Use of a Fluorescence Imaging Device to Detect Elevated Bacterial Loads, Enhance Antimicrobial Stewardship, and Increase Communication Across Inpatient Complex Wound Care Teams

Wounds are a so-called silent epidemic affecting patients both inside and outside of the hospital setting. A point prevalence study of wound audits in 13 Canadian hospitals in 2006 and 2007 reported that an average of 41.2% of hospitalized patients had wounds.¹ The increasing number and complexity of wounds within the hospital system results in both a financial and resource burden, with wound care accounting for 2% to 3% of total health care spending worldwide.^2,3 Because wound care is not limited to a single department, hospital systems are increasingly utilizing specialized, interprofessional wound care teams to optimize the care of acute and chronic wounds. These teams often perform resource-intensive care comprising comprehensive wound assessments and facilitate interventions by recommending the best treatment and care plan. Even with this dedicated staffing, wound care continues to place a significant strain on hospitals.

Appropriate detection and management of bacterial burden are critical to improving wound healing and preventing further serious infection or escalation of sepsis. Although all wounds contain some level of bacteria, studies have shown that bacterial loads exceeding 10⁴ CFU/g contribute to delayed healing^4,5 and potentially increase the cost of caring for these wounds.^2,6 The concept of the infection continuum published by the International Wound Infection Institute provides context on the effect of bacteria on wounds and wound healing as wounds increase in number and virulence.⁷ Low bacteria levels represent stages of contamination or colonization that should not delay healing and typically do not require antimicrobial agents. However, the bacterial burden reaches a tipping point, or level of critical colonization, at which point the number of bacteria and their virulence begin to stall wound healing and initiate a host response.⁷ This point is reached between 10⁴ CFU/g and 10⁵ CFU/g of bacteria.^4,8 Past this tipping point, an infection can occur. Local infection can occur as the bacteria move deeper into the wound and increase in number.⁷ A local infection presents with subtle signs, and early detection and intervention are essential to help prevent further escalation. Left undetected or untreated, this local infection can escalate to spreading and systemic infection as bacteria begin to invade the surrounding tissue and more overt signs of infection present, eventually resulting in serious complications.

Clinical signs and symptoms (CSS) of infection such as redness, swelling, pain, and odor are subjective and can be mistaken for inflammation, making it difficult to identify elevated bacterial loads that may cause infection.^9-11 Furthermore, these CSS may be suppressed or absent in immunocompromised individuals and patients with comorbidities.^12,13 Because of the lack of point-of-care information to aid in the wound assessment, treatment decisions are routinely made without sufficient information about wound bacterial burden. The microbiologic analysis of wound samples may take 2 to 3 days to acquire, thus contributing to treatment delays or forcing a prophylactic approach. This can lead to overprescribing antimicrobial agents or antibiotics “just in case,” which may result in antimicrobial resistance.

Antimicrobial resistance is a major threat because it can produce clinical unresponsiveness to treatment and rapid evolution to sepsis and septic shock.¹⁴ Antimicrobial stewardship programs can successfully reduce the use of antimicrobial agents without negatively affecting clinical outcomes in acute care settings.¹⁵ In line with this goal and others, many countries have engaged with the Choosing Wisely campaign, with many hospitals adopting the Choosing Wisely recommendations.¹⁶ This program aims to reduce unnecessary tests and treatments in health care, emphasizing evidence-based patient care. Choosing Wisely highlights several reasons why unnecessary tests or treatments may occur, including ingrained practice habits, outdated decision support systems, patient insistence, and defensive medicine practices. All of these reasons can be applied to decisions around antimicrobial and antibiotic use. A goal of this campaign and the current study is to promote the judicious use of antimicrobial agents or antibiotics with evidence-based prescribing. However, the challenge of how best to implement this evidence-based prescribing remains.

Fluorescence imaging has emerged as a diagnostic tool to aid clinicians in determining the presence of elevated levels of bacterial burden (>10⁴ CFU/g) in acute and chronic wounds. Point-of-care fluorescence imaging has been researched extensively to validate its diagnostic accuracy and utility for bacterial detection and treatment planning.^11,17-19 Multiple clinical trials have demonstrated that fluorescence imaging has a positive predictive value of greater than 95% for detecting bacteria at loads greater than 10⁴ CFU/g^11,17-20 and that it increases the sensitivity of detecting these bacterial loads by threefold to fourfold compared with assessment based on CSS alone.^11,17 The point-of-care fluorescence imaging device studied herein (MolecuLight i:X; MolecuLight Corp.) provides objective, diagnostic information that has been shown to alter treatment plans in approximately 70% of wounds,^11,17 including treatment decisions specifically concerning antimicrobial stewardship.^11,17,18

Both early detection of bacteria and dedicated patient care are essential to improve wound healing rates. Sunnybrook Health Sciences Center, Toronto, Canada, instituted a wound care program that consists of a practice structure including unit-based advanced practice nurses, clinical educators, wound champions, and the Complex Wound Services (CWS) team to support care for wounds with high complexity. The CWS is a specialized team that includes 3 advanced practice nurses and 1 registered nurse who specialize in managing wounds. The vision is to provide excellent, state-of-the-art care for patients with complex wounds. Specialty trained nurses with NSWOC (nurse specialized in wound, ostomy, and continence) certification, working in partnership with unit-based teams, engage in complex consultations, independent follow-up, patient and family education, adjunct therapy, transition planning, and capacity building of local leaders and teams. The Wound, Ostomy, and Continence Steering Committee at Sunnybrook led best-practice implementation and evaluation. This structure develops and maintains best-practice policies, guidelines, and procedures that promote wound, ostomy, and continence management, providing leadership and oversight to enhance overall skin health and monitor outcome measures.

This study describes the implementation of a fluorescence imaging device at 2 Sunnybrook locations to aid the CWS team and unit-based advanced practice nurses. The purpose of this study was to determine if fluorescence imaging could detect more wounds with elevated bacterial loads compared with CSS alone, provide objective information to support evidence-based treatment decisions, and help facilitate increased communication between sites for improved continuation of care. Because the device has been thoroughly researched,^11,17-20 the aim of this study was not to revalidate its bacterial detection capabilities but instead to determine how that information could inform treatment decisions. The objective of the evaluation was to measure how often fluorescence imaging affected or changed the care plan, in particular in response to the selection of antimicrobial agents.

Study population and design

This observational study was performed at 2 Sunnybrook Hospital locations (Bayview Campus and St John’s Rehab [SJR], Toronto, Canada) between November 2020 and January 2021 with COVID-19 infection protocols in place. Twenty-one adult patients with wounds of mixed etiology were recruited to this study. A total of 26 wound assessments were made, including 5 follow-up assessments. Consent was obtained from all participants. Institutional ethics review was not required for this project as determined by the Sunnybrook Health Sciences Ethics Review Self-Assessment Tool and confirmed by the Ethics Office.

Patients 18 years and older with a wound and the willingness and ability to consent were eligible to participate in the study. To be considered for inclusion, a wound must have exhibited 1 of the following factors: suspected infection, stalled and not progressing as expected, deteriorating, or present in a new patient admitted for inpatient unit care. Patients were excluded from the study if they were unwilling or unable to consent, had contraindications to routine wound monitoring, or had wounds in an anatomic location that prevented wound imaging, such as areas of extreme contours creating unavoidable shadows, areas where the device is too big to fit, or patients who could not be easily positioned for imaging of certain anatomic areas (eg, the coccyx). Each wound was managed with standard treatment as well as fluorescence imaging using the fluorescence imaging device.

CSS and fluorescence imaging assessment

A clinical assessment was recorded, including CSS of infection and whether the clinician believed moderate to heavy bacterial loads (>10⁴ CFU/g) were present in each wound. Based on the CSS assessment, clinicians were asked to develop and record a proposed treatment plan for the wound.

Fluorescence imaging was performed on each wound. The patient was appropriately positioned, and a standard image of the wound was acquired for documentation and anatomic reference. Room lights were turned off prior to capturing a fluorescence image. If adequate darkness could not be achieved, a portable single-use drape was used to provide the requisite level of darkness needed to capture fluorescence images. In some instances, a disposable sterile sleeve was placed over the fluorescence imaging device before the clinician entered the patient’s room, in accordance with infection prevention and control guidance for organisms including COVID-19, methicillin-resistant Staphylococcus aureus (MRSA), and vancomycin-resistant enterococci (VRE). The fluorescence imaging device was positioned 8 to 12 cm away from and parallel to the wound. A fluorescence image was then acquired under violet light illumination.

The presence of red fluorescence in the image is indicative of bacterial loads of greater than 10⁴ CFU/g, and cyan fluorescence is indicative of the presence of Pseudomonas aeruginosa.^11,17-19 This is distinct from wound tissue fluorescence signals, which appear as shades of green (ie, skin, slough) and dark maroon (ie, blood or granulation tissue). Figure 1 provides examples of wounds from this study with either red or cyan fluorescence that indicates the presence of bacteria at those loads. 

After obtaining the fluorescence images, the clinician examined and interpreted them. The presence of red or cyan fluorescence was indicative of elevated bacterial loads based on the extensive research done on the technology; thus, microbiologic analysis was not performed to confirm the bacterial presence in this study.

Clinicians recorded any changes to the treatment plan based on their assessment combined with additional information provided by fluorescence imaging and the clinicians’ own clinical judgment. Of particular interest were antimicrobial prescribing decisions, including systemic antibiotics, topical antibiotics, and antimicrobial dressings. This study defines antimicrobial dressings as iodine-based, silver-based, polyhexamethylene biguanide (PHMB)-based, or medical-grade honey. Topical antibiotics were ointments with antimicrobial agents with targeted action against the pathogen. Polyhexamethylene biguanide gel was recommended for its other mechanisms of action but was included in the study tracking because it has antimicrobial properties. Based on the clinical judgment of the study authors, a change in the treatment plan in 10% of assessments was deemed the threshold for a meaningful change when using this device. In cases in which the presentation of the wound bed required initiation of therapy or antimicrobial-impregnated dressing, such treatment was applied.

Patient demographics

Table 1 reports the baseline characteristics of all 21 patients across the 2 sites. There were 14 male and 7 female patients, and the average patient age was 60 years (range, 37–92 years). Twelve of the 26 wounds assessed (46%) were considered complex wounds of mixed etiology that could be categorized as 2 or more wound types (eg, diabetic foot ulcer with a surgical component).

Extensive use of antimicrobial agents during standard of care

One study aim was to assess the potential for fluorescence imaging to influence the evidence-based selection of wound care dressings and topical antimicrobial agents. To assess potential changes in antimicrobial prescribing, the baseline extent of the use of antimicrobial agents was determined. Based on the proposed treatment plan recorded before fluorescence imaging, antimicrobial dressings (iodine-, silver-, or PHMB-based, or medical-grade honey) were prescribed to manage 16 of 26 wounds (62%). Furthermore, for 4 of 26 wounds (15%), topical antibiotics or systemic antibiotics were prescribed. Overall, a total of 88% of wounds were managed (dressings, topical or oral antibiotics) (Figure 2A). More than half of the wounds (58%) managed with an antimicrobial dressing were assessed as CSS negative or not suspicious for infection (Figure 2B).

Treatment changes prompted by fluorescence imaging

Based on fluorescence imaging, 17 of 26 wounds (65%) had evidence of bacterial fluorescence (red or cyan) (Figure 3A). The clinical team identified CSS in 10 wounds found under fluorescence, with the fluorescence imaging identifying an additional 7 wounds (an additional 70%). In 62% of cases (16 wounds), fluorescence imaging correlated with CSS assessment (Figure 3B), whether positive or negative. One false-negative result based on fluorescence imaging was observed; it was confirmed by microbiologic analysis. The remaining 35% of wounds (9 wounds) represent the largest opportunity for treatment changes because they were negative for CSS but displayed bacterial fluorescence.

In this study, treatment plans were changed 35% of the time (9 of 26 assessments) after fluorescence imaging, which is higher than the a priori level of 10% that the authors considered would be meaningful. Most treatment changes focused on obtaining samples for microbiology cultures in an area of red or cyan fluorescence (3 assessments) or the additional prescription of antibiotics (2 assessments). Fluorescence imaging information also prompted extended debridement and cleaning (Figure 3C). Antimicrobial dressing decisions were not influenced by fluorescence imaging, likely because of the high level of preventive prescribing done before fluorescence imaging based on clinical judgment alone.

Because antimicrobial dressings were prescribed for 19 wounds (73%) from the outset, few patients were available for evaluating whether the fluorescence imaging information could prompt a change in treatment. Instead, the authors examined the number of patients for which a change could have been implemented (Figure 3D). This area of impact represents 12 of 26 wounds (46%). These wounds were either fluorescence positive (indicating elevated bacterial loads) and no antimicrobial dressing was prescribed (5 of 12) or were negative for fluorescence but an antimicrobial dressing was prescribed (7 of 12). Antimicrobial dressings may be prescribed for several reasons, including as a preventive measure, which may have been the case for patients in this study. However, this study highlights the potential for reduction in antimicrobial dressing use when wounds lack both CSS of infection and bacterial fluorescence and when there is not a strong case for such dressing use as a preventive measure. This was the case in 7 of 12 wounds, although potential preventive reasons may have also been applied in the decision-making process. For wounds that exhibit bacterial fluorescence, either the judicious use of antimicrobial dressings may be applied, or alternative wound care practices could be used, including debridement and cleaning followed by repeat imaging to determine the effect of the intervention.

Case study

A 30-year-old female with an incision wound on the left residual limb consented to inclusion in this study. The wound had a complex etiology, including arterial disease, diabetes, and the surgical nature of the wound. Other comorbidities unrelated to the wound included hepatitis C, hypertension, and obesity. Wound healing was thought to be delayed, with evidence of swelling and pain at the wound site. The clinician did not suspect infection or elevated bacterial loads based on clinical assessment. Prior to fluorescence imaging, the treatment plan consisted of wound cleansing, applying an iodine-containing non-sticky antimicrobial dressing, and offloading. Standard and fluorescence images were obtained using the fluorescence imaging device with a portable single-use drape. The fluorescence image displayed red (bacterial) fluorescence areas at the incision site (Figure 4, Table 2). Slough at the incision site appeared bright green. The fluorescence images were shown to the most responsible physician for review. Based on the fluorescence imaging information, the most responsible physician prescribed broad-spectrum antibiotics in addition to the planned antimicrobial dressing and offloading. A swab sample was taken from an area of red fluorescence and sent for microbiologic analysis to determine bacterial species and resistances. After 2 days, the microbiologic analysis confirmed elevated bacterial loads, including group B Streptococcus. Based on the microbiology results, the strength of the antibiotics was increased.

Eight days later, the clinician deemed the wound to be CSS positive while the patient was still on antibiotic therapy (cefazolin). Fluorescence imaging continued to display red fluorescence circumferentially around the edge of the incision and at the suture sites. Upon reviewing this fluorescence image, the physician changed the course of the antibiotic therapy and scheduled surgical debridement of the wound by an orthopedic surgeon.

In this case, fluorescence imaging provided information on bacterial burden at the point of care before CSS became evident. Based on this early knowledge, the treatment plan was changed to address the bacterial presence. Microbiologic analysis both confirmed the finding on fluorescence imaging and provided information on bacterial species and resistance that allowed the physician to choose the appropriate antibiotic course. Monitoring the wound over time provided continual information on the bacterial burden within the wound.

The implementation of this fluorescence imaging device into routine use by the CWS and SJR advanced practice nurses team demonstrated strong clinical benefits for fluorescence imaging. Use of such imaging aided in detecting bacteria-laden wounds and influenced treatment decisions. Much was learned throughout the process about institutional use of antimicrobial dressings, situations in which fluorescence imaging was the most useful, and the benefits of bacterial fluorescence images for documentation and communication between wound care teams.

In this study, fluorescence imaging increased the detection of wounds with elevated bacterial loads by 70% compared with clinical assessment alone. These findings are consistent with numerous studies reporting the substantial benefit of fluorescence imaging to increase the detection of wounds with elevated bacterial loads missed based on CSS alone^11,17,18 and are most similar to the findings reported in a study by Hill and Woo,¹⁸ which also focused on wounds within a hospital setting. Factors unique to the inpatient hospital setting, such as increased care time and more advanced wound care education, may have contributed to the better performance of CSS assessment in this study vs the performance of CSS assessment in studies that were performed in the outpatient setting.^11,17

A surprising discovery was the extensive use of antimicrobial dressings in this patient population, regardless of CSS or fluorescence imaging information. Based on CSS assessment alone, an antimicrobial dressing was prescribed for 73% of wounds, even though almost half of those wounds did not have noticeable symptoms. This prophylactic approach may have been chosen because many study wounds were often in areas that could easily become colonized with bacteria (eg, sacrum). In addition, given the known subjectivity and unreliability of CSS,^9-11 clinicians may have opted to err on the side of caution and thus prescribed antimicrobial dressings to prevent or minimize the likelihood of infection. A recent study that examined antimicrobial prescribing practices in 350 chronic wounds in the outpatient wound care setting found that antimicrobial agents (dressings, topical and oral antibiotics) were prescribed at the same rate between wounds that presented with CSS of infection and those that presented without CSS of infection, resulting in overprescribing and underprescribing.²¹ This tendency for prophylactic prescribing has been identified as a key challenge in antimicrobial stewardship in wound care,⁸ and recent articles have suggested fluorescence imaging as a potential solution.^18,22-24 Additional research validating the need for prophylactic antimicrobial dressings based on wound location would be of value in the further development of evidence-based guidelines for prophylactic antimicrobial dressing use.

Support for the use of fluorescence imaging to aid in antimicrobial decision-making has increased in recent years.²² In clinical trials, fluorescence imaging has influenced antibiotic stewardship decisions in more than 50% of wounds,¹¹ and real-world data indicate that fluorescence imaging supports more judicious prescribing^18,23 and decreases the use of antimicrobial agents or antibiotics when warranted.^11,23 In a recent retrospective analysis, a hospital outpatient wound care center demonstrated a 49% decrease in the prescribing of antimicrobial dressings with the use of fluorescence imaging in their typical clinical practice compared with the previous year in which fluorescence imaging was not used; this change had a positive effect on healing outcomes.²⁴ Not only did the use of fluorescence imaging support antimicrobial stewardship goals, it also resulted in a cost savings of 33% on antimicrobial agents for the year; a 10% total savings per patient also was projected. Furthermore, these results were observed in a year with a 27% increase in the number of wounds seen compared with the previous year. The cost of antimicrobial agents is one of the most fluid in wound care; thus, such use of fluorescence imaging to improve the targeted use of antimicrobial dressings translates to improved patient care and potential cost savings for the patient and the health care system. The findings of the current study have the potential to support cost savings through continued use of fluorescence imaging and enhanced decision-making related to wound care products, antimicrobial agents, and debridement. It could be informative to conduct a similar retrospective analysis following 6 to 12 months of routine fluorescence imaging use in the current study authors’ system.

The use of interprofessional health care teams to provide wound care has been widely adopted.^25,26 Best-practice guidelines stress the importance of the accurate and multidirectional flow of meaningful information between the wound care team members and across the various care settings.²⁷ Hill and Woo¹⁸ demonstrated the utility of the fluorescence imaging device to improve communication between members of a interprofessional wound care team, improve treatment planning, and expedite interventions when required. In the current study, increased communication and standardization of care between the principal hospital site (Bayview Campus) and rehabilitation center (SJR) within the Sunnybrook Hospital system were observed. Current access to the specialized CWS was limited across organizational campuses in this organization. A recent enhancement to the team resources has been approved to enable the CWS to further support remote consultations in postacute clinical settings. Implementation of fluorescence imaging at the SJR site, review of the images through interprofessional team rounds, and virtual consultation across sites improved assessment accuracy and standardization of treatment interventions. Future scaling will include implementation on all postacute campuses.

This study has several limitations, foremost of which is the small sample size. Although fluorescence imaging was shown to be successful as an initial evaluation, data on wound follow-up were recorded for all wounds, and information on patient outcomes is lacking. This inhibits any conclusions on how such imaging may affect patient care in the long term. The authors would like to perform a retrospective study, similar to the study performed by Price,²⁴ to determine the effect on various outcome measures, including wound healing and the use of antimicrobial dressings. Although the fluorescence imaging device discussed herein is generally easy to implement, a limitation is the time required to perform the imaging and interpret the images correctly. The study was performed amid restrictions caused by the COVID-19 pandemic and a specific COVID outbreak at one of the sites. This inhibited the rapid recruitment of patients to this study and resulted in additional restrictions and duties performed by the study team. For example, a sterile sleeve was used for patients under precaution because of COVID-19, patients under investigation for COVID-19, or other issues such as MRSA or VRE, which added time to the evaluation process. In addition, the device cannot capture fluorescence signatures from tunneling or undermining in the wounds because of lack of sight lines and limited depth penetration of the illumination light.²⁸

Another limitation of this study was the lack of microbiologic analysis to confirm the presence of elevated bacterial loads. To compensate for this limitation, the clinicians chose to rely on the large body of evidence that supports a high positive predictive value of red or cyan fluorescence to determine the presence of elevated bacterial loads in wounds. Future research on how the device is implemented into the authors’ wound care program would be useful. A larger sample size and additional patient follow-up would be necessary to better assess the overall benefit of the device in the authors’ health care system. A final limitation is that the evidence supporting antimicrobial stewardship is based on studies of systemic antibiotics rather than on topical use of antimicrobial dressings and focuses on the economic value rather than on patient outcomes. As such, more research is needed.

This small evaluation study demonstrates a strong case for fluorescence imaging to support clinical decision-making in wound care. Fluorescence imaging detected 70% more wounds with elevated bacterial loads compared with clinical judgment alone and led to treatment changes in 35% of cases. Although antimicrobial dressing use was high in this patient population, results from this study suggest that the incorporation of fluorescence imaging into antimicrobial decision-making may support a decrease in antimicrobial dressing use. The use of fluorescence imaging can also support a more standardized approach to wound care as well as better communication and collaboration between hospital sites throughout the course of a patient’s care.