Criteria for Identifying Wound Infection — Revisited
This article first appeared in the British Journal of Community Nursing. 2004;9(3 Suppl);S6–S16.
T he accurate identification of wound infection is a challenge for any clinician involved in this area of care and can have a significant impact on patient morbidity. The more obvious infection signs, such as purulent discharge and spreading erythema, are generally recognized as diagnostic. However, these features are not always present in the early stages when diagnosis is important for treatment and the avoidance of complicating sequelae.
A number of subtle signs (clinical indicators) that herald the onset of infection have been proposed. Following a review of the available literature, Cutting and Harding1 attempted to collate the indicators of infection. The aim was to include all clinical indicators of acute and chronic wound infection that had been used by colleagues in the field and had been shown to be of value, either by being generated through research or through empirical findings. These collated criteria appear to have gained acceptance not only in the UK but also elsewhere.
In addition to the use of clinical criteria to assist in the identification of wound infection, strong opinion exists that quantitative bacteriology is of significance.2-4 However, reliance on this approach has been challenged by Bowler,5 who maintains that the types of organisms, their interactions with not only each other but with the wound environment, and the local conditions are relevant factors for consideration together with host resistance.
Cutting and Harding1 brought to the attention of many clinicians subtle criteria (see Table 1) that may not have previously been considered. For an explanation, the reader is referred to the original paper. These criteria provide a reminder or checklist and it is likely that many clinicians have put the criteria to the test in their own clinical practices. For any clinical tool to be of proven value it needs to be tested and there have been two validation studies challenging the 1994 criteria; these are Cutting6 and Gardner et al.7
Cutting6 tested the criteria by asking ward nurses to view patients’ wounds and to make a decision on the infection status of the wound by using their own criteria. These decisions were then compared with the researcher’s verdict, using the 1994 criteria, and a microbial assay of the wound taken via wound swab. A consultant microbiologist also took decisions on the infection status of the wound from results of the cultures. A total of 20 nurses took part in the study. Two nurses at a time viewed four wounds, so a total of 40 different patients’ wounds were seen which allowed 80 opportunities for separate decisions to be made. Although the types of wounds included in the study were not made explicit in the publication, all of the wounds were healing by secondary intention and did not include burns or leg ulcers. The findings in this study indicated that the criteria have a high degree of validity. Thirty-nine of the 40 decisions (97.5%) made by the researcher on the infected status of the wounds were corroborated by the wound swab culture.
Gardner et al7 examined the validity of the classic signs of infection (pain, erythema, edema, heat and purulence) and “… signs specific to secondary intention wounds (ie, serous exudate, delayed healing, discoloration of granulation tissue, friable granulation tissue, pocketing at the base of the wound, foul odor, and wound breakdown).” The wound types against which the criteria were tested were described in the study as “a mix of chronic wounds,” and subjects were enrolled who had a “nonarterial chronic wound.” These types of wounds are defined as “… wounds caused by prolonged pressure, venous insufficiency, peripheral neuropathy, surgical incision (healing by secondary intention) or trauma.”
One of the criteria included by Cutting and Harding1 — “bridging of the epithelium or soft tissue” —was omitted from the Clinical Signs and Symptoms Checklist (CSSC) used in this study and substituted with “erythema.” The list of 12 features included in CSSC is presented in Table 2.
The results show that the signs specific to secondary wounds are more accurate as indicators of infection than the classic signs. The only sign that did not demonstrate validity was pocketing of the wound base. This is an unexpected finding, as pocketing at the base of the wound was found by Marks et al8 to be indicative of deficient granulation tissue due to an uneradicated infection, and therefore diagnostic of infection in pilonidal sinus excisions left to heal by secondary intention. To the authors’ knowledge, this feature (pocketing at the base of a wound) has not been observed or reported as an indicator of infection in a pressure ulcer or ulcer of the lower limb. The signs showing validity (based on the four parameters of sensitivity, specificity, discriminatory power, and positive predictive value) were increasing wound pain, friable granulation tissue, foul odor, and wound breakdown.
Cutting and Harding1 produced a set of objective and subjective criteria to help identify wound infection. While criteria that are subjective in nature are thought to be of limited value, Gardner et al7 found that the inter-rater reliability of the subjective elements (touch, color, and smell) were capable of assessment in a consistent manner and required little training.
Although most infected wounds exhibit common “classical” criteria, there are subtle differences between wound types. A weakness of Cutting and Harding1 was that the original set of criteria developed did not differentiate between different wound types. This was because of a failure, at the time, to recognize that diverse wound types could exhibit different criteria. It is interesting to note that this tenet appears to have been also omitted from a joint consensus meeting of the European Tissue Repair Society and the European Wound Management Association considering chronic wound infection.9 Focus on different wound types is now considered to be essential so that these understated and possibly misunderstood variations are recognized, enabling accurate identification of infection.
The following sets of criteria have been developed using a similar process to Cutting and Harding.1 The literature has been reviewed and, where possible, criteria deemed to be indicative of infection in a particular wound type have been generated. The purpose of this exercise is to draw attention to the variations in criteria that may occur between different wound types. There is undoubtedly some overlap between groups and, if this were not the case, the Cutting and Harding1 criteria would have been of limited benefit in the intervening years. Reviewing the literature to assist in identifying infection in fungating wounds was not pursued. These lesions present enormous challenges to the clinician and the effects of the underlying pathology results in devitalized tissue and polymicrobial activity, the effects of which are almost impossible to determine.
Criteria for Identifying Infection
Diabetic foot ulcers. Foot ulceration and infection are major causes of hospitalization in people with diabetes.10 Signs of infection in these lesions are likely to be “masked” as people with diabetes may not show typical inflammatory response to infection (pain, erythema, swelling, and leucocytosis). Despite this inherent difficulty, diagnosis of infection is essentially clinical.11 In addition, infection of the diabetic foot often involves superficial and deep tissues, including bone. This is unlike other common chronic wound types and will accordingly manifest in different criteria for infection (see Table 3 and Figure 1). The current standard, known as PEDIS (perfusion, extent, size/depth/tissue loss, infection, and sensation), is an infection classification system.12 The PEDIS system is complimentary to the approach taken with the diabetic foot Delphi panel.13
Pressure ulcers. Heggers14 suggested that pressure ulcers might not only become infected but also that infection may be implicated in etiology. The rationale for this is that the localization of pressure encourages bacterial concentration in that area and bacterial enzymes and toxins15 precipitate loss of superficial skin.
The identification of wound malodor is reliant on the individual’s acuity of smell1 and depends on the microbial species present.16 Anaerobes are traditionally linked to malodor but the influence of aerobic or anaerobic synergy should not be ruled out as a cause of malodor, particularly when Gram-negative anaerobes are involved.16
There is controversy over the association between undermining and pressure ulcer infection (undermining refers to the extension of the lesion under the skin such that the wound cavity is larger than the surface opening). Bliss17 has stated that undermining is not related to infection but is caused by proteolytic lysis of dermal tissues. However, Bliss did not consider that the proteases might be of bacterial origin. Conversely, Sibbald18 argues that undermining may be the result of bacterial digestion or prevention of granulation tissue. Some bacteria produce proteolytic enzymes known as invasins15 that digest protein and extracellular matrix components, so enlarging the wound.
The criterion of pain requires careful consideration with respect to infection. Damage or irritation to peripheral nerves (neuropathic pain) may be a component of pressure ulcer development.19 If infection intervenes, the pain is likely to increase and erythema and induration more than 2 cm from the wound edge may well be concurrent manifestations19 (see Table 4 and Figure 2).
Acute and surgical wounds. Many clinicians rely on the classical signs of infection (pain, erythema, edema, heat, and purulence) to identify infection in this type of wound. Although these are often relevant, the infection detection rate is dependent on the criteria used. This fact is clearly demonstrated by Ayliffe et al,20 who showed that a wound infection prevalence rate of 15.6% (of a sample of 3,354 wounds) dropped to 6.9% when pus was the only criterion for infection. Suggested criteria for acute or surgical wound infection are listed in Table 5 with provision made for wounds that are healing through either primary or secondary intention (see Figure 3).
Venous leg ulcers. The presence of lipodermatosclerosis and haemosiderin staining can mask features such as erythema, and pain is a common characteristic of these ulcers even without infection. Therefore, the permutation of signs and symptoms must be interpreted in context of patient familiarity, notably changes in presentation (see Table 6 and Figure 4).21
Arterial leg ulcers. Arterial ulcers of purely ischemic etiology exist independently of cutaneous haemosiderin deposits (as is the case with venous and complex ulcers). Consequently the skin is ostensibly normal in appearance with erythema clearly evident (Figure 5). As with ischemic diabetic foot ulcers, the redness of erythema may be difficult to distinguish from the erythema of cellulitis. Furthermore, infection (see Table 7) may present as bluish-purple discoloration of soft tissues, owing to the increased metabolic demands of infection, combined with reduced blood flow to the skin secondary to septic vasculitis of the cutaneous circulation.21
Full-thickness burns. Third-degree burns (see Table 8), particularly after tangential excision, are generally anesthetic; therefore, pain is not always an infection criterion. Similarly, erythema as a result of infection may not be distinct from normal wound inflammation or where areas of the wound are of partial thickness. Increase in the level of exudate from infection is also difficult or impossible to determine because of the high volume of fluid produced in full thickness burns.23
The development of the above criteria may be considered as an interim measure as they reflect current observation and understanding. The next step is to review the development of wound infection criteria in light of current knowledge. In 1994, Cutting and Harding’s understanding of infection did not include critical colonization, a key stage in the infection continuum.24-26 Critical colonization has traditionally been perceived as a theoretical situation, but it is now gaining scientific acceptance. The wound infection continuum26,27 is a construct that illustrates the possible transition of a wound from sterility to infection. Research-based publications (eg, Fumal et al,28 Wall et al,29 and Stephens et al30) and a review31 have lent support to critical colonization, although a consensus definition of the clinical manifestation has yet to be generated.
Current thinking indicates that, at the point of critical colonization, the wound has become indolent.32 At this stage, the wound is not demonstrating any visible clues to this compromised healing situation. But is this correct? Is the wound on the brink of infection and only demonstrating this with indolence or recalcitrance, or, are clues present that we have not yet recognized? Critical colonization implies the inability to maintain a balance between the increasing numbers of bacteria and an effective immune system. Perhaps this circumstance is influencing the situation in ways that we do not yet understand. There is a pressing need to investigate, develop, and subsequently validate clinical signs of infection that will assist in identifying not only infection but also incorporate critical colonization. It should be remembered that the identification of infection is dependent on the criteria used, a notion which is supported by Wilson et al.33 Detection is clearly the impetus for treatment and therefore will have a major impact on morbidity.
If the wound progresses to the final stage in the infection continuum, the clinical signs become more obvious although some may remain subtle in nature.
The above collation of criteria according to wound type has been prepared for a number of reasons: 1) raising awareness that criteria for infection may vary with wound type, 2) applying these criteria provides the potential for early detection of wound infection, and 3) encouraging reflection on individual use of clinical criteria, stimulating debate, and ultimately generating consensus. It is anticipated that familiarity with these criteria, combined with a systematic approach to wound management,34 will enable the accurate assessment and appropriate treatment of wounds.
Although the 1994 wound infection criteria remain a useful tool in the identification of wound infection, they currently offer a “catch all” approach. It is considered that the development of criteria specific to six different wound types is essential if improvements or advancements are to be made in the identification of wound infection, in order to decrease the time spent between the onset of infection and its effective treatment, and to decrease patient morbidity. Whether the early detection of critical colonization is feasible remains to be seen.
The development of wound infection criteria has already taken a step forward. Criteria to identify infection in different wound types are currently being developed through the process of collating and refining opinions through a Delphi approach.35 A Delphi panel is a group of experts who respond to statements or questions on a particular topic. The responses are collated and returned for rating by the panel members. This process continues until a consensus is reached. The Delphi approach differs from other consensus groups by relying on the fact that panel members are not aware of the identities of their co-members. Statements and responses are sent and returned via mail or email so that the panel members never get to meet during the development of the consensus statement. This process allows each member to freely voice his/her opinions and avoids dominance by any one person.
In this instance, the panel consists of 54 international, multidisciplinary experts who are all involved in specialist areas of wound management. The Delphi panel has recently completed their work on criteria for arterial ulcers, diabetic foot ulcers, venous leg ulcers, pressure ulcers, acute/surgical wounds, and partial-thickness/full-thickness burn wounds. The data generated by the individual Delphi panels has been fed back to the members and will be published in due course. Following this development, the need will arise to validate the data based on expert opinion. Ideally this would involve comparing the clinical “picture” with microbial assay and treatment outcome of a large number of wounds. - OWM