Periprosthetic joint infection with Ureaplasma parvum detected by 16S rRNA PCR

Abstract

Usual culture methods to identify pathogenic bacteria can be unsuccessful, particularly when working with fastidious organisms. We present a case of early periprosthetic knee joint infection with Ureaplasma parvum only identified using 16S ribosomal RNA PCR. This case represents the impact molecular methods of bacterial identification can have on clinical care allowing for more targeted antimicrobial therapy; something which is imperative in an era of increasing antimicrobial resistance.

Background

Ureaplasmas are difficult to grow, small, pleomorphic bacteria that are colonisers of the genital tract. Although Ureaplasma urealyticum is a recognised cause of pneumonia in infants it is unusual, but possible, for ureaplasmas to cause extragenital infection in the adult immunocompetent population.^1–4 Historically, Ureaplasma sp have rarely been implicated in the context of deep infection (including those affecting prosthetic material), however, it is unknown whether this is due to their genuine absence or our inability to identify the organisms using traditional culture techniques.⁵

A crucial aspect of the Ureaplasma morphology is their lack of a cell wall.³ This is particularly significant in the context of infection where the empirical antibiotic to be administered prior to or without microbiological identification would be a beta-lactam-based or glycopeptide-based medication. These antibiotics work by inhibiting cell wall formation and would therefore have no impact on the growth of Ureaplasma.

An increasing number of patients are requiring joint replacement surgery with the UK National Joint Registry reporting a total of 273 364 primary total knee replacements between January 2017 and December 2019.⁶ Rates of periprosthetic joint infection have been estimated at between 1% and 2% with significant associated morbidity including poorer patient outcomes, longer inpatient stays and increased healthcare costs.⁷

Case presentation

A 72-year-old-man with a background of anaemia, chronic kidney disease, asthma and hypertension presented 15 days post uncomplicated left total knee replacement (for osteoarthritis) with a 24-hour history of swelling and erythema at the wound site. He was systemically well, apyrexial and haemodynamically stable on admission. He was independent, had no known drug allergies, was a non-smoker and did not consume any alcohol.

On examination, the left knee was hot, swollen and erythematous with a fluctuant mass under the suture line and bloody exudate. Full range of movement was preserved and the leg was neurovascularly intact. Intravenous flucloxacillin was initially commenced for wound cellulitis. On day 7 of flucloxacillin (postoperative day 22), the patient proceeded to have a debridement, antibiotics and implant retention (DAIR) procedure as the wound had continued to deteriorate with a purulent discharge and continued skin erythema. As per local standard practice, empirical treatment with meropenem and vancomycin was commenced while awaiting further culture results.

Five deep tissue samples were sent for analysis. Gram stains were negative after 7 days of incubation and no growth was seen on culture plates. At this point, antibiotics were consolidated to teicoplanin as per standard practice for culture- negative periprosthetic infection in our trust. Due to the complex nature of this infection with prosthetic material in situ, the samples were sent for further testing with 16S ribosomal RNA (16S rRNA) PCR. Within a 2-week period, both superficial and deep tissue samples returned positive for Ureaplasma parvum. The teicoplanin that the patient had been receiving prior to the 16S rRNA PCR result would not be appropriate therapy for U. parvum so this was changed to doxycycline.

Due to poor soft tissue coverage postoperatively, a gastrocnemius flap was considered. In preparation for this, a repeat aspirate was performed 3 weeks post DAIR to ascertain if the knee was clear of infection. Fortunately, the soft tissue improved without intervention but, interestingly, U. parvum was identified by 16S rRNA PCR on the repeat aspirate. This helped to confirm genuine infection with U. parvum.

Outcome and follow-up

A total 12-week course of doxycycline was administered to the patient with no further surgical intervention. A revision knee replacement was fortunately avoided and the patient experienced no flare after stopping antibiotics. The patient was followed up for a total of 28 months and remained infection free throughout this time. Treatment of U. parvum infection with a DAIR followed by targeted antibiotics has resulted in a successful cure for this patient.

Discussion

Ureaplasmas have been implicated as a pathogen in neonatal pneumonia but they may also be associated with chorioamnionitis leading to preterm birth and colonisation of the neonatal respiratory tract leading to bronchopulmonary dysplasia.² They are also known to cause infections of the urogenital tract including non-gonococcal urethritis.⁸ They are significantly less well-established pathogens of other anatomical sites and the finding in this case of U. parvum from a prosthetic site was unusual.

The use of molecular diagnostics allowed prompt identification of the causative organism and timely administration of targeted antimicrobial therapy. As the patient was not septic, it would have been ideal to send microbiological samples prior to commencing antibiotics, however, the admitting team felt the cellulitis was severe enough to warrant starting therapy in the absence of a septic picture. From a microbiological perspective, this is unfortunate as the administration of antimicrobials prior to sampling increases the risk of negative cultures.⁹ Ideally, microbiological samples should have been acquired before commencing antibiotics, however, the isolation of U. parvum using standard culture techniques is not routine. The identification of U. parvum on separately taken, multiple superficial and deep tissue samples using molecular testing was convincing for this being the true pathogen.

To the best of our knowledge, there have only been three other reported cases of Ureaplasma periprosthetic joint infection in the immunocompetent adult population—one involving late infection of a total knee arthroplasty and two others describing infection post hip arthroplasty.^10–12 One report was found describing periprosthetic joint infection with U. urealyticum in an immunocompromised patient.¹³ Unfortunately, in two of the previously reported cases, difficulties identifying the pathogen resulted in delayed administration of effective antibiotics.^{10 11} These cases also document the difficulty in identifying Ureaplasma sp by traditional culture methods and the advantage of using 16S rRNA PCR to identify the organism.

Identification of a pathogen has a significant impact on our choice of systemic antimicrobial therapy but should also be considered in the local delivery of antibiotics. Cement impregnated with clindamycin, vancomycin or gentamicin is usually used in the process of joint revision to cover the most common pathogens causing periprosthetic infections. These include Staphylococcus aureus, coagulase-negative staphylococci and less frequently streptococci and enterococci.¹⁴ Unfortunately, Ureaplasma are not susceptible to the aforementioned antibiotics and the antibiotics to which Ureaplasma are susceptible (tetracyclines/macrolides/fluoroquinolones) cannot be used in cement due to heat instability.³ If our patient had required revision surgery he would have received gentamicin containing surgical cement; this would not have been active against U. parvum but would have covered other potentially infecting organisms.

16S rRNA PCR is widely present in all bacteria and variable sequences within this gene are used to identify bacteria to the species level. A huge advantage of PCR over other automated identification systems is that it does not require the growth of the bacteria from a clinical specimen. It is therefore useful in the identification of slow growing or difficult to grow bacteria and in circumstances where the initiation of antibiotics has rendered the bacterial DNA non-viable.¹⁴ PCR is often performed at dedicated reference laboratories and this contributes to some of the limitations of this investigation mainly availability and a delay in receiving results. Another disadvantage concerning the identification of unusual organisms with molecular diagnostics is the potential for contamination of the sample leading to false diagnosis and inappropriate treatment. In these circumstances, the result of the sample must be interpreted in the appropriate clinical context.

The use of 16S rRNA PCR was extremely valuable in the clinical case described. Identification of the pathogen allowed prompt and targeted treatment of an infection which had the potential to cause significant morbidity. The efficient treatment of periprosthetic joint infection is imperative to avoid future complications; particularly when organisms not susceptible to empirical therapy are involved.

Learning points.

• Identifying the causative organism in an infective process allows for more targeted antibiotic therapy and reduces selective pressure.

• When possible and in the context of a non-septic patient, every effort should be made to gain tissue samples from an infected site prior to commencing antibiotics.

• The development of and access to molecular methods of bacterial identification allows a more targeted approach to antimicrobial therapy.