Park JW et al.[21]	80/female	Not specified	Knee	AMK + CTX + CLR (16)	REA	16 weeks	No relapse
Singh D et al.[22]	72/female	Not specified	Hip	CLR + IMP + AMK (8) CLR + TOB (16)	REA. Debridement		Persistence of infection

Abbreviations: AMK = amikacin, AZM = azithromycin, CPT = ceftaroline, CFX = cefoxitin, CTX = cefotaxime, CAZ-AVI = ceftazidime-avibactam, CIP = ciprofloxacin, CLR = clarithromycin, CFZ = clofazimine, COL = colistine, DOX = doxycicline, EMB = ethambutol, ERV = eravacycline, FCZ = fluconazole, IMP = imipenem, INH = isoniazid, LEV = levofloxacin, LNZ = linezolid, MEM = meropenem, MOX = moxifloxacin, PE = polyethylene, PZA = pyrazinamide, REA = resection arthroplasty, RFB = rifabutin, RIF = rifampicin, TGC = tigecycline, TOB = tobramycin, VAN = vancomycin.

^*Not reporting empirical therapy prior to NTM isolation.

^§Not reporting surgical interventions prior to NTM isolation.

We present a case of M. abscessus early-onset PJI after TKA highlighting therapeutic management and reporting a summary of MABC PJI cases described in the literature.

2. Case report

A 68-years-old female patient presented on July 22, 2021 to the ICCS Hospital in Milan lamenting fever, aggravating pain, limited range of motion and swelling of her right knee, with secreting fistula formation. Less than 2 months earlier, she underwent right TKA.

On admission, past medical history included left TKA, hypercholesterolemia, hyperuricemia, arterial hypertension, bilateral glaucoma, seasonal affective disorder and diverticular disease. Drug history included statin, allopurinol, angiotensin receptor blocker, topical beta-blocker and benzodiazepine. She reported a questionable allergy to penicillin. The physical examination evidenced swelling, redness and warmth of her right knee with tenderness and functional restriction; a fistula was present on the surgical wound. Laboratory investigation demonstrated raised C-reactive protein (CRP) without leukocytosis (Fig. 1) in conjunction with fever. Radiograph evidenced only joint effusion, without signs of periprosthetic bone resorption. Arthrocentesis was performed in the emergency room (ER), with synovial fluid culture testing negative. Therefore, a revision surgery was planned five days after admission, with debridement and polyethylene (PE) insert replacement. Empirical broad-spectrum antibiotic therapy (including levofloxacin, meropenem and daptomycin) was started on the same day. Given the lack of clinical improvement, with persistent signs of local inflammation and no microbiological isolates from routine bacterial culture (synovial fluid and periprosthetic tissue), she underwent resection arthroplasty (REA) on August 13, where the prosthesis was extracted and replaced with gentamicin and vancomycin impregnated cement spacer. Bacterial culture from sonification fluid tested negative and antibiotic therapy was interrupted on September 20, after almost two months, due to ineffectiveness. Periprosthetic fluid aspirate was collected the day after with acid-fast bacilli (AFB) growth on standard bacterial culture, identified as M. abscessus. The RGM was isolated from liquid (MGIT) and solid culture (Lowenstein-Jensen), identified using GenoType NTM-DR (Arnika kit). Genotypic drug susceptibility testing (DST) was conducted using the latter method and phenotypic DST using Sensititre™ Myco RAPMYCOI AST Plate. Results were evaluated according to the breakpoints reported in the 2018 CLSI document [25] (Table 2). On October 15 targeted antimycobacterial therapy was started with imipenem/cilastatin 500/500 mg QID plus linezolid 600 mg BID plus amikacin 750 mg QD (10 mg/kg). Six days later another revision procedure was performed, with debridement plus gentamicin loaded spacer renewal, collecting another synovial fluid sample for microbiological examination, which again tested positive for M. abscessus. On November 11 the patient was transferred to the Infectious Diseases Clinic of San Raffaele Hospital in Milan. Dosage of linezolid and amikacin was adjusted on therapeutic drug monitoring (TDM) for better tolerance and efficacy. To further implement the bactericidal capacity of the therapeutic regimen in use, cefoxitin 3 g TID was added for 14 days, after assessing the negativity of total IgE and beta-lactam-specific IgE. To optimize therapeutic management, phenotypic DST was extended to other antibiotics, according to the latest evidence [26], [27], [28], [29], [30] (Table 2). During hospitalization physical examination of the knee showed no signs of infectious relapse, inflammatory markers and leukocyte count were consistently normal (Fig. 1). The patient became highly intolerant of hospitalization, therefore on December 16 imipenem/cilastatin was discontinued in favor of clofazimine 100 mg QD and the patient was discharged, continuing amikacin 1500 mg (20 mg/kg) 3 times per week and oral linezolid (dosage on TDM) in our Day Hospital (DH) Service and starting physiotherapy at home. On March 30, 2022 after five and a half months of targeted therapy, the patient was readmitted to the ICCS Hospital, where the spacer was removed and arthrodesis was performed. Intraoperative specimen (bone sample) was collected for microbiological examination and tested negative. In the postoperative period, antibiotic therapy was interrupted on May 6 (almost 7 months of antimycobacterial therapy). At 6-months of follow up, no signs or symptoms of relapse were observed. With reference to drug toxicities, transient worsening of renal function, peripheral neuropathy eventually reversed with therapy discontinuation as well as persistent sensorineural hypoacusis and were documented.

Fig. 1.

Laboratory investigations, medical and surgical interventions. Abbreviations: AMK = amikacin, CFX = cefoxitine, CFZ = clofazimine, CRP = C-reactive protein, eGFR = estimated glomerular filtration rate, IMP = imipenem, LNZ = linezolid, PE = polyethylene, REA = resection arthroplasty.

Table 2.

M. abscessus: phenotypic DST and genetic sequencing. Clinical breakpoints according to 2018 CLSI.

Antibiotics	MIC (µg/ml)	Antibiotics	MIC (µg/ml)
Amikacin	16 (S)	Linezolid	8 (S)
Imipenem	8 (I)	Cefoxitin	32 (I)
Ciprofloxacin	4 (R)	Moxifloxacin	8 (R)
TMP-SMX	8/132 (R)	Doxicycline	> 16 (R)
Rifampicin	8	Rifabutin	8
Clarithromycin	8 (R)
Clofazimine	0.5
erm(41)wild-type: detected

Abbreviations: MIC = minimum inhibitor concentration, TMP-SMX = trimethoprim-sulphametoxazole.

3. Discussion

M. abscessus, as the others RGM, is a ubiquitous environmental organism, quite resistant to disinfectants and potentially able to cause post-surgical and post-procedural infections [6], [23], [24]. RGM are increasingly being recognized as a possible cause of early-onset PJI, with a clinical presentation that does not differ significantly from non-mycobacterial PJI [5], [24]. Their growth requires specific media, yet they can develop also into routine bacterial plates or broth culture media in about one week. Indeed, in our case no AFB culture was submitted for examination. However, their isolation can be missed when cultures are incubated for only 48 h, so it is necessary to advise microbiologist in case of a high degree of suspicion [23]. Definitive identification to the subspecies level is no longer possible without molecular techniques, since the recent changes in the RGM’s taxonomy [6]. Consequently, identification of RGM in most laboratories is either incomplete or imprecise (Table 1). In vitro antimicrobial susceptibility testing should be performed on all clinically significant isolates of RGM [23], [25]. Regimen selection depends upon the presence and activity of the erm(41) gene, which confers inducible macrolide resistance despite in vitro susceptibility at three to five days. The majority of M. abscessus and some M. abscessus subsp. bolletii contain an erm(41) gene. However, some isolates have a T-to-C polymorphism at nucleotide 28 of the erm(41) gene, which inactivates it [31]. In our case mycobacterial isolate was identified and erm(41) gene was detected by GenoType NTM-DR. Phenotypic DST showed favorable MIC for amikacin and linezolid and intermediate for imipenem and cefoxitin, while clinical breakpoints were not available for clofazimine and rifampicin/rifabutin. Extended incubation with clarithromycin was performed to assess the functionality of the gene (Table 2). Regarding antimycobacterial therapy, there are no randomized controlled trials or comparative observational studies examining the treatment of MABC PJI nor NTM PJI. Thus, treatment guidance is primarily based on DST data, as well as case reports of NTM PJI and small retrospective reviews on MABC pulmonary infections, which have contributed to the recent guidelines on NTM pulmonary infections [23], [25], [32]. For most NTM soft tissue and skeletal disease, medical therapy should be continued for a minimum of six months and is sometimes given for 12 months or longer [23]. There is a high grade of variability in the number of molecules and duration of therapy used to treat MABC PJI. As reported in Table 1 [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22] in all but two [10], [19] cases of successfully treated MABC PJI, a multidrug regimen containing at least 3 molecules was used, at least for the intensive phase. The mean duration of antimycobacterial therapy of successfully treated MABC PJI was 39 ± 23 weeks. In our case, antimycobacterial therapy was administered for 7 months with a regimen containing at least 3 drugs for the entire duration of treatment (Fig. 1). For a brief period, we combined cefoxitin with imipenem/cilastatin to increase the efficacy of the multidrug regimen in use, considering the in vitro synergism of dual β-lactams against M. abscessus proved by recent studies [28], [29]. Based on the results of the initial phenotypic DST, the range of drugs tested was broadened to maintain 3 active molecules in the simplified regimen (Table 2). Subsequently, imipenem/cilastatin was replaced with clofazimine in outpatient treatment, considering multiple evidence, despite the absence of established MIC breakpoints [26], [27]. Actually, MICs suggested for susceptibility were higher than our [26]. TDM was performed in order to minimize toxicities, nevertheless the patient still developed side effects attributable to the long treatment with the multi-drug regimen, in particular aminoglycoside-induced hearing loss [33] and reversible linezolid-related polyneuropathy [34]. Although linezolid is typically used once-daily for mycobacterial infections, PK-PD data do not exist for RGM infections, as they do for TB [35], [36], [37]. In addition, linezolid MICs are typically much lower for Mycobacterium tuberculosis (usually <1 µg/ml, even in cases of XDR-TB) than for RGMs. Considering the exposure-dependent antibacterial activity of linezolid and T > MIC at least > 85% as a PK-PD index that can predict clinical efficacy [38], it is reasonable to assume that linezolid dosing used for TB is inadequate for RGM infections and dosage should be individualized based on the MIC of the isolate (Table 2). Given the above and despite the concentration-related toxicity reported by various studies [39], but not others [40], [41], we tried to maintain C_min around 8 mg/L through a twice-daily dosing. PK-PD studies in RGM infections are warranted to determine the most appropriate linezolid dosing. We did not ask the reference laboratory to test synthetic tetracyclines (tigecycline and eravacycline) as the intravenous regimen was already adequate and we needed instead effective oral drugs in order to manage the patient with triweekly DH admissions. Note that omadacyclin is not available in Europe yet. Still, this class of antibiotics has proven to be a viable option in the treatment of M. abscessus infections [42]. As for surgical interventions, aggressive ones in the early stages of the disease may shorten the course of antimycobacterial therapy and improve infection control rate. Prosthesis removal has been associated with a favorable outcome at 6–12 months, and in NTM PJI is required in almost 90% of cases, but the optimal surgical strategy for the treatment of RGM PJI has yet to be defined [5], [24]. As reported in Table 1 [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22] 14 of 17 (82 %) successfully treated MABC PJI cases required a 2-stage revision arthroplasty, one patient [8] underwent hemipelvectomy and one [13] underwent arthrodesis. Only one patient [17] was successfully treated with debridement, antibiotics and implant retention (DAIR). Our patient underwent a revision surgery, with debridement and spacer renewal, approximately one week after the start of targeted therapy. Arthrodesis was performed after five and a half months of antimycobacterial treatment, due to significant bone loss and soft tissue compromise, collecting a sample for microbiological examination, which resulted negative. The multidrug regimen was then continued for another 5 weeks (Fig. 1).

In summary, M. abscessus is a rare but emerging cause of PJI and should be suspected particularly when dealing with early-onset, culture negative and/or refractory PJI. In the management of MABC PJI, the critical points are the combination of antimycobacterial drugs, the correlation between DST and clinical outcomes, the optimal type and timing of surgery, and the duration of antimycobacterial therapy. Based on the reported case, the proposal for the management of this uncommon infection is to tailor medical and surgical therapy according to the DST data and the clinical course.

Financial support

This work has been supported by the Infectious Disease Unit of San Raffaele Hospital.

Authors’ contribution

GM and MG conceived the report. GM performed the literature review, wrote the draft and edited the manuscript for content and wording. All co-authors were responsible for critically reviewing the draft. All authors read and approved the final manuscript.

Patient consent

The patient described in this report has signed informed consent to the use of her clinical data. No formal approval by an ethics committee is required for this type of paper.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article: