Molecular Pathogenesis and Clinical Impact of Biofilms in Surgery

Abstract

Biofilms are responsible for chronic persistent infections and are a major problem in implant surgery. The microbial pathogenesis, treatment and prevention of biofilm infections is reviewed.

1. Introduction

Though the “discovery” of biofilms is ascribed to Anton van Leeuwenhoek in 1676 using a novel magnifying device, and possibly to Robert Hooke two decades earlier, and biofilms were recognised in a marine setting about a century ago, they were of no medical interest until two studies described them in a medical device and in sputum in 1972 and 1974 respectively. The latter was a description of aggregates of Pseudomonas aeruginosa in secretions from the lungs of people with cystic fibrosis [1], and led to a burgeoning of research into Ps aeruginosa infection in that field. Through a meeting with Costerton, Højby studied these aggregates and the term “Biofilm” was made popular by Costerton in 1987 [2], though the term was originally used by Mack et al. [3] to describe “biofilm” on a water filter. However, many biofilm infections occur in association with implanted materials and devices, and their use has become much more common since the middle 1900’s. The first biofilm reported in a medical device was found in a shunt to treat hydrocephalus in 1972 [4]. This discovery explained the difficulty in successfully treating these infections non-surgically with antibiotics alone, and the report demonstrated the extracellular matrix of the biofilm in vitro and in vivo and carried out investigations to suggest that it was a glycosaminoglycan. This was later confirmed by important studies in 1996 [5]. Implantable biomaterials and devices are now widely used in modern surgery, and the list is extensive (Figures 1 and 2).

Figure 1.

Examples of implantable devices.

Figure 2.

Anatomical sites of common implantable devices.

2. Biofilm definitions

Many definitions of “biofilm” are found in the literature, and they can be based on either structure or function. Many of the definitions and their accompanying images are derived from in vitro models, and the appearance of mushroom-like structures and water-channels are not seen in biofilms occurring in vivo [6]. A definition based on functional aspects of biofilms is more useful in a medical context. This could be reduced to a population of bacteria or other micro-organisms, often associated with a surface, and enveloped in an extracellular matrix, showing insusceptibility to antimicrobials and to the host immune system, and ability to persist for long periods.

2.1 Biofilm phenotypes

The basis of this functional definition is the paucity of nutrients, including iron, and oxygen in the depths of the biofilm leading to a bacterial stress response caused by a crisis in energy generation and transport [7]. The bacterial stress response is mediated by the intracellular signal sigma-B. The bacterial response to this is to downregulate all synthetic functions not needed in biofilm mode, such as cell wall material, toxin and other non-essential protein synthesis, and DNA replication. These are the targets for common antibiotics, and beta-lactams, glycopeptides, aminoglycosides, macrolides and fluoroquinolones all become significantly less effective against biofilm bacteria. Other factors contribute to the lack of effect of antibiotics, including a slowing of their penetration into the biofilm, though this is rarely a major factor. The bacterial stress response results in significantly reduced cell metabolic activity and loss of some synthetic activities leading to auxotrophy for heme and menadione, and sometimes other substances such as thymidine. This biofilm phenotype is crucial to the clinical impact of biofilm infections; the colonies of biofilm bacteria when grown from clinical samples in the laboratory are typically less than ten times the size of their planktonic counterparts, and are known as small colony variants or SCV. The molecular control and regulation of biofilm phenotype has been described in detail by Proctor et al. [8]. SCV are important in biofilm infections not only because their metabolism leads to antibiotic insusceptibility, but because, though they can be internalised by professional and non-professional phagocytes, they are not killed and survive inside the phagocytic cells. Auxotrophic SCV of Staphylococcus aureus for heme and menadione, that do not produce alpha-toxin, are more able to survive intracellularly, and supplementation of intracellular populations of S aureus in vitro with menadione resulted in restoration of alpha-toxin production and reduced intracellular survival [9, 10]. SCV are not always auxotrophic and considerable variation occurs, but intracellular survival is a common feature. Many also show reduced susceptibility to aminoglycosides, and exposure to gentamicin can induce SCV formation [11]. Some SCV are the result of mutations in the genes concerned with electron transport, and these do not revert to parent forms whereas other forms of SCV appear to be phenotypic variants that revert to parent forms when the stress factor is withdrawn [8, 12]. SCV of gram negative bacteria have been known for decades, having been produced in the laboratory from exposure to antibacterial chemicals [13, 14]. However, more recently capnophilic (carbon dioxide—dependent) SCV of Escherichia coli have been isolated from a patient with a urinary tract infection, though no information on biofilm involvement was given [15]. A report of septic shock in a patient from whose urine capnophilic Proteus mirabilis SCV were isolated again did not state that biofilms were involved [16] but the patient had chronic renal stones, known to be associated with biofilms [17]. P mirabilis is an important uropathogen as it is highly motile and is capable of enzymatically hydrolysing urea into ammonia, thus being highly inflammatory as well as alkalinising the urine. The rising pH causes crystallisation of calcium and magnesium phosphates [18], and the P mirabilis biofilm typically consists of a mesh of bacteria, their extracellular matrix and phosphate crystals. These biofilms are obviously different in composition from those consisting mainly of bacteria and their products, and another example of such complex biofilms is the vegetations found in native valve endocarditis. Here the lesion consists largely of a matrix of platelets and fibrin, with bacteria, usually viridans streptococci, embedded in it. The lesion usually begins as a response to damage to the endocardium, which is then colonised by bacteria from the bloodstream, becoming progressively built up of fibrin and platelets with rafts of bacteria interspersed [19, 20]. A similar situation arises with prosthetic heart valves. In both cases, SCVs have been reported [21, 22] as well as other auxotrophic variants [23].

The biofilm phenotype, and SCV in particular, are important in treatment of biofilm infections. Surviving intracellular bacteria are protected from further immune assault and from most therapeutic antibiotics, which do not accumulate inside host cells sufficiently to kill SCV [24]. These factors mean that the amount of antibiotic required to kill bacteria in biofilm mode is typically 500–1000 times the minimum inhibitory concentration as measured in the clinical laboratory. Such concentrations are not achievable by intravenous or oral therapy, and eradication of biofilm infection usually requires extensive surgery to debride the site and to remove all surgical hardware.

2.2 Biofilm development

Development of biofilms in surgery depends on a sequence of events. Initially, the causative bacteria must be able to gain access to the site of biofilm formation, usually an implantable device. In modern surgery most device pathogens originate on the patient’s skin or mucous membranes, consisting mainly of coagulase-negative staphylococci (CoNS), typically Staphylococcus epidermidis, and Cutibacterium acnes. Conventional pre-operative skin preparation reduces but does not eradicate these bacteria, and the importance of relatively small numbers of bacteria in the operation field has been shown by an experiment in human volunteers, where various “doses” of S aureus were inoculated into incisions to determine how many bacteria were necessary to produce an abscess [25]. In one group, “foreign” material in the form of sutures were also introduced into the incision, and the number of bacteria required to form an abscess in those cases was 10,000 times fewer. This study, which is unlikely to be repeated in a modern setting, is extremely important in illustrating the role played by implantable materials and devices in infection in modern surgery.

The sequence of events involved in development of a biofilm infection involving a surgically implanted device are (Figure 3):

Access to the device from the source. Though heavy contamination of the air in the operating environment has historically been associated with surgical infection, modern operating room design and ventilation has meant that this source has declined in importance, and most surgical infections are caused by bacteria originating on the patient’s skin or mucous membranes. Bacteria reach the incision from the cut edges of the skin, or from contamination from surrounding skin surfaces, during surgery. The causative bacteria are therefore often present when the device is implanted.

Attachment to the device. Many bacteria possess adhesins on their surfaces that allow them to attach to biomaterials (vitronectin—binding protein etc) but more often they employ specific adhesins for the glycoproteins, platelets and other host-derived materials that rapidly coat all implanted materials [26, 27]. S aureus possesses specific adhesins for fibrinogen, fibronectin, laminin, thrombospondin, bone sialoprotein and other host-derived components of the conditioning film. These bacterial surface adhesins are known as MSCRAMMs (Microbial Surface Component Recognising Adhesive Matrix Molecules) [28] and they can be found in other organisms such as S epidermidis and enterococci [29]. Gram negative bacteria often attach by means of swarming or twitching motility over the new surface [30], some using twitching motility by Type IV pili [31, 32], and this might be particularly important in biofilm formation on urinary catheters. In addition, Ps aeruginosa uses a von Willebrand Factor-like surface factor in twitching motility over biomaterial surfaces [33].

Once bacteria have attached to the surface or conditioning film, they begin to proliferate and to develop intercellular adhesins such as polysaccharide intercellular adhesin (PIA) in staphylococci. This substance is integral to further development of biofilm, and is encoded by the ABDC operon, and regulated by icaR. At this stage, bacterial stress responses are operating in response to limitation of nutrients and oxygen and the biofilm phenotype is appearing [34]. It is important to note that the bacterial stress response, mediated by Sigma B, downregulates icaR and increases PIA production, and the stress response can be provoked by external factors such as antibiotics as well as nutrient starvation. Once the biofilm phenotype has developed, the biofilm is stable and is not susceptible to host immune activity or to antimicrobials. There is often a lag phase of about 14–28 days before the biofilm reaches functional maturity, during which it might be more susceptible to antimicrobials [35].

Figure 3.

Sequence of events in development of biofilm infection. Here implant has an antimicrobial coating, but within minutes this is covered by a glycoprotein conditioning film produced by the patient. This usually prevents the activity of the coating and bacteria now adhere to the conditioning film. Within a few hours the attached bacteria begin to produce an extracellular matrix and to multiply. Powerful antibacterial activity is essential now, as after this point, it is almost inevitable that a biofilm will develop, within a few weeks.

Clear understanding of the sequence of events and periods of risk is essential for effective planning of preventative measures.

3. Prevention of biofilm infections

3.1 Surgical considerations

Since the days of Semmelweis, Lister and others in the mid–to late 1800s, personal hygiene of the surgeon, aseptic technique and antisepsis have become accepted norms. Since the 1950s, when bacteria-laden operating room air was identified as a major factor in surgical infection [36], greatly improved practices and ventilation systems have made this a minor source. Two main forms of ventilation are in use in modern operating rooms: plenum, and laminar flow with high efficiency particulate air (HEPA) filtration. While it is clear that the numbers of airborne bacteria are significantly reduced when laminar flow is used [37] there has never been a clear causative link between either this reduction or the actual bacteria and surgical infection, leading the USA CDC to downgrade their initial recommendation [38]. More recently, reports have appeared of small but significantly increased infection rates when laminar flow is used [39, 40] and this appears to be due to flaws in its design and manner of use [41]. For most types of implant surgery, plenum (conventional) ventilation appears to be satisfactory so long as other precautions are taken (Figure 4).

Figure 4.

Sequence of surgical preventative events.

Care bundles have been proposed for infection reduction in various healthcare settings. A bundle is a collection of interventions that are expected to contribute to reduced risk of infection, but which singly might have weak or no evidence base. A measure such as ensuring that only three people are present in the operating room during a procedure is not supported by any clear evidence but it is intuitively likely to be beneficial if only in reinforcing operating room discipline. A bundle must be directed towards behaviour change on the part of relevant staff members, and it works best if they contribute to its content, and formally agree to abide by it. Some bundles insist on contents being evidence-based, but the quality of evidence is usually very weak for individual components. However, when bundles are properly applied, they are often very effective in reducing surgical infection [42, 43] and in any case they and their contents should form part of a well-managed surgical discipline. Usually no single component can be identified to explain their success, but clinical trial evidence has shown that violations of the bundle are associated with re-emergence of infection [43].

As the major source of pathogens is the patient’s skin, attention has been directed towards the effectiveness of preoperative skin preparation. Two main antiseptics are in use: chlorhexidine and povidone iodine. Each can be formulated in water or 70% alcohol. A report by the World Health Organisation (WHO) favouring chlorhexidine [44] has been called into question on the basis of quality of evidence [45]. However, sampling is usually by swabbing of the skin surface, and almost none of the many studies on surgical skin preparation explore the effectiveness of any agent on bacteria resident in the dermis, though an early study showed that full thickness skin biopsy was necessary [46]. This has since been confirmed [47, 48]. When skin biopsy is used, neither antiseptic in alcohol is able to eradicate resident skin bacteria, and though reduced, the remaining numbers are often sufficient to cause a biomaterial-associated infection [25]. Two studies on the penetration of both aqueous and alcoholic chlorhexidine into human skin using full thickness biopsy have found it to be minimal [49, 50]. Further measures are therefore necessary. Some researchers have investigated the effect of antiseptic-soaked material to protect the incision from the skin edges during surgery, and while this is commonly used, there have been no quantitative studies to show benefit. Intravenous antibiotics are almost universally used in surgery, ideally as a single dose 30–60 min before incision, but extra doses are commonly used postoperatively though they offer no benefit over that of the single pre-operative dose. Antibiotic prophylaxis is undoubtedly highly effective in reducing infection risk in many types of surgery, including colorectal surgery [51] and orthopaedic surgery [52] but probably less so in neurosurgery due to limited penetration of systemic antibiotics intracranially. However, it is probably inevitable that a small number of bacteria will reach the implant during operation, and further measures have been directed to attempts to eradicate these. As knowledge of attached bacteria and biofilms has shown that very high concentrations of antibiotics are necessary, some surgeons have used either antiseptic or antibiotic irrigation [53, 54], or have simply added antibiotic powder to the incision before closure [55, 56, 57] with successful reduction in infection rates and complications. This intervention gives extremely high local antibiotic levels not reachable by systemic administration, yet avoids most of the complications associated with the latter method.

3.2 Antimicrobial biomaterials

Other methods of prevention accept that despite efforts, bacteria will reach the implant, and aim to prevent their attachment or to kill them when attached. Various “anti-fouling” surfaces have been investigated with the aim of allowing host cell and tissue proliferation but preventing bacterial attachment [58, 59] but none of these has yet reached clinical application, largely because of the complex relationship between implant surface, host tissue environment, and bacterial surface adhesins. Biomaterials designed to kill bacteria that do attach to them have generally included coatings of silver, antiseptic or antibiotic and combinations of these, often with a vehicle to bind the antimicrobial to the biomaterial surface. Such coatings have several disadvantages. The normal host reaction to the implant of deposition of plasma proteins [26, 27] also obliterates the antimicrobial coating in many cases, making it ineffective. Silver is susceptible to this due its avidity for proteins [60], and it can also be inactivated by chloride [61] which is abundant in the human body. Silver ions have also been shown to be cytotoxic in certain conditions [62]. Clinical studies on silver-processed devices give very variable results, and there is doubt about their cost-effectiveness in wound dressings [63]. A recent randomised controlled trial of silver-containing catheters intended to reduce ventriculitis in people with hydrocephalus shunts found no difference from plain catheters [64]. Another randomised controlled trial of silver-processed urinary catheters again found no significant difference from plain catheters [65]. In both of these clinical settings, biofilms play a key role, and the goal is to prevent bacterial proliferation and biofilm development on the catheters. Both have fluid containing proteins and chloride flowing through them.

Another approach has been impregnation of catheter material with antimicrobials. Though the impregnation processes differ, two catheter types can be considered: those containing rifampicin and minocycline, and those containing rifampicin and clindamycin. The first type has been used in central venous catheters [66] and external ventricular drains [67]. The second type has been used in hydrocephalus shunts and external ventricular drains. In all cases they have shown effectiveness in reducing device -related infection. The advantage of impregnation over coatings is that they give a long duration of activity: coatings are usually washed away by fluid after a few days, whereas the surface of an impregnated material is continually replenished by migrating antimicrobials until the depot in the material is depleted, usually several weeks later (Figure 5). This is important when the implantable device is at risk of contamination for an extended period.

Figure 5.

Principle of impregnated biomaterial. Antimicrobial molecules are motile within the device matrix and can migrate to the surface to replace those removed by fluid flow.

3.3 Importance of source of infection and period of risk

In order to formulate an effective preventive strategy, knowledge of the source and nature of device pathogens and the period during which the device is at risk is essential (Table 1). As many biofilm infections are caused by micro-organisms originating in or on the patient, a knowledge of the distribution of these is useful. The normal bacterial flora of the skin differs according to age and sex, but particularly depending on the anatomical site. The most common bacteria found on the skin are staphylococci, particularly members of the CoNS. These are typified by S epidermidis which is broadly distributed over the body surfaces, but other species such as Staphylococcus capitis have preferred sites such as the head and neck. C acnes is an important pathogen in the context of implant infections, but it is a good example of the importance of specific topographical distribution in determining the important pathogens in particular implants. C acnes is found on the upper body and head (Figure 6) [68], and it is therefore not surprising that devices implanted in these areas show a significantly higher incidence of C acnes infection. Examples are neurosurgical shunts and drains [69, 70], spine instrumentation [71], breast implants [72] and shoulder arthroplasty [73, 74]. Implants in other sites such as urinary catheters are at risk from a different microbial profile, as the pathogens originate in the large intestine, and E coli, Klebsiella pneumoniae and P mirabilis are the most common.

Implant/device	Duration of use	Main source of pathogens	Period of risk
			At insertion	During use
Hydrocephalus shunt	indefinite	Patient’s skin	++	—
External ventricular drain	Few days-weeks	Patient’s skin/environment	±	++
Joint replacement	Indefinite	Patient’s skin	++	±
Urinary catheter 1	<28 days	Patient/environment	±	++
Urinary catheter 2	~90 days	Patient/environment	±	++
Peritoneal dialysis catheter	Indefinite	Patient/environment	±	++
Vascular graft	Indefinite	Patient	++	+
Prosthetic heart valve	Indefinite	Patient	+	++
Spinal instrumentation	Indefinite	Patient	++	±
Venous access device	Days—months	Patient/environment	±	++
Sutures	Days	Patient/healthcare worker	+	±

Table 1.

Periods of risk of infection of common implantable devices.

Figure 6.

Topographical distribution of common biofilm pathogens (after Grice et al. [68]).

The time at which the implant is at risk of microbial contamination also varies. While there is always a risk at the time of implantation, in some implants this is the main time, and the risk of subsequent contamination is proportionally small. Examples of this are hydrocephalus shunts and joint replacements. In other implants the risk at insertion is significantly outweighed by that during use. Examples are external ventricular drains (EVD) for raised intracranial pressure, urinary catheters, venous access catheters and peritoneal dialysis catheters, all of which can be contaminated from environmental sources or from the hands of staff or users during use. Other examples are vascular grafts and prosthetic heart valves, which are at risk from hematogenous seeding from bacteria entering the bloodstream at a distant site.

When planning strategies for prevention of biofilm infections involving antimicrobials, it is therefore important to match the antimicrobial to the most likely pathogen(s). If systemic antimicrobial prophylaxis is contemplated, then the adverse effects of this must be taken into consideration if there is a need for prolonged administration due to extended period of risk. If antimicrobial materials or devices are to be used, these must address not only the likely pathogen(s) but also the duration of protective activity required.

International guidelines indicate that for most surgical procedures, any systemic antimicrobial prophylaxis should be administered as one dose 30–60 min before start of surgery [75, 76]. Extension of this prophylaxis beyond 24 hours does not reduce surgical infection further, but it does increase the incidence of acute kidney injury and Clostridioides difficile infection [77], which is a life-threatening colitis associated with over-use of antibiotics. Where the period of risk extends beyond the insertion procedure, such as in EVD, long courses of systemic antibiotics are often given until the drain is removed. This has been shown in some cases to reduce brain infections, but at a cost. A randomised study comparing the use of plain catheters and prolonged systemic antibiotics with antimicrobial-impregnated catheters and one dose of antibiotic at insertion found no difference in the brain infection rate, which was low in each group, but there were three cases of C difficile infection in the prolonged antibiotics group, one patient requiring total colectomy [78].

4. Treatment of biofilm infections

The difficulty in treating biofilm infections in surgery emphasises the importance of effective prevention. However, this is not always possible. The nature of the biofilm phenotype and its implications for antibiotic treatment mean that further surgery is almost inevitable, and this usually involves removal of the device. This might be relatively simple, as in the case of a venous access catheter or a urinary catheter, but it can be both surgically complicated and hazardous, as in the case of spinal instrumentation or prosthetic heart valves.

Attempts to eradicate established biofilm with antibiotics usually fail. A comparison of treatment regimens for hydrocephalus shunt infections showed that results with shunt removal and antibiotics were significantly superior to those with antibiotics alone [79]. Successful treatment of joint replacement infections relies on device removal and extensive debridement of infected tissue, with prolonged antibiotic therapy. However, understanding of biofilm biology has led to advances in this area. The biofilm phenotype takes a few weeks to “mature” to the point where full insusceptibility to antibiotics is expressed, and this has been exploited in development of a regimen for treatment of prosthetic joint infection when the diagnosis can be made within 3–4 weeks of insertion [80]. In this regimen, known as Debridement, Antibiotics and Implant Retention (DAIR), surgical treatment of the infected joint prosthesis is carried out on a planned basis after careful investigation to establish the causative micro-organism and its antimicrobial susceptibilities, to allow consultation with specialists including Microbiology/Infectious Diseases, and to determine that the implant is stable (Figure 7). Infections due to multi-drug-resistant bacteria, fungi or multiple bacteria are not suitable for this approach. During the operation, the prosthetic components are exposed and the acetabular module is removed, leaving the main metal prosthesis in place. All infected tissue is removed and samples are sent for microbiological examination. Copious irrigation with antiseptic is applied, and biodegradable antibiotic—eluting beads can be inserted to provide high local concentrations. The choice of antibiotic in the beads should be made in consultation with a microbiologist. The joint is then closed and a long postoperative course of suitable antibiotics is then started [81]. The success rate of DAIR compared to conventional full implant removal and replacement is slightly lower. Moreover, despite the very thorough surgical debridement and long courses of antibiotics, often for over a year, relapse can occur [82], illustrating the difficulty in eradication of biofilms. DAIR spares the patient the much more extensive surgical removal of the main implant components, and the second surgery to inert fresh implants a few weeks later.

Figure 7.

Possibility of retention of infected implant based on knowledge of biofilm phenotype maturation (based on Zimmerli and Trampuz, 2004) [80].

5. Diagnosis of biofilm infections

5.1 Clinical features

Most biofilm infections in surgery are chronic and persistent, sometimes for many years [83]. It is important to distinguish between “late infection,” implying an infection contracted long after surgery, such as hematogenously, and “delayed infection,” meaning that the infection appears long after surgery even though it was contracted at the operation. Delayed infection in spine instrumentation is usually due to infection with CoNS or C acnes [84, 85]. A similar situation is found in shoulder arthroplasty [86]. Generally, more virulent bacteria such as S aureus are associated with either early-presenting or with hematogenous infections. The delay of months or years between initial surgical implantation and appearance of symptoms [84] has led to doubt about the surgical origins of some infections but this has now been largely dispelled. However, the need for prolonged follow-up and vigilance must be emphasised.

Acute postoperative biofilm infections usually appear within days or weeks of surgery, with failure of wound healing, drainage of pus or other fluid from the wound, local pain and swelling, fever and general illness. Delayed or chronic infections of joint prostheses present with persistent pain and restricted mobility, local swelling and sometimes a sinus. In the absence of a sinus, diagnosis might be delayed as it is often difficult to distinguish infective from mechanical complications. Aspiration of synovial fluid often gives a diagnosis but sensitivity is low [87, 88]. Delayed infection in spine instrumentation similarly presents with persistent pain, tenderness and possibly a draining sinus. Delayed infections in hydrocephalus shunts are very uncommon now that the preferred route of drainage is to the abdomen (ventriculoperitoneal, VP), but the ventriculo-atrial (VA) route is still used in some cases. In VP shunts infection usually presents within a few months as it leads to obstruction, but this does not happen in VA shunts and symptoms might not appear, or at least become recognisable, for several years. During this time, bacteria are being discharged from the biofilm in the shunt into the bloodstream, and this might give rise to periods of ill-health or sporadic fevers. It also provokes production of antibodies to the bacteria, and eventually the concentrations of circulating antigen and antibody, and therefore immune complexes, become so high that they precipitate on basement membranes of joints, renal glomeruli, alveoli and microvascular system. The presenting clinical picture can therefore be a confusing array of disorders from hematuria, hemorrhagic skin rashes, arthropathy, and chronic cough [89, 90]. Clinical diagnosis can therefore be very difficult, and a high level of suspicion is needed. Aspiration of cerebrospinal fluid from the shunt often gives the diagnosis, but blood cultures can be negative in the later stages.

5.2 Laboratory methods

Depending on the site of the infection and presence of an implant, sometimes blood cultures are positive, indicating systemic spread of the infection, and risk of sepsis. Blood inflammatory markers such as erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) levels are usually raised. Swab cultures from the wound might yield the infecting pathogen, but they might be misleading due to contamination [91]. Surgical exploration of the incision and deeper layers allows tissue samples to be taken and these are more likely to yield the pathogen(s). Such samples should always be taken during debridement surgery [92], using fresh instruments for each of up to six separate samples [81, 93]. In view of the anaerobic preference of C acnes and its slow growth, cultures should be incubated anaerobically for up to 10 days [94]. The way in which tissue samples are processed in the laboratory is important. Simply rubbing them on a culture plate or incubating them in a fluid culture is prone to contamination and gives poor yield, leading to under-diagnosis of infection. Tissue should be homogenised but the method of doing this is also important [95]. When hardware such as joint replacement or spinal instrumentation components are removed, these should be seen as valuable samples. Sonication to remove the biofilm has been shown to significantly increase the culture positivity rate [96, 97]. A further aid to laboratory diagnosis has been PCR [98] especially when applied to tissue homogenates or hardware sonicates. However, if PCR is used in an attempt to certify eradication of infection before re-insertion of a prosthesis, residual DNA from bacteria successfully killed by antibiotic therapy can give false positive results suggesting ongoing active infection. This can be overcome by use of a modified PCR method that detects DNA only from live bacteria [99].

6. Conclusions

The impact of biofilm infections in surgery on healthcare systems, economies and personal lives of patients is immense. The financial cost can only be estimated and published figures do not usually take into account “unseen” costs such as loss of earnings due to disability, increased dependency, and financial burden on carers.

The physical and mental trauma of surgery such as joint replacement, reconstructive breast implant or hydrocephalus treatment can be made unimaginably worse by postoperative biofilm infection.

The significant difficulty in successfully treating biofilm infections with antibiotics, due largely to the biofilm phenotype, is now well recognised, and the importance of commensal bacteria previously thought to be harmless, such as S epidermidis and C acnes, is becoming more widely known. However, surgical device removal remains the mainstay of treatment, and new approaches that allow implant retention are needed. Prevention of biofilm infections is crucial, and biomaterials that either reduce bacterial attachment, such as those coated with novel synthetic polymers [100] or those designed to kill bacteria on contact [66, 67] are now in clinical use. Many other biomaterials approaches are in development, and considerable strides have been made in this direction but further progress is being slowed by unrealistic commercial and regulatory barriers [101].