Thermal Images Enhance Pressure Ulcer Risk Detection

Increased temperature is a classic sign of infection, but it often heralds skin breakdown due to other tissue injury as well. The late Dr. Paul Brand first observed that skin warmth induced by local inflammation preceded ulceration of insensate skin in individuals with leprosy or diabetes.¹   Nearly 40 years ago, Bergtholdt and Brand² used infrared thermography to identify insensate skin areas likely to break down, as potential opportunities to prevent further deterioration with appropriate offloading and care. In a physician-blinded study of subjects with prior diabetic neuropathic foot ulcers (DNFU), Lavery et al³ confirmed that patients using infrared skin thermometers to monitor the temperature of 6 standardized foot skin sites at high-risk of developing a new DNFU could reduce recurrence rates from more than 29% to 8%. Subjects were instructed to rest and contact the study nurse if any of the 6 high-risk sites monitored on 1 foot remained > 2.2°C warmer than the corresponding site on the other foot for 2 consecutive days. Those using the skin thermometers remained DNFU-free for more than 7 weeks longer than controls. If monitoring skin temperature is so powerful for subjects at risk for a DNFU, what might it do for patients at risk of developing a pressure ulcer? The 2 studies^4,5 reviewed here are hot on the trail to discovering the answer.

Reference: Judy D, Brooks B, Fennie K, Lyder C, Burton C. Improving the detection of pressure ulcers using the TMI ImageMed system. Adv Skin Wound Care. 2011;24(1):18-24.

Rationale: Monitoring pressure ulcer (PU) risk at the patient level helps focus professionals on reducing the incidence and severity of PUs, but monitoring the skin for hot spots over bony prominences may be even more useful.

Objective: A prospective repeated measures study evaluated predictive capacity of an infrared thermography device integrated with software to predict PU development in a cohort of patients admitted to Duke University Medical Center.

Methods: With appropriate approval and patient consent, 100 consecutive qualifying PU-free patients at high risk of developing a PU participated in the study over a period of 1.5 years. Unit nurses and 2 research nurses, each with more than 5 years experience and 95% interrater and intrarater Braden Scale reliability, administered the Braden Scale on admission and daily until the patient was discharged from the hospital. Consistent with unit policy, subjects with a Braden score < 16 were assigned “high Braden risk” status and all others assigned “low Braden risk” status. All Braden scores were derived blinded to same-day thermography results obtained using an infrared camera with crosshairs centered over the bony prominences of each patient’s heels and sacrum using a target area of 3 in x 3 in. Before thermography, each participant remained in a standardized lateral position, offloading heels and sacrum for 3 minutes to reduce reactive hyperemia artifacts. Thermography system software displayed mean temperatures within 64 sectors of a matrix of 8 rows and 8 columns to allow assessment of skin temperature differences within the 9 square-inch target area. Subjects with a 1.5°C or greater temperature difference within the target area were defined as “high PU risk.” All others were termed “low PU risk.” Patient demographics, Braden scores, thermography results, and PU development were tabulated as means and standard deviations or as frequencies for categorical variables. Odds ratios of PU prediction models were calculated from the “high” and “low” risk Braden and thermographic categories using generalized estimating equations with mixed models for non-normal data.

Results: Of 399 consecutive patients screened, most were reportedly “too sick” to participate. During a mean 3.9 day length of stay,100 qualifying consenting participants developed 3 stage 1 PUs and 2 stage II PUs as categorized by the National Pressure Ulcer Advisory Panel (NPUAP).⁶ Thermography correctly classified as “high risk” all 5 anatomical sites that eventually developed a PU, using a temperature difference in the target area calculated as (mean - minimum temperature) > 1.5°C. Two participants rated “low risk” on the Braden Scale developed a PU. Unit nurses consistently assigned Braden scores about 1 unit higher (ie, at lower risk) than research nurses, whose risk ratings more closely approached those of thermography. All participants identified as “low risk” by thermographic imaging were also rated “low risk” by both unit and research nurses.

Authors’ Conclusions: Thermal imaging identified sites at risk of developing a PU that were not identified by the Braden scores. Though risk assessment techniques should never replace clinical judgment, thermography may provide valuable information about specific sites at PU risk.

Reference: Farid KJ, Winkelman C, Rizkala A, Jones K. Using temperature of pressure-related intact discolored areas of skin to detect deep tissue injury: an observational, retrospective, correlational study. Ostomy Wound Manage. 2012;58(8):20-31.

Rationale: Many clinicians mistakenly believe that all blanching erythema (not a currently accepted pressure ulcer stage) and nonblanching erythema (NPUAP stage I PU) will resolve with pressure relief. An accurate assessment is needed to discriminate between these visible skin changes and deep tissue injury (DTI) which progresses to a stage III or stage IV PU.⁶ Research has shown that temperature of unbroken skin may help identify underlying nonviable tissue better than the human eye and earlier than the 7-14 days when the skin appears necrotic.

Objective: A retrospective observational correlational study explored the hypothesis that a patient’s pressure-related intact discolored skin (PRIDS) temperature relative to that of surrounding skin can predict subsequent PRIDS necrosis. Additionally, the study explored how this predictive validity is affected by patient variables such as age, gender, and skin color.

Methods: With appropriate ethics board approval, de-identified skin-integrity consult routine records were abstracted for PRIDS capillary refill status (obtained per the NPUAP method, by blanching to fingertip pressure); size, site, and color; and skin infrared thermography of the PRIDS center and normal adjacent skin 5-10 cm² from the PRIDS margin. All 85 qualifying patients were hospitalized for at least 6 days with a PRIDS of at least a 4 cm² area in the medical/surgical, ventilator, and critical care units of the Staten Island University Hospital, Staten Island, NY, between August 2009 and February 2011. Pressure-related intact discolored skin on scar tissue or on extremities of subjects with diabetes or with peripheral vascular disease were excluded. Skin temperature was measured at maximum sensitivity in the range 26.8-37.8°C for the largest PRIDS on each subject, compared to surrounding skin temperature, and correlated with patient Braden Risk core, serum albumin, body mass index, skin color, gender, and age. Findings were correlated to DTI-related outcomes (ie, the subsequent 7-day to 14-day PRIDS color, demarcation, and necrosis). Predictive validity of capillary refill status and PRIDS warmth vs coolness relative to surrounding skin were tested using Chi square statistics. Logistic regression tested effects of PRIDS temperature on progression to necrosis, with P < 0.05 considered statistically significant.

Results: Patients with warm or cool PRIDS centers were similar in pressure ulcer risk factors at baseline, but more cool PRIDS than warm PRIDS subjects were on ventilators or admitted with a diagnosis of infection. Pressure-related intact discolored skin temperature were a mean of 1.20˚C warmer than surrounding skin in 30 patients and 1.20˚C cooler than surrounding skin in 55 patients (P < 0.001 for each difference).

On admission, 29 (53%) of the cool-center PRIDS progressed to necrosis, most within 8 days, compared to 1 (3%) of the warm-center PRIDS (P < 0.0001). Most of the cool PRIDS progressed to typical signs of DTI by 7-14 days. Capillary refill in 3 seconds (ie, blanchable erythema) occurred in 28 (93%) of warm-center PRIDS, and 20 (36.4%) of cool-center PRIDS (P < 0.001). Capillary refill in < 3 seconds accurately predicted absence of necrosis in 45 of the 55 (78%) PRIDS showing blanchable erythema. A warm PRIDS center had stronger negative predictive validity of 97% (ie, 29 of 30 warm-center PRIDS healed without necrosis). Positive predictive validity of a cool-center PRIDS on admission was only 52%, with 29 of 55 PRIDS becoming necrotic, though many proceeded to show signs of DTI. Darker skin was 7.7 times more likely to progress to necrosis, approaching statistical significance (P = 0.07). The odds ratio of a cool-center PRIDS progressing to necrosis was 32.4 (P < 0.01), unaffected by skin color, age, gender, or anatomical site.

Authors’ Conclusions: This study demonstrated the value of PRIDS surface thermography in identifying previously undetected DTI. Large multicenter validation studies are needed to explore the relationships of stage I pressure ulcers and DTI.

Judy et al⁴ reported good diagnostic validity for computer-assisted thermography: 100% sensitivity, identifying all 5 incipient PUs, with 78% specificity (ie, temperature readings correctly identifying “low risk” sites that did not develop a PU). Thermography had a 17% false positive rate identifying as “high risk” sites that remained free of PUs, including 2 PUs not predicted by Braden Scale ratings, suggesting the value of adding thermography to the PU prevention arsenal.

  Farid et al⁵ reported that warm skin can serve as an early alert to inflammation in time to prevent further skin injury. Alerted by the skin integrity consult, 97% of skin hot spots in their study healed without necrosis. This suggests the merit of a PU randomized controlled trial (RCT) similar to the DNFU RCT conducted by Lavery et al.³ Would providing thermal feedback to care-givers and patients about warmth of their bony prominences at PU risk help prevent PU occurrence or recurrence? Farid et al⁵ also suggest the value of monitoring blanching erythema, another recognized sign of inflammation, validated in 1982 as a histopathologic precursor to PU development⁷ and recently recommended as an indicator of a pre-stage I PU.⁸ For those without access to thermography, feeling warmth of the bony prominence with a gloved hand or pressing gently and waiting 3 seconds to assess blanching erythema may assess these inflammation signs to warn of PU development.⁵ Research has been telling us for more than 30 years that these signs of inflammation warn of incipient tissue breakdown if we don’t act quickly to stop the cause of the observed inflammation before the inflamed skin becomes an open wound.

  Just as with the Braden Scale, the act of monitoring PU risk may focus resources on preventing a PU, thus reducing the likelihood of PU development. Identifying a hot spot alerts the caregiver or clinician to provide it with special preventive care and, if addressed in time, it may not develop at all. This may deflate screening and diagnostic validity statistics, but, as with the findings of Lavery et al,³ it can do wonders for clinical outcomes. The real message worthy of further research is that hot spots indicate incipient pressure,^4,5 diabetic^1-3 or venous⁹ ulcers, before they appear on the clinical radar screen. These results are not a criticism of current risk assessment or staging systems. Rather, they are an invitation to identify incipient ulcers earlier in their development and avert them before they add to the patient, clinical, and economic burdens of chronic ulcer management.

Laura Bolton, PhD
Adjunct Associate Professor
Department of Surgery,
Rutgers Robert Wood Johnson Medical School,
New Brunswick, NJ

This article was not subject to the WOUNDS peer-review process.