Dear Readers: In this installment of the Evidence Corner, we venture into the world of the infected wound. How do we know when a wound is infected? Is it a matter of “I’ll know it when I see it?” Do the symptoms of infection differ between acute and chronic wounds? Theoretical definitions of infection usually describe it as invasion of viable tissue by one or more types of microorganisms influenced by compromised host defenses, virulence of the organisms, and local tissue risk factors.[1] Yet practical measures of infection often fail to meet scientific standards of reliability and either diagnostic or predictive validity. The two articles below explore evidence addressing the reliability and validity of surgical and chronic wound infection measures with some surprising conclusions.

Quality of Surgical Wound Infection Measurements: A Systematic Review

Reference: Bruce J, Russell EM, Mollison J, Krukowski ZH. The quality of measurement of surgical wound infections as the basis for monitoring: A systematic review. J Hosp Infection 2001;49:99–108.

Rationale: Attempts to standardize the definition of surgical wound infection are challenging for both surgeons and infection control personnel.

Objective: Systematically assess the validity and reliability of definitions and methods of measuring surgical wound infection.

Methods: Authors searched the MEDLINE, CINAHL, EMBASE, Cochrane Library, and HealthSTAR databases from 1993 to 1999 for studies of inter-rater and intra-rater reliability, criterion, construct, and content or face validity of measures defining surgical wound infection. Studies were included in the analysis if a definition of surgical wound infection and/or details of wound assessment were described or if reliability and/or validity of the definition were assessed.

Results: Of 2,490 abstracts identified, 90 prospective studies from 20 countries qualified for inclusion in the analysis. These included 41 separate definitions of wound infection and eight articles formally assessing validity or reliability of the definition of surgical wound infection. Five “standard” definitions of wound infection were identified: the Public Health Laboratory Service (NPS), the Surgical Infection Society Study Group, the Second UK National Prevalence Survey, and the Centers for Disease Control (CDC) 1988 and 1992 definitions. No single symptom was common to all definitions. The most common stand-alone criteria of infection were purulent discharge (54 studies, with positive bacterial culture mandatory in 12 of these 54 studies) and erythema or swelling (8 studies). Content, criterion, and construct validity studies of the different infection scales suggested that the ASEPSIS and Southampton Wound Assessment scales are more sensitive than the NPS and CDC definitions. Reliability of infection diagnoses ranged from 0.73 (physician inter-rater correlations with patient diagnosis of purulent discharge) to 0.96 (physician correlated with nurse diagnosis using ASEPSIS), with the “classic signs and symptoms” of infection (purulent discharge, tenderness, erythema, swelling, induration, warmth) yielding eight-percent false positives and 48-percent false negatives by patients as compared to nurses and doctors. The basic core symptoms in most studies included wound discharge (purulent or otherwise), redness or erythema, swelling or edema, pain, tenderness, heat, pyrexia, dehiscence, and separation of wound edges. Bacteria can be isolated from wounds that are healing without clinical signs of infection and conversely are often not isolated when cultured from early wound infections.

Conclusions: Surgical wound infection diagnosis is often subjective, based variably on the presence and severity of several properties, with no single objective gold standard test, not even isolating the organisms from tissue biopsies or wound fluid.

Validity of Signs and Symptoms of Chronic Wound Infection

Reference: Gardner SE, Frantz RA, Doebbeling BN. The validity of the clinical signs and symptoms used to identify localized chronic wound infection. Wound Repair Regen 2001;9:178–86.

Rationale: Classic signs of infection, pain, erythema, edema, heat, and purulent exudate may be diminished or altered in chronic wounds, despite high wound bioburdens. Conversely, continued tissue injury can trigger inflammatory responses that mimic these signs of infection without high bioburdens causing false positive diagnoses of chronic wound infection. Cutting and Harding[2] suggest using secondary wound infection symptoms, such as serous drainage with concurrent inflammation, delayed healing, discolored granulation tissue, pocketing at the wound base, foul odor, and wound breakdown, as diagnostic signs of infection.

Objective: Examine the validity (sensitivity and specificity) of clinical signs and symptoms of infection used to identify localized chronic wound infection in chronic wound patients as compared to the “gold standard,” which is quantitative cultures of viable wound tissue.

Methods: A cross-sectional study of chronic pressure, venous, or diabetic foot ulcer patients without arterial disease in two Midwestern US Veterans’ Administration facilities, a state-owned long-term-care facility, and a university clinic serving both outpatients and inpatients explored classic and secondary signs and symptoms of wound infection using a Clinical Signs and Symptoms Checklist (CSSC) with inter-rater reliability Kappa item values ranging from 0.53 to 1.00. CSSC item scores were correlated with wound infection defined as colony-forming units (CFU) of aerobic and anaerobic microbial populations >105 CFU per gram of viable tissue or b-hemolytic Streptococcus at any level in tissue harvested in standard biopsies. Dry gauze was left in place on the wounds for one hour before evaluating the CSSC and harvesting the biopsy. Patient risk factors for chronic wound infection also measured included age, presence of diabetes mellitus, current use of steroid or antibiotic medications, white blood cell count, type and amount of wound bed tissue, stage and size of wound, tissue oxygenation (TcPO2), patient nutritional status, and type of wound dressing.

Results: Of 36 patients enrolled in the study, 11 (31%) had at least one wound meeting the presumed definition of infection (>105CFU/g of tissue or any b-hemolytic Streptococcus), which, as would be expected, was negatively correlated with patient systemic antibiotic therapy. Wounds with >105 CFU had lower TcPO2 and more necrotic tissue. The most sensitive classic sign of infection was erythema (0.54), while the most sensitive secondary wound signs of infection were friable granulation tissue (0.82), delayed healing (0.81), and discoloration (0.64). The classic signs were more specific (all >0.64) than sensitive. Pain and wound breakdown were most specific (1.00), i.e., not having the symptom was associated with less than 105CFU per gram, and both had 1.00 positive predictive values, i.e., all wounds with these signs had >105CFU per gram or some b-hemolytic Streptococcus.

Conclusions: No sign or symptom achieved 100-percent sensitivity, suggesting that none are crucial or necessary in diagnosing clinical infection; in other words, heavy colonization often occurs without a sign or symptom of infection. However, increasing wound pain or tissue breakdown were both perfect predictors of tissue invasion by organisms, the absence of which was perfectly correlated with the disappearance of these two symptoms, so pain and tissue breakdown may be considered sufficient signs of infection. The authors conclude that if either of these symptoms is present, the wound may be regarded as infected.

Clinical Perspective

While both surgical and chronic wound infection appear to have no gold standard for diagnosis, at least two symptoms, increasing pain and wound breakdown, were common core symptoms of infection in both surgical and chronic wounds. These were 100-percent predictors of microbial invasion into tissue surrounding chronic wounds, with friable granulation tissue and foul odor also having a better than 50:50 chance of predicting microbial invasion. Alert clinicians may note either increasing pain or tissue breakdown in chronic or surgical wounds as a call to act quickly to alleviate causes of further tissue breakdown or reduce bacterial burden or both to avoid the progressive tissue damage accompanied by microbial invasion that we call “wound infection.”

These findings raise the question, “Which comes first: factors causing pain and tissue breakdown, which predispose the tissue to microbial invasion, or organism invasion into the tissue. which initiate the pain and tissue breakdown?” Bacteria are rarely regarded as necessary and sufficient causes of infection, but rather as opportunistic invaders after host immune defenses and predisposing factors place tissue at risk. Local factors, such as oxygen and tissue perfusion,[3] or use of gauze dressings, which desiccate the wound and may leave foreign bodies in it to potentiate infection,[4] can play roles in delaying healing or allowing the development of infection. Indeed, artifacts due to the one hour of gauze used to standardize CSSC measurement by Gardner, et al., mentioned previously, may have limited capacity to observe the previously reported association of hydrocolloid dressings with lower clinical infection rates of chronic and acute wounds.[5]

The review by Bruce, et al., suggests that the criterion of >105CFU in a tissue biopsy may not be the gold-standard definition of infection. Earlier work using this criterion showed 35.7-percent sensitivity or 64.3-percent false positives for biopsies in diagnosing invasive infection in burn patients.[6] The same study showed that a biopsy with >105CFUs predicted clinical symptoms of infection 89.7 percent of the time, raising the question of whether the biopsy needle carries organisms deeper into compromised tissue, potentially increasing the zone of microbial invasion. Like a variation on the Heisenberg Uncertainty Principle, harvesting the biopsy may increase the likelihood of invasive infection in biopsied tissue. Quantitative swabs or irrigation aspiration may be less invasive options for isolating organisms in chronic or acute wounds with high reported validity.[7,8]