Pulsatile Lavage Is Not Enough to Remove Implant Biofilm

Abstract

Background:

During periprosthetic joint infection revision surgeries, intraoperative irrigation is performed to clear debris, blood, purulence, and microbes from the surgical wound. Dental evidence suggests motorized agitation is effective in removing established biofilms. The purpose of this study was to investigate how pulsatile irrigation and sonicated brushing affect mature staphylococcal biofilm on real arthroplasty components.

Methods:

Five identically sized, never implanted tibial base plates (TBPs) underwent keel removal with a wire electrical discharge machine. Implants were passivated in 25% nitric acid, autoclaved, and submerged in Methicillin sensitive Staphylococcus aureus (MSSA) infected tryptic soy broth. Biofilm was grown for 72 hours, with media replaced every 24 hours. Following growth, TBPs were assigned to 6 experimental conditions: control (no treatment), low-speed pulsatile lavage, high-speed pulsatile lavage, sonication brushing, combination of low-speed pulsatile lavage followed by sonication brushing, and combination of high-speed pulsatile lavage followed by sonication brushing. Pulsatile lavage lasted 5 seconds, while sonication brushing lasted 20 seconds using a modified commercial brush. Experiments were performed in sextuplicate. Posttreatment TBPs were either put in a sonication bath to dislodge remaining bacteria to count colony-forming units (CFUs) or stained with crystal violet to quantify residual biofilm biomass.

Results:

All mechanical methods significantly reduced CFU counts. Low-speed pulsatile lavage, high-speed pulsatile lavage, sonication brushing, and brushing without sonication reduced CFU counts by 64%, 68%, 87%, and 82%, and reduced biofilm biomass by 74%, 68%, 78%, and 77%, respectively, as compared with controls. The combination of pulsatile lavage and brushing lowered CFU counts by 99%, and biofilm biomass was reduced by 86%. Scanning electron microscope (SEM) imaging confirmed biofilm removal from the locking mechanism by sonication brushing only.

Conclusions:

Combining pulsatile lavage and mechanical debridement methods more effectively removes biofilm from implant surfaces compared with either method alone.

Clinical Relevance:

Mechanical methods including pulsatile lavage and sonication brushing remove biofilm from orthopaedic implants. Clinicians should be aware of these tools and consider using them.

Introduction

Periprosthetic joint infection (PJI) is a potentially devastating complication that can occur following total joint arthroplasty. Central to the pathogenesis of PJI is formation of biofilm on implant surfaces which shields pathogens from immune responses and antibiotics¹. When attempting to retain infected implants, principles of successful PJI management include debriding infected tissue and removing as much biofilm as possible. Current methods used by surgeons include chemical sterilization using antiseptic solutions² and mechanical evacuation using different tools and irrigation³. Despite these interventions, PJI recurs often when implants are retained^4-6, indicating that contemporary biofilm removal techniques are inadequate.

Although research by our group^7,8 and others^9,10 has quantified the impact of antiseptic solutions on biofilm, very little orthopaedic literature exists pertaining to mechanical methods. Pulsatile lavage has been an intraoperative omnipresent tool to evacuate the organic material, but its effect on removing biofilm remains unknown, and there are concerns that pathogens could be pushed deeper into host tissues¹¹. Scrub brushes have been recently recommended³, but these devices are not sterile, can break apart when rubbed against implant surfaces, and are too bulky for use in tight spaces.

The existing literature pertaining to mechanical methods originates from dentistry, where electrically powered brushing has been shown to be superior to manual methods¹². A type of electrical brushing involves sonication, where brush bristles vibrate at very high frequencies. These frequencies create cavitation bubbles which implode, creating shockwaves that detach bacteria from their underlying biofilm matrices¹³ and hydrodynamic effects that force fluid into hard-to-reach areas¹⁴. These collective properties make sonication brushing more effective in removing plaque from teeth¹⁵, and potentially an effective tool for removing biofilm from orthopaedic implants, whose locking mechanisms and complex topography provide ample places for pathogens to hide¹⁶.

To date, no studies have investigated the effects of sonication brushing on biofilm grown on orthopaedic surfaces. Consequently, the purpose of this study was to determine how pulsatile lavage and sonication brushing compare in removing mature biofilm formed on clinically used orthopaedic implants. We hypothesized that (1) sonication brushing would remove more viable bacteria and biofilm matrix compared with pulsatile lavage and (2) sonication brushing would leave less visible biofilm in the locking mechanisms of a tibial implant compared with pulsatile lavage.

Materials and Methods

Bacterial Preparation

Methicillin sensitive Staphylococcus aureus (MSSA) Xen36 (Perkin Elmer) was used for all experiments. This organism was chosen because (1) Staphylococcus aureus is a common pathogen in PJI¹⁷, (2) this strain is resistant to kanamycin and can culture in kanamycin-infused broth which minimizes contamination risk, and (3) our laboratory has experience using this strain⁸. Tryptic soy broth (TSB; Becton Dickinson) with kanamycin (Sigma-Aldrich; 200 μg/mL) wase used for growth media and plating. A single colony of Xen36 was subcultured, grown to a known density of 3 × 10⁸ colony-forming units (CFUs) per ml as detected by a cuvette spectrophotometer (Bio-Rad SmartSpec Plus), and diluted in TSB to achieve 1 × 10⁷ CFU/mL.

Implant Preparation and Establishing Biofilm

Five never used tibial implants of identical dimension (size D Persona; Zimmer Biomet) had their stem and keel removed using wire electrical discharge machining (Figs. 1-A through 1-C) so that the tibial baseplate (TBP) could lie flat in a petri dish. TBPs were autoclaved (Steris, 121°C, 45 minutes) and placed in individual petri dishes (100 mm diameter, 20 mm depth).

Fig. 1.

Tibial implant before (Fig. 1-A) and after (Fig. 1-B) keel removal through electrical discharge machining. All implants were the same size (Fig. 1-C). Fig. 1-D Mechanical debridement setup to provide standardized pulsatile lavage conditions. Fig. 1-E Modified sonication toothbrush.

To produce mature biofilm, TBPs were submerged in 45 mL of TSB containing 1 × 10⁷ CFU/mL Xen36 and maintained at 37°C in a static incubator for 72 hours. This duration was based on previous literature^8,18,19 and pilot work (data not shown) demonstrating that CFU counts plateaued at 72 hours, indicating surface confluence. TSB media was exchanged every 24 hours.

Mechanical Treatment

Following mature biofilm formation, TBPs were subjected to a control intervention (gently rinsed with sterile saline) or one of the 5 mechanical treatment methods: (1) pulsatile lavage on low setting (5.6 pounds per square inch [psi]), (2) pulsatile lavage on high setting (10.4 psi), (3) sonication brushing, (4) combination of low-speed pulsatile lavage and sonication brushing, and (5) combination of high-speed pulsatile lavage and sonication brushing. Each condition was tested 6 times.

Pulsatile lavage treatment was performed with sterile saline for a 5-second duration using a Pulsavac Plus Wound Debridement Gun (Zimmer Biomet) that was clamped to a titration stand at a height of 5 inches above the TBP (Fig. 1-D). Treatment was enclosed within a large biohazard bag to protect research personnel. Sonication brushing treatment was performed using a commercially available brush (Sonic Electric Toothbrush; Initio) modified with assistance from a precision prototyping engineering firm (Jaktool) by attaching an external RioRand PWM Motor Speed controller to modulate vibrating motor speed (Fig. 1-E). The motor was set to a maximum setting of 40,000 strokes per minute (667 Hz). The brush heads contained nylon DuPont-engineered standard bristles, which are nonabsorbent, of varying lengths, and engineered to be less susceptible to wear²⁰. Sonicating brush treatment lasted for 20 seconds and was performed by a single research individual to minimize operator confounding.

Combination treatment was performed first with pulsatile lavage (high-speed or low-speed) for 5 seconds, then sonicating brushing. Following all conditions, TBPs were dipped in sterile phosphate buffered saline (PBS) to remove planktonic bacteria and processed for either CFU quantification or biofilm matrix. Used TBPs were washed with soap and water, scrubbed with a bristled brush, placed in 25% nitric acid solution for 24 hours to repassivate titanium surfaces according to American Society for Testing and Materials (ASTM) F86-21²¹, and then autoclaved for reuse.

Quantifying Viable Bacteria, Biofilm Matrix

To quantify viable bacteria, TBPs were carefully inverted and positioned in a petri dish containing enough TSB to submerge only the articular surface of the baseplate. Sealed petri dishes were placed in an ice-water mixture and sonicated for 30 minutes at 52 kHz, 137 watts (550 HT Aquasonic; VWR Scientific Products) to dislodge viable bacteria²². Resulting sonicated fluid was serially diluted in TSB, plated in 50 μL aliquots on TSB-agar and grown overnight at 37°C in a static incubator. CFUs from each plate were counted by 2 observers blinded to treatment.

To quantify biofilm matrix formation, TBPs were submerged in 0.1% crystal violet stain in deionized water for 15 minutes, washed 3 times and dried for 2 hours. TBPs were then photographed head on and at a 60° angle to visualize the locking mechanism. TBPs were then placed face down in 15 mL of 33% glacial acetic acid for 15 minutes to dissolve adhered dye. One hundred fifty microliters of the resulting mixture was plated into a 96-well dish, and crystal violet was quantified using a spectrophotometer (SpectraMax iD5; Molecular Devices) measuring an absorbance of 590 nm.

Direct Visualization of Biofilm

A larger tibial component (size F) was autoclaved and immersed in TSB containing Xen 36 for 72 hours. The right half of the TBP was treated with sonication brushing. The left side was not treated. The TBP was rinsed 3 times with sterile PBS, placed in 2.5% glutaraldehyde in 0.1 M cacodylate buffer for 20 minutes, washed 3 times in deionized water, dehydrated in a series of ethanol mixtures, dried for 3 hours, sputter coated (Denton Vacuum, Desk V) with gold, and imaged with a scanning electron microscope (Zeiss Gemini 300 FE-SEM). Standardized images were taken of the locking mechanism at 30, 10, and 2 μm. Images were scrutinized for the presence of 1 μm spheres (bacteria) clustered with fibrin connections⁸.

Statistics

The primary study outcome was quantifying viable bacteria within biofilm among control and mechanically treated TBPs. The secondary study outcome was quantifying biofilm matrix among TBPs. Data normality was determined using the Shapiro-Wilk test. Individual and combination mechanical TBP treatments were compared with control conditions and then each other.

For comparisons involving 3 or more groups, a one-way analysis of variance was applied followed by a post hoc Tukey test. Two-group comparisons were completed using a 2-way Student t test. Results were presented as mean ± 95% confidence intervals. An a priori power analysis based on pilot data from 2 sample TBPs determined that a minimum of 4 TBPs per group would provide 80% power (β = 0.20) to detect a 50% difference in viable bacterial counts between treatments, assuming a 2-sided α = 0.05 and an estimated standard deviation of 25% in bacterial reduction based on pilot data. All analyses were performed using GraphPad Prism software (version 9.0a).

Results

Mechanical Treatment and Viable Bacteria

Mature MSSA biofilm was successfully grown on TBPs, with controls producing colony counts of 5.58 ± 2.06 × 10⁶ CFU/mL. Mechanical interventions reduced CFU counts compared with controls (Fig. 2). Specifically, low-speed and high-speed pulsatile lavage reduced CFUs by 64.04% (2.01 ± 0.56 × 10⁶ CFU/mL; p < 0.001) and 68.30% (1.77 ± 0.48 × 10⁶ CFU/mL; p < 0.001), respectively. Sonication brushing significantly lowered CFU counts further, producing an 86.95% reduction (7.28 ± 4.79 × 10⁵ CFU/mL; p < 0.001) and 9.68 ± 4.60 × 10⁵ CFU/mL (p < 0.001). The largest effects were observed with combination treatments, producing a 99.59% reduction (2.28 ± 1.49 × 10⁴ CFU/mL; p < 0.001; Fig. 2) observed with low-speed pulsatile lavage plus brushing and 99.67% reduction (1.86 ± 5.52 × 10⁴ CFU/mL; p < 0.001) observed with high-speed pulsatile lavage plus brushing.

Fig. 2.

Bar graph demonstrating quantitative bacterial counts on tibia component surfaces after control and mechanical treatment conditions. CFU = colony-forming units.

When comparing efficacy among mechanical methods, no significant difference was found between high-speed and low-speed pulsatile lavage (p = 0.1509; Fig. 2). However, the use of sonication brushing was associated with a significant CFU reduction when compared with low-speed pulsatile lavage (63.7%; p = 0.005) and trended to lower CFUs compared with high-speed pulsatile lavage (48.5%; p = 0.061) pulsatile lavage. Furthermore, the combination of high-speed pulsatile lavage and sonication brushing led to significant CFU reductions compared with low-speed pulsatile lavage alone (99.1%; p = 0.037) and high-speed pulsatile lavage alone (98.9%; p = 0.037).

Mechanical Treatment and Biofilm Matrix

Mature MSSA biofilm matrix was detected on all control TBPs. All mechanical interventions significantly reduced the amount of crystal violet compared with controls (Fig. 3). Low-speed pulsatile lavage and high-speed pulsatile lavage reduced biofilm matrix by 73.89% (p < 0.001) and 68.13% (p < 0.001), respectively. Sonication brushing further reduced biofilm matrix by 78.24% (p < 0.001) compared with controls. The most substantial reductions in biofilm were consistently observed with combination mechanical treatment. Low-speed pulsatile lavage and sonication brushing reduced biofilm matrix by 87.26% (p < 0.001), while high-speed pulsatile lavage and sonication brushing reduced matrix by 86.26% (p < 0.001).

Fig. 3.

Bar graph demonstrating quantitative crystal violet biofilm staining on tibia component surfaces after control and mechanical treatment conditions. CV = crystal violet.

When comparing mechanical methods, no significant difference was found between low-speed and high-speed pulsatile lavage (p = 0.438). Sonication brushing yielded a significant improvement in removing biofilm matrix vs. high-speed pulsatile lavage (p = 0.002) and trended positively vs. low-speed pulsatile lavage (p = 0.192). A combination of lavage and brushing did significantly reduce biofilm matrix vs. high-speed (p < 0.001) and low-speed pulsatile lavage (p = 0.01), as well as sonication brushing (p = 0.07).

Visual Confirmation of Biofilm Removal

Control TBPs stained with crystal violet demonstrated accumulation throughout (Fig. 4-A). TBPs treated with pulsatile lavage demonstrated less staining over flat surfaces, but heavy staining persisted in the locking mechanism (Fig. 4-B). Conversely, TBPs treated with sonication brushing yielded a visually distinguishable decrease in staining in both the flat surface and within the peripheral locking mechanism compared with pulsatile lavage treated and control TBPs (Fig. 4-C).

Fig. 4.

Top-down photographs of tibial components stained with crystal violet after no treatment (Fig. 4-A), high-speed pulsatile lavage (Fig. 4-B) and sonication brushing (Fig. 4-C). Arrows indicate residual violet staining denoting the presence of residual biofilm.

SEM inspection of the locking mechanism in the control half of the TBP confirmed confluent staphylococci connected with extracellular fibrin (Fig. 5-C). Conversely, the treated half demonstrated no identifiable staphylococci (Fig. 5-D).

Fig. 5.

Fig. 5-A In-chamber of view of tibial baseplate positioned for imaging by scanning electron microscopy. Half of the tibial component was treated with sonication brushing and half received no treatment. Fig. 5-B Regions of interest (boxes) within the locking mechanism in each half of the baseplate were imaged. High-magnification images were produced of the locking mechanism of the untreated (Fig. 5-C) and treated (Fig. 5-D) halves. Biofilm was identified as 1 μm spherical shapes connected in clusters.

Discussion

This manuscript describes the first investigation of how different mechanical interventions compare in removing viable bacteria and mature biofilm from orthopaedic implants. Through using commercially available implants and the most frequently offending pathogen, this study demonstrates that pulsatile lavage removes 2 thirds of viable bacteria and mature biofilm matrix. Furthermore, given clinical enthusiasm for studying other mechanical techniques, such as low frequency sound waves, which can dislodge bacteria from bone and soft tissue^23,24, we investigated the effect of high frequency sonication brushing on biofilm clearance. We confirmed our hypothesis that sonication brushing is more effective than pulsatile lavage in clearing bacterial organisms and biofilm. Strikingly, we found that the combination of pulsatile lavage and brushing enhanced pathogen removal further, involving 99% of viable bacteria and 85% of biofilm matrix. The observed effects are due to known hydrodynamic forces and cavitation bubbles that occur when sonication processes are performed when an aqueous solution is present. These electrically powered processes destroy biofilm more consistently than mechanical brushing^25-27 and work independently of need for applying pressure by the operator²⁸.

We do acknowledge several study limitations. First, the experiments conducted in a controlled in vitro environment do not replicate the complexity of in vivo conditions which include host immune responses, tissue interactions, and heterogeneity in biofilm composition. Second, our study focused on monomicrobial Staphylococcus aureus biofilms, which may not exhibit the same structural characteristics as polymicrobial biofilms encountered clinically. Third, our experimental conditions restricted pulsatile lavage and brushing times which underestimate what is used clinically. Fourth, investigations were restricted only on the articular surface of the tibial component. Fifth, owing to the need to make our study replicable, we had to standardized things such as the order of brushing and washing and the distance of the pulsatile lavage. This could have an effect on the results. Further work is needed to determine how effective brushing is on other components as well as soft tissues. Finally, we did not compare sonication brushing with surgical scrub brushes that have been previously recommended³.

In conclusion, our findings demonstrate that mechanical modalities can effectively, but not completely remove biofilm from contaminated orthopaedic implants. By demonstrating an enhanced effect with pulsatile lavage, sonication brushing deserves further investigation to determine if it could evolve into a usable surgical tool. Future efforts are required to determine what optimal mechanical interventions are and if combining them with standardized debridement methods and targeted bactericidal products can lead to improved surgical outcomes when performing implant-retaining infection surgery.