Lack of Reliability of Clinical/Visual Assessment of Chronic Wound Infection: The Incidence of Biopsy-Proven Infection in Venous
For centuries, physicians have diagnosed skin and soft tissue infection using the classic signs and symptoms of erythema (rubor), warmth (calor), tenderness, purulent drainage, pain (dolor), and swelling (tumor). For the most part, these findings are reliable indicators of infection in the majority of acute and post-surgical wounds. However, Robson and Heggers1 demonstrated that surgeons evaluating wounds preoperatively by visual inspection prior to flap closure were unable to reliably determine which wounds were infected. The wounds with ≥ 106 colony-forming units (cfu) per gram of tissue on quantitative biopsy had significantly less flap survival. In 2001, Gardner et al2 evaluated several different types of wounds and found little correlation between wound bed infection and the classic signs of infection. The signs that had some positive predictive value were delayed wound healing over time, friability and discoloration of granulation tissue, pocketing at the base of the wound, foul odor, wound breakdown, and increased pain. Unfortunately, these signs and symptoms are largely subjective and require examination of the wound over a period of time. In short, the traditional signs of infection are unreliable in chronic wounds.
Despite this information, a survey of wound care practitioners in the United States revealed that the physical characteristics of the ulceration are most commonly used for diagnosing wound infection.3 The reason for this disparity is likely to be the lack of a readily available objective diagnostic test and the ongoing belief, passed on by tradition in medical and nursing schools, that observation alone is a valid assessment. The gold standard for diagnosing infection in chronic wounds is a quantitative tissue biopsy with > 105 cfu per gram of tissue of any organism with the exception of beta-hemolytic Streptococci for which any level of the organism indicates infection.1,4 However, quantitative biopsies are invasive, expensive, painful, and difficult for the inexperienced clinician to perform appropriately and are unavailable to many practitioners. Some physicians have resorted to using swab cultures; however, there are several problems associated with swab cultures that make the information obtained of little value. First, a standard technique for obtaining a swab culture that is used routinely across the United States does not exist. In addition, practitioners often swab the wound without first removing nonvitalized and necrotic material. Furthermore, there is little correlation between routine swab cultures and tissue biopsies. Swab cultures of the wound reflect bacteria that have colonized or contaminated the wound surface and not bacteria that have infected the tissue within the wound—a pivotal criterion in the definition of wound infection. Moreover, the results of swab cultures are reported in a nonquantitative manner, which oftentimes makes treatment decisions haphazard and, in some cases, excessive. There have been reports of the use of semi-quantitative swab cultures.5 However, this test is more rigorous than a routine swab culture and requires a specific type of swab and procedure, which is not widely available and is used infrequently. In the absence of quantitative techniques, the clinician is left to evaluate the presence or absence of physical characteristics, which is unreliable and inaccurate.
Relying solely on the physical characteristics of the wound has at least 2 consequences. Clinicians tend to treat the inflammation associated with chronic wounds with systemic antibiotics when potentially no infection is present, which has led to the overuse of antibiotics and promoted bacterial resistance. The opposite is also true: clinicians, particularly wound care professionals accustomed to treating chronic wounds, sometimes fail to treat infection when infection is present because the classic signs of infection are absent.
Venous leg ulceration (VLU) creates the most confusion in regard to bacterial burden. The characteristics of chronic venous disease (intense inflammation surrounding the wound, induration and warmth of the skin, pain, swelling, tenderness to the touch, copious wound drainage, and foul odor) in combination with the typical appearance of a venous ulceration are easily confused with infection. In the authors’ experience, practitioners outside the wound care field routinely prescribe systemic antibiotics as the primary treatment. Antibiotics are frequently employed prior to compression therapy, the current standard of care.6 However, this practice is also found in wound care clinics as evidenced by the ubiquitous use of silver as a topical antimicrobial. It has become common practice to treat all chronic wounds empirically with topical antimicrobials. This is not optimal patient care, as it leads to the over-treatment of a large number of patients. Moreover, this practice is costly, and a growing body of evidence exists that the use of topical antimicrobials may injure healthy cells.7 Sole reliance on clinical signs of infection in VLUs can lead to the overuse of systemic antibiotics, which promotes bacterial resistance, and the inappropriate treatment of patients with topical antimicrobials, which is expensive and may be detrimental to wound healing or, in some cases, unnecessary.
Conversely, clinicians may not treat infected VLUs because they lack signs of infection. The wound care clinician may correctly identify the distinctive presentation of a VLU. However, it is still possible that the wound is infected. The mere absence of the cardinal signs and symptoms of infection in these ulcers is equally unreliable. In this case, it is possible that patients will not receive appropriate antimicrobial treatment, which could lead to worsening infection and prolonged wound healing.
To date, infection in VLUs has not been studied in a large group of patients. The recent clinical trial testing the safety and efficacy of repifermin (KGF-2) in the treatment of VLUs provided the opportunity to objectively evaluate the incidence of infection in VLUs.
The KGF-2 clinical trial was conducted by 55 investigators across the United States with 614 individuals screened and 352 ultimately enrolled. Candidates for the trial had VLUs within the gaiter region of the leg. The ulcers had to be present for > 1 month. The size of the ulcers was between 3 cm2 and 25 cm2 in area. Patients with clinical evidence of infection were excluded. Venous insufficiency was documented by venous duplex scan. As part of the screening process, a 6-mm quantitative biopsy of the wound bed was obtained to rule out infection. The biopsies were evaluated by a single laboratory. Investigators were instructed on the standardized technique for obtaining quantitative biopsies.3 Moreover, patients with infected ulcerations were not enrolled until the infection had resolved as determined by a subsequent biopsy. Subjects could undergo as many as 3 biopsies with treatment episodes interspersed prior to being disqualified as candidates for inclusion in the trial. Treatment during the interval between tissue biopsies was at the discretion of the investigator and could be topical or systemic. Patients with greater than 105 cfu or the presence of 1 beta-hemolytic Streptococcus per gram of tissue were excluded from the study if, after 3 consecutive biopsies, none were found to be negative. The results of this clinical trial have been published elsewhere.8 In this article, the authors focus on the ability of wound care physicians to determine, based on clinical characteristics, the infection status of VLUs.
Materials and Methods
The objective of the KGF-2 phase IIb clinical trial was to compare the safety and efficacy of 2 doses of the truncated form of the growth factor, repifermin (KGF-2), applied twice weekly under standardized compression (Dynaflex, Johnson & Johnson Wound Management Worldwide, Somerville, NJ) versus standardized compression (Dynaflex) alone in the treatment of VLUs. Adaptic (Johnson & Johnson Wound Management Worldwide), a nonadherent dressing, was employed as a primary wound dressing. The primary end point was complete closure (100% closure with no drainage and no need for a dressing) at 20 weeks. A screening period of up to 48 days was permitted. During this time, patients underwent venous duplex examination to demonstrate venous insufficiency and arterial Doppler to rule out concomitant peripheral arterial disease. Standardized wound care was performed including debridement. A quantitative biopsy was obtained to determine if infection was present.
Principle investigators were instructed on the proper standardized technique to obtain the tissue biopsy. The wound bed was prepped with sterile normal saline. Local anesthetic was injected into the wound bed. Topical analgesics were not employed. Debridement of any fibrinous exudate or biofilm was recommended. A punch 6-mm biopsy was obtained and immediately placed in the preservative media provided. Investigators were instructed not to enroll patients who showed signs of infection. If the biopsy did not demonstrate infection, the patient was placed in compression and the screening process continued. If the biopsy revealed an active infection (> 105 cfu or any level of beta-hemolytic streptococci), the patients were treated with systemic or topical antibiotics for at least 10 days and re-biopsied up to 3 times until the results were negative.
Patients were enrolled and randomized only after all inclusion and exclusion criteria were met.
Results
A total of 614 patients underwent screening biopsies of their VLUs. Of these, 352 patients were eventually enrolled in the study by 55 investigators. The mean age of the patients was 61.2 years. There were more men (61.3%) than women. Quantitative biopsies were positive for infection in 122 of the patients who underwent a screening biopsy (19.9%). If only enrolled subjects were included in the analysis, 92 of 352 biopsies were positive (26%). This may be a more accurate reflection of the rate of infection given that at least 10% of the screen failures may not have had ulcers of venous origin as evidenced by a lack of ultrasonic evidence of venous insufficiency.
If the first biopsy was positive (ie, demonstrated infection), the protocol allowed for repeat biopsy after treatment. A total of 82 patients underwent 2 biopsies and 10 patients had 3 biopsies performed prior to enrollment. There was a trend toward lower healing rates in the patients who underwent more than 1 biopsy: In the single biopsy group, 60% of the patients healed in 20 weeks; in the 2 biopsy group, 34 patients went onto complete closure (51%); and in the 3 biopsy group, only 4 patients (40%) healed (Figure 1). These differences demonstrated a trend toward poorer healing in the multiple biopsy groups, but they were not statistically significant (P = 0.15).
The majority of the VLU infections were monomicrobial. Polymicrobial infections occurred in 5% of ulcers overall and represented 23% of the infected ulcers among the enrolled patients. Staphylococcus aureus was the infecting organism in 41.1% of patients. No beta-hemolytic Streptococci were cultured, and all streptococcal species accounted for 14% of invasive infections. Pseudomonas aeruginosa represented the second most common infecting organism (14%). The Enterobacteriaceae family represented 35.6% of all of infected ulcers (Figure 2).
The polymicrobial infections included 2 infecting organisms 64% of the time and 3 infecting organisms 18% of the time. There were no ulcers with more than 4 organisms at levels > 105 cfu. The polymicrobial infections contained the same bacteria seen in monomicrobial infections: 16 biopsies contained Staphylococcus aureus (72%), 10 Pseudomonas aeruginosa (45%), and 19 cultured members of the Enterobacteriaceae family (86%).
In biopsies without infection, the prevalence of the different bacterial species was relatively similar. In a comparison between organisms at > 106 versus those at 105, the prevalence of the bacterial species was remarkably similar (Figure 3).
Discussion
The diagnosis of infection in chronic wounds is problematic because accurate diagnostic techniques are not currently available, and the physical characteristics of various chronic wound types are misleading and unreliable. This dilemma is perhaps most pronounced in the evaluation of VLUs. The wound care professional encounters patients on a daily basis who have received systemic antibiotics for VLUs that were presumed to be infected when, in fact, no objective test for infection was obtained. This frequent practice has led to increasing development of resistant bacteria, increased cost for the care of these wounds, and prolonged healing. However, it is also clear that the practitioner cannot rely on physical examination to rule out infection as demonstrated by the evidence from this multicenter clinical trial.
In this large, multicenter trial, investigators were instructed to screen patients for infection status based on their clinical judgment. Their assessment was, however, verified through an objective value provided to them following a tissue biopsy obtained from within the wound. The results indicated that 74% of the time experienced practitioners participating in a clinical trial were able to determine infection status of a VLU by clinical examination. This has far reaching implications for the treatment of the millions of patients with VLUs who present to wound care centers and physician offices across the United States. The clinical trial setting represents an ideal patient encounter where patients undergo an in-depth examination far more rigorous than a routine patient visit in the wound clinic setting. One can reasonably conclude that, in addition to varying levels of expertise of wound care professionals as well as the brevity of time a patient is seen in the busy wound clinic, the ability to accurately diagnose infection in a VLU falls well below 74%.
One might state that if the signs and symptoms of infection are not present, high levels of tissue bacteria are irrelevant. However, it has been demonstrated that every process of healing is inhibited by high levels of bacteria, and VLU healing is inhibited in the presence of high levels of bacteria.9,10
The recently completed Wound Healing Society’s Guidelines for the Treatment of Venous Leg Ulcerations recommends patients undergo a quantitative biopsy if a high suspicion of infection exists or if the wound has failed to progress after 2 weeks of standard therapy (unpublished data). In light of the data from this trial, this is a reasonable recommendation. However, for investigators and clinical sponsors evaluating new pharmaceuticals and devices to promote the healing of VLUs, a negative quantitative biopsy prior to enrollment is essential to ensure that the data obtained from the trial are reliable and not confounded by unrecognized bacterial infection.
Analysis of the bacteriology of the VLUs studied in this trial demonstrates principally monomicrobial infections. This correlates with what has been previously reported when tissue biopsies are used for diagnosis instead of swab cultures.11Staphylococcus aureus was predominantly found followed by the Enterobacteriaceae family, particularly Pseudomonas aeruginosa. When polymicrobial infections were encountered, there were only 2 or 3 organisms present. These bacteria were the same organisms seen in the monomicrobial infections with Staphylococcus aureus and the Enterobacter family present in the vast majority of the biopsies with > 1 organism at levels greater than 105 cfu. Interestingly, at bacterial counts < 106, there is a similar prevalence of bacterial species. This suggests that replicating organisms present within ulcers at levels below 106 are likely to be the source of the ensuing infection in this particular ulcer type if the wound is not properly cared for in the interim.
The presence of wound infection prior to enrollment in the trial resulted in a trend toward poorer healing even after the wound bed infection had been treated. The patients who had 3 biopsies had the worst healing rates compared to those undergoing a single biopsy or 2 biopsies. The best explanation for this finding may be that active infection led to changes in the wound microenvironment that persisted after antimicrobial treatment or infection that had to be overcome before the salubrious effect(s) of any treatment occurred.
